article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
in this paper , we study the almost gorenstein property of stanley reisner rings .
let @xmath6 be a field , @xmath7 a graded @xmath6-algebra of krull dimension @xmath8 , and let @xmath9 be its canonical module .
the algebra @xmath10 is said to be _ almost gorenstein _
if @xmath10 is cohen macaulay and there is a short exact sequence of degree @xmath11 @xmath12 such that the multiplicity of @xmath13 is equal to the number of minimal generators of @xmath13 , where @xmath14 is the _
@xmath0-invariant _ of @xmath10 and where @xmath15 denotes the graded module @xmath16 with grading shifted by degree @xmath17 .
note that if @xmath10 is gorenstein , then it is almost gorenstein and @xmath13 is the zero module .
the almost gorenstein property was first introduced by barucci and frberg @xcite for @xmath18-dimensional local rings which are analytically unramified .
then the definition was extended to all @xmath18-dimensional local rings by goto
phuong @xcite and to all local and graded rings by goto
taniguchi @xcite .
since the almost gorenstein property for graded rings was introduced very recently , not many are known on this property and it is expected to understand what kinds of rings are almost gorenstein and to find various examples of almost gorenstein rings . in this paper , we address this problem for stanley reisner rings .
we first recall basics on simplicial complexes and stanley reisner rings .
let @xmath2 be a simplicial complex on @xmath19=\{1,2,\dots , n\}$ ] .
thus @xmath2 is a collection of subsets of @xmath19 $ ] such that , for any @xmath20 and @xmath21 , one has @xmath22 .
elements of @xmath2 are called _ faces _ of @xmath2 and the maximal faces ( under inclusion ) are called _ facets _ of @xmath2 .
the _ dimension _ of @xmath2 is the integer @xmath23 , where @xmath24 denotes the cardinality of a finite set @xmath25 . for every simplicial complex @xmath2 on @xmath19 $ ]
, we write @xmath26 for its natural geometric realization which is the union of convex hulls of @xmath27 for all @xmath20 , where @xmath28 is the @xmath29th unit vector of @xmath30 .
next , we recall stanley reisner rings .
let @xmath6 be an infinite field and @xmath31 $ ] the polynomial ring with each @xmath32 .
for a simplicial complex @xmath2 on @xmath19 $ ] , its _
stanley reisner ring _
( over @xmath6 ) is the quotient ring @xmath33=s/(x_f : f \subset [ n],\ f \not \in \delta),\ ] ] where @xmath34 .
we say that a simplicial complex @xmath2 is _ cohen macaulay _ ( cm for short ) over @xmath6 if @xmath35 $ ] is a cohen
macaulay ring @xcite .
the cm property depends on the base field @xmath6 , but we will omit information on @xmath6 when it is clear . to study the almost gorenstein property , it is convenient to know which graded @xmath6-algebra @xmath10 admits the short exact sequence , that is , when there is an injection @xmath36 . if @xmath10 is a domain , then it is clear that such an injection exists . however , if @xmath10 is not a domain , then the existence of such an injection is non - trivial .
the first main result of this paper is a simple combinatorial criterion of simplicial complexes whose stanley reisner rings admit such an injection with @xmath37 . for a simplicial complex @xmath2 on @xmath19 $ ] and a face @xmath20 , the simplicial complex @xmath38
is called the _ contrastar _ of @xmath3 in @xmath2 .
we say that a @xmath1-dimensional simplicial complex @xmath2 is _ uniformly cohen macaulay _ if @xmath2 is cm and , for any facet @xmath20 , @xmath39 is cm of dimension @xmath1 .
( we note that if @xmath2 is cm and @xmath3 is its facet , then @xmath40 has dimension @xmath1 unless @xmath2 is a simplex . ) for topological spaces ( or simplicial complexes ) @xmath41 , we write @xmath42 and @xmath43 for the @xmath17th homology group of @xmath25 and of the pair @xmath44 with coefficients in @xmath6 , respectively . [ main1 ] let @xmath2 be a cm simplicial complex on @xmath19 $ ] of dimension @xmath45 .
the following conditions are equivalent .
* @xmath2 is uniformly cm .
* there is an injective @xmath46-homomorphism @xmath47 \to \omega_{k[\delta]}$ ] of degree @xmath11 .
* for any facet @xmath20 , @xmath48 .
* for any point @xmath49 , the map @xmath50 induced by the inclusion @xmath51 is non - zero .
the uniformly cm property is a new property of simplicial complexes and is itself an interesting property .
since the cm property is a topological property , the fourth condition of the theorem shows that the uniformly cm property is a topological property , that is , only depends on the geometric realization and a base field @xmath6 .
the second condition implies an interesting enumerative consequence .
it implies that if @xmath52 is the @xmath53-vector of the uniformly cm simplicial complex , then we have @xmath54 for all @xmath17 .
this is an analogue of stanley s inequality for ehrhart @xmath55-vectors @xcite .
also , the third condition gives a simple way to check the uniformly cm property .
the definition of the uniformly cm property is similar to that of the @xmath56-cm property introduced by baclawski @xcite .
a @xmath1-dimensional simplicial complex @xmath2 is said to be @xmath56-cm if , for every vertex @xmath57 of @xmath2 , @xmath58 is cm of dimension @xmath1 . in particular , for @xmath18-dimensional simplicial complexes , the 2-cm and uniformly cm property are the 2-vertex and 2-edge connectivity of graphs . as we see later ,
it is not hard to see that @xmath56-cm simplicial complexes are uniformly cm ( proposition [ 3.4 ] ) .
these fundamental properties of uniformly cm simplicial complexes will be discussed in the first part of the paper . in the latter half of the paper , we study the almost gorenstein property of stanley reisner rings .
we are interested in the following special class of almost gorenstein stanley reisner rings .
we say that a simplicial complex @xmath2 is _ almost gorenstein * _ ( over @xmath6 ) if @xmath35 $ ] is almost gorenstein and @xmath59 , where @xmath60 is the @xmath17th reduced homology group of @xmath2 with coefficients in @xmath6 .
these are natural generalizations of gorenstein * simplicial complexes which are simplicial complexes @xmath2 such that @xmath35 $ ] is gorenstein and @xmath59 .
since this non - vanishing condition on the homology implies that the @xmath0-invariant of @xmath35 $ ] is zero by hochster s formula ( * ? ? ?
* ii theorem 4.1 ) , theorem [ main1 ] and the existence of a short exact sequence say that almost gorenstein * simplicial complexes are uniformly cm .
we say that an almost gorenstein * simplicial complex @xmath2 is _ indecomposable _ if @xmath61 .
our main result is that every almost gorenstein * simplicial complex can be decomposed into indecomposable ones in the following sense .
let @xmath62 and @xmath63 be @xmath64-dimensional simplicial complexes .
we say that @xmath2 is the _ ridge sum _ of @xmath65 and @xmath63 if @xmath66 and @xmath67 is a @xmath68-dimensional simplex . as we prove later , if @xmath2 is the ridge sum of @xmath65 and @xmath63 , then @xmath2 is almost gorenstein * if and only if @xmath65 and @xmath63 are almost gorenstein * ( lemma [ 4.5 ] ) .
[ main2 ] let @xmath2 be an almost gorenstein * simplicial complex of dimension @xmath45 . if @xmath2 is not indecomposable , then there are @xmath64-dimensional almost gorenstein * simplicial complexes @xmath65 and @xmath63 such that @xmath2 is a ridge sum of @xmath65 and @xmath63 .
the above theorem says that every almost gorenstein * simplicial complex is a ridge sum of ( many ) indecomposable ones .
thus , to study the almost gorenstein * simplicial complexes , it is essential to consider indecomposable ones
. this fact gives some insight on almost gorenstein * simplicial complexes of dimension @xmath5 .
it is not hard to see that @xmath18-dimensional indecomposable almost gorenstein * simplicial complexes are gorenstein * ( in particular , they are cycles ) . by using this ,
we classify almost gorenstein stanley
reisner rings of krull dimension @xmath56 ( proposition [ 4.7 ] ) .
we also show that , in two dimensional case , the indecomposable almost gorenstein * simplicial complexes are exactly the uniformly cm simplicial complexes @xmath2 with @xmath69 ( proposition [ 4.8 ] ) .
this provides a combinatorial criterion of the @xmath56-dimensional almost gorenstein * simplicial complexes by theorems [ main1 ] and [ main2 ] .
in this section , we study the uniformly cm property .
we first introduce necessary notations . throughout the paper ,
@xmath31 $ ] is the polynomial ring over an infinite field @xmath6 .
let @xmath70 be the @xmath29th unit vector of @xmath71 .
we consider the @xmath71-grading of @xmath46 defined by @xmath72 . for a @xmath71-graded @xmath46-module @xmath16 and for @xmath73 , let @xmath74 be the graded component of @xmath16 of degree @xmath75 and let @xmath76 : a_i \ne 0\}$ ] . for a subset
@xmath77 $ ] , we write @xmath78 . the key idea to prove theorem [ main1 ] is grbe s structure theorem for canonical modules of stanley reisner rings .
we quickly recall grbe s result .
fix a simplicial complex @xmath2 of dimension @xmath45 . for faces
let @xmath80 be the map appearing in the long exact sequence of the triple @xmath81 . grbe ( * ? ? ?
* theorem 4 ) proved the isomorphism @xmath82 } \cong \bigoplus_{{\mathbf{a}}\in { \ensuremath{\mathbb{z}}}^n_{\geq 0},\ { \mathrm{supp}}({\mathbf{a } } ) \in \delta } h_{d-1}\big(\delta,{\mathrm{cost}}_\delta\big({\mathrm{supp}}({\mathbf{a } } ) \big ) \big),\end{aligned}\ ] ] where the @xmath75th graded component of the right - hand side is @xmath83 and where we consider that @xmath84 if @xmath85 , and determined the multiplication structure of the right - hand side as follows . [ 2.1 ] with the same notation as above , the multiplication of @xmath86 from the @xmath75th component of the right - hand side of to its @xmath87th component is the following map @xmath88 in particular , lemma [ 2.1 ] says that the multiplication @xmath89})_0 \to ( { \omega}_{k[\delta]})_{{\mathbf{e}}_f}$ ] can be identified with the map @xmath90 which appears in the long exact sequence of the pair @xmath91 . by using this fact ,
we prove the following statement which proves the equivalence of ( ii ) , ( iii ) and ( iv ) in theorem [ main1 ] . [ 2.2 ] let @xmath2 be a simplicial complex on @xmath19 $ ] of dimension @xmath8 .
the following conditions are equivalent .
* @xmath92 is non - zero for any face @xmath20 .
* there is an injective @xmath46-homomorphism @xmath93 \to { \omega}_{k[\delta]}$ ] of degree @xmath11 .
* @xmath94 for any face ( equivalently , facet ) @xmath20 . * for any point @xmath49 , the natural map @xmath95 induced by the inclusion @xmath96 is non - zero . since both @xmath35 $ ] and @xmath97}$ ] are squarefree @xmath46-modules ( c.f .
@xcite ) , a degree preserving injection @xmath93 \to \omega_{k[\delta]}$ ] exists if and only if there is an element @xmath98}$ ] of degree @xmath11 such that @xmath99 for any @xmath20 . since @xmath6 is infinite , the latter condition is equivalent to saying that , for any face ( equivalently , facet ) @xmath20 , the multiplication map @xmath100})_0 \to ( { \omega}_{k[\delta]})_{{\mathbf{e}}_f}\end{aligned}\ ] ] is non - zero .
then lemma [ 2.1 ] proves the equivalence of ( i ) and ( ii ) . also , the long exact sequence of the pair @xmath91 @xmath101 guarantees the equivalence of ( i ) and ( iii ) .
the equivalence of ( i ) and ( iv ) is standard in topology .
observe that , for any point @xmath49 , if @xmath20 is the smallest face ( w.r.t .
inclusion ) such that @xmath102 lies in the convex hull of @xmath103 , then @xmath102 line in its relative interior ( note that the relative interior of a point is the point itself ) and @xmath104 ( see ( * ? ? ?
* lemma 1.3 ) and ( * ? ? ?
* lemma 3.3 ) ) .
consider the commutative diagram @xmath105 induced by the inclusion .
then @xmath106 is identity and @xmath107 is an isomorphism .
this proves the equivalence of ( i ) and ( iv ) . to prove theorem [ main1 ] , we need one more technical lemma .
let @xmath2 be a simplicial complex .
for a face @xmath20 , the simplicial complex @xmath108 is called the _ link _ of @xmath3 in @xmath2 . we write @xmath109 for simplicity . a simplicial complex is said to be _ pure _ if all its facets have the same cardinality . [ 2.3 ] let @xmath2 be a pure simplicial complex of dimension @xmath110 and @xmath57 a vertex of @xmath2 .
if @xmath2 satisfies the condition ( i ) in lemma [ 2.2 ] then so does @xmath111 .
let @xmath112 and let @xmath3 be a face of @xmath111 . by the purity of @xmath2 , @xmath113 .
thus , what we must prove is that the map @xmath114 is non - zero . by the natural isomorphisms @xmath115 and
@xmath116 the map is identified with the map @xmath117 in the long exact sequence of the triple @xmath118 .
( indeed , the natural isomorphisms send an element corresponding to @xmath119 to that of @xmath120 in the homology , so they commute with @xmath121 . )
this map is non - zero since its composition with the map @xmath122 is the map @xmath123 which is non - zero by the assumption .
we also recall reisner s criterion of cm simplicial complexes @xcite .
a @xmath64-dimensional simplicial complex @xmath2 is cm if and only if , for any face @xmath20 ( including the empty face @xmath124 ) , @xmath125 for all @xmath126 . we now prove theorem [ main1 ] . for subsets @xmath127 of @xmath19 $ ] ,
let @xmath128 : g \subset f_k \mbox { for some } k \in \{1,2,\dots , r\}\big\}\ ] ] be the simplicial complex generated by @xmath127 .
also , for @xmath77 $ ] , let @xmath129 the equivalence ( ii ) @xmath130 ( iii ) @xmath130 ( iv ) follows from lemma [ 2.2 ] .
we prove ( i ) @xmath130 ( iii ) .
observe that @xmath131 for all @xmath132 since @xmath2 is cm .
let @xmath3 be a facet of @xmath2 .
then we have the mayer
vietoris long exact sequence @xmath133 note that @xmath134 for all @xmath17 .
we first prove ( i ) @xmath135 ( iii ) .
suppose that @xmath40 is cm of dimension @xmath136 . then @xmath137 .
since @xmath138 , the exact sequence implies @xmath139 . we prove ( iii ) @xmath135 ( i ) by induction on @xmath1 .
when @xmath140 , the implication easily follows from the fact that a @xmath18-dimensional simplicial complex is cm if and only if it is connected .
suppose @xmath141 .
we prove that @xmath40 is cm of dimension @xmath136 .
since @xmath2 has the non - zero top homology by the assumption , @xmath2 is not a simplex and @xmath142 . since @xmath143 , @xmath144 and @xmath138 , the map @xmath106 in is surjective .
hence @xmath137 .
the exact sequence also shows @xmath145 for @xmath146 since @xmath147 for @xmath148
. then reisner s criterion says that , to prove that @xmath40 is cm , it is enough to prove that @xmath149 is cm for any vertex @xmath57 of @xmath40 .
a routine computation implies @xmath150 by reisner s criterion , @xmath111 is cm .
also , @xmath111 satisfies condition ( iii ) by lemmas [ 2.2 ] and [ 2.3 ] .
then @xmath111 is uniformly cm by the induction hypothesis and proves that @xmath149 is cm , as desired .
a link of a uniformly cm simplicial complex is uniformly cm by .
there are non - cm simplicial complexes satisfying the equivalent conditions ( ii)(iv ) of theorem [ main1 ] .
for example , buchsbaum * complexes introduced by athanasiadis and welker @xcite satisfy the conditions .
see ( * ? ? ?
* proposition 2.3 ) . in the rest of this section
, we discuss a few easy properties of uniformly cm simplicial complexes .
first , we consider an enumerative property . for a @xmath64-dimensional simplicial complex @xmath2 , its _
@xmath151-vector _ is the sequence @xmath152 defined by @xmath153 , where @xmath154 , and its _ @xmath53-vector _ @xmath155 is defined by the relation @xmath156 it is well - known that the @xmath53-vector of @xmath2 and the hilbert series of @xmath35 $ ] are related by @xmath157_k ) t^k = \frac 1 { ( 1-t)^d } \big(h_0(\delta)+h_1(\delta)t+ \cdots + h_d(\delta)t^d \big).\ ] ] let @xmath2 be a @xmath64-dimensional uniformly cm simplicial complex with the @xmath53-vector @xmath158 . by theorem [ main1 ] , there exists an injection @xmath47 \to \omega_{k[\delta]}$ ] .
then the module @xmath159 is cm of krull dimension @xmath136 ( * ? ? ?
* lemma 3.1 ) unless @xmath160 .
also , since the hilbert series of @xmath97}$ ] is given by @xmath161 the short exact sequence @xmath162 \to \omega_{k[\delta ] } \to c \to 0 $ ] says that the hilbert series of @xmath13 is given by @xmath163 since the @xmath53-vector of a cm @xmath46-module is non - negative , these facts imply the following statement , which is an analogue of stanley s inequality for ehrhart @xmath55-vectors ( * ? ? ?
* proposition 4.1 ) .
[ 2.6 ] if @xmath2 is a @xmath64-dimensional uniformly cm simplicial complex with the @xmath53-vector @xmath158 , then @xmath164 for all @xmath165 .
next , we discuss 2-cm simplicial complexes .
a @xmath1-dimensional simplicial complex @xmath2 is said to be _ doubly cohen macaulay _
( @xmath56-cm for short ) if it is cm and , for any vertex @xmath57 of @xmath2 , @xmath58 is cm of dimension @xmath1 . note that it is known that a cm simplicial complex @xmath2 is @xmath56-cm if and only if @xmath35 $ ] is level and has @xmath0-invariant zero @xcite .
it is not hard to see that 2-cm simplicial complexes are uniformly cm ( see e.g. the proof of ( * ? ? ?
* iii proposition 3.3 ) ) .
the following result gives a more precise relation between these properties .
[ 3.4 ] if a simplicial complex @xmath2 is @xmath56-cm , then @xmath40 is cm for any face @xmath20 . to prove the above proposition ,
we need the following well - known fact ( see , e.g. , the proof of ( * ? ? ?
* theorem 5.1.13 ) ) .
[ 3.1 ] let @xmath2 and @xmath65 be @xmath1-dimensional cm simplicial complexes . if @xmath166 is cm of dimension @xmath136 , then @xmath167 is cm . for a simplicial complex @xmath2 and a new vertex @xmath57 which is not in @xmath2 , let @xmath168 be the _ cone _ over @xmath2 with apex @xmath57 .
note that if @xmath2 is cm then so is @xmath169 .
we use induction on @xmath170 .
the statement is obvious when @xmath171 .
suppose @xmath172 .
let @xmath20 be a face with @xmath173 .
we prove that @xmath40 is cm .
let @xmath174 .
a routine computation says @xmath175 since a link of a @xmath56-cm simplicial complex is 2-cm ( * ? ? ?
* proposition 9.7 ) , by the induction hypothesis , @xmath176 is cm of dimension @xmath177 .
then , since @xmath58 is cm of dimension @xmath136 and since @xmath178 it follows from lemma [ 3.1 ] that @xmath40 is cm as desired . the proof of proposition 2.8 can be simplified by using stellar subdivisions ( see ( * ? ? ?
* definition 2.22 ) ) .
indeed , if @xmath2 is @xmath56-cm and @xmath3 is its face , then the stellar subdivision @xmath179 of @xmath2 with respect to @xmath3 is @xmath56-cm since @xmath56-cm property is a topological property .
also , we have @xmath180 , where @xmath181 is the new vertex that is added when we take a stellar subdivision , so @xmath182 is cm .
in this section , we study almost gorenstein * simplicial complexes .
recall that a simplicial complex @xmath2 is almost gorenstein * if @xmath35 $ ] is almost gorenstein and @xmath183 is non - zero .
in other words , @xmath2 is almost gorenstein * if it is cm and there is an injection @xmath184 \to \omega_{k[\delta]}$ ] of degree @xmath11 such that the multiplicity of @xmath185 is equal to the number of its minimal generators .
let @xmath2 be a @xmath64-dimensional cm simplicial complex on @xmath19 $ ] .
we write @xmath186 for the _ type _ of @xmath35 $ ] , that is , @xmath187,k).\ ] ] also , we define @xmath188 suppose that @xmath2 is uniformly cm
. then we have a short exact sequence @xmath189 \stackrel \phi \longrightarrow \omega_{k[\delta ] } \longrightarrow c={\ensuremath{\mathrm{coker}}\hspace{1pt}}\phi \longrightarrow 0.\end{aligned}\ ] ] by , @xmath190 is the multiplicity of @xmath13 . since @xmath186 is equal to the number of minimal generators of @xmath97}$ ] , the number of minimal generators of @xmath13 is @xmath191 .
then , since the number of minimal generators of a cm @xmath46-module is smaller than or equal to its multiplicity , we have the following property . [ 4.3 ] let @xmath2 be a @xmath64-dimensional almost gorenstein * simplicial complex on @xmath19 $ ] with the @xmath53-vector @xmath158 .
then @xmath194,k)_{n - k}= ( h_d+ \cdots + h_{d - k})-(h_0+\cdots+h_k)\ ] ] for all @xmath195 . with the same notation as in
, we have @xmath196,k)_{n - k } \cong { \ensuremath{\mathrm{tor}}\hspace{1pt}}_0^s ( \omega_{k[\delta]},k)_k \cong { \ensuremath{\mathrm{tor}}\hspace{1pt}}_0^s(c , k)_k\ ] ] for all @xmath195 ( see ( * ? ? ? * i 12 ) for the first isomorphism ) .
let @xmath197 be a linear system of parameters of @xmath13 . since @xmath13 is a cm @xmath46-module whose multiplicity is equal to the number of its minimal generators
, we have @xmath198 . then implies @xmath194,k)_{n - k}=\dim_k ( c/\theta c)_k=(h_d+ \cdots + h_{d - k})-(h_0+\cdots+h_k)\ ] ] for all @xmath195 , as desired . in this subsection , we prove theorem [ main2 ] . for a simplicial complex @xmath2 on @xmath19
$ ] and a subset @xmath199 $ ] , let @xmath200 be the _ restriction _ of @xmath2 to @xmath201 .
the hochster s formula for @xmath202-algebras ( * ? ? ?
* theorem 5.5.1 ) gives a way to compute @xmath186 from restrictions of @xmath2 .
[ 4.4 ] let @xmath2 be a simplicial complex on @xmath19 $ ] .
then , for any non - negative integer @xmath203 , @xmath204,k)=\bigoplus_{f \subset [ n ] } ( { \ensuremath{\mathrm{tor}}\hspace{1pt}}_i^s(k[\delta],k))_{{\mathbf{e}}_f}$ ] , and for any @xmath77 $ ] one has @xmath205,k)_{{\mathbf{e}}_f } \cong \widetilde h_{\#f - i-1}(\delta|_f).\ ] ] next , we study basic properties of ridge sums . recall that a @xmath64-dimensional simplicial complex @xmath2 is the ridge sum of @xmath64-dimensional simplicial complexes @xmath65 and @xmath63 if @xmath66 and @xmath67 is a @xmath68-dimensional simplex . see figure 1 . 1 .
2 . @xmath210 .
3 . @xmath2 is cm if and only if both @xmath65 and @xmath63 are cm .
@xmath2 is uniformly cm if and only if both @xmath65 and @xmath63 are uniformly cm .
suppose that @xmath2 is cm . if neither @xmath65 nor @xmath63 is a simplex , then @xmath211 .
@xmath2 is almost gorenstein * if and only if both @xmath65 and @xmath63 are almost gorenstein*. let @xmath212 . note that @xmath213 and @xmath214 \cong s/(x_i : i\not\in w)$ ] .
then the mayer
vietoris short exact sequence @xmath215 \longrightarrow k[\gamma ] \bigoplus k[\sigma ] \longrightarrow k[\langle w \rangle ] \longrightarrow 0\end{aligned}\ ] ] says @xmath216 the above equation implies ( i ) and ( ii )
. we prove ( iii ) . the `
if ' part is a special case of lemma [ 3.1 ] .
we prove the ` only if ' part .
recall that a graded @xmath6-algebra @xmath217 of krull dimension @xmath1 is cm if and only if @xmath218 for @xmath219 .
suppose that @xmath2 is cm .
since @xmath35 $ ] is cm of krull dimension @xmath1 and @xmath220 $ ] is cm of krull dimension @xmath136 , the short exact sequence says @xmath221,k)= { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{i}^s(k[\sigma],k)=0 $ ] for @xmath222 .
what we must prove is @xmath223,k)= { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d+1}^s(k[\sigma],k)=0 $ ] .
the short exact sequence induces the long exact sequence @xmath224,k ) \bigoplus { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d+1}^s(k[\sigma],k ) \longrightarrow { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d+1}^s(k[\langle w \rangle],k ) \\ \label{3 - 2 } & \longrightarrow { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d}^s(k[\delta],k ) \longrightarrow { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d}^s(k[\gamma],k ) \bigoplus { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d}^s(k[\sigma],k ) \\ & \nonumber \stackrel{\psi}{\longrightarrow } { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d}^s(k[\langle w \rangle],k ) \longrightarrow \cdots.\end{aligned}\ ] ] since @xmath225 \cong s/(x_i : i \not \in w)$ ] , we have @xmath226,k)={\ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d+1}^s(k[\langle
w \rangle],k)_{n - d+1}.\ ] ] then the long exact sequence says @xmath227,k)={\ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d+1}^s(k[\gamma],k)_{n - d+1}\ ] ] and @xmath228,k)={\ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d+1}^s(k[\sigma],k)_{n - d+1}.\ ] ] on the other hand , for any graded ideal @xmath229 , one has @xmath230 if and only if @xmath231 contains @xmath232 linearly independent linear forms . since @xmath65 and @xmath63 have dimension @xmath136 , if we write @xmath233=s / i_{\gamma}$ ] and @xmath234=s / i_\sigma$ ] then @xmath235 and @xmath236 contain at most @xmath237 linearly independent linear forms
. thus @xmath223,k)= { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d+1}^s(k[\sigma],k)=0 $ ] .
next , we prove ( v ) .
suppose that @xmath2 is cm
. then @xmath65 and @xmath63 are cm by ( iii ) .
we claim that the map @xmath107 in is the zero map . since @xmath225 \cong s/(x_i : i \not \in w)$ ] , @xmath242,k)$ ] has non - zero elements only in degree @xmath237 .
thus @xmath243,k ) = \bigoplus_{\substack{f \subset [ n ] \\ \ # f = n - d } } \left ( { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d}^s(k[\langle w \rangle],k ) \right)_{{\mathbf{e}}_f}.\end{aligned}\ ] ] on the other hand , hochster s formula says that , for any @xmath77 $ ] with @xmath244 , @xmath245,k)_{{\mathbf{e}}_f } \cong \widetilde h_{-1}(\gamma |_f ) \mbox { and } { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d}^s(k[\sigma],k)_{{\mathbf{e}}_f } \cong \widetilde h_{-1}(\sigma |_f).\ ] ] since @xmath65 and @xmath63 are not simplexes , @xmath246 and @xmath247 if @xmath248 .
thus we have @xmath245,k)_{{\mathbf{e}}_f } = { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d}^s(k[\sigma],k)_{{\mathbf{e}}_f}=0\ ] ] for all @xmath77 $ ] with @xmath249 .
this fact and say that @xmath107 is the zero map .
then , by , we have @xmath250,k)\\ = & \dim_k \big({\ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d}^s ( k[\gamma],k ) \bigoplus { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d}^s(k[\sigma],k ) \big)+ \dim_k { \ensuremath{\mathrm{tor}}\hspace{1pt}}_{n - d+1}^s(k[\langle w \rangle],k)\\ = & { \mathrm{type}}(\gamma)+{\mathrm{type}}(\sigma ) + 1,\end{aligned}\ ] ] as desired .
it remains to prove ( vi ) . by ( iv ) , we may assume that @xmath2 , @xmath65 and @xmath63 are uniformly cm .
note that this guarantees that @xmath65 and @xmath63 are not simplexes . by ( ii ) and
( v ) , @xmath251 and @xmath252 then the assertion follows from lemma [ 4.1 ] .
[ ridgesum ] let @xmath2 be a @xmath64-dimensional cm simplicial complex on @xmath19 $ ] .
if @xmath253,k)_{n - d+1 } \ne 0 $ ] then there are @xmath64-dimensional simplicial complexes @xmath65 and @xmath63 such that @xmath2 is the ridge sum of @xmath65 and @xmath63 .
since hochster s formula says that @xmath254,k)_{n - d+1}\cong \bigoplus_{{f \subset[n],\ \#f = n - d+1 } } \widetilde h_0(\delta|_f),\ ] ] there is a subset @xmath199 $ ] with @xmath213 such that @xmath255\setminus w}$ ] has more than two connected components .
let @xmath256 be a connected component of @xmath255\setminus w}$ ] and let @xmath257\setminus w } : f \not \in \gamma'\}$ ] be the complement of @xmath256 in @xmath255\setminus w}$ ] .
let @xmath258 be the vertex set of @xmath256 and @xmath259 the vertex set of @xmath260 .
we claim that @xmath2 is the ridge sum of @xmath261 and @xmath262 .
we first prove @xmath66 .
it is enough to prove that @xmath263 .
let @xmath3 be a facet of @xmath2 . since @xmath264 is a connected component of @xmath255 \setminus w}$ ]
, one has either @xmath265 or @xmath266 , which implies @xmath267 .
thus @xmath263 .
next , it is clear that @xmath65 and @xmath63 have dimension @xmath136 since all facets of @xmath2 have dimension @xmath136 .
it remains to prove that @xmath67 is generated by @xmath201 . to prove this
, what we must prove is that @xmath268 .
let @xmath3 and @xmath269 be facets of @xmath2 such that @xmath270 and @xmath271 .
note that such facets exist since @xmath65 and @xmath63 are not empty .
since @xmath2 is cm , @xmath2 is strongly connected ( * ? ? ?
* proposition 11.7 ) .
thus there is a sequence of facets @xmath272 such that @xmath273 for @xmath274 .
then , since @xmath65 is a simplicial complex on @xmath275 and @xmath63 is a simplicial complex on @xmath276 , if we choose an integer @xmath277 such that @xmath278 and @xmath279 , then we have @xmath280 .
hence @xmath268 .
suppose @xmath281 .
then @xmath282 .
since @xmath283 by reisner s criterion , we have @xmath284,k)_{n - d+1}= h_d-1>0\ ] ] by lemma [ 4.3 ] .
then the desired statement follows from lemmas [ 4.5](vi ) and [ ridgesum ] .
a simplicial complex @xmath2 is said to be _ gorenstein * _ if @xmath35 $ ] is gorenstein and @xmath285 .
lemma [ 4.5](vi ) says that a ridge sum of gorenstein * simplicial complexes are almost gorenstein*. in particular , this gives an infinite family of stanley reisner rings which are almost gorenstein but not gorenstein . by theorem [ main2 ] , to study almost gorenstein * simplicial complexes it is crucial to study indecomposable ones .
a typical example of an indecomposable almost gorenstein * simplicial complex is a gorenstein * simplicial complex . for @xmath18-dimensional simplicial complexes , that is graphs , it is not hard to see that
any indecomposable almost gorenstein * simplicial complex is actually gorenstein * ( i.e. , it is a cycle ) .
indeed , such a simplicial complex @xmath2 has the @xmath53-vector @xmath286 , which implies that @xmath287 and @xmath35 $ ] is gorenstein .
thus any almost gorenstein * graph is a ridge sum of cycles , that is , obtained from a cycle by taking a ridge sum with a cycle repeatedly ( see figure 2 ) .
this fact and ( * ? ? ?
* theorem 10.4 ) give the following characterization of the almost gorenstein stanley
reisner rings of krull dimension @xmath56 .
if @xmath288 is non - zero , then @xmath35 $ ] is almost gorenstein if and only if @xmath2 is a ridge sum of cycles . on the other hand
, @xmath2 is a tree if and only if @xmath2 is cm and @xmath289
. then the assertion follows from ( * ? ? ?
* theorem 10.4 ) which proved that if @xmath233 $ ] is cm and @xmath290 , then @xmath233 $ ] is almost gorenstein . in dimension
@xmath56 , there is an indecomposable almost gorenstein * simplicial complex which is not gorenstein * ( see example [ 5.6 ] ) .
thus an almost gorenstein * simplicial complex is not always a ridge sum of gorenstein * simplicial complexes . on the other hand , the next statement and theorem [ main2 ]
provide a combinatorial criterion of the @xmath56-dimensional almost gorenstein * simplicial complexes .
the only if part is obvious .
suppose that @xmath2 is uniformly cm and @xmath291 .
since @xmath292 , the @xmath53-vector of @xmath2 is the vector of the form @xmath293 .
let @xmath184 \to \omega_{k[\delta]}$ ] be an injection and @xmath159 . by , the hilbert series of @xmath13 is @xmath294 , which implies that @xmath13 has @xmath295 generators of degree @xmath18 .
since @xmath296 is the multiplicity of @xmath13 , @xmath35 $ ] is almost gorenstein .
we close this paper with a few questions and examples .
let @xmath2 be an almost gorenstein * simplicial complex with the @xmath53-vector @xmath158 . for @xmath297 , we define @xmath298 and @xmath299 thus , as we saw in , @xmath300 is the @xmath53-polynomial of the cokernel of the injection
@xmath47 \to \omega_{k[\delta]}$ ] and @xmath301 . by the definition ,
@xmath300 is symmetric , that is , @xmath302 for all @xmath17 . since the cokernel of @xmath303 looks to be of fundamental in the study of the almost gorenstein *
property , to find various types of almost gorenstein * simplicial complexes , we ask the following problem . [ 5.6 ] we give a @xmath56-dimensional almost gorenstein * simplicial complex with @xmath308 .
let @xmath309 ( see figure 3 . )
then @xmath63 is uniformly cm and @xmath310 .
thus @xmath63 is almost gorenstein * by proposition [ 4.8 ] . since @xmath311 , we have @xmath312 .
below , we give more specified questions relating to problem [ 5.1 ] . in the above example
, we have @xmath313 .
since @xmath190 is even if @xmath2 is a ridge sum of gorenstein * simplicial complexes , we suggest the following question .
note that question [ q1 ] is obvious in odd dimensions since @xmath300 is symmetric .
but the question is open even for @xmath56-dimensional simplicial complexes .
we think that question [ q2 ] for @xmath316 would not be hard , but for @xmath317 the question seems not to be easy .
indeed , we do not have an example of an almost gorenstein * simplicial complex of dimension @xmath318 with @xmath319 .
the almost gorenstein * property is not a topological property .
for example , @xmath320 is almost gorenstein * since it is a ridge sum of two boundaries of a simplex . on the other hand
, @xmath321 is a subdivision of @xmath2 ( subdivide the edge @xmath322 by a new vertex @xmath323 ) but is not almost gorenstein * since it is not a ridge sum although @xmath324 .
the link of an almost gorenstein * simplicial complex may not be almost gorenstein*. let @xmath325 then @xmath2 is a subdivision of @xmath63 in example [ 5.6 ] ( subdivide the edge @xmath326 by a new vertex @xmath323 ) , and is almost gorenstein * by proposition [ 4.8 ] .
however , @xmath327 is not almost gorenstein * by proposition [ 4.7 ] .
|
in this paper , we study simplicial complexes whose stanley reisner rings are almost gorenstein and have @xmath0-invariant zero .
we call such a simplicial complex an almost gorenstein * simplicial complex . to study the almost gorenstein * property
, we introduce a new class of simplicial complexes which we call uniformly cohen macaulay simplicial complexes .
a @xmath1-dimensional simplicial complex @xmath2 is said to be uniformly cohen
macaulay if it is cohen macaulay and , for any facet @xmath3 of @xmath2 , the simplicial complex @xmath4 is cohen macaulay of dimension @xmath1 .
we investigate fundamental algebraic , combinatorial and topological properties of these simplicial complexes , and show that almost gorenstein * simplicial complexes must be uniformly cohen macaulay . by using this fact ,
we show that every almost gorenstein * simplicial complex can be decomposed into those of having one dimensional top homology .
also , we give a combinatorial criterion of the almost gorenstein * property for simplicial complexes of dimension @xmath5 .
|
measurements of the depletion of heavy elements , relative to the solar abundances , along the line of sight to the stars show that about @xmath4 of the metals in the galaxy is bound up in dust grains ( whittet 1992 ) .
this raises the questions : is this fraction the same for every galaxy ?
what are the processes that so efficiently lock metals into dust grains ?
although the presence of dust manifests itself in a multitude of ways - extinction , reddening , the polarisation of starlight , to name but three - the best way , in principle , to estimate the mass of dust in a galaxy is from the submm flux density .
the advantages of this are that ( a ) the emission from dust is optically thin and ( b ) the emission depends mainly on the mass of dust and is only weakly dependent on the temperature of the dust . the practical obstacle to making the full use of this technique has been the difficulty of estimating the constant of proportionality between the emission and the mass of dust , the mass - extinction or the mass - absorption coefficient .
there are two methods that have been used to estimate this constant .
the most commonly used technique to calibrate this constant is by observations of galactic reflection nebulae ( hildebrand 1983 ; casey et al .
1991 ) . by making optical and far - infrared / submillimetre observations
, one can estimate @xmath5/@xmath6@xmath7 in a reasonably straightforward way . if one assumes a value of _
@xmath8 _ and a gas - to - dust ratio , one can estimate the mass - absorption coefficient , @xmath0 . for observations at long wavelengths such as with the scuba submillimetre camera ( holland et al .
1999 ) , which operates mostly at 850@xmath9 m , there is also the practical problem that this technique has only been applied at relatively short ( @xmath10 @xmath11 400@xmath9 m ) submillimetre wavelengths , which means that the coefficient has to be extrapolated to longer wavelengths . since the dust emissivity index , @xmath12
, @xmath13 is uncertain by a factor of two , there is an immediate uncertainty of a factor of two in the long - wavelength value of @xmath0 on top of all the other uncertainties .
this problem , of course , could be solved by scuba observations of reflection nebulae , but there are additional problems with this technique .
first , there is evidence that the properties of dust vary with environment ( e.g. cardelli et al .
1996 ) , and so the dust in a reflection nebula may not be representative of a galaxy as a whole .
second , the technique relies on the assumption of spherical symmetry for the reflection nebulae and on the values for properties that are not themselves simple to measure , such as the dust - to - gas ratio and _ @xmath8 _ ( although there are variations in this technique that do not require these , there are other assumptions that have to be made , such as the typical sizes and densities of dust grains ) .
hughes et al .
( 1993 ) estimate that the factor of uncertainty in @xmath14 at @xmath15 is @xmath16 .
alton ( 2000 ) employed hildebrand s ( 1983 ) technique using scuba observations of an external galaxy to derive @xmath17 , essentially using a galaxy rather than a reflection nebula to calibrate @xmath0 .
this method obviously avoids the problems of extrapolation to longer wavelengths and the fact that nebular dust may not be the same as dust in the rest of the interstellar medium but still requires assumptions to be made about the physical properties of the dust .
the other technique that has been used to estimate @xmath0 is more loosely linked to submillimetre observations . in this technique one
constructs a model for dust ( its chemical composition , distribution of grain sizes etc . ) that agrees with all the available data , for example the shape of the optical extinction curve , and then uses the model to estimate the value of the submillimetre mass - absorption coefficient ( draine & lee 1984 ; hughes et al . 1993 ) . in this paper
we suggest an alternative way of estimating the mass - absorption coefficient based on the global properties of galaxies rather than on individual properties within the galaxy .
this technique is based on the assumption that the fraction of metals within the interstellar medium of a galaxy that is bound up in dust is a constant .
there is some evidence for this from the correlation between dust - to - gas ratio and metallicity in nearby galaxies ( issa , maclaren & wolfendale 1990 ) .
the high depletion of many elements within our own ism ( whittet 1992 ) also suggests that the mechanism forming dust is efficient - that is , if there are metals present , they are efficiently incorporated into dust grains .
finally , there is evidence in the galaxy that although the dust extinction curve is quite variable from place to place , the gas phase carbon and oxygen abundances are remarkably constant over a wide range of ism density ( cardelli et al .
1996 ; meyer et al .
1998 ) , implying ( as these elements are the most important constituents of dust ) that the fraction of metals in dust is a constant .
in contrast to this evidence , lisenfeld & ferrara ( 1998 ) found a non - linear relationship between metallicity and dust - to - gas ratio for 28 dwarf irregular galaxies , which suggests that @xmath2 , the fraction of metals locked up in dust , is not the same for all types of galaxy . in this paper
, we will present new data that shows that lisenfeld & ferrara s technique missed cool dust , and when this is taken into account @xmath2 does appear to be remarkably constant from galaxy to galaxy . to apply the technique
we need a galaxy which has submillimetre and far - infrared flux measurements , a measure of its total gass mass , and a measure of its metallicity .
the dust mass of the galaxy is then given by two equations : @xmath18 where @xmath19 is the mass of gas in the galaxy , @xmath20 is the metallicity relative to solar , @xmath2 is the ratio of the mass of metals in the dust to the total mass of metals and @xmath21 is the ratio of the mass of metals to the mass of gas for gas with solar metallicity , we estimate @xmath22 , and ; @xmath23 in which @xmath24 is the flux density at 850@xmath9 m , @xmath25 is the distance of the galaxy , @xmath26 is the value of the planck function at 850@xmath9 m , t is the dust temperature and @xmath27 is the mass - absorption coefficient at 850@xmath9 m .
we then set these two equations equal to one another , and as long as we know @xmath2 , we can estimate @xmath0 .
the arrangement of this paper is as follows . in section 2
we present the results of a search through the archive of the james clerk maxwell telescope for submm data for galaxies with known metal abundances . in section 3
we derive a value for @xmath2 based on observations of the local interstellar medium .
in section 4 we apply the technique to estimate @xmath0 and investigate whether @xmath2 is a universal constant .
we assume throughout that @xmath28 .
we looked through the jcmt archive for all the objects for which there are both abundance measurements and observations with the scuba camera ( holland 1999 ) .
table 1 is a list of galaxies in the archive for which there are published metallicity measurements but no published submm results .
the data for these five objects was reduced in the usual way using the surf ( jenness & lightfoot 1998 ) software package .
images were first flat - fielded to remove inhomogeneity in bolometer sensitivity .
corrections were then made for atmospheric absorption using the opacities ( @xmath29 ) derived from skydip measurements .
noisy bolometers were masked and large spikes in the bolometer time - stream removed by applying a 4-@xmath30 clip .
residual sky noise was removed with the surf task remsky . following removal of the sky noise ,
the bolometer time - stream was clipped again at the 3-@xmath30 level .
all the datasets in the archive which were taken with the same chop throw were then made into a map using the surf program rebin .
the maps were calibrated using jiggle maps of both primary and secondary calibrators reduced in the same way .
figure 1 shows contour plots of the maps of these objects overlaid on images from the digitised sky survey ( dss ) .
the lowest level contour is @xmath31 and increasing in @xmath31 steps to @xmath32 .
although it appears that the submm emission is concentrated in the centre of ngc5253 , the dynamical range of the optical data is much greater than the submm data , thereby making it difficult to draw any firm conclusions as to whether the optical emission is indeed more extended than the dust emission .
the flux was measured by using the dss image as a guide to chose an aperture which contained as much of the galaxy and as little of the sky as possible .
the flux errors were estimated using the procedure in dunne et al .
the fluxes and the apertures used are given in table 1 .
since the first draft of this paper another submm map of ngc1569 has been published ( lisenfeld et al .
the new data was taken in much better conditions than the data we retrieved from the scuba archive and the flux measured is significantly higher than ours .
we repeated our reduction of this object but the flux determined remained @xmath33 lower than the flux measured by lisenfeld et al .
the discrepancy may arise from the galaxy being larger than the scuba array , for which lisenfeld et al ( 2001 ) have tried to correct .
@xmath34 + + @xmath35 @xmath36 +
the fraction of metals incorporated in dust in the interstellar medium can be estimated from @xmath37 spectroscopic observations of nearby stars , which show that a significant fraction of the metals that one would expect , from cosmic abundances , to be in the ism are not there - they have been bound up in dust . if @xmath38 is the number of atoms of a given chemical element relative to the number of hydrogen atoms , @xmath39 is the atomic weight of the element , and @xmath40 is the proportion of that element bound up in dust , then @xmath41 recent @xmath37 observations of nearby stars with the hubble space telescope have made it possible to calculate @xmath2 more accurately because they have provided accurate depletion measurements for two of the most important elements , oxygen and carbon , strong evidence that the depletion is the same along different sight lines , and evidence that the cosmic metal abundance is about two thirds the solar metal abundance ( cardelli et al .
1996 ; meyer et al .
1998 ) . to calculate @xmath2
, we have assumed the solar metal abundance given in pagel ( 1997 ) and then multiplied this by two thirds to obtain the cosmic metal abundances .
it should be noted that the recent re - determination of the solar oxygen abundance by allende prieto et al .
( 2001 ) , suggests that the solar metallicity is considerably closer to the ism value than has been conventionally assumed .
we have assumed the depletions given by whittet ( 1992 ) , except that we have preferred the recent measurements for carbon and oxygen obtained from the hubble space telescope observations ( cardelli et al .
1996 ; meyer et al . 1998 ) .
we obtain a value for @xmath2 of 0.456 .
table 2 lists the submm results for all the galaxies that we are aware of which have metallicity measurements and submillimetre measurements .
many of the objects in the table are taken from the scuba local universe galaxy survey ( slugs ) of dunne et al .
distances for objects with radial velocities @xmath42 were taken from the nearby galaxies catalogue ( tully 1988 ) .
table 3 gives the metallicity measurements for the sample and also measurements of the masses of the atomic and molecular gas .
metallicities for some slugs objects were estimated using strong line ratios and the calibration method of edmunds & pagel ( 1984 ) , relating @xmath43 and @xmath44/h_{\alpha}$ ] and oxygen abundances in hii regions .
line ratios were taken fron keel et al .
( 1985 ) and veilleux et al .
although more recent work has shown that the edmunds & pagel calibrator may over - estimate metallicities above solar , it is probably a reasonable estimator in the abundance range of these galaxies .
there are two complications to applying the simple method we have outlined in the introduction : ( a ) the requirement of equation ( 2 ) for a temperature model ; ( b ) the fact that estimates of gas mass depend to some extent on assumptions about metallicity .
we will now discuss these in turn .
dunne @xmath45 eales ( 2001 ) , in their investigation of the fir / submm spectral energy distributions ( seds ) of galaxies in the scuba local universe galaxy survey ( slugs ) , found that the 450/850@xmath9 m flux ratio is remarkably constant for galaxies with a wide range of luminosities .
they conclude that the dust emissivity index @xmath12 must be @xmath462 .
they then showed that with this emissivity index , there must be at least two dust components with different temperatures in galaxies , even in ultraluminous iras galaxies like arp220 .
we fitted the dunne & eales two - component model to the five objects for which we have new scuba data .
since we only have three fluxes for each galaxy ( 60 , 100 and @xmath1 m ) , we fixed the temperature of the cold component to be 20k , the average of the values found by dunne & eales ( 2001 ) .
figure 2 shows the fits for the five galaxies and table 2 gives the temperature of the warm component and the ratio of the cold - to - hot dust ( @xmath47 ) .
we have also given the results of the two component fits for the slugs galaxies by dunne & eales ( 2001 ) and values for a few other galaxies for which we either made two - component fits or for which we took two - component fits from the literature .
from the results of these fits we have estimated dust masses using equation ( 4 ) : @xmath48\ ] ] in which @xmath24 is the flux density at 850@xmath9 m , @xmath26 is the value of the planck function at 850@xmath9 m , @xmath27 is the mass - absorption coefficient at 850@xmath9 m and d is the distance found for @xmath49 .
we have initially adopted the value for @xmath50 m ) of @xmath51 used by dunne et al .
( 2000 ) , although of course one of the goals of the present paper is to measure a new value for this .
these initial estimates of the dust masses are given in table 2 .
the mass of hi can be unambiguosly calculated from the intensity of the 21-cm line .
following lisenfeld & ferrara ( 1998 ) , we have calculated the dust mass within the optical disk , since the metallicity ( and thus the dust content ) outside the optical disk is likely to be low .
cold @xmath52 is detected indirectly by means of a tracer such as co. the @xmath53 conversion factor is however highly variable depending on temperature , density and metallicity .
it is usual to determine the @xmath52 mass in a galaxy by applying a standard conversion factor , @xmath54 , where i(co ) is the velocity - integrated co intensity and @xmath55 is the molecular hydrogen column density .
a complication here is that the conversion factor almost certainly depends on metallicity .
we investigated the dependence of our method on this uncertainty by trying a number of suggested relationships between x and metallicity .
molecular hydrogen ( @xmath52 ) masses for the five objects from the scuba archive were calculated from the co fluxes taken from the literature using @xmath56 ( see tinney et al .
1990 ) where @xmath57 is in @xmath58 .
@xmath59 values were taken from sage , salzer , loose and henkel ( 1992 ) and from taylor , kobulnicky and skillman ( 1998 ) .
@xmath57 was calculated from the relation @xmath60 where @xmath61 is the radius of the telescope beam on the galaxy in @xmath62 .
gas masses for the slugs sample were taken from dunne et al .
( 2000 ) . for those objects whose
@xmath1 m fluxes were obtained from amure ( priv .
comm . ) and bianchi et al . ( 1998 , 2000 ) we took @xmath63 values from young et al .
( 1995 ) .
we investigated the dependence of the method on the metallicity - dependence of the x - factor by trying two different suggested relationships for the dependence of x on metallicity .
arimoto , sofue and tsujimoto ( 1996 ) propose that x has a strong dependence on metallicity .
@xmath64/[h])+9.30\ ] ] where @xmath65 .
israel ( 2000 ) notes the probable dependence on ultra - violet ratiation field , but also suggests a simplified approximate metallicity dependence for x , @xmath66/[h]\ ] ] we have used both of these relations to determine @xmath52 masses for our sample .
these are listed in table 3 . as an additional test of the sensitivity of the method to the x - factor
we have also used a `` standard '' conversion factor , @xmath67 $ ] , from bloemen et al .
( 1986 ) derived from observations of the galaxy .
table 3 also gives the dust masses calculated using equation 1 .
in figure 3 we have plotted the dust masses calculated by the two different methods against each other .
we first performed a simple least - squares fit on the data points .
all of the least squares solutions were consistent with a slope of unity .
this suggests that the properties of dust , in particular @xmath0 and @xmath2 , are similar for dwarf galaxies and large galaxies like our own . on the assumption that this is true
, the value of @xmath0 can be determined in the following way . if the value of @xmath0 we have initially assumed for plotting fig .
3 is correct , the points should lie on a line of slope unity passing through the origin . by measuring the actual offset of the points from the line
, we can calculate a new value of @xmath0 .
we obtain values of 0.07 @xmath68 0.02 , [email protected] and [email protected] @xmath69 using arimoto , israel and galactic @xmath53 x - factors respectively .
the method clearly does not depend critically on the metallicity - dependence of the co - to-@xmath52 x - factor , presumably because either the overall contribution of @xmath52 to the gas - mass is in all cases relatively small , or , where the molecular content is high , the x - factors do not vary greatly at these metallicities .
the main uncertainty in the method is the basic assumption that the fraction of metals in dust ( @xmath2 ) is a universal constant .
if this is not correct , our estimate of this constant , from observations of the interstellar medium in our own galaxy will not be applicable to other galaxies . however , figure 3 is itself evidence that this is not too bad an assumption .
the scatter around the best fit lines is only a factor of 2 in the ordinate .
this scatter includes all observational errors , as well as galaxy - to - galaxy variations in @xmath0 and @xmath2 . on the extreme assumption that all the scatter is caused by variation in @xmath2
, we estimate that the galaxy - to - galaxy variation in @xmath2 can only be a factor of @xmath70 .
if this extreme assumption is correct , then our estimate of @xmath71 , would have an error of a factor of @xmath72 for an individual galaxy .
our conclusion that the fraction of metals incorporated in dust is the same for giant and dwarf galaxies is in contradiction with the conclusions of lisenfeld & ferrara ( 1998 ) , who found that metals are less effectively incorporated in dust in dwarfs than in giant galaxies .
figure 4 shows metallicity plotted against dust - to - gas mass ratio for our sample and that of lisenfeld & ferrara .
the lisenfeld & ferrara points suggest a non - linear relationship between dust - to - gas ratio and metallicity .
however we have re - observed some of the objects from their sample with scuba and find that their dust - to - gas mass ratios are increased when estimated from submm rather than far - infrared flux .
this is expected since in their study , lisenfeld et al .
( 1998 ) used the _ iras _
@xmath73 m flux ratio to determine a dust temperature and hence may have underestimated the dust mass by missing cold dust at longer wavelengths .
the re - observed points are consistent with a slope of unity .
very recently lisenfeld et al .
( 2001 ) have suggested that a low dust mass is appropriate for ngc1569 , whether such a model is applicable to all dwarf galaxies , or is unique to ngc1569 is not yet clear . we have also plotted two of the galaxy dust evolution models of edmunds ( 2001 ) in figure 4 .
the solid line shows a model in which supernovae play an important role in dust production and the dashed line a model in which dust is produced in some kinds of evolved low and intermediate mass stars .
our current dataset is unable to make a distinction between these two models but further submm observations at low metallicities @xmath74 should help demonstrate whether or not condensation of dust grain cores in supernova ejecta is an important dust source .
such observations will be quite challenging as the number of objects suitable for observation at @xmath1 m which have very low metallicities is small . in figure 5
we have reproduced a plot from alton ( 2001 ) that shows estimates of the dust mass absorption coefficient in the ir - submm from a variety of sources .
sopka et al .
( 1985 ) made observations of the thermal emission from the dust envelopes of evolved stars .
rengarajan ( 1984 ) reports results from observations of centrally heated infrared sources deeply imbedded in molecular clouds .
boulanger et al .
( 1996 ) studied the dust - to - gas correlation at high galactic latitude using cobe data .
the draine & lee ( 1984 ) points come from a model for diffuse galactic dust .
both the hildebrand ( 1983 ) and casey ( 1991 ) points are from galactic reflection nebulae .
the alton ( 2000 ) measurement is from scuba observations of the spiral galaxy ngc891 while bianchi et al .
( 1999 ) used cobe & iras data .
finally , agladze et al . (
1994 ) performed laboratory experiments on forsterite @xmath75 .
the line shows @xmath76 . our data ( open triangle )
is consistent with the cobe data of boulanger et al .
( 1996 ) as well as that of draine & lee ( 1984 ) , hildebrand ( 1983 ) and bianchi et al .
we have presented a new method for determining the @xmath1 m dust mass absorption coefficient , and obtain a value of [email protected] @xmath69 . our method for determining this coefficient avoids the problems encountered by previous investigators ( hildebrand 1983 , bianchi 1999 , alton 2000 ) .
using scuba observations at @xmath1 m we can probe the dust emission directly .
we do not make any assumptions about the properties of the dust nor do we rely on similarity between dust in galactic reflection nebulae and that in the general ism . our only assumption ( admittedly an important one )
is that the fraction of metals incorporated in dust is a universal constant .
this seems a justified assumption however since the small scatter in fig .
3 suggests that the fraction of metals in dust is the same for a wide range , from dwarf to high mass galaxies .
having used both metallicity - dependent ( israel 2000 ; arimoto et al .
1996 ) and independent ( bloemen et al . 1986 ) @xmath53 conversion factors to derive a molecular gas mass we have concluded that the metallicity - dependence of the @xmath53 x - factor is not an important consideration in this derivation of @xmath0 . the cobe data of boulanger et al .
( 1996 ) are consistent with our value for @xmath27 as are the data of draine & lee , bianchi and hildebrand .
the @xmath27 value of alton ( 2000 ) that was derived from scuba observations of the edge on galaxy ngc891 is @xmath77 times higher than ours .
they do however admit that uncertainty in their technique means that their quoted value for the dust emissivity could be in error by a factor of 2 either way .
agladaze et al . (
1994 ) laboratory experiments yield a @xmath27 that is @xmath72 times higher than ours , however they use @xmath78 for wavelength extrapolations .
sopka et al .
1985 also use @xmath79 .
unfortunately , the metallicity range of our sample does not extend low enough to make a distinction between the dust formation models proposed by edmunds ( 2001 ) .
we intend to acquire @xmath1 m data for the rest of the lisenfeld & ferrara ( 1998 ) sample where possible .
guest user , canadian astronomy data centre , which is operated by the dominion astrophysical observatory for the national research council of canada s herzberg institute of astrophysics .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
agladze n. , sievers a. , jones s. , burlitch j. & beckwith s. 1994 , nature , 372 , 243 allende prieto c. , lambert d.l . and asplund m. 2001 , apj , 556 , 63 alton p.b . , lequeux j. , bianchi s. , churches d. , davies d. and combes f. 2001 , a&a , 366 , 451 alton p.b . , xilouris e.m . , bianchi s. , davies j. , kylafis n. 2000 , a&a , 356 , 795 arimoto n.,sofue y. and tsujimoto t. 1996 , publ .
japan , 48 , 275 bianchi s. , alton p.b . , davies j.i .
and trewhella m. 1998 , mnras , 298 , l49 bianchi s. , davies j.i . , alton p.b . 1999 ,
a&a , 344 , l1 bianchi s. , davies j.i . ,
alton p.b . , gerin m. , 2000 bolemen j.b.g.m .
1986 , a&a , 154 , 25 bonatto c.j . &
pastoriza m.g .
1990 , apj , 353 , 445 boulanger f. , abergel a. , bernard j.p . , et al .
1996 , a&a , 312 , 256 cardelli j.a . ,
meyer d.m .
, jura m. & savage b.d .
1996 , ap , 467 , 334 casey s.c .
1991 , apj , 371 , 183 davies j.i .
, alton p. , trewhella m. , evans r. , bianchi s. 1999 , mnras , 304 , 495 draine b.t .
& lee h.m .
1984 , apj , 285 , 89 dunne l. & eales s.a .
2001 , mnras , 327 , 697 dunne l. , eales s. edmunds m. , ivison r. , alexander p. , clements d.l . 2000 ,
mnras , 315 , 115 edmunds m.g .
2001 , mnras , 328 , 223 edmunds m.g . &
pagel b.e.j 1984 , mnras , 211 , 507 gonzalez d. , & rosa m. 1995 , pasp , 107 , 1130 hildebrand r.h .
1983 , qjras , 24 , 267 holland w.s .
et al . 1999 ,
mnras , 303 , 659 hughes d. , robson e. , dunlop j. , gear w. , 1993 , mnras , 263 , 607 israel f.p .
2000 , astro - ph/0001250 issa m.r .
, mclaren i. and wolfendale a.w . 1990 ,
a&a , 236 , 237 jenness t. & lightfoot j.f .
1998 , astronomical data analysis software and systems vii , conf .
145 , 216 , r. albrecht , r.n .
hook , h.a .
bushouse eds .
keel w.c . , kennicutt r.c.jr , hummel e. , van der hulst j.m . 1995 ,
aj , 90 , 708 kenney j.d.p . , & young j.s .
1989 , apj , 344 , 171 lisenfeld u. & ferrara a. 1998 , apj , 496 , 145 lisenfeld u. , israel f.p .
, stil j.m . and
sievers a. 2001 , astro - ph/0112212 meyer d.m .
, jura m. , and cardelli j.a .
1998 , apj , 493 , 222 pagel b.e.j , simonson e.a .
, terlevich r.j .
, edmunds m.g .
1992 , mnras , 255 , 325 pagel b.e.j 1997 , nucleosynthesis and chemical evolution of galaxies , cambridge university press petrosian a.r . & burenkov a.n .
1993 , a&a , 279 , 21 rengarajan t. 1984 , a&a , 140 , 213 sage l.j .
, salzer j.j . ,
h , henkel c. 1992 , a&a , 265 , 19 skillman e.d .
, kennicutt r.c . , and hodge p.w .
1989 , apj , 347 , 875 sopka r. , hildebrand r. , jaffe d. et al .
1985 , apj , 294 , 242 taylor c.l .
, kobulnicky h.a .
, skillman e.d .
1998 , apj , 116 , 2746 tinney c.g .
, scoville n.z .
, sanders d.b . , soifier b.t .
1990 , apj , 362 , 473 tully 1988 , nearby galaxies catalog , cambridge university press veilleux s. , kim d .- c .
, sanders d.b . , mazarella j.m . , soifer b.t .
1995 , apjs , 98 , 171 whittet d 1992 , dust in the galactic environment ( iop publishing ) young j.s .
1995 , apjs , 98 , 219 zaritsky d. , kennicutt r.c . , huchra j.p
. 1994 , apj , 420 , 87
|
using a new technique we have determined a value for the constant of proportionality between submillimetre ( submm ) emission and dust mass , the dust mass absorption coefficient ( @xmath0 ) at @xmath1 m .
our method has an advantage over previous methods in that we avoid assumptions about the properties of dust in the interstellar medium .
our only assumption is that the fraction of metals incorporated in the dust ( @xmath2 ) in galaxies is a universal constant . to implement our method we require objects that have submillimetre and far - infrared ( fir ) flux measurements as well as gas mass and metallicity estimates .
we present data for all the galaxies with suitable measurements , including new submm maps for five galaxies .
we find @xmath3 .
we have also been able to use our sample to investigate our assumption that @xmath2 is a universal constant .
we find no evidence that @xmath2 is different for dwarf and giant galaxies and show that the scatter in @xmath2 from galaxy to galaxy is apparently quite small .
dust , extinction - galaxies : dwarf - galaxies : abundances - galaxies : evolution
|
to account for the vast majority of mass in the observable universe and to explain its accelerated expansion , the physical cosmology of today requires two outstanding concepts : ( i ) the matter which does not interact with the electromagnetic force - _ dark matter _ , and ( ii )
the hypothetical energy that tends to increase the rate of expansion of the universe - _ dark energy_. zwicky @xcite , using virial theorem , had suggested for a possible existence of dark matter long ago , which was later on supported by the studies of rotation curves @xcite , gravitational lensing @xcite , cmb anisotropy @xcite and bullet clusters @xcite .
the advent of inflationary theory @xcite led to the convincing belief that @xmath4 of matter content of the universe is hidden mass constituted by @xmath5 dark matter and @xmath6 dark energy @xcite .
dark matter plays a central role in early universe during structure formation and galaxy evolution because of its nature to clump in sub - megaparsec scales .
cobe and cmb experiments suggest that baryonic dark matter is not more than a small fraction of the total dark matter present in the universe @xcite .
more to it , observational constraint regarding neutrino mass and relic neutrino density @xcite eliminate the possibility of hot dark matter in favor of cold non - baryonic dark matter @xcite .
however , warm dark matter ( wdm ) which is neither hot nor cold , rather an intermediate one between them , may be considered as a dark matter candidate along with cold dark matter .
thus , @xmath0-cwdm model consisting of a mixture of cold and warm dark matter may be a possible alternative to @xmath0-cdm model .
a favourite choice of wdm is sterile neutrinos @xcite .
it has been shown @xcite that oscillations of active neutrinos in the primeval universe can produce sterile neutrinos .
boyarsky et al .
@xcite have used recent results on @xmath0-cwdm models to show that for each mass above 2 kev , the presence of at least one model of sterile neutrino can serve as dark matter .
it has also been observed @xcite that sterile neutrinos have the necessary features of a dark matter candidate .
observational results for an accelerating universe @xcite favored the idea of an accelerating agent that had been referred to as dark energy . in accordance with the data , the already introduced standard cold dark matter ( scdm ) model
is found giving way to @xmath0-cdm model that has an advantage of assuming a nearly scale - invariant primordial perturbations and a universe with no spatial curvature as predicted by the inflationary theory @xcite .
it is found in good agreement with various observational results @xcite . in the @xmath0-cdm scenario
the present acceleration of the universe can not be a permanent feature because , structure formation can not proceed during acceleration if it is assumed that dark energy is not coupled to dark matter .
thus , the universe must have undergone a decelerating phase prior to the present accelerating phase @xcite .
observational evidence @xcite also supports this idea .
so , the deceleration parameter must have undergone a flip in sign during cosmic evolution .
one of the candidates among the dark energy models is related to dynamic @xmath0 ( for an overview see @xcite and @xcite ) .
in fact , the concept of dark energy and the physics of accelerating universe appears to be inherent in the @xmath0-term of einstein s field equations .
herein , we perform a study of accelerating universe in context of the time - dependent @xmath0 in the field equations and reveal the importance of dynamic @xmath0 while addressing various key issues and some known physical observable like dark matter , dark energy , geometry of the field , deceleration parameter and its sign flip , age of the universe and equation of state parameter .
for the present purpose we consider the einstein s field equations , @xmath7,\ ] ] which yield friedmann equation @xmath8 for the spherically symmetric flrw metric .
we also get raychaudhuri equation @xmath9 in connection to the evolution scenario for expansion of null or time - like geodesic congruences . here
, @xmath10 is the time - dependent function of the erstwhile _ cosmological constant _ as introduced by einstein , @xmath11 is the scale factor of the universe and @xmath12 is the curvature constant .
the raychaudhuri equation ( [ akr ] ) at once shows that for @xmath13 , acceleration is initiated by the @xmath0-term only that seems to relate @xmath0 with dark energy .
it also shows that for a positive @xmath0 , the universe may accelerate with the condition @xmath14 i.e. @xmath15 is negative for a positive @xmath16 with a definite contribution of @xmath0 in the acceleration .
the placement of @xmath0 in the field equations itself suggests for it to be a part of total energy momentum @xmath17 of the universe .
in fact , for a variable @xmath0 , solutions of field equations are possible only , if instead of @xmath18 , the total energy momentum @xmath19 is conserved @xcite and in that case the second term @xmath20 of the above equation acts as an additional source term in the field equations . in the observational front , the data set coming from the supernova legacy survey ( snls ) during its first year of observation show that dark energy behaves in the same manner as that of a cosmological constant to a precision of @xmath21 @xcite .
thus , @xmath0 being a time - dependent function , its effect on field equations in regard to the accelerating universe is a pertinent study .
using friedmann equation ( [ fried ] ) for energy density @xmath22 and the deceleration parameter @xmath23 in raychaudhuri equation ( [ akr ] ) , one may arrive at following expression for the pressure @xmath24 .
\label{p01}\ ] ] now we consider the phenomenological assumption @xmath25 @xcite .
theoretical and observational values of @xmath0 and @xmath16 for the present day universe , respectively @xmath26 @xcite and @xmath27 @xcite , suggest a more general form @xmath28 with @xmath29 .
as shown later , the physical constraint on @xmath2 requires @xmath30 . in view of this wide observational uncertainty in @xmath2 with regard to its relationship with the time variable @xmath0
should thus be taken carefully while studying field equations for the evolution of the universe . under the same assumption @xmath31 , for a flat universe , equation ( [ rho01 ] ) reads as @xmath32 giving thereby @xmath33 .
similarly , equation ( [ p01 ] ) reads as @xmath34 here , subscript @xmath35 refers to the values of present day universe .
not ignoring the fact that the flat universe would continue evolving with variable @xmath36 , the present value , @xmath37 , and other related quantities ( @xmath38 , @xmath39 , @xmath40 and @xmath41 ) may be treated as dynamic and time variable .
equations ( [ rho02 ] ) and ( [ p02 ] ) yield the barotropic equation of state @xmath42 representing a straight line @xmath43 with @xmath44 . for @xmath45 , equation of state parameter
equals the deceleration parameter and therefore is expected to obey the same physical conditions .
more to it , as a direct consequence of friedmann and raychaudhuri equations , we get while adding equations ( [ rho02 ] ) and ( [ p02 ] ) @xmath46 it is well known that after the initial introduction of the @xmath0 term by einstein in his cosmological model , a number of times that particular term has been accepted as well as rejected . in the 1960 s zeldovich @xcite revived the same term as quantum fluctuations of the vacuum .
he showed that as a slow - rolling approximation of a scalar field in robertson - walker space - time , the stress energy tensor takes the form of a cosmological constant .
this idea of zeldovich played a pivotal role on inflationary theory of guth @xcite .
in fact if we consider a scalar field @xmath47 which obeys the klein - gordon equation , then as slow - rolling approximation we have , @xmath48 . the same situation has been obtained in the present work via equation ( 11 ) for @xmath49 .
this , for the vacuum equation of state , @xmath50 @xcite , provides the condition of constant acceleration ( @xmath51 ) @xcite . in general , the matter - energy density being positive the counterpart negative pressure acts as a repulsive agent and hence the vacuum equation of state has a deep implication in the case of accelerating universe scenario .
obviously , for a collapsing universe with positive @xmath15 , we find @xmath52 and for an accelerating universe we have @xmath53 . its higher limit may be positive depending upon the density .
it is evident from equation ( [ vacuu ] ) that the fate of the universe depends on @xmath41 .
bearing in mind the hubble s law and the assumption @xmath54 , we may directly arrive from equation ( [ d1 ] ) at @xmath55 .
this demonstrates that for @xmath56 or for @xmath57 constant , the universe has been evolving through a constant acceleration as indicated earlier @xcite .
it is interesting to note here that einstein initially obtained an expanding universe with @xmath58 ( and hence to counteract the dynamical effects of gravity , which would cause the matter - filled universe to collapse , he later on adopted a non - zero @xmath0 to obtain a static model ) while de sitter obtained a similar expanding universe with constant @xmath0 and devoid of any ordinary matter ( which ultimately made einstein to drop the cosmological constant from his general relativistic field equations ) .
we are also curious about the situation @xmath56 which , by virtue of the equation ( [ rho02 ] ) , makes the energy density to vanish .
it seems to correspond to the special relativistic universe of milne @xcite under the zero - density limit of the expanding flrw metric with no cosmological constant @xcite but with @xmath59 .
the available observational data for redshift and scale factor have got flexibility , though limited , to distinguish between a time varying and a constant equation of state @xcite .
it , therefore , supports a time variable @xmath60 . such a time variable @xmath60 has been used in literature by many authors @xcite to predict various physical observables of the universe .
some useful limits on @xmath60 was suggested by snia data , @xmath61 @xcite whereas refined values come from combined snia data with cmb anisotropy and galaxy clustering statistics which are @xmath62 @xcite .
moreover , inflation at an early stage scales the parameter @xmath60 , which combined with the above data and dark energy constraint ( @xmath63 ) suggests a physical condition , @xmath64 @xcite . in the light of our previous discussion ,
@xmath41 too must obey the same physical conditions as @xmath60 but for @xmath45 .
the experimental limits , @xmath65 @xcite does fall in this range , which supports the view point that a variable @xmath60 provides physical reason for a nonzero value of @xmath38 and for a limited time variability for it . moreover , the above ranges of the values of @xmath60 and @xmath41 can be achieved if @xmath2 is constrained to lie in the range @xmath66 . for an accelerating universe
, @xmath38 must be positive , so is @xmath2 .
we plot the linear relationship ( [ d2 ] ) for different values of @xmath2 within its physical range in fig .
we choose @xmath60 from 0.0 to -1.0 covering the range @xmath67 . to reproduce experimental @xmath41 ,
we require smaller values of @xmath2 for smaller values of @xmath60 .
the hypotenuse and the base of the bold face triangle , respectively represent @xmath68 ( or @xmath56 ) and @xmath69 , thereby representing the full physical region .
we may compare @xmath68 plot with other lines of the figure representing positive values of @xmath2 ( or respective values of @xmath38 ) to extract the effect of @xmath38 on @xmath41 .
the thin solid line shows the case , @xmath45 .
we find @xmath41 negative within the bold face triangular region demarcated by physical conditions @xmath70 and @xmath71 .
the values @xmath72 ( or @xmath73 ) lies below the experimental @xmath41 . towards the smaller values of @xmath2 ( @xmath74 ) and higher values of @xmath60 ( @xmath75 ) , @xmath41 is observed to flip the sign in the accelerating epoch .
before the present cosmic acceleration , which had started only recently ( a few gyr earlier ) , the universe was expanding with deceleration .
so , at the turnover stage ( from deceleration to acceleration ) , the deceleration parameter must have changed its sign .
it is worth noting here that @xmath41 and @xmath2 ( hence @xmath76 determine the fate of the universe .
one may arrive at the same conclusion through recognizing the fact that density is directly proportional to @xmath38 .
hence , time variability of @xmath38 can not be put aside while addressing the geometry of the universe .
one of the predictions of the inflation theory is a flat universe with a large value of @xmath0 in the early stages of the universe .
thus , one may argue that it is the cosmological parameter that determined the geometry of the universe and made it flat during inflation . in this connection
we mention that the equation ( 10 ) is nothing but a special form of the general equation which can be obtained easily from the equations ( 4 ) and ( 6 ) for flat universe ( @xmath77 ) and using the ansatz @xmath31 .
it is also to be noted here that the fig .
[ fig1 ] demonstrates the evolution of the universe for a number of selected values of @xmath60 .
this implies that general form of equation ( 10 ) will show the same behavior in the @xmath78 versus @xmath79 plot .
the equation ( 10 ) ( or it s general form ) seems to be telling us that the cosmological constant scales , via @xmath80 , roughly as the ` velocity ' squared of the universe .
so , equation ( 10 ) then will remain valid for a wide range of cosmological epoch so far as the values of @xmath79 and @xmath78 are concerned .
we now focus on the implications of @xmath38 on the hidden mass .
it should be mentioned here that a combination of type ia supernova data and cmb anisotropy gives a best fit value for the dark energy density parameter @xmath81 as @xmath82 @xcite .
this value of @xmath81 yields , @xmath83 @xcite .
we may , from equation ( [ rho02 ] ) , deduce that @xmath84 is equivalent to @xmath85 as have been suggested for other type of phenomenological @xmath0-model @xcite .
it may also be shown that the ratio @xmath86 gives a measure for the hidden mass of the universe .
on the basis of the phenomenological model @xmath87
@xcite , the present density has been obtained as @xmath88 , which is close to the lower limit of the value obtained by guth @xcite : @xmath89 .
therefore , knowing the measured galactic mass density , @xmath90 @xcite , one finds @xmath91 suggesting thereby that the galactic mass density is about @xmath92 - @xmath93 of the total mass density of the present universe .
hence , according to the present model there is hidden mass ranging from @xmath94 - @xmath95 , which seemingly approves the value of ref .
variation of @xmath86 is shown in the lower panel of fig .
[ fig2 ] with respect to @xmath38 and @xmath2 .
higher values of @xmath38 and lower values of @xmath2 demonstrate large missing mass . on the basis of the most recent observations weinberg @xcite reports the age of the universe as @xmath96 gyr .
dynamic @xmath0 model has also been used to estimate the age @xcite , however it seems to suffer either from low - age like @xmath97 gyr @xcite that is less than the estimated globular cluster ages , @xmath98 gyr @xcite or it is as high as @xmath99 gyr @xcite . with the variations of @xmath2 and @xmath38
we plot our calculations in the upper panel of fig .
we notice that all the values of @xmath2 within its physical range reproduce observed age of the universe but with a suitable @xmath38 .
one requires , smaller values of @xmath38 , for the smaller values of @xmath2 in order to be close with observation .
this sheds light on the important time variability of @xmath37 and its correlation to @xmath37 , @xmath100 .
the present work may be considered as a part of a series of papers dealing with the behaviour of some phenomenological @xmath0-dark energy models under specific assumptions on the physical parameters @xmath0 , @xmath79 , g etc .
@xcite . in those works various aspects of @xmath0 models
have already been explored .
so , the present work has been carried out in the same line and is guided by the motivation of revealing some new features of dark energy in the framework of the present accelerating universe . in the present paper , it has been possible to arrive at various interesting physical ideas of modern cosmology through simple considerations of time variability of the observables of flat universe , specially @xmath38 .
we have considered the phenomenological assumption , @xmath101 , for the full physical range of @xmath2 and hence the relevance of time variable @xmath0 with regard to various key issues like dark matter , dark energy , geometry of the field , age of the universe , deceleration parameter and barotropic equation of state has been trivially addressed .
interestingly , the deceleration parameter @xmath41 and the barotropic equation of state parameter @xmath60 have been found to obey a straight line relationship ( equation ( [ d2 ] ) ) with slope @xmath44 for a flat universe described by friedmann and raychaudhuri equations . for @xmath45 ,
both the parameters , @xmath60 and @xmath41 , are equal and hence are expected to obey the same physical conditions .
the assumption , @xmath59 , seems to represent the milne universe .
it has been shown that within the physical limits of accelerating universe , @xmath41 may flip its sign towards lower end values of @xmath2 and higher end values of @xmath60 .
the authors ( sr & aau ) would like to express their gratitude to the authority of iucaa , pune for providing them the visiting associateship under which a part of this work was carried out .
sr is personally grateful to andrew debenedictis , simon fraser university , canada for helpful discussion via electronic correspondence .
we all are thankful to the referee for valuable suggestions which have enabled us to improve the manuscript substantially .
zwicky , f. : astrophys .
j. * 86 * 217 ( 1937 ) roberts , m.s . , rots , a.h .
: astron . astrophys .
* 26 * 483 ( 1973 ) einasto , j. , kaasik , a. , saar , e. : nature * 250 * 309 ( 1974 ) tyson , j.a . , et al .
j. * 91 * 1274 ( 1986 ) subramanian , k. , chitre , s.m .
: astrophys .
j. * 313 * 13 ( 1987 ) bode , p.w . ,
bertschinger , e. : astron .
. suppl . * 186 * 0805 ( 1995 ) doroshkevich , a.g . , schneider , r. : astrophys .
j. * 469 * 445 ( 1996 ) clowe , d. , randall , s.w . ,
markevitch , m. : nucl .
. suppl . *
173 * 28 ( 2007 ) bradac , m. , et al .
: astrophys .
j. * 681 * 187 ( 2008 ) guth , a.h .
d * 23 * 347 ( 1981 ) linde , a. : phys .
b * 108 * 389 ( 1982 ) albrecht , a. , steinhardt , p.j . :
lett . * 48 * 1220 ( 1982 ) amsler , c. , et al .
b * 667 * 1 ( 2008 ) sahni , v. : lect .
notes phys .
* 653 * 141 ( 2004 ) elgary , . , et al .
rev . lett . * 89 * 061301 ( 2002 ) spergel , d.n . , et al .
: astrophys .
* 148 * 175 ( 2003 ) dodelson , s. , widrow , l.m .
72 * 17 ( 1994 ) shi , x .- d .
, fuller , g.m .
. lett . * 82 * 2832 ( 1999 ) dolgov , a.d . ,
hansen , s.h .
: astroparticle phys .
* 16 * 339 ( 2002 ) abazajian , k. , fuller , g.m . , patel , m. : phys .
d * 64 * 023501 ( 2001 ) abazajian , k. , fuller , g.m . ,
tucker , w.h .
: astrophys .
j. * 562 * 593 ( 2001 ) asaka , t. , blanchet , s. , shaposnikov , m. : phys .
b * 631 * 151 ( 2005 ) asaka , t. , shaposnikov , m. : phys .
b * 620 * 17 ( 2005 ) boyarsky , a. , lesgourgues , j. , ruchayskiy , o. , viel , m. : phys .
102 , 201304 ( 2009 ) gu , p .- h . ,
hirsch , m. , sarkar , u. , valle , j.w.f .
d * 79 * 033010 ( 2009 ) perlmutter , s. , et al .
: nature * 391 * 51 ( 1998 ) riess , a.g . , et al .
j. * 116 * 1009 ( 1998 ) guth , a.h . , pi , s .- y .
lett . * 49 * 1110 ( 1982 ) hawking , s.w .
b * 115 * 295 ( 1982 ) bardeen , j.m . , steinhardt , p.j . ,
turner , m.s .
d * 28 * 679 ( 1983 ) tegmark , m. , et al .
: astrophys .
j. * 606 * 702 ( 2004 ) amendola , l. : mon . not .
* 342 * 221 ( 2003 ) riess , a.g .
: astrophys .
j. * 560 * 49 ( 2001 ) overduin , j.m . ,
cooperstock , f.i .
d * 58 * 043506 ( 1998 ) sahni , v. , starobinsky , a. : int .
d * 9 * 373 ( 2000 ) vishwakarma , r.g . : class .
quantum gravit .
* 17 * 3833 ( 2000 ) astier , p. , et al .
: astron . astrophys . * 447 * 31 ( 2006 ) hsu , s.d.h .
b * 594 * 13 ( 2004 ) ray , s. , mukhopadyay , u. , meng , x .- h .
: grav . cosmol .
* 13 * 142 ( 2007 ) carmeli , m. : int .
a * 17 * 4219 ( 2000 ) zeldovich , ya.b .
: jetp lett .
* 6 * 316 ( 1967 ) zeldovich , ya.b .
* 11 * 381 ( 1968 ) guth , a.h .
: the inflationary universe ( new york , addison wesley , p22 , 1997 ) davies , p.c.w . : phys .
d * 30 * 737 ( 1984 ) hogan , c. : nature * 310 * 365 ( 1984 ) kaiser , n. , stebbins , a. : nature * 310 * 391 ( 1984 ) blome , j.j .
, priester , w. : naturwissenshaften * 71 * 528 ( 1984 ) weinberg , s. : cosmology ( oxford univ . press , p39 , 2008 ) milne , e.a . : relativity , gravitation and world structure ( oxford univ . press , oxford , 1935 ) kujat , j. , et al . :
j. * 572 * 1 ( 2002 ) bartelmann , m. , et al .
: new astron . rev . *
49 * 199 ( 2005 ) nojiri , s. , odintsov , s.d .
b * 649 * 440 ( 2007 ) brevik , i. , gorbunova , o.g . , timoshkin , a.v .
j. c * 51 * 179 ( 2007 ) usmani , a.a . , et al .
: mon . not .
r. astron .
soc . lett . *
386 * , l92 ( 2008 ) knop , r. , et al .
: astrophys .
j. * 598 * 102 ( 2003 ) catton , c. , visser , m. : phys .
d * 78 * 063501 ( 2008 ) efstathiou , et al .
: mon . not .
r. astron .
. lett . * 303 * , l37 ( 1999 ) peebles , p.j.e . , partridge , r.b .
: astrophys .
j. * 148 * 713 ( 1967 ) vishwakarma , r.g . : class .
quantum gravit . * 19 * 4747 ( 2002 ) cayrel , r. , et al . :
nature * 409 * 691 ( 2001 ) ray , s. , et al .
. phys . * 48 * 2499 ( 2009 ) mukhopadhyay , u. : ph .
d. thesis , jadavpur university , india ( 2009 ) mukhopadhyay , u. , ray , s. : n. b. u. math . j.
* ii * 51 ( 2009 ) mukhopadhyay , u. , ray , s. , rahaman , f. : int . j. mod .
d * 19 * 475 ( 2010 )
|
we perform a deductive study of accelerating universe and focus on the importance of variable time - dependent @xmath0 in the einstein s field equations under the phenomenological assumption , @xmath1 for the full physical range of @xmath2 .
the relevance of variable @xmath0 with regard to various key issues like dark matter , dark energy , geometry of the field , age of the universe , deceleration parameter and barotropic equation of state has been trivially addressed .
the deceleration parameter and the barotropic equation of state parameter obey a straight line relationship for a flat universe described by friedmann and raychaudhuri equations .
both the parameters are found identical for @xmath3 .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore
|
several observations suggest that the andromeda galaxy could be the remnant of a major merger .
m31 has a robustly classical and not pseudo bulge , which is an argument in favor of such a merger scenario , according to @xcite .
@xcite also argued that due to the high metallicity of the m31 halo and the @xmath0 luminosity profile of the bulge , m31 is likely the result of a merger of two massive metal - rich ancestors .
conversely , there are several difficulties in reproducing the giant stream by a recent minor merger , according to @xcite . for instance
, they do not succeed in reproducing the metal - poor ( [ fe / h]@xmath1 ) stellar population .
in addition , a recent merger should have produced a fraction of young stars that is not detected .
a last argument is given by @xcite .
around the andromeda galaxy , several other features like shelves or clumps have been detected using deep exposures with hst / acs .
@xcite found a moderate range of age and metallicity in these structures , which suggest a simultaneous formation process . although a major merger seems to be a likely event in the m31 lifetime , other mechanisms are likely to occur in parallel such as minor mergers or gas accretion , and be at the origin of some m31 sub - structures
we propose , from figure [ fig1 ] , a chronological history of the different structures in m31 , as this figure can be used as a clock for determining the occurrence of merger phases as follows . the star formation history of a merging system is enhanced during the first passage until the fusion and then at the fusion itself ( see * ? ? ?
during such a major merger event , most of the gas and stars in the remnant outskirts is deposited by tidal tails formed during the first passage and later during the fusion of the cores .
a few hundred billion years after its formation , the tidal tail dilutes , provoking a natural quenching of the residual star formation ( see * ? ? ?
thus , the age of the material brought by tidal tails provides , with a relatively small delay , possible dates for both the first passage and fusion times . in the following
, we assumed that the first passage occurred from 8.5 to 9 gyr ago , and that the corresponding tidal tails are responsible for the halo enrichment seen in the 21 and 35 kpc fields , without significant star formation more recent than 8.5 gyr .
the thick disk has a star formation history comparable to that of the giant stream and is also generated by material returning to the galaxy mostly from tidal tails generated at the fusion . because their youngest significant population of stars has ages of 5.5 gyr , the delay between the first passage and fusion ranges between 3 and 3.5 gyr .
this could be accommodated for by relatively large impact parameters ( 20 - 30 kpc ) . .
squares represent fields observed by @xcite , and are linked to their measurements by arrows .
figure extracted from @xcite . ]
the gadget2 hydrodynamical code @xcite supplemented by star formation , feedback and cooling prescriptions @xcite was used . for simulating large scale aspects ( thin and thick disk ,
b / t ratio , 10 kpc ring ) , low resolution are enough , typically @xmath2 particles .
however , we had to increase the number of particles for reproducing the giant stream , i.e , more than @xmath3 .
other structures could not be simulated with such a resolution ( like the double nucleus ) because they require a much better spatial resolution .
in addition , some features can be formed by minor mergers , or other mechanisms completely independent of the major merger process assessed here .
observations strongly constrain the parameters of the simulations .
the first constraints are the rotation velocity curve of the thin disk and the b / t ratio . according to @xcite , a mass ratio from @xmath43:1 to @xmath42:1
is required to produce a remnant bulge with b / t @xmath4 0.3 .
another constraint is the total baryonic mass for the progenitors , which must be close to the m31 baryonic mass , i.e. , @xmath5 m@xmath6 .
a prograde - retrograde orientation for the spin axis was chosen , which is favourable to rebuild disk ( see * ? ? ?
considering the gas fraction , reforming an sb - like thin disk demands high gas fraction just before the fusion , i.e more than 50% . as a result , low star formation and gas - rich progenitors , with more than 65% of gas , were used in order to keep enough gas before the fusion .
these assumptions have theoretical grounds .
it has been estimated that many high - z galaxies are gas rich , hence m31 progenitors should have a high gas fraction .
low star formation could be caused either by higher feedback efficiency in primordial medium , or because the gas in the progenitors is less concentrated than in present - day spirals as it is in present - day low - surface brightness galaxies , or because cooling is less efficient in a relatively pristine medium , or a combination of all these factors .
more generally , the expected increase of the gas metal abundance ( expected to be slow before the fusion but very efficient during the fusion ) may help to increase the molecular gas fraction , the optical depth of the gas , and the radiation pressure effects , all contributing to a change in the star formation history during the interaction ( t. j. cox 2010 , private communication ) .
another structure to be reproduced was the 10 kpc ring , which suggests a polar orbit .
the last feature , the giant stream , is caused by particles coming back from the tidal tail formed just before fusion .
figure [ fig2 ] describes the formation of such a structure that is aligned along the trajectory of the satellite which falls into the mass center at the fusion , 4.5 gyr after the beginning of the simulation .
we verified that , within the family of orbits we chose , the strength of the tidal tail ( see figure [ fig2 ] , panel at 4.2 gyr ) depends on the inclination of the progenitors relative to the orbital angular momentum .
the lmc has a particular status in the local group : it is the only dwarf irregular within 250 kpc of the mw . with new velocity measures , close to 380 km / s @xcite , and a distance from the milky way equal to 50 kpc ,
the idea that the lmc go past for the first time around the milky way was revived . by tracing back the lmc
, it goes back far away from the mw .
it results many configurations for which the lmc was close to m31 in the past ; and could have been formed as a tidal dwarf during the merger event .
the radial velocity between the mw and m31 , @xmath4 136 km / s , is known but the tangential one is not measured for the moment because it is too small . to test this idea ,
simulations were run with different values for the m31 tangential velocity and we were looked for which one the lmc would go back to m31 @xcite .
|
we investigated a scenario in which m31 could be the remnant of a major merger and at the origin of the lmc .
galaxy merger simulations were run in order to reproduce some m31 properties .
we succeeded in reproducing some of the most important m31 large - scale features like the thick disk or the polar ring , and gave a possible explanation for the formation of the giant stream .
we also found that the lmc could be expelled by this high energetic phenomenon .
|
rare @xmath3 decays are very important in particle physics , because they are important windows in testing the standard model ( sm ) and they are sensitive to new physics . as a rather simple method ,
factorization approach is accepted , because it explained many decay branching ratios successfully @xcite .
recently , some efforts have been made to improve their theoretical application .
one of these methods is the perturbative qcd approach ( pqcd ) , in which many rare branching ratios such as @xmath4 @xcite , @xmath5 @xcite , @xmath6 @xcite were predicted . in recent calculations , @xmath7 ,
@xmath8 have been analyzed in the pqcd approach @xcite , leaving @xmath9 not calculated . in decay @xmath10 , none of quarks in the final states is the same as one of the @xmath3 meson .
this decay is a pure annihilation type decay , which is described as @xmath3 meson annihilating into vacuum and @xmath11 and @xmath12 produced from vacuum afterwards .
some information about pqcd picture can be get in @xcite . in pqcd
, the decay amplitude is usually separated into soft(@xmath13 ) , hard(@xmath14 ) and harder(@xmath15 ) dynamics by different scales .
the factorization theorem allows us to write the decay amplitude as convolution , @xmath16 .
\label{eq : convolution1}\ ] ] in the function , @xmath17 are momenta of light quarks in each meson .
@xmath18 is wilson coefficient which comes from the qcd radiative corrections to the four quark operators .
@xmath19 is the wave function which describes the inner information of meson @xmath20 .
@xmath14 describes the four - quark operator and the quark pair from the sea connected by a hard gluon whose scale is at the order of @xmath21 , so the @xmath14 can be perturbatively calculated .
the hard part @xmath14 is channel dependent , while @xmath19 is independent of the specific processes .
some analytic formulas for the decay amplitudes of @xmath22 decays will be given in the next section . in section [ sc : neval ] , we give the numerical results and discussion . finally , we conclude this study in section [ sc : concl ] .
for simplicity , we consider @xmath3 meson at rest . in the light - cone
coordinate , the @xmath3 meson momentum @xmath23 , @xmath24 momentum @xmath25 and @xmath12 momentum @xmath26 are taken to be : @xmath27 where @xmath28 and @xmath29 . putting the light ( anti-)quark momenta in @xmath30 , @xmath11 and @xmath12 mesons as @xmath31 , @xmath32 , and @xmath33 , respectively
, we can choose @xmath34 the @xmath12 meson s longitudinal polarization vector @xmath35 and transverse polarization vector @xmath36 are given by : @xmath37 then , integration over @xmath38 , @xmath39 , and @xmath40 in eq.([eq : convolution1 ] ) leads to : @xmath41 , \label{eq : convolution2}\end{gathered}\ ] ] where @xmath42 is the conjugate space coordinate of @xmath43 , and @xmath44 is the largest energy scale in @xmath14 , as the function in terms of @xmath45 and @xmath42 .
the large logarithms ( @xmath46 ) coming from qcd radiative corrections to four quark operators are included in the wilson coefficients @xmath18 .
the large double logarithms ( @xmath47 ) on the longitudinal direction are summed by the threshold resummation @xcite , and they lead to a jet function @xmath48 which smears the end - point singularities on @xmath45 .
the last term , @xmath49 , contains two kinds of logarithms .
one of the large logarithms is due to the renormalization of ultra - violet divergence @xmath50 , the other is double logarithm @xmath51 from the overlap of collinear and soft gluon corrections .
this sudakov form factor suppresses the soft dynamics effectively @xcite .
thus it makes perturbative calculation of the hard part @xmath14 reliable . the heavy @xmath3 and @xmath52 meson wave functions are restricted by heavy quark symmetry . in the heavy quark limit , we may use only one independent distribution amplitude for each of them @xcite .
@xmath53 \phi_m(x , b),\ ] ] where @xmath54 , @xmath52 . for the light @xmath55 meson ,
only the longitudinal wave function for outgoing state is relevant , which is written as : @xmath56.\ ] ] unlike the heavy mesons , there are three different distribution amplitudes . in the decay @xmath0 ,
the effective hamiltonian at the scale lower than @xmath57 is the same as @xmath58 decay @xmath59 , \\
o_1 = ( \bar{b}\gamma_\mu p_l d ) ( \bar{c}\gamma^\mu p_lu ) , \quad o_2 = ( \bar{b}\gamma_\mu p_lu ) ( \bar{c}\gamma^\mu p_ld),\end{gathered}\ ] ] where the projection operator is defined by @xmath60 .
@xmath61 are the products of the ckm matrix elements , and @xmath62 are the wilson coefficients . according to the effective hamiltonian , the lowest order diagrams of @xmath63 are drawn in fig.[fig : diagrams1 ] . as stated above
, this decay only has annihilation diagrams .
after perturbative qcd calculations , we get the factorizable contribution @xmath64 , and the nonfactorizable contribution @xmath65 illustrated by fig.1(a)(b ) and fig.1(c)(d ) , respectively .
@xmath66 , \label{eq : fa}\end{gathered}\ ] ] @xmath67 .
\label{eq : ma}\end{gathered}\ ] ] in our work , @xmath68 , @xmath69 and @xmath70 in numerators are neglected . in the functions
, @xmath71 is the group factor of @xmath72 gauge group , and the functions @xmath73 , @xmath74 , @xmath75 are given in the appendix of @xcite . comparing with the previous calculated @xmath76 decay @xcite
, we find that the leading twist contribution , which is proportional to @xmath77 , is almost the same .
however the sub - leading twist contribution , which is proportional to @xmath78 in ( [ eq : fa ] ) and ( [ eq : ma ] ) , is quite different .
the total decay width for @xmath0 decay is given as @xmath79 conjugated model , @xmath80 , is the same value as @xmath81 , just replace @xmath61 with @xmath82 . since there is only one kind of ckm phase involved in this decay , there is no cp violation in the standard model .
we use the same @xmath3 and @xmath52 meson wave functions as before @xcite @xmath83,\ ] ] @xmath84.\ ] ] the @xmath55 meson s distribution amplitudes are given by light cone qcd sum rules @xcite : @xmath85 + 1.68 c_4^{1/2}(1 - 2x ) \right .
\nonumber \\ & & \left.+0.06(1 - 2x)^2[5(1 - 2x)^2 - 3]+0.36[1 - 2(1 - 2x)(1+\ln(1-x)]\right\ } , \\
\phi_{k^*}^s(x ) & = & \frac{f_{k^*}^t}{2\sqrt{2 n_c}}\left\{3 ( 1 - 2x ) [ 1 + 0.2 ( 1 - 2x)+0.6 ( 10x^2 - 10x+1 ) ] \right.\nonumber \\ & & \left .
-0.12x(1-x)+0.36[1 - 6x-2\ln(1-x)]\right\},\end{aligned}\ ] ] with the gegenbauer polynomials , @xmath86 in addition , we use the following input parameters , for meson decay constants and the ckm matrix elements and the lifetime of @xmath30 @xcite , @xmath87 the branching ratio obtained from the analytic formulas may be sensitive to many parameters especially those in the meson wave functions @xcite . similar to
the @xmath88 decays @xcite , we found that the branching ratio is not sensitive to the parameter change of the @xmath55 meson wave function .
but they do be sensitive to the heavy b and @xmath52 meson wave function parameters .
for illustration of the uncertainties of the branching ratios , we choose the following @xmath3 and @xmath52 meson wave function parameters @xmath89 using the range stated above , the branching ratio normalized by the decay constants and the ckm matrix elements results in : @xmath90 considering the uncertainty of @xmath91 , @xmath92 and the ckm matrix elements , the branching ratio of the @xmath93 decay is at the order of @xmath2 .
this is still far from the current experimental upper limit @xcite , @xmath94 .decay branching ratios calculated in pqcd approach using the same parameters for @xmath3 and @xmath52 wave functions , and @xmath95mev . [
cols="^,^,^,^ " , ] for comparison , we list the branching ratios of @xmath96 and @xmath97 calculated in pqcd approach , using the same parameters for @xmath3 and @xmath98 wave functions . for simplicity
, we set @xmath99mev . from the above result
, we can see that the branching ratio of @xmath97 is a little larger than that of @xmath100 @xcite . from equation ( [ eq :
fa],[eq : ma ] ) and formulas in ref.@xcite , we find the coefficients of sub - leading twist become negative , but they are all proportional to @xmath78 which is a little bit small . because of the suppression of ckm matrix elements , the branching ratio of @xmath93 is much smaller than that of the neutral decay @xmath101 .
since pure annihilation type b decays are suppressed comparing to spectator diagram decays , the soft final state interactions may be important @xcite . in our case , @xmath30 meson can decay into @xmath102 and @xmath103 , the secondary particles then exchanging a @xmath104 , scatter into @xmath11 , @xmath105 through final state interaction afterwards .
this picture is depicted in fig.[fig : diagrams2 ] , which is difficult to calculate accurately , since final state interaction is purely non - perturbative @xcite . in ref.@xcite ,
the results from pqcd approach for @xmath106 decay were consistent with the experiment , which shows that the soft final state interaction may not be important .
in hadronic two - body @xmath3 meson decays , the energy release is larger than 1 gev .
the final state mesons are moving very fast .
the soft final states interaction may not be important in the two - body @xmath3 meson decays .
the pqcd approach based on @xmath1 factorization theorem is applicable to the calculation of b meson decays . in this work , we calculate the @xmath107 decay in the pqcd approach .
since neither of quarks in @xmath30 appeared in the final state mesons , this process occurs only via annihilation type diagrams . from our pqcd work ,
the branching ratio of @xmath107 is small in sm , which is around @xmath2 .
this branching ratio will be measured in the lhc - b in future .
this small branching ratio predicted in the sm , makes the channel sensitive to any new physics contribution .
99 m. wirbel , b. stech , m. bauer , z. phys .
c29 , 637 ( 1985 ) ; m. bauer , b. stech , m. wirbel , z. phys .
c34 , 103 ( 1987 ) ; a. ali , g. kramer and c.d .
l , phys .
d58 , 094009 ( 1998 ) ; c.d .
l , nucl .
74 , 227 - 230 ( 1999 ) ; y .- h .
chen , h .- y .
cheng , b. tseng , k .- c .
yang , phys .
d60 , 094014 ( 1999 ) .
k. ukai , preceedings for 4th international conference on b physics and cp violation ( bcp4 ) , japan , 2001 ; c .- d .
l and k. ukai , hep - ph/0210206 , to appear in eur .
j. c ; y. li and c .- d .
l , hep - ph/0304288 .
review of particle physics , k. hagiwara _ et al .
d66 , 010001 ( 2002 ) .
zheng , phys .
b356 , 107 ( 1995 ) ; z .- h .
li , f. yuan , h .- q .
zheng , commun .
theor . phys .
34 , 567 ( 2000 ) ; j .- w .
li , m .- z .
yang , d .- s .
du , preprint , hep - ph/0206154 ; m. ablikim , d .- s .
du , m .- z .
yang , preprint , hep - ph/0211413 .
|
the rare decay @xmath0 can occur only via annihilation diagrams in the standard model .
we calculate the branching ratio in perturbative qcd approach based on @xmath1 factorization theorem .
we found that the branching ratio of this decay is about of order @xmath2 , which may be sensitive to new physics .
( 0,0)(0,0 ) ( 340,420)bihep - th-2003 - 19
|
r0.42 a significant fraction of @xmath0 collisions at the lhc will involve ( quasi - real ) photon interactions , at center - of - mass ( c.m.s ) energies well beyond the electroweak energy scale .
hence , the lhc can to some extend be considered as a high - energy photon - photon or photon - proton collider .
the equivalent photon approximation ( epa ) can be successfully used to describe the majority of processes involving photon exchange , provided that the amplitude of a given process can be factorised into the photon exchange and interaction parts @xcite . the photon - photon and photon - proton cross sections , @xmath1 and @xmath2 , must be convoluted with the photon spectra @xmath3 to obtain the @xmath0 cross sections .
this paper considers the low pile - up conditions available at start - up and focuses on the irreducible backgrounds to the presented analyses .
deeper studies involving inclusive background will be the object of later communications .
many @xmath4 cross sections at the lhc are sufficiently large to yield interesting measurements ( see figure [ fig : gammagammaxsec ] ) . in particular , the expected high statistics for exclusive @xmath5 pair production should allow precise measurement of the @xmath6 quartic couplings .
the existence of new massive charged particles can also be probed using the two - photon pair production .
the particular case of super - symmetric pairs is considered here . moreover
, the two - photon exclusive production of muon pairs will be an excellent luminosity monitoring tool @xcite .
two - photon exclusive production of muon pairs at the lhc has a well known cross section , and requires very small hadronic corrections .
small theoretical uncertainties and a large cross section ( @xmath7 pb at lhc energies ) makes it a perfect candidate for the lhc absolute luminosity measurement @xcite .
the selection procedure is very simple : two opposite charge muons within the central detector acceptance ( @xmath8 ) , with transverse momenta above some low thresholds ( @xmath9 or @xmath10 gev ) . requiring one proton tagged would yield @xmath11 muon pairs detected in a @xmath12 h run at the average luminosity of @xmath13 , allowing run by run calibration of the very forward detectors ( vfd ) .
two - photon production of @xmath5 boson pairs provides a unique opportunity to investigate anomalous gauge boson couplings , in particular the quartic gauge couplings ( qgc ) @xmath6 . for 1 @xmath14 , the obtained limits of @xmath15 , @xmath16 , @xmath17 and @xmath18 for @xmath19 , @xmath20 , @xmath21 , @xmath22 , respectively , are about 10,000 times better than the best limits established at lep2 @xcite clearly showing large and unique potential of such studies at the lhc . the unique signature with large lepton acoplanarity ( and large missing transverse momentum )
drastically reduces the backgrounds .
r0.42 the interest in the two - photon exclusive production of pairs of new charged particles is three - fold : it provides a new , complementary and very simple production mechanism ; constraints on the masses of new particles could be obtained using double vfd tagged events ; finally , final states involving super - symmetric particles are usually produced without cascade decays .
the fully leptonic final state consists of two acoplanar charged leptons and large missing energy has low background .
the corresponding feynman diagrams are shown at figure [ fig.diag.susy ] .
the only irreducible background for this type of processes is the exclusive @xmath5 pair production . indeed
, direct lepton pairs @xmath23 can be suppressed using large acoplanarity cuts .
discovery of super - symmetry via @xmath24 interaction is _ a priori _ difficult because of production cross sections smaller than 1 @xmath25 before acceptance cuts .
however , the measured energy of the two scattered protons in vfds could allow for the distinction between various contributions to the photon - photon invariant mass @xmath26 @xcite .
the high luminosity and the high c.m.s .
energy of photo - production processes at lhc offer interesting possibilities for the study of electroweak interaction and for searches beyond the standard model ( bsm ) up to the tev scale .
the cross sections for various electroweak and bsm reactions together with their irreducible ( standard model ) background processes are shown in figure [ fig : gammapartonxsec ] .
several processes have been studied in detail to illustrate the discovery potential of photo - production at lhc .
in contrast to photon - photon reactions , the photo - production processes involve topologies with hard jets in the final state .
the effect of jet algorithms and the efficiency of event selection was taken into account using a fast simulation of a typical multipurpose lhc detector response @xcite .
the associated @xmath27 photo - production cross section reaches 23 @xmath25 for a 115 gev higgs boson and 17.5 @xmath25 for a 170 gev higgs boson .
it represents more than @xmath28 of the total inclusive @xmath27 production at the lhc .
r0.45 when the decay branching ratio of the higgs boson into a @xmath5 pair becomes dominant , the same sign lepton signature coming from leptonic decays of two out of the three produced @xmath5 seems promising .
it has a signal to noise ratio of about one , which is unique at lhc .
an integrated luminosity of a few tens of inverse femptobarns could directly reveal the @xmath29 gauge coupling .
the @xmath30 coupling , very challenging to assess in parton - parton processes , could be probed for a light higgs boson if sufficient integrated luminosity without pile - up can be collected .
photo - production of single top is dominated by t - channel amplitudes when the top quark is produced in association with a @xmath5 boson .
in contrast to proton - proton deep inelastic scattering where the ratio of @xmath31 associated production cross section to the sum of all top production cross sections is only about @xmath32 , it is about 10 times higher in photo - production as illustrated in table [ vistop ] . this provides a unique opportunity to study this reaction at the start phase of lhc .
cross sections after the application of acceptance cuts are shown in table [ xsec_visible ] .
for the signal , a value of 0.1 was chosen for @xmath33 while @xmath34 was set at zero .
the resolved @xmath35 is negligible .
l0.4 l||c| c cross section [ fb ] & @xmath36 & @xmath37 + @xmath38 & 440 & 103.7 + @xmath39 & 34.1 & 8.69 + + @xmath39 @xmath40 & 46.37 & 2.80 + @xmath41 & 15.61 & - + @xmath42 & 1.01 & - + @xmath43 & - & 0.18 + @xmath39 total & 62.99 & 2.99 + while the overall photoproduction of top quark is sensitive to the , yet unmeasured , top quark electrical charge , the @xmath31 associated photoproduction amplitudes are all proportional to the ckm matrix element @xmath44 .
preliminary studies show that the di - leptonic channel could be competitive with the inclusive channel as detailed in @xcite .
single top photoproduction could reveal phenomena beyond the standard model , and in particular flavour changing neutral currents ( fcnc)@xcite .
the effective lagrangian for this anomalous coupling can be written as @xcite : @xmath45 where @xmath46 is defined as @xmath47 , @xmath48 being the photon 4-vector and @xmath49 an arbitrary scale , conventionally taken as the top mass .
the best upper limit on @xmath33 is around 0.14 , depending on the top mass @xcite while the anomalous coupling @xmath34 has not been probed yet .
the final state is composed of a @xmath50-jet and a @xmath5 boson .
the studied topology is therefore @xmath51 .
main irreducible background processes come from @xmath52 interactions producing a @xmath5 boson and a jet mistagged as a @xmath50-jet .
r0.3 l||c cross section [ fb ] & @xmath51 + @xmath53 @xmath54 & 769.0 + @xmath39 & 144.0 + + @xmath39 @xmath55 & 56.2 + @xmath56@xmath57 & 82.8 + @xmath39 total & 139.0 + the present selection together with the assumption that no other background contribution will interfere , would lead to the expectation of a five sigma discovery just below @xmath58 of integrated luminosity .
the extracted limits on the anomalous couplings k@xmath59 and k@xmath60 are reported in table [ cl.anotop ] .
r0.45 [ cols="^,^,^ " , ]
a survey of high energy @xmath24 and @xmath52 interactions at lhc has been presented .
the large cross sections show the potential of the study of massive electroweakly interacting particles in a complementary way to the usual , parton - parton processes , at the very start of lhc . at high luminosity ,
the efficient selection of photon induced processes is conditioned by the capacity of experiments to tag forward protons by means of dedicated forward detectors . in this context ,
r&d programs such as fp420@xcite and other developments related to the instrumentation of forward regions of cms and atlas experiments , are particularly important .
budnev , i.f .
ginzburg , g.v .
meledin and v.g .
serbo , phys .
rept . *
15 * , 181 ( 1974 ) .
x. rouby , to appear in the proceedings of xliiird rencontres de moriond on qcd and hard interactions , arxiv:0805.4406 [ hep - ex ] g. abbiendi _ et al .
_ ( opal collaboration ) , phys . rev .
d * 70 * ( 2004 ) 032005 .
n. schul and k. piotrkowski , to appear in the proceedings of workshop on high energy photon collisions at the lhc , geneva , switzerland , arxiv:0806.1097 [ hep - ph ] s. ovyn , to appear in the proceedings of workshop on high energy photon collisions at the lhc , geneva , switzerland , arxiv:0806.1157 [ hep - ph ] s. ovyn and j. de favereau de jeneret to appear in the proceedings of top 2008 , international workshop on top physics , arxiv:0806.4841 [ hep - ph ] j. de favereau de jeneret and s. ovyn , to appear in the proceedings of workshop on high energy photon collisions at the lhc , geneva , switzerland , arxiv:0806.4886 [ hep - ph ] t. han and j.l .
hewett , phys .
rev . d**60 * * , ( 1999 ) 074015 s. chekanov et al .
lett . b**559 * * , ( 2003 ) 153 - 170 .
albrow et al.(fp420 r&d collaboration),arxiv:0806.0302 [ hep - ph ]
|
experimental prospects for studying high - energy photon - photon and photon - proton interactions at the lhc are discussed . assuming a typical lhc multipurpose detector , various signals and their irreducible backgrounds are presented after applying acceptance cuts .
selection strategies based on photon interaction tagging techniques are presented .
prospects are discussed for the higgs boson search , detection of susy particles and of anomalous quartic gauge couplings , as well as for the top quark physics .
|
rubbers and elastomers are highly deformable solids which have the remarkable property of preserving their volume through any deformation .
this permanent isochoricity can be incorporated into the equations of continuum mechanics through the concept of an _ internal constraint _
, here the constraint of _ incompressibility_. mathematically , the formulation of the constraint of incompressibility has led to the discovery of several exact solutions in isotropic finite elasticity , most notably to the controllable or universal solutions of rivlin and co - workers ( see for example rivlin ( 1948 ) ) . subsequently , ericksen ( 1954 ) examined the problem of finding all such solutions .
he found that there are no controllable finite deformations in isotropic _ compressible _ elasticity , except for homogeneous deformations ( ericksen 1955 ) .
the impact of that result on the theory of nonlinear elasticity was quite important and long - lasting , and for many years a palpable pessimism reigned about the possibility of finding exact solutions at all for compressible elastic materials .
then currie & hayes ( 1981 ) showed that one could obtain interesting classes of exact solutions , beyond the homogeneous universal deformations , if one restricted their attention to certain special classes of compressible materials .
a string of results about the search for exact solutions in nonlinear elasticity followed .
now a long list exists of classes of exact solutions which are universal only relative to some special strain - energy functions ( for a recent presentation of such classes see fu and ogden ( 2001 ) . )
these solutions can help us to understand the structure of the theory of nonlinear elasticity and to complement the celebrated solutions of rivlin .
in the same vein , some recent efforts focused on determining the maximal strain energy for which a certain deformation field , fixed a priori , is admissible .
this is a sort of _ inverse problem _ : find the elastic materials ( that is , the functional form of the strain - energy function ) for which a given deformation field is controllable ( that is , for which the deformation is a solution to the equilibrium equations in the absence of body forces ) .
a classical example illustrating such an approach is obtained by considering deformations of _ anti - plane shear _ type .
knowles ( 1977 ) shows that a non - trivial ( non - homogeneous ) equilibrium state of anti - plane shear is not always ( universally ) admissible , not only for compressible solids ( as expected from ericksen s result ) but also for incompressible solids ( horgan ( 1995 ) gives a survey of anti - plane shear deformations in nonlinear elasticity ) . only for a special class of incompressible materials ( inclusive of the so - called ` generalized neo - hookean materials ' )
is an anti - plane shear deformation controllable .
let us consider for example the case of an elastic material filling the annular region between two coaxial cylinders , with the following boundary value problem : hold fixed the outer cylinder and pull the inner cylinder by applying a tension in the axial direction .
it is well established that a solution to this problem , valid for every isotropic incompressible elastic solid , is obtained by assuming a priori that the deformation field is a pure axial shear .
now consider the corresponding problem for _ non - coaxial _
cylinders , thereby losing the axial symmetry . then it is clear that we can not expect the material to deform as prescribed by a pure axial shear deformation .
knowles s result tells us that now the boundary value problem can be solved with a general anti - plane deformation ( not axially symmetric ) _ only for a subclass _ of incompressible isotropic elastic materials .
of course this restriction does not mean that for a generic material it is not possible to deform the annular material as prescribed by our boundary conditions , but rather that in general , these lead to a deformation field which is more complex than an anti - plane shear . and so
, we expect _ secondary _ in - plane deformations .
the theory of non - newtonian fluid dynamics has generated a substantial literature about secondary flows , see for example fosdick & serrin ( 1973 ) . in solid mechanics
it seems that only fosdick & kao ( 1978 ) and mollica & rajagopal ( 1997 ) produced some significant and beautiful examples of secondary deformations fields for the non - coaxial cylinders problem , although this topic is clearly of fundamental importance , not only from a theoretical point of view but also for technical applications . in this paper
we consider a complex deformation field in isotropic incompressible elasticity , to point out by an explicit example the situations just evoked and to elaborate on their possible impact on solid mechanics .
our deformation field takes advantage of the radial symmetry and therefore we find it convenient to visualize it by considering an elastic cylinder .
let us imagine that a corkscrew has been driven through a cork ( the cylinder ) in a bottle .
the inside of the bottleneck is the outer rigid cylinder and the idealization of the gallery carved out by the corkscrew constitutes the inner coaxial rigid cylinder .
our first deformation is purely radial , originated from the introduction of the cork into the bottleneck and then completed when the corkscrew penetrates the cork ( a so - called _ shrink fit problem _ , which is a source of elastic residual stresses here ) .
we call @xmath0 , @xmath1 the respective inner and outer radii of the cork in the reference configuration and @xmath2 , @xmath3 their current counterparts .
then we follow with a _ simple torsion _ combined to a _
helical shear _
, in order to model pulling the cork out of the bottleneck in the presence of a contact force .
figure [ fig_deformation ] sketches this deformation .
of course we are aware of the shortcomings of our modelling with respect to the description of a ` real ' cork - pulling problem , because no cork is an infinitely long cylinder , nor is a corkscrew perfectly straight .
in addition , traditional corks made from bark are anisotropic ( honeycomb mesoscopic structure ) and possess the remarkable ( and little - known ) property of having an infinitesimal poisson ratio equal to zero , see the review article by gibson _
et al . _ ( 1981 ) .
however we note that _ polymer corks _ have appeared on the world wine market ; they are made of elastomers , for which incompressible , isotropic elasticity seems like a reasonable framework ( indeed the documentation of these synthetic wine stoppers indicates that they lengthen during the sealing process ) .
we hope that this study provides a first step toward a nonlinear alternative to the linear elasticity testing protocols presented in the international standard iso 9727 .
we also note that low - cost _ shock absorbers _ often consist of a moving metal cylinder , glued to the inner face of an elastomeric tube , whose outer face is glued to a fixed metal cylinder ( hill 1975 ) .
the plan of the paper is the following .
section 2 is devoted to the derivation of the governing equations and to a detailed description of the boundary value problem . in
3 we specialize the analysis to the _ neo - hookean strain - energy density _ , and find the corresponding exact solution .
we use it to show that it is advantageous to add a twisting moment to an axial force when extruding a cork from a bottle .
the neo - hookean strain - energy density is linear with respect to the first principal invariant of the cauchy - green strain tensor .
it is much used in finite elasticity theory , although it captures poorly the basic features of rubber behaviour ( saccomandi 2004 ) .
we thus investigate the consequence of departing from that strain - energy density .
first , in 4 we consider the _
generalized neo - hookean strain - energy density _
non - linear with respect to the first principal invariant of the cauchy - green strain tensor to show that in this case torsion is explicitly present in the solution for the axial shear displacement , but it is a second - order dependence .
next we consider in 5 the _ mooney rivlin strain - energy density _
linear with respect to the first and second principal invariants of the cauchy - green strain tensor and find that it is then also possible to obtain an exact solution to our boundary value problem .
its expression is too cumbersome to manipulate and we resort to a small parameter asymptotic expansion from the neo - hookean case .
section 6 concludes the paper with some remarks on the limitations of the semi - inverse method .
consider a long hollow cylindrical tube , composed of an isotropic incompressible nonlinearly elastic material . at rest
the tube is in the region @xmath4 where @xmath5 are the cylindrical coordinates associated with the undeformed configuration , and @xmath0 and @xmath1 are the inner and outer radii of the tube , respectively .
consider the general deformation obtained by the combination of radial dilatation , helical shear , and torsion , as @xmath6 where @xmath7 are the cylindrical coordinates in the deformed configuration , @xmath8 is the amount of torsion and @xmath9 is the stretch ratio in the @xmath10 direction
. here , @xmath11 and @xmath12 are yet unknown functions of @xmath13 only ( the classical case of pure torsion corresponds to @xmath14 , see the textbooks by ogden ( 1997 ) or by atkin and fox ( 2007 ) , for instance . ) hidden inside is the _ shrink fit deformation _
@xmath15 which is without any torsion nor helical shear ( @xmath16 ) .
the physical components of the deformation gradient @xmath17 and of its inverse @xmath18 are then @xmath19 respectively
. note that we used the incompressibility constraint in order to compute @xmath18 ; it states that @xmath20 , so that @xmath21 in our first deformation , the cylindrical tube is pressed into a cylindrical cavity with inner radius @xmath22 and outer radius @xmath23 .
it follows by integration of the equation above that @xmath24 where now @xmath25 we compute the physical components of the left cauchy - green strain tensor @xmath26 from and find its first three principal invariants @xmath27 , @xmath28 , and @xmath29 as @xmath30 and of course , @xmath31 . for a general incompressible hyperelastic
solid , the cauchy stress tensor @xmath32 is related to the strain through @xmath33 where @xmath34 is the lagrange multiplier introduced by the incompressibility constraint , @xmath35 is the strain energy density , and @xmath36 .
having computed @xmath37 from , we find that the components of @xmath32 are @xmath38 , \notag \\[2pt ] & t_{\theta \theta } = -p + 2 w_1 \left[(rg^{\prime})^2 + ( r / r)^2 + ( r\tau)^2 \right ] -2 w_2 ( r / r)^2 , \notag \\[2pt ] & t_{z z } = -p + 2 w_1 [ \lambda^2 + ( w^{\prime})^2 ] - 2 w_2 \left[(1/\lambda)^2 + ( r \tau / \lambda)^2 \right ] , \notag \\[2pt ] & t_{r \theta } = 2 w_1 ( r r^{\prime}g^{\prime } ) - 2 w_2 ( w^{\prime}\tau - g^{\prime}\lambda)r , \notag \\[2pt ] & t_{r z } = 2 w_1 ( r^{\prime}w^{\prime } ) -2 w_2 \left[r r g^{\prime}\tau - r r w^{\prime}\tau^2/\lambda - r w^{\prime}/(\lambda r ) \right ] , \notag \\[2pt ] & t_{\theta z } = 2 w_1 ( r w^{\prime}g^{\prime}+ r \lambda \tau ) + 2 w_2 ( r^{\prime}r \tau).\end{aligned}\ ] ] finally the equilibrium equations , in the absence of body forces , are : @xmath39 ; for fields depending only on the radial coordinate as here , they reduce to @xmath40 now consider the inner face of the tube : we assume that it is subject to a vertical pull , @xmath41 say
. then we can integrate the second and third equations of equilibrium @xmath42 ; we find that @xmath43 the outer face of the tube ( in contact with the glass in the cork / bottle problem ) remains fixed , so that @xmath44 say .
in addition to the axial traction applied on its inner face , the tube is subject to a resultant axial force @xmath45 ( say ) and a resultant moment @xmath46 ( say ) , @xmath47 note that the traction @xmath48 of is not arbitrary but is instead determined by the _ shrink fit pre - deformation _
, by requiring that @xmath49 when @xmath50 ( this process is detailed in the next section for the neo - hookean material . )
therefore @xmath48 is connected with the stress field experienced by the cork when it is introduced in the bottleneck . in the rest of the paper we aim at presenting results in dimensionless form . to this end
we normalize the strain - energy density @xmath51 and the cauchy stress tensor @xmath32 with respect to @xmath52 , the infinitesimal shear modulus ; hence we introduce @xmath53 and @xmath54 defined by @xmath55 similarly we introduce the following non - dimensional variables , @xmath56 so that @xmath57
. also , we find from that @xmath58 turning to our cork or shock absorber problems , we imagine that the inner metal cylinder is introduced into a pre - existing cylindrical cavity ( this precaution ensures a one - to - one correspondence of the material points between the reference and the current configurations ) . in our upcoming numerical simulations ,
we take @xmath59 so that @xmath60 ; we consider that the outer radius is shrunk by 10% : @xmath61 , and that the inner radius has doubled : @xmath62 ; finally we apply a traction whose magnitude is half the infinitesimal shear modulus : @xmath63 .
this gives @xmath64 at this point it is possible to state clearly our main observation .
a first glance at the boundary conditions , in particular at the requirements that @xmath11 be zero on the outer face of the tube , gives the expectation that @xmath65 everywhere is a solution to our boundary value problem , at least for some simple forms of the constitutive equations .
in what follows we find that for the so - called ` neo - hookean ' solids , @xmath65 is indeed a solution , whether there is a torsion @xmath66 or not .
however if the solid is not neo - hookean , then it is necessary that @xmath67 when @xmath68 , and the picture becomes more complex .
for this reason , we classify as ` purely academic ' the question : _ which is the most general strain - energy density for which it is possible to solve the above boundary value problem with @xmath65 _ ?
indeed there is no ` real world ' material whose behaviour is ever going to be described _
exactly _ by that strain - energy density ( supposing it exists ) .
instead a more pertinent issue to raise for ` real word applications ' is whether we are able to evaluate the importance of latent ( secondary ) stress fields , because they are bound to be woken up ( triggered ) by the deformation .
first we consider the special strain energy density which generates the class of neo - hookean materials , namely @xmath69 note that here and hereafter , we use the non - dimensional quantities introduced previously , from which we drop the overbar for convenience . hence the components of the ( non - dimensional ) stress field in a neo - hookean material reduce to @xmath70 substituting into we find that @xmath71 and by integration , using , that @xmath72 in figure 2a , we present a rectangle in the tube at rest
. it is delimited by @xmath73 and @xmath74 .
then it is subject to the deformation corresponding to the numerical values of . to generate figure 2b ,
we computed the resulting shape for a neo - hookean tube , using , , and . now that we know the ful deformation field , see and , we can compute @xmath75 from and deduce @xmath76 by integration of @xmath77 , with initial condition @xmath78 .
then the other field quantities follow from the rest of . in the end
we find in turn that @xmath79 \right\ } + t_{0 } , \notag \\[2pt ] & t_{\theta \theta } = t_{rr } + \left(\frac{r^2}{\lambda } + \alpha \right ) \left(\frac{1}{r^2 } + \tau^2\right ) - \frac{r^2}{\lambda ( r^2 + \alpha \lambda ) } , \notag \\[2pt ] & t_{zz } = t_{rr } + \lambda^2 \left(1 + \frac{r_1 ^ 2 ( t_0^a)^2}{r^2 } \right ) - \frac{r^2}{\lambda(r^2 + \alpha \lambda ) } , \label{ttt}\end{aligned}\ ] ] ( where we used the identity @xmath80 , see with @xmath81 ) , and that @xmath82 the constant @xmath83 is fixed by the shrink fit pre - deformation , imposing that @xmath49 when @xmath84 , or @xmath85 r \text{d}r = 0.\ ] ] using this , and , , , we find the following expressions for the resultant moment , @xmath86 and for the axial force , @xmath87 we now have a clear picture of the response of a neo - hookean solid to the deformation , with the boundary conditions of 2@xmath88 .
first we saw that here the contribution @xmath89 is not required for the azimuthal displacement , whether there is a torsion @xmath66 or not .
also , if a moment @xmath90 is applied , then the tube suffers an amount of torsion @xmath91 proportional to @xmath46 . on the other hand ,
if the tube is pulled by the application of an axial force only ( @xmath92 ) and no moment ( @xmath93 ) , then @xmath94 and no azimuthal shear occurs at all .
when we try to apply our results to the extrusion of a cork from the neck of a bottle , the following remarks seem to be relevant . from the elementary theory of coulomb friction , it is known that the pulled cork starts to move when , in modulus , the friction force exerted on the neck surface is equal to the normal force times the coefficient of static friction . in our case
this means that @xmath95 where @xmath96 is the coefficient of static friction . using and
, we find that the elements of the left handside of this inequality are @xmath97 now , our main concern is to understand if it is better to apply a moment @xmath98 when uncorking a bottle , than to pull only . to address this question
we note that the left handside of inequality increases when @xmath99 increases ; on the other hand , combining and , we have @xmath100/(2 \pi \lambda r_1 ^ 2 |\ln \eta| ) , \label{cc2bis}\ ] ] it is now clear , that for a fixed value of @xmath101 , in the case @xmath102 , it is necessary to apply an axial force whose intensity is less than the one in the case @xmath93 .
moreover , the equation above shows that @xmath103 grows linearly with @xmath45 but quadratically with @xmath46 . with respect to efficient cork - pulling
, the conclusion is that adding a twisting moment to a given pure axial force is more advantageous than solely increasing the vertical pull .
moreover , we observe that a moment is applied by using a lever and this is always more convenient from an energetic point of view .
recall that we made several simplifying assumptions to reach these results : not only infinite axial length , incompressibility , and isotropy , but also the choice of a truly special strain energy density . in the next two sections we depart from the neo - hookean model .
as a first broadening of the neo - hookean strain - energy density , we consider _ generalized neo - hookean materials _ , for which the strain - energy density is a nonlinear function of the first invariant @xmath104 only , @xmath105 say . to gain access to the cauchy stress components in this context , it suffices to take @xmath106 and @xmath107 in equations . in particular , @xmath108 , and the integrated equation of equilibrium @xmath109 gives @xmath110 .
integrating , with as an initial value , we find that @xmath111 hence , just as in the neo - hookean case , azimuthal shear can be avoided altogether , whether there is a torsion @xmath66 or not .
we are left with an equation for the axial shear , namely @xmath112 , which here reads @xmath113 obviously the same steps as those taken for neo - hookean solids may be followed here for any given strain energy density , but now by resorting to a numerical treatment .
horgan & saccomandi ( 2003@xmath114 ) show , through some specific examples of hardening generalized neo - hookean solids , how rapidly involved the analysis becomes , even when there is only helical shear and no shrink fit .
instead we simply point out some striking differences between our present situation and the neo - hookean case .
we remark that @xmath104 is of the form @xmath112 at @xmath65 that is , @xmath115 ^ 2.\ ] ] it follows that is a _ nonlinear _
differential equation for @xmath116 , in contrast with the neo - hookean case .
another contrast is that the axial shear @xmath12 is now intimately coupled to the torsion parameter @xmath66 , and that this dependence is a _ second - order effect _
( @xmath66 appears above as @xmath117 ) . a similar problem where the azimuthal shear has not been ignored , but the axial shear has been considered null i.e. @xmath118 has been recently considered by wineman ( 2005 ) .
in this section we specialize the general equations of 2 to the mooney rivlin form of the strain - energy density , which in its non - dimensional form reads @xmath119 where @xmath120 is a material parameter , distinguishing the mooney rivlin material from the neo - hookean material , and also allowing a dependence on the second principal strain invariant @xmath121 , in contrast to the generalized neo - hookean solids of the previous section .
then the integrated equations of equilibrium read @xmath122 first we ask ourselves if it is possible to avoid torsion during the pulling of the inner face . taking @xmath94 above gives @xmath123 it follows that here it is indeed possible to solve our boundary value problem .
we find @xmath124 , \qquad g = 0.\ ] ] however if @xmath91 , then it is necessary that @xmath67 , otherwise @xmath109 gives @xmath125 while @xmath112 gives @xmath126 , a contradiction .
this constitutes the first departure from the neo - hookean and generalized neo - hookean behaviours : _ torsion ( @xmath127 ) is necessarily accompanied by azimuthal shear ( @xmath67)_. in the case @xmath91 , we introduce the function @xmath128 defined as @xmath129 ( recall that @xmath130 is given explicitly in . ) we then solve the system for @xmath131 and @xmath132 as @xmath133 making clear the link between @xmath11 and @xmath66 .
thus for the mooney rivlin material , the azimuthal shear @xmath11 is a _ latent _ mode of deformation ; it is _ woken up _ by any amount of torsion @xmath66 .
recall that at first sight , the azimuthal shear component of the deformation seemed inessential to satisfy the boundary conditions , especially in view of the boundary condition @xmath134 .
however , a non - zero @xmath135 term in the constitutive equation clearly couples the effects of a torsion and of an azimuthal shear , as displayed explicitly by the presence of @xmath66 in the expression for @xmath136 above .
it is perfectly possible to integrate equations in the general case , but to save space we do not reproduce the resulting long expressions . with them , we generated the deformation field picture of figure 3a and figure 3b .
there we took the numerical values of for @xmath137 , @xmath138 , @xmath101 ; we took a mooney
rivlin solid with @xmath139 ; we imposed a torsion of amount @xmath140 ; and we looked at the deformation field in the plane @xmath141 ( reference configuration ) and @xmath142 ( current configuration ) .
although the secondary fields appear to be slight in the picture , they are nonetheless truly present and can not be neglected . to show this we consider a perturbation method to obtain simpler solutions and to understand the effect of the coupling , by taking @xmath143 small . then integrating
, we find at first order that @xmath144 , \notag \\ & \dfrac{g}{r_1 t_0^a } \simeq \lambda^2 \tau m \ln { r}.\end{aligned}\ ] ] hence , the secondary field @xmath11 exists even for a nearly neo - hookean solid ( if @xmath143 is small , then @xmath11 of order @xmath143 . ) interestingly we also note that the azimuthal shear @xmath11 in varies in a homogeneous and linear manner with respect to the torsion parameter @xmath66 and in a quadratic manner with respect to the axial stretch @xmath138 , showing that that the presence of this secondary deformation field can not be neglected when the effects of the prestress and of the torsion are both taken into account .
to complete the picture , we use the first - order approximations @xmath145 to obtain the stress field as @xmath146 \left ( \lambda r_{1}t_{0}^{a}\right ) ^{2}/r^{2}\right\ } , \notag \\[2pt ] & t_{\theta \theta } \simeq -p+\left ( 1-m\right ) \left [ ( r / r)^{2}+(r\tau ) ^{2}\right ] -m(r / r)^{2 } , \notag \\[2pt ] & t_{zz}\simeq -p+\left(\lambda t_0^a r_1\right)^2 \left[\left(1 + 2m\lambda^2\right)\frac{1}{r^2 } - \frac{2}{r}\left(\frac{\tau^2r^2}{r } + \frac{r^2}{r^3}+\frac{\lambda^2}{r}-\frac{1}{2r}\right)m \right ] \notag \\ & \hspace{2.02cm}+\left ( 1-m\right ) \lambda ^{2 } -m\left [ ( 1/\lambda ) ^{2}+(r\tau /\lambda ) ^{2}\right ] , \notag \\[2pt ] & t_{r\theta } \simeq rr^{\prime } g^{\prime } -m\lambda r_{1}t_{0}^{a}\tau , \notag \\[2pt ] & t_{rz}\simeq \left ( 1-m\right ) ( r^{\prime } w^{\prime } ) + m\lambda r_{1}t_{0}^{a}\left [ r r\tau ^{2}/\lambda + r/(\lambda r)\right ] /r , \notag \\[2pt ] & t_{\theta z}\simeq \left ( 1-m\right ) r\lambda \tau + \lambda rr_{1}t_{0}^{a}g^{\prime } /r+m(r^{\prime }
r\tau ) .\end{aligned}\ ] ] using this stress field it is straightforward , but long and cumbersome , to derive the analogue for a mooney
rivlin solid with a small @xmath143 of relation ( which was established for neo - hookean solids .
) however nothing truly new is gained from these complex formulas with respect to the simple neo - hookean case , and we do not pursue this aspect any further .
in non - newtonian fluid mechanics and in turbulence theory , the existence of shear - induced normal stresses on planes transverse to the direction of shear is at the root of some important phenomena occurring in the flow of fluid down pipes of non - circular cross section ( fosdick & serrin 1973 ) . in other words ,
pure parallel flows in tubes without axial symmetry are possible only when we consider the classical theory of navier - stokes equations or the linear theory of turbulence or tubes of circular cross section . in nonlinear elasticity theory ,
similar phenomena are reported .
hence fosdick & kao ( 1978 ) and mollica & rajagopal ( 1997 ) show that for general isotropic incompressible materials , an anti - plane shear deformation of a cylinder with non axial - symmetric cross section causes a secondary in - plane deformation field , because of normal stress differences .
horgan & saccomandi ( 2003@xmath147 ) give a detailed discussion of how the anti - plane shear deformation field couples with the in - plane deformation field in a generalized neo - hookean solid .
the appearance of what we called here _ latent deformations _ is quite general and common .
for example it is known in compressible nonlinear elasticity that pure torsion is possible only in a special class of materials , but we know that torsion plus a radial displacement is possible in all compressible isotropic elastic materials ( polignone & horgan 1991 ) ( here we signal that ` possible in all materials ' is not equivalent to ` universal ' , because the corresponding radial deformation differs from one material to another . ) in this paper we give an example where axial symmetry holds , where the boundary conditions suggest that an axial shear deformation field is sufficient to solve the boundary value problem , and where nevertheless , the normal stress difference wakes up a latent azimuthal shear deformation . moreover , because we are able to find some explicit exact solutions by some perturbation techniques , we are able to evaluate the _ importance _ of the latent deformation . indeed , we show that if a certain constitutive parameter @xmath143 ( distinguishing a neo - hookean solid from a mooney
rivlin solid ) is zero or if the torsion parameter @xmath66 is zero , then the solution to the boundary value problem can be found only in terms of the axial shear deformation field ; if these two parameters are not zero even if they are small then the latent mode of deformation is quantitatively appreciable .
in conclusion we suggest that it is not crucial to determine the class of materials for which a given deformation field is possible .
rather it is crucial to classify all the latent deformations associated with a given deformation field in such a way that this field is controllable for the entire class of materials . indeed , no _ real _ material , even when we accept that its mechanical behaviour is purely elastic , is ever going to be described exactly by a special choice of strain - energy . looking for special classes of materials for which special deformations fields are admissible may mislead us in our understanding of the nonlinear mechanical behaviour of materials .
to finish the paper on a light note , we evoke a classic wine party dilemma : which kind of corkscrew system requires the least effort to uncork a bottle ?
figure 4 sketches the two working principles commonly found in commercial corkscrews .
the most common ( on the left ) relies on pulling only ( directly or through levers ) and the other type ( on the right ) relies on a combination of pulling and twisting .
notwithstanding the shortcomings of this paper s modelling with respect to an actual uncorking , the authors are confident that they have provided a scientific argument to those wine amateurs who favour the second type of corkscrews over the first .
horgan , c. o. , saccomandi , g. 2003@xmath147 coupling of anti - plane shear deformations with plane deformations in generalized neo - hookean isotropic , incompressible , hyperelastic materials .
_ j. elast . _ * 73 * , 221235 .
|
in an attempt to describe cork - pulling , we model a cork as an incompressible rubber - like material and consider that it is subject to a helical shear deformation superimposed onto a shrink fit and a simple torsion . it turns out that this deformation field provides an insight into the possible appearance of secondary deformation fields for special classes of materials .
we also find that these latent deformation fields are woken up by normal stress differences .
we present some explicit examples based on the neo - hookean , the generalized neo - hookean , and the mooney
rivlin forms of the strain - energy density . using the simple exact solution found in the neo - hookean case , we conjecture that it is advantageous to accompany the usual vertical axial force by a twisting moment , in order to extrude a cork from the neck of a bottle efficiently
. then we analyse departures from the neo - hookean behaviour by exact and by asymptotic analyses . in that process
we are able to give an elegant and analytic example of secondary ( or latent ) deformations in the framework of nonlinear elasticity .
|
the @xmath0-dibaryon , predicted by r. l. jaffe in 34 years ago @xcite .
is one of most famous candidates of exotic - hadron .
this prediction was based on the observations that the quark exclusion can be completely avoided due to the essentially flavor - singlet(@xmath3 ) nature of @xmath0-dibaryon , together with the large attraction from one - gluon - exchange interaction between quarks suggested in the quark model @xcite .
search for the @xmath0-dibaryon is one of the most challenging theoretical and experimental problems in the physics of strong interaction and quantum chromodynamics ( qcd ) .
still it is not clear whether there exists the @xmath0-dibaryon in nature .
although deeply bound @xmath0-dibaryon with the binding energy @xmath4 mev from the @xmath5 threshold has been ruled out by the discovery of the double @xmath6 hypernuclei , @xmath7he @xcite , there still remains a possibility of a shallow bound state or a resonance in this channel @xcite .
several lattice qcd calculations on @xmath0 have been reported as reviewed in @xcite ( see also recent works @xcite ) .
however , there is a serious problem in studying dibaryons on the lattice : to accommodate two baryons inside the lattice volume , the spatial lattice size @xmath8 should be large enough .
once @xmath8 becomes large , however , energy levels of two baryons become dense , so that isolation of the ground state from the excited states is very difficult ( quite a large imaginary - time @xmath9 is required ) .
all the previous works on dibaryons more or less face this issue . in this paper , we employ our original approach distinct from one in previous studies , and search the @xmath0-dibaryon in lattice qcd @xcite . because the flavor @xmath1 breaking complicate calculation , we start with the flavor @xmath1 limit .
to study interaction or property of a hadron systems in lattice qcd , we utilize the nambu - bethe - salpeter ( nbs ) wave function and the corresponding potential , instead of the energy eigenstate . within the non - relativistic approximation ,
potential of a two - body interaction can be defined from the nbs wave function @xmath10 through the schrdinger equation in the euclidean space - time @xcite , @xmath11 \psi(\vec r , t ) + \int \!\ !
d^3 \vec r ' \ , u(\vec r , \vec r ' ) \ , \psi(\vec r ' , t ) = - \frac{\partial}{\partial t } \psi(\vec r , t)\ ] ] where a non - local but energy - independent potential @xmath12 is introduced , and @xmath13 is the reduced mass of the system with masses @xmath14 and @xmath15 .
the potential @xmath12 can be determined by solving this equation by using data of @xmath10 measured on the lattice . since
only a single data of @xmath10 is available from one simulation , we can not construct whole the potential @xmath12 , but can obtain the leading term of the velocity expansion of it .
we describe the potential @xmath16 . as shown in ref .
@xcite , one obtains a sink - time independent @xmath16 even without isolating energy eigenstates .
it is also shown that we can obtain a volume independent @xmath16 , if we setup the volume larger than the minimum to accommodate the interaction , e.g. @xmath17 fm for the nucleon - nucleon interaction at @xmath18 mev .
once we obtain a volume independent potential , any observable of the system , such as binding energy and scattering phase shift , can be obtained by solving the schrdinger equation in the infinite volume .
note that , in contrast to the conventional lscher s method @xcite , we do not calculate the energy shift of two hadrons at finite @xmath8 to access the observables in the infinite volume .
finally , the nbs wave function can be obtained by summing up the hadron four - pint function @xmath19 which can be evaluated numerically in lattice qcd .
.[tbl : lattice]summary of lattice parameters and hadron masses . [ cols="^,^,^,^,^,^,^",options="header " , ] for dynamical lattice qcd simulations , we have generated five ensembles of gauge configuration on @xmath20 lattice . we have employed the renormalization group improved iwasaki gauge action with the coupling constant @xmath21 , and the non - perturbatively @xmath22 improved wilson quark action .
quark propagators are calculated for the quark wall source with the dirichlet boundary condition in the temporal direction . the point type octet - baryon field operators are used at sink .
the sink @xmath2 operator is projected to the @xmath23 representation of the cubic group , so that the nbs wave function is dominated by the s - wave component . for the time derivative of the nbs wave function
, we adopt the symmetric difference on the lattice . to enhance signal over noise , we average 16 measurements for each configuration , together with the average between forward and backward propagation in time .
statistical errors are evaluated by the jackknife method .
lattice parameters such as lattice spacing @xmath24 , lattice size @xmath8 , the hopping parameter @xmath25 , the number of gauge configurations @xmath26 , together with the hadron masses are summarized in table [ tbl : lattice ] .
the left panel of fig.[fig1 ] shows the flavor - singlet @xmath2 potential at each quark mass , extracted from our numerical simulation of full qcd .
one see that it has an attractive core " and its range is well localized in space .
this entirely attractive potential prove that the quark model prediction to the flavor - singlet @xmath2 channel is essentially correct .
one can see that the long range part of the attraction tends to increase as the quark mass decreases . by solving the schrdinger equation involving this potential
, we can obtain any information of the system , for instance , energy of the ground state .
it turns out that there is only one bound state in each quark mass @xcite .
the right panel of fig.[fig1 ] shows the energy and the root - mean - square ( rms ) distance of the bound state .
the result for @xmath27 mev is only preliminary because of insufficient number of independent gauge configuration .
these bound state correspond to the @xmath0-dibaryon predicted by jaffe .
our lattice qcd calculation shows that a stable @xmath0-dibaryon certainly exists in the flavor @xmath1 limit world with the binding energy of 2050 mev for the present quark masses . despite
that the attractive potential become stronger as quark mass decreases , the resultant binding energies of the @xmath0-dibaryon decrees in the present range of the quark masses .
this is due to the fact that the increase of the attraction toward the lighter quark mass is compensated by the increase of the kinetic energy for the lighter baryon mass .
the rms distance @xmath28 is a measure of spacial distribution of baryonic matter in the @xmath0-dibaryon , which can be compared to the point matter rms distance of the deuteron in nature , @xmath29 fm .
one may be able to get feeling of the @xmath0-dibaryon from this comparison : @xmath0-dibaryon is much compact compared to usual nuclei .
when we solve the schrdinger equation , we use the potential expressed in terms of an analytic function fitted to the data .
error from the choice for the analytic function to fit is negligible ( less than few % ) .
small systematic error arise from choice for the sink - time - slice .
the final result of the binding energy of the @xmath0-dibaryon in the @xmath1 limit described as @xmath30 becomes , for example , @xmath31 with statistical error ( first ) and systematic error ( second ) .
one may think such a deeply bound @xmath0-dibaryon is ridiculous or ruled out by the discovery of the double @xmath6 nuclei .
the present @xmath32 should be interpreted as the binding energy from the @xmath2 threshold averaged in the @xmath33 sector .
the three @xmath2 thresholds in this sector split largely in the real world with the flavor @xmath1 breaking .
therefore , we expect that the binding energy of the @xmath0-dibaryon in the real world , measured from the @xmath5 threshold , is much smaller than the present value or even @xmath0-dibaryon goes to above the @xmath5 threshold . to make a definite conclusion on this point , however , we need ( 2 + 1)-flavor lattice qcd simulations and the analysis in @xmath34 coupled channel .
study along this direction is in progress @xcite .
we thank k .-
ishikawa and pacs - cs group for providing their ddhmc / phmc code @xcite , and authors and maintainer of cps++ @xcite for a modified version of it used in this paper .
numerical computations of this work have been carried out at univ . of tsukuba supercomputer system ( t2k ) .
this research is supported in part by mext grant - in - aid for scientific research on innovative areas(no.2004:20105001 , 20105003 ) .
s. r. beane _ et al . _
[ nplqcd collaboration ] , phys .
lett . * 106 * , 162001 ( 2011 ) [ arxiv:1012.3812 [ hep - lat ] ] .
t. inoue _ et al .
[ hal qcd collaboration ] , phys . rev .
* 106 * , 162002 ( 2011 ) [ arxiv:1012.5928 [ hep - lat ] ] .
n. ishii , s. aoki and t. hatsuda , phys .
lett . * 99 * , 022001 ( 2007 ) [ arxiv : nucl - th/0611096 ] .
h. nemura , n. ishii , s. aoki and t. hatsuda , phys .
b * 673 * , 136 ( 2009 ) [ arxiv:0806.1094 [ nucl - th ] ] .
s. aoki , t. hatsuda and n. ishii , prog .
phys . *
123 * , 89 ( 2010 ) [ arxiv:0909.5585 [ hep - lat ] ] .
|
existence of the @xmath0-dibaryon in the flavor @xmath1 symmetric limit is studied by full qcd simulations on the lattice , in the approach recently developed for the baryon - baryon ( @xmath2 ) interactions .
potential of the flavor - singlet @xmath2 channel is derived from the nambu - bethe - salpeter wave function , and a bound @xmath0-dibaryon is discovered from it , with the binding energy of 2050 mev for the pseudo - scalar meson mass of 4691171 mev .
address = nihon university , college of bioresource sciences , kanagawa 1866 , japan
|
topological quantum computation@xcite is immune to certain types of noise .
it has attracted much attention since it was proposed and becomes active again in recent years due to the progress in experiments .
the gates used for quantum computation can be conducted by creating quasi - particles , braiding them and measuring their states .
two well - known types of quasi - particles are fibonacci anyons and ising anyons ( majorana bound states ) .
the former are capable of offering universal topological quantum computation , while the latter can not form an universal set of gates by braiding operation only .
so the non - topologically protected gates have to be introduced in the computation .
this type of gate always requires coupling between majorana bound states ( cmbs ) .
although there is great progress in theories @xcite and experiments @xcite , the question how to couple two mbss is barely explored .
recently , schmidt and his co - workers @xcite presented proposals to couple mbss by putting the system into a microwave cavity , where the microwave field can effectively drive population transfer between mbss .
the authors found that if the microwave frequency approaches the band gap of the topologically trivial region , the coupling between the mbss is exponentially enhanced .
in other words , by modulating microwave frequency or changing number of photons in the cavity , the cmbs can be controlled .
inspiring by this idea , in this work , we study how the central chain affects cmbs .
we realize the universal majorana qubit rotation ( umqr ) by controlling the tunable local gate voltage ( tlgv ) . the paper is organized as follows . in sec .
[ ii ] we first briefly review the kitaev model and calculate cmbs in presence of pairing coupling of the central chain itself . in sec .
[ iii ] we explore the performance of cmbs when there exist pairing couplings at the boundaries . in sec .
[ iv ] we extend our calculations to the central chain with long - range interactions . in sec .
[ v ] we develop a scheme to control cmbs by modulating the amplitude or the frequency of the driving field .
[ vi ] is devoted to discussions and conclusions .
we here consider an inhomogeneous kitaev chain @xcite which can be divided into three homogeneous parts .
the total hamiltonian of the whole chain reads @xmath0 where @xmath1 ( @xmath2 ) denotes the hamiltonian of left ( right ) chain with sites from @xmath3 ( @xmath4 ) to @xmath5 ( @xmath6 ) , and @xmath7 represents the hamiltonian of central chain with sites from 0 to @xmath8 .
@xmath9 ( @xmath10 ) denotes the hamiltonian of the coupling between the left ( right ) chain and the central chain . to be specific
the hamiltonian of the three homogeneous kitaev chains can be expressed as @xmath11 where @xmath12 and @xmath13 are the spinless fermion creation and annihilation operators at site @xmath14 with chemical potential @xmath15 ( @xmath16 ) . @xmath17 and
@xmath18 label the beginning and end sites of the left , central , and right chain , respectively . @xmath19 and @xmath20 are the hopping and pairing amplitudes , respectively . since the phase of the pairing amplitude can be removed from the hamiltonian by a gauge transformation , we here and hereafter assume both @xmath19 and @xmath21 to be real . physically , the kitaev model can be exactly mapped into a spin-@xmath22 chain with xy interactions by jordan - wigner transformation @xcite and can be realized in the semiconductor nanowire proximity coupling to a @xmath23-wave superconductor @xcite .
this paradigm model shows plentiful fascinating topological properties and there are two distinct phases , i.e. , the topologically trivial and nontrivial phases .
the critical points lie at @xmath24 and @xmath25 @xcite . by choosing @xmath26 ( the chemical potential can be modulated by tlgv @xmath27 shown in fig .
[ fig:02 ] ) and @xmath28 for the left and right chains , both chains are in the topologically nontrivial phase . as a result there exist mbss at the ends of both chains , depicted by the stars in fig .
[ fig:02 ] .
one can define majorana operators @xmath29 after substituting them into eq.([2a ] ) , the hamiltonian of left ( right ) chain becomes @xmath30 clearly , the majorana operators @xmath31 and @xmath32 are coupled with different sites .
in particular the majorana operators @xmath33 , @xmath34 , @xmath35 , and @xmath36 do not appear in eq.([4a ] ) , which indicate there are four mbss in our system .
since the four mbss are decoupled to the remaining sites and only locate at the end of both left and right chains , we can ignore the hamiltonian of the remaining sites when we study cmbs . , and the cmbs can be modulated by the voltage @xmath37 . ]
note that to couple mbss one should choose the parameters @xmath38 to make sure the central chain is in the topologically trivial phase . otherwise the mbss @xmath39 and @xmath40 would disappear in this inhomogeneous chain . as we just consider the nearest - neighbour interactions in eq.([2a ] ) , only the mbss @xmath39 and @xmath40 can be coupled to the central chain .
so the hamiltonian of interest is given by @xcite @xmath41 in order to obtain the energy spectrum of the central chain , it is convenient to carry out fourier transform @xmath42 by imposing periodic boundary condition ( we have set the lattice spacing to be unit and fourier transform will be precise when the number of sites is very larege , @xmath43 ) .
the validity of fourier transform is also verified by the numerical calculations in fig .
[ fig:03 ] .
the hamiltonian of the central chain becomes @xmath44 one can rewrite the above hamiltonian in a normal bogoliubov - de gennes ( bdg ) form by defining a two component operator @xmath45 $ ] , i.e. , @xmath46 where @xmath47 and @xmath48 .
consequently the hamiltonian is simplified as ( up to a constant ) @xmath49 where the quasi - particle operator @xmath50 is defined as @xmath51 and the energy spectrum of the central chain is given by @xmath52 , @xmath53 , @xmath54 . using the above notations
, we can rewrite the hamiltonian of interest more explicitly as @xmath55 \nonumber\\ & & -\sum_{k}\frac{t_2}{2}[e^{-ikn } ( u_k - v_k)b_{k}^{\dag}+e^{ikn}(u_k^{*}-v_k^{*})b_{k}]\gamma_{n+1 } ' .
\nonumber\\\end{aligned}\ ] ] since we are particularly interested in the low - energy dynamics of system , we apply schrieffer - wolff transformation @xcite to eliminate high - energy spectrum .
we denote the first term in eq.([11 ] ) by @xmath56 and the remaining terms by @xmath57 . by choosing the unitary transformation @xmath58 to satisfy the relation @xmath59=\mathcal{h}_1 $ ] , one has @xmath60 making use of the baker - hausdorff formula @xmath61+\frac{1}{2}[[h_i,\mathcal{s}],\mathcal{s}]+\cdot\cdot\cdot,\end{aligned}\ ] ] and keeping the terms up to first order , we obtain an effective hamiltonian between the adjacent mbss @xmath39 and @xmath40 , i.e. , @xmath62 here @xmath63 is the coherence length and @xmath64 .
eq.([12 ] ) demonstrates that the effective cmbs can be induced by the central chain .
physically , @xmath65 represents an amplitude for the mbss to tunnel across the central chain .
this is a virtual co - tunneling process since there exists energy gap in the central chain and real electrons and holes can not tunnel from one mbs to another .
in addition the coupling strength @xmath65 depends exponentially on the length @xmath8 , which is the reason why the mbss are topologically protected . in fig .
[ fig:03 ] , we plot the coupling strength @xmath65 versus the chemical potential @xmath66 and the pairing amplitude @xmath67 of the central chain
. one can find that the analytical solutions are in agreement with the numerical solutions when the chemical potential is much larger than the hopping amplitude .
it is also readily observed that the coupling strength @xmath65 decreases with the increasing of chemical potential @xmath66 and is enhanced by the pairing amplitude @xmath67 of the central chain , reminiscent of the results where the majorana qubit setup is placed in a microwave cavity @xcite .
physically , it originates from the fact that the chemical potential associates with the energy gap , and the pairing amplitude is related to the feature of the energy spectrum ( cf .
@xmath68 in eq.([4 ] ) ) . as we know , a larger chemical potential @xmath66 gives rise to a larger energy gap .
the larger energy gap would prohibit the electron co - tunneling , leading to a weak cmbs .
on the other hand , a larger pairing amplitude @xmath67 gives rise to wider of energy spectrum , making more quasi - particles participate in the cotunneling process .
as a result the coherence length @xmath63 also be large , which can be found in fig .
[ fig:03](d ) .
therefore the cmbs gets strong with large pairing amplitude @xmath67 . as a function of the chemical potential @xmath66 with the different number of sites .
( a ) @xmath69 .
( b ) @xmath70 .
( c ) @xmath71 .
the circle and solid lines are the exact numerical solutions of the whole kitaev chain ( see eq.([9 ] ) ) with and without periodic boundary condition , respectively .
it shows that it almost makes no difference by imposing periodic boundary condition .
the dot line is the approximate analytical solutions ( see eq.([12 ] ) ) .
all parameters are in units of hopping amplitude , i.e. , @xmath72 .
( d ) the coherence length versus the pairing amplitude with @xmath73 . ]
in this section we will study cmbs when the central chain exists the pairing coupling at the boundaries .
the hamiltonian of interest then becomes @xmath74 where @xmath75 is the same as in eq.([10 ] ) . by using the majorana operators representation in eq.([3a ] ) , the hamiltonian can be written as , @xmath76 the second term in the hamiltonian @xmath9 @xmath77 can be ignored if we only consider cmbs , but this term affects the spatial distribution of the mbss @xmath39 and @xmath40 .
it is interesting to find that the pairing coupling is equivalent to adjust the effective hopping coupling between the central chain and the left ( right ) chain with an effective hopping @xmath78 .
[ fig:04 ] demonstrates the coupling strength @xmath65 as a function of pairing amplitude in the central chain .
it can be observed that the analytical solutions are in agreement with the numerical solutions when @xmath66 is large enough , and the coupling strength @xmath65 can reach an relatively high value even though the central chain is long .
note that cmbs would in turn modify the spatial distribution of the mbss @xmath39 and @xmath40 . especially , the stronger the coupling strength is , the larger modification the spatial distribution will be .
so the coupling strength should not be very large for this consideration . noting that it would cost long time to perform the operation @xmath79 ( cf .
eq.([21 ] ) ) when the coupling strength is too weak , one should trade off those factors to choose an reasonable coupling strength to implement universal majorana qubit rotation ( umqr ) . as a function of the pairing amplitude @xmath67 with the different number of sites .
( a ) @xmath69 .
( b ) @xmath70 .
( c ) @xmath71 .
the solid line and dot line denote the exact numerical solutions ( without periodic boundary condition ) and the approximate analytical solutions , respectively .
the hopping amplitude @xmath80 are revised as @xmath81 .
all parameters are chosen in units of hopping amplitude in the central chain .
@xmath82 ( d ) the exact numerical solutions of coupling strength @xmath65 versus the pairing amplitude @xmath67 when @xmath83 .
@xmath84 . ]
in this section we turn to study cmbs mediated by trivial chains with long - range interactions .
the hamiltonian is described by @xmath85 that is , the next - to - nearest neighbor interactions are added in the central chain . with the help of raman laser
, this model can be realized by the system of trapped fermi atoms in optical lattice with the zigzag structure coupling to a 3d bec reservoir , as shown in fig .
[ fig:05 ] .
the relative strength of hopping amplitudes @xmath86 and @xmath87 can be modulated by changing the zigzag geometry . for details , we refer the readers to @xcite . the hamiltonian of the hopping coupling between the left ( right ) chain and the central chain becomes @xmath88 following the similar derivation procedures in eqs.([10])-([12 ] ) , one can estimate the effective hamiltonian of cmbs , @xmath89 where @xmath90 ( @xmath91 ) is the root of the quartic equation @xmath92 . obviously , the hamiltonian returns to eq .
( [ 12 ] ) if one sets @xmath93 .
[ fig:06 ] depicts the relation between the coupling strength and the parameters of the central chain .
one observes that the analytical solutions are in well agreement with the numerical solutions when @xmath94 , i.e. , @xmath95 . in the presence of long - range interactions , the coupling strength @xmath65 is also enhanced since the value of coupling strengh @xmath65 takes minimum when @xmath93 , as shown in fig .
[ fig:06](b ) .
it is not surprising since there are multiple channels for electrons co - tunneling in the long - range interactions , i.e. , the next - to - nearest hopping amplitude also makes contributions to the electrons co - tunneling process . as a function of the chemical potential @xmath66 . @xmath70 . @xmath96 . @xmath97 . @xmath98 .
the solid line and dot line denote the exact numerical solutions and the approximate analytical solutions , respectively .
( b ) the coupling strength @xmath65 as a function of the hopping amplitude and pairing amplitude . @xmath83
. @xmath100 . ]
as we have shown , cmbs is closely related to the chemical potential , the hopping amplitude , and the pairing amplitude of the central chain ( cf .
( [ 12 ] ) ) .
this implies the coupling strength @xmath65 can be manipulated through controlling those parameters .
considering the hopping amplitude and the pairing amplitude can not be easily manipulated in practice , we first explore the relation between the coupling strength @xmath65 and the chemical potential @xmath66 .
we do not plot directly the dependence of @xmath65 on @xmath66 .
instead , we show the dependence of rabi oscillation between two mbss on @xmath66 .
this rabi oscillation can show not only how the coupling strength changes with the chemical potential ( since distinct coupling strengths can reflect on distinct rabi frequencies , cf .
eq.([21 ] ) ) , but also how well the majorana qubit works . to show the rabi oscillation , we define the bdg hamiltonian @xmath101 of our system as follows .
@xmath102 , where @xmath103^{t}$ ] . by this definition
, we can solve the following equation @xcite @xmath104 with the initial condition @xmath105 .
the label @xmath106 denotes the vector ( in basis @xmath107 ) with components @xmath108 for @xmath109 , and so on .
@xmath110 denotes the total number of sites . in the heisenberg picture ,
@xmath111 stands for the initial operator @xmath112 , which is the left mbs in vicinity of the central chain . with these notations
, @xmath113 clearly represents the operator evolving at time @xmath114 , and @xmath115 defines the distance between @xmath113 and @xmath111 .
[ fig:07 ] shows the distance as a function of evolution time .
if there is no cmbs , the rabi oscillation can not appear in the system , as depicted in fig . [
fig:07](a ) . in fig .
[ fig:07](b)-(d ) , different chemical potentials @xmath66 result in different rabi frequencies .
when the chemical potential is large , the coupling strength would be small , leading to a small rabi frequency .
that is , the rabi frequency increases with the decreasing of the chemical potential .
similarly , the coupling strength can be manipulated by changing the pairing amplitude , as the blue - dark lines show in fig .
[ fig:08 ] . besides
, the coupling strength can be modulated purely by the phase of pairing amplitude , e.g. , refs .
@xcite . as a function of evolution time .
the system is in the mbs @xmath39 initially .
the other parameters are @xmath116 , @xmath117 , @xmath118 , @xmath119 , @xmath120 , @xmath70 .
( a ) @xmath121 .
( b ) @xmath73 .
( c ) @xmath122 .
( d ) @xmath83 . ] as a function of evolution time .
the system is in the mbs @xmath39 initially .
the other parameters are @xmath117 , @xmath118 , @xmath83 , @xmath120 , @xmath70 .
( a ) @xmath123 .
( b ) @xmath124 .
( c ) @xmath119 .
( d ) @xmath125 .
the blue - dark lines are the exact numerical results obtained by eq .
( [ 9 ] ) , and the green - grey lines are plotted by the hamiltonian given in eq .
( [ 19 ] ) .
the effective pairing amplitude are the same for two lines in each panels . ]
since the pairing amplitude is inherently determined by the property of superconductors , it may be difficult to directly modulate the pairing amplitude with the current techniques . in following
we show that this goal can be reached by driving the central chain with periodic field .
the hamiltonian of the periodic driving field reads @xmath126 where @xmath127 and @xmath128 are the amplitude and the frequency of the driving field , respectively . in an realistic situation , this driving field can be achieved by applying an external ac electric potential to tlgv , since the on - site chemical potential can be modulated by tlgv . to get an effective hamiltonian
, we work in an rotation frame defined by the unitary transformation , @xmath129 . the effective hamiltonian of the whole chain becomes , @xmath130u(t)-iu^{\dag}(t)\dot{u}(t ) , \nonumber\\ & = & h_{l}+h_{r}+h_{c}'+h_{lc}'+h_{rc } ' , \nonumber\\ h_{c}'&=&\sum_{n=0}^{n}\mu_{c } a_{n}^{\dag}a_{n}-\sum_{n=0}^{n-1}(\frac{t_{c}}{2}a_{n}^{\dag}a_{n+1 } \nonumber\\ & & + e^{i\frac{2\mu_0}{\omega}\sin(\omega t)}\frac{\delta_{c}}{2 } a_{n}^{\dag}a_{n+1}^{\dag}+h.c . ) , \nonumber\\ h_{lc}'&=&-e^{-i\frac{\mu_0}{\omega}\sin(\omega t)}\frac{t_1}{2}a_{-1}^{\dag}a_0+h.c .
, \nonumber\\ h_{rc}'&=&-e^{-i\frac{\mu_0}{\omega}\sin(\omega t)}\frac{t_2}{2}a_{n+1}^{\dag}a_n+h.c.,\end{aligned}\ ] ] where @xmath131 and @xmath132 are invariant under this rotation . by making use of the identity @xmath133 with the @xmath14-order bessel function @xmath134 ,
the time - dependent effective hamiltonian in the high - frequency limit ( i.e. , @xmath135 ) becomes @xmath136 under this approximation , the pairing amplitude of the central chain is modulated by the amplitude and the frequency of the driving field through the zero - order bessel function , i.e. , @xmath137 .
we plot the time evolution of the mbss by driving field in fig .
[ fig:08 ] .
the green - grey lines show that the rabi frequency in the presence of the driving field is a bit different from the exact results .
this originates from the fact that the effective hopping amplitude is also changed by the driving field ( see @xmath138 and @xmath139 in eq.([22 ] ) ) , rendering the correction of rabi frequency .
we would like to address that , this correction does not make difference for umqr as it only change the time to complete the operation .
interestingly , we can also modulate the effective hopping amplitude at the boundaries by the driving field to control cmbs .
it is demonstrated in fig .
[ fig:09](a ) that the adjacent mbss @xmath39 and @xmath40 can be decoupled by making the effective hopping amplitude at boundaries ( @xmath138 and @xmath139 ) vanish , i.e. , setting @xmath140 .
otherwise , cmbs can be really induced , manifesting in the rabi oscillation shown in fig .
[ fig:09](b)-(d ) . as a function of time .
the mbs @xmath39 locates at the site @xmath5 initially .
( a ) @xmath140 .
( b ) @xmath141 .
( c ) @xmath142 .
( d ) @xmath143 .
the other physical parameters are the same as in fig .
[ fig:08 ] .
the three part of kitaev chain are decoupled from each other when @xmath140 , and there exists rabi oscillation between the adjacent mbss @xmath39 and @xmath40 when @xmath144 , where the mbss @xmath40 locates at site @xmath145 . ]
as well known , an ordinary spinless fermion can be used to encode a logical qubit because it can span a two - dimensional hilbert space ( occupy or empty ) .
however it is not true for mbss since the operators satisfy @xmath146 and @xmath147 . by recombining the operators @xmath148 , one can use two mbss to construct a dirac fermion , e.g. , @xmath149 and @xmath150 in fig .
[ fig:02 ] .
it seems that a logical qubit can be encoded by two mbss . nevertheless , for the system with parity conservation ( calculated through the dirac fermions formed by mbss )
, the coherent superposition of mbss with different parities is prohibited .
therefore it can not encode a logical qubit by two mbss in the parity conservation system . to guarantee the two computational bases having the same parity , four mbss ( see @xmath151 , @xmath39 , @xmath40 , and @xmath152 in fig . [ fig:02 ] )
is necessary to construct a logical qubit .
e. g. , we can construct the majonara - based qubit in the odd parity subspace , @xmath153 is the vacuum state of the dirac fermions .
the topologically protected single qubit operations are achieved by exchanging spatial positions of the mbss @xmath151 and @xmath39 , which exhibit the non - abelian statistics .
this braiding operation , i.e. , the @xmath154 phase gate , can be represented by the following unitary operation @xmath155 where @xmath156 .
it can be achieved in the one - dimensional semiconducting wires with t - junction by controlling tlgv adiabatically @xcite .
note that the braiding operation is insufficient for realizing universal quantum computation since it is not able to perform arbitrary single qubit rotations @xcite , which are usually not topologically protected ( rotation angle @xmath157 , @xmath14 is integer ) .
when there exists cmbs in the system ( it has been also investigated in the continuous model instead of lattice model @xcite ) , the effective hamiltonian takes @xmath158 .
we can obtain the following unitary operation for a fixed evolution time @xmath159 where @xmath160 . together with the braiding operation @xmath161
, one can implement umqr via the successive operations @xmath162 @xcite . to realize the operation @xmath163 , exact control over cmbs with a fixed evolution time is required .
this is crucial since the procedure is usually not topologically protected .
the other consideration is that the mbss should decouple instantaneously from each other after ( before ) the coupling of the adjacent mbss . as shown in figs .
( [ fig:07])-([fig:09 ] ) , the two considerations are sufficiently solved since the modulation of the amplitude or frequency of the electric potential on tlgv can be controlled with high precision . as an example , in order to manipulate the chemical potential , we adopt two distinct voltages , e.g. , @xmath164 and @xmath165 as shown in fig .
[ fig:07](a ) and fig .
[ fig:07](c ) .
we first apply a low voltage @xmath166 for periods of time @xmath167 to realize the operation @xmath168
. then we switch it to a high voltage @xmath169 to realize the braiding operation @xmath161 .
next we change the gate to the low voltage @xmath166 for periods of time @xmath170 to realize the operation @xmath171 .
finally we back to the high voltage @xmath169 to cancel cmbs .
this suggests that the umqr can be realized by changing tlgv in a square - wave form , composed by @xmath169 and @xmath166 .
in fact , the result is the same as in the case where the driving frequency is switched to @xmath172 or @xmath173 in sequence , see fig . [ fig:09](a ) and fig . [
fig:09](c ) . in conclusion ,
we have presented a proposal to couple majorana bound states mediated by a topologically trivial chain .
both cases of short - range and long - range interactions in the chain are considered .
we demonstrate that cmbs depends sharply on the pairing amplitude and the chemical potential of the central chain .
particularly , cmbs can be enhanced when the pairing amplitude of the chain is large .
when a driving field is applied to the chain , we explore the dependence of the cmbs on the frequency and amplitude of the driving field .
these results suggest that there are many ways to change cmbs . one way is to manipulate the chemical potential of the central chain . the other way is to modulate the effective pairing amplitude of the central chain , which can be realized by changing the frequency of periodic driving field .
finally , we have demonstrated the application of tunable cmbs into the realization of quantum gates , i.e. , implementing the unitary rotation operator @xmath174 for quantum computation .
this work is supported by the national natural science foundation of china ( grants no .
11175032 , no . 61475033 , and no .
11204028 ) .
n. read and d. green , phys .
b * 61 * , 10267 ( 2000 )
. d. a. ivanov , phys .
lett . * 86 * , 268 ( 2001 ) .
l. fu and c. l. kane , phys .
lett . * 100 * , 096407 ( 2008 ) .
m. z. hasan and c. l. kane , rev .
mod . phys . * 82 * , 3045 ( 2010 ) .
j. d. sau , r. m. lutchyn , s. tewari , and s. das sarma , phys .
* 104 * , 040502 ( 2010 ) .
x. l. qi and s. c. zhang , rev .
* 83 * , 1057 ( 2011 ) .
j. alicea , rep .
phys , * 75 * , 076501 ( 2012 ) . c. w. j. beenakker , annu .
con . mat .
phys . * 4 * , 113 ( 2013 ) .
a. das , y. ronen , y. most , y. oreg , m. heiblum , and h. shtrikman , nat .
* 8 * , 887 ( 2012 ) .
v. mourik , k. zuo , s. m. frolov , s. r. plissard , e. p. a. m. bakkers , and l. p. kouwenhoven , science , * 336 * , 1003 ( 2012 ) . m. t. deng , c. l. yu , g. y. huang , m. larsson , p. caroff , and h. q. xu , nano lett .
* 12 * , 6414 ( 2012 ) .
l. p. rokhinson , x. liu , and j. k. furdyna , nat . phys .
* 8 * , 795 ( 2012 ) .
s. n. perge , i. k. drozdov , j. li , h. chen , s. jeon , j. seo , a. h. macdonald , b. a. bernevig , and a. yazdani , science , * 346 * , 602 ( 2014 ) .
e. j. h. lee , x. jiang , m. houzet , r. aguado , c. m. lieber , and s. de franceschi , nat . nanotech .
* 9 * , 79 ( 2014 ) .
l. jiang , t. kitagawa , j. alicea , a. r. akhmerov , d. pekker , g. refael , j. i. cirac , e. demler , m. d. lukin , and p. zoller , phys .
lett . * 106 * , 220402 ( 2011 ) . c. v. kraus , s. diehl , p. zoller , and m. a. baranov , new j. phys .
* 14 * , 113036 ( 2012 ) .
a. a. zyuzin , d. rainis , j. klinovaja , and d. loss , phys .
lett . * 111 * , 056802 ( 2013 ) .
q. j. tong , j. h. an , j. b. gong , h. g. luo , and c. h. oh , phys . rev .
b * 87 * , 201109 ( 2013 )
. z. y. xue , m. gong , j. liu , y. hu , s. l. zhu , and z. d. wang , sci .
rep . * 5 * , 12233 ( 2015 ) .
m. cheng , r. m. lutchyn , v. galitski , and s. das sarma , phys .
lett . * 103 * , 107001 ( 2009 ) .
j. d. sau , d. j. clarke , and s. tewari , phys . rev .
b * 84 * , 094505 ( 2011 ) .
s. d. sarma , j. d. sau , and t. d. stanescu , phys .
b * 86 * , 220506 ( 2012 ) .
a. cottet , t. kontos , and b. doucot , phys .
b * 88 * , 195415 ( 2013 ) .
a. a. kovalev , a. de , and k. shtengel , phys .
lett . * 112 * , 106402 ( 2014 ) .
|
we study the coupling between majorana bound states(cmbs ) mediated by a topologically trivial chain in the absence / presence of pairing couplings at the boundaries .
both situations of short - range and long - range interactions in the chain are considered .
our results show that cmbs can be enhanced by the pairing coupling and long - range interaction in the trivial chain . with a driving field applied to the chain
, we calculate the dependence of cmbs on the frequency and amplitude of the driving field .
discussion on the application of the tunable cmbs in quantum computation is made .
|
many studies over the past decade have presented strong evidence for the coevolution of black holes and galaxies based on samples dominated by the low - density field ( e.g. * ? ? ?
* ; * ? ? ?
* and references therein ) .
it is interesting to determine if similar coevolution between agn and galaxies is present in groups and clusters because the physical processes that drive galaxy evolution , such as the available cold gas to fuel star formation and black hole growth , are substantially different from the field .
in addition , agn in groups and clusters at low - redshift appear to play the critical role in maintaining the temperature of their hot atmospheres ( e.g. * ? ? ? * ) , while agn heating at the epoch of cluster assembly remains a viable explanation for the minimum entropy level in the intracluster medium .
these questions have motivated my collaborators and i to systematically search for agn in groups and clusters of galaxies and led to the discovery of large numbers of agn in these dense environments in the local universe @xcite . in this contribution
i summarize some recently published work on the evolution of agn in clusters of galaxies @xcite and present a new , preliminary analysis of the evolution of agn in the lower - density group environment .
i conclude with a brief discussion of several future directions .
the key requirement to measure the evolution of agn in groups and clusters , as well as perform quantitative comparisons to their galaxy populations , is systematic and unbiased selection .
we select agn with x - ray observations obtained by the _
chandra _ and _ xmm _ satellites because x - ray emission provides a relatively unbiased measure of accretion for luminous objects .
redshifts of the counterparts to the x - ray sources , obtained from our work or the literature @xcite , are then used to identify the subset of x - ray sources associated with the groups and clusters .
our final sample includes 17 clusters with @xmath1 and 15 clusters with @xmath2 that are reasonably well - matched in spite of their heterogeneous selection based on available archival data .
we systematically characterize the agn population via the agn fraction , which is defined as the fraction of all galaxies brighter than some absolute magnitude that host agn above some luminosity threshold .
given the sensitivity limits of x - ray data and multi - object spectroscopy , the agn fraction in clusters is defined here as the fraction of galaxies brighter than @xmath3 that host agn with hard x - ray [ 2 - 10 kev ] luminosities above @xmath4 erg s@xmath5 . while redshift measurements are nearly complete for the x - ray sources , this is not always the case for the entire galaxy population of the group or cluster .
in many cases the number of galaxies above the luminosity threshold is estimated from the velocity dispersion and an empirical relation derived from the sdss @xcite . throughout this work
the agn fraction is only measured for agn and other galaxies that fall within the projected @xmath6 radius , that is the radius within which the group or cluster is a factor of 200 overdensity .
further details are provided in @xcite .
the main result of this analysis is presented in the figure , which demonstrates that the cluster agn fraction increases by a factor of eight from the present to @xmath0 .
this evolution is based on 2 luminous agn in 17 clusters at @xmath1 and 17 luminous agn in 15 clusters at @xmath7 .
this dramatic evolution is similar to the evolution of the star - forming galaxy population in clusters known as the butcher - oemler effect @xcite .
here we have parametrized the agn evolution as @xmath8 where @xmath9 ( _ dashed line _ in the figure ) .
this power - law index is consistent with the value of @xmath10 recently measured for star - forming galaxies from mid - infrared observations @xcite and supports the hypothesis that black holes and galaxies coevolve in dense environments .
to @xmath11 from a sample 17 clusters at @xmath1 and 15 clusters at @xmath7 ( _ filled circles _ ) .
these points represent the average fraction of cluster members more luminous than @xmath3 that host agn more luminous than @xmath12 erg / s .
the increase in the agn fraction is consistent with a power - law of the form @xmath13 ( _ dashed line _ ) and is approximately an order of magnitude below the field agn fraction ( _ open symbols _ ) over this redshift range .
a preliminary estimate of the group agn fraction suggests their rate of evolution is similar ( _ gray triangle _ at @xmath14 and _ upper limit _ at @xmath15 ) .
the group agn fraction is defined as the fraction of @xmath16 erg / s agn in galaxies with @xmath17 mag ( see _ right - hand axis _
label ) . ]
the rate of evolution of the agn fraction also appears similar to the lower - density field ( _ open symbols _ ) , although the cluster agn fraction is approximately an order - of - magnitude lower over the entire redshift range .
the lower agn fraction in clusters relative to the field strongly suggests that there must be a substantial change in the agn fraction in the intermediate - density group environment .
observations of high - redshift groups find a substantial decrease in the agn fraction with increasing overdensity at @xmath0 @xcite and a measured group agn fraction of @xmath18% for agn with broad - band x - ray luminosities @xmath19 erg s@xmath5 in galaxies more luminous than @xmath17 mag ( * ? ? ?
* _ gray triangle _ in figure ) .
studies of low - redshift groups find that the agn fraction is somewhat higher but consistent with the value measured in clusters @xcite ; however , the group sample contains far fewer galaxies and provides only an upper limit on agn as luminous as those in the high - redshift sample ( the _ arrow _ at @xmath15 in the figure represents the 90% confidence limit on the group agn fraction ) .
the observations summarized here have shown that the agn fraction in clusters increases by approximately an order of magnitude from the present to @xmath0 .
this evolution is also quantitatively similar to the evolution of the star - forming galaxy fraction in clusters , which suggests that cluster galaxies and their central black holes also coevolve , although at a different rate from their counterparts in the field .
a simple comparison between the agn fraction in clusters and the field indicates that the cluster agn fraction is approximately an order of magnitude lower over the entire observed redshift range .
one future direction for this research is to push further into the past toward the epoch of cluster galaxy assembly at @xmath20 .
numerous lines of evidence suggest that cluster galaxies formed earlier than their field counterparts , and the hypothesis of black hole and galaxy coevolution predicts that the the agn fraction in dense environments should exceed the field value by this point . at lower redshifts ,
the identification of the properties of the `` transition '' group environment , in which the agn fraction drops from the field to the cluster value , can potentially shed substantial light on the nature of the mechanism(s ) responsible for triggering and fueling agn .
finally , measurements of star formation and black hole accretion rates in the same cluster samples could reveal the extent of the correlation exhibited by individual galaxies and provide new insights into the physical processes that drive the observed coevolution .
i am grateful for the contributions of t. arnold , a. berti , t. e. jeltema , d. d. kelson , j. s. mulchaey , g. r. sivakoff , and a. i. zabludoff to the research that i have presented in this conference article .
i also appreciate the opportunity to speak at the conference and thank the organizers for this stimulating meeting .
my research was supported in part by nasa through chandra award number ar8 - 9014x and spitzer grant 1364997 , and nsf via award ast-0705170 , and the department of astronomy at osu .
|
the evolution of agn in groups and clusters provides important information about how their black holes grow , the extent to which galaxies and black holes coevolve in dense environments , and has implications for feedback in the local universe and at the epoch of cluster assembly .
i describe new observations and analysis that demonstrates that the agn fraction in clusters increases by a factor of eight from the local universe to @xmath0 and that this evolution is consistent with the evolution of star - forming galaxies in clusters .
the cluster agn fraction remains approximately an order of magnitude below the field agn fraction over this entire range , while a preliminary analysis of groups indicates that they too undergo substantial evolution .
address = department of astronomy and center for cosmology and astroparticle physics , + the ohio state university , columbus , oh , usa
|
it has been known for a long time that the line connecting the leading and following portions of a bipolar sunspot group usually tilts with respect to the solar equator @xcite .
the tilt angle has been regarded as an important parameter of the magnetic field since the publication of the solar dynamo models by @xcite and @xcite .
the longest available tilt angle data bases are derived from white - light photographic observations taken at mount wilson observatory ( @xmath2 ) in 1917 - 1985 and kodaikanal solar observatory ( @xmath3 ) in 1906 - 1987 @xcite .
the magnetograms available since 1974 also allow the study of tilt angle of active regions @xcite .
numerous characteristics of tilt angles and their changes have been investigated based on these data bases ( e. g. @xcite ) .
several papers have also confirmed that the average tilt angle increases with increasing latitude known as joy s law reported by @xcite at first .
recently new methods and data ( e.g. tilt angles derived from space - based magnetograms ) have been included into the studies .
for example , @xcite initiated the study of cycle - to - cycle dependence of the average tilt angles . @xcite developed a new method to derive tilt angles from mw and soho / mdi magnetograms .
they started to investigate the median tilt angle instead of the mean tilt angle , and they found a persistent asymmetry between the median tilt angles measured in the northern and southern hemispheres in all latitudinal ranges . @xcite also paid attention to the differences between hemispheres and cycles and they suggested a revision of joy s law equations .
it is widely accepted that the tilt angle is primarily due to the coriolis force acting on initially untilted rising magnetic flux tubes ( e.g. @xcite ) but some authors ( @xcite ) have questioned this explanation arguing that the tilt is already established in the source region inside the sun .
@xcite revealed a new dynamo pattern by separating the tilt angles of large and small solar bipoles .
@xcite studied the tilt angle scatter in the framework of babcock - leighton dynamo models by using the @xmath4 data bases .
they concluded that the tilt angle scatter has a significant impact on the variability of the solar cycle strength .
these examples show that a number of new ideas and questions have arisen concerning the tilt angles recently . to reveal more details and answer the open questions ,
the studies need extended homogeneous long - term measurements of tilt angles or cross - calibration of various data sets derived with different methods .
the longest sunspot group tilt angle data set available after the termination of @xmath4 data bases has been derived recently from @xmath0 .
the @xmath0 tilt angles can be also compared with the tilt angles derived from the soho / mdi - debrecen data ( @xmath1 ) catalogue with or without considering the magnetic polarity information . in this study , we analyse and compare sunspot group data available in various data bases ( @xmath4 , @xmath0 , @xmath1 ) , and some new data derived for the purpose of comparison presented in this paper .
the aim of this study is to reveal the special characteristics of these data bases and to look for various filters that can be used to decrease the differences between them .
the @xmath0 data ( gyri , baranyi & ludmny 2011 ) are measured on daily white - light full - disk photographic plates , which are mainly taken at debrecen heliophysical observatory and its gyula observing station but , in some cases , the data derived from observations of cooperating ground - based observatories or from space - borne images .
@xmath0 covers the years starting with 1974 with one observation / day time resolution and it contains the position and ( whole and umbral ) area data for all sunspot groups and each spot in them .
these data allow deriving tilt angle data from them . by using the umbral data of @xmath0 , the umbral tilt angle ( @xmath5 ) of the sunspot groups can be determined in a similar way as it was done at @xmath4 after measuring only umbral areas .
in addition , a new type of tilt angle ( @xmath6 ) can be derived from @xmath0 by using the whole spot area ( ws ) of the spots and pores ( @xmath6 ) .
the calculation of the tilt angle data follows the traditional way in both cases @xcite : @xmath7 where @xmath8 / @xmath9 are the area - weighted latitude of the following / leading portion , @xmath10 , @xmath11 are the area - weighted longitude of the following / leading portion , and @xmath12 is the latitude of the centroid of the whole group .
the separation of the leading and following portions can be derived as @xmath13 the @xmath0 does not contain magnetic polarity information on spots but the available magnetograms are frequently taken into account while grouping spots .
thus , the @xmath0 tilt angles based on estimated or measured polarities may be closer to the magnetic tilt data than they would be in the case of automatic grouping based simply on proximity . the mount wilson solar observatory has taken full - disk broad - band observations called white light directs once per day since 1906 .
the images for 1917 - 1985 were measured to determine the sunspot group tilt angles with the method described by @xcite .
the umbral area and mean position of each spot present within @xmath14 longitude from the central meridian ( lcm ) were measured by measuring the position of two successive vertices of the superposed quadrilateral . in the next step , a program based on their proximity grouped the spots .
a spot was included in a group if its position was within a distance of @xmath15 in latitude or @xmath16 in longitude from another spot in the group .
the area - weighted position of the leading ( following ) portion was derived from the data of umbrae in the portion located to the west ( east ) of the area - weighted centroid of the sunspot group @xcite .
the same measuring method was applied to the white light images of kodaikanal solar observatory @xcite to determine the @xmath3 tilt angles of sunspot groups and their daily changes for the year 1906 - 1987 .
these traditional sources of tilt angle data ( @xmath17 and @xmath18 ) are available at noaa / ngdc .
the @xmath2 directs was digitized and published in the frame of mt .
wilson solar photographic archive digitization project @xcite .
at present , the @xmath2 photo archive contains scans of white - light observations for the intervals 1917 - 18 and 1961 - 67 overlapping with the @xmath2 tilt angle data base .
these images can be evaluated with the software developed for @xmath0 .
we selected the directs available for 1961 - 67 , and the position and area data of spots were measured with the software of @xmath0 .
the new tilt angle data set ( @xmath19 ) was created by using the umbral data like in the case of @xmath5 . the mount wilson sunspot polarity drawings ( see e.g. @xcite ) are often available close in time to
the @xmath2 directs .
thus , by comparing the @xmath2 drawings and images , we can usually determine or estimate the polarity of the spots involved in the leading and following portions of the groups in @xmath17 and @xmath19 . the greenwich photoheliographic results ( @xmath20 ) catalogue ( 1874 - 1976 ) contains position and area data of sunspot groups measured in photographic observations taken at royal observatory of greenwich and at a few other observatories
. it can provide additional information for us on the correct separation of groups and on the nearby sunspot groups within the overlapping interval with the @xmath2 data set .
we used the digital version of @xmath20 available at ngdc .
concerning further details on @xmath20 data and their digital versions see e.g. the paper by @xcite .
for the time interval 1996 - 2010 , some comparisons can be made by using the @xmath1tilt angle data base @xcite derived from soho / mdi quasi - continuum images .
four types of tilt angle can be calculated based on this data set but we used only two of them .
the @xmath21 is the `` traditional '' tilt angle like @xmath5 while in the case of @xmath22 the separation of leading and following portion is based on the polarity information of umbrae derived from mdi magnetograms . in the case of @xmath1 and @xmath0 , the additional selection criterion of @xmath23 was applied because the measurements were confined to this region at @xmath4 .
to reveal the major characteristics of @xmath2 data , we compared two tilt angle data sets ( @xmath19 and @xmath17 ) based on the @xmath2 directs .
we used the @xmath20 to match the data of these two data set . in @xmath19
, we identified the sunspot groups according to @xmath20 .
we also tried to match the sunspot groups in @xmath17 to the groups in @xmath20 by using their heliographic latitude and lcm data corrected to the time difference of observations .
the comparison of the available data after jan . 1
, 1965 revealed that the time of observation of @xmath17 is in ut but the correct date is the next day . before 1965 ,
the date of @xmath17 is correct but the published time has to be subtracted from @xmath24 to derive the correct time of observation in ut . after the needed corrections ,
the data of sunspot groups in @xmath17 and @xmath19 can be matched quite well .
( this is also valid for the @xmath3 data set because of the same measuring method and software . )
identifying the matching cases , we generated the combined data set @xmath25 containing 506 sunspot groups with tilt angles for 1961 - 67 .
this pilot data set together with the directs and polarity drawings allows us to determine such characteristics of the measured sunspot groups which are not published in @xmath17 .
the pilot data are used to extrapolate the result for the whole interval of @xmath17 ( and @xmath18 ) and for the following era .
the correctness of the extrapolation is checked by using @xmath21 in some cases .
the position data of leading and following portions published in @xmath17 allowed us to estimate which spots were measured as members of the given portions in the directs . when we compared these portions in the image with their matching parts in the @xmath2 polarity drawings , the estimation of the magnetic polarities of spots also became available . in this way
, we can determine whether a given portion of a group indicated as leading ( following ) portion published in @xmath17 corresponds to the magnetic portion of a bipolar group correctly .
table 1 shows that in the most cases , the identification of the portions of bipolar groups was correct in @xmath17
. however , in about 26% of the cases , the measured tilt angles refer to unipolar groups that contain only spots of one polarity , these data necessarily distort the statistics .
the estimated number of these cases based on the @xmath2 observations is about the same as the percentage based on the @xmath1 tilt data derived by comparing the number of the traditional tilt angles and that of the tilts with polarity information .
the percentage of tilt angle of unipolar groups is probably similar in any tilt angle data base in any time interval .
the other source of distorted statistics is the false grouping of spots . in the pilot data set ,
about 14% of the @xmath17 data refers to such groups which are unipolar because of the measuring method . in these cases ,
both the leading and following parts of a group are measured but they are separated and assigned as two independent unipolar groups .
these divisions are caused by the automatic separating algorithm in those cases when the opposite polarity portions have larger distance than the threshold for grouping .
a further possible source of false tilt data may be the accidental close proximity of nearby groups . by examining the studied sample we can estimate that @xmath26 4% of @xmath17 belongs to this type of cases
. the percentage of such cases may be probably smaller in @xmath0 because of the use of magnetograms but a few percent of incorrect grouping can also be estimated both in the case of @xmath5 and @xmath21 mainly in their preliminary versions .
.percentages of the various cases of sunspot group data in tilt angle data bases @xmath17 ( 1961 - 67 ) and @xmath21 ( 1996 - 2010 ) . [ cols="<,<,<",options="header " , ] in three separation ranges in latitudinal bins of @xmath16 of @xmath27.,width=264 ] for the groups of @xmath28 .
the time interval is divided into six parts to see the temporal variation .
upper panel : years around solar maximum .
lower panel : years before and after maximum years.,width=264 ]
we have compared the traditional tilt angle data bases of mount wilson ( @xmath2 ) , kodaikanal ( @xmath3 ) observatories with the recently published data bases ( @xmath0 , @xmath1 ) of debrecen observatory in different time intervals by statistical methods and case studies .
new tilt angles have been derived from @xmath2 white - light observations for 196167 by using the software of @xmath0 .
we have compared the positions of the leading and following portions of the @xmath2 tilt angle data base with the positions in the photographic images and with the polarity information on that part of the group in the @xmath2 polarity drawings . based on these data
, we have revealed several differences between the data bases coming from the different measuring methods .
we have tested a few new methods to filter out the majority of incorrect tilt angles and to decrease the deviations because of random errors .
we have calculated the calibration factors between umbral area of @xmath2 , @xmath3 , @xmath20 and @xmath0 .
our results show that there is no systematic deviation between the traditional and new tilt angle data sets .
the filters and calibrations enhance the reliability and homogeneity of the tilt angle data bases .
because of these revisions , improved curves of the joy s law have been obtained by applying various filters on the tilt angle data .
our results show that the applied filters increase the slope of the joy s law ; in addition , the slopes derived from different data sets get closer to each other .
the decrease of the unambiguous tilt data may allow recognizing new types of characteristics of tilt angles .
the most unexpected result based on the new debrecen data is that the curves representing the joy s law exhibit an indication for a conspicuous feature : a plateau in the domain around the mean latitude of the active region belt at @xmath29 .
this may imply a still unknown contributor to the fine structure of the joy s law based on sunspot group tilt angles .
further research is required to provide much clearer evidence or a refutation of the possible significance of this feature but the advantages of the new data may be seen well .
i thank n. krivova , s. solanki , and m. dasi - espuig for initiating the study of dpd tilt angles and for helpful discussions on the data .
i thank a. ludmny and y .- m .
wang for improving the manuscript .
i am grateful to the referee for suggesting some extension of the study .
this study includes data from the synoptic program at the 150-foot solar tower of the mt .
wilson observatory . the mt .
wilson 150-foot solar tower is operated by ucla , with funding from nasa , onr and nsf , under agreement with the mt .
wilson institute .
wilson solar photographic archive digitization project is based upon work supported by the national science foundation under grant no.0236682 .
soho is a project of international cooperation between esa and nasa .
we wish to acknowledge the several generations of observers at the various observatories and institutes who took the ground - based observations covering many decades and made them public .
the research leading to these results has received funding from the european community s seventh framework programme ( fp7/2007 - 2013 ) under grant agreement eheroes ( project no
. 284461 ) . 99 babcock h.w . , 1961 , apj , 133 , 572 babu g. j. , feigelson e. d. , 1992 , comm .
simulation comput .
21 , 533 balmaceda l.a .
, solanki s.k . ,
krivova n.a . , foster s. , 2009 , jgr , 114 , a07104 , 12 baranyi , t. , gyri l. , ludmny a. , coffey h.e . , 2001 , mnras , 323 , 223 baranyi t. , kirly s. , coffey h.e . , 2013 , mnras , 323 , 223 bertello l. , ulrich r.k .
, boyden j.e . , 2010 , solar phys . , 264 , 31 dasi - espuig m. , solanki s.k .
, krivova n.a . , cameron r. , penuela t. , 2010 , a&a , 518 , a7 dasi - espuig m. , solanki s.k . , krivova n.a . , cameron r. , penuela t. , 2013 , a&a , 556 , c3 dsilva s. , choudhuri a. r. , 1993 , a&a , 272 , 621 foukal p. , 2014
, solar phys .
289 , 1517 .
gyri l. , baranyi t. , ludmny a. , 2011 , iau symp .
, 273 , 403 hale g.e . ,
ellerman , f. , nicholson , s.b . , joy , a.h .
, 1919 , apj , 48 , 153 howard r. , 1991a , solar phys . , 132 , 257 howard r. , 1991b , solar phys . , 136 , 251 howard r. , 1993 , solar phys . , 145 , 105 howard r. , 1996a , solar phys . , 167 , 95 howard r. , 1996b , solar phys . , 169 , 293 howard r. , gilman p.a . ,
gilman p.i . , 1984 , apj , 283 , 373 howard , r f. , gupta , s. s. , sivaraman , k. r. , 1999 , solar phys . , 186 , 25 howard , r f. , sivaraman , k. r. , gupta , s. s. , 2000 , solar phys . , 196 , 333 ivanov v. g. , 2012 , geomagn .
aeron . , 52 , 999 jiang j. , cameron r. h. , schssler m. , 2014 , apj , 795 , 5 kosovichev a. g. , stenflo j. o. , 2008 , apj , 688 , l115 leighton , r.b . , 1969 , apj , 156 , 1 li j. , ulrich r. k. , 2012 , apj , 758 , 115 mcclintock , b. h. , norton a. a. , 2013 , solar phys .
, 287 , 215 scherrer p. h. , bogart r. s. , bush r. i. , et al .
1995 , solar phys . , 162 , 129 sivaraman k. r. , gupta s. s. , howard r. f. solar phys .
, 146 , 27 sivaraman k. r. , gupta s. s. , howard r. f. solar phys . , 189 , 69 sivaraman k. r. , gokhale m. h. , sivaraman h. , gupta s. s. , howard r. f. apj . , 657 , 592 stenflo j. o. , kosovichev , a. g. , 2012 , apj , 745 , 129 tlatov a.g . , illarionov e. , sokoloff d. , pipin v. , 2013 , mnras , 432 , 2975 tlatov a. g. , tlatova k.a .
, vasileva v.v . ,
pevtsov a.a .
, mursula k. , 2014 , adv .
space res . , accepted willis d. m. , coffey h. e. , henwood r. , erwin e. h. , hoyt d. v. , wild m. n. , denig w.f , 2013a , solar phys .
, 288 , 117 wang , y .- m . ,
sheeley , n.r . : 1989 , solar phys . , 124 , 81 wang , y .- m . ,
sheeley , n.r . : 1991 , apj , 375 , 761
|
the study of active region tilt angles and their variations in different time scales plays an important role in revealing the subsurface dynamics of magnetic flux ropes and in understanding the dynamo mechanism . in order to reveal the exact characteristics of tilt angles ,
precise long - term tilt angle data bases are needed .
however , there are only a few different data sets at present , which are difficult to be compared and cross - calibrate because of their substantial deviations . in this paper , we describe new tilt angle data bases derived from the debrecen photoheliographic data ( @xmath0 ) ( 1974 ) and from the soho / mdi - debrecen data ( @xmath1 ) ( 1996 - 2010 ) sunspot catalogues .
we compare them with the traditional sunspot group tilt angle data bases of mount wilson observatory ( 1917 - 85 ) and kodaikanal solar observatory ( 1906 - 87 ) and we analyse the deviations .
various methods and filters are investigated which may improve the sample of data and may help deriving better results based on combined data . as a demonstration of the enhanced quality of the improved data set a refined diagram of the joy s law
is presented .
[ firstpage ] ( sun : ) sunspots methods : data analysis .
|
since the beginning of quantum theory a century ago , the study of the frontier between classical and quantum mechanics has been a constant topic of debate @xcite . in our opinion ,
the difficulties in the progress of this debate are also due to the deep - rooted use of different languages in classical and quantum mechanics : while classical system are usually described using trajectories @xcite , quantum descriptions involve wave functions @xcite .
these two languages are so different that it is hard ( and sometimes misleading ) to directly compare trajectories and wave functions . but
despite its pervasiveness , this clash of languages can be avoided by using other readily available formulations , where both theories can be compared on an equal footing .
for instance , bohmian mechanics provides a deterministic description of quantum systems where particles have definite positions , and their trajectories are _ choreographed _ by the wave function @xcite . on the other hand , ensembles of classical particles can be described in terms of waves using the hamilton
jacobi formalism , where a classical schrdinger ( wave ) equation can be derived @xcite . in this work , we discuss about the classical schrdinger equation . in the first part of the paper , in sec .
[ sec : hj ] , we will review the path from newtonian mechanics to the classical schrdinger equation , through the use of the hamilton
jacobi formulation @xcite .
we will notice that a single - valued condition on the hamiltons principal function is assumed during such development .
this additional assumption implies that the trajectories can not cross in configuration space , which leads to unexpected non - classical features not present in the newtonian formalism . as a conclusion ,
there are inherent non - classical features on the dynamics of the classical schrdinger equation . in the second part of the paper , in section sec .
[ sec : bm ] , we analyze different paths to arrive to the classical schrdinger equation from quantum systems by invoking the quantum - to - classical transition within the bohmian formalism .
in particular , we show that the center of mass of quantum systems with a very large number of identical particles tends to behave classically , following the classical schrdinger equation mentioned above .
we notice that this second derivation demands a very narrow wave packet .
this additional requirement avoids the inherent non - classical features mentioned in the first part of the paper for the classical schrdinger equation . in sec .
[ sec : sim ] we provide some numerical results obtained from the quantum and classical schrdinger equations justifying the theoretical discussions mentioned previously .
we conclude in sec .
[ sec : conc ] .
in what follows we present the usual way of arriving to the classical schrdinger equation to describe an ensemble of classical particles starting from newton s law . the mathematical derivation for a single - particle system
is done in sec .
[ sec : hjderiva ] , while sec .
[ sec : sunival ] discusses one of the relevant topics of this work : the requirement to the reach the classical schrdinger equation is a single - valued hamilton s principal function . according to classical mechanics @xcite , after specifying the initial position and velocity , the trajectory @xmath0 $ ] of a particle
is fully determined by newton s second law : @xmath1 } { dt^2 } = - \left.\frac { \partial u(x , t ) } { \partial x } \right|_{x = x[t]},\ ] ] where @xmath2 is the particle s mass and @xmath3 is a ( classical ) scalar potential @xcite .
it is well known that the previous newtonian trajectory @xmath0 $ ] is compatible with the hamilton
jacobi equation @xcite : @xmath4 where @xmath5 is the hamiltonian of the single - particle system , containing a kinetic energy plus a potential energy : @xmath6 the function @xmath7 can be identified with hamilton s principal function ( also known as the action @xcite ) , from which the particle s velocity is derived as @xmath8 here , the additional parameter @xmath9 is written to emphasize that each particle has its own velocity in the physical point @xmath10 .
the velocity in a physical point depends , not only on the point itself , but also on the particle . in other words
, two different particles can have two different values of their velocity @xmath11 ( i.e. two different values of hamilton s principal function @xmath12 ) at one unique point of physical space @xmath10 .
in general , finding the direct solution of eq .
( [ hamiltoncla ] ) is more difficult than solving newton law in eq .
( [ newtoncla ] ) .
one of the great merits of the hamilton - jacobi formalism is its ability to arrive to a wave formulation of classical systems .
although it describes the dynamics of a classical particle in 1d physical space , the hamilton s principal function @xmath13 is defined mathematically as a function in a 2-dimensional ( plus time ) space . on the contrary , for example , the ( wave ) function that describes the dynamics of a quantum particle in a 1d physical space , @xmath14 ,
is indeed defined in position - time dimensional space .
therefore , in order to connect the classical and quantum world , one is tempted to use the function @xmath15 neglecting its dependence on @xmath9 , to describe classical systems within a wave formulation .
we will see that this elimination has dramatic consequences on the ability of the classical schrdinger equation to correctly model classical dynamics .
once we neglect @xmath9 , the particle s velocity in eq .
( [ eq : v_cl1 ] ) can be rewritten as : @xmath16 we emphasize that the new equation eq .
( [ eq : v_cl ] ) implies that the velocity in a point @xmath10 does not depend on a the particle itself .
there is only one possible velocity in @xmath10 for all particles .
not all ensemble of classical system satisfy this requirement .
however , it is a requirement to be able to arrive to a wave description of of a classical ensemble of particles .
let us assume that we have a classical ensemble of single - particle experiments , each experiment is described by exactly the same hamiltonian , but with a slightly different initial conditions for the particles . in this case , we can define an ensemble of different initial positions of the particles of different experiments , distributed following @xmath17 . each trajectory in this ensemble
will evolve according to eq .
( [ newtoncla ] ) so that the function @xmath18 describes the distribution of particles , in the phase - space , at any time .
since all these classical particles will move in a _ continuous _ way , that is , from one unit of volume of the physical space to another , we can ensure that the ensemble of trajectories accomplishes the following _ local _ conservation law @xcite : @xmath19 it can be shown that the two previous ( real ) equations , eqs .
( [ hamiltoncla ] ) and ( [ conservationcla ] ) , for @xmath15 and @xmath20 are equivalent to the following ( complex ) classical schrdinger equation @xcite : @xmath21 where @xmath22 is defined as a classical ( complex ) wave function .
the additional function @xmath23 on eq .
( [ 1schocla ] ) is the so - called quantum potential : @xmath24 the reader can be , perhaps , surprised that this function , appearing in a purely classical context , is referred to as a quantum entity .
the justification is based on historical arguments @xcite , as it appeared first in the hydrodynamic or bohmian interpretation of quantum mechanics . in the transition from newton s equation , eq .
( [ newtoncla ] ) , to the classical schrdinger equation , eq .
( [ 1schocla ] ) , we have made an additional ( and usually unnoticed ) but critical assumption .
we have assumed that hamilton s principal function @xmath7 becomes a single - valued function ( having all velocities of an ensemble of particles the same velocity in each position @xmath10 ) .
let us explain with detail the critical problem mentioned here .
newton s second law can be used to describe any trajectory occurring in a given experiment .
we can therefore use it to describe one experiment defined by a trajectory @xmath0 $ ] and another realization of the same experiment with a different trajectory @xmath25 $ ] .
however , a problem arises when we define @xmath22 , as we force the trajectories velocities of @xmath0 $ ] and @xmath25 $ ] to fulfil eq .
( [ eq : v_cl ] ) for the same action function .
the use of the wave function implicitly assumes that @xmath15 is single - valued , which in turn implies that different trajectories ( described by the same @xmath15 ) will not cross in the physical ( configuration ) space .
however , in newtonian mechanics , trajectories actually cross with each other .
this fact is intimately related with how a distribution of positions and momenta in a classical ensemble is represented with the classical schrdinger equation .
not any general probability distribution of positions and momenta , e.g. , @xmath26 , can be represented with this formalism , as the it imposes a very strong restriction on the momenta . in order for the classical wave function to have a well - defined single - valued phase ,
each position is only allowed to have a single momentum ( the momentum is obtained from the derivative of the phase of the wave function ) . therefore
, only those distributions where all the particles in a particular position have the same velocity can be described by the classical schrdinger equation , i.e. , @xmath27 .
however , the formalism is not exempt of problems even in the case where these particular distributions are used , as we explain in the following and illustrate in the simulations in sec .
[ sec : sim ] . to understand better the physical implications of this additional requirement of single - valuedness of @xmath15 ,
let us illustrate the problem in very simple system where this issue becomes relevant : a particle falling under the action of a constant potential .
consider a particle in a one - dimensional physical system , under a potential @xmath28 , where @xmath2 is the mass of the particle and @xmath29 is a positive constant , e.g. gravity . from newton s law
we get that the acceleration is constant and equal to @xmath30 such that the trajectory for a particle with initial position @xmath31 $ ] and velocity @xmath32 is @xmath33=x[0]+v_0 t - \frac{1}{2 } g t^2 .
\label{eq : freefall}\ ] ] some trajectories starting at different initial positions @xmath31 $ ] are shown in fig .
[ fig : david ] the relevant point to our discussion is the presence of trajectories with different velocities ( one positive and one negative ) at a unique point in configuration space ( see , for example , \{@xmath34 m , @xmath35 s } in fig .
[ fig : david ] ) .
this duality of the velocity in a single point is not possible when we assume a single - valued function @xmath15 . ) for different initial positions @xmath31 $ ] with @xmath36 m / s and @xmath37 m / s@xmath38 . ]
let us now use the hamilton
jacobi formalism @xcite to tackle the same simple problem .
since the hamiltonian of the problem is @xmath39 and does not explicitly depend on time , then @xmath40 , and the hamilton
jacobi equation can be partially solved giving for the momentum of the particle @xcite .
@xmath41 where @xmath42 depends on the initial conditions of the particle , and in this case it corresponds to the total mechanical energy of the system .
the two possible signs in eq .
( [ eq:3 ] ) imply that ( at a particular point of @xmath10 ) one particle @xmath9 can have a positive velocity while going up , and another particle @xmath43 can have a negative velocity when going down .
therefore , the hamilton
jacobi formulation is capable of describing the trajectories in fig .
[ fig : david ] with a multivalued solution for @xmath7 without any problem . now suppose we want to describe an ensemble of trajectories with a classical wave function @xmath44 . at each position
, their distribution will be given by some single - valued @xmath20 , and their velocity described by some single - valued @xmath15 .
as we force @xmath15 to be single - valued in eq .
( [ 1schocla ] ) , the trajectories obtained from the classical schrdinger equation will deviate from the ( true ) ones given by newton s law in order to avoid crossing .
we will see many examples of this in sec .
[ sec : sim ] .
the conclusion of this discussion are quite dramatic . in the construction of the classical schrdinger equation ,
we use @xmath15 as a single - valued function , neglecting the parameter @xmath9 that provide a velocity for each particle at each physical point . this additional requirement on @xmath15 produces inherent non - classical features on the classical schrdinger equation , not present in the newtonian formulation of the problem . as a final note
, we must add that the single - valued @xmath15 implies non - crossing property for the trajectories .
one can understand the non - crossing requirement as if the different classical trajectories in different experiments @xcite have some type of non - crossing interaction among them .
an non - crossing interaction between the trajectories of different experiments is something unexpected classically , but it is actually the essential characteristic of quantum mechanics , the so - called `` quantum wholeness '' where the dynamics of one particle in one experiment is governed by the probability distribution @xmath18 of the ensemble of positions in all experiments @xcite .
this section contains a completely different way of arriving to the classical schrdinger equation .
we depart from the quantum world by invoking the quantum - to - classical limit in bohmian mechanics .
we start by introducing the basic ideas of bohmian mechanics in sec .
[ sec : bmintro ] , followed by a discussion of classicality in single - particle bohmian mechanics in sec .
[ sec : bmclass ] and its drawbacks . in sec .
[ sec : bmcom ] , we briefly explain a natural path to obtain classical behavior for the center of mass of a quantum system with a very large number of identical particles , and in sec .
[ sec : bmnarrow ] we discuss that this path imposes a condition on the shape of the quantum wave packet to reach classical dynamics .
we present here a single - particle description of a quantum ( sub)system according to the bohmian formalisms .
more detailed derivations , also including discussions on many - particle systems can be found elsewhere @xcite .
our starting point is the single - particle schrdinger equation : @xmath45 by casting into it the polar form of the wave function , @xmath46 one arrives at two equations .
the first equation is : @xmath47 this equation ( [ hamilton ] ) is the so - called quantum hamilton jacobi equation because of its similarity with the ( classical ) hamilton jacobi equation in eq .
( [ hamiltoncla ] ) but with one additional term , the quantum potential defined in eq .
( [ quantumpotential ] ) , which accounts for the quantum features of the system @xcite . from the similarities between the classical eq .
( [ hamilton ] ) and the quantum eq .
( [ hamiltoncla ] ) , a velocity _ guiding _ field can be defined as @xcite : @xmath48 once the velocity is defined , a quantum trajectory ( initially located at @xmath31 $ ] ) , can be defined as easily as in classical mechanics as : @xmath49 & = x[0]+ \int_0^t v(x[t^\prime],t^\prime ) dt^\prime . \label{trajint}\end{aligned}\ ] ] the second equation obtained from eq .
( [ 1scho ] ) is : @xmath50 this eq .
( [ conservation ] ) is a continuity equation for the conservation of a probability density @xmath18 .
it ensures _ equivariance _ for an ensemble of trajectories : if the initial positions of such ensemble are distributed according to @xmath51 , the positions at any other future time @xmath52 will be distributed according to @xmath53 .
the selection of the initial distribution of particles according to @xmath51 is based on a _ quantum equilibrium hypothesis_. some authors argue that it is an additional postulate of the theory , while others argue that this is just a particular fact verified in virtually all systems under study @xcite in summary , the development performed up to here in secs . [
sec : hj ] and [ sec : bm ] provides a common language for classical and quantum theories , in terms of either wave functions or trajectories .
we emphasize that one has to compare either classical and quantum wave functions or classical and quantum ensembles of trajectories ( not a single classical trajectory with a quantum wave function ) .
the presence of @xmath23 in the quantum versions of the hamilton
jacobi in eq .
( [ hamilton ] ) , implies that each bohmian trajectory of a single - particle experiment depends explicitly on the shape of @xmath20 . on the contrary , each classical trajectory can be computed ( from newton s law or the hamilton
jacobi equation ) independently of the shape of the classical ensemble .
therefore , the differences between quantum and classical ensembles of trajectories do not correspond to differences between waves and particles , because both descriptions can be used to both kinds of systems .
the difference is on the dependence of each trajectory on @xmath20 or not .
we note , finally , that as seen in sec .
[ sec : sunival ] , the classical schrdinger equation forces somehow a dependence of the trajectories on @xmath20 because eq .
( [ 1schocla ] ) imposes the additional non - crossing condition for the ensemble of trajectories .
it can be easily demonstrated using eq .
( [ conservation ] ) and eq .
( [ hamilton ] ) that a newton - like equation can be developed for the quantum ( bohmian ) trajectories @xcite : @xmath54 } { dt^2 } = \left . -\frac { \partial } { \partial x } \big[u(x , t ) + q(x , t ) \big ] \right|_{x = x[t]}.\ ] ] here , again , we confirm that each unique quantum ( bohmian ) trajectory depends on the quantum potential @xmath23 that in turns depends on @xmath20 . in 1964 , nathan rosen @xcite used the bohmian formalism to develop the conditions needed to reach a classical regime for a one - particle system .
we follow here his single - particle arguments . comparing eqs .
( [ newtoncla ] ) and ( [ newton ] ) it is easy to see that bohmian trajectories with classical behavior can be obtained by imposing @xcite @xmath55 i.e. , that @xmath23 is constant everywhere .
this strategy to reach a classical equation , however , can be quite problematic because most of the time it is incompatible with a well - defined wave function solution of eq .
( [ 1scho ] ) .
the reason is because we have three equations , eqs .
( [ hamilton ] ) , ( [ conservation ] ) and ( [ quantumpotentialzero ] ) , imposed on only two unknowns , @xmath20 and @xmath15 .
the number of unknowns is sometimes artificially increased in the literature by also assuming the ( classical ) potential @xmath3 as an unknown to be fixed by these equations .
some very exotic solutions can be found where quantum and classical solutions are the same @xcite .
a simple example is a plane wave of momentum @xmath56 in free space , @xmath57 , such that @xmath58 and @xmath59 .
another , more phenomenological , way of arriving to the classical schrdinger equation was presented by richardson _
et al . _
they define @xmath60 , a _ degree of quantumness _
, into the schrdinger equation such that it reads @xmath61 \psi(x , t ) .\end{aligned}\ ] ] for @xmath62 this equation becomes the schrdinger equation and for @xmath63 it becomes the classical schrdinger equation .
this equation has the virtue that for intermediate values of @xmath60 allows the study of the appearance of quantum behavior as @xmath64 becomes more and more relevant ( and thus the system becomes more classical ) .
however , an explanation of the origin or physical meaning of @xmath60 is still missing in the literature . in what follows
we summarize our recent work on the quantum - to - classical transition through a generalization of rosen s attempt @xcite by analyzing the dynamics of the center of mass of a many - particle system with a large number of identical particles @xcite .
we show that classicality appears as a _
natural _ quantum limit of the center of mass of most macroscopic objects , i.e. the effect of the quantum potential associated to the center of mass becomes negligible while still retaining a well - defined wave function .
we start by providing some simple arguments on why it is better to formulate the quantum - to - classical transition in terms of the center of mass rather than in terms of individual particles .
first , it is important to notice that a center of mass of a quantum object can have a classical behavior , while the fundamental particles that form the object ( and used to compute the center of mass ) are still fully quantum .
second , it can be shown that the quantum potential of the center of mass is the sum of the quantum potentials of the particles of the quantum object ( which can take positive and negative values ) .
therefore , one can expect that the value of the quantum potential averaged over a large number of particles will tend to become small ( even if the effect of the quantum potential in each particle is not negligible ) . let us emphasize the quantum potential acting on a particle is , in general , much more complex than the classical potential .
therefore , the averaged classical potential over the object will basically remain equal to the classical potential assigned to one particle , while the averaged quantum potential tends to be much smaller .
finally , by looking the quantum - to - classical transition through the center of mass of a large number of particles , we are imposing a `` natural '' type of coarse - grained procedure : only macroscopic objects ( whose center of mass involves a very large number of particles ) will become classical .
after these qualitative arguments , next we briefly summarize our work of ref .
we consider a quantum system defined by the many particle wave function @xmath65 where @xmath66 for @xmath67 are the positions of @xmath68 particles with mass @xmath2 .
we consider a change of variables , from @xmath69 to the center of mass @xmath70 and a set of relative coordinates @xmath71 , @xmath72 by rewriting the many - particle schrdinger equation as a function of the center of mass @xmath70 and a set of relative coordinates @xmath71 , one arrives at the equation for @xmath73 @xcite @xmath74 with @xmath75 .
following similar development as the ones above , the force acting on the ( bohmian ) center of mass trajectory is given by @xmath76}{dt^2 } & = -\left .
\frac{\partial } { \partial x_{cm } } \left(u(x_{cm},\vec{y},t ) + q_{cm } + \sum_{j=2}^n q_{j } \right ) \right|_{x_{cm}=x_{cm}[t ] \atop \vec y=\vec y[t]}\end{aligned}\ ] ] with the quantum potentials
@xmath77 the wave function @xmath73 depends on the center of mass @xmath70 and it includes correlations with all relative coordinates @xmath78 . a single particle wave function of the center of mass can be easily constructed from bohmian mechanics through the concept of conditional wave function @xcite .
the conditional wave function for the center of mass can be obtained by evaluating the many - particle wave function , @xmath73 , at the positions of the bohmian trajectories for the rest of degrees of freedom , e.g. , @xmath79,t)$ ] .
then , by construction , the conditional wave function will yield the same bohmian velocities for the center of mass as would the full wave function , and by extension the same trajectory . following refs .
@xcite , the equation of motion for the conditional wave function is @xmath80}\nonumber\\ & - i\hbar \sum_{j = 2}^n v_j^h[t ] \left .
\frac{\partial \psi(x_{cm},\vec{y},t)}{\partial y_j}\right|_{\vec{y}=\vec{y}[t]}+ u(x_{cm},\vec{y},t)\psi , \end{aligned}\ ] ] a particular case single - particle ( conditional ) wave equation of the type @xmath81,t ) \right .
\nonumber \\ & \qquad \left . \vphantom{\frac
{ \partial}{\partial x^2 } } + g_{a}(x_a,\vec y[t],t)+ i j_{a}(x_a,\vec y[t],t ) \right ] \psi .\end{aligned}\ ] ] the additional potentials @xmath82,t)$ ] and @xmath83,t)$ ] are defined through the many - body wave function @xcite .
the relevant point here in this many - particle generalization of rosens attempt ( see sec .
[ sec : bmclass ] ) is that the new equation eq .
( [ eq : conditional ] ) has the additional environment degrees of freedom @xmath82,t)$ ] and @xmath83,t)$ ] included .
now , contrarily to the single - particle case , the number of equations and unknowns in the quantum - to - classical transition is well balanced and it is possible to recover classical dynamics . to obtain a classical behavior for the ( quantum ) center of mass ( not other degrees of freedom ) using eq .
( [ realcond ] ) , an additional requirement is needed , @xmath84 along the path @xmath85 $ ] . therefore , we have again three real equations to get a classical solution and , now , we have three unknowns @xmath86 , @xmath87 and @xmath88 @xcite .
then , when eq .
( [ condition6 ] ) is satisfied , we arrive to the classical schrdinger equation in eq .
( [ 1schocla ] ) for the center of mass of a macroscopic system @xmath70 . in summary
, it is possible to obtain classical dynamics for the center of mass of a ( very large ) quantum system in a single experiment @xcite .
this result can also be interpreted as a single - experiment generalization of the well - known multiple - experiment ehrenfest theorem . in our recent work
@xcite , we presented the conditions required for a quantum system of identical particles to ensure that its center of mass @xmath70 has a classical behavior .
the qualitative arguments are mentioned in sec .
[ sec : bmcom ] and the mathematical development can be found in the appendix of our ref .
in particular , we discussed when and why it is reasonable to expect that the spatial derivatives of the classical potential in eq .
( [ condition6 ] ) are expected to be larger than those of the quantum potentials .
we notice that a relevant requirements for satisfying the condition in eq .
( [ condition6 ] ) is that the ( center of mass conditional ) wave function has to be very narrow .
we argue that the requirement of a narrow wave packet is a natural requirement for the center of mass of a quantum object with a large number of particles . by using the double nature of @xmath89 , as a probability distribution and as a guiding field @xcite , one can roughly anticipate the shape of such ( center of mass conditional ) wave packet .
we consider that we repeat the experiments with the same @xcite macroscopic object .
according to the quantum equilibrium hypothesis mentioned in sec .
[ sec : bmintro ] , each of these experiments will have a different distribution of bohmian particles .
however , all these distribution of bohmian particles will imply very similar center of mass .
we can reasonable assume that each position @xmath90 has a distribution with a standard deviation @xmath91 .
when @xmath92 , ( a kind of central limit theorem analyzed in the appendix of our ref .
@xcite ) the distribution @xmath93 of the center of mass can be roughly approximated by a normal distribution ( around a central value @xmath94 ) , @xmath95 and the dispersion of the center of mass , @xmath96 , follows @xcite then @xmath97 we clearly see from eq .
( [ variance ] ) that when the number of particles of the macroscopic object increases , the wave packet becomes narrower . the combination of the classical schrdinger equation and a narrow wave packet has some unexpected beneficial effects in our final goal of describing the quantum - to - classical transition .
first , as we will see in sec .
[ sec : sim ] , a narrow wave packet governed by the classical schrdinger equation tends to remain narrow all the time ( this is due to the non - linearity of this equation ) .
this effect in addition implies that there is a negligible ( quantum ) uncertainty on the values of the position and momentum of the center of mass . in sec .
[ sec : sim ] we will see that these narrow shape of the wave packet imply , in fact , that the solutions of the classical schrdinger equation in eq .
( [ 1schocla ] ) are fully classical , where the problems due to the single - valued @xmath15 becomes irrelevant .
in this section , we will compare three types of trajectories associated to three equations of motions discussed in this work for the same physical problem ( i.e. the same hamiltonian ) .
first , the quantum / bohmian trajectories obtained from the ( quantum ) schrdinger equation .
second , the trajectories obtained from the classical schrdinger equation ( cse ) .
third , the ( actual ) classical trajectories , i.e. , obtained from newtonian mechanics . we will check several physical problems with several potentials @xmath98 and initial wave functions , and we will use units where @xmath99 throughout . the evolution of the wave functions is performed by means of a spectral split - step integration @xcite of the ( classical and quantum ) schrdinger equations .
however , some care needs to be taken into account when integrating the cse as the quantum potential can easily be the source of numerical noise . because of its @xmath100 dependence
, the numerically - calculated @xmath64 is very prone to having ( unphysical ) quick oscillations and discontinuities where @xmath101 is small .
therefore , before calculating @xmath64 using finite differences for the second derivative , @xmath101 is smoothed out using a @xmath102-point smoothing function ( depending on the case ) .
we consider an ensemble of ( classical or quantum ) @xmath103 experiments .
the trajectories in each experiment are selected according to the modulus squared of the initial wave function @xmath104 . in order to obtain nicely spaced trajectories ,
we fix the initial position set @xmath105 according to @xmath106 for @xmath107 .
the evolution of each quantum or cse trajectory is obtained by integrating eq .
( [ trajint ] ) with the velocity in eq .
( [ bohmvel ] ) ( or eq . ( [ eq : v_cl ] ) ) . ) with an initial momentum .
the initial wave function is given by eq .
( [ eq : gauss ] ) with @xmath108 , @xmath109 , and @xmath110 .
top row shows @xmath111 for the integration of the ( a ) quantum and ( b ) classical schrdinger equations .
bottom row shows the trajectories corresponding to the ( c ) quantum , ( d ) cse and ( 3 ) classical dynamics .
[ fig : fs_gauss ] ] the first case one can consider is free - space dynamics , i.e. @xmath112 . for simplicity we will consider an initial gaussian wave packet of width @xmath113 , centered around @xmath114 and with an initial momentum @xmath115 . @xmath116 the results of the simulations for @xmath108 , @xmath109 , and @xmath110 are shown in fig .
[ fig : fs_gauss ] .
let us start by comparing the wave dynamics .
both the classical and quantum wave packets propagate forward with a mean velocity @xmath115 .
however , while the quantum wave packet expands as it propagates forward , fig .
[ fig : fs_gauss](a ) , the cse one retains its shape during the whole simulation , fig .
[ fig : fs_gauss](b ) .
because of this , the quantum trajectories do not follow straight lines and separate during the evolution , fig .
[ fig : fs_gauss](c ) , while the cse ones follow straight lines , fig .
[ fig : fs_gauss](d ) , as one would expect for a free particle with fixed momentum , as shown in fig .
[ fig : fs_gauss](e ) .
this separation of the quantum trajectories ( and the non - expansion of the classical wave packet ) can be seen due to the presence ( or absence ) of the quantum potential ( in this case , an inverted parabola centered on the wave packet ) in their equations of motion . in this particular case , since the different trajectories in the classical case do `` interact '' among them ( i.e. they do not tend to cross ) the problem of having a single - valued @xmath15 does not appear here ; and the cse trajectories perfectly match the classical ones .
but for two wave packet colliding in free space ( @xmath112 ) .
the initial wave function is given by eq .
( [ eq : twogauss ] ) with @xmath108 , @xmath117 , and @xmath118 .
[ fig : fs_twogauss ] ] a more complicated case where one can expect trajectories to cross is considering the collision of two wave packets , i.e. , an initial wave function @xmath119 i.e. , the superposition of two gaussian wave packets centered at @xmath120 with momenta @xmath121 .
the results are shown in fig .
[ fig : fs_twogauss ] .
this is a similar case to the one studied in ref .
@xcite , but because its authors did not consider initial momenta , they could not see the interference and shape of the wave packets after them . for both the quantum and classical wave functions , fig .
[ fig : fs_twogauss](a , b ) , before the two wave packets interact ( for @xmath122 ) , we can see each one doing similar dynamics to the previous case . however , when they find each other we can see interferences happening which disappear once the packets fly past each other . after the interferences the wave packets continue with the same dynamics as before the collision .
since each trajectory represents an independent single - particle experiment , newtonian mechanics will yield straight ( crossing ) trajectories , fig .
[ fig : fs_twogauss](e ) .
however , because of the single - valuedness of @xmath15 discussed in sec .
[ sec : sunival ] , the cse trajectories , fig .
[ fig : fs_twogauss](d ) , do not follow this straight paths and bounce off each other as in the quantum case , fig .
[ fig : fs_twogauss](c ) .
this effects can also be seen in the classical wave function , fig .
[ fig : fs_twogauss](b ) , by displaying interferences and some population around @xmath123 after the wave packets have crossed .
this is clearly a drawback of the classical schrdinger equation .
let us notice that a very narrow wave - packet ( for example , one whose initial support does only include the trajectories with positive initial positions ) as discussed in sec .
[ sec : bmnarrow ] will avoid this spurious interference effect of the cse .
another text - book example of quantum dynamics is a wave packet impinging on a barrier . in this case
, the potential will be modeled by a ( gaussian ) barrier of width @xmath124 and height @xmath125 centered at @xmath126 , i.e. , @xmath127 with @xmath128 and @xmath129 .
the initial wave function will be described by eq .
( [ eq : gauss ] ) with @xmath130 , @xmath131 and @xmath132 , therefore the classical particle will have energy @xmath133 in the quantum case , however , the total energy contains an additional term , @xmath134 and therefore the conditions for transmission / reflection on the barrier will differ slightly .
we will consider two different cases , depending on whether the wave packet has more or less energy than the potential barrier .
but for a wave packet impinging on a ( low ) barrier .
the initial wave function is given by eq .
( [ eq : gauss ] ) with @xmath108 , @xmath135 , and @xmath132 .
the barrier potential is described by eq .
( [ eq : barrier ] ) with @xmath128 , @xmath129 , and @xmath136 .
[ fig : barrier_more ] ] the first of these cases , with @xmath136 , is depicted in fig .
[ fig : barrier_more ] . we know that a quantum particle will get partially reflected and partially transmitted
this is shown in fig .
[ fig : barrier_more](a ) , where quantum wave function split in two parts with a small fraction getting reflected and the rest going over the barrier .
this is confirmed by the quantum trajectories , fig .
[ fig : barrier_more](c ) , some of which get reflected on the barrier . because @xmath137 , a classical particle will climb up the barrier and slide down the other side without any net change in its energy , as shown by the classical trajectories in fig .
[ fig : barrier_more](e ) . again , as these classical trajectories do not cross , the cse trajectories follow them perfectly , see fig . [
fig : barrier_more](d ) .
it is interesting to see how this is visualized with the classical wave function , fig .
[ fig : barrier_more](b ) : the particle slows down and gets narrower around the barrier position , and after transversing it , recovers its previous shape and velocity .
but for a wave packet impinging on a ( high ) barrier .
the initial wave function is given by eq .
( [ eq : gauss ] ) with @xmath108 , @xmath135 , and @xmath132 .
the barrier potential is described by eq .
( [ eq : barrier ] ) with @xmath128 , @xmath129 , and @xmath138 .
[ fig : barrier_less ] ] the other case , where @xmath138 , is depicted in fig .
[ fig : barrier_less ] .
again , the quantum particle gets mostly reflected and partially transmitted , seen in both the wave function and the trajectories in fig .
[ fig : barrier_less](a , c ) .
the classical case does not have enough energy to surpass the barrier and it is forced to be completely reflected .
however , one would expect in this case the classical trajectories to cross ( the first ones arriving at the barrier should be the first ones reflected ) , as in fig .
[ fig : barrier_less](e ) .
therefore , the cse trajectories equation are forced to deviate from this behavior : the first ones to arrive at the barrier linger longer at the classical turning point , while the ones which arrive later turn around earlier , see fig .
[ fig : barrier_less](d ) .
this also produces a narrowing of the classical wave packet while it is `` bouncing '' off the barrier , fig .
[ fig : barrier_less](b ) . again
, narrowing the classical wave packet , as indicated in sec .
[ sec : bmnarrow ] , will make this feature totally insignificant .
finally , let us notice that there is no tunneling in the classical wave function .
but for a displaced wave packet in a harmonic potential .
the initial wave function is given by eq .
( [ eq : gauss ] ) with @xmath108 , @xmath139 , and @xmath140 .
the harmonic potential is described by eq .
( [ eq : harmonic ] ) with @xmath141 .
[ fig : harm_displaced ] ] @xmath142 @xmath143 [ fig : harm_narrow ] ] the last case we will study is that of a harmonic oscillator , i.e. , a potential of the form @xmath144 with an initial gaussian wave function described by eq .
( [ eq : gauss ] ) . the first case we study is shown in fig .
[ fig : harm_displaced ] , where the initial wave function is the ( quantum ) harmonic oscillator ground state ( a gaussian of width 1 ) , but displaced from the minimum of the potential . the simulation for the ( centered ) ground state of this system can be found in @xcite .
the quantum dynamics show the wave packet not changing shape as it oscillates around the origin , see fig .
[ fig : harm_displaced](a ) . therefore , the corresponding trajectories , fig .
[ fig : harm_displaced](c ) , also perform oscillations keeping a fixed distance between them . on the other hand , the classical trajectories shown in fig .
[ fig : harm_displaced](e ) , all cross the origin at the same time because the oscillation frequency is independent of the oscillation amplitude . because of this crossings , the cse trajectories are again affected by the single - valuedness of @xmath145 , see fig .
[ fig : harm_displaced](d ) : trajectories starting closer to the center of the trap would oscillate with smaller amplitudes than those starting further away . in order to avoid the crossing , they seem to bounce off each other around @xmath146 . similar behavior can be seen in fig .
[ fig : harm_narrow ] , where the initial wave function is now centered around the potential minimum , but it has a width narrower than the harmonic oscillator ground state . because the wave packet is narrower than the ground state , the quantum wave packets starts oscillating by expanding and compressing around a width of 1 , see fig .
[ fig : harm_narrow](a ) .
the quantum trajectories , fig .
[ fig : harm_narrow](c ) , then follow suit : spreading out when the wave function expands and getting closer together when the wave function compresses . the classical wave function in fig .
[ fig : harm_narrow](b ) , however , shows the opposite behavior : because classical dynamics should in principle by unaffected by the shape of the wave function , a centered gaussian will always start getting compressed no matter its initial width . as in the previous case
, we can see the cse trajectories trying to oscillate around the origin , but , again , because they can not cross they are forced to repel each other , see fig . [
fig : harm_narrow](d ) .
this behaviour by the cse trajectories is similar to the classical trajectories in fig .
[ fig : harm_narrow](e ) .
however , as cse trajectories can not cross , they seem to swap places when they are around @xmath146 such that they appear to completely oscillate around the origin ( like the classical ones do ) .
this is also the case for the previous example .
again , narrowing the classical wave packet , with @xmath92 which implies @xmath147 for the wave packet of the center of mass as indicated in eq .
( [ variance ] ) , will make this non - classical feature totally insignificant .
we have presented a study of the classical schrdinger equation .
first , we have shown that newtonian mechanics leads itself naturally to a reformulation of classical systems in terms of the so - called classical schrdinger equation .
we emphasize that it is usually unnoticed that reaching the classical schrdinger equation requires an additional single - valued requirement for the action @xmath15 not present in the original hamilton - jacobi formulation .
this restriction introduces non - classical ( non - newtonian ) features on the trajectories obtained from the classical schrdinger equation .
second , we have shown how the classical schrdinger equation can be reached from quantum theories by invoking the quantum - to - classical transition .
we have shown rosen s proposal for the recovery of classical formalism from bohmian mechanics of a single particle quantum system , and then we have presented a generalization of his work applicable to a more general type of experiments where a classical behavior emerges naturally at the center of mass of a quantum object ( not at each individual particle ) with a large number of identical particles .
the continuity between quantum and classical systems expressed here is an attempt to bridge the gap between these two theories . by providing a description of quantum mechanics in terms of trajectories , and their encompassing wave functions , bohmian mechanics provides a common formalism which can be applied to both .
the transition from one regime to another under plausible conditions is given in sec .
[ sec : bm ] , where we show that not only the common language makes it _ a priori _ an adequate vehicle for relating these theories , but that classical dynamics actually emerge naturally from purely quantum behavior . in this work ,
we warn about two difficulties that have to be acknowledged in this quantum - to - classical bohmian path .
first , the possible unphysical features of the classical schrdinger equation when dealing with ensembles of particles because of imposing a single - valued condition on the action @xmath15 .
second , the several arguments that give strong support to the idea that it is better to formulate the quantum - to - classical bohmian path in terms of the center of mass rather than in terms of individual particles .
see the recent work of the authors in ref .
this work has been partially supported by the fondo europeo de desarrollo regional ( feder ) and ministerio de economa y competitividad through the spanish project nos .
tec2012 - 31330 and tec2015 - 67462-c2 - 1-r , the generalitat de catalunya ( 2014 sgr-384 ) , by the european union seventh framework program under the grant agreement no : 604391 of the flagship initiative `` graphene - based revolutions in ict and beyond '' and the okinawa institute of science and technology graduate university .
the development of the classical formalism done here is for a single - particle in a one - dimensional space .
the generalization of the conclusions extracted here to system with higher dimensionality is not trivial .
the single - valued @xmath15 in the 1d the classical schrdinger equation , implies that the trajectories do not cross in the physical space .
however , the single - valued hamilton s principal function @xmath148 in the @xmath68-dimensional classical schrdinger equation , implies that the trajectories do not cross in the ( @xmath68-dimensional ) configuration space , but such trajectories can indeed cross in physical space .
in addition , structures such as vortices ( that require physical spaces with dimensions greater than 1 ) are not present in our 1d analysis . in any case
, the 1d case is enough to reveal important features and drawbacks of the classical schrdinger equation .
one could very well argue that @xmath149 is an additional unknown .
however , one can usually fix @xmath149 by fixing a conservation of probability ( although this is not necessary for a conditional wave function ) .
nevertheless , and unlike the single - particle case , the number of unknowns for the many - body system is equal or higher than the number of equations to satisfy .
|
in this paper , the classical schrdinger equation , which allows the study of classical dynamics in terms of wave functions , is analyzed theoretically and numerically .
first , departing from classical ( newtonian ) mechanics , and assuming an additional single - valued condition for the hamiltons principal function , the classical schrdinger equation is obtained .
this additional assumption implies inherent non - classical features on the description of the dynamics obtained from the classical schrdinger equation : the trajectories do not cross in the configuration space .
second , departing from bohmian mechanics and invoking the quantum - to - classical transition , the classical schrdinger equation is obtained in a natural way for the center of mass of a quantum system with a large number of identical particles .
this quantum development imposes the condition of dealing with a narrow wave packet , which implicitly avoids the non - classical features mentioned above .
we illustrate all the above points with numerical simulations of the classical and quantum schrdinger equations for different systems .
|
the transport properties of matter at extreme high temperature and energy density , well proved in high energy accelerator experiments at the relativistic heavy - ion collider ( rhic ) and at the large hadron collider ( lhc ) , raised some fundamental questions related to the smallest distance scale in physics .
theories , motivated by conformal field theory and higher dimensional dual gravity , in fact predicted a lower limit for the shear viscosity over entropy density ratio being around @xmath0 in natural units @xcite .
although this limitation can be overcome in more sophisticated , nonlinear dual gravity models @xcite or other approaches @xcite , the actual calculation @xcite and measurement of the viscosity and related properties of elementary matter remained in the focus of research interest .
the more , because the scientific evaluation of experimental signals , extracted from analyses of particle spectra and correlations , can be interpreted in terms of thermal concepts only if our knowledge about the material quality of a ( strongly ) interacting quark - gluon plasma is well established @xcite . besides the interpretation of temperature @xcite
, the whole hydrodynamical and statistical approach relies on our basic assumptions about the equation of state ( eos ) and the dissipative capacities of this new stage of strongly interacting matter @xcite .
modern statistical physics methods to the theoretical description of high - energy phenomena include a consideration of non - extensive thermodynamics @xcite .
besides giving a more general framework than boltzmann did for the entropy formula , also the mesoscopic background including langevin , fokker - planck or boltzmann type equations can be extended to involve deviations from the classical nineteenth century picture @xcite .
most of these generalizations depend on a single parameter , denoted by @xmath1 , which also influences the canonical equilibrium state leading from an exponential gibbs distribution in energy to a power - law tailed one .
these are observable in experimental transverse momentum spectra and multiplicity fluctuations @xcite , therefore the important question arises , how are the kinetic and hydrodynamical approaches affected by the value of this parameter . are there qualitative or only quantitative , minor deviations to be expected when @xmath2 ? fits to the relativistic heavy - ion collider ( rhic ) and the large hadron collider ( lhc ) spectra make a hadronic matter value of @xmath3 @xcite and a quark matter value of @xmath4 @xcite probable .
these fits point out a deviation from the standard boltzmann - gibbs ( bg ) statistics for @xmath5 .
the degree of non - equilibrium in a system is defined relative to a fixed local @xmath1-equilibrium .
assuming @xmath2 means that the local equilibrium differs from the standard bg equilibrium . on the other hand
it may be regarded as non - equilibrium compared to the classical @xmath5 , if one considers a time evolution of this parameter .
however , we do not consider this possibility in the present paper .
we investigate a general non - equilibrium system which - for any given @xmath1 - dissipates energy and produces entropy .
this we call @xmath1-non - equilibrium . similarly to classical bg systems this @xmath1-non - equilibrium is related to the response of the system to gradients of different thermodynamical intensives .
this @xmath1-non - equilibrium state does not relax to a standard bg equilibrium , but to a local @xmath1-equilibrium . in this paper
we derive the relativistic fluid - dynamical equations of motion from a non - extensive relativistic boltzmann transport equation ( nebe ) and calculate the transport coefficients employing the relaxation time approximation for the collision integral .
such calculations were done in the non - relativistic limit @xcite with a non - extensive boltzmann equation @xcite slightly different from ours .
the @xmath1-generalized first order relativistic navier - stokes - fourier equations of relativistic dissipative fluid dynamics are also derived within this framework .
we use the metric of flat space - time @xmath6diag@xmath7 and natural units throughout this work , @xmath8 .
we follow lavagno @xcite and recall the @xmath1-modified non - extensive boltzmann equation ( nebe ) compatible with tsallis suggestion for a generalized non - extensive entropy formula . extensions of the classical boltzmann equation , like the one we adopt here from @xcite , are put forward without any derivation from microscopic quantum field dynamics , with the purpose to study classical statistical effects due to correlations possibly induced by finite system size compared to the characteristic interaction range @xcite . for the sake of simplicity we neglect external field effects which simplifies the vlasov operator in the following general boltzmann - type equation : @xmath9 , \]]where @xmath10 is the single particle phase - space distribution function and @xmath11 , with @xmath12 .
@xmath13 $ ] is the collision integral .
the space - time coordinates are @xmath14 and the particle four - momentum @xmath15 is normalized to the mass @xmath16 .
the nebe is a closed equation assuming that the number of binary collisions around space - time coordinates @xmath17 is proportional to an expression @xmath18 $ ] , where the explicit form of the collision integral involving binary collision with incoming momenta @xmath19 , @xmath20 and outgoing momenta @xmath21 , @xmath22 is given as , @xmath23 & = \frac{1}{2}\int dk^{\prime } dpdp^{\prime } \notag \\ & \times w_{kk\prime \rightarrow pp\prime } \left ( h_{q}\left [ f_{p},f_{p^{\prime } } \right ] - h_{q}\left [ f_{k},f_{k^{\prime } } \right ] \right ) .\end{aligned}\ ] ] here the factor @xmath24 takes into account the symmetry of the initial or final momentum configuration in case of identical particles if we interchange the labels i.e. , @xmath25 @xmath26 or @xmath27 @xmath28 .
furthermore , @xmath29 is the lorentz - invariant momentum - space volume , with @xmath30 being the number of internal degrees of freedom , e.g. spin degeneracy , while @xmath31 is the lorentz - invariant transition rate .
the transition rate is symmetric with respect to the sequence of final states @xmath32 , as well as it is symmetric for time reversed processes @xmath33 . in this way it fulfils the detailed balance property .
the collision integral is positive for any @xmath34 and it does not change sign under time reversal . the @xmath1-generalized version of the so - called stosszahlansatz , the assumption of molecular chaos , @xmath35 \equiv \exp _ { q}\left [ \ln _ { q}\left ( f_{k}\right ) + \ln _ { q}\left ( f_{k^{\prime } } \right ) \right ] , \]]uses the @xmath1-deformed exponential and logarithm functions , @xmath36 ^{1/\left ( 1-q\right ) } , \\
\ln _ { q}\left ( x\right ) & \equiv & \frac{x^{1-q}-1}{1-q}.\end{aligned}\]]for @xmath37 the above expressions are the standard exponential and natural logarithm functions . here
we stress that for @xmath5 the detailed balance solution of the above equations obey the classical boltzmann - gibbs statistics , not the usual quantum statistics .
the entropy four - current has to be properly defined as @xmath38 , \label{kinetic_s_mu}\]]demanding positive entropy production according to the second law of thermodynamics : @xmath39 \geq 0 .
\label{kinetic_entropy_production}\]]note that lavagno defined the entropy four - current as @xmath40 $ ] , while the current form was introduced , correcting lavagno s definition , by osada and wilk @xcite .
we shall return to this point later . in order to elucidate the relation to a fluid dynamical description ,
let us define now the following momentum - space integrals of the nebe , @xmath41 , \label{kinetic_n_mu_cons } \\ \partial _ { \mu } t^{\mu \nu } & \equiv & \int dk\ k^{\nu } k^{\mu } \partial _ { \mu } \tilde{f}_{k}=\int dkk^{\mu } c\left [ f\right ] , \label{kinetic_t_mu_nu_cons}\end{aligned}\]]where we introduced the @xmath1-modified particle four - current and symmetric energy - momentum tensor , @xmath42it is straightforward to show that the left hand side of eqs .
( kinetic_n_mu_cons ) and ( [ kinetic_t_mu_nu_cons ] ) vanishes when the particle number , the energy and momentum in individual collisions are conserved , that is , if @xmath43 holds .
this property is expressed by specifying the so - called collision invariants by using , @xmath44 , such that @xmath45 & = 0 , \\
\int dk\beta _ { \mu } k^{\mu }
c\left [ f\right ] & = 0.\end{aligned}\]]it implies the macroscopic conservation laws of particle four - current and energy - momentum tensor for any solution of the nebe , @xmath46besides a non - negative entropy production from eq .
( @xmath47 ) .
the entropy production vanishes in local @xmath1-equilibrium and it leads to the collision invariant @xmath48 . the distribution function which satisfies @xmath49 is the canonical equilibrium one .
this result is equivalent with the requirement that @xmath50 , whence the @xmath1-equilibrium distribution is given by @xmath51here @xmath52 and @xmath53 are the collisional invariants given previously , and @xmath54 is a four - vector normalized to one , @xmath55 .
these quantities will be identified with the inverse temperature @xmath56 , the chemical potential over temperature , @xmath57 , and the fluid dynamical four - velocity of matter @xmath54 .
the formula in eq .
( [ q_equilibrium ] ) reduces to the well known jttner distribution or relativistic maxwell - boltzmann distribution for @xmath5 , that is @xmath58 .
therefore it is clear that in local @xmath1-equilibrium there are only five fields which completely characterize the system for any given @xmath1 . for the purpose of a quick reference
let us introduce @xmath59 , used to project an arbitrary four - vector into another four - vector orthogonal to @xmath54 .
any four - vector , @xmath60 , and in particular the four - momenta of particles can be decomposed into two parts using an arbitrary fluid dynamic flow velocity , @xmath54 : @xmath61 , where @xmath62 is the local rest frame ( lrf ) energy of the particle and @xmath63 contains the lrf momenta @xcite .
the lrf or co - moving frame is defined with @xmath64 .
moreover , the space - time gradient or four - divergence , @xmath65 , should also be decomposed into parts parallel and orthogonal to the flow
. here , the co - moving or proper time - derivative , @xmath66 , is also denoted by an over - dot , @xmath67 , and @xmath68 denotes the gradient .
for second - rank tensors the orthogonal and traceless projection is defined as @xmath69 , where @xmath70 .
now , using the local @xmath1-equilibrium distribution function we are prepared to introduce the following generalized thermodynamic integrals : @xmath71where @xmath72 are natural numbers and @xmath73 denotes the double factorial . at a given fixed temperature @xmath74 and @xmath75 ,
whence it follows that @xmath76 the above integrals for @xmath77 run over a finite range in energy , @xmath78 , and are always convergent . for @xmath79
the integrands have power - law tails satisfying a convergence criterion : @xmath80 is required for the @xmath81 integrals to be finite .
correspondingly the maximal index @xmath82 is one higher for the @xmath83 and two higher for the @xmath84 integrals .
finally , for @xmath5 the familiar jttner distribution makes all integrals convergent by its exponential tail for any finite indices . in this case
@xmath85 we retrieve the familiar relativistic thermodynamic integrals , see refs .
@xcite . note that in the paper by lavagno @xcite , the @xmath86-modified bessel function of second kind was introduced , @xmath87 , where @xmath88 , which in our notation corresponds to @xmath89 in the lrf .
it is also straightforward to show by partial integration that the following recursive relations hold in equilibrium : @xmath90furthermore , one can also show that @xmath91as well as@xmath92and similarly @xmath93 let us recall the definition of the @xmath1-modified entropy four - current and calculate the equilibrium entropy density , @xmath94 with @xmath95 for the @xmath1-equilibrium distribution function from eq .
( [ q_equilibrium ] ) . making use of the previously introduced @xmath1-generalized thermodynamic integrals
we obtain @xmath96identifying the momentum integrals , or shortly `` the moments '' , as the particle density @xmath97 , energy density @xmath98 , and pressure @xmath99 . here
we have used eq .
( [ jq_nk_rec ] ) to obtain @xmath100 which translates into the familiar ideal gas eos , @xmath101 .
this is the reason why the definition of the entropy four - current has to be given in the specific form of eq .
( [ kinetic_s_mu ] ) .
hence the fundamental thermodynamic equation follows directly from eq .
( [ eq27 ] ) : @xmath102note that @xmath103 , because @xmath104 .
however , if the number of particles is conserved , the integrals of @xmath105 and @xmath83 should lead the same number of particles per unit volume , although the normalization of the distribution functions differ .
the definition of entropy leads to the fundamental thermodynamic relation and the well known gibbs - duhem relations , @xmath106 the last equation follows from eqs .
( [ kq_nk_rec],[d_j_nk ] ) .
note that , the above thermostatic relations resemble those of classical bg thermodynamics .
kinetic theory does not specify the zeroth law of thermodynamics , therefore , whether the thermodynamic temperature is different from the temperature in kinetic theory is an open question @xcite .
however , choosing a single parameter for the temperature , the equations of @xmath1-fluid dynamics become formally identical with that of classical fluid dynamics for @xmath5 . before going further we discuss the fluid dynamical equations in @xmath1-equilibrium . the particle four - current and energy - momentum tensor
are calculated from eqs .
( [ kinetic_n_mu],[kinetic_t_mu_nu ] ) using the local @xmath1-equilibrium distribution function from eq .
( [ q_equilibrium ] ) , hence @xmath107these decompositions are formally identical with that of a perfect fluid , hence some authors called it a perfect q - fluid @xcite .
moreover , the conservation laws from eqs .
( [ n_mu_cons ] , t_mu_nu_cons ) lead , at least formally , to the well known euler equations of perfect fluid dynamics , @xmath108making a closed system of equations by supplementing with an eos .
these equations imply a vanishing local entropy production in @xmath1-local equilibrium , @xmath109 for smooth initial conditions , i.e. without discontinuities , and eos s without first order phase transitions .
note that in general the eos is not restricted to be an ideal gas eos . in our case
the eos is given as , @xmath110 , which in the massless limit returns the familiar @xmath111 and hence the speed of sound is , @xmath112 , irrespective of the @xmath1-parameter @xcite . on this account the general formula can be derived rewriting the above equations with the help of the thermodynamic integrals and relations , @xmath113 and @xmath114 . after some algebra we get that the speed of sound squared @xmath115 at fixed entropy per particle , @xmath116 ,
is given as,@xmath117where @xmath118 is the enthalpy per particle and @xmath119 , @xmath120 .
if a system is out - of - equilibrium the distribution function is different from the local@xmath121-equilibrium distribution , @xmath122 .
this will lead to additional terms in the thermodynamic quantities and increase the entropy until the system relaxes to local @xmath1-equilibrium according to the microscopic dynamics described by a nebe .
once local @xmath1-equilibrium is reached these additional dissipative quantities vanish .
making use of the previously introduced notations , the macroscopic fields , such as the particle four - current and energy - momentum tensor , can be decomposed in a general frame the following way , @xmath123where the round brackets around the greek indices denote symmetrization : @xmath124 .
therefore , we can uniquely identify the fundamental fluid dynamical quantities such as the particle density , energy density and isotropic pressure : @xmath125similarly the particle diffusion current , the energy - momentum current and the stress tensor are given by @xmath126these macroscopic fields , expressed above as momentum integrals over the non - equilibrium distribution function , @xmath127 , are part of the particle four - current and energy momentum tensor of a non - equilibrated fluid .
some of these quantities are related to their equilibrium thermodynamical pendants through the so called matching conditions .
it is standard practice to assume that the particle density and energy density are unchanged from their equilibrium values , @xmath128while the isotropic pressure , @xmath129 , separates into two parts : the thermodynamical pressure , @xmath130and the bulk viscous pressure , @xmath131such that @xmath132 .
here we introduced the deviation from equilibrium , @xmath133 .
the matching conditions require that , @xmath134 , while @xmath135 , @xmath136 and @xmath137 .
now , it is clear that only the non - equilibrium deviations lead to dissipation , since the moments of the equilibrium distribution function lead to vanishing dissipative quantities @xmath138 . for completeness
, here we calculate the entropy four - current from eq .
( kinetic_s_mu ) up to first order in deviations from equilibrium @xmath139 , by expanding the entropy four - current around @xmath140 .
this expression is given as , @xmath141where @xmath142 and we made use of the definitions from eqs .
[ def_charge_diffusion ] and [ def_momentum_diffusion ] .
note that the higher order deviations from @xmath1-equilibrium can only be calculated once @xmath143 is specified .
the first order result is independent of such details .
we recall that eckart @xcite was the first to discuss relativistic dissipative fluids in a proper way . the energy - momentum tensor decomposition proposed by him
contained no bulk viscous pressure , @xmath144 , and no particle diffusion current , @xmath145 , hence the fluid dynamical flow velocity was fixed to the conserved particles , @xmath146 .
moreover the energy - momentum current in this case was identified with the heat - flow , @xmath147 .
later landau and lifshitz @xcite introduced a different decomposition where the flow is tied to the flow of energy - momentum @xmath148 .
this defines the flow as the only time - like eigenvector of the energy - momentum tensor .
therefore the energy - momentum diffusion current vanishes , @xmath149 , in this frame .
these physically different choices are related by the general expression for the heat - flow @xmath150 .
here we showed that using the nebe we obtain formally the same results as in the case of the classical boltzmann equation , but they are generalized to include the non - extensivity of entropy .
since the equations of dissipative fluid dynamics are not closed due to the fact that @xmath127 is unknown , we need to find additional equations and relations which close the system of equations .
let us assume , following chapman and enskog @xcite , that the non - equilibrium contributions from the streaming term are negligible near the equilibrium @xmath151 @xmath140 , that is @xmath152 . in this way
@xmath153 @xmath154 and therefore all non - equilibrium moments vanish from the left hand side of the transport equation .
hence we obtain the gradients of thermodynamic quantities multiplied by transport coefficients .
we calculate the streaming term in @xmath1-equilibrium , @xmath155 from here we follow the pioneering works of anderson and witting @xcite and employ the relaxation time approximation .
this simplifies the collision integral of the nebe , @xmath156 \simeq -e_{k}\frac{\left ( \tilde{f}_{k}-\tilde{f}% _ { 0k}\right ) } { \tau _ { c } } , \label{coll_int}\]]where @xmath157 is the relaxation time .
the relaxation time can be interpreted as a mean time between collisions .
furthermore , this approximation fixes the local rest frame so that the energy flow vanishes , as it will be shown below .
note that these approximations are made in order to simplify the collision term , which in the nebe is even more complicated than in the classical case .
nevertheless , these model equations and solutions can be used to make very reasonable first estimates for the transport coefficients . replacing @xmath158 into the nebe and using the previous assumptions we get the following transport equation
, @xmath159utilizing eqs .
( [ d_f0 ] ) and ( [ coll_int ] ) we arrive at , @xmath160 .
\label{delta_f}\end{aligned}\]]here we replaced @xmath161 where @xmath162 is the expansion scalar , @xmath163 shear stress tensor and @xmath164 vorticity . in the last term @xmath165 , due to the fact that the vorticity is antisymmetric while@xmath166 is symmetric .
note that we made use of the conservation laws of perfect fluids to calculate the proper time derivatives and hence we express them in terms of gradients .
thus , applying eq .
( [ d_j_nk ] ) for different @xmath167 values we obtain @xmath168 \theta , \label{alpha_dot } \\
\dot{\beta}_{0 } & = & n_{0}\mathcal{d}_{q\left ( 2,0\right ) } ^{-1}\left [ h_{0}% \mathcal{k}_{q\left ( 1,0\right ) } -\mathcal{k}_{q\left ( 2,0\right ) } \right ] \theta , \label{beta_dot } \\
\dot{u}^{\mu } & = & \beta _ { 0}^{-1}\left [ h_{0}^{-1}\nabla ^{\mu } \alpha _ { 0}-\nabla ^{\mu } \beta _ { 0}\right ] . \label{beta_grad}\end{aligned}\ ] ] knowing the non - equilibrium distribution function , we can calculate now dissipative quantities .
the stress tensor can be obtained in a straightforward manner from eq .
( [ def_stress_tensor ] ) , together with the definitions of the integral from eq .
( [ k_i_j ] ) , as being @xmath169whence the traditional newton - navier - stokes relation between the stress tensor and shear tensor emerges : @xmath170the shear viscosity coefficient is given by @xmath171the relativistic generalization of the newtonian shear @xmath172 , and the navier - stokes relation between the stress and shear have the same form for the @xmath1-modified statistics as the ones obtained using the classical bg statistics . similarly to the shear viscosity the heat - conductivity coefficient can be calculated .
one can show that @xmath173 in the anderson - witting relaxation time - approximation . using eq .
( [ def_charge_diffusion ] ) and@xmath174 the result is @xmath175 \notag \\ & = \tau _ { c}\left ( \nabla ^{\mu } \alpha _ { 0}\right ) \left [ h_{0}\mathcal{k}% _ { q\left ( 1,1\right ) } -\mathcal{k}_{q\left ( 2,1\right ) } \right ] .\end{aligned}\]]the fourier - navier - stokes law becomes @xmath176where @xmath177 and the coefficient for heat - conductivity or thermal conductivity is given by @xmath178 finally the bulk viscous pressure also known as the volume or second viscosity , can be calculated from eq .
( [ def_bulk_pressure ] ) where we used eqs .
( [ alpha_dot],[beta_dot ] ) to express the proper time derivatives in eq .
( [ delta_f ] ) , hence @xmath179 \notag \\ & = -\tau _ { c}\frac{\theta } { 3}\left [ 5\beta _ { 0}\mathcal{k}_{q\left ( 3,2\right ) } + 3n_{0}\left ( \frac{\mathcal{k}_{30}-h_{0}\mathcal{k}_{20}}{% \mathcal{d}_{20}}\right ) \mathcal{k}_{q\left ( 2,1\right ) } \right .
\notag \\ & \left .
-3n_{0}\left ( \frac{\mathcal{k}_{20}-h_{0}\mathcal{k}_{10}}{% \mathcal{d}_{20}}\right ) \mathcal{k}_{q\left ( 3,1\right ) } \right ] .\end{aligned}\]]from this the relativistic version of the classical stokes result follows immediately : @xmath180with the bulk viscosity coefficient , @xmath181 .
\label{bulk_visc_coeff}\end{aligned}\]]once again we note that the above results formally resemble those of the bg statistics , the only explicit @xmath1-dependence occurs in the definitions of the @xmath1-modified thermodynamic integrals .
all formulas show that the transport coefficients are directly proportional to the mean free time between collisions and so inversely proportional to the cross section .
having defined the dissipative quantities , the deviation from the @xmath1-equilibrium eq .
[ delta_f ] can be written as , @xmath182where @xmath183 , @xmath184 and @xmath185 are tedious expression inversely proportional to the respective transport coefficients , and may be calculated from the above formulas .
in this section we investigate the previously obtained results in the lrf where @xmath186 .
the thermodynamic integrals are calculated in hyperspherical coordinates , @xmath187 , @xmath188 , @xmath189 , @xmath190 and @xmath191 .
note that the integrals for @xmath192 are calculated up to infinity while for @xmath77 the upper limit of integration is @xmath193 .
furthermore , the chemical potential is set to zero @xmath194 , while the spin degeneracy is @xmath195 .
distribution as function of longitudinal momenta for different @xmath1-parameters : @xmath196 ( thick line ) , @xmath5 ( thin line ) , @xmath197 ( dashed line ) and @xmath198 ( dotted line ) .
the mass of particles as well as the temperature are @xmath199 mev . ] in fig .
[ fig : fq_ek ] we plot the natural logarithm of the @xmath1-th power of the @xmath1-equilibrium distribution function from eq .
( [ q_equilibrium ] ) in the lrf , with @xmath200 mev and @xmath201 mev for longitudinal momenta @xmath202 $ ] gev and transverse momenta @xmath203 .
the functions are plotted for @xmath204 ( thick line ) , @xmath5 ( thin line ) , @xmath197 ( dashed line ) and @xmath198 ( dotted line ) .
note that for @xmath196 the function is practically cut - off at @xmath205 . as a function of temperature for different @xmath1-parameters : @xmath196 ( thick line ) , @xmath5 ( thin line ) , @xmath197 ( dashed line ) and @xmath198 ( dotted line ) . ] ( a ) and the shear viscosity coefficient over entropy density divided by the relaxation time @xmath206 ( b ) as function of temperature for different @xmath1-values.,title="fig : " ] ( a ) and the shear viscosity coefficient over entropy density divided by the relaxation time @xmath206 ( b ) as function of temperature for different @xmath1-values.,title="fig : " ] next we plot the speed of sound squared @xmath207 as a function of temperature and @xmath1-parameter from eq .
( [ cs_2 ] ) .
the mass of particles is fixed at @xmath200 mev , while the temperature is @xmath208 $ ] mev .
we see that for any value of @xmath1 the speed of sound is increasing with temperature while it saturates for @xmath209 to the ultra - relativistic limit , @xmath210 independently of @xmath1 .
this is a straightforward result from eq .
( [ cs_2 ] ) . the speed of sound is larger for larger @xmath1 at any given temperature
. -values .
, title="fig : " ] -values .
, title="fig : " ] the shear viscosity divided by the relaxation time from eq .
( [ shear_visc_coeff ] ) , @xmath211 , is shown in fig .
[ fig : eta]a . on the other hand
the shear viscosity to entropy ratio divided by the relaxation time @xmath212 can be inspected for different values of @xmath1 in fig .
[ fig : eta]b .
we observe that the viscosity increases with increasing temperature for all @xmath213-values .
this behaviour is more pronounced for larger @xmath1-values , hence there is a monotonic increase in the viscosity coefficient with increasing @xmath214 at a given temperature . at the same time it turns out that the shear viscosity to entropy ratio at a given temperature is almost insensitive to different values of @xmath1 .
[ fig : eta]b demonstrates that for a given relaxation time @xmath157 the @xmath215 ratio does not change significantly and at high temperatures the @xmath1-dependence diminishes .
this also means that the ratio of dissipative to equilibrium quantities such as @xmath216 does not change at high temperatures .
this is due to the fact that the entropy density also increases with increasing @xmath1-values by about the same factor . in the classical maxwell - boltzmann massless limit @xmath217 and @xmath218
we recover the familiar results @xcite . similarly to the viscosity coefficient
, we also plot the coefficient for the heat - conductivity from eq .
( [ heat_cond_coeff ] ) , as well as the ratio of heat - conductivity to entropy density divided by the relaxation time . both of these quantities , shown in figs .
[ fig : kappa]a and [ fig : kappa]b , are sensitive to the value of the parameter @xmath1 .
furthermore , we observe that the heat - conductivity increases with increasing @xmath1-values , and - other than before - the ratio @xmath219 increases in a noticeable fashion . the last figures , figs .
[ fig : zeta]a and fig .
[ fig : zeta]b show the bulk viscosity coefficient divided by the relaxation time , and the bulk viscosity coefficient to entropy density ratio divided by the relaxation time .
-values.,title="fig : " ] -values.,title="fig : " ] we observe that the value of bulk viscosity coefficient is very small .
it is about @xmath220 times smaller than the coefficient of shear viscosity @xmath221 , which corroborates the classical result that the bulk viscosity in the non - relativistic or ultra - relativistic massless limit vanishes , while it is much smaller than the shear viscosity in between for massive particles @xcite . on the other hand
the bulk viscosity to entropy density ratio does show a monotonic increase with increasing @xmath1 but only for @xmath222 .
for @xmath223 the behaviour is reversed due to the fact that the entropy density increases with temperature at a slower rate compared to the viscosity .
in this paper we have calculated the transport coefficients from a @xmath1-modified boltzmann equation .
the results show that all transport coefficients increase monotonically with increasing @xmath1-values for relativistic ideal gases .
the speed of sound is also increasing with increasing @xmath1 but it saturates quickly to the well known ultra - relativistic limit independent of @xmath1 .
furthermore , we derived the @xmath1-generalized versions of the classical navier - stokes - fourier equations of relativistic dissipative fluid dynamics from a @xmath1-generalized boltzmann transport equation .
these equations relate the dissipative quantities to the thermodynamical forces linearly with positive transport coefficients which dependent on parameter @xmath1 .
the resulting equations , are nevertheless parabolic , hence acausal and unstable @xcite .
hyperbolicity related problems are not solved by introducing the parameter @xmath1 .
this manqu can be resolved by an appropriate choice of the relaxation time in the evolving fluid @xcite .
superior but more complicated equations of motion may be derived from the nebe as in the classical theory , with methods different from the one presented in this work , see for example refs . @xcite .
the authors thank for g.s .
denicol , h. niemi , l.p .
csernai , m. gyulassy for discussions and p. vn for reading the manuscript and for valuable comments .
this work was supported by the hungarian national scientific fund otka k68108 .
e.m . was supported by the national development agency otka / nf 81655 .
a. jakovac , phys . rev . *
d81 * , 045020 ( 2010 ) .
l. p. csernai , j. .i .
kapusta , l. d. mclerran , phys .
lett . * 97 * , 152303 ( 2006 ) .
z. xu , c. greiner , phys .
lett . * 100 * , 172301 ( 2008 ) .
j. noronha - hostler , j. noronha , c. greiner , phys .
lett . * 103 * , 172302 ( 2009 ) .
t. koide , e. nakano , t. kodama , phys . rev . lett . * 103 * , 052301 ( 2009 ) .
p. chakraborty , j. i. kapusta , phys .
rev . * c83 * , 014906 ( 2011 ) .
m. bluhm , b. kampfer , k. redlich , phys .
rev . * c84 * , 025201 ( 2011 ) .
h. niemi , g. s. denicol , p. huovinen , e. molnar , d. h. rischke , phys .
* 106 * , 212302 ( 2011 ) .
t. kodama , t. koide , eur .
j. * a40 * , 289 - 297 ( 2009 ) ; g. wilk , z. wlodarczyk , eur .
j. * a40 * , 299 - 312 ( 2009 ) ; w. m. alberico , a. lavagno , eur .
j. * a40 * , 313 - 323 ( 2009 ) ; t. s. biro , g. purcsel , k. urmossy , eur . phys .
j. * a40 * , 325 - 340 ( 2009 ) .
t. s. biro , a. jakovac , phys .
lett . * 94 * , 132302 ( 2005 ) .
t. s. biro , g. purcsel , phys .
lett . * 95 * , 162302 ( 2005 ) .
g. wilk , braz .
j. phys .
* 37 * , 714 ( 2007 ) .
m. shao , l. yi , z. tang , h. chen , c. li , z. xu , j. phys .
g * g37 * , 085104 ( 2010 ) .
z. tang , y. xu , l. ruan , g. van buren , f. wang , z. xu , phys . rev . *
c79 * , 051901 ( 2009 ) .
f. siklr epj web of conferences * 13 * , 03002 ( 2011 ) .
t. osada , g. wilk , phys .
rev . * c77 * , 044903 ( 2008 ) .
t. osada , g. wilk , prog .
suppl . * 174 * , 168 ( 2008 ) .
t. osada , g. wilk , central eur .
j. phys .
* 7 * , 432 ( 2009 ) .
t. osada , phys .
rev . * c81 * , 024907 ( 2010 ) .
j. l. anderson , j. math .
phys . * 15 * no.7 , 1116 ( 1974 ) .
|
we derive equations for fluid dynamics from a non - extensive boltzmann transport equation consistent with tsallis non - extensive entropy formula .
we evaluate transport coefficients employing the relaxation time approximation and investigate non - extensive effects in leading order dissipative phenomena at relativistic energies , like heat conductivity , shear and bulk viscosity .
|
capillarity is an important physical phenomena , which occurs when two different materials contact and do not mix . given a container @xmath0 with an incompressible liquid drop @xmath1 in it , the interface of the liquid and
the air is a capillary surface @xmath2 . in absence of gravity ,
the interface @xmath2 is of constant mean curvature and the contact angle of @xmath2 to the boundary @xmath3 is constant .
one should compare this setting with soap bubble ( resp .
soap film ) , where the surface has no boundary ( resp .
fixed boundary ) and constant mean curvature .
the literature for the study of capillarity is extensive and we refer to the book of finn @xcite , where the treatment of the theory is mainly in the nonparametric case and in the more general situation of presence of gravity .
also we mention @xcite for a more recent survey about this topic .
in this paper we are concerned with the special case that the container @xmath0 is a unit euclidean ball and no gravity is involved .
we study the ( weak ) stability for capillary hypersurfaces .
this problem has been discussed by ros and souam @xcite , where they dealt with surface case and obtained some topological and geometrical restrictions . for the hypersurface case with free boundary ( the contact angle is @xmath4 ) ,
ros and vergasta also proved some interesting results in @xcite . applying the same argument as in the proof of proposition 1.1 in @xcite , we know that the hyperplanes and the spherical caps in a unit euclidean ball are capillarily stable .
recently , marinov @xcite proved that , for surface case , all other known examples are unstable .
we generalize marinov s result to the hypersurface case .
in fact we prove the following theorem .
[ thm1 ] let @xmath5 be an immersed capillary hypersurface in a unit euclidean ball @xmath6 and @xmath7 the wetted part of the boundary of the ball .
denote by @xmath1 the generalized body enclosed by @xmath8 and @xmath7 .
if the mass center of @xmath1 is at the origin , the capillary hypersurface @xmath2 is unstable . here
since we assume @xmath2 is immersed , @xmath8 may have self - intersections .
thus we need to consider the generalized body @xmath1 .
when @xmath2 is embedded , @xmath1 is understood in the common sense .
see remark [ rem1 ] below for more explanation . for @xmath9
, our theorem [ thm1 ] reduces to marinov s result in @xcite .
we also note that his argument relies on the conformal coordinates , which can not be generalized to the higher dimensional case . applying theorem [ thm1 ] to delaunay hypersurfaces , we get the following corollary [ cor1 ] . recall that delaunay hypersurfaces are the ones of revolution with constant mean curvature . by proposition 4.3 in @xcite ,
delaunay hypersurfaces are classified as : an unduloid , cylinder , nodoid , sphere , catenoid , or a hyperplane . to guarantee the portion of a delaunay hypersurface in a euclidean ball
is also capillary , it shall have some symmetry .
see section [ sec2 ] below for more details . and in that case , we call it delaunay capillary hypersurface . from theorem [ thm1 ]
we have [ cor1 ] the only stable delaunay capillary hypersurface @xmath10 in a unit euclidean ball @xmath6 is a totally geodesic hyperplane or a spherical cap .
our approach for proving theorem [ thm1 ] is as follows . in higher dimensional case
, we find that we can construct a conformal killing vector field @xmath11 $ ] for any fixed @xmath12 from the natural conformal transformation family on @xmath6 .
using the normal part @xmath13,n\ra$ ] as the test function we can define a symmetric quadratic form @xmath14 . by summing @xmath15 over @xmath16 coordinate directions we find @xmath15 has at least one negative eigenvalue .
this summation technique can be compared with j. simons work @xcite . at last , under the hypothesis of theorem [ thm1 ] we can derive the instability of the hypersurface .
our argument indicates that this conformal field is very important and we can use it to conclude that the mass center of minimal submanifolds with free boundary in a unit euclidean ball is at the origin ( see proposition [ prop1 ] ) .
we refer the readers to @xcite for the very recent work on the minimal submanifolds with free boundary .
at last , as an application of our argument , we give a new proof of the classical result due to barbosa and do carmo @xcite , which states that the only closed stable immersed hypersurface of constant mean curvature in @xmath17 is the round sphere .
an outline of this paper is as follows . in section [ sec2 ] after fixing some notations and definitions , we prove the stability of hyperplanes and spherical caps
. then we construct the crucial conformal vector field .
we also review some known facts about the delaunay hypersurfaces . in section [ sec3 ]
we give the proof of theorem [ thm1 ] . in last section
, we discuss some applications of our method .
let @xmath5 be an orientable immersed hypersurface in the unit euclidean ball @xmath18 with @xmath19 and @xmath20 .
suppose @xmath21 such that @xmath22 . and denote by @xmath23 the part of ball satisfying @xmath24 .
[ rem1 ] if @xmath8 has self - intersections , @xmath1 may be viewed as the finite union of some domains @xmath25 , i.e. @xmath26 . here @xmath27 may intersect with each other . if there are not one choices for @xmath28 , choose one and fix it . in the proof we will see that only the property @xmath24 is needed . and if there is no confusion , we will write @xmath2 ( resp . @xmath29 ) for @xmath8 ( resp .
@xmath30 ) for simplicity . let @xmath31 be the unit normal of @xmath2 pointing inwards to @xmath1 and @xmath32 the unit outward normal of @xmath33 .
denote by @xmath34 and @xmath35 the conormals of @xmath29 in @xmath2 and @xmath7 , respectively .
let @xmath36 ( resp .
@xmath37 ) be the connection of @xmath17 ( resp .
@xmath2 ) .
then the second fundamental form of @xmath2 in @xmath17 is given by @xmath38 for @xmath39 .
when taking an orthonormal basis @xmath40 on @xmath41 , we also denote by @xmath42 the components @xmath43 .
so the mean curvature @xmath44 of @xmath2 is @xmath45 . and
the second fundamental form of @xmath3 in @xmath17 is given by @xmath46 for @xmath47 .
at last let @xmath48 be the angle between @xmath34 and @xmath35 .
see figure [ pic ] for an illustration . ] following @xcite
, we discuss the variation of @xmath2 .
an admissible variation of @xmath5 is a differentiable map @xmath49 so that @xmath50 given by @xmath51 is an immersion satisfying @xmath52 and @xmath53 for all @xmath54 , and @xmath55 .
now for given @xmath56 , we define a energy functional @xmath57 where @xmath58 denotes the area function . and the volume functional can be defined as @xmath59\times m } x^*dv,\ ] ] where @xmath60 is the standard volume element of @xmath17 . under these constraints
, we have an immersed hypersurface @xmath5 is called capillary if @xmath61 for any admissible volume - preserving variation of @xmath62 . note
that we have the following formulas , @xmath63 where @xmath64 is the variational vector field @xmath65 , @xmath66 its normal component @xmath67 , and @xmath68 and @xmath69 are the corresponding area elements . from these formulas
we see that @xmath2 is capillary if and only if it has constant mean curvature and makes constant contact angle @xmath70 with @xmath3 .
furthermore , one can compute the second derivative at @xmath71 with respect to an admissible volume - preserving variation to get ( see e.g. the appendix of @xcite ) @xmath72 where @xmath73 , @xmath74 is the ricci curvature of the ambient space and @xmath75 in our setting , @xmath76 and @xmath77 . a capillary hypersurface @xmath2 is called ( weakly ) stable if @xmath78 for all @xmath79 .
in the sequel we will denote by @xmath80 the quantity @xmath81 for a given function @xmath82 .
first we prove the stability of totally geodesic capillary hypersurfaces and spherical caps .
the proof is similar to that of proposition 1.1 in @xcite .
we include it for completeness .
let @xmath18 be a unit euclidean ball .
then totally geodesic capillary hypersurfaces and spherical caps are stable .
first assume @xmath2 is a totally geodesic capillary hypersurface , i.e. , an @xmath83-dimensional ball @xmath84 with radius @xmath85 in @xmath6 .
then the contact angle @xmath70 satisfies @xmath86 . by the definition of stability
, we have to prove @xmath87 consider now the @xmath16-dimensional ball @xmath88 of radius @xmath85 having @xmath2 as an equatorial totally geodesic hypersurface .
then by @xcite , @xmath2 is area minimizing for partitioning problem in @xmath88 .
thus @xmath2 is stable in @xmath88 , which is equivalent to the inequality .
next assume @xmath2 is a spherical cap in @xmath6 with @xmath85 the radius of the sphere containing @xmath2 and @xmath70 the contact angle .
consider the @xmath83-dimensional hyperplane @xmath89 containing @xmath29 .
then @xmath2 is a capillary hypersurface in a halfspace with a contact angle @xmath90 .
by @xcite , @xmath2 is stable in the halfspace , which means @xmath91 elementary calculation leads to @xmath92 now and together yield the stability of @xmath2 in @xmath6 .
now we construct a conformal vector field .
fix a vector @xmath93 .
then ( see e.g. section @xmath94 in @xcite ) @xmath95 defines a map from @xmath6 to @xmath6 and from @xmath96 to @xmath96 , since we have @xmath97 moreover @xmath98 is conformal .
in fact , by a direct calculation we can check that @xmath99 note that @xmath100 , @xmath101 , @xmath98 fixes two points @xmath102 and @xmath103 is an identity .
next fix @xmath104 .
let @xmath105 with @xmath106 .
then @xmath107 is a family of conformal transformations with parameter @xmath54 .
thus @xmath108 determines a conformal vector field @xmath11 $ ] as follows .
@xmath109=\f{d}{dt}\bigg|_{t=0}f_t(x)=-(1+|x|^2)\xi+2\langle \xi , x\rangle x.\ ] ] note that @xmath11 $ ] is tangential along the sphere @xmath96 , since for @xmath110 , @xmath111,x\ra=-(1+|x|^2)\la \xi , x\ra+2\la \xi , x\ra |x|^2=0.\ ] ] in this subsection , following @xcite we review some facts about delaunay hypersurfaces , which is rotational and of constant mean curvature @xmath44 .
these hypersurfaces are the models we are concerned with in theorem [ thm1 ] .
let @xmath112 be a hypersurface which is invariant under the action of the orthogonal group @xmath113 fixing the @xmath114-axis .
assume @xmath2 is generated by a curve @xmath115 contained in the @xmath116-plane .
then it suffices to determine the curve @xmath115 .
parametrize the curve @xmath117 by arc - length @xmath118 .
denote by @xmath119 the angle between the tangent to @xmath115 and the positive @xmath114-direction and choose the normal vector @xmath120 .
then @xmath121 satisfies the following system of ordinary differential equations @xmath122 the first integral of this system is given by @xmath123 where the constant @xmath124 is called the force of the curve @xmath115 and it together with @xmath44 will determine the curve as follows .
( see proposition 4.3 in @xcite ) the curve @xmath115 and the hypersurface @xmath2 generated by @xmath115 have the following several possible types . 1 .
if @xmath125 then @xmath115 is a periodic graph over the @xmath114-axis .
it generates a periodic embedded unduloid , or a cylinder .
2 . if @xmath126 then @xmath115 is a locally convex curve and @xmath2 is a nodoid , which has self - intersections .
3 . if @xmath127 and @xmath128 then @xmath2 is a sphere .
if @xmath129 and @xmath130 we obtain a catenary which generates an embedded catenoid @xmath2 with @xmath131 if the normal points down and @xmath132 if the normal points up . 5 .
if @xmath129 and @xmath127 then @xmath115 is a straight line orthogonal to the @xmath114-axis which generates a hyperplane . 6 . if @xmath2 touches the @xmath114-axis , then it must be a sphere or a hyperplane .
the curve @xmath115 is determined , up to translation along the @xmath114-axis , by the pair @xmath133 . from this proposition ,
it is easy to see if @xmath10 is the portion of an unduloid , cylinder , nodoid or a catenoid in a unit euclidean ball @xmath6 with revolution axis @xmath114 and moreover @xmath2 is symmetric with respect to the hyperplane @xmath134 , then @xmath2 is a capillary hypersurface in @xmath6 . in that case
we call them delaunay capillary hypersurfaces in @xmath6
. furthermore the generalized body @xmath1 enclosed by @xmath2 and the wetted part of the sphere has the mass center at the origin .
so theorem [ thm1 ] is applicable .
with the preparations above , we can define a `` test function '' @xmath135=\langle y[\xi],n\rangle=\langle -(1+|x|^2)\xi+2\langle \xi , x\rangle x , n\rangle.\ ] ] we mention that we will also use the following expression of @xmath136 $ ] @xmath137=\langle \xi , -(1+|x|^2)n+2\langle x , n\rangle x\rangle.\ ] ] recall the second variational formula @xmath138 where @xmath139 and @xmath140 .
now we can prove the following lemmas .
[ lem3 ] @xmath34 is a principal direction for @xmath141 along @xmath29 .
in particular , @xmath142 .
it suffices to prove that , for @xmath143 , @xmath144 .
in fact , we have @xmath145 where we used @xmath70 is constant , @xmath35 and @xmath32 are unit vectors , and @xmath3 is totally umbilical
. thus we complete the proof of lemma [ lem3 ] .
[ lem1 ] along @xmath29 , there holds @xmath146 first from and lemma [ lem3 ] we have @xmath147 where in the third line we used @xmath148 along @xmath29 .
next noticing that @xmath149 , we get @xmath150 on the other hand , there holds @xmath151 where again in the second line we used @xmath148 along @xmath29 .
hence we obtain @xmath146 the next lemma , which indicates the geometric meaning of lemma [ lem1 ] , may have its own interest .
thus we also include it here . under the flow @xmath108 , there holds @xmath152 in particular , since @xmath108 is conformal ( angle preserving ) , @xmath153 . following @xcite ,
we denote by a `` prime '' the convariant derivative @xmath154 . also by the appendix of @xcite
, we have @xmath155 where @xmath156 denotes the gradient on @xmath29 , @xmath157 ( resp .
@xmath158 ) the tangent part of the variational vector field @xmath64 of @xmath2 ( resp . to @xmath29 ) , @xmath159 the shape operator of @xmath2 in @xmath17 with respect to @xmath31 , and @xmath160 ( resp .
@xmath161 ) the shape operator of @xmath29 in @xmath2 ( resp .
@xmath3 ) with respect to @xmath34 ( resp .
@xmath35 ) . note that @xmath162 , which implies @xmath163 taking into account that @xmath164 we have @xmath165 or @xmath166 again from the appendix of @xcite , there hold @xmath167 now plugging these equalities into , the lemma follows immediately .
@xmath168 the proof is a direct calculation using moving frame method .
this method is very powerful in differential geometry .
take an orthonormal basis @xmath169 .
then we have the structure equations : ( see e.g. @xcite ) @xmath170 where @xmath171 is the dual forms and @xmath172 the connection forms . thus there holds @xmath173 note that @xmath174 and using codazzi equation @xmath175 we have @xmath176 moreover we can get @xmath177 now substituting all these terms into gives rise to @xmath178 therefore , @xmath179 thus we obtain @xmath180 to analyze @xmath80 , we define a quadratic form @xmath181 for @xmath182 .
denote by @xmath183 the standard coordinate vectors in @xmath17 .
then we have the following lemma . [ lem2 ] @xmath15 has the following properties . 1
. @xmath15 is symmetric .
2 . @xmath184 with equality if and only if @xmath185 on m. \(1 ) first we prove @xmath15 is symmetric
. note that in fact @xmath15 is defined as @xmath186)\cdot \phi[\xi_2 ] da.\ ] ] then green s formula implies @xmath187\cdot l(\phi[\xi_2 ] ) da+\int_{\pt m } \left(\phi[\xi_1](\phi[\xi_2])_\nu-(\phi[\xi_1])_\nu\phi[\xi_2]\right)ds.\\\end{aligned}\ ] ] but lemma [ lem1 ] yields @xmath188)_\nu = q\phi[\xi_i]$ ] , @xmath189 .
so the boundary term vanishes and then @xmath190 \(2 ) next we calculate @xmath191 .
@xmath192 also we have @xmath193 .
consequently , @xmath194 where we have used @xmath148 on @xmath29 to remove the boundary term . and
it is easy to see @xmath195 if and only if @xmath185 .
so we complete the proof .
thus @xmath15 has at least one negative eigenvalue .
but on the other hand , @xmath196=\sum_{a=1}^{n+1}\la d_{\pt_a}y[\xi],\pt_a\ra=2(n+1)\langle \xi , x\rangle,\ ] ] which by integration implies @xmath197,n\rangle da\nonumber\\ & = -\int_t div_{\r^{n+1}}y[\xi ] dv+\int_\omega \langle y[\xi],\b{n } \rangle da\nonumber\\ & = -2(n+1)\int_t \langle \xi , x\rangle dv.\end{aligned}\ ] ] so generally @xmath198 .
that means @xmath136 $ ] is not a test function .
however , under the hypothesis of theorem [ thm1 ] that the mass center of @xmath1 is at the origin , we have @xmath199 for @xmath200 .
so if we choose @xmath201 as an eigenvector corresponding to the negative eigenvalue of @xmath15 , we have @xmath202)=q(\xi,\xi)<0 $ ] , which implies that @xmath2 is unstable .
this completes the proof of theorem [ thm1 ] .
in this section we shall give several applications of the above argument and propose a conjecture on the topic .
assume @xmath15 is diagonalized such that @xmath203 and @xmath204 are the eigenvectors corresponding to the two negative eigenvalues .
then for real numbers @xmath205 and @xmath206 with @xmath207 , @xmath208 on the other hand , @xmath209 da&=-2(n+1)\int_t \langle c_1\xi_1+c_2\xi_2,x\rangle dv\\ & = -2(n+1)(c_1\int_t \langle \xi_1,x\rangle dv+c_2\int_t \langle \xi_2,x\rangle dv).\end{aligned}\ ] ] so we can always choose proper @xmath205 and @xmath206 with @xmath207 such that @xmath210 da=0 $ ] . then using @xmath211 $ ] as a test function , from we know @xmath2 is unstable .
the significance of the above proposition is as follows . for a given concrete capillary hypersurface @xmath2 in @xmath6 ,
the quadratic form @xmath15 is computable in principle .
then if @xmath15 has two negative eigenvalues , we can assert its instability . also from this proposition
we know that for hyperplanes and spherical caps @xmath15 has exactly one negative eigenvalue . along
@xmath213 choose the orthonormal basis @xmath214 such that @xmath215 .
then we have @xmath216^t = div_m(y[\xi]-\sum_\alpha \la y[\xi],e_\alpha \ra e_\alpha)=2k\la\xi , x\ra+\la y[\xi ] , k\vec{h}\ra=2k\la\xi , x\ra.\end{aligned}\ ] ] by divergence theorem , we have @xmath217^t,\nu\ra ds=\int_{\pt m } \la y[\xi],x\ra ds=0,\ ] ] where we have used the fact @xmath11 $ ] is tangential to @xmath33
. by translation , assume the mass center of generalized body @xmath1 enclosed by @xmath2 is at the origin .
so @xmath221da=0 $ ] for all @xmath104 .
if @xmath2 is the round sphere , we are done .
otherwise @xmath222 .
so by lemma [ lem2 ] the quadratic form @xmath15 has a negative eigenvalue .
choosing @xmath201 as an eigenvector corresponding to the negative eigenvalue , we have @xmath223)=-\int_m l\phi[\xi]\cdot \phi[\xi]da = q(\xi,\xi)<0,\ ] ] which shows that @xmath2 is unstable . 1 . for @xmath225 , @xmath129 and @xmath226 ,
@xmath2 must be totally geodesic @xcite .
2 . for @xmath9 and @xmath226
, @xmath2 is a totally geodesic disk , a spherical cap or a surface of genus @xmath227 with embedded boundary having at most two connected components @xcite .
3 . for @xmath9 and @xmath129 ,
@xmath2 is a totally geodesic disk or a surface of genus @xmath227 with at most three connected boundary components @xcite .
|
we study the stability of capillary hypersurfaces in a unit euclidean ball .
it is proved that if the mass center of the generalized body enclosed by the immersed capillary hypersurface and the wetted part of the sphere is located at the origin , then the hypersurface is unstable .
an immediate result is that all known examples except the totally geodesic ones and spherical caps are unstable .
|
in this paper , we will discuss selected aspects of the effects of quantum spacetime geometry fluctuations , using the riemann tensor correlation function as our basic tool .
it is first useful to make a distinction between active fluctuations , which arise from the dynamical degrees of freedom of gravity itself , and passive fluctuations , which are driven by quantum stress tensor fluctuations of matter fields . in general , both types of fluctuations are present .
we will be concerned with the case where the fluctuations around a classical background spacetime are small , so we can consider each type separately and add their effects .
of course , a full treatment of active fluctuations would require a more complete theory of quantum gravity than currently exists , but we restrict our attention to quantized linear perturbations of the background spacetime .
various aspects of spacetime geometry fluctuations have been discussed by several authors in recent years , for example @xcite . in particular ,
stochastic gravity @xcite has been studied as a systematic approach to go beyond the semiclassical theory . in the present paper , we will concentrate on two issues .
the first is some phenomenological effects of geometry fluctuations in a nearly flat background , which will be discussed in section [ sec : rtf ] . here
we will be summarising work that was previously published @xcite .
the second issue will be the possibility of enhancing horizon fluctuations by use of squeezed states , discussed in section [ sec : enhanced ] . here
we will be summarising work that will developed in more detail in a forthcoming paper @xcite .
in this section , we will briefly review some of the operational signatures of spacetime geometry fluctuations , and describe how these may be given a geometrical description in terms of the riemann tensor correlation function . here
we direct our attention to the case of a nearly flat background , but the same techniques can be used in more general spacetimes . the image of a source viewed through a fluctuating medium can undergo variations in apparent brightness . this effect is known to astronomers as scintillation , and to the general public as `` twinkling '' .
the most familiar example arises when stars are viewed through the earth s atmosphere , which is undergoing density fluctuations . in the case of fluctuations of the spacetime geometry
, this effect may be studied using the raychaudhuri equation as a langevin equation @xcite .
one calculates fluctuations of the expansion @xmath0 . in the case
that the shear , vorticity , and squared expansion can be neglected , this equation becomes @xmath1 where @xmath2 is an affine parameter , @xmath3 is the tangent to the geodesic congruence , and @xmath4 is the ricci tensor . if the expansion vanishes at @xmath5 , then the variance of the expansion becomes @xmath6 where @xmath7 is the ricci tensor correlation function .
it is obtained by appropriate contractions of the riemann tensor correlation function and is determined by the stress tensor correlation function .
thus we see that luminosity fluctuations , in lowest order , are signatures of passive fluctuations .
the results of several explicit models are given in @xcite .
two other effects which can arise when a source is viewed through a fluctuating spacetime geometry are broadening of spectral lines and blurring of images . both of these effects may be given a unified geometric treatment , which we summarise here .
consider a source which is sending successive pulses to a detector as illustrated in figure [ fig : spacetime ] .
+ the effect of the spacetime geometry is to cause a shift @xmath8 between successive pulses .
this shift can be expressed as an integral of the riemann tensor over the region illustrated : @xmath9 a shift in the time component is observed as a frequency shift , whereas as shift in a spatial component produces a shift in position of the image .
fluctuations of the geometry lead to variations in each of these quantities . denote the fractional line broadening by @xmath10 .
its variance can be expressed as @xmath11 here @xmath12 is the riemann tensor correlation function .
similarly the fluctuations of an image s angular position in a direction defined by the spacelike vector @xmath13 are given by .
several explicit examples of these effects , including enhanced fluctuations due to gravitons in squeezed states , are discussed in @xcite .
+ in this section , we will briefly review hawking s derivation @xcite of black hole evaporation .
the basic idea is to consider the spacetime of a black hole formed by gravitational collapse , as illustrated in figure [ fig : collapse ] . a quantum field propagating in this spacetime
is assumed to be in the in - vacuum state , that is , containing no particles before the collapse . in the case of a massless field ,
a purely positive frequency mode proportional to @xmath15 leaves @xmath16 , propagates through the collapsing body , and reaches @xmath17 after undergoing a large redshift in the region outside of the collapsing matter . at @xmath17 ,
the mode is now a mixture of positive and negative frequency parts , signalling quantum particle creation .
of special interest are the modes which leave @xmath16 just before the formation of the horizon , which is the @xmath18 ray .
these modes give the dominant contribution to the outgoing flux at times long after the black hole has formed . after passing through the collapsing body , they are @xmath19= constant rays , where @xmath20 where @xmath21 is the black hole s mass , and @xmath22 is a constant .
the logarithmic dependence leads to a planckian spectrum of created particles .
it also leads to what some authors call the `` transplanckian problem '' .
we prefer to refer it as the `` transplanckian issue '' , as in our view there may not be any problem . in any case , this is the enormous frequency which the dominant modes must have when they leave @xmath16 .
the typical frequency of the radiated particles reaching @xmath17 midway through the evaporation process is of order @xmath23 , but the typical frequency of these modes at @xmath16 is of order @xmath24 where @xmath25 is the planck mass . another way to state this is to note that the characteristic value of @xmath19 for these modes is of order @xmath26 a geodesic observer who falls from rest at large distance from the black hole will pass from @xmath27 to the horizon at @xmath28 in a proper time of @xmath29 which is far smaller than the planck time . in this sense , the outgoing modes are much less than a planck length outside the horizon .
several authors , especially unruh @xcite and jacobson @xcite have noted that it is possible to avoid the use of transplanckian modes provided that the dispersion relation for the quantum field is suitably modified .
this can lead to the phenomenon of `` mode regeneration '' , whereby the modes which become occupied by the thermal radiation do not arise by red - shifting of transplanckian modes , but rather are generated by nonlinear effects just before they are due to be occupied .
this possibility has the benefit of avoiding transplankian frequencies , but the drawback that it requires violation of local lorentz invariance , and hence would depend upon as yet undiscovered physics . for the case of a bosonic field ,
spontaneous emission implies the possibility of stimulated emission .
however , in the picture described by the original hawking derivation , stimulated emission into modes with wavelengths of order @xmath23 require one to start with an initial state which is populated by particles with transplanckian frequencies .
the possibility of stimulated emission in the case of the hawking effect was first discussed by wald @xcite .
we now turn to the central issue of this paper , that of black hole horizon fluctuations . in classical general relativity ,
the event horizon is the history of the light ray which is marginally trapped by the black hole .
it gives a sharp boundary between events which are visible to the outside , and those which are not .
however , in quantum physics , one does not expect such sharp boundaries , but rather a smeared horizon due to quantum fluctuation effects .
several authors have discussed this possibility from different viewpoints @xcite .
one approach is to calculate fluctuations in the horizon area @xcite .
another @xcite is to study the variation in the times at which rays just outside the classical horizon arrive at a distant observer .
it is the latter approach which will adopted in the present paper .
a crucial question is whether quantum horizon fluctuations will significantly alter the hawking derivation outlined above . given that the outgoing modes which eventually become populated with the thermal radiation spend part of their history less than a planck length outside the mean horizon , it would seem at first that horizon fluctuations could drastically change the result .
fluctuations on the planck scale would seem to cause a significant fraction of the outgoing modes either to fall into the black hole and never reach @xmath17 , or else to be ejected prematurely . in either case
, the thermal character of the hawking radiation would seem to be greatly changed .
this would upset the elegant connection between black hole physics , thermodynamics , and quantum theory .
in fact , the situation is not so dire as it first seems .
ford and svaiter @xcite gave an estimate of the effects of the active fluctuations when the graviton field is in its vacuum state .
they found that the effects are quite small for black holes whose mass is above the planck scale .
specifically , they calculated variations in the time @xmath30 given in ( [ eq : del - tau ] ) due to the quantum metric fluctuations and found that these fractional variations are small so long as @xmath31 .
squeezed states are capable of enhancing quantum fluctuations beyond the vacuum level .
this raises the possibility of increasing the horizon fluctuations of a black hole by sending in quantum fields in a squeezed vacuum state . in the remainder of this paper
, we will discuss three models for the source of the fluctuations .
however , first we outline a geometric approach to study the effects of geometry fluctuations on the horizon .
the basic idea is to study the geodesic deviation of the outgoing null geodesics just outside the mean horizon .
+ this is illustrated in figure [ fig : blackhole2 ] . here
@xmath32 is the separation vector which describes the peeling of the outgoing geodesics from the horizon ( which is hidden in a conformal spacetime diagram ) .
this vector satisfies the geodesic deviation equation , @xmath33 where @xmath34 is the tangent to the horizon .
we first consider the solution of this equation in the classical schwarzschild spacetime .
this is conveniently done in null kruskal coordinates , in which the schwarzschild metric is @xmath35 the advantage of these coordinates is that @xmath36 , which is constant on ingoing null rays , is an affine parameter for the horizon .
let @xmath37 be an initial point at which @xmath38 , where @xmath39 is a constant .
then the solution of ( [ eq : geod ] ) to lowest non - trivial order in @xmath40 , when @xmath41 , is @xmath42 that is , the vector @xmath32 which initially has only a @xmath43-component , develops a @xmath36-component with no change in the @xmath43-component to this order .
the vector is null at the starting point , but afterwards is spacelike with squared norm @xmath44 the growth of @xmath45 describes the peeling of the outgoing geodesic from the horizon .
we now wish to consider the effects of quantum geometry fluctuations .
this may be done by treating ( [ eq : geod ] ) as a langevin equation with a fluctuating part of the riemann tensor , @xmath46 .
the separation vector now becomes a fluctuating vector written as @xmath47 where the contribution from the fluctuations , to lowest order in @xmath40 , is @xmath48 our measure of the effects of the quantum fluctuations will be the variance of @xmath49 .
to lowest order in @xmath40 , this variance may be expressed as @xmath50.\end{aligned}\ ] ] here we have used the facts that @xmath51 , and that each @xmath52 , as will be demonstrated in each of the explicit models of fluctuations to be considered below .
note that ( [ eq : variance ] ) involves a double integral of the riemann tensor correlation function , as each factor of @xmath53 is an integral over the fluctuating riemann tensor , ( [ eq : del_p ] ) .
first we discuss a simplified model which captures the essential physics of active fluctuations without the complications entailed in treating tensor perturbations in a black hole spacetime . here
the perturbed metric is taken to have the form @xmath54 where @xmath55 is the unperturbed schwarzschild background metric and @xmath56 is a free quantum scalar field , which we take to satisfy the minimally coupled wave equation @xmath57 in the schwarzschild spacetime .
this is sometimes called a dilaton field , but for our purposes it models the effects of gravitons .
we can expand the operator @xmath56 in terms of a complete set of positive norm wave packets solutions @xmath58 as @xmath59 where the @xmath60 are annihilation operators .
following hawking @xcite , we define the ingoing wave packets by @xmath61 where @xmath62 here @xmath63 is the usual advanced time coordinate .
the function @xmath64 approaches unity for @xmath65 and approaches a constant of magnitude less than unity , the transmission coefficient , near the horizon .
the integer @xmath66 controls where in frequency space the wave packet is peaked , while @xmath67 controls the width of the wave packet and has units of frequency .
the integer @xmath68 describes which wave packet is under consideration .
this construction allows for wave packets to be sent in at regular intervals of @xmath69 with various frequencies .
thus @xmath70 is the @xmath71 wave packet sent in with component frequencies ranging from @xmath72 to @xmath73 .
a corresponding set of outgoing packets can be defined by substitution of @xmath19 for @xmath74 in ( [ eq : monochromatic ] ) .
we now wish to take the quantum state of the scalar field to be a multi - mode squeezed vacuum state for some finite set of wave packet modes , which are sent into the black hole well after the collapse .
let the squeeze parameter of a given mode be @xmath75 , where @xmath66 labels the various excited modes .
we are interested in the effects of the excitation , so we take the difference between the given state and the vacuum state , and also assume that @xmath76 , corresponding to highly excited states .
the function @xmath64 becomes the transmission coefficient near the horizon , where we need to evaluate it .
this transmission coefficient is of order unity for wavelengths shorter than the size of the black hole , @xmath77 , and is approximately @xmath78 when @xmath79 , where @xmath80 is a constant of order unity . in the case that @xmath81 , we find the fractional fluctuations to be approximately @xmath82 ^ 2\,,\end{aligned}\ ] ] where @xmath83 is the planck length .
the corresponding result for the low frequency limit , @xmath84 , is @xmath85 ^ 2.\end{aligned}\ ] ] these results may be further simplified to yield the estimates @xmath86 and @xmath87 here @xmath88 is the bandwidth , a characteristic value of the @xmath67 , and we assume that @xmath89 , where @xmath90 is the typical peak frequency of the packets . in principle
, we can make @xmath91 and hence the fractional fluctuations arbitrarily large .
however , we have assumed that the average background spacetime is close to schwarzschild .
this assumption will fail to be valid once the perturbation of the riemann tensor becomes of the same order as the background riemann tensor .
this occurs when @xmath92 if we combine this result with ( [ eq : sg1 ] ) , we see that the back reaction becomes large when the fractional fluctuations in @xmath49 are of order unity .
this is expected , as these fractional fluctuations are determined by the fractional change in curvature . in this section
, we briefly outline a more realistic model involving gravitons propagating on the schwazlrschild background .
the gravitons are quantized perturbations of the schwarzschild geometry .
classical black hole perturbations have been discussed by many authors , beginning with regge and wheeler @xcite .
we adopt the regge - wheeler gauge explicit calculations , but as our key object is the gauge invariant riemann tensor correlation function , the results do not depend upon this choice . as in the previous section ,
we consider graviton wave packets in squeezed vacuum states which are sent into the black hole after the collapse .
the details of the calculation are considerably more complicated than in the case of the scalar graviton model , but the results @xcite are essentially of the same order . specifically , both ( [ eq : sg1 ] ) , ( [ eq : sg1 ] ) , and ( [ eq : sglimit ] ) still apply to the graviton model .
again , one can have the fractional fluctuations in @xmath49 of order unity . in this model ,
the geometry fluctuations arise from the quantum stress tensor fluctuations of a massless , minimally coupled scalar field . in general ,
the stress tensor correlation function , and hence the riemann tensor correlation function , are singular in the coincidence limit even if the expectation value of the stress tensor has been renormalized .
specifically , one can decompose the stress tensor correlation function into a sum of a fully normal - ordered part , a cross term and a vacuum term .
the latter two terms are singular in the coincidence limit , but can be defined as distributions by an integration by parts procedure @xcite , or by dimensional regularization @xcite . for our purposes , it is sufficient to include only the fully normal - ordered term , as this term gives the dominant contribution in the limit of highly excited states . in this limit ,
the approximation of keeping only the fully normal - ordered term is equivalent to that of including only those modes which are excited .
perturbation of the riemann tensor comes entirely from ricci tensor parts in this model and is given by @xmath93\beta}-g_{\beta[\mu}t_{\nu]\alpha}- \frac{2}{3 } g_{\alpha[\mu}g_{\nu]\beta}t\right ] \,,\ ] ] where the scalar field stress tensor is @xmath94 the brackets denote anti - symmetrisation and the parentheses symmetrisation .
the scalar field operator @xmath56 is given by ( [ eq : phi ] ) and ( [ eq : wavepacket ] ) .
the results of lengthy calculations for the fractional fluctuations are @xmath95 ^ 2 \nonumber \\ & \times & \frac{e^{2\rho_k } \ell_p^2 m}{\pi\varepsilon_k } \left [ \int_{k\varepsilon_k}^{(k+1)\varepsilon_k}d\omega'\ , ( \omega')^{1/2}\cos\omega'(v_0+\delta_k)\right]^2 \,,\end{aligned}\ ] ] for the case @xmath96 , and @xmath97 ^ 2 \nonumber \\ & \times & \frac{e^{2\rho_k } \ell_p^2}{\pi \varepsilon_k m } \left [ \int_{k\varepsilon_k}^{(k+1)\varepsilon_k}d\omega'\ , \sin\omega ' v_0 \right]^2\,,\end{aligned}\ ] ] for the case @xmath98 .
note that in contrast to the scalar graviton and graviton models , we now have a double sum over the excited wave packet modes .
this arises from the fact that the riemann tensor correlation function is quartic in the field operators .
we can estimate the fractional fluctuations for the case @xmath99 as @xmath100 comparison with ( [ eq : sg1 ] ) , which hold for both the scalar graviton and graviton models , reveals that the active fluctuations are proportional to @xmath101 , whereas the passive fluctuations are proportional to @xmath102 . in this sense ,
active fluctuations tend to dominate over passive ones .
however , the passive fluctuations grow more rapidly for large @xmath91 .
the point at which the mean background spacetime ceases to be schwarzschild is again when @xmath103
we have discussed the problem of spacetime geometry fluctuations using the riemann tensor correlation function , first near flat spacetime , and then in black hole spacetime . in the former case , we noted some of the operational signatures of spacetime geometry fluctuation : luminosity fluctuations , line broadening , and angular blurring .
however , the main interest is in the problem of black hole horizon fluctuations .
the hawking derivation of black hole radiance invokes transplanckian modes , which must remain extremely close to the event horizon for a very long time , as measured by an external observer . here
extremely close means far less than one planck length as measured by an infalling observer .
this suggests that quantum fluctuations of the horizon might drastically alter black hole radiance .
if this were the case , then the connection between black hole physics and thermodynamics might only be preserved by going to a `` mode - regeneration '' picture based upon a non - linear , non - lorentz invariant dispersion relation . however , a previous study @xcite of active fluctuations from gravitons in a vacuum state indicated that the vacuum level horizon fluctuations will not upset the hawking derivation for black holes more massive than the planck mass . in the present paper
, we have summarised an investigation @xcite which goes further and considers both active and passive fluctuations from squeezed states .
this has the advantage that the level of the fluctuations can be increased by increasing the squeeze parameter .
indeed , we did find that one can the fractional fluctuation in the geodesic deviation vector approach order unity .
however , in all of the models discussed in this paper , the contribution to the quantity @xmath104 come only from the @xmath36-component of @xmath32 , not the @xmath43-component .
if we refer to figure [ fig : blackhole2 ] , we see that it would require @xmath43-component fluctuations to cause serious problems with modes being prematurely ejected or captured by the black hole .
fluctuations in the @xmath36-component do influence when a given wave packet arrives at @xmath105 .
however , this is likely to be unobservable when the quantum state before collapse is the vacuum state .
these fluctuations could alter stimulated emission , and hence are in principle observable , but would require a state containing particles with transplanckian energies if the stimulated emission is to be observed well after the collapse . in summary ,
although it is possible to enhance spacetime geometry fluctuations by use of squeezed states , these enhanced fluctuations do not seem to alter the thermal hawking radiation .
thus we conclude that hawking s derivation of black hole radiance is remarkably robust , and that there may be no transplanckian `` problem '' .
this work was supported in part by the national science foundation under grant phy-0555754 .
|
we will discuss selected physical effects of spacetime geometry fluctuations , especially the operational signatures of geometry fluctuations and their effects on black hole horizons .
the operational signatures which we discuss involve the effects of the fluctuations on images , and include luminosity variations , spectral line broadening and angular blurring .
our main interest will be in black hole horizon fluctuations , especially horizon fluctuations which have been enhanced above the vacuum level by gravitons or matter in squeezed states .
we investigate whether these fluctuations can alter the thermal character of a black hole .
we find that this thermal character is remarkably robust , and that hawking s original derivation using transplanckian modes does not seem to be sensitive even to enhanced horizon fluctuations .
|
for decades , astronomers have struggled to determine the properties of the high - velocity clouds ( hvcs ) , a population of objects identified by 21 cm hi emission with velocities inconsistent with galactic rotation .
recent improvements in available radio survey data ( e.g. the hi parkes all - sky survey and the leiden dwingeloo survey ; barnes et al . 2001 and hartmann & burton 1997 respectively ) and uv instrumentation ( e.g. stis and fuse ) , have substantially increased our understanding of these numerous ( @xmath6 2000 in the south alone , putman et al .
2002 ) clouds .
however , the origin of the vast majority of the hvc population has remained controversial .
several explanations for their presence have been suggested , including : a galactic fountain ; tidal or ram pressure stripping from dwarf galaxies ; and low mass dark matter halos predicted by currently favored cosmologies but not observed in optical surveys ( blitz et al .
1999 , braun & burton 2000 , among others ) .
a combination of these , and other , scenarios is likely necessary to explain the wide range of observed hvc properties ( see wakker & van woerden 1997 for overview of suggested origins and discussion ) .
one of the principle limitations in distinguishing between these scenarios is the unknown distances to the clouds ( e.g. gibson et al .
the most reliable distance estimates to date have resulted from absorption line studies toward background stars at known distances placing concrete upper and lower limits on the distances to a handful of nearby hvcs ( wakker 2001 ) .
unfortunately , this method can not be applied to clouds at extragalactic ( i.e. @xmath7 100 kpc ) distances due to a lack of background stars . alternatively , if the hvcs host a detectable stellar population , their distances could be measured directly using the location of the tip of the red giant branch or rr lyrae stars . identifying stellar populations associated with hvcs would therefore provide the measurements of both stellar content and distance necessary for constraining their possible origins .
in particular , stellar associations would provide strong support for the theory that some hvcs may be extremely low mass galaxies and , as such , may likely host stars ( see 4.2 for discussion ) .
several groups are currently conducting deep , pointed surveys toward compact hvcs ( chvcs ) in an attempt to identify resolved stars associated with the clouds ( grebel , braun & burton 2000 , gibson et al .
although these observations are excellent probes for stars close to the position of the cloud , there are notable drawbacks to this method .
observations of local group dwarf spheroidals ( dsphs ) suggest that hi associations with dsphs are not necessarily coincident with the galaxies positions ( carignan et al 1998 , st - germain et al .
1999 , blitz & robishaw 2000 ) .
furthermore , although only chvcs have been targeted in existing searches , putman & moore ( 2001 ) have shown that the chvc distribution is not consistent with the predicted distribution of dark matter halos around the milky way and that other subsets of the hvc family possess properties more similar to those predicted for the halos .
the only existing large - area search for stellar counterparts used digitized poss plates to examine one square degree areas around 264 clouds ( simon & blitz 2002 ) .
although their survey is sensitive to stellar associations throughout the local group , all 264 clouds were classified as non - detections . however , the survey used a nonuniform data set with a poorly characterized surface brightness limit that is no fainter than the known local group dwarf galaxies .
substantial improvements in survey sensitivity and uniformity can be made using the sloan digital sky survey ( sdss ) .
sdss is a unique tool with which to search for stellar counterparts to hvcs , providing the means both to survey sufficiently large areas to detect displaced stars and to efficiently sample a large number of clouds .
we are currently using sdss data to conduct a survey for resolved low surface brightness satellites out to the milky way s virial radius ( willman et al . 2002 , hereafter paper i ) .
our survey also provides a way to statistically investigate the possibility of stellar associations with the high - velocity clouds on very large scales .
our detection limits ( calculated in paper i ) provide a template with which to interpret the results , whether positive or null , in a quantitative framework . in this paper , we present an analysis of 325 square degrees of the sdss early data release ( edr ; stoughton et al .
2002 ) imaging database .
this zero - declination strip contains 13 zero - declination hvcs , as cataloged by putman et al .
we briefly summarize our survey technique in [ sec : tech ] , discuss the analysis of the sdss early data release in [ sec : application ] , use the analysis results and survey detection limits to constrain the stellar nature of the hvcs in [ sec : constraints ] , and discuss our findings in 4 .
the sloan digital sky survey ( york et al . 2000 ) will eventually image nearly one - quarter of the sky . to date
, roughly 460 deg@xmath9 of imaging data have been released to the general community in the early data release ( stoughton et al .
these imaging data are taken in drift - scan mode with a camera consisting of 30 2048x2048 ccds ( see gunn et al .
1998 ) on a dedicated 2.5 m telescope at apache point observatory .
the sky is observed nearly simultaneously through 5 broad - band filters ( @xmath10 and @xmath11 , fukugita et al .
1996 ) and the data are @xmath12 complete for stellar sources to limiting magnitudes of roughly 22.0 , 22.2 , 22.2 , 21.3 and 20.5 respectively ( stoughton et al . 2002 ) . as the sdss photometric system is not yet finalized , the edr magnitudes used in this paper will be denoted as @xmath13,@xmath14,@xmath15,@xmath16 , and @xmath17 .
the current system will differ absolutely from the final system by only a few percent in every band except for @xmath18 , where it will differ by no more than @xmath19 ( stoughton et al .
2002 ) . a combination of accurate astrometry ( see pier et al .
2002 for details on the astrometric calibration ) , color and morphological information produces robust star - galaxy separation to a limiting magnitude of r@xmath20 = 21.5 ( ivezi et al .
see hogg et al .
( 2001 ) for details on the photometric calibration of the data and smith et al .
( 2002 ) for a description of the photometric system . because we are looking for extragalactic objects , all magnitudes presented in this paper have been corrected for reddening due to galactic extinction with the maps of schlegel , finkbeiner & davis ( 1998 ) .
our survey takes advantage of sdss s major strengths ( large sky coverage , uniformity , depth , and accurate photometry in five bands ) to uniformly probe for very low surface brightness ( lsb ) dwarf galaxies , based only on resolved stars .
paper i describes both our survey technique and detection limit calculations in detail , but we briefly summarize our method here .
sdss s faint photometric limits and robust star - galaxy separation are such that about a magnitude of a stellar population s red giant branch can be resolved out to a distance of 350 kpc , roughly equivalent to the milky way s virial radius .
this allows us to identify candidate dwarf galaxies as spatial overdensities of resolved stars .
we apply a red color cut to stellar sources brighter than r@xmath20 = 21.5 .
the color cut , in g@xmath20 - r@xmath20 and r@xmath20 - i@xmath20 , is designed to select stars whose colors are consistent with red giant branch colors .
although our cut does not uniquely distinguish red giant branch stars from foreground red main sequence stars , it does reduce the stellar foreground by @xmath21 , increasing the contrast of extragalactic sources .
we then consider bins in r@xmath20 magnitude , further reducing the foreground .
stars which pass the magnitude and color criteria are binned into a spatial density array , which is then smoothed with a circularly symmetric exponential filter with a scale length of either h=3@xmath5 or 7@xmath5 .
candidates are automatically identified as 5.5 and 5@xmath22 statistical overdensities ( for the 3@xmath5 and 7@xmath5 filter , respectively ) , in the smoothed sky map .
each field of sky is analyzed multiple times , once for each @xmath15 magnitude bin . in paper
i , we calculated our detection limits for purely old stellar populations with a range of angular scale lengths ( @xmath23 ) and distances ( 23 - 350 kpc ) .
figure 5 in paper i displays our survey s detailed detection limits for the stellar foreground at @xmath24 .
the surface brightness limits are a function of galaxy size and distance , and range between @xmath25 and @xmath26 mag arcsec@xmath27 .
( for comparison , sextans is the lowest surface brightness milky way satellite , with @xmath28 ( mateo 1998 ) . )
our detection efficiency rapidly drops off for small angular size ( @xmath29 ) systems , and is insensitive to systems more distant than @xmath6 350 kpc .
the limits are not sensitive to seeing , because our technique uses only resolved stars brighter than the limit for robust star - galaxy separation .
we have analyzed @xmath6 325 square degrees of data in the sdss edr around zero declination between @xmath30 and @xmath31 .
this region of sky contains 10 hvcs , 2 : hvcs ( a cloud that could not unambiguously be identified as a chvc ) , and 1 chvc as cataloged by putman et al .
we have chosen the putman et al .
catalog as a reference for the present analysis because it is the most uniform and high resolution hvc catalog currently available .
braun and burton ( 1999 ) have also cataloged 1 chvc in the zero - declination region covered by the edr .
the braun and burton chvc is not coincident with the putman chvc , but rather is coincident with a putman `` : hvc '' included in this analysis .
our analysis produced 20 detections that we are currently investigating .
it is difficult to say conclusively that these stellar detections are not associated with any particular hvc , because the gas and stars in low mass local group galaxies are not necessarily spatially coincident .
however , we can show that the spatial displacements between the stellar detections and the hvcs are no different than we would expect for completely uncorrelated populations . to test this , in figure 1
we show a comparison of the distances between the detections and their nearest hvc centers with the distances between randomly generated positions and their nearest hvc centers .
a kolmogorov - smirnov test of these data shows that , statistically , none of the candidate detections are apt to be associated with any of the hvcs in the region .
this suggests that any stellar population associated with the hvcs must have a surface brightness fainter than the limits of our survey .
we now couple our null detections with the survey s detection limits to place quantitative limits on the stellar content of the hvcs , assuming they are located within our virial radius .
as the detection limits vary with galaxy angular size and distance , we have selected two fiducial distances ( 50 and 300 kpc ) and sizes ( exponential scale lengths of 3 and 7@xmath5 ) for which to calculate the minimum hi mass - to - light ratios of the clouds .
we did not simply use the semi - major axes of the hvcs 21 cm emission for the fiducial sizes because the radial extents of neutral hydrogen and stars within dwarf galaxies are often quite different ( young & lo 1997 , salpeter & hoffman 1996 ) .
table 1 displays both the hi and the constrained stellar properties of the 13 zero - declination hvcs covered in our analysis . columns 1 - 5 of table 1 contain the name , @xmath32 ( in degrees ) , @xmath33 ( in degrees ) , peak @xmath34 ( in units of @xmath35 ) and semi - major axis ( in degrees ) for each cloud as published in putman et al .
column 6 contains the assumed distance used to calculate : the semi - major axis in kpc ( column 7 ) ; @xmath36 ( in units of @xmath37 ; column 8) ; and the optical detection limits for 3@xmath5 and 7@xmath5 sized systems ( quantified as v - band central surface brightnesses of the old stellar population ; columns 9 an 11 ) .
@xmath36 was calculated using @xmath38 ( where s is the integrated line flux in jy km sec@xmath27 from putman et al .
( 2002 ) ) .
we have recalculated the detection limits from paper i for a range of stellar foreground levels corresponding to the range of galactic positions for the zero - declination hvcs . to get a sense of how the amount of hi contained in these hvcs compares to that contained in known gas - rich dwarf galaxies , figure 2 displays the hi masses of the zero - dec hvcs ( for assumed distances of 50 and 300 kpc ) , the estimated hi masses of 9 gas - rich local group dsphs ( tucana oosterloo et al .
1996 ; phoenix
st - germain et al 1999 ; ddo 210 , leo i , lgs 3 , pegasus , sextans , sculptor blitz & robishaw 2000 ; antlia barnes & de blok 2001 ) , and the hi masses of the pss - ii lsb dwarf irregular ( dirr ) sample of schombert , mcgaugh and eder ( 2001 ) .
we would like to note that the published hi detections of a few of the local group dsphs are not definitive associations , so the plotted hi masses are upper limits .
the 9 dsph hi detections in figure 2 are either those that are conclusive associations or those that do not have conclusive evidence that shows otherwise . from figure 2 , we see that the dirr galaxies have much higher hi masses than the hvcs , and thus are not close analogs for a local group population of hvcs .
in contrast , the hvcs could contain hi masses that correspond to the approximate hi masses of the gas - rich local group dsphs .
we calculate lower limits on the neutral hydrogen mass - to - light ratio ( @xmath36/@xmath2 ) of the hvcs ( columns 10 and 12 ) using the limiting surface brightnesses from columns 9 and 11 .
specifically , we calculate a lower limit on the @xmath39 for all of these clouds with @xmath40 , where we derive @xmath41 by integrating over an exponential galaxy profile normalized to the limiting central surface brightness .
m@xmath1 and luminosity both scale as @xmath42 , so the limiting @xmath36/@xmath2 ratio is independent of cloud distance , aside from the dependence of @xmath43 on distance .
figure 3 displays a comparison between our lower limits to m@xmath1/@xmath2 for zero - dec hvcs , the estimated m@xmath1/@xmath2 of local group dsphs and the m@xmath1/@xmath2 of the pss - ii dirrs .
the available pss - ii data is in the i - band , but we convert to the corresponding v - band measurement for an old stellar population by using @xmath44 for elliptical galaxies from pildis et al .
( 1997 ) .
the m@xmath1/@xmath2 lower limits for the hvcs range from about 0.2 to 250 .
the calculated minimum @xmath39 for many of the hvcs are similar to those of known gas - rich dwarf galaxies , despite our faint detection limits ( i.e. they may not be particularly faint in the optical compared to their very low hi masses ) .
even so , figure 3 shows that @xmath45 of the hvcs with assumed distances of 300 kpc have @xmath39 greater than that of all known local group dsphs , and that 3/5 of that 40@xmath46 also have @xmath39 greater than the entire pss - ii dwarf sample .
if shown to be within dark matter halos , these hvcs would be among the most gas rich self - gravitating galaxies observed to date .
the import of a non - detection clearly depends on whether or not a cloud is closer than 350 kpc .
if an hvc is more distant than 350 kpc , then we can place absolutely no constraints on its stellar content .
it is possible that a fraction of the hvcs in our sample are indeed at large distances .
although upper distance limits from absorption line studies and observed head - tail morphologies would imply inner halo locations , we do not know of any such published analyses applied to any cloud in this sample ( see wakker 2001 for compendium of absorption line estimates ; br " uns et al . 2000 ; quilis & moore 2001 ) . however , if one believes the many indirect estimates of hvc distances that place a large fraction of the clouds closer than several hundred kiloparsecs ( bland - hawthorn & maloney 1999 , weiner et al .
2001 , blitz et al . 1999 ,
quilis & moore 2001 ) , it is logical to expect at least some of these clouds to be within our virial radius . as further evidence that at least some of the 13 clouds are nearby ,
surveys for hi clouds outside of the local group have failed to identify starless clouds with m @xmath47 ( zwaan 2001 , zwaan & briggs 2000 ) .
this observation suggests that the hvcs must either be sufficiently close that their hi masses fall below the zwaan limit , or should host a stellar population detectable by existing local group surveys .
if located beyond our virial radius , four of the 13 hvcs have masses that rival the zwaan limit but do not appear to have a detectable stellar population .
therefore , it is quite likely at least all four of those clouds fall within our survey volume . clouds in our analysis closer than 350 kpc either have : _ i ) _ no associated stars ; _ ii ) _ a stellar population fainter than our detection limits and a mass - to - light ratio , m@xmath1/l@xmath48 , higher than that given in table 1 ; or _
iii ) _ a stellar population with a very small angular scale length ( @xmath49 ) but brighter than our given detection limits . for clouds that do contain stars , as in cases _
ii ) _ and _ iii ) _ , most other survey methods would also not be able to efficiently uncover such faint or small populations .
pointed searches can probe deeper and to greater distances than large - scale surveys of resolved stars , but they are much less efficient and may have difficulty associating very small scale length systems with an hvc .
large - scale surveys based of resolved stars could reach fainter limits only with prohibitively deeper exposures , with a priori assumptions about the undetected stellar populations ( odenkirchen et al .
2001 ) , or with surveys for very specific stellar types ( e.g. horizontal branch and rr lyrae ; yanny et al .
2000 , ivezi ' c et al .
2000 ) that are only effective in the inner halo . therefore , although our sample size is small , our detection limits are faint enough to conclude that stellar populations are unlikely to be an efficient or effective distance determination technique to hvcs within our virial radius .
a large - scale survey for overdensities of diffuse light using more recent and uniform data ( such as sdss ) than available for past surveys ( simon & blitz 2002 ; armandroff , jacoby & davies 1998 ; dalcanton et al . 1997 ) holds promise for efficiently improving existing limits on the possible stellar content of hvcs located beyond our virial radius .
there is a substantial body of circumstantial evidence to support the theory that a subset of high - velocity clouds inhabit some of the numerous low mass dark matter halos predicted by the current paradigm of structure formation , yet notably missing from our observations of the local group of galaxies ( blitz et al .
1999 , braun & burton 2000 , among others ) .
for example , blitz and robishaw ( 2002 ) show that the masses and hi morphologies of many hvcs are consistent with the hi associated with local group dsphs .
furthermore , low mass analogues to known extremely gas - rich lsb galaxies ( mcmahon et al .
1990 ; oneil , bothun & schombert 2000 ; schombert , mcgaugh & eder 2001 ) may easily have gone undetected in past surveys for local group galaxies
. we can use existing evidence to evaluate the likelihood of a gas - rich dark matter halo occurring within our survey s volume , and discuss the prospects for such halos to host stars .
this will allow us to show that our non - detection of stars places few limits on missing galaxy explanations for the hvcs .
@xmath51cdm simulations predict @xmath6 300 sub - halos within the virial radius of the milky way with masses greater than or equal to the least massive known milky way companions ( @xmath52 km sec@xmath27 , where @xmath53 ; moore et al .
the expected number of halos increases steeply with decreasing mass , so @xmath6 200 more sub - halos are expected with 5 km sec@xmath54 km sec@xmath27 ( moore et al .
1999 ) , and perhaps yet more at smaller masses . assuming a uniform distribution within the 325 square degree survey area , this translates to roughly 2 - 3 halos similar to those of known dwarf satellites , and a comparable number of lower mass halos .
thus , it is reasonable to expect a small number of low mass dark matter sub - halos within our survey volume in a @xmath51cdm cosmology . these low mass
, dark matter halos may or may not contain a detectable amount of neutral hydrogen .
numerous physical arguments can be invoked to suggest that v@xmath55 km sec@xmath27 halos ( particularly those that contain stars ) are unlikely to host neutral gas ( quinn , katz & efstathiou 1996 ; bullock , kravtsov , & weinberg 2000 ; ferrara & tolstoy 2000 ) .
however , empirical evidence suggests that low mass halos can host both stars and gas .
3 of the milky way s 9 dsphs ( leo i , sculptor , and sextans ) potentially have hi associations ( blitz and robishaw 2000 ) .
all 3 of these dsphs are less massive than @xmath56 km sec@xmath27 and closer than @xmath6 250 kpc .
the sculptor detection is somewhat ambiguous because it is part of a larger cloud complex that may be mixed in with the magellanic stream ( putman 2000 ) , and the sextans detection is only a few sigma and has yet to be confirmed by other radio observations .
these existing observations imply that @xmath57 of milky way sub - halos are gas - rich .
of course , the true gas - rich fraction is complicated by several observational effects , including : halos that are gas - rich due to inefficient star formation may not have been detected in local group surveys , and the lowest mass halos will more easily be stripped of their gas by feedback and tidal forces . despite these uncertainties , it is reasonable to expect no more than 2 - 3 of the 4 - 6 sub - halos expected in our survey volume to be detectable as hvcs , if any ( assuming poisson statistics ) .
thus , although it is possible that a dark matter bound cloud exists in our survey , the volume sampled is not yet large enough to make this a statistical certainty . as discussed in
4 , any hi rich dark matter halo within our survey volume is either starless , contains a stellar population many times fainter than all known milky way satellites , or contains a very small scale length stellar population . according to benson et al .
( 2002 ) , an m@xmath58 ( @xmath59 fainter than the faintest known milky way dsph ) galaxy is expected to have a half - mass angular radius of 1.5 at a distance of 300 kpc . therefore , the prospect of a small angular scale length system much brighter than our detection limits seems unlikely . with our small sample ,
it is neither productive nor possible to further consider which of those three options would be the most likely for any of these 13 clouds , were they shown to inhabit a dark matter halo . studies which directly probe for dark matter associated with the clouds , such as that suggested by lewis et al .
( 2000 ) , may ultimately be the best test of theories that place hvcs within dark matter halos , although our current sample is too small to say one way or the other .
it is quite possible that one / some of the predicted 2 - 6 gas - poor dark matter halos in the surveyed region do contain stars .
we are currently investigating whether any of our 20 detections actually represent a new low luminosity dwarf galaxy .
we have used our non - detections of stars associated with 13 hvcs to place new constraints on their stellar content .
we have shown that many of the 13 clouds have hi masses and ( minimum ) hi mass - to - light ratios similar to those of known dsphs , if the clouds are closer than 350 kpc . also , 3 of the hvcs have minimum m@xmath1/l that would place them among the most gas - rich galaxies , if shown to reside in a dark matter halo . given our results , and those of simon and blitz ( 2002 ) , stellar populations appear unlikely to be a promising distance determination technique for a large number of hvcs . however , the continued pursuit of stellar associations with hvcs is a worthwhile endeavor , as strong circumstantial evidence suggests that some fraction , however small , of the hvcs are likely to host a detectable low surface brightness ( lsb ) galaxy .
as there are only 1 - 3 known gas - rich milky way dsph satellites , the discovery of _ any _ nearby gas - rich stellar association would substantially increase the sample and provide insights into extremely low - mass galaxy evolution .
we are currently pursuing the prospect that some of our 20 detections may actually be a new lsb hi poor milky way satellite .
as our present survey volume is too small to unquestionably contain any of the `` missing '' low mass , dark matter halos predicted by @xmath51cdm cosmology , we can not place limits on missing galaxy explanations for the hvcs .
however , the analysis performed in this work will be applied to survey quality sdss data as it becomes available , and then correlated with the current and future hvc catalogs of putman et al .
( hipass ) and de heij et al . (
2002 ) ( leiden / dwingeloo hi survey ) . the completed sloan survey will cover @xmath60 more area than we have presented in this paper , guaranteeing that `` missing '' dark matter halos would be well represented in our survey volume ( provided @xmath51cdm is correct ) . our future analysis of the sdss will enable us to apply the arguments presented here to hundreds of clouds , a sample large enough for us to do a concrete statistical assessment of their nature .
we would like to thank james schombert for helping us access his dirr catalogs , mary putman for pointing us toward her hvc catalog , and christian br " uns for sending us a list of the hvcs he observed to have head - tail morphologies .
funding for the distribution and creation of the sdss archive has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , princeton university , the united states naval observatory , and the university of washington .
armandroff , t. e. , davies , j. e. , & jacoby , g. h. 1998 , , 116 , 2287 barnes , d. g. et al .
2001 , , 322 , 486 .
barnes , d. g. & de blok , w. j. g. 2001 , , 122 , 825 benson , a. j. , frenk , c. s. , lacey , c. g. , baugh , c. m. , & cole , s. 2002 , , 333 , 177 bland - hawthorn , j. & maloney , p. r. 1999 , , 510 , l33 blitz , l. @xmath61 robishaw , t , 2000 , apj , 541 , 675 .
blitz , l. , spergel , d. n. , teuben , p. j. , hartmann , d. @xmath61 burton , w. b. 1999 , , 514 , 818 .
braun , r. @xmath61 burton , w.b .
1999 , , 341 , 437 braun , r. & burton , w. b. 2000 , , 354 , 853 .
br " uns , c. , kerp , j. , kalberla , p. m. w. , & mebold , u. 2000 , , 357 , 120 bullock , j. s. , kravtsov , a. v. , & weinberg , d. h. 2000 , , 539 , 517 carignan , c. , beaulieu , s. , c^ ot ' e , s. ; . , demers , s. , & mateo , m. 1998 , , 116 , 1690 dalcanton , j. j. , spergel , d. n. , gunn , j. e. , schmidt , m. , & schneider , d. p. 1997
, , 114 , 635 de heij , v. , braun r. , & burton w. b. 2002 a&a submitted ( astro - ph/0201249 ) ferrara , a. & tolstoy , e. 2000 , , 313 , 291 fukugita , m. , ichikawa , t. , gunn , j. e. , doi , m. , shimasaku , k. , & schneider , d. p. 1996 , aj , 111 , 1748 gibson , b. k. , fenner , y. , maddison , s. t. , & kawata , d. 2001 , to appear in _ extragalactic gas at low redshift _ asp conference series ( astro - ph 0107496 ) grebel , e. k. , braun , r. , & burton , w. b. 2000 , american astronomical society meeting , 196 , 2809 .
gunn , j. e. , et al .
1998 , aj , 116 , 3040 hartmann , d. & burton , w. b. 1997 , cambridge ; new york : cambridge university press , isbn 0521471117 .
hogg , d. w. , finkbeiner , d. p. , schlegel , d. j. , & gunn , j. e. 2001 , , 122 , 2129 ivezi ' c , v zeljko et al . 2000 , , 120 , 963 .
lewis , g. f. , bland - hawthorn , j. , gibson , b. k. , & putman , m. e. 2000 , , 112 , 1300 mateo , m. l. 1998 , , 36 , 435 mcmahon , r. g. , irwin , m. j. , giovanelli , r. , haynes , m. p. , wolfe , a. m. , & hazard , c. 1990 , , 359 , 302 moore , b. , ghigna , s. , governato , f. , lake , g. , quinn , t. , stadel , j. , @xmath61 tozzi , p. 1999
, , 524 , 19l odenkirchen , m. et al .
2001 , , 548 , l165 oneil , k. , bothun , g. d. , & schombert , j. 2000 , , 119 , 136 oosterloo , t. , da costa , g. s. , & staveley - smith , l. 1996 , , 112 , 1969 pier , j.r .
2002 , aj , submitted pildis , r. a. , schombert , j. m. , & eder , j. 1997 , , 481 , 157 putman , m. e. 2000 , publications of the astronomical society of australia , 17 , 1 putman , m. , et al .
2002 , aj , 123 , 873 putman , m. & moore , b. 2001 , to appear in _ extragalactic gas at low redshift _ asp conference series ( astro - ph 0110417 ) quilis , v. & moore , b. 2001 , , 555 , l95 quinn , t. , katz , n. , & efstathiou , g. 1996 , , 278 , l49 salpeter , e. e. & hoffman , g. l. 1996 , , 465 , 595 schlegel , d.j . ,
finkbeiner , d.p . , @xmath61 davis , m. 1998 , , 500 , 525 schombert , j. m. , mcgaugh , s. s. , & eder , j. 2001 , , 121 , 2420 simon , j. and blitz , l. apj submitted , 2002 , astro - ph/0204042 smith , j. a. et al .
2002 , , 123 , 2121 st - germain , j. , carignan , c. , c^ ote , s. ; . , & oosterloo , t. 1999 , , 118 , 1235 stoughton , c. et al .
2002 , aj , 123 , 485 wakker , b. p. 2001 , apjs , 136 , 463 wakker , b. p. & van woerden , h. 1997 , , 35 , 217 weiner , b. j. , vogel , s. n. , & williams , t. b. 2001 , asp conf .
ser . 240 : gas and galaxy evolution , 515 willman , b. , dalcanton , j. , ivezi , z. , jackson , t. , lupton , r. , brinkmann , j. , henessey , g. , & hindsley , r. , 2002 , apj , 123 , 848 yanny , b. et al .
2000 , , 540 , 825 york , d. g. et al , 2000 , aj , 120 , 1579 young , l. m. & lo , k. y. 1997 , , 490 , 710 zwaan , m. a. 2001 , , 325 , 1142 zwaan , m. a. & briggs , f. h. 2000 , , 530 , l61 lrrrrrrrrrrr 1 : hvc 113.0 - 63.2 - 310 & 8.40 & -0.67 & 0.03 & 0.30 & 50 & 0.26 & 0.02 & 28.0 & 0.6 & 28.9 & 0.3 & & & & & 300 & 1.57 & 0.57 & 29.5 & 2.3 & 30.4 & 1.1 2 hvc 124.7 - 61.9 - 28 & 13.70 & 0.92 & 0.02 & 0.40 & 50 & 0.35 & 0.01 & 28.0 & 0.4 & 29.0 & 0.2 & & & & & 300 & 2.09 & 0.38 & 29.5 & 1.6 & 30.5 & 0.7 3 hvc 158.1 - 58.6 - 298 & 30.33 & -0.57 & 0.04 & 0.20 & 50 & 0.17 & 0.03 & 28.1 & 1.2 & 29.0 & 0.5 & & & & & 300 & 1.05 & 1.08 & 29.6 & 5.0 & 30.6 & 2.3 4 hvc 160.6 - 57.5 - 296 & 32.03 & -0.38 & 0.06 & 0.30 & 50 & 0.26 & 0.12 & 28.1 & 5.1 & 29.0 & 2.2 & & & & & 300 & 1.57 & 4.46 & 29.6 & 20.8 & 30.6 & 9.6 5 hvc 167.0 - 55.6 - 172 & 36.00 & -0.85 & 0.02 & 0.20 & 50 & 0.17 & 0.01 & 28.1 & 0.4 & 29.0 & 0.2 & & & & & 300 & 1.05 & 0.38 & 29.6 & 1.8 & 30.6 & 0.8 6 hvc 170.7 - 52.2 - 193 & 39.88 & 0.38 & 0.36 & 0.20 & 50 & 0.17 & 0.36 & 28.1 & 15.1 & 29.1 & 6.7 & & & & & 300 & 1.05 & 13.08 & 29.6 & 63.0 & 30.6 & 28.9 7 : hvc 171.0 - 53.6 - 237 & 39.05 & -0.82 & 0.41 & 0.50 & 50 & 0.44 & 1.19 & 28.1 & 49.3 & 29.1 & 21.8 & & & & & 300 & 2.62 & 42.82 & 29.6 & 205.4 & 30.6 & 94.4 8 chvc 173.3 - 52.1 - 225 & 41.12 & -0.60 & 0.22 & 0.30 & 50 & 0.26 & 0.25 & 28.1 & 10.4 & 29.1 & 4.6 & & & & & 300 & 1.57 & 9.01 & 29.7 & 43.5 & 30.6 & 20.0 9 hvc 254.1 + 53.3 + 105 & 166.15 & 0.88 & 0.25 & 0.30 & 50 & 0.26 & 1.18 & 28.0 & 43.0 & 28.8 & 18.0 & & & & & 300 & 1.57 & 42.48 & 29.4 & 160.0 & 30.4 & 72.7 10 hvc 261 + 55 + 124 & 170.60 & -0.08 & 0.23 & 3.50 & 50 & 3.05 & 1.85 & 27.9 & 65.9 & 28.8 & 27.4 & & & & & 300 & 18.33 & 66.59 & 29.4 & 244.6 & 30.3 & 110.6 11 hvc 339.5 + 56.6 - 110 & 212.00 & -0.68 & 0.06 & 0.30 & 50 & 0.26 & 0.04 & 27.5 & 0.9 & 28.3 & 0.4 & & & & & 300 & 1.57 & 1.44 & 29.0 & 3.7 & 29.8 & 1.5 12 hvc 081.2 - 54.8 - 279 & 350.35 & 0.57 & 0.09 & 0.30 & 50 & 0.26 & 0.12 & 27.9 & 4.2 & 28.8 & 1.7 & & & & & 300 & 1.57 & 4.46 & 29.3 & 15.7 & 30.3 & 7.0 13 hvc 091.7 - 58.3 - 273 & 357.08 & 0.67 & 0.10 & 0.20 & 50 & 0.17 & 0.10 & 27.9 & 3.4 & 28.8 & 1.4 & & & & & 300 & 1.05 & 3.46 & 29.4 & 12.7 & 30.3 & 5.8
|
in this paper , we limit the stellar content of 13 high - velocity clouds ( hvcs ; including 1 compact hvc ) using the detection limits of a new survey for resolved milky way satellite galaxies in the sloan digital sky survey early data release ( edr ) .
our analysis is sensitive to stellar associations within the virial radius of the milky way that are up to 50@xmath0 fainter than the faintest known milky way satellites .
statistically , we find no stellar overdensity associated with any of the clouds .
these non - detections suggest lower limits of m@xmath1/@xmath2 that range between @xmath3 and @xmath4 , assuming cloud distances of 50 and 300 kpc and using fiducial optical scale lengths of 3@xmath5 and 7@xmath5 .
we explore the implication of these non - detections on the origin of the hvcs .
|
the upgraded mipp experiment is designed to measure the properties of strong interaction spectra form beams @xmath0 and @xmath1 for momenta ranging from 1 gev / c to 120 gev / c .
the layout of the apparatus in the data taken so far can be seen in figure [ layout ] .
the centerpiece of the experiment is the time projection chamber , which is followed by the time of flight counter , a multi - cell cerenkov detector and the rich detector .
the tpc can identify charged particles with momenta less than 1 gev / c using de / dx , the time of flight will identify particles below approximately 2 gev / c , the multi - cell cerenkov detector is operational from 2.5 gev / c to 14 gev / c and the rich detector can identify particles up to 120 gevc . following
this is an em and hadronic calorimeter capable of detecting forward going neutrons and photons .
the experiment has been busy analyzing its data taken on various nuclei and beam conditions .
the table [ runtable ] shows the data taken by mipp i to date .
we have almost complete acceptance in the forward hemisphere in the lab using the tpc .
the reconstruction capabilities of the tpc can be seen in figure [ tpc ] .
the particle identification capabilities of the tpc can be seen in figure [ tpcdedx ] .
the time of flight system provides further measurement of the particles with momenta less than 2 gev / c .
figure [ tof ] shows the time of flight data where a kaon peak is clearly visible .
mipp has published forward going neutron spectra off various nuclei and compared it with monte carlo predictions @xcite it can be seen clearly that the monte carlo predictions differ significantly from each other and the data . figure [ numidata ] shows ( in a paper in preparation , @xcite ) preliminary results of particle production from the full numi target as a function of momenta .
the mipp upgrade proposal consists primarily of an electronics upgrade of the whole detector so that the detector reads out at 3 kilohz as compared to 20 hz in run i. all the new electronics modules have been prototyped .
the tpc has the largest data rate and we display one of its elctronics modules in figure [ tpcchips ] .
the document @xcite contains the details of the upgrade proposal .
the mipp upgrade proposal will acquire high quality data of particle production with high statistics and much improved particle identification .
the available beam momenta and the calculated beam rates can be seen in table [ mippbeam ] and range approximately 1 gev / c to 120 gev / c .
as such the upgrade can straddle the resonance region and the scaling region of strong interaction physics using 6 beam species .
the mipp upgrade intends to acquire data for a variety of interactions .
these include * a systematic study of missing baryon resonances between 1 gev and 3 gev .
* an investigation of fundamental scaling laws in inclusive reactions . * a thorough investigation of forward neutron spectra . * a systematic acquisition of particle production data and comparison and tuning of existing monte carlo generators . * a measurement of particle production data from full scale neutrino targets and using the data to improve the target design .
* a systematic study of hypernuclei * measurement of cross section off nitrogen targets which is essential to understanding the cosmic ray showers and atmospheric neutrino production for experiments such as ino .
this is made possible since mipp upgrade will be able to collect 5 millions events per day assuming a beam spill rate of one 4 second spill every 2 minutes , .calculated beam rates for @xmath2 protons on primary target for mipp upgrade[mippbeam ] [ cols="^,^,^,^,^,^,^,^ " , ]
single strange hypernuclei ( with one @xmath3 or one @xmath4 ) have production cross sections in the 100 to 250 nbarn / steradian at 1 to 2 gev / c proton beams on nuclear targets as can be seen for figure [ single ] .
doubly strange hypernuclei ( with two @xmath3 ) are 10 to 50 nbarn / sr with a kaon or hyperon beam on nuclear target and are more difficult to measure . the diagrams for doubly strange hypernuclei can be found in figure [ double ] .
doubly strange hypernuclei with one @xmath5 are not discovered yet , but 7 candidate events are known to exist .
the binding energies for nuclei as a function of atomic weight can be found in figure [ binding ] .
the binding energies decrease with atomic weight indicating that the hypernuclei are less tightly bound as atomic weight increases .
this behavior is also reflected in the lifetimes plot in figure [ lifetimes ] where the lifetimes of the bound lambda is seen to decrease with increasing atomic weight .
@xmath6 beams can study the background to this process @xcite , scaledwidth=60.0% ] , scaledwidth=60.0% ] using mipp upgrade , we have complete forward acceptance using the tpc and full particle identification ( using the tpc and the time of flight ) for particles of momentum less than 1.5 gev / c . in the backward hemisphere in the lab frame
, we have the plastic ball that catches recoil particles .
we investigate the channels that are likely to yield hypernuclei .
we employ the high energy physics notation to denote the reactions .
the channels involving a @xmath7 beam can be denoted by @xmath8 in the above equation , @xmath7 beam remains a @xmath7 beam and fragments a proton target to form a @xmath6 and a @xmath3 .
such reactions are called diffractive dissociations of the target and these maintain a large cross section as the beam energy increases .
the @xmath6 in the final state is detected by the tpc and the time of flight systems and the @xmath3 is absorbed in the nucleus to make the hypernucleus .
there may be other particles in the finally state , which the particle identification of mipp will ensure contain a net strangeness of zero .
the counterpart of the above reaction with a neutron target can be written @xmath9 the @xmath10 in the final state can be detected by the tpc .
similarly with a proton beam @xmath11 @xmath12 these form the proton counter part . using a k@xmath13 beam
, one can look for the reaction @xmath14 @xmath15 one can test the backgrounds to this reaction using k@xmath16 beams which should not produce hypernuclei .
the high acceeptance and particle identification of mipp upgrade promises to produce a high quality hypernucleus experiment .
this work was done at fermilab operated by fermi research alliance , llc under contract no .
de - ac02 - 07ch11359 with the united states department of eneegy .
9 `` forward neutron production at the fermilab main injector '' , t. s. nigmanov et al phys .
d83 ( 2011)012002 .
`` measurement of particle production on the numi target with a 102 gev / c proton beam in the mipp experiment '' , paper in preparation . a copy of the mipp upgrade proposal may be found at https://mipp-docdb.fnal.gov:440/cgi-bin/showdocument?docid=1046 v. n. fetisov , nucl .
a639 ( 1998 ) , p177 . h. c. bhang et al
a639 ( 1998 ) , p269 and 270 s. ajimura et al . ,
a639 ( 1998 ) p 93 - 99
|
we describe the state of analysis of the mipp experiment , its plans to upgrade the experiment and the impact such an upgraded experiment will have on hypernuclear physics .
|
light scattering lies at the heart of optics .
to harvest weak signals , collecting a large fraction of scattered light is essential .
this is most directly achieved using a lens with a large numerical aperture .
an alternative and more powerful approach , however , is to couple the scattering object to an optical resonator , even if the relevant light mode covers only a small solid angle .
the potential offered by the resonator comes from the purcell effect @xcite , generally associated with the enhanced spontaneous emission rate of an excited particle in electromagnetically confined space .
although spontaneous emission is a quantum - mechanical phenomenon , the modification of its rate by the resonator can be explained classically as a light interference effect @xcite .
this makes the purcell enhancement a universal phenomenon , independent of the internal structure of the particles under investigation .
the purcell effect therefore occurs both for light scattering from quantum objects with discrete energy levels and classical objects described as oscillating dipoles . while the purcell effect has extensively been studied for quantum - mechanical objects over the last several decades , experiments involving classical objects are rare . despite its conceptual simplicity
, the effect of a cavity on the classical light - scattering properties has been scrutinized only recently using relatively complex systems such as subwavelength - sized fiber tips @xcite or silicon nanocrystals @xcite .
moreover , the simplicity of the classical description has not been emphasized in @xcite , where the scattering of far - detuned light into the cavity mode was described by a semiquantum model for the interaction between the classical oscillating dipole and the quantized modes of the light field .
hence , this leaves open the question about the classical nature of the purcell effect .
equally important , a demonstration of the purcell effect for the detection of extremely weak signals in the far - detuned , classical regime is still lacking .
the enhancement of such rayleigh scattering from atomic or molecular gases by means of an optical resonator would enable one to observe in a sensitive and non - destructive way particles for which a closed two - level system or a near - resonant laser is not available . here
we report on an experiment where the enhancement of rayleigh scattering into an optical fabry - perot resonator is quantitatively studied using various thermal gases consisting of either atoms ( xe ) , homonuclear ( n@xmath0 ) or heteronuclear molecules ( cf@xmath1h ) .
the light is detuned by hundreds of nanometers from the nearest electronic or vibrational transition .
this detuning by more than one octave makes excitation of the system negligible . for different values of the cavity finesse , we observe an increased rate of rayleigh scattering , in agreement with the predictions of a completely classical model based on the interference of intracavity light waves
firstly , this proves that the purcell effect is classical , and secondly , it shows that an explanation in terms of a modified local mode density is not needed .
comparing the power scattered into the fundamental cavity mode to that scattered into the same mode but under free - space conditions , we find an enhancement by a factor of up to 38 .
. scattered light leaking out of the cavity through the oc is mode matched into a single - mode fiber , and then detected by an avalanche photodiode ( apd ) . ]
the experimental setup is depicted in figure [ fig : expsetup ] .
a 10w single - frequency laser at wavelength @xmath2=532 nm was used .
this wavelength was chosen as a compromise between the @xmath3 dependence of the rayleigh scattering cross section on the laser wavelength , available laser power and sufficient detuning from deep - ultraviolet electronic transitions of the used species .
the cavity mirrors are separated by 0.6 cm , and the mirrors radius of curvature is 4.5 cm .
this combination results in a spacing of transverse modes of 4.1ghz , larger than the observed doppler widths of the gases used .
the cavity length is tuned by a piezoelectric tube separating the two mirrors . to modify the cavity finesse ,
the mirror on the outcoupling side ( oc ) is exchanged , while the other mirror ( hr , @xmath4=99.7% ) is used for all measurements . for the various combinations of mirrors with measured intensity reflectivities
@xmath4=(99.7% , 98.9% , 95.9% ) , a cavity finesse of @xmath5=(1000 , 400 , 100 ) is determined from the cavity line width as observed in transmission .
the cavity is placed inside a vacuum chamber , which is pumped out to @xmath6 mbar before the various gases are introduced .
the pump beam was focussed to a waist of about 50@xmath7 as measured by a beam profiler .
this value was chosen to match the waist of the fundamental cavity mode , which was calculated to be @xmath8 . since we pump with powers of up to 5w and expect a scattered power into the cavity mode of only a few fw , suppression of stray light is crucial . towards this end ,
light leaving the cavity through the outcoupling mirror is send to a single - mode fiber which is aligned for optimum transmission of the @xmath9 fundamental cavity mode . behind the fiber ,
the light is detected by an avalanche photodiode operated in single - photon - counting mode .
mode are separated by the cavity s free spectral range of 24.9ghz .
the signal of the individual gases is normalized to the signal obtained with xe .
the line profile originates from the doppler broadening of the light scattered by the thermal gases . ] in the experiment , we align the pump beam perpendicular to the cavity axis as indicated in figure [ fig : expsetup ] , scan the cavity over at least one free - spectral range and monitor the intensity of the light leaking out of the cavity . as shown in figure [ phd : fig : rayleigh : modespec ] , we observe mainly light scattered into the @xmath9 mode , which is selected by the single - mode fiber .
a small amount of light scattered into other cavity modes is visible as well due to imperfect mode matching into the fiber . for the various gases used ,
the peaks show different heights and widths .
the different widths originate from the different masses and , hence , the different doppler broadenings of the thermal gases .
the different heights are caused by the specific polarizabilities of the different particles as well as by the frequency overlaps of the cavity mode with the doppler profiles . within experimental accuracy ,
the background signal observed for n@xmath0 and cf@xmath1h agrees with the one obtained for xe .
doppler broadening and the multitude of thermally populated internal molecular states prevent the observation of a raman spectrum , which could benefit from a cavity enhancement as well @xcite .
polarization dependence to the data . for light scattering from a classical oscillating dipole
, such a @xmath10 polarization dependence is expected . ]
a signature of rayleigh scattering is the @xmath10 polarization dependence when driven by linearly polarized light . here
, @xmath11 denotes the azimuthal angle of the pump beam s polarization direction with respect to the cavity axis .
figure [ phd : fig : rayleigh : poldep ] shows the polarization dependence of light scattered into the @xmath9 mode of the cavity , when @xmath11 is rotated using a pockels cell .
it follows a @xmath10 dependence , as expected for emission from a classical oscillating dipole .
when the polarization direction is aligned with the cavity axis @xmath12 , no rayleigh scattering into the cavity mode is expected due to the small solid angle of the cavity covering only the nodal line of the dipole pattern .
the residual signal amplitude of 12% can be explained by imperfect linear polarization of the pump beam .
we emphasize that the @xmath10 dependence , observed for all the gases investigated , is the characteristic feature for a scattering process involving an oscillating dipole .
h. the data of the cavity scans obtained for the different pressures have been individually normalized to the maximal value . in the measurement with the lowest pressure ,
the theoretically predicted doppler profile is shown in red . at higher pressure , sidebands caused by brillouin scattering appear . ] at low densities , the line shape is determined by doppler broadening as shown in figure [ phd : fig : rayleigh : pressdep ] . due to the @xmath13 scattering geometry , we observe a doppler width @xmath14=@xmath15 , with @xmath16 being the standard doppler width of absorption spectroscopy . by varying the pressure ,
we observe a linear dependence of the scattered power on the gas density , thereby ruling out scattering off the surfaces as the signal source .
we also find the expected linear dependence of the rayleigh scattered light on the pump beam power over the entire examined range of 0.15w .
however , when increasing the gas pressure from 100 mbar , where we typically operate , to values of up to 1bar , we observe the appearance of a substructure on the doppler - broadened peaks .
these sidebands are caused by brillouin scattering on density waves in the gas @xcite .
to model the experiment , we describe the scattering from a polarizable particle , a classical oscillating dipole , into a single cavity mode as follows .
the polarizable particle is pumped from the side by the electric field @xmath17 and scatters an electric field into the cavity as indicated in figure [ fig : rayleigh : theory ] . to derive an expression for the intracavity light field and the power scattered into the cavity mode
, we follow the intracavity light field on one round trip through the cavity . along its way
, the intracavity light field grows due to light scattering from the pump beam into the cavity mode and decreases due to transmission losses through the mirrors .
as schematically shown in figure [ fig : rayleigh : theory ] , the scatterer emits an electromagnetic field into a right - traveling and a left - traveling wave , which can interfere .
the right - running intracavity field @xmath18 at the position of the scatterer is @xmath19 here , @xmath20 is a proportionality factor for the scattering process , and @xmath21 @xmath22 are the mirror s amplitude reflection coefficients ( @xmath23 is the mirror reflectivity ) .
the distance between the mirrors is denoted by @xmath24 , and @xmath25 is the wave number of the light field .
the first term of , @xmath26 , is the field directly scattered into the right direction .
the second term of , @xmath27 , is the field initially scattered into the left direction . after being reflected by the left cavity mirror
, it arrives at the position of the scatterer as a right - traveling electric field .
the additional phase factor @xmath28 accounts for a translation @xmath29 of the scatterer from the cavity centre .
the last term of , @xmath30 , describes the right - traveling intracavity field after one round trip . by solving ,
one finds for the right - traveling intracavity field @xmath31 in the following discussion , we assume a resonant cavity , @xmath32 . as a first step , we use to calculate the right - traveling intensity , @xmath33 since the experiment is performed with thermal gases , the scattering particles are randomly distributed . therefore , an average over the position of the particles , i.e. , @xmath34 , must be taken , resulting in @xmath35 in the limit of high mirror reflectivities , @xmath36 , the right - traveling intracavity intensity can be expressed in a form with a clear physical meaning . with the cavity finesse @xmath37{\mathrm{r}_1\mathrm{r}_2}/(1-\sqrt{\mathrm{r}_1\mathrm{r}_2 } ) \approx\pi/(1-\sqrt{\mathrm{r}_1\mathrm{r}_2})$ ] one finds @xmath38 in an intuitive picture , @xmath39 is the number of reflections in the resonator .
the right - traveling electric field increases linearly with this number .
the intensity is proportional to the square modulus of the electric field , which results in @xmath40 . as a next step ,
we quantify the enhancement of the power emitted from the cavity .
we compare the power of the scattered light leaving the resonator through one of the cavity mirrors to the power of the light scattered into the same mode but under free - space conditions . for this and the following discussions ,
it is convenient to convert all intensities @xmath41 into powers @xmath42 taking the area of the cavity mode as a reference .
from , the power leaving the cavity through the right mirror is found to be @xmath43 when the taylor series expansion @xmath44 and @xmath45 are used . in the case of a symmetric cavity , @xmath46
, this can be simplified to yield @xmath47 without the cavity in place , the right - traveling scattered field is given by @xmath48 . the power scattered into the right - traveling mode defined by the cavity but under free - space conditions
is therefore given by @xmath49 .
comparing this with the power leaving the cavity through the right mirror , , the cavity enhances the detectable power by a factor @xmath50 .
half of the light , which has been scattered into the cavity mode , leaks out of the left cavity mirror . taking this into account , the total power scattered into the cavity mode
is given by @xmath51 for particles maximally coupled to the cavity , i.e. , no averaging over particle position , an additional factor 2 comes in , @xmath52 this has to be compared to the total power scattered under free - space conditions into the mode @xmath53 defined by the cavity but without the cavity enhancement , which is @xmath54 therefore , the cavity enhances the power scattered into the same mode of the electromagnetic field by a factor @xmath55 as compared with the free - space situation .
so far , only a single mode of the electromagnetic field was considered .
thereby , the power scattered into the cavity mode was compared to the power scattered into the same mode in the free - space situation .
such a scenario refers to our experiment .
it is , however , also interesting to compare the power @xmath56 scattered into the cavity mode with the power @xmath57 scattered into the full @xmath58 solid angle under free - space conditions . in the end
, this calculation allows to establish a connection between the classical wave interference model and the well - known purcell factor @xcite . of the lowest - order transverse hermite - gaussian cavity mode @xmath9 along its propagation direction @xmath59 .
@xmath60 is the beam waist , and @xmath61 is the rayleigh length .
( b ) emission pattern of the classical oscillating dipole oriented perpendicular to the @xmath59-axis . ] as a starting point , we calculate which fraction of the light scattered by the oscillating dipole is radiated into the mode defined by the cavity but under free - space conditions . for this purpose ,
the overlap integral between the intensity - normalized electric fields of the dipole mode and the cavity mode are evaluated . the cavity mode and the dipole emission pattern
are schematically shown in figure [ fig : rayleigh : modes ] .
the oscillating dipole is oriented perpendicular to the cavity axis , which results in maximal light scattering into the cavity mode . in the far field
, the scalar dipole mode @xmath62 can be defined in spherical coordinates as @xmath63 where the phase variation @xmath64 along the propagation direction @xmath65 has been omitted .
the dipole mode is normalized to its intensity , @xmath66 in the experiment , the lowest - order transverse hermite - gaussian mode @xmath9 was used .
the electric field of the fundamental cavity mode @xmath67 in cylindrical coordinates is given by @xmath68 again , all phase variations have been omitted , as justified below .
@xmath69 is a normalization constant , and @xmath70 is the beam waist .
this cavity mode traveling in one direction is normalized in cylindrical coordinates , @xmath71 now , the overlap integral between the intensity - normalized electric fields of the dipole mode @xmath72 , , and the fundamental hermite - gaussian cavity mode @xmath73 , , is evaluated in the far field at a fixed value @xmath59 .
there , both the dipole mode and the cavity mode have phase fronts lying on spheres centered at the position of the scatterer .
all phase factors therefore cancel in the evaluation of the overlap integral between the two modes , which justifies the initial disregard of these . for the real - valued dipole and cavity modes ,
the overlap in one propagation direction @xmath59 is given by @xmath74 which is evaluated in cylindrical coordinates .
due to the small transverse extent of the cavity mode , the dipole mode @xmath72 is approximated by its value on the @xmath59-axis , @xmath75 in the limit @xmath76 , this results in @xmath77 using , the overlap between the electric fields of the dipole mode and the cavity mode was evaluated .
now , the power scattered into the cavity mode @xmath78 and the total power scattered into the dipole mode @xmath79 are compared . since power scales with the square of the electric field , one finds @xmath80 in the evaluation of the overlap integral only one propagation direction of the cavity mode was taken into account .
the factor 2 in accounts for the two independent propagation directions of the cavity mode .
since there is no interference between the fields scattered into these two directions , intensities are summed up , not fields . using and , the total power scattered into the dipole mode is found to be @xmath81 to retrieve the purcell factor , the power @xmath56 scattered into the cavity mode by a particle maximally coupled to the cavity is compared with the power @xmath79 scattered by the particle into the dipole mode under free - space conditions .
one finds @xmath82 in its most common form , the purcell factor @xcite is defined as @xmath83 with @xmath84 being the quality factor of the cavity and @xmath85 the mode volume . for an atomic system ,
the free - space decay rate @xmath86 of an excited state is changed to the value @xmath87 by the presence of a cavity being resonant with the atomic transition frequency .
for a two - mirror fabry - perot cavity with mirror separation @xmath24 , the mode volume @xmath85 is well approximated by @xmath88 the @xmath84 factor is defined as @xmath89 with @xmath90 being the resonance frequency of the cavity and @xmath91 the cavity linewidth .
for a two - mirror fabry - perot cavity , this can be related to the cavity finesse @xmath5 and the mirror separation @xmath24 by @xmath92 and @xmath93 therefore , the cavity @xmath84 factor can be expressed in terms of the cavity finesse @xmath5 and mirror separation @xmath24 as @xmath94 putting all of the above together , one finds for the purcell factor of a standing - wave fabry - perot cavity at its antinode @xmath95 this is exactly the same expression as the one derived from the classical wave - interference model .
this shows that the purcell factor is indeed fully explained by interference of scattered fields @xcite .
to test the expected dependence of the scattered power on the cavity finesse , we measured the cavity output power for the three available values @xmath5=(1000,400,100 ) .
the spectral overlaps of a doppler - broadened thermal xe gas with the cavity modes are ( 4.2,10.1,33.4)% for the used cavities , respectively . from the measured signals
@xmath96 , we calculate the power scattered from particles at rest and plot it in figure [ phd : fig : rayleigh : finessedep ] . as predicted by the classical wave - interference model and by the purcell theory , a linear dependence on the cavity finesse is found .
the scattered power also depends on the polarizability of the particles . for fixed cavity finesse
, we find a ratio between scattered powers for xe , cf@xmath1h , and n@xmath0 of ( 1:0.35:0.1 ) . based on their polarizabilities @xcite and doppler broadenings
we expect a ratio of ( 1:0.36:0.09 ) , which is in good agreement with the observations . finally , we show the enhancement of the scattered power leaving the cavity as compared with the free - space situation . for this measurement ,
the single - mode fiber was used to select the same mode in a free - space scattering experiment . under otherwise identical conditions ( 100 mbar xe pressure , 1w pump power ) , but without the cavity in place , we observe a scattered power of @xmath97 .
we estimate a factor 2 for the accuracy of this measurement , constrained by mode - matching .
this power measured in free - space scattering has to be compared with the value of @xmath98 measured with a cavity finesse of 1000 , an enhancement factor of 38 . to test the classical model for the cavity enhancement we take into account the limited overlap of the doppler - broadened spectral profile with the cavity mode of 4.2% , which does not occur in the free - space scattering , and the enhancement by the cavity with a finesse of 1000 .
putting things together , we expect a value of @xmath99 for the free - space scattering measurement without the cavity , taking @xmath98 measured with the cavity as a reference . within the experimental uncertainties ,
the calculated and measured enhancements are in agreement , showing that the process is well described by the classical model .
the experiment shows the potential of an optical cavity to enhance weak signals in light - scattering experiments .
the large detuning of the laser light from any optical transition allows the method to be applied to different species as demonstrated with the use of xe , n@xmath0 and cf@xmath1h , independent of their specific internal level structure .
this opens a new pathway to optical detection of deeply - bound ultracold molecules , for which the efficient production was reported recently @xcite .
so far , however , detection of these molecules requires either dissociation to unbound atom pairs or ionization . here , cavity - enhanced rayleigh scattering might be an attractive in - situ detection technique since it does not rely on closed cycling transitions .
although in our experiment room - temperature gases at typical densities of @xmath100@xmath101 were used , the number of particles contributing to the scattering into the cavity is only about @xmath102 for a cavity finesse of 1000 .
the aforementioned molecule production techniques can typically prepare @xmath103 ultracold molecules at cloud sizes compatible with the cavity - mode diameter in the presented experiment .
since doppler broadening is absent for light scattered by these trapped ultracold molecules , the cavity finesse could be increased to an experimentally realistic value of @xmath103 .
furthermore , alkali dimers have a more than 10@xmath104 larger polarizability @xmath20 as compared with xe ( static polarizabilities @xmath20 , which should give a lower bound for the polarizabilities at optical frequencies : krb : 502a.u . , cs@xmath0 : 670a.u . ,
xe : 27.3a.u . , 1a.u.=0.148@xmath105 )
@xcite , and the rayleigh - scattering cross section @xmath106 scales proportionally to @xmath107 .
altogether , ensemble scattering rates into the cavity of the order of 100khz can therefore be anticipated for such an ultracold molecular sample .
hence , on average , every molecule scatters photons at a total rate of the order of 1hz , with a ratio between scattering into the cavity and scattering into free space @xmath108 for the given parameters .
this allows for a nearly demolition - free detection with almost no photon - recoil heating . as an extension of the present setup ,
the use of cavities with degenerate transverse modes could boost the total power scattered into the cavity even further @xcite .
collective enhancement effects @xcite , observed for atomic ensembles @xcite , could result in an additional increase of signal .
the authors thank t.j .
kippenberg , m. kowalewski and k. murr for fruitful discussions .
the support of the dfg through the excellence cluster `` munich centre for advanced photonics '' and euroquam ( cavity - mediated molecular cooling ) is acknowledged .
|
we demonstrate purcell - like enhancement of rayleigh scattering into a single optical mode of a fabry - perot resonator for several thermal atomic and molecular gases .
the light is detuned by more than an octave , in this case by hundreds of nanometers , from any optical transition , making particle excitation and spontaneous emission negligible .
the enhancement of light scattering into the resonator is explained quantitatively as an interference effect of light waves emitted by a classical driven dipole oscillator .
applications of our method include the sensitive , non - destructive in - situ detection of ultracold molecules .
|
for all terms and definitions , not defined specifically in this paper , we refer to @xcite . unless mentioned otherwise , all graphs considered here are simple , finite and have no isolated vertices .
let @xmath7 denote the set of all non - negative integers . for all @xmath8 , the sum of these sets
is denoted by @xmath9 and is defined by @xmath10 .
the set @xmath9 is called the sumset of the sets @xmath11 and @xmath12 . an _ integer additive set - indexer _ ( iasi , in short ) is defined in @xcite as an injective function @xmath0 such that the induced function @xmath1 defined by @xmath2 is also injective .
the cardinality of the labeling set of an element ( vertex or edge ) of a graph @xmath6 is called the _ set - indexing number _ of that element .
[ l - card ] @xcite let @xmath11 and @xmath12 be two non - empty finite subsets of @xmath7 . then , @xmath13 .
therefore , for an integer additive set - indexer @xmath3 of a graph @xmath6 , we have @xmath14 , where @xmath5 .
@xcite an iasi @xmath3 is said to be a _ weak iasi _
if @xmath4 for all @xmath5 . a graph which admits a weak iasi may be called a _ weak iasi graph_. a weak iasi @xmath3 is said to be _ weakly @xmath15-uniform iasi _ if @xmath16 , for all @xmath5 and for some positive integer @xmath15 .
@xcite an element ( a vertex or an edge ) of graph which has the set - indexing number 1 is called a _
mono - indexed element _ of that graph .
@xcite the _ sparing number _ of a graph @xmath6 is defined to be the minimum number of mono - indexed edges required for @xmath6 to admit a weak iasi and is denoted by @xmath17 .
first , we recall the following theorems proved in @xcite .
[ t - sb1 ] @xcite if a connected graph @xmath6 admits a weak iasi , then @xmath6 is bipartite or @xmath6 has at least one mono - indexed edge .
[ t - sb2 ] @xcite let @xmath6 be a bipartite graph which admits a weak iasi and let @xmath18 be two non - adjacent vertices in @xmath6 .
then , @xmath19 is a weak iasi graph if and only if @xmath19 is bipartite or @xmath20 is a mono - indexed edge . due to theorem [ t - sb1]and thorem [ t - sb2 ] , we have the following theorem .
[ t - sngb1 ] let @xmath6 be a weak iasi graph and let @xmath21 be a minimal subset of @xmath22 such that @xmath23 is bipartite .
then , @xmath24 .
let @xmath6 be a non - bipartite weak iasi graph .
let @xmath25 be a minimal set of edges of @xmath6 such that @xmath26 is bipartite .
then , @xmath26 is a maximal bipartite subgraph of @xmath6 . by theorem [ t - sb1 ]
, @xmath27 admits a weak iasi without any mono - indexed edge .
now , @xmath28 is a non - bipartite graph .
then , by theorem [ t - sb2 ] , @xmath29 is a mono - indexed edge in @xmath6 , for @xmath30 .
hence , in each edge in @xmath21 must be mono - indexed .
therefore , @xmath24 .
the size of the maximal bipartite subgraph of a graph @xmath6 is denoted by @xmath31 .
then , as a consequence of theorem [ t - sngb1 ] , we have [ c - t - sngb1 ] for a graph @xmath6 , the size of its maximal bipartite subgraph is @xmath32 .
let @xmath6 be a non - bipartite graph with the edge set @xmath33 .
let @xmath34 such that @xmath23 is the maximal bipartite subgraph of @xmath6 .
then , @xmath35 . by theorem [ t - sngb1 ]
, we have @xmath36
. therefore , @xmath32 .
the most interesting question in this context is whether the sparing number of a graph is related to other parameters like chromatic number , matching number etc .
the following results establish some relations between these parameters of certain graph classes .
let @xmath6 be a @xmath15-chromatic graph with colors @xmath37 .
then , @xmath6 is a @xmath15-partite graph in which the @xmath38-th partition , denoted by @xmath39 , consists of all the vertices of @xmath6 which has the color @xmath40 .
the set @xmath41 is called the _
@xmath38-th color class_. hence we have , the sparing number of a @xmath15-chromatic graph @xmath6 is the total number of vertices in all the color classes of @xmath6 except the two maximal color classes .
let @xmath42 be the @xmath15 colors that are used for coloring the vertices @xmath6 and let @xmath41 be the color class in which the vertices are assigned the color @xmath40 . without loss of generality ,
let @xmath43 be in the descending order of their cardinality .
label all the vertices in the color class @xmath44 by distinct singleton sets of non - negative integers and label all the vertices in the color class @xmath45 by non - singleton singleton sets of non - negative integers .
since @xmath44 and @xmath45 are maximal color classes , each vertex , say @xmath46 , in other color classes must be adjacent to at least one vertex of @xmath44 and at least one vertex of @xmath45 . since @xmath6 is a weak iasi graph ,
@xmath46 can be labeled only by a singleton set , which is not used for labeling before .
hence , the sparing number of @xmath6 is the total number of vertices in all the color classes of @xmath6 except the two maximal color classes .
now , recall the following theorem proved in @xcite .
[ t - snc1 ] @xcite the sparing number of an odd cycle is @xmath47 and that of a bipartite graph is @xmath48 . due to theorem [ t - snc1 ] , we have , if @xmath6 is path or cycle , then @xmath49 , where @xmath50 is the chromatic number of @xmath6 . if @xmath6 is path or an even cycle , then @xmath51 and by theorem [ t - snc1 ] , @xmath52 .
hence , @xmath49 . if @xmath6 is an odd cycle , then @xmath53 and by theorem [ t - snc1 ] , @xmath54 .
hence , @xmath49 .
this completes the proof .
a _ matching _
@xmath55 in a given graph @xmath6 is a set of pairwise non - adjacent edges .
a _ maximum matching _ is a matching that contains the largest possible number of edges .
the matching number @xmath56 of @xmath6 is the size of a maximum matching .
the following theorem establishes the relation between the sparing number and the matching number paths and cycles .
[ t - snmncp ] if @xmath6 is a path or a cycle on @xmath57 vertices , then @xmath58 , where @xmath56 is the matching number of @xmath6 .
if @xmath6 is a path , then by theorem [ t - snc1 ] , @xmath52 .
if @xmath6 is even , then @xmath59 and if @xmath6 is odd , then @xmath60 .
therefore , the matching number of @xmath6 is @xmath61 .
therefore , @xmath58 . if @xmath6 is an even cycle , then by theorem [ t - snc1 ] , @xmath52 and the matching number of @xmath6 is @xmath59 . here , @xmath62 .
if @xmath6 is an odd cycle , then by theorem [ t - snc1 ] , @xmath54 and the matching number of @xmath6 is @xmath63 . here ,
@xmath64 .
this completes the proof .
now , recall the following theorem .
[ t - snug ] @xcite let @xmath65 and @xmath66 be two weak iasi graphs . then , @xmath67 . as a consequence of theorem [ t - snmncp ] and theorem [ t - snug ] , we have [ t - snfuoc ] if @xmath6 can be decomposed into finite number of odd cycles , then @xmath68 , where @xmath69 is the number of edge - disjoint cycles and @xmath70 is the size of the cycle @xmath71 in @xmath6 .
let @xmath6 can be decomposed into finite number of odd cycles .
let @xmath72 be the distinct edge disjoint odd cycles in @xmath6 so that @xmath73 .
then , @xmath74 by theorem [ t - snmncp ] , @xmath75 .
since , all @xmath76 are edge - disjoint , by theorem [ t - snug ] , @xmath77}}\\ & = & \displaystyle{\sum_{i=1}^m{\lceil{\frac{n_i}{2}}\rceil } } - \displaystyle{\sum_{i=1}^m{\nu(c_i)}}\\ & = & \displaystyle{\sum_{i=1}^m{\lceil{\frac{n_i}{2}}\rceil } } - \nu(g ) \end{aligned}\ ] ] this completes the proof .
[ r - snc1]we observe that theorem [ t - snfuoc ] , does not hold for edge disjoint bipartite graphs since equation [ myeq1 ] does not hold for edge disjoint even cycles .
it can also be verified that equation [ myeq1 ] will not hold for those graphs , which can be decomposed into edge disjoint cycles , having more than one even cycle .
now , recall the well - known theorem on eulerian graphs .
[ t - eg1 ] @xcite a graph @xmath6 is eulerian if and only if it can be decomposed into edge disjoint cycles . in view of theorem
[ t - eg1 ] and remark [ r - snc1 ] , theorem [ t - snfuoc ] can be rewritten as follows . if @xmath6 is an eulerian graph that has at most one even cycle then @xmath68 , where @xmath69 is the number of edge - disjoint cycles and @xmath70 is the size of the cycle @xmath71 in @xmath6 .
a _ vertex cover _ of a graph @xmath6 is a subset @xmath78 of @xmath79 such that each edge of @xmath6 has at least one end vertex in @xmath78 .
the number of vertices in a minimum vertex cover of a graph @xmath6 is known as the _ vertex covering number _ or simply the _ covering number _ of @xmath6 and is denoted by @xmath80 .
the relations of weak iasi and sparing number with its minimal vertex cover and covering number are discussed in the following theorem .
[ t - snvc ] the minimum number of mono - indexed vertices in a weak iasi graph @xmath6 is equal to the covering number of @xmath6 .
moreover , the sparing number of @xmath6 is the number edges of @xmath6 which have both of their end vertices in the minimal vertex cover of @xmath6 .
let @xmath78 be the minimal vertex cover of @xmath6 .
then , @xmath78 and @xmath81 are two partitions of @xmath79 .
since every edge of @xmath6 has at least one end vertex in @xmath78 , no two vertices in @xmath81 are adjacent in @xmath6 and some edges in @xmath6 may have both of their end vertices in @xmath78 . since @xmath6 is a weak iasi graph , the vertices in @xmath78 can not be labeled by non - singleton vertices .
hence , the set - labels of the vertices in @xmath6 are singleton sets of non - negative integers . that is , the number of mono - indexed vertices in @xmath6 is equal to its covering number .
since no two vertices in @xmath81 are adjacent each other , we can label all the vertices in @xmath81 by non - singleton sets of non - negative integers .
hence , an edge having both of its end vertices in @xmath78 is mono - indexed .
therefore , the sparing number of @xmath6 is the number of edges of @xmath6 which have both of their end vertices in @xmath78 .
this completes the proof .
an _ independent set _ or _ stable set _ of a graph @xmath6 is a subset @xmath82 of @xmath79 in a graph , no two vertices in @xmath82 are adjacent .
it is to be noted that a subset @xmath82 of @xmath79 is independent if and only if its complement @xmath83 is a vertex cover .
the number of vertices in a maximal independence set is called the independence number of @xmath6 and is denoted by @xmath84 .
the following theorem establishes the relation between the independence number and covering number of a given graph @xmath6 .
[ t - rincn ] @xcite for any connected non - trivial graph @xmath6 , @xmath85 .
note that if @xmath78 is a vertex cover of a graph @xmath6 , then the set @xmath81 is an independent set of @xmath6 .
hence , if @xmath78 is a minimal vertex cover of @xmath6 , then @xmath81 is a maximal independent set of @xmath6 .
hence we have , let @xmath6 be a weak iasi graph on @xmath57 vertices .
then , the number of mono - indexed vertices in @xmath6 is @xmath86 , where @xmath84 is the independence number of the graph @xmath6 . by theorem [ t - snvc ] ,
the number of mono - indexed vertices of a graph @xmath6 is equal to @xmath80 , the covering number of @xmath6 . by theorem [ t - rincn ]
, we have @xmath87 .
therefore , the number of mono - indexed vertices = @xmath86 . if @xmath78 is a minimal vertex cover of @xmath6 and if all vertices in @xmath78 are labeled by distinct singleton sets of non - negative integers , @xmath81 is a maximal independent set in which all elements are non - adjacent to each other and the set - labels of all of them are distinct non - singleton sets of non - negative integers . hence , we have the following proposition . if @xmath6 is a weak iasi graph , then the maximum number of vertices that are not mono - indexed in @xmath6 is equal to the independence number of @xmath6 .
in view of theorem [ t - sngb1 ] , we discuss the sparing number of some standard non - bipartite graphs .
the sparing number of petersen graph is @xmath88 .
name the vertices of the petersen graph @xmath6 as shown in figure [ g - pgwl ] . in @xmath6 , both the external cycle @xmath89 and the internal cycle @xmath90 are odd and hence @xmath6 is not bipartite .
hence , remove one edge , say @xmath91 , from the cycle @xmath92 and remove one edge , say @xmath93 , from the cycle @xmath94 . in the resultant graph @xmath95 , the cycles @xmath96 and @xmath97 are odd .
hence , remove the common edge @xmath98 from @xmath99 and @xmath100 .
let @xmath101 , which is shown in figure [ g - pg - mbsg ] .
all the cycles in @xmath102 are of even length and hence @xmath102 is the maximal bipartite subgraph of @xmath6 .
let @xmath103 .
then , by theorem [ t - sngb1 ] , @xmath104 . 0.45 0.45 for the peterson graph @xmath6 , @xmath105 .
hence , by corollary [ c - t - sngb1 ] , the number of edges in a maximal bipartite subgraph of @xmath6 is @xmath106 .
one other well - known non - bipartite graph is the _ frucht graph _ , which is is a @xmath88-regular graph with @xmath107 vertices , @xmath108 edges and every vertex of which can be distinguished topologically from every other vertex .
hence we have , the sparing number of frucht graph is @xmath88 .
name the vertices of the frucht graph @xmath6 as shown in figure [ g - fgwl ] . in @xmath6 ,
the cycle @xmath109 is of odd length and has a common edge @xmath110 with the even cycle @xmath111 and has a common edge @xmath112 with the even cycle @xmath113 .
hence , remove the edge @xmath114 from @xmath6 .
the cycle @xmath115 in @xmath116 has a common edge @xmath117 with the even cycle @xmath111 and has a common edge @xmath118 with the odd cycle @xmath119 .
hence , remove the edge @xmath117 from @xmath116 .
the cycle in @xmath120 has a common edge @xmath121 with the even cycle @xmath122 and has a common edge @xmath123 with the odd cycle @xmath124 .
hence , remove the edge @xmath123 from @xmath120 .
now all the cycles in @xmath125 are of even length , which is shown in figure [ g - fg - mbsg ] .
therefore , @xmath102 is the maximal bipartite subgraph in @xmath6 . if @xmath126 , then by theorem [ t - sngb1 ] , @xmath104 . 0.45 0.45 for the frucht graph @xmath6 , @xmath127 .
hence , by corollary [ c - t - sngb1 ] , the number of edges in a maximal bipartite subgraph of @xmath6 is @xmath128 .
the @xmath129 is a triangle - free graph with @xmath130 vertices , @xmath131 edges .
the following theorem establishes the sparing number of @xmath132 graph .
the sparing number of the @xmath132 graph is @xmath133 .
name the vertices of @xmath132 graph @xmath6 as shown in figure [ g - ggwl ] .
remove the edge @xmath134 common to the odd cycles @xmath135 and @xmath136 , the edge @xmath137 common to @xmath138 and @xmath139 , the edge @xmath140 common to @xmath141 and @xmath142 , the edge @xmath143 common to @xmath144 and the edge @xmath145 and @xmath146 common to @xmath147 and @xmath148 . the resultant graph @xmath149 contains no odd cycles , as shown in [ g - ggmbsg ] , and hence @xmath102 is a maximal bipartite subgraph of @xmath6 . let @xmath150 .
then , the sparing number @xmath151 . 0.45 0.45 for the @xmath132 graph @xmath6 , @xmath152 . hence , by corollary [ c - t - sngb1 ] , the number of edges in a maximal bipartite subgraph of @xmath6 is @xmath153 .
another well - known non - bipartite graph is the @xmath154 , which is a @xmath88-regular graph with @xmath107 vertices , @xmath108 edges and every vertex of which can be distinguished topologically from every other vertex .
hence we have , the sparing number of _ @xmath155 _ is @xmath156 .
name the vertices of the _ @xmath155 _ @xmath6 as shown in figure [ g - dgwl ] . in @xmath6 ,
the cycle @xmath157 and @xmath158 are of odd length . hence , remove the edge @xmath159 from @xmath92 and remove the edge @xmath160 from @xmath94 .
the cycles @xmath161 and @xmath97 in @xmath162 are odd cycles and have a common edge @xmath98 and the cycles @xmath163 and @xmath164 in @xmath162 are odd cycles and have a common edge @xmath165 .
hence , remove the edges @xmath98 and @xmath165 from @xmath162 .
now all the cycles in @xmath166 are of even length , which is shown in [ g - dg - mbsg ] .
therefore , @xmath102 is the maximal bipartite subgraph in @xmath6 . if @xmath167 , then by theorem [ t - sngb1 ] , @xmath168 .
0.4 0.45 for the @xmath169 graph @xmath6 , @xmath127 . hence , by corollary [ c - t - sngb1 ] , the number of edges in a maximal bipartite subgraph of @xmath6 is @xmath170 .
a _ dodecahedron _ is another popular non - bipartite graph .
it is a @xmath88-regular graph with @xmath131 vertices and @xmath171 edges and it contains @xmath107 pentagons ( the cycle @xmath172 ) .
hence , we have the sparing number of the dodecahedron is @xmath173 .
name the vertices of dodecahedron @xmath6 as shown in figure [ g - dgwl ] . from @xmath6 , remove the edge @xmath174 common to the odd cycles @xmath175 and @xmath176 , the edge @xmath177 common to @xmath178 and @xmath179 , the edge @xmath180 common to @xmath181 and @xmath182 , the edge @xmath183 common to @xmath184 and @xmath185 , the edge @xmath186 common to @xmath187 and @xmath188 and the edge @xmath91 common to @xmath189 and @xmath190 .
the resultant graph @xmath191 contains no odd cycles , as shown in [ g - ddg - mbsg ] , and hence @xmath102 is a maximal bipartite subgraph of @xmath6 .
let @xmath192 .
then , the sparing number @xmath193 . for a dodecahedron @xmath6 , @xmath194 .
hence , by corollary [ c - t - sngb1 ] , the number of edges in a maximal bipartite subgraph of @xmath6 is @xmath195 . 0.4 0.4
in this paper , we have established some results on the sparing number of certain graphs and established some relations between the sparing number and some other parameters of certain graph classes .
the admissibility of weak iasi by various graph classes , graph operations and graph products and finding the corresponding sparing numbers are open .
more properties and characteristics of different types of iasis , both uniform and non - uniform , are yet to be investigated .
the problems of establishing the necessary and sufficient conditions for various graphs and graph classes to have certain iasis are also open . 25 j a bondy and u s r murty , ( 2008 ) .
* graph theory * , springer .
g chartrand and p zhang , ( 2005 ) .
* introduction to graph theory * , mcgraw - hill inc .
n deo , ( 1974 ) .
_ graph theory with applications to engineering and computer science _ , phi learning .
j a gallian , ( 2011 ) . _ a dynamic survey of graph labelling _ , the electronic journal of combinatorics ( ds 16 ) . k a germina and t m k anandavally , ( 2012 ) . _ integer additive set - indexers of a graph : sum square graphs _ ,
journal of combinatorics , information and system sciences , * 37*(2 - 4 ) , 345 - 358 .
k a germina and n k sudev , _ on weakly uniform integer additive set - indexers of graphs _ , int .
math forum * 8*(37 ) , 1827 - 34 .
k a germina and n k sudev , ( 2013 ) . _ some new results on strong integer additive set - indexers _ , communicated .
f harary , ( 1994 ) .
* graph theory * , addison - wesley publishing company inc .
n k sudev and k a germina , ( 2014 ) . _ on integer additive set - indexers of graphs _ , to appear in int .
appl . , * 8*(2 ) , .
n k sudev and k a germina , ( 2014)._a characterisation of weak integer additive set - indexers of graphs _ , ispacs j. fuzzy set valued analysis , * 2014*(2014 ) ( ) , article id : jfsva-00189 , 7 pages .
n k sudev and k a germina , ( 2014 ) . _
weak integer additive set - indexers of certain graph operations _ , to appear in global j. math sci . :
theory and practical .
n k sudev and k a germina , _ on the sparing number of certain graph structures _ , communicated .
n k sudev and k a germina , _ weak integer additive set - indexers of graph products _ , communicated .
d b west , ( 2001 ) . *
introduction to graph theory * , pearson education inc .
|
an integer additive set - indexer is defined as an injective function @xmath0 such that the induced function @xmath1 defined by @xmath2 is also injective .
an iasi @xmath3 is said to be a weak iasi if @xmath4 for all @xmath5 .
a graph which admits a weak iasi may be called a weak iasi graph .
the set - indexing number of an element of a graph @xmath6 , a vertex or an edge , is the cardinality of its set - labels .
the sparing number of a graph @xmath6 is the minimum number of edges with singleton set - labels , required for a graph @xmath6 to admit a weak iasi . in this paper
, we study the sparing number of certain graphs and the relation of sparing number with some other parameters like matching number , chromatic number , covering number , independence number etc . *
key words * : weak integer additive set - indexers , mono - indexed elements of a graph , sparing number of a graph . +
* ams subject classification : 05c78 *
|
x - ray absorption variability is a common feature in active galactic nuclei ( agn ) .
an analysis of a sample of nearby x - ray obscured agn with multiple x - ray observations , performed a few years ago ( risaliti et al .
2002 ) revealed that column density ( n@xmath0 ) variations are almost ubiquitous in local seyfert galaxies .
more recent observations performed with _ xmm - newton _ , _ chandra _ and _ suzaku _ further confirmed this finding .
the physical implications of these measurements are that the circumnuclear x - ray absorber ( or , at least , one component of it ) must be clumpy , and located at sub - parsec distances from the central source .
the comparison between different observations , typically performed at time distances of months - years , only provides upper limits to the intrinsic time scales of @xmath3 variations .
an improvement of these estimates can only be obtained through observational campaigns within a few weeks / days , and/or through the search for n@xmath0 variations within single long observations .
such short time - scale studies have been already performed for a handful sources : ngc 1365 ( risaliti et al .
2005 , 2007 , 2009 ) , ngc 4388 ( elvis et al .
2004 ) , ngc 4151 ( puccetti et al .
in particular , in the case of the agn in ngc 1365 we revealed extreme spectral changes , from compton - thin ( n@xmath0 in the range 10@xmath4 @xmath5 ) to reflection - dominated ( n@xmath0@xmath610@xmath7 @xmath5 ) in time scales from a couple of days to @xmath810 hours .
such rapid events imply that the absorption is due to clouds with velocity v@xmath610@xmath1 km s@xmath2 , at distances of the order of 10@xmath9 gravitational radii ( assuming that they are moving with keplerian velocity around the central black hole ) .
the physical size and density of the clouds are of the order of 10@xmath10 cm and 10@xmath11 - 10@xmath12 @xmath13 , respectively .
all these physical parameters are typical for broad line region ( blr ) clouds , strongly suggesting that the x - ray absorber and the clouds responsible for broad emission lines in the optical / uv are one and the same . here
we present more evidence , based on time resolved x - ray spectroscopy , of variable x - ray absorption due to blr clouds , and we show how in the best cases we can use x - ray observations to directly probe the nature and physical state of blr clouds .
variations on time scales of a few days discovered with campaigns of snapshot observations.,width=510 ] x - ray absorption due to blr clouds crossing the line of sight to the observer implies variability time scales from hours to weeks , depending of the physical size of the x - ray source . we can obtain a rough estimate of the occultation times assuming a cloud moving with keplerian velocity around the central black hole , and a size of the x - ray source of 10 r@xmath14 : t@xmath83@xmath1510@xmath16 m@xmath17 v@xmath18 , where m@xmath17=m/(10@xmath19 m@xmath20 ) and v@xmath21=v/(10@xmath1 km s@xmath2 ) .
two observational methods have been used to investigate @xmath3 variations at these time scales : + 1 ) snapshot observations of the same source , repeated every few days .
after the first successful campaign of six short ( 10 ks ) _ chandra _ observations of ngc 1365 , we repeated this approach for other bright compton - thin sources with column densities between 10@xmath4 and 10@xmath7 @xmath5 , using both archival data and dedicated observations . in this way we discovered clear cases of @xmath3 variations in ngc 4151 and ugc
4203 ( one of the few `` changing look '' sources , known for having been observed in the past in both compton - thin and reflection - dominated states ) .
a further case of such variations has been discovered by bianchi et al .
( 2009 ) in ngc 7582 ( fig . 1 ) .
+ 2 ) study of the spectral variations of bright sources during long observations .
this method consists of a two - phase analysis : we first use the hardness - ratio light curve to select the time intervals where strong spectral variations occurred ; we then perform a complete analysis of the spectra obtained from these intervals , in order to measure possible n@xmath0 varations ( and to check if the spectral changes are due to other effects , such as variations of the slope of the continuum emission ) . this approach is illustrated in fig . 2 for the long _ xmm - newton _ observation of mrk 766 .
a complete analysis of these data is presented in turner et al .
2007 and miller et al . 2007 .
with respect to these works , our analysis is in many aspects less detailed ( though the results are in full agreement ) , but is more effective in isolating the effects of possible n@xmath0 variations .
we note that mrk 766 is a narrow line seyfert 1 , so on average we do not expect to observe complete x - ray absorption of the x - ray source .
however , as we show below , isolated clouds occasionally cross the line of sight , producing measurable absorption in the x - ray spectrum .
the upper panel of fig . 2 shows the standard 2 - 10 kev flux light curve for this observation , with the well known strong variability on time scales of thousands of seconds , or even shorter .
the lower panel shows the light curve of the ( 6 - 10 kev)/(2 - 5 kev ) flux ratio .
in general , this light curve shows much smaller variations , indicating that the continuum shape remains the same during most of the luminosity variations .
however , clear exceptions are observed in at least three intervals , highlighted in fig . 2 .
during these intervals it is possible that a cloud with n@xmath0 of the order of 10@xmath4 @xmath5 has covered the central source , strongly decreasing the observed flux in the soft band , without affecting the hard band , and therefore increasing the observed hardness ratio . following this interpretation , in the first interval we should be observing a cloud uncovering the x - ray source ( with the covering phase occurred before the beginning of the observation ) ; in the second interval another cloud shoud be covering the x - ray source , with
the uncovering phase not observed due to the `` dead time '' between two consecutive _ xmm - newton _ orbits .
finally , in the third interval we should be observing the whole eclipse . , and covering factor . , width=510 ] in order to check this scenario , we performed a complete analysis of the spectra obtained from the three highlighted intervals , and of those obtained from the third and fourth orbit , representing the standard spectral state of the source . in this analysis
we allowed all the main spectral parameters of the model to vary among the different intervals .
the results of the this study , illustrated in fig . 3 are the following : 1 ) the 2 - 10 kev spectrum obtained from the third and fourth orbit ( the `` standard '' state ) is well reproduced by a typical model for type 1 agns , consisting of a power law , a reflection component and an iron emission line ; 2 ) the spectral variations observed in the three intervals discussed above are completely reproduced by three absorption components with column densities in the range 1 - 3 10@xmath4 @xmath5 . the analysis briefly summarized here has been applied to several sources with long archival observations by _ xmm - newton _ and _ suzaku_. up to now , we obtained unambiguous evidence for @xmath3 variations in ngc 4388 , mcg-6 - 30 - 15 , mrk 766 , ngc 1365 .
a further case of similar variations has been found in a _
observation of ngc 3227 ( liu et al .
2010 , subm . ) .
a summary of all the positive detections obtained so far is shown in table 1 . in all cases
the inferred physical parameters for the obscuring clouds clearly indicate that the x - ray absorption is due to blr clouds .
the analysis presented above can be further refined in the few cases where the statistics is high enough to perform a more detailed study , or where multiple occultation event occur .
in particular , we briefly summarize two results from the analysis of the two so - far best studied sources , mrk 766 and ngc 1365 .
+ * 1 ) velocity distribution of blr clouds . *
the analysis of the spectral variations in mrk 766 ( fig . 2 - 4 ) directly provides contraints on the distribution of blr cloud velocities , from the comparison of the durations of the observed occultations .
the actual values of the cloud velocity in each case can not be precisely determined , being dependent on the exact size of the x - ray source .
however , the ratio between the durations of the different eclipses is an estimate of the velocity ratios . in particular , the duration of the two first occultation events ( fig .
2 ) , is about 4 - 5 times longer than that of the third eclipse .
this suggests a similar spread in the velocity of blr clouds .
+ * 2 ) structure of blr clouds .
* the model adopted to reproduce the observed occultation events consists of an absorption component with a constant ( during the eclipse ) n@xmath0 , and a variable covering fraction , @xmath22 .
this is an extreme simplification of the cloud structure , which is assumed to have constant column density , and very sharp edges .
the model can be improved by releasing the assumption of constant n@xmath0 .
doing so , in most of the cases analyzed so far we found that it is impossible to significantly constrain the values of n@xmath0 and @xmath22 during the different phases of an eclipse , due to their strong degeneracy .
in other words , it is only possible to study the evolution of one parameter , freezing the other one .
the only exceptions found so far , i.e. spectra with enough signal - to - noise to analyze the evolution of the two parameters at the same time , are mrk 766 and ngc 1365 . in fig .
4a we show the results of a recent analysis of a _ suzaku _ observation of ngc 1365 ( maiolino et al .
2010 , subm ) , where the hardness ratio light curve clearly shows at least two possible occultation events .
we performed a complete spectral analysis of several short intervals before , during and after the eclipses , and we obtained the results shown in the two lower panels of fig . 4 , showing the light curves of the covering factor and the column density .
the most interesting results come from the second eclipse : the covering factor @xmath22 increases during the occultation phase , with a constant n@xmath0@xmath83@xmath1510@xmath4 @xmath5 .
then , in the subsequent phase , @xmath22 remains compatible with 1 ( i.e. complete covering ) while the column density decreases .
this behaviour can only be explained with a `` cometary shape '' of the obscuring cloud , with a high column density head , and a tail with decreasing n@xmath0 .
an analogous result has been obtained ( though with smaller statistical significance ) for the first two eclipses in mrk 766 ( fig .
the third eclipse of mrk 766 is instead too fast to allow a detailed spectral analysis ( the spectral counts in such short time intervals are not enough to break the degeneracy between @xmath22 and @xmath3 , as shown in fig . 5 ) .
however , a strong indication of a cometary tail comes from the residuals with respect to the best fit model during and after the eclipse , showing strong absorption lines due to fe xxv in outflow with a velocity of 10 - 15000 km s@xmath2 .
these features completely disappear after @xmath840 ks .
a complete analysis of these spectra , where we demonstrate the high statistical significance of the lines detections , is presented in a dedicated paper ( risaliti et al .
2010 , subm . )
the highly ionized component revealed by the absorption lines strongly suggest the presence of a ionized outflowing tail associated to the obscuring cloud .
we note that this highly ionized component is not present in the first two occultation events in mrk 766 , but only in the third , much faster event .
even if one case is clearly not enough to build a complete model , the data suggest a scenario where the obscuring clouds are distributed in a large range of distances from the center ( if we assume keplerian velocity , the factor @xmath84 - 5 of spread in velocities corresponds to a factor of @xmath820 of spread in distances ) , and where the clouds closer to the center are also the more ionized ones .
thsi simple scheme is illustrated in fig .
.list of sources with n@xmath0 variations on short time scales [ cols="<,^,^,^,^",options="header " , ]
we have presented new time - resolved spectral studies of several bright agns , showing column density variability on time scales for a few hours to a few days .
the main results of our analysis are : + 1 ) fast ( hours - days ) column density variability is common among agns .
this implies that the observed variable x - ray absorption is due to clouds with velocities , densities , sizes and distances from the central black hole of the same order of those of blr clouds .
+ 2 ) in the highest signal - to - noise studies , it is possible to investigate the structure and shape of the single obscuring clouds in detail .
this reveals a `` cometary '' profile , with a high column density head and a tail with decreasing n@xmath0 .
+ these results show that the x - ray absorption is at least in part due to blr clouds , and that x - ray spectroscopy can be a powerful tool to diretly measure the physical properties of the broad line region in agns .
+ 3@xmath1510@xmath4 @xmath5 in the head , to a few 10@xmath23 @xmath5 in the tail ( maiolino et al .
2010 ) . in mrk 766 , we suggest that the fast cloud with the ionized tail is much closer to the center than the two slower , neautral clouds .
the ionized tail in the fast cloud is probably in outflow ( not shown here ; a complete analysis is presented in risaliti et al .
2010).,width=510 ] the work presented here is only the first part of an on - going comprehensive analysis of all the bright agns with long high quality x - ray observations . at present , the evidence of `` common '' short time scale n@xmath0 variations presented here is based on a sparse sample of sources , with no homogeneous selection criteria .
the next major step is therefore the selection and the homogeneous analysis of a representative sample of the local agn population .
this work will provide a quantitative estimate of the occurrence of n@xmath0 variability on short time scales , and hopefully , will lead to the discovery of more high signal to noise spectra of occultations , such as the ones found in ngc 1365 and mrk 766 , in order to perform other studied of the physical propoerties of blr clouds . + * acknowledgements.*. this work has been partially supported by nasa grants nnx08ax78 g and g08 - 9107x , and by grant asii - inaf i/088/06/0
|
we present new evidence of x - ray absorption variability on time scales from a few hours to a few days for several nearby bright agns .
the observed n@xmath0 variations imply that the x - ray absorber is made of clouds eclipsing the x - ray source with velocities in excess of 10@xmath1 km s@xmath2 , and densities , sizes and distances from the central black hole typical of blr clouds .
we conclude that the variable x - ray absorption is due to the same clouds emitting the broad emission lines in the optical / uv .
we then concentrate on the two highest signal - to - noise spectra of eclipses , discovered in two long observations of ngc 1365 and mrk 766 , and we show that the obscuring clouds have a cometary shape , with a high density head followed by a tail with decreasing @xmath3 .
our results show that x - ray time resolved spectroscopy can be a powerful way to directly measure the physical and geometrical properties of blr clouds .
|
the problem of the cosmological constant is salient yet unsettled in cosmology .
the smallness of the effective cosmological constant recently observed @xmath2 poses the most difficult problems involving cosmology and elementary particle physics theory . to explain the striking cancellation between the `` bare '' cosmological constant and the ordinary vacuum energy contributions of the quantum fields ,
many mechanisms have been proposed @xcite .
the `` cosmological constant problem '' can be expressed as the discrepancy between the negligible value of @xmath1 for the present universe as seen by the successes of newton s theory of gravitation @xcite whereas the values @xmath3 larger is expected by the glashow - salam - weinberg model @xcite and by grand unified theory ( gut ) it should be @xmath4 larger @xcite .
the cosmological term @xmath1 is then small at the present epoch simply because the universe is too old .
the problem in this approach is to determine the right dependence of @xmath1 upon @xmath5 or @xmath6 .
models with a relic cosmological constant @xmath1 have received ample attention among researchers recently for various reasons ( see refs .
@xcite - @xcite and references therein ) . some of the recent discussions on the cosmological constant `` problem '' and on cosmology with a time - varying cosmological constant by ratra and peebles @xcite , dolgov et al .
@xcite , dolgov @xcite , and sahni and starobinsky @xcite point out that in the absence of any interaction with matter or radiation , the cosmological constant remains a `` constant '' , however , in the presence of interactions with matter or radiation , a solution of einstein equations and the assumed equation of covariant conservation of stress - energy with a time - varying @xmath1 can be found .
for these solutions , conservation of energy requires decrease in the energy density of the vacuum component to be compensated by a corresponding increase in the energy density of matter or radiation .
earlier researches on this topic , are contained in zeldovich @xcite , weinberg @xcite and carroll , press and turner @xcite .
recent observations by perlmutter et al .
@xcite and riess et al .
@xcite strongly favour a significant and positive value of @xmath1 .
their findings arising from the study of more than @xmath7 type ia supernovae with redshifts in the range @xmath8 suggest friedmann models with negative pressure matter such as a cosmological constant @xmath9 , domain walls or cosmic strings ( vilenkin @xcite , garnavich et al .
@xcite ) . recently , carmeli and kuzmenko @xcite , behar and carmeli @xcite have shown that the cosmological relativistic theory predicts the value for cosmological constant @xmath10 .
this value of `` @xmath1 '' is in excellent agreement with the measurements recently obtained by the high - z supernova team and supernova cosmological project ( garnavich et al .
@xcite , perlmutter et al .
@xcite , riess et al .
@xcite , schmidt et al .
@xcite ) the main conclusion of these observations is that the expansion of the universe is accelerating .
several ans@xmath11tz have been proposed in which the @xmath1 term decays with time ( see refs .
gasperini @xcite , berman @xcite , berman et al .
@xcite , @xmath12zer and taha @xcite , freese et al .
@xcite , peebles and ratra @xcite , chen and wu @xcite , abdussattar and viswakarma @xcite , gariel and le denmat @xcite , pradhan @xcite , pradhan et al .
@xcite ) . of the special interest
is the ans@xmath11tz @xmath13 ( where @xmath14 is the scale factor of the robertson - walker metric ) by chen and wu @xcite , which has been considered or modified by several authors ( abdel - rahaman @xcite , carvalho et al .
@xcite , silveira and waga @xcite , vishwakarma @xcite ) . in classical electromagnetic theory ,
the electromagnetic field has two independent invariants @xmath15 and @xmath16 . the classification of the field is characterized by the property of the scalar @xmath17 = @xmath18 + @xmath19 .
when @xmath20 = @xmath21 the field is said to be null and for any observer @xmath22 = @xmath23 and @xmath24.@xmath25 = 0 where e and h are the electric and magnetic vectors respectively . when @xmath20 @xmath26 , the field is non - null and there exists an observer for which @xmath27 = @xmath28 , @xmath29 and @xmath30 being scalars .
it is easy to see that if @xmath16 = 0 , @xmath15 @xmath26 then either @xmath24 = @xmath21 or @xmath25 = @xmath21 .
these we call the magnetic and electric fields , respectively .
if @xmath15 = @xmath21,@xmath16 @xmath26 then @xmath24 = @xmath31 .
this we call the ` electromagnetic wrench ' with unit ` pitch ' . in this case
maxwell s equations lead to an empty electromagnetic field with constant electric and magnetic intensities . in the case of gravitational field , the number of independent scalar invariants of the second order is fourteen .
the independent scalar invariants formed from the conformal curvature tensor are four in number . in the case of petrov type
d space - times , the number of independent scalar invariants are only two , viz . @xmath32 and @xmath33 .
analogous to the electromagnetic case , the electric and magnetic parts of free gravitational field for an observer with velocity @xmath34 are defined as @xmath35 and @xmath36 ( @xcite ) .
it is clear from the canonical form of the conformal curvature for a general type d space - time that there exists an observer for which @xmath37 , where @xmath38 being integers and @xmath39 .
the field is said to be purely magnetic type for @xmath40 , @xmath39 . in this case
we have @xmath41 and @xmath42 . the physical significance for the gravitational field of being magnetic type
is that the matter particles do not experience the tidal force .
when @xmath39 and also @xmath43 , we call that there is a ` gravitational wrench ' of unit ` pitch ' @xmath44 in the free gravitational field @xcite .
if ` pitch ' is unity then we have @xmath45 .
+ + the space - time having a symmetry property is invariant under a continuous group of transformations .
the transformation equations for such a group of order r is given by @xmath46 which satisfy the differential equations @xmath47 where @xmath48 are r essential parameters .
the vectors @xmath49 are the killing vectors for the group @xmath50 of isometry satisfying the killing s equation @xmath51 a subspace of space - time is said to be the surface of transitivity of the group if any point of this space can be transformed to another point of it by the action of this group . a space - time is said to be spatially homogeneous if it admit a group @xmath50 of isometry which is transitive on three dimensional space - like hyper - surfaces .
the group @xmath52 of isometry was first considered by @xcite who obtained nine different types of isometry group known as the bianchi types . the space - time which admits @xmath53 group of isometry
is known as locally rotationally symmetric ( lrs ) which always has a @xmath52 as its subgroup belonging to one of the bianchi type provided this @xmath52 is simply transitive on the three dimensional hyper - surface @xmath6 = constant .
considerable work has been done in obtaining various bianchi type cosmological models and their inhomogeneous generalization .
barrow @xcite pointed out that bianchi @xmath0 models of the universe give a better explanation of some of the cosmological problems like primordial helium abundance and they also isotropize in a special sense .
looking to the importance of bianchi type @xmath0 universes , many authors @xcite@xmath54@xcite have studied it in different context .
recently bali et al .
@xcite have obtained some lrs bianchi type @xmath0 cosmological models imposing two types of conditions over the free gravitational fields . in this paper
we have revisited and generalized the solutions obtained by bali et al . @xcite .
we have considered an lrs bianchi type @xmath0 space - time and obtained models with free gravitational field of purely ` magnetic type ' and also in the presence of ` gravitational wrench ' of unit ` pitch ' in the free gravitational field .
it is found that the ` magnetic ' part of the free gravitational field induces shear in the fluid flow , which is zero in the case of a ` electric ' type free gravitational field representing an unrealistic distribution in this case .
this paper is organized as follows .
the introduction and motivation are laid down in sec . @xmath55 .
the metric and the field equations are given in sec .
@xmath56 . in sec .
@xmath57 , solutions representing lrs bianchi type @xmath0 cosmological models with perfect fluid and bulk viscous fluid are obtained imposing the condition when the free gravitational field is purely magnetic type ( @xmath58 . in sec .
@xmath59 , we obtain the solution in presence of perfect fluid imposing the condition when there is a ` gravitational wrench ' of unit ` pitch ' in the free gravitational field i.e. @xmath45 .
discussion and concluding remarks are given in the last sec @xmath60 , .
we consider an lrs bianchi type @xmath0 universe for which @xmath61 where @xmath62 , @xmath63 , @xmath64 , @xmath65 .
+ the energy - momentum tensor for a perfect fluid distribution with comoving flow vector @xmath34 is given by @xmath66 where @xmath67 , @xmath68 and @xmath69 being respectively , energy density and thermodynamic pressure of the fluid .
here we obtain @xmath70 the einstein s field equations ( in gravitational units c = 1 , g = 1 ) read as @xmath71 for the line element ( [ eq4 ] ) has been set up as @xmath72 @xmath73 @xmath74 here , and also in the following expressions a dot indicates ordinary differentiation with respect to @xmath6 .
+ + the average scale factor @xmath14 for lrs bianchi type @xmath0 model is defined by @xmath75 a volume scale factor is given by @xmath76 the generalized mean hubble parameter @xmath25 is given by @xmath77 where @xmath78 , @xmath79 + the expansion scalar @xmath80 and shear scalar @xmath81 are obtained as @xmath82 and @xmath83 respectively .
the average anisotropy parameter is given by @xmath84 where @xmath85 .
+ the deceleration parameter ( @xmath86 ) is defined as @xmath87 the non - vanishing physical components of @xmath88 for the line ( [ eq18 ] ) are given by @xmath89 @xmath90,\ ] ] @xmath91.\ ] ] equations ( [ eq8])-([eq10 ] ) are three relations in five unknowns @xmath92 , @xmath93 , @xmath69 , @xmath68 and @xmath1 . for complete solutions of equations ( [ eq8])-([eq10 ] )
, we need two extra conditions .
the research on exact solutions is based on some physically reasonable restrictions used to simplify the einstein equations .
although the distribution of matter at each point determines the nature of expansion in the model , the latter is also affected by the free gravitational field through it s effect on the expansion , vorticity and shear in the fluid flow .
a prescription of such a field may therefore be made on an _ a priori _ basis .
the cosmological models of friedman robertson walker , as well as the universe of einstein - de sitter , have vanishing free gravitational fields . in the following two cases we impose different conditions over the free gravitational field to find the deterministic solutions .
in this section we have extended the solution obtained by bali et al .
@xcite by revisiting their solution .
when free gravitational field is purely magnetic type ( @xmath58 ) , we have @xmath94 and @xmath41 . from ( [ eq18 ] ) , we obtain @xmath95 equation ( [ eq8 ] ) together with ( [ eq9 ] ) reduce to @xmath96 from eqs .
( [ eq20 ] ) and ( [ eq21 ] ) , we obtain two independent equations @xmath97 @xmath98 let @xmath99
. then eqs .
( [ eq22 ] ) and ( [ eq23 ] ) take the form @xmath100 and @xmath101 respectively , where @xmath102 is defined by @xmath103 eqs .
( [ eq24 ] ) and ( [ eq25 ] ) lead to @xmath104 on substituting @xmath105 , the above second order differential equation reduces to the form @xmath106 which gives @xmath107 where @xmath108 and @xmath109 are arbitrary constants .
hence , we obtain @xmath110},\ ] ] where @xmath111 stands for @xmath102 . therefore , from eqs .
( [ eq25 ] ) , ( [ eq25 ] ) and ( [ eq30 ] ) , we obtain @xmath112}}{(k_{1 } - 2t)},\ ] ] @xmath113 where @xmath111 is given by @xmath114},\ ] ] @xmath115 , @xmath116 and @xmath117 being arbitrary constants
. therefore the geometry of the universe ( [ eq4 ] ) reduces to the form @xmath118}}{(k_{1 } - 2t)^{2}}\biggl[-dt^{2 } + dx^{2 } + \frac{\exp{[2t(t - k_{1})]}}{k_{3}^{2}}\ ] ] @xmath119.\ ] ] the expressions for pressure @xmath69 and density @xmath68 for the model ( [ eq34 ] ) are given by @xmath120\exp{[2t(t - k_{1 } ) ] } - \lambda(t),\ ] ] @xmath121\exp{[2t(t - k_{1 } ) ] } + \lambda(t),\ ] ] for the specification of @xmath122 , we assume that the fluid obeys an equation of state of the form @xmath123 where @xmath124 is a constant . + using ( [ eq37 ] ) in ( [ eq35 ] ) and ( [ eq36 ] ) , we obtain @xmath125}\ ] ] eliminating @xmath68 between eqs .
( [ eq36 ] ) and ( [ eq38 ] ) , we obtain @xmath126 } + \frac{3(1 - \gamma)}{k_{4}^{2}}(k_{1 } - 2t)^{2}\exp{\left[2t(t - k_{1})\right]}\ ] ] vs. time , width=302,height=302 ] vs. time , width=302,height=302 ] from eq .
( [ eq38 ] ) , we observe that @xmath127 is a decreasing function of time and @xmath128 for all times .
figure @xmath55 shows this behaviour of energy density . from eq .
( [ eq39 ] ) , we note the cosmological term @xmath1 is a decreasing function of time and it approaches a small positive value with increase in time . from figure @xmath56 ( @xmath129 are in geometrical units in entire paper ) , we note the same character of @xmath1 .
this is to be taken as a representative case of physical viability of the model .
+ the behaviour of the universe in this model will be determined by the cosmological term @xmath1 , this term has the same effect as a uniform mass density @xmath130 which is constant in space and time .
a positive value of @xmath1 corresponds to a negative effective mass density ( repulsion ) .
hence , we expect that in the universe with a positive value of @xmath1 the expansion will tend to accelerate whereas in the universe with negative value of @xmath1 the expansion will slow down , stop and reverse . in a universe with both matter and vacuum energy ,
there is a competition between the tendency of @xmath1 to cause acceleration and the tendency of matter to cause deceleration with the ultimate fate of the universe depending on the precise amounts of each component .
this continues to be true in the presence of spatial curvature , and with a nonzero cosmological constant it is no longer true that the negatively curved ( `` open '' ) universes expand indefinitely while positively curved ( `` closed '' ) universes will necessarily recollapse - each of the four combinations of negative or positive curvature and eternal expansion or eventual recollapse become possible for appropriate values of the parameters .
there may even be a delicate balance , in which the competition between matter and vacuum energy is needed drawn and the universe is static ( non expanding ) .
the search for such a solution was einstein s original motivation for introducing the cosmological constant .
recent cosmological observations ( garnavich et al .
@xcite , perlmutter et al .
@xcite , riess et al .
@xcite , schmidt et al .
@xcite ) suggest the existence of a positive cosmological constant @xmath1 with the magnitude @xmath131 .
these observations on magnitude and red - shift of type ia supernova suggest that our universe may be an accelerating one with induced cosmological density through the cosmological @xmath1-term .
thus the nature of @xmath1 in our derived model of the universe is consistent with recent observations ( garnavich et al .
@xcite , perlmutter et al .
@xcite , riess et al .
@xcite , schmidt et al .
+ + * some physical and geometric features of the model * + the expressions for kinematical parameters i. e. the scalar of expansion @xmath80 , shear scalar @xmath81 , average scale factor @xmath14 , proper volume @xmath132 and average anisotropy parameter @xmath133 for the model ( [ eq36 ] ) are given by @xmath134\exp{[t(t - k_{1})]},\ ] ] @xmath135},\ ] ] @xmath136 @xmath137}}{k_{3}^{2}(k_{1 } - 2t)^{4}},\ ] ] @xmath138 the directional hubble s parameters @xmath139 , @xmath140 and @xmath141 are given by @xmath142\exp{[t(t - k_{1}}],\ ] ] @xmath143,\ ] ] where the mean hubble s parameter is given by @xmath144}\ ] ] vs. time , width=302,height=302 ] the model ( [ eq36 ] ) starts expansion with a big - bang singularity from @xmath145 and it goes on expanding till @xmath146 respectively correspond to the cosmic time @xmath147 and @xmath148 .
the is found to be realistic everywhere in this time interval for @xmath149 .
the model behaves like a steady - state de - sitter type universe at late times where the physical and kinematic parameters @xmath68 , @xmath69 , @xmath80 tend to a finite value , however shear vanishes there .
the model has a point type singularity at time @xmath150 .
the singular behaviour may be close to cosmic origin or outside the evolution .
the average anisotropy parameter @xmath133 remains uniform and isotropic through out the evolution of the universe .
this would depend on physical properties of matter and radiation .
this may need detailed study to make better quantifiable view . from figure @xmath57
, it can be seen that in the early stages of the universe , _
, @xmath151 , the scale factor of the universe had been approximately constant and had increased very slowly . at specific time
the universe had exploded suddenly and expanded to large scale .
this is good matching with big bang scenario .
this is indicated in first part ( top ) of figure @xmath57 .
later singular behaviour depends on @xmath152 .
astronomical observations of large - scale distribution of galaxies of our universe show that the distribution of matter can be satisfactorily described by a perfect fluid .
but large entropy per baryon and the remarkable degree of isotropy of the cosmic microwave background radiation , suggest that we should analyze dissipative effects in cosmology .
further , there are several processes which are expected to give rise to viscous effect .
these are the decoupling of neutrinos during the radiation era and the recombination era @xcite , decay of massive super string modes into massless modes @xcite , gravitational string production @xcite and particle creation effect in grand unification era @xcite .
it is known that the introduction of bulk viscosity can avoid the big bang singularity .
thus , we should consider the presence of a material distribution other than a perfect fluid to have realistic cosmological models ( see grn @xcite for a review on cosmological models with bulk viscosity ) .
a uniform cosmological model filled with fluid which possesses pressure and second ( bulk ) viscosity was developed by murphy @xcite .
the solutions that he found exhibit an interesting feature that the big bang type singularity appears in the infinite past .
+ in presence of bulk viscous fluid distribution , we replace isotropic pressure @xmath69 by effective pressure @xmath153 in eq .
( [ eq35 ] ) where @xmath154 where @xmath102 is the coefficient of bulk viscosity .
+ the expression for effective pressure @xmath153 for the model eq .
( [ eq34 ] ) is given by @xmath155\exp{[2t(t - k_{1 } ) ] } - \lambda(t).\ ] ] thus , for given @xmath156 we can solve for the cosmological parameters .
in most of the investigation involving bulk viscosity is assumed to be a simple power function of the energy density ( pavon @xcite , maartens @xcite , zimdahl @xcite , santos @xcite ) @xmath157 where @xmath158 and @xmath30 are constants . for small density
, @xmath30 may even be equal to unity as used in murphy s work @xcite for simplicity . if @xmath159 , eq . ( [ eq24 ] ) may correspond to a radiative fluid ( weinberg @xcite ) .
near the big bang , @xmath160 is a more appropriate assumption ( belinskii and khalatnikov @xcite ) to obtain realistic models .
+ for simplicity sake and for realistic models of physical importance , we consider the following two cases ( @xmath161 ) : vs. time , width=302,height=302 ] vs. time , width=302,height=302 ] when @xmath40 , eq .
( [ eq50 ] ) reduces to @xmath162 constant . with the use of eqs .
( [ eq36 ] ) , ( [ eq37 ] ) and ( [ eq40 ] ) , eq .
( [ eq49 ] ) reduces to @xmath163 } + \ ] ] @xmath164}.\ ] ] eliminating @xmath127 between eqs .
( [ eq36 ] ) and ( [ eq51 ] ) , we obtain @xmath165 } + \frac{3(1 - \gamma)}{k_{4}^{2}}(k_{1 } - 2t)^{2}\exp{[2t(t - k_{1})]}\ ] ] @xmath166}.\ ] ] vs. time , width=302,height=302 ] vs. time , width=302,height=302 ] when @xmath159 , eq .
( [ eq50 ] ) reduces to @xmath167 . with the use of eqs .
( [ eq36 ] ) , ( [ eq37 ] ) and ( [ eq40 ] ) , eq .
( [ eq49 ] ) reduces to @xmath168}}{4\pi k_{4}\left[k_{4}(1 + \gamma ) - \xi_{0}\{6 + ( k_{1 } - 2t)^{2}\}\exp{[t(t - k_{1})]}\right]}\ ] ] eliminating @xmath127 between eqs . ( [ eq36 ] ) and ( [ eq53 ] )
, we obtain @xmath169 } - \frac{3(k_{1 } - 2t)^{2}\exp{[2t(t - k_{1})]}}{k_{4}^{2}\bigl[k_{4}(\gamma + 1)- \xi_{0}\{6 + ( k_{1 } - 2t)^{2}\}\exp{[t(t - k_{1})]}\bigr]}\times\ ] ] @xmath170}\bigr]\ ] ] where @xmath108 , @xmath116 , @xmath158 , and @xmath171 are constants . + from eqs .
( [ eq51 ] ) and ( [ eq53 ] ) , we observe that @xmath127 in both models are a decreasing function of time and @xmath128 for all times .
figures @xmath59 and @xmath172 show this behaviour of energy density . from eqs .
( [ eq52 ] ) and ( [ eq54 ] ) , we see that the cosmological terms @xmath1 in both models are a decreasing function of time and they approach a small positive value at late time . from figures
@xmath60 and @xmath173 , we note this behaviour of cosmological term @xmath1 in both models i and ii .
thus , our models are consistent with the results of recent observations ( garnavich et al .
@xcite , perlmutter et al .
@xcite , riess et al .
@xcite , schmidt et al .
+ the effect of bulk viscosity is to produce a change in perfect fluid and therefore exhibits essential influence on the character of the solution . a comparative inspection of figures @xmath174 show apparent evolution of time due to perfect fluid and bulk viscous fluid .
it is apparent that the vacuum energy density ( @xmath68 ) decays much fast in later case .
also shows effect of uniform viscosity model and linear viscosity model . even in these case decay of vacuum energy density
is much faster than uniform .
so the coupling parameter @xmath175 would be related with physical structure of the matter and provides mechanism to incorporate relevant property . in order to say more specific
, detailed study would be needed which would be reported in future .
similar behaviour is observed for the cosmological constant @xmath1 .
we also observe here that murphy s @xcite conclusion about the absence of a big bang type singularity in the infinite past in models with bulk viscous fluid in general , is not true .
the results obtained by myung and cho @xcite also show that , it is not generally valid since for some cases big bang singularity occurs in finite past . for both models , it is observed that the effect of viscosity prevents the shear and the free gravitational field from withering away .
in this case there is a ` gravitational wrench ' of unit ` pitch ' in the free gravitational field i.e. @xmath176 .
therefore we have @xmath177 in this case , bali et al .
@xcite have investigated the solution given by @xmath178 @xmath179.\ ] ] the expressions for pressure @xmath69 and energy density @xmath68 for the model ( [ eq56 ] ) are obtained as @xmath180 @xmath181 for the specification of @xmath182 , we assume that the fluid obeys an equation of state of the form ( [ eq37 ] ) . using eqs .
( [ eq37 ] ) in ( [ eq57 ] ) and ( [ eq58 ] ) , we obtain @xmath183\exp{\left\{-\frac{3\kappa}{\sqrt{7}}\tan^{-1 } { \left(\frac{4\tau + 3\kappa}{\sqrt{7}}\right)}\right\}}.\ ] ] eliminating @xmath68 between ( [ eq57 ] ) and ( [ eq59 ] ) , we obtain @xmath184\times\ ] ] @xmath185 from preliminary study , we find that both energy density and cosmological constant are negative . hence it will not be studied .
it also shows singular behaviour in energy density at later stage of the evolution .
hence , the model is unphysical for further study .
in this paper , we have studied properties of the free gravitational field and their invariant characterizations and obtained lrs bianchi type @xmath186 cosmological models imposing different conditions on the free gravitational field . in first case , where fee gravitational field is purely magnetic type , we observe that the energy density@xmath187 and cosmological constant @xmath188 are well behaved . also the effect of bulk viscous fluid distribution in the universe is compared with perfect fluid model .
we observe that due to presence of bulk viscous fluid , the rate of decrease in energy density is faster compared to perfect fluid model .
the linear relation of the coefficient of bulk viscosity with mass density provides further enhance decrease rate ( see figures @xmath189 ) .
since we had considered extension of the various models , detailed physical parameter study would be reported in future .
important incorporation is time - dependence of cosmological constant @xmath188 .
the similar enhance decrease rate is also observed for @xmath188 ( see figures @xmath190 ) . the scale factor dependence is also reported .
the nature of @xmath1 in our derived models of the universe is found to be consistent with recent observations ( garnavich et al .
@xcite , perlmutter et al .
@xcite , riess et al .
@xcite , schmidt et al .
authors would like to thank iucaa , pune , india for providing facility and support where part of this work was carried out . the financial support ( grant no.f.6 - 2(29)/2008(mrp / nrcb))of university grants commission , new delhi
is gratefully acknowledged by a. pradhan .
000 s. weinberg , rev .
phys . * 61 * , 1 ( 1989 ) .
s. weinberg : _ gravitation and cosmology _ ( wiley , new york , 1972 ) .
e. s. abers and b. w. lee , phys . rep .
* 9 * , 1 ( 1973 ) .
p. langacker , phys . rep . * 72 * , 185 ( 1981 ) . j. a. frieman and i. waga , phys .
d * 57 * , 4642 ( 1998 ) .
r. carlberg et al . , astrophys .
j. * 462 * , 32 ( 1996 ) .
m. @xmath12zer and m. o. taha , nucl . phys .
b * 287 * , 776 ( 1987 ) .
k. freese , f. c. adams , j. a. frieman and e. motta , ibid .
b * 287 * , 1797 ( 1987 ) .
j. c. carvalho , j. a. s. lima and i. waga , phys .
d * 46 * , 2404 ( 1992 ) . v. silviera and i. waga , ibid .
d * 50 * , 4890 ( 1994 ) .
b. ratra and p. j. e. peebles , phys .
d * 37 * , 3406 ( 1988 ) .
a. d. dolgov : in _ the very early universe _ ,
gibbons , s. w. hawking and s. t. c. siklos ( cambridge university press , cambridge 1983 ) .
a. d. dolgov , m. v. sazhin and ya .
b. zeldovich , _ basics of modern cosmology _ ( editions frontiers 1990 ) .
a. d. dolgov , phys .
d * 55 * , 5881 ( 1997 ) .
v. sahni and a. starobinsky , int . j. mod
d * 9 * , 373 ( 2000 ) 373 . ya .
b. zeldovich , sov .
uspekhi * 11 * , 381 ( 1968 ) .
s. m. carroll , w. h. press and e. l. turner , ann .
. astrophys . *
30 * , 499 ( 1992 ) .
s. perlmutter et al . , astrophys .
j. * 483 * , 565 ( 1997 ) .
+ s. perlmutter et al . ,
nature * 391 * , 51 ( 1998 ) .
+ s. perlmutter et al . , astrophys . j. * 517 * , 565 ( 1999 ) .
a. g. riess et al .
j. * 116 * , 1009 ( 1998 ) .
+ a. g. riess et al .
, astrophys .
j. * 607 * , 665 ( 2004 ) .
a. vilenkin , phys . rep . * 121 * , 265 ( 1985 ) .
p. m. garnavich et al . , astrophys .
j. * 493 * , l53 ( 1998a ) .
+ p. m. garnavich et al . , astrophys .
j. * 509 * , 74 ( 1998b ) .
m. carmeli and t. kuzmenko , int .
41 * , 131 ( 2002 ) .
s. behar and m. carmeli , int .
* 39 * , 1375 ( 2002 ) .
b. p. schmidt et al .
, astrophys .
j. * 507 * , 46 ( 1998 ) .
m. gasperini , phys .
b * 194 * , 347 ( 1987 ) .
+ m. gasperini , class .
* 5 * , 521 ( 1988 ) .
m. s. berman , int .
* 29 * , 567 ( 1990 ) .
+ m. s. berman , int .
* 29 * , 1419 ( 1990 ) .
+ m. s. berman , phys .
d * 43 * , 75 ( 1991 ) .
m. s. berman and m. m. som , int .
* 29 * , 1411 ( 1990 ) .
+ m. s. berman , m. m. som and f. m. gomide , gen .
. grav . * 21 * , 287 ( 1989 ) .
+ m. s. berman and f. m. gomide , gen .
* 22 * , 625 ( 1990 ) .
p. j. e. peebles and b. ratra , astrophys .
j. * 325 * , l17 ( 1988 ) .
w. chen and y.s .
wu , phys .
d * 41 * , 695 ( 1990 ) .
abdussattar and r. g. vishwakarma , pramana j. phys . * 47 * , 41 ( 1996 ) .
j. gariel and g. le denmat , class .
quant . grav .
* 16 * , 149 ( 1999 ) .
a. pradhan , fizika b ( zagreb ) , * 16 * , 205 ( 2007 ) .
+ a. pradhan , commun .
. phys . * 51 * , 367 ( 2009 ) .
a. pradhan and a. kumar , int .
d * 10 * , 291 ( 2001 ) .
+ a. pradhan and v. k. yadav : int .
j. mod phys .
d * 11 * , 983 ( 2002 ) .
+ a. pradhan and h. r. pandey , int .
d * 12 * , 941 ( 2003 ) .
+ a. pradhan , s. k. srivastava and k. r. jotania , czech . j. phys .
* 54 * , 255 ( 2004 ) .
+ a. pradhan and p. pandey , czech .
* 55 * , 749 ( 2005 ) .
+ a. pradhan and p. pandey , astrophys .
space sci . * 301 * , 127 ( 2006 ) .
+ a. pradhan , p. pandey and s. k. singh , int .
phys . * 46 * , 1584 ( 2007 ) .
+ a. pradhan , k. jotania and a. singh , braz .
* 38 * , 167 ( 2008 ) .
+ a. pradhan , v. rai and k. jotania , comm .
theor . phys . *
50 * , 279 ( 2008 ) .
m. abdel - rahaman , gen .
* 22 * , 655 ( 1990 ) 655 ; phys .
d * 45 * , 3492 ( 1992 ) .
r. g. vishwakarma , class .
* 17 * , 3833 ( 2000 ) .
g. f. r. ellis , _ general relativity and cosmology _ , ed .
r. k. sachs ( new york : academic press , 1971 ) .
s. r. roy and s. k. banerjee , astrophys .
space sci . * 167 * , 81 ( 1990 ) .
l. bianchi , _ lezioni sull a teoria deigruppi continui di transformazioni pisa _ : ( editiozi speern , 1918 ) .
j. d. barrow , mon . not .
211 * , 221 ( 1984 ) .
s. r. roy and j. p. singh , acta physica austriaca * 55 * , 57 ( 1983 ) .
r. tikekar and l. k. patel , pramana - j. phys .
* 42 * , 483 ( 1994 ) .
r. bali , r. banerjee and s. k. banerjee , astrophys .
space sci . * 317 * , 21 ( 2008 ) .
r. bali , a. pradhan and a. hassan , int .
. phys . * 47 * , 2594 ( 2008 ) .
a. pradhan and r. bali , ejtp * 19 * , 91 ( 2008 ) .
r. bali , r. banerjee and s. k. banerjee , ejtp * 21 * , 165 ( 2009 ) .
e. w. kolb and m. s. turner , _ the early universe _ ( addison - wesley ,
u s a , 1990 ) .
s. myung and b. m. cho , mod .
. lett . * a 1 * , 37 ( 1986 ) .
n. turok , phys .
* 60 * , 549 ( 1988 ) .
j. d. barrow , nucl .
* b 310 * , 243 ( 1988 ) . c. wolf , s .- afr .
* 14 * , 68 ( 1991 ) .
grn , astrophys .
space sci . * 173 * , 191 ( 1990 ) .
g. l. murphy , phys . rev . *
d 8 * , 4231 ( 1973 ) .
d. pavon , j. bafaluy and d. jou , class .
* 8 * , 357 ( 1996 ) .
r. maartens , class quantum gravit .
* 12 * , 1455 ( 1995 ) .
w. zimdahl , phys . rev . * d 53 * , 5483 ( 1996 ) . n. o. santos , r. s. dias and a. banerjee , j. math . phys . * 26 * , 878 ( 1985 ) .
u. a. belinskii and i. m. khalatnikov , sov .
jetp * 42 * , 205 ( 1976 ) .
s. myung , b. m. cho , mod .
lett . a * 1 * , 37 ( 1986 ) .
|
a new class of lrs bianchi type @xmath0 cosmological models with free gravitational fields and a variable cosmological term is investigated in presence of perfect fluid as well as bulk viscous fluid . to get the deterministic solution we have imposed the two different conditions over the free gravitational fields . in first case we consider the free gravitational field as magnetic type whereas in second case ` gravitational wrench ' of unit pitch "
is supposed to be present in free gravitational field .
the viscosity coefficient of bulk viscous fluid is assumed to be a power function of mass density .
the cosmological constant @xmath1 is found to be a decreasing function of time and positive which is corroborated by results from recent supernovae ia observations .
the physical and geometric aspects of the models are discussed .
+ + + + + + + + pacs : 98.80.cq , 98.80.-k +
keywords : cosmology .
bianchi models . gravitational fields
|
reconstruction of the photon statistics , @xmath0 , of optical states provides fundamental information on the nature of any optical field and finds relevant applications in foundations of quantum mechanics , quantum state engineering by postselection @xcite , quantum information @xcite , and quantum metrology .
indeed , detectors with the capability of counting photons @xcite are currently under investigation . among these ,
photomultiplier tubes ( pmt s ) @xcite and hybrid photodetectors @xcite are promising devices , though they have the drawback of a low quantum efficiency . on the other hand ,
solid state detectors with internal gain are still under development .
highly efficient thermal detectors have also been used as photon counters , though their operating conditions are still extreme to allow common use @xcite .
the advent of quantum tomography provided an alternative method to measure photon number distributions @xcite .
however , tomography needs the implementation of homodyne detection , which involves challenging mode matching , especially in the case of pulsed optical fields . in principle , in a photodetector each photon ionizes a single atom , and the resulting charge is amplified to produce a measurable pulse . in practice , however , available photodetectors are usually characterized by a quantum efficiency lower than unity , which means that only a fraction of the incoming photons leads to an electric pulse .
if the resulting current is proportional to the incoming photon flux we have a linear photodetector .
this is , for example , the case of the high - flux photodetectors used in homodyne detection . on the other hand photodetectors operating at very low intensities usually resort to avalanche process or very high amplification in order to transform a single ionization event into a recordable pulse . due to the gain instability in the process of amplification
it is generally difficult to discriminate the number of detected photons as far as this number becomes larger . as
a consequence the output of a photodetector consists of a current pulses whose charge statistics is a generalized bernoulli convolution of the actual photon distribution with the possibility to perform a discrimination between the number of detected photons only up to a finite number .
we define @xmath1 as the maximum number @xmath2 of detected photons that can be distinguished from @xmath3 , @xmath1 represent the counting capability of the detectors : @xmath4 detectors that works in geiger mode have @xmath5 because form the output is possible to discriminate from @xmath6 , one detected photon , to @xmath7 , dark , but is not possible to discriminate the output for @xmath8 from that for to @xmath6 .
the inversion of such a convolution may be performed in several ways , though , in general it is possible only when the quantum efficiency is larger than @xmath9 @xcite , and it is inherently inefficient , as it requires a large data sample to provide a reliable result . on the other hand , maximum - likelihood methods assisted by measurements taken at different quantum efficiencies have been proved to be both effective and statistically reliable @xcite . in particular , it is possible to obtain a method to reconstruct @xmath10 _ without any a priori information _ on the state of light under investigation .
an alternative approach is based on the so called _ photon chopping _
i.e. _ on a network of beam splitter followed by an array of on / off detectors .
photon chopping allows to send at most one photon on each detector , however , with the drawback of increasing the overall complexity of the detection scheme . in this paper , we extend previous analysis on maximum - likelihood on / off - based schemes @xcite , and present a detailed comparison among the performances of detectors characterized by realistic quantum efficiencies and different values of @xmath1 .
the comparison is made using real data and by an extensive set of monte carlo simulations , performed on different states of the radiation field including quantum and semiclassical states associated to single - peaked as well as multi - peaked photon distributions .
the reconstruction is obtained by maximum - likelihood methods assisted by measurements taken at different quantum efficiencies . in particular , as the statistics of the detected photons is linearly dependent on the photon distribution ( see below ) , the inversion can be well approximated by using the expectation - maximization algorithm ( em ) which leads to an effective and reliable iterative solution .
our results indicates that detectors with @xmath11 discriminating between zero , one and more than one photon , are generally enough to provide a reliable reconstruction of the photon statistics for single - peaked distributions , while detectors with higher value of @xmath1 do not lead to further improvements . on the other hand ,
multi - peaked distributions requires a higher quantum efficiency whereas for semiclassical states even on / off detectors are enough to provide a good reconstruction .
the paper is structured as follows . in the next section
we describe the statistics of the detected photons and illustrate the reconstruction algorithm . in section [
s : sim ] we report the results of a set of monte carlo simulated experiments performed on different kinds of signals , whereas in section [ s : exp ] we report experimental results for coherent signals .
section [ s : out ] closes the paper with some concluding remarks .
using a photodetector with quantum efficiency @xmath12 and @xmath13 , that corresponding to an unlimited counting capability , the probability of obtaining @xmath2 detected photons at the output is given by the convolution @xmath14 where @xmath15 is the actual photon statistics of the signal under investigation and @xmath16 if @xmath1 have a finite value only @xmath17 outcomes are possible , which occur with probabilities @xmath18 once the value of the quantum efficiency is known , eqs .
( [ kappa1 ] ) and ( [ kappa2 ] ) provide @xmath1 relations among the statistics of detected photons and the actual statistics of photons . at first sight
the statistics of a detector with low counting capability ( @xmath1 in the range @xmath19 ) appears to provide quite a scarce piece of information about the state under investigation .
however , if the statistics about @xmath20 is collected for a suitably large set of efficiency values , then the information is enough to reconstruct the whole photon distribution @xmath21 of the signal , upon a suitable truncation at @xmath22 of the hilbert space .
we adopt the following strategy : by placing in front of the detector @xmath23 filters with different transmissions , we may perform the detection with @xmath23 different values @xmath24 , @xmath25 , ranging from zero to a maximum value
@xmath26 equal to the nominal quantum efficiency of the detector . by denoting the probability of having @xmath2 detected photons in the experiment with quantum efficiency @xmath27 by @xmath28 we can rewrite eqs .
( [ kappa1 ] ) and ( [ convoluzio0 ] ) as @xmath29 where @xmath30 and @xmath31 .
let us now suppose that the @xmath21 s are negligible for @xmath32 and that the @xmath24 s are known , then @xmath33 eq . ( [ convoluzio ] ) may be rewritten a finite sum over @xmath34 from @xmath2 to @xmath22 .
overall , we obtain a finite linear system with @xmath35 equations in the @xmath22 unknowns @xmath36 .
unfortunately , the reconstruction of @xmath21 by matrix inversion can not be used in practice since it would require an unreasonable number of experimental runs to assure the necessary precision @xcite .
this problem can be circumvented by considering eqs .
( [ convoluzio ] ) as a statistical model for the parameters @xmath36 to be solved by maximum - likelihood ( ml ) estimation .
the likelihood functional is given by the global probability of the sample _
@xmath37 the ml estimates of @xmath36 are the values that maximizes @xmath38 . in eq .
( [ lik ] ) @xmath39 denotes the number of events `` @xmath2 detected photons '' obtained with quantum efficiency @xmath24 .
the maximization of @xmath38 under the conditions @xmath40 , @xmath41 , can be well approximated by using the expectation - maximization ( em ) algorithm @xcite , which leads to the iterative solution @xmath42}\:.\end{aligned}\ ] ] in eq .
( [ iterazio ] ) @xmath43 denotes the @xmath34-th element of reconstructed statistics at the @xmath44-th step , @xmath45 $ ] the theoretical probabilities as calculated from eq .
( [ convoluzio ] ) at the @xmath44-th step , whereas @xmath46 represents the frequency of the event `` @xmath2 detected photons '' with quantum efficiency @xmath24 , being @xmath47 the total number of runs performed with @xmath27 . in the following we assume equal @xmath47 , @xmath48 .
( [ iterazio ] ) provides a solution once an initial distribution is chosen . in the following
we will always consider an initial uniform distribution @xmath49 @xmath50 in the truncated fock space @xmath51 .
results from monte carlo simulated experiments shows that any distribution with @xmath52 @xmath50 performs equally well , _
i.e. _ the choice of the initial distribution does not alter the quality of reconstruction , though it may slightly affect the convergence properties of the algorithm .
in this section we report the photon statistics for different states of a single - mode radiation field as obtained by ml reconstruction from monte carlo simulated photodetection performed with the same low quantum efficiency and different resolution threshold .
we consider semiclassical states as well as highly nonclassical states and show results for different values of the maximal quantum efficiency @xmath53 .
@xmath54 [ cols="^,^,^ " , ] let us now illustrate our results about the accuracy of the reconstructions . as mentioned above we performed simulated experiments , on semiclassical states like coherent and thermal ones as well as highly nonclassical states such fock states , by using three values for the maximum quantum efficiency @xmath55 .
simulations have been performed by using @xmath56 data at each value of the quantum efficiency .
this value of @xmath47 is relatively small and well within the realm of quantum optical experiment . in a real experiment performed under the same conditions , @xmath47 of
this order would allow on - line reconstruction of the photon statistics .
simulations performed using different values of the parameters lead to similar conclusion . in fig .
[ fig : tanti ] we show the fidelity of reconstruction @xmath57 at the last iterations , @xmath58 , for different input signals and different resolution thresholds as a function of the average number of photons of the signal . as it is apparent from the plots , for low quantum efficiency and semiclassical states on / off detectors ( @xmath5 )
are enough to achieve a good reconstruction . on the other hand ,
the accuracy for nonclassical states largely improves for @xmath59 . for higher values of the quantum efficiency ( the middle and the right plots ) on / off
detectors becomes sufficient for a good reconstruction also for nonclassical states having single - peaked distributions .
notice that a higher resolution , however with a low value of @xmath60 , does not guarantee a good reconstruction . also notice that the results reported in fig .
[ fig : tanti ] has been obtained by stop the algorithm according to the prescriptions mentioned above .
we therefore do not expect that the fidelity is optimal for _ any _ state . indeed the fidelity @xmath58 in panel c5 ) and c8 ) of fig .
[ fig : tanti ] slightly decreases for high number of photons , though remaining close to unit value . the decreases in panels t2 ) , t5 ) and t8 ) of fig .
[ fig : tanti ] for greater mean values is due to the choice of a small dimension of the truncated hilbert space : by increasing n the trend disappear .
the low values of @xmath58 for the superposition states are due to the noise in @xmath61 , also for the high quantum efficiencies .
indeed , by increasing @xmath47 the reconstructions greatly improves ; besides , for high quantum efficiency the detectors perform good reconstructions also for @xmath62 , see the inset of panel s2 ) in fig .
[ fig : tanti ] .
the error bars in fig .
[ fig : tanti ] are standard deviations of @xmath58 as calculated from 40 different monte carlo runs .
they may appear large , but this is due to the high value of @xmath58 , which in turn determines the scale of the plot .
look at panels s) for comparison .
in order to confirm the monte carlo results for single - peaked distributions we have performed the reconstruction of the photon statistics of a coherent signal obtained from a nd : ylf laser .
the experimental data have been recorded with a hybrid photo detector , hamamatsu h8236 - 40 , placed on the second harmonics ( 523.5 nm ) of a cw mode - locked nd : ylf laser regeneratively amplified at a repetition rate of 5 khz ( high q laser ) .
the frequencies @xmath63 , until @xmath64 , have been extrapolated from the response of the detector with the following procedure . at first a gaussian best - fit of the response peaks at 0, .. ,4 photoelectrons has been performed @xcite , then the the frequencies has been obtained by choosing three thresholds , whose optimal choice turns out to be the mean point between photoelectron peaks@xcite .
after the frequencies @xmath63 have been obtained we used the algorithm of the previous section .
the results of the reconstruction are shown in fig .
[ fig : sperim ] , panel a ) .
the reconstructed distribution at the last iteration @xmath65 have been then compared with a poissonian best - fit . as it is apparent from fig .
[ fig : sperim ] , the agreement between @xmath65 and the best - fit is very good .
in addition , we notice that the shape of @xmath57 and @xmath66 are very similar to those coming from simulation and that the @xmath67 obtained by the experimental data with m=1 is close to the one obtained with m=2 and m=3 , as predicted by simulations .
@xmath54 and coherent best fit , @xmath68 ; b ) : total absolute error @xmath66 ; c ) : fidelity @xmath57 calculated _ a posteriori_. the reconstruction is performed with a hybrid photo detector operated by taking @xmath5 ( on / off detector ) , @xmath11 and @xmath69 ; @xmath70 different quantum efficiencies @xmath27 distributed in @xmath71 $ ] , @xmath72 ; the hilbert space is truncated at @xmath73 ; @xmath56 number of runs have been performed for each @xmath12 ; the algorithm is stopped at iteration @xmath74.,title="fig:",scaledwidth=25.0% ] and coherent best fit , @xmath68 ; b ) : total absolute error @xmath66 ; c ) : fidelity @xmath57 calculated _ a posteriori_. the reconstruction is performed with a hybrid photo detector operated by taking @xmath5 ( on / off detector ) , @xmath11 and @xmath69 ; @xmath70 different quantum efficiencies @xmath27 distributed in @xmath71 $ ] , @xmath72 ; the hilbert space is truncated at @xmath73 ; @xmath56 number of runs have been performed for each @xmath12 ; the algorithm is stopped at iteration @xmath74.,title="fig:",scaledwidth=25.0% ] and coherent best fit , @xmath68 ; b ) : total absolute error @xmath66 ; c ) : fidelity @xmath57 calculated _ a posteriori_. the reconstruction is performed with a hybrid photo detector operated by taking @xmath5 ( on / off detector ) , @xmath11 and @xmath69 ; @xmath70 different quantum efficiencies @xmath27 distributed in @xmath71 $ ] , @xmath72 ; the hilbert space is truncated at @xmath73 ; @xmath56 number of runs have been performed for each @xmath12 ; the algorithm is stopped at iteration @xmath74.,title="fig:",scaledwidth=25.0% ]
we have compared the reconstruction of the photon statistics as obtained applying ml algorithm to data coming from detectors with different counting capability and different quantum efficiencies . we found that using ml methods , detectors with high quantum efficiency does not need to have high counting capability , since on / off detection assisted by ml methods
already provides good state reconstructions . on the other hand ,
a small quantum efficiency makes the counting capability a crucial parameter .
overall , our results indicate that development of future photodetectors may be focused on increasing the quantum efficiency rather than the counting capability .
the authors are grateful to alessandra andreoni for encouragement and support .
they also thank maria bondani , fabio ferri , zdenek hradil and jarda ehek for useful discussions .
this work has been partially supported by miur through the projects prin-2005024254 - 002 and firb rbau014clc-002
. 99 e. knill , r. laflamme and g.j .
milburn , nature ( london ) m. dakna , j. clausen , l. knoll , d .-
welsch , phys .
a * 59 * , 1658 ( 1999 ) ; m. g. a. paris , phys .
a. * 62 * , 033813 ( 2000 ) ; m. g. a. paris , m. cola , r. bonifacio , phys .
a * 67 * , 042104 ( 2003 ) ; j. laurat et al .
lett . * 91 * , 213601 ( 2003 ) ; s. a. babichev , b. brezger , a. i. lvovsky , phys . rev . lett . * 92 * , 047903 ( 2004 ) .
e. knill , r. laflamme and g.j .
milburn , nature ( london ) * 409 * , 46 ( 2001 ) d. rosenberg , a.e .
lita , a.j .
miller , s.w .
nam , phys .
a * 71 * , 61803 ( 2005 ) .
r. desalvo , nucl . instr . and meth . a 387 ( 1997 ) 92 - 96 .
g. zambra , m. bondani , a. s. spinelli , a. andreoni , rev .
instrum . * 75 * , 2762 ( 2004 ) .
e. hergert , single photon detector workshop , gaithersburg , nist ( 2003 ) .
j. kim , s. takeuchi , y. yamamoto , and h.h .
hogue , appl .
phys . lett . *
74 * , 902 ( 1999 ) ; a. peacock , p. verhoeve , n. rando , a. van dordrecht , b. g. taylor , c. erd , m. a. c. perryman , r. venn , j. howlett , d. j. goldie , j. lumley , and m. wallis , nature * 381 * , 135 ( 1996 ) . g. di giuseppe , a. v. sergienko , b. e. a. saleh , and m. c. teich in _ quantum information and computation _ , e. donkor , a. r. pirich , and h. e. brandt eds .
, proceedings of the spie * 5105 * , 39 ( 2003 ) . m. munroe , d. boggavarapu , m. e. anderson , and m. g. raymer , phys . rev .
a * 52 * , r924 ( 1995 )
; y. zhang , k. kasai , and m. watanabe , opt . lett . * 27 * , 1244 ( 2002 ) .
t. kiss , u. herzog , u. leonhardt , phys .
a * 52 * , 2433 ( 1995 ) ; g. m. dariano , c. macchiavello , phys .
a 57 , 3131 ( 1998 ) ; t. kiss , u. herzog , and u. leonhardt , phys . rev . a 57 , 3134 ( 1998 ) .
d. mogilevtsev , opt .
comm * 156 * , 307 ( 1998 ) ; acta phys . slov . *
49 * , 743 ( 1999 ) .
a. r. rossi , s. olivares , and m. g. a. paris , phys .
a * 70 * , 055801 ( 2004 ) .
g. zambra et al , phys .
* 95 * , 063602 ( 2005 ) .
j. @xmath75eh@xmath76ek , z. hradil , o. haderka , j. pe@xmath77ina , jr . , and m. hamar , phys . rev .
a * 67 * , 061801(r ) ( 2003 ) ; o. haderka , m. hamar , j. pe@xmath78ina , eur .
d * 28 * , ( 2004 ) .
h. paul at al .
lett . * 76 * , 24642467 ( 1996 ) a.p .
dempster , n.m .
laird , d.b .
rubin , j. r. statist .
b * 39 * , 1 ( 1977 ) ; y. vardi and d. lee , j. r. statist .
b * 55 * , 569 ( 1993 ) .
r. a. boyles , j. r. statist .
b * 45 * , 47 ( 1983 ) ; c. f. j. wu , the annals of statistics * 11 * , 95 ( 1983 ) .
k. banaszek , acta phys . slov . * 48 * , 185 ( 1998 ) ; phys . rev .
a * 57 * , 5013 ( 1998 ) .
m. bondani , _ self consistent characterization of light statistics _ ,
manuscript in preparation .
|
the output of a photodetector consists of a current pulse whose charge has the statistical distribution of the actual photon numbers convolved with a bernoulli distribution .
photodetectors are characterized by a nonunit quantum efficiency , _
i.e. _ not all the photons lead to a charge , and by a finite resolution , _
i.e. _ a different number of detected photons leads to a discriminable values of the charge only up to a maximum value .
we present a detailed comparison , based on monte carlo simulated experiments and real data , among the performances of detectors with different upper limits of counting capability . in our scheme
the inversion of bernoulli convolution is performed by maximum - likelihood methods assisted by measurements taken at different quantum efficiencies .
we show that detectors that are only able to discriminate between zero , one and more than one detected photons are generally enough to provide a reliable reconstruction of the photon statistics for single - peaked distributions , while detectors with higher resolution limits do not lead to further improvements .
in addition , we demonstrate that , for semiclassical states , even on / off detectors are enough to provide a good reconstruction .
finally , we show that a reliable reconstruction of multi - peaked distributions requires either higher quantum efficiency or better capability in discriminating high number of detected photons .
|
almost @xmath3 years after the discovery of the organic superconductors @xcite , the problem of determination of their pairing state has not yet found a definite solution .
the critical magnetic field @xmath4 exceeds the pauli paramagnetic limit and indicates that the pairing symmetry is triplet @xcite .
the knight shift does not change between the normal and superconducting states and it is a signature for triplet pairing state @xcite . however the absence of hebel - slichter peak and the power - low decay of @xmath5 below @xmath6 @xcite is an indication of the presence of nodes of the pair potential on the fermi surface .
the scattering theory can be used to distinguish the symmetry of the pair potential @xcite . in @xmath1-wave superconductors the pair potential changes sign under a @xmath7-rotation .
so under appropriate orientation of the @xmath8-axis within the @xmath9 plane of @xmath1-wave superconductor the transmitted quasiparticles feel different sign of the pair potential .
this results in the formation of bound states within the energy gap , which are detected as zero energy peaks ( zep ) in the conductance spectra @xcite .
the scattering theory has also been used for the determination of the pairing symmetry in ( tmtsf)@xmath10x by sengupta _
however their calculation is restricted to the presence or absence of zep at the surface , which can not distinguish the @xmath0 from the @xmath2-wave case . in a more realistic calculation tanuma _
@xcite used the extended hubbard model on a quasi - one - dimensional lattice at quarter - filling , to study the quasiparticle states near the surface of a quasi - one dimensional organic superconductor , where the pairing symmetry can actually be distinguished from the overall line shape of the surface density of states ( sdos ) and the presence or absence of zep . in this paper
we extend the btk formula to calculate the tunneling conductance in a normal - metal / insulator / quasi - one - dimensional organic superconductor , where the structure of the fermi surface is taken into account .
the pairing symmetry of the superconductor is assumed to be triplet @xmath0 , singled @xmath1 , @xmath11 , and triplet @xmath2-wave .
in particular , we find that the zep appears in the electron tunneling along the @xmath8 or @xmath12 axis , when the transmitted quasiparticles experience different sign of the pair potential .
also the line shape of the spectra is sensitive to the presence or absence of nodes of the pair potential on the fermi surface .
the present calculation is useful for ( tmtsf)@xmath10pf@xmath13 , although this salt needs pressure to show superconductivity , but is not very suitable for ( tmtsf)@xmath10clo@xmath14 , which is the only material exhibiting superconductivity in the bechgaard salts at ambient pressure , but its unit cell is doubled due to anion ordering @xcite . also tunneling spectroscopy on the quasi two - dimensional organic superconductor @xmath15-(bedt - ttf)@xmath10cu(ncs)@xmath10
has been performed by using scanning tunneling microscopy @xcite . in this compound
the conducting cu(ncs)@xmath10 layers run parallel to the @xmath16 plane , along the @xmath8-axis , and the in - plane tunneling data strongly suggest @xmath1-wave pairing state .
these features can be used to distinguish the pairing symmetry in quasi - one - dimensional organic superconductor .
the motion of quasiparticles in inhomogeneous superconductors is described by the four component bogoliubov de gennes ( bdg ) equations .
the bdg equations are decoupled into two sets of ( two component ) equations , one for the spin up electron , spin down hole quasiparticle wave functions ( @xmath17 ) and the other for ( @xmath18 ) . the bdg equation for spin index @xmath19 or @xmath20 ,
read @xcite @xmath21 the single - particle hamiltonian is given by @xmath22 , @xmath23 is the external magnetic field , @xmath24 is the energy measured from the fermi energy @xmath25 . for a given spin projection @xmath26 the magnetic field shifts the energy by @xmath27 .
@xmath28 is the pair potential , after a transformation from the position coordinates @xmath29 to the center of mass coordinate @xmath30 and the relative vector @xmath31 .
after fourier transformation the pair potential depends on the relative wave vector @xmath32 and @xmath33 . in the weak coupling limit @xmath32
is fixed on the fermi surface ( @xmath34 ) , and only its direction is variable . after applying the quasi - classical approximation , i.e. , @xcite
@xmath35 so that the fast oscillating part , of the wave function is divided out , the bdg equations are reduced to the andreev equations @xcite @xmath36 where the quantities @xmath37 and @xmath38 are electron - like and hole - like quasiparticles with spin index @xmath26 , and @xmath39 respectively , and @xmath40 is the fermi velocity .
we consider the normal - metal / insulator / superconductor junction shown in fig .
[ fig1.fig ] .
the electron momentum parallel to the interface @xmath41 is conserved . for the interface that is normal to the @xmath8-axis
, the insulator is modelled by a delta function , located at @xmath42 , of the form @xmath43 .
the temperature is fixed to @xmath44 k. we take the pair potential as a step function , i.e. , @xmath45 . for the geometry shown in fig .
[ fig1.fig ] , eqs .
[ ndrv ] take the form @xmath46 where @xmath60 , @xmath61}=\frac{2 z_0}{k_{fx } } $ ] .
the bcs coherence factors are given by @xmath62/2,\ ] ] @xmath63/2,\ ] ] and @xmath64 .
the internal phase coming from the energy gap is given by @xmath65 $ ] , where @xmath66 ( @xmath67 ) , is the pair potential experienced by the transmitted electron - like ( hole - like ) quasiparticle .
for the spin triplet pairing state the cooper pairs have spin @xmath74 degree of freedom .
the gap function is a @xmath75 symmetric matrix which in the spin space can be written as @xmath76 where @xmath77 denotes the pauli matrices and @xmath78 is a vector which defines the axis along which the cooper pairs have zero spin projection . in the following
we will take @xmath79 , i.e. , parallel to the chains . in that case @xmath80 , while @xmath81 .
the fermi surface ( fs ) consists of two branches and is open in the @xmath82 direction , as shown in fig .
[ fs.fig ] .
we consider the following pairing cases \a ) in case of @xmath83-wave , @xmath84-wave superconductor @xmath85 @xmath86 changes its sign along the fs in the @xmath87 direction while the @xmath88 changes its sign along the fs in the @xmath82 direction as seen in figs .
[ fs.fig](a ) and [ fs.fig](d ) .
\b ) in case of @xmath89-wave , @xmath90-wave superconductor @xmath91 @xmath92 changes its sign along the fs in the @xmath87 direction while the @xmath93 changes its sign along the fs in the @xmath82 direction as seen in figs .
[ fs.fig](b ) and [ fs.fig](e ) .
\c ) in case of @xmath94-wave , @xmath95-wave superconductor @xmath96 @xmath97 changes its sign along the fs in the @xmath87 direction while the @xmath98 changes its sign along the fs in the @xmath82 direction as seen in figs .
[ fs.fig](c ) and [ fs.fig](f ) .
\d ) in case of @xmath11-wave , @xmath99-wave superconductor @xmath100 @xmath101 , @xmath102 change sign along the fs as seen in figs .
[ fsdxy.fig](a ) and [ fsdxy.fig](b ) .
in figs . [ p.fig]-[dxy.fig ] we plot the tunneling conductance @xmath103 as a function of @xmath104 for various values of @xmath105 , for different orientations normal to the @xmath8 or @xmath12 axis .
the pairing symmetry of the superconductor is @xmath83-wave , @xmath84-wave , in fig .
[ p.fig ] , @xmath89-wave , @xmath90-wave , in fig .
[ d.fig ] , @xmath94-wave , @xmath95-wave , in fig .
[ f.fig ] , @xmath11-wave , @xmath99-wave , in fig .
[ dxy.fig ] .
the conductance peak is formed in the electron tunneling along the @xmath8 or @xmath12 axis , when the transmitted quasiparticles experience different sign of the pair potential .
also the line shape of the spectra is sensitive to the presence or absence of nodes of the pair potential on the fermi surface . for the @xmath83-wave case , when the @xmath8-axis of the crystal is at right angle to the interface , the horizontal line in fig .
[ fs.fig ] ( a ) , representing the scattering process , connects points of the fs with opposite sign of the pair potential for @xmath106 . as a result a peak exists in the conductance spectra , at @xmath107 as seen in fig .
[ p.fig ] ( a ) for @xmath108 .
the height of the zep is proportional to the range of @xmath82 for which sign change occurs and it is expected to have its maximum value when the @xmath8-axis of the crystal is normal to the interface , since for this orientation the transmitted quasiparticles feel a different sign of the pair potential for all @xmath106 . on the other hand , when the @xmath8 axis is along the interface , then the vertical line in fig .
[ fs.fig ] ( a ) connects points of the fs with the same sign of the pair potential for @xmath109 , i.e. , there is no @xmath87 for which , the transmitted quasiparticles feel the sign change of the pair potential , and no zep is formed as seen in fig .
[ p.fig ] ( b ) . for
the @xmath83-wave case the nodes of the pair potential do not intersect the fs and the line shape of the spectra is u - like as in the case of the @xmath26-wave superconductor .
the situation is opposite in the @xmath84-wave case , where the scattering process for the surface orientation normal to the @xmath8 ( @xmath12)-axis , described by the horizontal ( vertical ) line in fig .
[ fs.fig ] ( d ) connects points of the fs with the same ( opposite ) sign . as a consequence for the surface orientation normal to the @xmath8-axis
, there is no @xmath82 for which the transmitted quasiparticles feel the sign change of the order parameter , and no zep is formed , as seen in fig .
[ p.fig ] ( c ) , while for the interface along the @xmath8-axis the transmitted quasiparticles feel the different sign of the pair potential , for all @xmath110 , and a zep exists as seen in fig .
[ p.fig ] ( d ) . for the @xmath84-wave case the nodes of the pair potential intersect the fs and
the spectra has a v - shaped form as in the case of @xmath1-wave superconductors . for the @xmath1-wave case , and for tunneling along the @xmath8 or @xmath12-axis , the scattering process , described with the horizontal or vertical line in figs .
[ fs.fig](b ) and [ fs.fig](e ) connects points of the fs with the same sign in all cases .
this means that the pair potential does not change sign and no zep is formed as seen in fig .
[ d.fig ] . also due to the presence of nodes of the pair potential along the fs ,
the line shape of the spectra is v - like . for the @xmath2-wave case , although the pair potential has a different structure , the lines in figs . [
fs.fig](c ) and [ fs.fig](f ) connect points of the fs with the same sign - change of the pair potential as in the @xmath0-wave case . as a result
the tunneling shows zep as in the @xmath0-wave case , due to the sign change of the transmitted quasiparticles .
however , in all cases in the @xmath2-wave case the pair potential intersects the fs and nodes are formed . as a consequence
the line shape of the spectra is v - like , as seen in fig .
[ f.fig ] unlike to @xmath83-wave case , where the pair potential is nodeless and the tunneling spectra has a u - shaped form .
the conclusion is that the line shape of the spectra can be used to distinguish the @xmath83 from the @xmath2-wave pairing state .
our results are comparable to that of tanuma _ et al . _
@xcite although their calculation was based on a different model , i.e. , the extended hubbard model on a quasi - one dimensional lattice at quarter - filling . for the @xmath11-wave ( @xmath99-wave ) case , and for tunneling along the @xmath8 or @xmath12-axis , the scattering process , described with the horizontal or vertical line in figs .
[ fsdxy.fig](a ) and [ fsdxy.fig](b ) connects points of the fs with different ( the same ) sign .
this means that the pair potential felt by the transmitted quasiparticles changes ( conserves ) sign during the scattering process and a zep ( no zep ) is formed at @xmath107 for the @xmath11-wave ( @xmath99-wave ) case as seen in fig .
[ dxy.fig ] . also due to the presence of nodes of the pair potential along the fs , the line shape of the spectra is v - like .
in addition the tunneling spectra for the @xmath11-wave , for tunneling direction along the @xmath8 or @xmath12 axis , is equivalent to the tunneling spectra for @xmath99-wave with surface orientation tilted by @xmath111 , i.e. , along the nodal direction . in the metallic limit ( @xmath112 )
, @xmath103 has the same form for each pairing potential independently from the orientation . in table
[ tables ] we summarize the results concerning the presence or absence of zep , and also the overall line shape of the tunneling spectra for various pairing potentials .
these features are explained if we calculate the energy of the midgap state , which is given for large @xmath105 by the value in which the denominator of eqs .
( [ ra ] ) and ( [ rb ] ) vanishes .
the equation giving the energy peak level is written as @xcite @xmath113 in the @xmath83-wave case , for surface orientation normal to @xmath8-axis , this equation has the solution @xmath107 , for @xmath114 , since @xmath115 , and also the transmitted quasiparticles feel a different sign of the pair potential , i.e. , @xmath116 .
when a midgap state exists the tunneling conductance @xmath117 is equal to @xmath118 and the peek in @xmath103 seen in fig .
[ p.fig](a ) , is due to the normal state conductance @xmath119 in eq .
[ sq ] , which depends inversely on the @xmath120 for large @xmath105 . for surface orientation normal to @xmath12-axis the range of @xmath87 for which eq . [ midgap ] has solutions collapses to zero , and
no bound states are formed .
then @xmath103 goes to zero as @xmath121 and there is no conductance peak . in the @xmath84-wave case for surface orientation normal to @xmath8-axis eq .
[ midgap ] has no solutions because @xmath122 , while for surface orientation normal to @xmath12-axis a bound state is formed for all @xmath87 in the interval @xmath110 . for the @xmath1-wave case and for tunneling along the @xmath8 or @xmath12 axis of the crystal ,
the condition @xmath116 is not satisfied and no bound states are formed . for the @xmath2-wave case
that condition is the same as in the @xmath0-wave case , and thus the zep is formed as in the @xmath0-wave pairing state . for the @xmath11-wave ( @xmath99-wave ) case
that condition @xmath116 is ( is not satisfyied for tunneling along the @xmath8 and @xmath12 axis and hence the zep ( no zep ) is formed for both tunneling directions .
in this section we describe the effect of the external magnetic field @xmath23 in the direction parallel to the chains , in the spectra .
the effect of the magnetic field depends on the spin , of the quasiparticles .
the tunneling conductance is given by @xmath123 in fig .
[ h.fig ] the tunneling conductance @xmath103 is plotted for a fixed magnetic field @xmath124 , and barrier strength @xmath108 .
the pairing symmetry of the superconductor is @xmath0 , @xmath1 , @xmath2 , @xmath11 respectively .
the magnetic field splits symmetrically the tunneling spectrum which is a linear superposition of the spectra for spin up(down ) quasiparticles .
the amplitude of the splitting depends linearly on the magnetic field @xmath23 . for the case of @xmath124 , seen in fig .
[ h.fig ] the spin up(down ) part of the spectra partially overlaps while for larger values of the magnetic field the spin up and down branches are well separated . in the latter case the right(left ) branch of
the spectra originates from spin up(down ) quasiparticle spectra @xmath125 .
the condition for the formation of bound states is slightly modified under the presence of a magnetic field to @xmath126 , for the spin - up region , and @xmath127 , for the spin - down , from the corresponding @xmath107 in the absence of any field
. so the multiplication of the bound states and the presence of magnetic field results into the appearance of double peak in the conductance spectra .
we calculated the tunneling conductance in normal - metal / insulator / quasi - one - dimensional superconductor , using the btk formalism .
we showed that the zep appears in the electron tunneling along the @xmath8 or @xmath12 axis , when the transmitted quasiparticles experience a different sign of the pair potential .
also the line shape of the spectra is sensitive to the presence or absence of nodes of the pair potential on the fermi surface , and results to a u - shaped structure for the @xmath83-wave case and to a v - shaped one for the @xmath1 and @xmath2-wave cases .
the zep are due to the formation of bound states within the gap .
the calculation of the conductance @xmath128 , for which bound state occurs , shows an enhancement at the bound state energy .
the effect of the magnetic field in the direction parallel to the chains is to split linearly the zep . throughout this paper the spatial variation of the order parameter near the surface , which depends on the boundary orientation ,
is ignored for simplicity . however , since the features presented here are intrinsic which are generated by the existence of surface bound states , the essential results do not change qualitatively .
also we assumed perfectly flat interfaces in the clean limit , so any impurity scattering and the effect of the surface roughness are ignored . .
presence ( yes ) or absence ( no ) of a zero - bias conductance peak in electron tunneling along the @xmath8 and @xmath12 axes for different symmetries of the pair potential .
also the overall line shape of the tunneling spectra is shown [ cols="^,^,^,^",options="header " , ]
|
the tunneling conductance spectra of a normal - metal / insulator / quasi - one - dimensional superconductor is calculated by using the blonder - tinkham - klapwijk formulation .
the pairing symmetry of the superconductor is assumed to be @xmath0 , @xmath1 , and @xmath2-wave .
it is found that there is a well defined zero energy peak in electron tunneling along the direction parallel to the chains or normal to these when the transmitted quasiparticles feel different sign of the pair potential .
the actual line shape of the spectra is sensitive to the nodes of the pair potential on the fermi surface .
key words : tunneling spectra , organic superconductors , andreev reflection , zero energy peak 2
|
this series of articles presents an improved solver for a challenging problem , the unified solution of compressible and incompressible fluids .
after harlow and amsden proposed the ice algorithm as a fully implicit solver for fluid equations in a conservative form @xcite , some approaches for this purpose have been investigated , such as the pressure - based semi - implicit algorithms @xcite , approaches based on the asymptotic expansion with respect to the local mach number @xcite , and the boundary condition capturing method that treats compressible and incompressible materials separately @xcite .
( see also recent reviews @xcite for more details . ) among them , the cip - cup method , a pressure - based semi - implicit algorithm proposed by yabe and wang @xcite , has already been applied to many practical multi - material problems such as the laser machining of a metal plate @xcite , comet shoemaker - levy 9 s collision with the planet jupiter @xcite , the interaction of a shock wave and a liquid drop @xcite , and the milk - crown formation on a liquid surface @xcite , and has been proven to be an efficient and robust solver for the unified solution @xcite . in this method ,
the euler ( or the navier - stokes ) equations for fluid flows in a non - conservative form are selected as the governing equations , and the convection terms in the equations are solved explicitly by the cip method @xcite , while the acoustic terms are solved implicitly by the cup method @xcite . in part
i of this series @xcite , an improved solver for the convection terms was constructed using both an interpolation and an extrapolation function to describe the spatial profile of the density of materials . in the improved method ,
the cubic interpolation function used in the cip scheme is replaced with the quadratic extrapolation functions constructed under some constraints for guaranteeing stability , only in a cell containing an interface between materials .
this method allows us to solve the density interface with no dissipation across the interface , and is applicable to compressible flow problems , i.e. , to variable - density problems . as part ii of this series
, this paper presents some further improvements for the cip - cup method , and also gives some application results . in sec .
[ sec3 ] , we attempt to incorporate the concept of the multi - time - step ( mts ) integration techniques into the cip - cup algorithm . the mts integration techniques , which allow us the efficient integration of dynamical systems involving some different time scales , have been used to solve a variety of problems , such as gravitational @xmath0-body problems @xcite , molecular dynamics @xcite , and atmospheric flow problems @xcite .
this technique reduces the computational effort by integrating only rapidly varying components of a system ( e.g. , strongly accelerated particles or waves having fast phase speeds ) with a small time interval , and others with a larger time interval . in the present study , based on this concept , different - nature terms in the fluid equations (
convection , acoustic , and other terms ) are solved using different time intervals .
this modification allows us a more flexible determination of the time interval , and improves the accuracy and efficiency of the cip - cup . as will be shown , the adaptation of the mts concept to the cip - cup integration is straightforward , because the cip - cup separately treats the terms having different natures , i.e. , those having different characteristic time scales .
section [ sec4 ] presents a modified scheme to update the spatial derivatives during the computational steps except for the convection parts .
as was shown in part i , the spatial derivatives of the dependent variables are used explicitly as additional dependent variables to solve the convection terms . in the convection process
, the derivatives are updated to obey the equations derived by taking derivatives of the governing equations @xcite .
those spatial derivatives should be updated by some way also in the non - convection processes . in refs .
@xcite , simple methods for this purpose have been proposed using classical centered differences . as was discussed in part
i , however , the use of such a classical discretization gives rise to the numerical dispersion and dissipation around the interfaces at which the spatial derivatives of the dependent variables tend to be discontinuous .
we modify the centered scheme by adopting a simple extrapolation . in sec .
[ sec5 ] , the applicability of the present methods to the compressible navier - stokes equations with a surface tension term is investigated , and in the appendixes , some additional discussions regarding the averaging of the density ( needed to solve the terms other than the convection ) and the treatment of the surface - tension term are given .
section [ sec6 ] presents some application results for the compressible and incompressible kelvin - helmholtz instabilities and the multibubble dynamics in an acoustic field , and sec .
[ sec7 ] presents concluding remarks .
the cip - cup method @xcite is a semi - implicit solver for the euler equations of fluid flows : @xmath1 @xmath2 @xmath3 where @xmath4 , @xmath5 , @xmath6 , and @xmath7 denote the density , the velocity vector , the pressure , and the local sound speed , respectively , and * f * may contain the viscosity , the surface tension , and external forces .
this method separately solves the terms in eqs .
( [ eq1])([eq3 ] ) of different natures by a time splitting technique .
the convection parts of these equations , @xmath8 @xmath9 @xmath10 are solved by the cip method @xcite ( in the present work , of course , the hybrid interpolation - extrapolation method proposed in part i is adapted ) , while the acoustic parts , @xmath11 @xmath12 @xmath13 are solved by the cup method , which is an implicit finite difference method . the remaining part , @xmath14 may be solved by some existing methods , such as the finite difference or finite volume method .
( we refer to eq .
( [ eq7 ] ) as the `` additional part '' . ) solving these parts successively completes one step of the cip - cup time integration . in the following
, the cup method for the acoustic parts is concretely reviewed .
discretizing the time derivatives on the lhs of eqs .
( [ eq4])([eq6 ] ) and estimating the spatial derivatives on the rhs with future values , one obtains @xmath15 @xmath16 @xmath17 where @xmath18 is the number of the time step , @xmath19 is the time interval , and the quantities with the superscript @xmath20 indicate the values after solving the parts other than the acoustic parts .
( the acoustic parts should be solved at the final stage of a time step , because , as is shown below , solving these parts enforces the divergence - free condition for an incompressible fluid . ) taking divergence of eq .
( [ eq9 ] ) and substituting it into eq .
( [ eq10 ] ) yield the following pressure equation : @xmath21 ( substituting eq .
( [ eq6 ] ) into this and rewriting as @xmath22 finds that eq .
( [ eq11 ] ) corresponds to a first - order approximation of the wave equation in terms of the pressure . )
almost the same pressure equation is given in ref .
@xcite , but discretization is done by a finite element technique . after solving eq .
( [ eq11 ] ) to get @xmath23 , the velocity is updated explicitly by eq .
( [ eq9 ] ) .
also , the density is updated explicitly by @xmath24 given by eqs .
( [ eq8 ] ) and ( [ eq10 ] ) .
generally , the spatial discretization for the above equations is performed using the 2nd - order centered finite differencing on the staggered grids .
when the sound speed is infinite , the pressure equation ( [ eq11 ] ) is reduced to an elliptic equation , @xmath25 which corresponds to that used in the smac algorithm @xcite for incompressible flows .
this result reveals that the cup method is applicable to both compressible and incompressible fluids . in the case where the cip or its variant is used to solve the convection parts
, we also need to update the spatial derivatives , used explicitly in those schemes , in the non - convection processes .
simple schemes for this purpose have already been proposed @xcite . for a two - dimensional case ,
the scheme is represented as @xmath26 @xmath27 where @xmath28 indicates an dependent variable , @xmath29 , @xmath30 , the subscripts @xmath31 and @xmath32 indicate @xmath33 , and @xmath34 is the grid spacing ( assumed to be constant for simplicity ) . in the case where the cross derivative @xmath35 is used in the convection process @xcite , the following equation is additionally used : @xmath36 equations ( [ eq14])([eq16 ] ) can be solved explicitly using the quantities obtained before and after solving the non - convection parts . in 1994
@xcite , ito proposed an improved variant of the pressure equation [ eq . ( [ eq11 ] ) ] by incorporating the concept of the exponential method for a heat - conduction equation @xcite .
the 1-d formula of the variant is represented by @xmath37 - \rho ^ * c_s^{*\,2 } u_x^ * , \ ] ] where @xmath38 is the weighting factor determined theoretically as @xmath39 @xmath40 and @xmath41 is the density at @xmath42 , determined approximately with @xmath43 and @xmath44 ( see appendix a ) . for @xmath45 ( i.e. , @xmath46 )
, @xmath38 converges to 1 and eq .
( [ eq17 ] ) is reduced to the elliptic equation ( [ eq13 ] ) ; therefore , this variant is applicable to incompressible flows . for @xmath47 , @xmath38 becomes @xmath48 , resulting in a higher resolution than that of the conventional cup @xcite .
we adopted this variant in the present study .
the 2-d formula of this variant ( not shown in ito s paper ) may be represented by @xmath49 where @xmath50 @xmath51 @xmath52 @xmath53 we simplify this as follows to reduce the computational efforts : @xmath54 \nonumber \\ & & - \rho ^ * c_s^{*\,2 } \nabla \cdot { \bf u}^ * , \end{aligned}\ ] ] where @xmath55 this formula requires only one weighting factor .
this simplification may be valid because @xmath56 generally . in ref .
@xcite , an alternative variant was proposed .
though this variant can provide higher resolution of the sound wave than that obtained by ito s variant , it is likely that the use of this variant as a part of the solver for fluid flows sometimes causes great violation of mass conservation @xcite .
we therefore did not adapt it .
in many fluid flow problems , the time scale of a significant phenomenon is determined by the flow velocity ( i.e. , the characteristic speed of the convection parts ) , rather than by the sound speed or other characteristic speed . in certain cases , however , the acoustic parts ( or others ) have much smaller time scales ( i.e. , much greater characteristic speeds ) , resulting in an excessive restriction on the time interval , @xmath19 . the cip - cup overcomes this problem by solving the acoustic parts implicitly .
such an approach , however , guarantees only the stability , and probably provides inaccurate results in certain situations .
for example , under an intermediate condition of flows such as weakly compressible flows , the acoustic parts may need to be solved by using a time interval of sufficiently small @xmath57 that makes @xmath58 very small , since the compressibility described by the acoustic parts plays an important role , where @xmath59 to achieve an efficient computation even in such a situation , we adopt the multi - time - step ( mts ) integration technique to the cip - cup method .
the adaptation of the mts integration to the cip - cup is straightforward because the components of different time scales ( convection , sound propagation , and others ) are treated separately .
the time propagation performed in the cip - cup can be represented schematically as follows : @xmath60 @xmath61 @xmath62 where @xmath63 and the operators @xmath64 , @xmath65 , and @xmath66 indicate the discrete propagators for the corresponding integration steps ( the convection , the additional , and the acoustic steps , respectively ) .
these equations can be summarized as @xmath67 if we factor @xmath65 and @xmath66 into some identical pieces , we get @xmath68^{m3 } [ l2(\delta t / m2)]^{m2 } l1(\delta t)\,{\bf u}^n , \ ] ] where @xmath69 and @xmath70 are positive integers .
this equation means that , after solving the convection parts with @xmath19 , the additional part is solved @xmath69 times with @xmath71 , and subsequently the acoustic parts are solved @xmath70 times with @xmath72 . if necessary
, we divide @xmath73 into two parts as @xmath74 , and modify eq .
( [ eq23 ] ) as @xmath75^{m3 } \,[l2^{(1 ) } ( \delta t / m2)]^{m2 } l1(\delta t)\,{\bf u}^n\ ] ] or @xmath76^{m2 } ] ^{m3 } l1(\delta t)\,{\bf u}^n .
\nonumber \\ { } \end{aligned}\ ] ] such a treatment of the additional parts is necessary , e.g. , when the surface - tension term exists .
( the surface - tension part , represented by @xmath77 , should be solved together with the acoustic parts because the surface tension always needs to balance with the pressure jump at the interface . )
some other forms of factorization adjusted for a governing system can be employed .
we determine the fundamental time interval , @xmath19 , by @xmath78 where the parameters are typically set to @xmath79 and @xmath80 .
@xmath70 is determined so that , e.g. , the following condition is satisfied : @xmath81 where the spatial difference of sound speed is used as a criterion . if the surface - tension term exists , the stability condition needed to solve it @xcite , @xmath82 should be taken into consideration to determine @xmath70 , where @xmath83 and @xmath84 are the densities of materials on different sides of the interface , and @xmath85 is the surface - tension coefficient .
also , @xmath69 is determined based on the stability condition of a method for the viscous term or others .
at the interfaces where the value or derivatives of physical quantities are in general discontinuous , the use of the conventional centered schemes ( [ eq14])([eq16 ] ) for updating the spatial derivatives is not suitable , providing less accurate results , as will be demonstrated below using a numerical example .
to modify the schemes , we recall here the extrapolation concept investigated in part i @xcite and by others @xcite . as has been pointed out in ref .
@xcite , an interpolation or differencing across the phase boundary may not physically valid , and gives rise to serious numerical errors . in the following ,
those schemes are modified by adapting a simple extrapolation .
equation ( [ eq14 ] ) can be rewritten as @xmath86 @xmath87 we introduce here a switching parameter @xmath88 , defined as @xmath89 @xmath90 where @xmath91 is the i d function ( a density function or a level set function ) defined as @xmath92 in the region occupied by a material and @xmath93 elsewhere , and is updated by solving @xmath94 using this , we modify eq .
( [ eq27 ] ) as @xmath95 this modification vanishes the derivative @xmath96 when @xmath97 , resulting in the advancement of @xmath98 using only the values of an identical material . this procedure corresponds to an extrapolation such as @xmath99 .
equation ( [ eq15 ] ) can be modified in the same manner , as @xmath100 the scheme ( [ eq16 ] ) for the cross derivative can be rewritten as @xmath101.\end{aligned}\ ] ] we modify this as @xmath102,\end{aligned}\ ] ] where @xmath103 the extrapolations used above are merely rough compared with those used to solve the convection parts @xcite ; these , however , might be sufficient , because the derivatives require lower accuracy than those required for the non - derivatives represented here by @xmath28 .
let us perform a numerical test to demonstrate the effectiveness of the above modification .
a one - dimensional nonlinear sound propagation is solved based on eqs .
( [ eq4])([eq6 ] ) .
the convection and other components are neglected for simplicity .
the initial condition is @xmath104 and the boundary condition is @xmath105 where the angular frequency @xmath106 is determined so that the wavelength of the emitted sound wave is @xmath107 .
the initial values of all derivatives are zero .
the square of sound speed is determined by @xmath108 other parameters are @xmath109 and @xmath110 .
( under this condition , @xmath57 is about 0.6 for @xmath111 and is about 0.15 elsewhere . )
the i d function is set to be a color function , @xmath112 the matrix equation given by discretizing eq .
( [ eq18 ] ) is solved by the red - black gauss - seidel method with the convergence criterion of @xmath113 , where @xmath114 and @xmath115 denotes the number of the gauss - seidel iteration .
figures [ fig1](a ) and [ fig1](b ) show the calculated pressure and density distributions at @xmath116 . also , figs . [ fig1](c ) and [ fig1](d ) show the density gradients at the same @xmath117 , calculated using the conventional and the modified schemes .
( in the acoustic step , the derivatives are not used to update the non - derivatives ; thus , the results shown in figs . [ fig1](a ) and [ fig1](b ) are irrespective of the schemes for the derivatives . )
as can be clearly seen , the result obtained by the conventional scheme has a strong overshoot at the interface , while the result obtained by the modified scheme is smooth . the result given using
the modified scheme with @xmath118 and @xmath119 ( the dashed lines ) is similar to the latter result .
these results can roughly prove the validity of the modified scheme , although more detailed discussions should be necessary .
in this section , as a summary of methods , we illustrate how to apply the methods discussed in this series to the compressible navier - stokes equations with a surface tension term : @xmath120 where @xmath121 is the deformation tensor , @xmath122 is the viscosity coefficient , and @xmath123 indicates the surface tension as a volume force .
we factor them into the following four systems : _ convection system : _ @xmath124 _ viscous system : _
@xmath125 _ surface - tension system : _
@xmath126 _ acoustic system : _
@xmath127 successively solving these four systems by the mts technique completes one step of the time integration . among the convection equations , one in terms of the density
is solved by the hybrid interpolation - extrapolation method introduced in part i , and the others are solved by the conventional cip ( the `` type b '' multidimensional formula @xcite is adapted ) .
the time integration of @xmath91 is performed together with the convection system .
the viscous system is solved by the conventional second - order centered finite differencing , and the surface - tension system is solved by the csf model @xcite , coupled with the smoothing procedure introduced in appendix b. the acoustic system is solved by the modified cup method reviewed in sec .
[ sec2:2 ] .
this section presents the application results for the multiphase flow problems , given using the present methods .
compressible and incompressible kelvin - helmholtz instabilities and multibubble dynamics in an acoustic field were selected as application examples .
* compressible and incompressible kelvin - helmholtz instabilities . *
it is well known that an interface between fluids of different densities is unstable .
when the velocity field around the interface is perturbed , the interface forms a complicated structure @xcite .
this instability , called the kelvin - helmholtz instability or the rayleigh - taylor instability , has been the subject of quite a number of theoretical , experimental , and numerical studies @xcite , and even the cip have been employed to investigate it @xcite .
using this example , the effectiveness of the extrapolation technique for the convection parts and that of the mts technique are demonstrated .
the initial arrangement of materials is shown in fig .
the density of the heavy material ( one for @xmath128 ) is @xmath129 , and that of the light one @xmath130 , and the initial pressure is @xmath131 .
the initial velocity field [ @xmath132 is set to @xmath133 @xmath134 @xmath135 \times [ 0,0.5 ] , \ ] ] which satisfies @xmath136 , where @xmath137 is the maximum velocity and @xmath138 ( @xmath139 ) is the wavelength of the perturbation .
the non - dimensional parameter @xmath140 is fixed to 1400 , and @xmath122 is assumed to be constant .
the surface tension and the gravity are neglected .
the slip boundary condition is adapted to the boundaries of @xmath141 and @xmath142 , and the neumann boundary condition ( @xmath143 ) to @xmath144 and @xmath145 .
figures [ fig3](a)[fig3](d ) show the time sequence of @xmath146 for @xmath147 , @xmath148 ( the mach number under this condition is about 0.6 ) , and @xmath34 @xmath149 , at @xmath150 . here , the linear monotone function @xcite is used for the extrapolation , and eq .
( [ eq23 ] ) is used for the mts integration .
the parameters for the time intervals are @xmath151 , @xmath152 , and @xmath153 .
( the inner iteration number for the acoustic parts , @xmath70 , is @xmath154 in this case . )
these results present sharp descriptions of the density interface , while the numerical diffusion and oscillation can be observed in the result given when the convection parts are solved by the conventional cip ( figs . [ fig3](e ) and [ fig4 ] ) .
figures [ fig5](a)[fig5](d ) show @xmath155 surfaces at @xmath150 for different @xmath156 .
( for @xmath157 , the mts integration is disabled , that is , a single time step is used . ) as is clearly seen , decreasing @xmath156 changes the result . a similar result to that for @xmath153 can be given , even using a smaller @xmath19 ( see figs .
[ fig5](e ) and [ fig5](f ) , which show the results for @xmath79 and different @xmath156 ) .
these results reveal that integrating the acoustic parts with a smaller time interval improves the accuracy of the solution , even if the time interval for the convection parts holds . while the above example can be solved by an explicit scheme without any difficulty , the next example is problematical .
figure [ fig6 ] shows the results given by artificially resetting the sound speed to @xmath158 for only the heavy material .
this setting results in a much greater sound speed than the flow velocity , and also results in the coexistence of materials of greatly different compressibilities .
the parameters for the time integration are @xmath159 , @xmath152 , and @xmath160 .
( under this setting , @xmath161 and @xmath162 . )
this result is not changed noticeably by decreasing the time interval , proving the applicability of the present method to a high @xmath163 condition and proving its robustness . using this example ,
we perform here a convergency test .
figures [ fig7](a)[fig7](c ) show the interfaces at @xmath150 given using the same parameters except for the grid width set to @xmath164 , @xmath165 , or @xmath166 , respectively , where @xmath167 .
the refined results are in good agreement with that for @xmath164 .
in contrast , the result given by solving the convection parts with only the interpolation ( i.e. by the conventional cip ) changes as the computational grids are refined ( see figs . [ fig7](d)[fig7](f ) ) ; the convergency of the solution is obviously lower than that obtained by the hybrid method .
these results show that the improvement given in part i accelerates the convergency , and that the hybrid scheme can accurately describe such deformable interfaces .
* two - bubble dynamics in an acoustic field .
* when a sound wave is applied , a bubble immersed in a liquid begins volume oscillation .
when other bubbles exist , they interact acoustically with each other , resulting in the changes in the oscillation amplitude and phase and the effective resonance frequencies @xcite .
moreover , it is known that an acoustic interaction force called the secondary bjerknes force acts between such pulsating bubbles @xcite .
the force is attractive when the bubbles pulsate in - phase , and is repulsive otherwise .
such effects resulting from the radiative interaction between bubbles have been important subjects not only for physicists and mechanical and acoustical engineers , but also for medical engineers and chemists @xcite , and novel insights regarding these effects have been provided even in very recent years @xcite .
( in refs .
@xcite , for example , the author discovered that in multibubble cases , the phase shift of the bubbles pulsation can take place not only at their natural frequencies , but also at some other driving frequencies . in a more recent paper @xcite
, it was found that the latter characteristic frequencies cause the sign reversal of the secondary bjerknes force @xcite . ) solving this problem numerically has some interesting points .
( 1 ) the gas inside the bubbles and the liquid surrounding them have quite different densities and compressibilities .
( 2 ) the compressibility of the gas is not negligible in principle , meaning that a solver for incompressible flows is not applicable .
( 3 ) especially for small bubbles , the viscosity of the surrounding liquid and the surface tension are not negligible .
( 4 ) many different time scales exist , determined by the flow velocity , the sound speed , the surface tension , the viscosity , the bubbles natural frequencies , and the frequency of an external sound .
( 5 ) the topology of the interfaces changes when the bubbles coalesce as a result of their radiative interaction .
based on these interesting factors , we consider this problem to be a very good example for demonstrating the abilities of our methods .
the axisymmetric coordinate @xmath168 is selected for the computational domain , and the centers of the bubbles are located on the central axis .
bubbles 1 and 2 , whose equilibrium radii are @xmath169 @xmath122 m and @xmath170 @xmath122 m , respectively , are filled with a gas of the constant specific heat ratio @xmath171 .
the equilibrium density and the viscosity coefficient are 1.226 kg / m@xmath172 and @xmath173 pa s for the gas phase , and 1000 kg / m@xmath172 and @xmath174 pa s for the liquid phase .
the surface tension coefficient is @xmath175 pa m , and the equilibrium pressure of the liquid is @xmath176 kpa .
the initial pressures inside the bubbles are thus determined by @xmath177 ( for @xmath178 and @xmath179 ) , in order to balance with the surface tension , where @xmath180 and @xmath181 .
the sound speed is set to @xmath182 for the gas , and to @xmath183 for the liquid .
the external sound is applied as the boundary condition to the pressure by @xmath184 , where @xmath185 denotes the boundaries of the computational domain except for @xmath186 , and @xmath187 is the angular frequency of the sound .
the boundary condition for the velocity is free .
the initial distance between the centers of the bubbles [ @xmath188 is fixed to @xmath189 @xmath122 m .
the grid width assumed to be constant is set to @xmath190 @xmath122 m , and the total number of grids is @xmath191 .
@xmath192 in eq .
( [ eq26 ] ) is set to 1000 kg / m@xmath172 (= the equilibrium density of the liquid ) , because the density of the gas is negligible , and that of the liquid can be assumed to be almost constant in time .
figure [ fig8 ] shows the bubbles mean radii and mass centers calculated for @xmath193 and @xmath194 ( @xmath195 ) , where @xmath196 is the linear , monopole natural frequency of bubble @xmath32 determined by @xmath197/\rho r_{j0}^2 } , \ ] ] and the parameters for the mts integration were set to @xmath79 , @xmath198 , and @xmath199 , resulting in @xmath200 . also
[ fig9 ] and [ fig10 ] display the interfaces at selected times .
we can observe the repulsion of the bubbles when @xmath193 , while the attraction when @xmath194 .
these results are validated by the theory presented in ref .
@xcite , and reexamined in refs .
@xcite . the theory determines the sign of the interaction force ( @xmath201 ) by @xmath202 where @xmath203\ ] ] with @xmath204 @xmath205 @xmath206 @xmath207 @xmath208 where @xmath209 , set to @xmath210 in the present study , is the damping coefficient @xcite , @xmath211 , @xmath212 , and @xmath213 are the viscosity coefficient , the equilibrium density and the sound speed , respectively , of the liquid , and exchanging 1 and 2 ( or 10 and 20 ) in the subscripts of these equations yields the expression for @xmath214 .
the positive sign of eq .
( [ eq39 ] ) indicates attraction , while the negative indicates repulsion .
figure [ fig11 ] shows @xmath215 as a function of @xmath187 for @xmath216 @xmath122 m , revealing that the bubbles repel each other when @xmath217 , while they attract when @xmath218 .
this theoretical prediction is well reproduced in the numerical results .
( more detailed discussions from the physical viewpoint are given in ref .
) meanwhile , in the result for @xmath194 , we can observe the coalescence of the bubbles as a result of the attraction .
this result proves the ability of the present methods to treat the topology change .
figure [ fig12 ] shows @xmath19 normalized by @xmath219 , @xmath220 , or @xmath221 , as functions of time .
this figure reveals that the fundamental time interval was mainly determined by the sound speed , whereas the flow velocity was responsible when a rapid motion of the interfaces , caused by the coalescence , occurred .
using the example for @xmath193 , we investigate here the effectiveness of the mts integration on this problem .
figure [ fig13 ] shows the bubbles mean radii as functions of time , for @xmath79 and different @xmath222 and @xmath156 .
large differences can not be seen between the results for @xmath223 , while the results for @xmath224 ( @xmath225 ) are obviously inaccurate .
this result proves that the mts integration has obviously contributed to the accurate solutions given for @xmath223 . during the computation for
@xmath194 introduced above , we sometimes observed negative @xmath6 near the interfaces .
in such a case , we adopted @xmath226 , where @xmath227 is set to @xmath228 ; the numerical results were insensitive to the magnitude of this parameter .
this problem would be overcome by employing the oscillation - free variants of the cip @xcite .
in this series of articles , we have proposed an improved unified solver for compressible and incompressible fluids involving free surfaces , based on the cip - cup method , by adapting several improvements and modifications .
( the most significant one given here in part ii is the adaptation of the mts integration technique , which makes the determination of the time interval very flexible . )
high accuracy and excellent robustness of the improved methods have been demonstrated by using examples of free - surface flows that contain both compressible and incompressible materials .
the present methods , however , face the following challenges : in part i of this series , we proposed five kinds of extrapolation functions used to solve the convection of the density .
although a concrete discussion has not been provided in the present paper , we observed that the most accurate result for a different problem was achieved by a different extrapolation function .
the optimization of the function based on some criterion is , therefore , sorely expected .
although in the present study we have empirically determined the parameters for the mts integration , an automatic , optimized determination should be useful in a practical application .
the optimal time interval for the acoustic parts might be determined by a criterion based on the compressibility of fluids .
if , for example , materials can be considered to be almost completely incompressible , as has been well known previously , one can use a time interval of an infinite cfl number with respect to the sound speed , whereas when a compressible material exists , as has been demonstrated in this paper , one needs a time interval depending on the sound speed .
this means that the maximum values of @xmath222 and @xmath156 sufficient for accurate computations are rightly dependent on the compressibility .
the author would like to thank dr .
n. masuda at gunma university for his helpful supply of materials on the multi - time - step integration methods ( or the individual timestep schemes ) used in the astrophysical community .
f. h. harlow , a. a. amsden , j. comput .
( 1971 ) 197 .
r. i. issa , j. comput .
62 ( 1985 ) 40 .
t. yabe , p. y. wang , j. phys .
jpn 60 ( 1991 ) 2105 .
o. c. zienkiewicz , j. wu , int . j. numer .
35 ( 1992 ) 457 .
r. klein , j. comput .
( 1995 ) 213 .
t. schneider , n. botta , k. j. geratz , r. klein , j. comput .
( 1999 ) 248 .
r. caiden , r. fedkiw , c. anderson , j. comput .
( 2001 ) 1 . b. mller , in : proc .
30@xmath229 comput .
fluid dynam .
( von karman institute for fluid dynamics , lecture series 1999 - 03 ) , march 1999 [ review ] .
t. yabe , in : proc .
eccomas 2000 , barcelona , september 2000 [ review ] .
t. yabe , t. mochizuki , h. hara , in : proc .
lamp 92 , nagoya .
june 1992 , p.387 .
t. yabe , f. xiao , d. zhang , s. sasaki , y. abe , n. kobayashi , t. terasawa , j. geomag .
46 ( 1994 ) 657 .
t. yabe , f. xiao , j. phys .
62 ( 1993 ) 2537 .
y. zhang , t. yabe , comput .
fluid dyman .
j. 8 ( 1999 ) 378 .
t. yabe , f. xiao , t. utsumi , j. comput . phys . 169 ( 2001 ) 556 [ review ] .
f. xiao , j. comput .
( 1999 ) 348 .
s. y. yoon , t. yabe , comput .
commun . 119
( 1999 ) 149 .
_ the special issue of the cip method _ , comput .
fluid dynam .
j. 8 ( 1999 ) .
t. yabe , t. aoki , comput .
66 ( 1991 ) 219 .
m. ida , comput .
132 ( 2000 ) 44 .
s. j. aarseth , mon .
notices roy .
astron . soc . 126
( 1963 ) 223 .
j. makino , publ .
43 ( 1991 ) 859 .
r. d. swindoll , j. m. haile , j. comput .
53 ( 1984 ) 289
. m. tuckerman , b. j. berne , g. j. martyna , j. chem .
97 ( 1992 ) 1990 .
r. zhou , e. harder , h. xu , b. j. berne , j. chem .
( 2001 ) 2348 .
w. c. chao , mon .
110 ( 1982 ) 1603 .
t. aoki , comput .
fluid dynam .
j. 4 ( 1995 ) 279 .
a. a. amsden , f. h. harlow , la-4370 , los - alamos scientific laboratory ( 1970 ) .
s. ito , in : proc .
the 43rd nat .
theor . & appl .
mech . , 1994 ,
[ in japanese ] . s. v. patankar , b. r. baliga , numer . heat transfer 1 ( 1978 ) 27 .
m. ida , y. yamakoshi , jpn . j. appl
40 ( 2001 ) 3846 .
m. ida ( unpublished ) .
j. u. brackbill , d. b. kothe , c. zemach , j. comput .
phys . 100 ( 1992 ) 335 .
m. sussman , e. fatemi , p. smereka , s. osher , comput .
fluids , 27 ( 1998 ) 663 .
p. charrier , b. tessieras , siam j. numer .
23 ( 1986 ) 461 .
d. mao , j. comput .
92 ( 1991 ) 422 . m. ida , in : proc .
10th symp .
comput . fluid .
dynam . , tokyo , japan , 1996 , p. 382
[ in japanese ] .
m. ida , in : proc .
10th comput .
conf . , tokyo , japan , 1997 , p. 17
[ in japanese ] .
d. h. sharp , physica d 12 ( 1984 ) 3 [ review ] .
j. glimm , x. l. li , r. menikoff , d. h. sharp , q. zhang , phys .
fluids a 2 ( 1990 ) 2046 .
d. youngs , phys .
fluids a 3 ( 1991 ) 1312 .
t. yabe , h. hoshino , t. tsuchiya , phys .
a 44 ( 1991 ) 2756 .
w. mulder , s. osher , j. a. sethian , j. comput .
( 1992 ) 209 .
s. k. zhdanov , physica d 87 ( 1995 ) 375 .
m. b. schneider , g. dimonte , b. remington , phys .
80 ( 1998 ) 3507 .
l. meignin , p. ern , p. gondret , m. rabaud , phys .
e 64 ( 2001 ) 026308 .
t. yabe , t. aoki , g. sakaguchi , p. y. wang , t. ishikawa , comput .
fluids 19 ( 1991 ) 421 .
m. strasberg , j. acoust .
25 ( 1953 ) 536 .
a. shima , trans .
asme , j. basic eng .
93 ( 1971 ) 426 .
j. f. scott , j. fluid mech .
113 ( 1981 ) 487 .
l. a. crum , j. acoust .
57 ( 1975 ) 1363 .
e. a. zabolotskaya , sov .
acoust . 30
( 1984 ) 365 .
j. r. blake , p. b. robinson , a. shima , y. tomita , j. fluid mech .
255 ( 1993 ) 707 .
p. a. dayton , k. e. morgan , a. l. klibanov , g. brandenburger , k. r. nightingale , k. w. ferrara , ieee trans . ultrason .
ferroelect . &
control 44 ( 1997 ) 1264 .
o. louisnard , n. lyczko , f. espitalier , m. urzedowski , y. vargas - hernandez , c. sanchez - romero , ultrason .
sonochem . 8
( 2001 ) 183 . c. feuillade , j. acoust .
109 , ( 2001 ) 2606 .
w. lauterborn , t. kurz , r. mettin , c. d. ohl , adv .
( 1999 ) 295 [ review ] .
z. ye , a. alvarez , phys .
rev . lett .
80 ( 1998 ) 3503 .
a. a. doinikov , phys .
e 64 ( 2001 ) 026301 .
m. ida , phys .
lett . a 297 ( 2002 ) 210 .
m. ida , j. phys .
71 ( 2002 ) 1214 .
m. ida ( submitted ) ; e - print , physics/0109005 .
a. harkin , t. j. kaper , a. nadim , j. fluid mech .
445 ( 2001 ) 377 .
a. prosperetti , ultrasonics 22 ( 1984 ) 69 [ review ] .
m. ida , e - print , physics/0111138 ( still in preparation ) .
f. xiao , t. yabe , t. ito , comput .
93 ( 1996 ) 1 .
f. xiao , mon .
128 ( 2000 ) 1165 .
m. ida , comput .
fluid dynam .
j. 10 ( 2001 ) 159 .
m. ida , comput .
commun . 143
( 2002 ) 142 . c. w. hirt ,
b. d. nichols , j. comput . phys .
39 ( 1981 ) 201 . m. rudman , int .
fluids 28 ( 1998 ) 357 . to discretize the acoustic , viscous , and surface tension terms on the staggered grids , we need the values of @xmath146 at the velocity positions , i.e. , @xmath230 and @xmath231 .
in this appendix , we briefly discuss how to estimate them .
as discussed in part i , the phase boundary is recognized using the zero level set of the i d function ( identification function ; the color or the level set function ) , @xmath91 , and materials are identified using the sign of the function . if @xmath232 is true , it is recognized that the call between ( @xmath233 ) and ( @xmath234 ) contains an interface .
in such a cell , we estimate the densities by the following vof @xcite - like procedure , which implies the weighted linear interpolation : where @xmath237 ( @xmath238 . note that the position of the interface is changed only when solving the convection parts . )
is the i d function after solving the convection part . in the remaining cells , we use the simple average , @xmath243 is the mean radius of the surface curvature , and @xmath244 denotes the color function defined as @xmath245 for a material , and @xmath246 otherwise .
the singularity appearing in @xmath247 at the interface is mollified by some smoothing procedures @xcite . in the present study ,
the smoothed @xmath244 is constructed as follows : * step 1 : * the non - smoothed color function is made from the i d function , which is an arbitrary function whose @xmath155 surface represents the interface @xcite , by setting @xmath248 for @xmath249 and @xmath250 for @xmath251 .
( @xmath91 always has a non - zero value at the gird points @xcite . )
steps 2 and 3 are necessary to approximately take into account the phase property of the interface , which is not contained in @xmath244 given at step 1 , but is latent between the grid points .
@xmath261 and @xmath260 are typically set to @xmath262 and @xmath263 .
|
an improved numerical solver for the unified solution of compressible and incompressible fluids involving interfaces is proposed .
the present method is based on the cip - cup ( cubic interpolated propagation / combined , unified procedure ) method , which is a pressure - based semi - implicit solver for the euler equations of fluid flows . in part
i of this series of articles [ m. ida , comput .
phys .
commun .
132 ( 2000 ) 44 ] , we proposed an improved scheme for the convection terms in the equations , which allowed us discontinuous descriptions of the density interface by replacing the cubic interpolation function used in the cip scheme with a quadratic extrapolation function only around the interface . in this paper , as part ii of this series , the multi - time - step integration technique is adapted to the cip - cup integration . because the cip - cup treats different - nature components in the fluid equations separately ,
the adaptation of the technique is straightforward .
this modification allows us flexible determinations of the time interval , which results in an efficient and accurate integration .
furthermore , some additional discussion on our methods is presented .
finally , the application results to composite flow problems such as compressible and incompressible kelvin - helmholtz instabilities and the dynamics of two acoustically coupled deformable bubbles in a viscous liquid are provided .
unified solution , cip , compressible fluid , incompressible fluid , multi - time - step integration , extrapolation , free - surface flow , bubble dynamics 02.70.bf , 47.11.+j , 47.20.ma , 43.25.yw
|
12 brhl , r. , kalinin , a. , kornilov , o. , toennies , j. p. , hegerfeldt , g. c. , & stoll , m. matter wave diffraction from an inclined transmission grating : searching for the elusive @xmath90he trimer efimov state .
lett . _ * 95 * , 063002 ( 2005 ) .
we thank e. braaten , c. greene , b. esry , h. hammer , and t. khler for many stimulating and fruitful discussions and e. kneringer for support regarding the data analysis . we acknowledge support by the austrian science fund ( fwf ) within spezialforschungsbereich 15 and within the lise meitner program , and by the european union in the frame of the tmr networks `` cold molecules '' and `` fastnet '' .
is supported within the doktorandenprogramm of the austrian academy of sciences . .
the shaded region indicates the scattering continuum for three atoms ( @xmath67 ) and for an atom and a dimer ( @xmath10 ) .
the arrow marks the intersection of the first efimov trimer with the three - atom threshold . to illustrate the series of efimov states , we have artificially reduced the universal scaling factor from @xmath93 to @xmath94 . for comparison ,
the dashed line indicates a tightly bound non - efimov trimer@xcite ( t ) which does not interact with the scattering continuum.,width=264 ] is plotted as a function of the scattering length @xmath0 .
the dots and the filled triangles show the experimental data from setup a for initial temperatures around 10nk and 200nk , respectively .
the open diamonds are from setup b at temperatures of 250nk .
the open squares are previous data@xcite at initial temperatures between 250 and 450nk .
the solid curve represents the analytic model from effective field theory@xcite with @xmath95 , @xmath96 , and @xmath97 .
the straight lines result from setting the @xmath98 and @xmath99-terms in the analytic theory to @xmath100 , which gives a lower recombination limit for @xmath67 and an upper limit for @xmath10 .
the inset shows an expanded view for small positive scattering lengths with a minimum for @xmath101 near @xmath66 .
the displayed error bars refer to statistical uncertainties only .
uncertainties in the determination of the atomic number densities may lead to additional calibration errors for @xmath102 of up to 20%.,width=370 ]
|
systems of three interacting particles are notorious for their complex physical behavior .
a landmark theoretical result in few - body quantum physics is efimov s prediction@xcite of a universal set of bound trimer states appearing for three identical bosons with a resonant two - body interaction .
counterintuitively , these states even exist in the absence of a corresponding two - body bound state .
since the formulation of efimov s problem in the context of nuclear physics 35 years ago , it has attracted great interest in many areas of physics@xcite .
however , the observation of efimov quantum states has remained an elusive goal@xcite . here
we report the observation of an efimov resonance in an ultracold gas of cesium atoms .
the resonance occurs in the range of large negative two - body scattering lengths , arising from the coupling of three free atoms to an efimov trimer .
experimentally , we observe its signature as a giant three - body recombination loss@xcite when the strength of the two - body interaction is varied .
we also detect a minimum@xcite in the recombination loss for positive scattering lengths , indicating destructive interference of decay pathways .
our results confirm central theoretical predictions of efimov physics and represent a starting point with which to explore the universal properties of resonantly interacting few - body systems@xcite .
while feshbach resonances@xcite have provided the key to control quantum - mechanical interactions on the two - body level , efimov resonances connect ultracold matter@xcite to the world of few - body quantum phenomena .
institut fr experimentalphysik , universitt innsbruck , technikerstrae 25 , a6020 innsbruck , austria james franck institute , physics department of the university of chicago , 5640 s. ellis ave .
chicago , illinois 60637 , usa institut fr quantenoptik und quanteninformation der sterreichischen akademie der wissenschaften , otto - hittmair - platz 1 , a6020 innsbruck , austria efimov s treatment of three identical bosons@xcite is closely linked to the concept of universality@xcite in systems with a resonant two - body interaction , where the s - wave scattering length @xmath0 fully characterizes the two - body physics .
when @xmath1 greatly exceeds the characteristic range @xmath2 of the two - body interaction potential , details of the short - range interaction become irrelevant because of the long - range nature of the wave function .
universality then leads to a generic behavior in three - body physics , reflected in the energy spectrum of weakly bound efimov trimer states . up to now , in spite of their great fundamental importance , these states could not be observed experimentally .
an observation in the realm of nuclear physics , as originally proposed by efimov , is hampered by the presence of the coulomb interaction , and only two - neutron halo systems with a spinless core are likely to feature efimov states@xcite . in molecular physics , the helium trimer
is predicted to have an excited state with efimov character@xcite .
the existence of this state could so far not be confirmed@xcite .
a different approach to experimentally study the physics of efimov states is based on the unique properties of ultracold atomic quantum gases .
such systems@xcite facilitate an unprecedented level of control to study interacting quantum systems .
the ultra - low collision energies allow to explore the zero - energy quantum limit .
moreover , two - body interactions can be precisely tuned based on feshbach resonances@xcite .
efimov s scenario@xcite can be illustrated by the energy spectrum of the three - body system as a function of the inverse scattering length @xmath3 ( fig .
[ efimov ] ) .
let us first consider the well - known weakly bound dimer state , which only exists for large positive @xmath0 . in the resonance regime ,
its binding energy is given by the universal expression @xmath4 , where @xmath5 is the atomic mass and @xmath6 is planck s constant divided by @xmath7 . in fig .
[ efimov ] , where the resonance limit corresponds to @xmath8 , the dimer energy @xmath9 is represented by a parabola for @xmath10 . if we now add one more atom with zero energy , a natural continuum threshold for the bound three - body system ( shaded region in fig .
[ efimov ] ) is given by the three - atom threshold ( @xmath11 ) for negative @xmath0 and by the dimer - atom threshold ( @xmath9 ) for positive @xmath0 .
energy states below the continuum threshold are necessarily three - body bound states .
when @xmath3 approaches the resonance from the negative-@xmath0 side , a first efimov trimer state appears in a range where a weakly bound two - body state does not exist . when passing through the resonance the state connects to the positive-@xmath0 side , where it finally intersects with the dimer - atom threshold .
an infinite series of such efimov states is found when scattering lengths and energies are increased in powers of universal scaling factors@xcite @xmath12 and @xmath13 ( @xmath14 ) , respectively .
resonant scattering phenomena arise as a natural consequence of efimov s scenario@xcite .
when an efimov state intersects with the continuum threshold at negative scattering lengths @xmath0 , three free atoms in the ultracold limit resonantly couple to a trimer .
this results in a triatomic efimov resonance . at finite collision energies
, the phenomenon evolves into a triatomic continuum resonance@xcite .
another type of efimov resonances@xcite is found at positive values of @xmath0 for collisions between a free atom and a dimer , when efimov states intersect with the dimer - atom threshold . while the latter type of efimov resonances corresponds to feshbach resonances in collisions between atoms and dimers@xcite ,
triatomic efimov resonances can be interpreted as a three - body generalization to feshbach resonances@xcite .
striking manifestations of efimov physics have been predicted for three - body recombination processes in ultracold gases with tunable two - body interactions@xcite .
three - body recombination leads to losses from a trapped gas with a rate proportional to the third power of the atomic number density .
these losses are commonly described@xcite in terms of a loss rate coefficient @xmath15 . in the resonant case ( @xmath16 ) , it is convenient to express this coefficient in the form @xmath17 , separating a general @xmath18-scaling@xcite from an additional dependence@xcite @xmath19 .
efimov physics is reflected in a logarithmically periodic behavior @xmath20 , corresponding to the scaling of the infinite series of weakly bound trimer states
. for negative scattering lengths , the resonant coupling of three atoms to an efimov state opens up fast decay channels into deeply bound dimer states plus a free atom .
triatomic efimov resonances thus show up in giant recombination loss .
this striking phenomenon was first identified in numerical solutions to the adiabatic hyperspherical approximation of the three - body schrdinger equation assuming simple model potentials and interpreted in terms of tunneling through a potential barrier in the three - body entrance channel@xcite .
a different theoretical approach@xcite , based on effective field theory , provides the analytic expression @xmath21 } + \sinh^2{\eta_- } \right)$ ] .
the free parameter @xmath22 for the resonance positions at @xmath22 , @xmath23 , @xmath24 depends on the short - range part of the three - body interaction and is thus not determined in the frame of the universal long - range theory . as a second free parameter , the dimensionless quantity @xmath25 describes the unknown decay rate of efimov states into deeply bound dimer states plus a free atom and thus characterizes the resonance width .
our measurements are based on the magnetically tunable interaction properties of cs atoms@xcite in the lowest internal state . by applying fields between @xmath26 and @xmath27 g
, we varied the s - wave scattering length @xmath0 in a range between @xmath28 to @xmath29 , where @xmath30 is bohr s radius .
accurate three - body loss measurements are facilitated by the fact that inelastic two - body loss is energetically forbidden@xcite .
the characteristic range of the two - body potential is given by the van der waals length@xcite , which for cs is @xmath31 .
this leaves us with enough room to study the universal regime requiring @xmath32 . for negative @xmath0 ,
a maximum value of @xmath33 is attainable for @xmath34 .
efimov s estimate @xmath35 for the number of weakly bound trimer states@xcite suggests the presence of one efimov resonance in the accessible range of negative scattering lengths .
our experimental results ( fig . [ rec_length ] ) , obtained with optically trapped thermal samples of cs atoms in two different set - ups ( see methods ) , indeed show a giant loss feature marking the expected resonance .
we present our data in terms of a recombination length@xcite @xmath36^{1/4}$ ] , which leads to the simple relation @xmath37 .
note that the general @xmath18-scaling corresponds to a linear behavior in @xmath38 ( straight lines in fig .
[ rec_length ] ) .
a fit of the analytic theory@xcite to our experimental data taken for negative @xmath0 at temperatures @xmath39nk shows a remarkable agreement and determines the resonance position to @xmath40 and the decay parameter to @xmath41 .
the pronounced resonance behavior with a small value for the decay parameter ( @xmath42 ) demonstrates a sufficiently long lifetime of efimov trimers to allow their observation as distinct quantum states .
all the results discussed so far are valid in the zero - energy collision limit of sufficiently low temperatures . for ultralow but non - zero temperatures the recombination length is unitarity limited@xcite to @xmath43 . for @xmath44nk
this limit corresponds to about @xmath45 and our sample is thus cold enough to justify the zero - temperature limit . for 250nk
, however , unitarity limits the recombination length to about @xmath46 .
the efimov resonance is still visible at temperatures of 200 and 250nk ( filled triangles and open diamonds in fig .
[ rec_length ] ) .
the slight shift to lower values of @xmath1 suggests the evolution of the zero - energy efimov resonance into a triatomic continuum resonance@xcite . in further experiments at higher temperatures ( data not shown ) we observed the resonance to disappear above @xmath47500nk
. for positive scattering lengths , we found three - body losses to be typically much weaker than for negative values .
our measurements are consistent with a maximum recombination loss of @xmath48 , or equivalently @xmath49 , as predicted by different theories@xcite ( straight line for @xmath10 in fig . [ rec_length ] ) . for @xmath0 below @xmath50
the measured recombination length significantly drops below this upper limit ( inset ) .
the analytic expression from effective field theory@xcite for @xmath10 reads @xmath51 + \sinh^2\eta_+ \right ) + 16.8\ , ( 1-e^{-4\eta_+})$ ] with the two free parameters @xmath52 and @xmath53 .
the first term describes recombination into the weakly bound dimer state with an oscillatory behavior due to an interference effect between two different pathways@xcite .
the second term results from decay into deeply bound states .
we use this expression to fit our data points with @xmath54 .
this somewhat arbitrary condition is introduced as a reasonable choice to satisfy @xmath55 for the validity of the universal theory .
the fit is quite insensitive to the value of the decay parameter and yields @xmath56 .
this result is consistent with the theoretical assumption@xcite of the same value for the decay parameter for positive and negative @xmath0 , which in our case is @xmath57 . for maximum loss ,
we obtain @xmath58 . according to theory@xcite
, the trimer state hits the dimer - atom threshold at @xmath59 .
the logarithmic periodicity of the efimov scenario suggests adjacent loss minima to occur at @xmath60 and at @xmath61 .
while the former value is out of our accessible range , the latter value ( @xmath62 ) is too small to strictly justify universal behavior in the resonance limit ( @xmath63 ) .
nevertheless , our experimental results ( inset to fig . [ rec_length ] ) indicate a minimum at @xmath64 and the analytic expression for @xmath19 is found to describe our data quite well down to this minimum .
the occurrence of the interference minimum in three - body loss is demonstrated more clearly in another set of experiments ( fig .
[ loss_fraction ] ) , where we simply measured the loss of atoms after a fixed storage time in the optical trap .
this minimum is located at @xmath65 in addition to a second minimum close to zero scattering length .
we point out that the existence of the minimum at @xmath66 is very advantageous for efficient evaporative cooling of cs as it combines a large scattering cross section with very low loss .
inadvertently , we have already benefitted of this loss minimum for the optimized production of a bose - einstein condensate of cs@xcite .
the comparison of our experimental results to available three - body theory shows remarkable agreement , although the collision physics of cs is in general a very complicated multi - channel scattering problem .
we believe that the particular nature of the broad , `` open - channel dominated '' feshbach resonance@xcite that underlies the tunability of our system plays a crucial role .
for such a resonance , the two - body scattering problem can be reduced to an effective single - channel model .
it is a very interesting question to what degree this great simplification of the two - body physics extends to the three - body problem .
here we in particular raise question how the regions of positive and negative scattering lengths are connected in our experiment , where @xmath0 is changed through a zero crossing , i.e. through a non - universal region , and not across the universal resonance region . in our case
, there is no obvious connection between the efimov state that leads to the observed resonance for @xmath67 and the states responsible for the behavior for @xmath10 . in our analysis of the experimental data
, we have thus independently fitted the data sets for negative and positive @xmath0 .
nevertheless , the resulting values for the two independent fit parameters @xmath22 and @xmath52 suggest a connection : for the ratio @xmath68 our experiment yields 1.25(9 ) , whereas universal theory@xcite predicts 0.96(3 )
. these numbers are quite close in view of the efimov factor of 22.7 .
if not an accidental coincidence , we speculate that the apparent relation between @xmath52 and @xmath22 may be a further consequence of universality in a system where the resonant two - body interaction can be modelled in terms of a single scattering channel . in general
, the multi - channel nature of three - body collisions near feshbach resonances@xcite leads to further interesting questions , like e.g. possible resonance effects beyond the efimov scenario .
advances in three - body theory are necessary to finally answer these questions and to provide a complete interpretation of our present observations . in the past few years , applications of feshbach resonances in ultracold gases and the resulting possibility to create
dimer states have set the stage for many new developments in matter - wave quantum physics .
the observation of an efimov resonance now confirms the existence of weakly bound trimer states and opens up new vistas@xcite to experimentally explore the intriguing physics of few - body quantum systems .
* methods * * magnetic tuning of the two - body interaction * for cs atoms in their energetically lowest state ( quantum numbers @xmath69 for the total spin and @xmath70 for its projection ) the s - wave scattering length @xmath0 varies strongly with the magnetic field@xcite .
between 0 and 150 g the dependence can in general be well approximated by the fit formula @xmath71 except for a few narrow feshbach resonances@xcite .
the smooth variation of the scattering length in the low - field region results from a broad feshbach resonance centered at about @xmath72 g ( equivalent to @xmath73 g in the state @xmath69 , @xmath74 ) . in all our measurements we excluded the magnetic field regions where the narrow feshbach resonances influence the scattering behavior through coupling to other molecular potentials .
the efimov resonance is centered at 7.5 g . *
trap setups and preparation of the cs gases * all measurements were performed with trapped thermal samples of cs atoms at temperatures @xmath75 ranging from 10 to 250nk .
we used two different experimental setups , which have been described elsewhere@xcite . in setup
a we first produced an essentially pure bose - einstein condensate ( bec ) with up to 250,000 atoms in a far - detuned crossed optical dipole trap generated by two 1060-nm yb - doped fiber laser beams@xcite .
we then ramped the magnetic field to 16.2 g where the scattering length is negative with a value of @xmath76 , thus inducing a collapse of the condensate@xcite .
after an equilibration time of 1s we were left with a thermal sample at typically @xmath77nk containing up to 20,000 atoms at peak densities ranging from @xmath78@xmath79 to @xmath80@xmath79 . alternatively , we interrupted the evaporation process before condensation to produce thermal samples at @xmath81nk in a crossed dipole trap generated by one of the 1060-nm beams and a @xmath82-@xmath83 m co@xmath84-laser beam .
after recompression of the trap this produced typical densities of @xmath85@xmath79 .
the measurements in the region of the loss minima as displayed in fig.[loss_fraction ] were taken after a storage time of 200ms at initial densities of @xmath86@xmath79 . in setup
b we used an optical surface trap@xcite in which we prepared a thermal sample of 10,000 atoms at @xmath87nk via forced evaporation at a density of @xmath88@xmath79 .
the dipole trap was formed by a repulsive evanescent laser wave on top of a horizontal glass prism in combination with a single horizontally confining 1060-nm laser beam propagating along the vertical direction .
* determination of three - body loss rate coefficients * we measured three - body loss rates in setup a by recording the time evolution of the atom number @xmath89 and the temperature @xmath75 .
a detailed description of this procedure has been given in ref .
@xcite . in brief
, the process of three - body recombination not only leads to a loss of atoms , but also induces `` anti - evaporation '' and recombination heating .
the first effect is present at any value of the scattering length @xmath0 .
the second effect occurs for positive values of @xmath0 when the recombination products remain trapped .
atom loss and temperature rise are modelled by a set of two coupled non - linear differential equations .
we used numerical solutions to this set of equations to fit our experimental data . from these fits together with measurements of the trapping parameters we obtained the rate coefficient @xmath15 . in setup
b we recorded the loss at sufficiently short decay times for which heating is negligible .
|
entanglement , a key concept in quantum - information science @xcite , has also been playing a pivotal role in the study of quantum ( or , generally , nonclassical ) correlations and related consequences in many - body systems @xcite , especially quantum phase transitions ( qpts ) induced by varying parameters of system hamiltonian at zero temperature @xcite ( see also refs . @xcite for geometric approaches to qpts . ) . accordingly
, many efforts have been made to find the possible relation between qpt in a system and nonanalyticities of , e.g. , pairwise entanglement of its ground state ( gs ) @xcite .
for example , under some fairly general conditions , a classification of qpts in terms of pairwise entanglement has been provided in ref .
besides , through density functional theoretical arguments , it has been shown that by an appropriate choice of entanglement measure ( either bi- or multipartite ) , one can study physical properties of a system near its critical points @xcite .
it is thus evident that having a sufficiently strong entanglement measure which is able to detect entanglement regions of the gs ( in terms of the hamiltonian parameters ) in a more subtle way is of paramount importance .
the difficulty of the issue , i.e. , quantifying quantum correlations , is mostly due to the existence of various types of entanglement in quantum many - body systems , such as free entanglement ( fe ) , bound entanglement ( be ) @xcite , and pairwise or multipartite entanglement @xcite , featuring also new phases for quantum matter ( e.g. , topological order @xcite ) . here
, however , we shall concentrate only on be . simply speaking , bound entangled states , in contrast to free entangled states , are those states which are not directly useful for quantum - information tasks such as teleportation @xcite .
because of the positive partial transpose ( ppt ) property of bound entangled states ( in the bipartite case ) @xcite , their entanglement can not be distilled by any local operation and classical communication ; for a general review , see , e.g. , ref . @xcite .
the possibility of the existence of non - ppt be , however , is still an open question @xcite .
nevertheless , bound entangled states are still useful because , for example , cryptographically secure keys can be constructed from them @xcite , they can enhance the fidelity of conclusive teleportation hence , the teleportation power of another bipartite state @xcite , they manifest irreversibility in the entanglement formation - distillation cycle ( hence , also of interest in thermodynamics ) @xcite , and they have applications in channel discrimination @xcite .
it is , therefore , intriguing to see how or whether be can appear in quantum many - body systems .
recently , it has been shown that a thermal environment can naturally induce the thermal transition from fe to be in quantum many - body systems @xcite .
additionally , in a simple @xmath0 spin chain , varying the magnetic field can generate bound entangled gss @xcite . in a recent experiment @xcite ,
an optical four - qubit state with be has been realized and then characterized fully .
there is vast literature on qpts induced by a transition from separability to fe or a transition in the fe region ( in the sense of bipartite reduced density matrices of the gss ) . in this work , in contrast
, we are interested in qpts in which the transition happens because of a change in the type of entanglement , specifically , from fe to be ( or conversely ) , or from separability to be ( or conversely ) .
this is particularly interesting , because one of the vastly used measures of bipartite entanglement in the context of qpts , negativity @xcite , is inherently unable of detecting be , hence it would fail to signal a corresponding criticality .
there are , however , some other computable ( but not necessarily conclusive ) inseparability tests for recognizing be , based on the cross norm @xcite and permutation @xcite criteria . in the case of quantum spin chains , a complete set of spin squeezing inequalities has also been introduced which can be used for experimentally detecting be of many - body thermal states @xcite .
more appealing and arguably a stronger identifier than entanglement measures for detecting qpts is provided by the ground - state fidelity ( gsf ) @xcite or its second derivative @xcite , fidelity susceptibility ( fs ) .
gsf and fs have been successfully used in the context of qpts in correlated quantum many - body systems @xcite .
we address the types of qpts we are interested in through a reverse - construction method . given a many - body density matrix @xmath1 , we construct a spin - one chain hamiltonian @xmath2 whose gs has @xmath1 as the reduced density matrix of specific adjacent sites .
thereby , by tuning the hamiltonian parameters in some suitable intervals , we will have gss representing qpts accompanied by the transition of their entanglement type from be to fe or separability .
the technique is in principle relatively general and is illustrated by three examples .
we should remark that we are not aiming to construct physically realizable hamiltonians .
thus , here we are not concerned with the possible appearance of many - body ( three - body or higher order ) interactions in the constructed hamiltonians .
this , although might seem to restrict the realistic applicability of the approach , is not a fundamental issue .
there exists a powerful perturbative ( gadget " ) machinery to construct arbitrary @xmath3-body ( effective ) interactions from two - body hamiltonians @xcite .
consider a closed chain of @xmath4 identical quantum systems , e.g. , spins , each of which has a @xmath5-dimensional hilbert space .
let us assume that we are given a global state @xmath6 whose two - site reduced density matrices ( tsrdms ) have some preset properties .
for example , for odd @xmath7 , @xmath8 $ ] must be compatible with an already given @xmath9 , where @xmath10 ( we denote the tsrdms for even @xmath7 by @xmath11 ) . for simplicity , let us suppose that there is a symmetry by which all @xmath9 ( @xmath12 ) are the same for all odd ( even ) @xmath7 .
that is , the matrix form of @xmath13 ( @xmath14 ) can be considered as a representative of all @xmath9 ( @xmath12 ) with odd ( even ) @xmath7 .
we also assume that @xmath15 where @xmath16 for all @xmath17 , @xmath18 , @xmath19 , and @xmath20 .
that is , eqs .
( [ rhoii1 ] ) are the spectral representations of @xmath9 and @xmath12 ( including degeneracies ) , where the vectors @xmath21 and @xmath22 constitute their null eigenspaces . our goal is to construct ( or , exactly speaking , reverse construct ) a corresponding hamiltonian @xmath2 for which this @xmath6 is a gs .
we follow similar steps as in ref . @xcite . in the following ,
we restrict ourselves only to positive semidefinite hamiltonians .
thus , for any arbitrary @xmath6 , we have @xmath23 , so that if for a @xmath6 , @xmath24 this is an eigenvector of @xmath2 with zero eigenvalue , namely , the gs of the hamiltonian .
moreover , we are interested in nearest - neighbor two - body ( or at most three - body ) interactions .
that is , the sought - after @xmath2 has the following general form : @xmath25 it then can be shown that @xmath26 + \sum_{i~\text{even } } \mathrm{tr}[h'_{i , i+1 } \varrho'_{i , i+1}].\nonumber\\ \label{php}\end{aligned}\ ] ] thus , if for odd ( even ) @xmath7 we construct their corresponding @xmath27 ( @xmath28 ) from the null eigenvectors ( or any combination or subset of the vectors in the null eigenspaces ) @xmath29 ( @xmath30 ) , as @xmath31}_{\jmath } |w_{\jmath}\rangle_{i , i+1 } \langle w_{\jmath}| \label{hii1-f}\end{aligned}\ ] ] ( and similarly for @xmath28 ) , where @xmath32 ( @xmath33 ) and @xmath34}_{\jmath}$ ] ( @xmath35}_{\jmath}$ ] ) are arbitrary nonnegative numbers , then all the terms in the summations of eq .
( [ php ] ) would vanish .
this of course is a sufficient condition , not always necessary .
overall , with this choice for @xmath2 , we have @xmath36 , implying that @xmath6 is a gs . put briefly , then , the construction boils down to finding the null eigenspaces of @xmath13 and @xmath37 .
as long as , for the given @xmath6 , the nullities of these density matrices are nonzero , one can find a nearest - neighbor two - body hamiltonian describing the interactions on the chain .
otherwise , the above recipe fails to provide coupling between all the neighboring particles on the chain .
this , however , does not necessarily imply nonexistence of other compatible two - body hamiltonians . to remedy the issue with our construction ,
given that @xmath6 is still the state with the desired properties , a possible way is to consider higher order interactions , e.g. , three body couplings .
we should consider @xmath38 , or in general @xmath39 ( for some @xmath40 ) , rather than @xmath9 ( for odd @xmath7 ) . now
if we repeat the hamiltonian construction recipe , we end up with @xmath40-local hamiltonians @xmath41 obtained from the null eigenspace of @xmath39 .
thus , for example , the following hamiltonian @xmath42 where @xmath7 could be odd or even , provides a many - body interaction for which @xmath6 is a gs . as a general example for the hamiltonian construction ,
let us consider the following global state : @xmath43 this specific form implies that except for @xmath9 , for odd @xmath7 , the reduced density matrices of all other pairs of sites are separable @xcite .
the proof is straightforward , e.g. , by using the schmidt decomposition for the bipartite states @xmath44}_{\ell } } |\ell^{[\imath]}\rangle_i |\ell^{[\imath]}\rangle_{i+1}$ ] , for odd @xmath7 .
thus , the state @xmath13 fully captures the pairwise ( i.e , two - site ) entanglement property of the whole chain .
for example , if @xmath13 is bound entangled , no two - site fe then can be distilled from @xmath6 . our examples in sec .
[ exs ] are all of the type of eq .
( [ e3 ] ) ; besides , they have zero - nullity @xmath14 , hence @xmath2 will be many bodied .
a few remarks about the recipe for the construction of the hamiltonian are in order here .
( i ) there could be many hamiltonians ( with a set of given requirements , such as symmetry , range of interactions , and so on ) for which a given @xmath6 is a gs .
our construction provides only one class of such hamiltonians .
( ii ) by construction , our recipe implies that underlying symmetries of the state @xmath1 would also carry over into the hamiltonian @xcite .
( iii ) note that @xmath6 is not necessarily the unique gs of the constructed @xmath2 .
for example , each term @xmath45 ( for any @xmath17 ) is also a gs ( one can construct other combinations which are gs ) .
thus , when @xmath1 is mixed ( i.e. , @xmath46 , which is a generic case ) , the gs is degenerate .
this means that if we consider the system with the hamiltonian @xmath2 , at zero temperature the system can be everywhere in the ground eigenspace spanned by @xmath47 and other possible gss . since this degeneracy is symmetry - driven , we can circumvent it by perturbing the system with a small symmetry - breaking term , i.e. , @xmath48 , and then allowing @xmath49 @xcite . for our purposes here and for the sake of clarity , however , the working gs is assumed to be @xmath6 ( which , for example , can be considered to be singled out by some selection rule ) .
( iv ) in most physically interesting cases , we have @xmath50 ( @xmath51-local hamiltonians ) .
however , since our goal in this paper is just to provide a framework for be qpts , we will not be concerned with the possible appearance of many - body interactions in the hamiltonians .
additionally , there exists a powerful perturbative gadget mechanism to construct arbitrary three - body hamiltonians from two - body interactions @xcite .
( v ) often ( but not necessarily always ) , if the state @xmath1 ( or @xmath6 ) depends on some controllable parameter ( or a set of parameters ) , say , @xmath52 , so will the hamiltonian @xmath53 . in the following , after briefly introducing gsf as the main tool for analyzing qpts , we shall illustrate our construction through three different examples featuring be .
entanglement is a rather universal means of studying qpts @xcite .
the ground - state fidelity ( gsf ) @xcite has been shown to be a very powerful tool that is capable of identifying various types of qpts , usually independent of symmetries , system - dependent details , or the nature of underlying quantum correlations ; hence , it can be used as a suitable order parameter . " recalling that negativity , is defined as @xmath54 , where @xmath55 ( in a given basis ) is the partial transposition and @xmath56 $ ] is the trace norm @xcite . ] a widely used entanglement measure , fails to signal be , and that other measures which may signal be are usually inconclusive , gsf or its derivatives seem appropriate supplemental tools for our analysis .
the rationale behind the notion of the gsf is that , at zero temperature , the gs describes the whole characteristics of a typical quantum system .
hence , when a qpt occurs at some point @xmath57 , a sudden change between the behavior of the gs slightly before ( @xmath58 ) and after ( @xmath59 ) this critical point may take place , which in principle can be captured by the following fidelity : @xmath60 the second derivative of the gsf with respect to @xmath61 , calculated at @xmath62 , is called the fidelity susceptibility hereafter denoted by @xmath63and _
per se _ is a powerful measure for detecting qpts @xcite . for the state @xmath64 as in eq .
( [ e3 ] ) , @xmath65 reads @xmath66 in which @xmath67 , and @xmath68 ( @xmath61 ) will be taken to be relatively large ( small ) .
here we remark that the _ reduced _ fidelity ( susceptibility)denoting the similarity between the density matrix of a subsystem of the gs before and after the qpt point can also be a good local order parameter for identifying symmetry - breaking @xcite and topological @xcite qpts .
there are also other , relatively less universal measures ( than the gsf , but in some sense stronger than negativity or concurrence @xcite ) to recognize ( particularly ) be associated with the tsrdm of the gs , e.g. , the realignment measure ( shortly , realignment ) @xmath69 @xcite .
it has been shown that if a bipartite density matrix @xmath1 is separable , then , for the trace norm of its realignment @xmath70 ( defined element - wise as @xmath71 ) , we should have @xmath72 .
it has been shown that in some cases , even if @xmath73 , the positivity of @xmath74 can be a signature of be @xcite .
nevertheless , the nonpositivity of @xmath69 does not necessarily imply separability @xcite .
realignment together with negativity have been shown to set a lower bound for another widely used entanglement measure , concurrence @xmath75 @xcite .
more specifically , it has been proven that for an @xmath76 ( @xmath77 ) bipartite quantum state @xmath1 , we have @xmath78 , where @xmath79}$ ] @xcite . moreover , realignment has been recently used in devising a scheme for directly measuring entanglement of arbitrary states @xcite .
recently , a criterion more powerful than realignment for @xmath68-partite systems ( when @xmath68 is even ) has been proposed @xcite . based on this criterion , which provides a sufficient condition for the separability ,
a sparable bipartite state @xmath1 is the one for which the relation @xmath80 is satisfied .
here , @xmath81 and @xmath82 denote the reduced density matrices associated with the subsystems of @xmath1 .
since in the special examples we study in the next section @xmath83 and @xmath84 are equal up to the multiplicative factor @xmath85 , in the rest of the paper we make use of @xmath83 as our main entanglement measure .
in the following , we choose three different bipartite bound entangled states , and construct their corresponding hamiltonians ( all spin - one chains ) as in sec .
[ consth ] .
our goal is to find qpts where , for some parameter regions , a transition from be to fe ( or vice versa ) or from be to separability ( or vice versa ) may occur .
consider a @xmath86 bipartite system , consisting of two spin-1 particles ( each of which can have spins @xmath87 , with the corresponding orthonormal states @xmath88 ) , whose joint density matrix @xmath89 has the following form @xcite : @xmath90 where @xmath91 $ ] .
this state , for all the parameter interval , has been shown to have ppt .
thus , negativity naturally fails to capture its entanglement , whereas closer inspection by the range criterion @xcite shows be for @xmath92 and separability for @xmath93 . to proceed with the construction of sec .
[ consth ] , we need the following nonzero eigenvalues and the corresponding eigenvectors of @xmath89 : @xmath94 , \label{eighor}\end{aligned}\ ] ] and @xmath95/\sqrt{z},\nonumber \label{vec}\end{aligned}\ ] ] where @xmath96 of example i , where @xmath97 and @xmath98.,width=283 ] since in this example nullity of @xmath99 is zero which is also the case for the next examples
we increase the range of the interaction to three , i.e. , @xmath100 , and calculate the null eigenvectors of @xmath101 .
this density matrix has @xmath102 null eigenvectors , from which only the following ones depend on @xmath52 : @xmath103 where @xmath104 , @xmath105 ( up to a normalization factor ) , @xmath106 ( also up to normalization ) , @xmath107 , and @xmath108 . here ,
@xmath109 and @xmath110 are the null eigenvectors of @xmath13 .
the other vectors in the null eigenspace can be rewritten as @xmath111 , @xmath112 , and @xmath113 . since
our aim is to study qpts relative to the external parameter @xmath52 , the relevant vectors in the null eigenspace of @xmath114 are the ones in eq .
( [ nulls ] ) . to have more symmetry , we construct the local hamiltonian only from ( the normalized ) @xmath115 and @xmath116 , with the same coupling strengths .
that is , @xmath117 ( for odd @xmath7 ) , hence @xmath118 , or more explicitly as the following : @xmath119{s^z_{i+2}}^2 , \label{hhor}\end{aligned}\ ] ] where @xmath120 , and @xmath121 is the spin-1 operator , and @xmath122 . and @xmath123 ( dashed line ) for example i. note that @xmath69 reaches its maximum at @xmath124.,width=275,height=154 ] we now use the gsf @xmath65 to find possible qpt points ( i.e. , criticalities ) of this model ( fig .
[ entfidhor ] ) .
there are two points at which @xmath65 shows nonanalyticity : @xmath125 and @xmath126 .
direct calculation shows that at both of these critical points , the behavior of the gsf for even and add @xmath68 is different : @xmath127/9 $ ] and @xmath128/11 $ ] ( notice the order of this limit taking ) . at @xmath129 the density matrix ( [ dehor ] )
is separable ; i.e. , this critical point corresponds to the transition from be to separability in the tsrdm of the gs of the system , eq .
( [ dehor ] ) , which , according to fig .
[ reahor ] , can also be identified by @xmath69 . moreover , from eq .
( [ dg ] ) , it is simple to find that at @xmath129 , @xmath130 and @xmath131 ; hence , @xmath132 and @xmath133 become separable and maximally entangled states , respectively . besides , here , @xmath134 and @xmath135 . on the other hand , @xmath1 at @xmath136
but , as clearly shown in fig
. [ reahor ] , this criticality can not be captured by @xmath69 ( or its derivatives ) . upon closer inspection
, the behavior of the gsf at very large ( small ) @xmath68 ( @xmath61 ) reveals another possibly critical point in this model at @xmath137 .
since @xmath1 becomes separable at this point , this criticality corresponds to the transition from separability to be in the tsrdm of the gs .
consider the following family of three - parameter @xmath86 states : @xmath138 where @xmath139 , @xmath140 ( @xmath141 ) , and @xmath142 .
the entanglement properties of @xmath143 have been studied , by using a geometric entanglement witness , in ref .
@xcite , and @xmath144 appears to have ppt be for @xmath145 .
of example ii [ eq . ( [ simstate ] ) ] vs @xmath52 , for @xmath146 and @xmath147 and @xmath148 from top to bottom.,width=268,height=162 ] to enable the method introduced earlier for the construction of the hamiltonian @xmath2 , it is required that the tsrdm @xmath1 have at least one null eigenvector that depends on a ( set of ) parameter ( parameters ) .
note , however , that since the eigenvectors of @xmath143 are parameter independent , this state can not be used as @xmath1 for our mechanism .
instead , we consider the following modification : @xmath149+(1-p)~{|\phi^+\rangle}{\langle \phi^+|}. \label{simstate}\end{aligned}\ ] ] the parameters must be chosen such that the positivity of the state @xmath1 is guaranteed .
also , by suitably choosing the parameter @xmath150 , we can make some of the eigenvalues of @xmath1 vanish .
in addition , introducing the parameter @xmath151 is another way by which we can adjust the amount of entanglement of @xmath1 so that a criticality occurs in our desired point .
because of the special form of @xmath144 , the eigenvectors of @xmath1 , apart from a dependence on the sign of @xmath152 , do not depend on the absolute values of @xmath151 and @xmath150 .
this means that the parent hamiltonian @xmath2 has a continuous degeneracy with respect to the parameters @xmath151 and @xmath150 . in the rest of the discussion , we assume @xmath153 .
there are two proper choices for the parameter @xmath150 : @xmath154 $ ] , either of which can make three eigenvalues of @xmath1 vanish . for @xmath155 ,
the sets of three null eigenvectors @xmath156 are @xmath157 in which @xmath158/(2 \alpha -\beta)$ ] and @xmath159 .
note that when we choose @xmath155 , the set @xmath160 corresponds to nonzero eigenvalues @xmath161
+ 2\pm3q~\text{sgn}(q)\sqrt{(\beta -2 \alpha ) ^2+a^2}\}/18 $ ] .
the other eigenvalues of @xmath1 , which we require to be nonzero ( otherwise the degeneracy of the parent hamiltonian would increase ) , are @xmath162p+9\}/9 $ ] and @xmath163/9 $ ] , corresponding to the following eigenvectors : @xmath164 .35 cm and @xmath69 of example ii - a vs @xmath52.,title="fig:",width=272 ] also note that , irrespective of the values of the parameters in eq . ( [ simstate ] ) , @xmath12 ( the tsrdm of sites @xmath7 and @xmath165 ) , for even @xmath7 has an empty null eigenspace .
thus , as explained in the previous example , we consider @xmath114 for the construction of the hamiltonian @xmath2 .
the null eigenspace of @xmath114 is spanned by @xmath166 vectors , from which only the following six are parameter dependent : @xmath167 where @xmath168 correspond to @xmath169 .
the other null eigenvectors are @xmath170 and @xmath171 . for odd @xmath7 , we write the local hamiltonian as @xmath172 and obtain the global hamiltonian @xmath173{(\openone-{s^z}^2)_{i+2}},\label{sb'sh}\end{aligned}\ ] ] where @xmath174 @xmath2 represents a three - parameter family of hamiltonians whose parameters are determined so that the positivity of @xmath1 is guaranteed . in the following , we study the entanglement properties of the gs of this hamiltonian in the vicinity of its critical points , for two special cases
of example ii - b vs @xmath52.,width=264,height=162 ] when @xmath175 , and @xmath176 [ i.e. , @xmath177 and @xmath178 , @xmath143 is reduced to another bound entangled state @xcite @xmath179/7 $ ] , in which @xmath180 , and @xmath181 is the swap of @xmath182 . for @xmath183
$ ] , @xmath184 has be .
let us assume that @xmath185 and @xmath186 .
the parent hamiltonian then will be @xmath2 in eq .
( [ sbsh ] ) with superscript @xmath187 .
the gs ( [ e3 ] ) is written with the set @xmath188 and its corresponding eigenvalues , @xmath189 , plus eigenvalues @xmath190 and eigenvectors of eqs .
( [ non - z ] ) .
in this way , the gsf ( [ gsf ] ) can be obtained readily .
figure [ simgsf ] shows the finite - size behavior of @xmath65 vs @xmath52 , for @xmath146 and different @xmath68 .
as is seen , @xmath65 has a minimum at @xmath191 for all values of @xmath68 .
when @xmath68 increases , @xmath65 decreases ; however , the rate of decreasing at @xmath191 ( @xmath192 ) is maximum
. this behavior can be a sign of criticality in the system . at this point ,
@xmath193 and negativities of @xmath194 , @xmath195 , and @xmath196 attain their maximum values .
on the other hand , as seen in fig .
[ simneg ] , @xmath191 is in the region where @xmath123 vanishes ; hence , @xmath197 fails to signal criticality in the system . from the positivity of @xmath198 around @xmath57 ,
it is evident that @xmath199 is the point where the bipartite entanglement of the gs experiences transition from be to separability@xmath69 detects this critical point .
furthermore , despite the nonanalyticity of the first derivative of @xmath69 at points @xmath1 goes from the fe to the be region , there is no critical behavior in @xmath65 at these points .
this implies that the transition from be to fe can not be responsible for qpt in this system
. 1.2 mm and @xmath69 of example ii - b vs @xmath52 .
inset : the first derivative of @xmath69 with respect to @xmath52.,title="fig:",width=253,height=154 ] suppose that @xmath200 , @xmath201 , @xmath202 , and @xmath203 . according to our previous explanations , the hamiltonian @xmath2 ( the gs @xmath6 )
is constructed from the set @xmath188 [ @xmath204 and its corresponding eigenvalues @xmath205 , together with the eigenvalues @xmath206 , and the eigenvectors ( [ non - z ] ) ] .
thus , the global hamiltonian of this model will be @xmath2 in eq .
( [ sbsh ] ) with superscript @xmath207 . in fig .
[ fs1 ] , @xmath208 has been plotted in terms of @xmath52 for different values of @xmath68 .
there exist two points at which this system may exhibit criticality . at @xmath209 , @xmath208 becomes nonanalytic .
indeed , for arbitrary values of @xmath151 at this point , we have @xmath210 p\}/11 $ ] . at @xmath129 , since @xmath211 , then @xmath212 .
therefore , vectors @xmath213 in eq .
( [ vs ] ) at @xmath129 are completely factorized .
furthermore , one should note that at this point , in the global hamiltonian @xmath2 , @xmath214 and then @xmath215 goes to zero .
as clearly shown in fig .
[ ss10 ] , this probable qpt occurs when @xmath1 is completely free entangled , and of course negativity does not show it ( note that this result is due to the existence of a trace norm in the definition of negativity since the first derivative of one of the eigenvalues of @xmath216 has a discontinuity at @xmath217 and therefore can signal it ) .
another possible qpt point of this system is @xmath218 . as the behavior of @xmath208 in fig .
[ fs1 ] shows , the rate of the drop in @xmath65 at this point is maximum , which may signal a qpt in the system . the
interesting fact is that this critical point corresponds to a change in the type of entanglement of the tsrdm of the gs .
in fact , since @xmath219 is the initial point of the range at which @xmath220 is zero and @xmath83 is positive ( see fig .
[ ss10 ] ) , it represents a transition from fe to be in the two - site states . as fig .
[ ss10 ] shows , the first derivative of @xmath83 at @xmath136 is discontinuous , implying that it detects this criticality . as the final example , consider the following @xmath221 density matrix : @xmath222 in which @xmath223 3.6 mm and @xmath69 of example iii vs @xmath52.,title="fig:",width=283,height=151 ] after some straightforward algebra , we find the following nonzero eigenvalues and their corresponding eigenvectors : @xmath224/(1+a),\nonumber\\ & { |v_3\rangle}=\bigl[a(a+1){|11\rangle}+b^2{|00\rangle}-a{|\bar{1}\bar{1}\rangle}\bigr]/b^2,\\ & { |v_4\rangle}={|1\bar{1}\rangle},~{|v_5\rangle}={|01\rangle},~{|v_6\rangle}={|0\bar{1}\rangle},\nonumber\\ & { |v_7\rangle}={|\bar{1}1\rangle},~{|v_8\rangle}={|\bar{1}0\rangle}.\nonumber\end{aligned}\ ] ] for the same reason as in the previous examples , i.e. , the nullity of @xmath37 constructed from eq .
( [ e3 ] ) is zero , we work with @xmath101 to construct the hamiltonian .
the null eigenspace of @xmath114 is spanned by @xmath225\otimes{|\imath\rangle},\nonumber\\ \hskip-6mm&{|w'\rangle}=\bigl[-(\frac{1+a}{a}){|11\rangle}+{|00\rangle } + \frac{1}{a}{|\bar{1}\bar{1}\rangle}\bigr]\otimes|0\rangle,\label{sahnull}\\ \hskip-6mm&{|u\rangle}=(a'|11\rangle+|\bar{1}\bar{1}\rangle)\otimes|0\rangle,\nonumber\end{aligned}\ ] ] together with @xmath226 , @xmath227 , @xmath228 , and @xmath229 , where @xmath230 and @xmath231 .
we consider the local hamiltonian as @xmath232 , for odd @xmath7 , hence @xmath233\nonumber\\ & + j_2({s^z_i}^2{s^z_{i+1}}^2+s^z_{i } s^z_{i+1})+ j_3({s^z_i}^2{s^z_{i+1}}+ { s^z_i}{s^z_{i+1}}^2)\bigr\}\nonumber\\&\times(\openone-{s^z}^2)_{i+2},\end{aligned}\ ] ] in which @xmath234 and @xmath235 . of example
iii vs @xmath52 , for @xmath236 and @xmath237.,width=268,height=162 ] figure [ sa - realign ] shows the behavior of negativity @xmath197 and realignment @xmath69 vs @xmath52 . at @xmath238 $ ]
, @xmath197 remains zero , whereas @xmath239 is negative ( with a nonanalyticity at @xmath240 ) .
as explained before , although for a separable state @xmath69 is nonpositive , the converse is not necessarily true .
we could not conclude whether any be can exist in this specific region of @xmath52 because we do not know of any other measure stronger than @xmath241 ( whose behavior as explained in sec .
[ orderparameter]is the same as @xmath83 up to the multiplicative factor @xmath85 ) for detecting be .
we , nonetheless , show that a criticality occurs in this region exactly at the nonanalyticity of @xmath69 .
figure [ sa - realign2 ] depicts the gsf ( [ gsf ] ) of this model , showing a qpt at @xmath240 . at this point ,
@xmath242/5 $ ] , and nonanalyticity at @xmath57 appears for odd @xmath68s . finally , we remark that we investigated the behavior of the single- and two - site fidelity ( susceptibility ) of our examples around their critical points .
the results show that in most of the cases , the total behavior of these local order parameters are similar to that of their global counterparts , hence they are also capable of detecting quantum critical points associated with our models . for example , the single - site fidelity in example i shows a sharp drop at @xmath129 and the two - site fidelity ( susceptibility ) in example ii - a ( example ii - b ) can identify all the qpt points .
nevertheless , there exist some critical points , e.g. , @xmath136 in example i , for which these local measures do not herald quantum criticality .
here we have investigated the role of bound entanglement in quantum phase transitions . to this aim ,
we have presented a method to construct quantum spin - chain hamiltonians in which a quantum phase transition can be accompanied by a change in the type of entanglement ( of the two - site reduced density matrices of the corresponding ground states ) from bound to free or separability .
this method _ per se _ is fairly general and can be applied to various scenarios . given a form for the desired two - site reduced density matrices ( with their engineered entanglement properties ) , by utilizing their spectral properties , we have outlined a reverse construction for a compatible pure state for the whole chain . additionally , we have suggested a method to obtain a class of hamiltonians for which the constructed pure state is a ground state . at zero temperature , then , by varying the hamiltonian parameters ( carried over from the very parameter dependence of the given reduced density matrices ) , a desired quantum phase transition at some engineered points can take place . to be able to detect possibly relevant criticalities ( due to the underlying entanglement - type changes ) ,
we have used the ground - state fidelity and the realignment criterion .
one direction for extending our results is to delve into a possible connection between our construction of hamiltonians and the methods of quantum inverse scattering @xcite .
although these methods have clear distinctions , there might be a rich common framework through which one can theoretically engineer quantum models with prescribed entanglement properties for ground states .
overall , our study may hopefully spur further interest in bound entanglement and the role it may play in quantum phase transitions or other areas in which a potential application or advantage could be anticipated .
for example , quantum communications @xcite and controllable generation of entanglement in engineered physical systems @xcite could be possible beneficiaries .
comments by h. johannesson , v. karimipour , and d. a. lidar are gratefully acknowledged .
g. alber , t. beth , m. horodecki , p. horodecki , r. horodecki , m. rttler , h. weinfurter , r. f. werner , and a. zeilinger , _ quantum information : an introduction to basic theoretical concepts and experiments _ , springer tracts in modern physics vol .
173 ( springer , new york , 2001 ) .
m. horodecki and p. horodecki , phys .
a * 59 * , 4206 ( 1999 ) ; d. p. divincenzo , p. w. shor , j. a. smolin , b. m. terhal , and a. v. thapliyal , _ ibid . _ * 61 * , 062312 ( 2000 ) ; w. dr , j. i. cirac , m. lewenstein , and d. bru , _ ibid . _ * 61 * , 062313 ( 2000 ) ; .
pankowski , m. piani , m. horodecki , and p. horodecki , e - print arxiv:0711.2613 ( 2007 ) .
|
we investigate quantum phase transitions in which a change in the type of entanglement from bound entanglement to either free entanglement or separability may occur .
in particular , we present a theoretical method to construct a class of quantum spin - chain hamiltonians that exhibit this type of quantum criticality . given parameter - dependent two - site reduced density matrices ( with prescribed entanglement properties )
, we lay out a reverse construction for a compatible pure state for the whole system , as well as a class of hamiltonians for which this pure state is a ground state .
this construction is illustrated through several examples .
|
stars form exclusively in molecular clouds , although the question whether molecular gas is a prerequisite or a byproduct of star formation is currently under debate ( e.g. , glover & clark 2012 ; kennicutt & evans 2012 ; krumholz 2012 ) . in either case ,
accurate measuruments of the physical properties of molecular clouds are critical to constrain the initial conditions for star and molecular gas formation . however , obtaining such measurements is hampered by the fact that molecular hydrogen ( h@xmath0 ) , the most abundant molecular species in the interstellar medium ( ism ) , is not directly observed under the typical conditions in molecular clouds . as a homonuclear diatomic molecule , h@xmath0 does not have a permanent electric dipole moment and its ro - vibrational states change only via weak quadrupole transitions . therefore ,
alternative tracers have been employed to infer the abundance and distribution of h@xmath0 .
carbon monoxide ( co ) is one of the most commonly used tracers of h@xmath0 due to its large abundance and low rotational transitions that are readily excited in molecular clouds through collisions with h@xmath0 . in particular , the @xmath12co(@xmath13 ) integrated intensity , @xmath3 , is often used to estimate the h@xmath0 column density , @xmath2(h@xmath0 ) , via the so - called `` @xmath14-factor '' , which is defined by @xmath15 accurate knowledge of @xmath1 is crucial to address some of the fundamental questions in astrophysics .
for example , one of the most intriguing properties of galaxies is a strong power - law relation between the surface density of star formation rate , @xmath16 , and the surface density of h@xmath0 , @xmath17 , generally known as the `` kennicutt - schmidt relation '' ( e.g. , schmidt 1959 ; kennicutt 1989 ; bigiel et al .
2008 ; schruba et al .
2011 ; rahman et al . 2012
; shetty et al .
while this empirical relation provides important insights into the physical process of star formation ( e.g. , a close connection between the chemical or thermal state of the ism and star formation ) , its precise form has been a subject of debate and strongly depends on @xmath1 . from an observational perspective
, @xmath1 is usually adopted as a conversion factor .
its estimate relies on the derivation of @xmath2(h@xmath0 ) using observational methods independent of co ( bolatto et al .
2013 for a review ) .
one of the methods to derive @xmath2(h@xmath0 ) utilizes dust as a tracer of total gas column density .
dust has been observed to be well mixed with gas ( e.g. , boulanger et al .
1996 ) and can be mapped through its emission at far - infrared ( fir ) wavelengths or its absorption at near - infrared ( nir ) wavelengths .
the procedure is to estimate the dust column density or the @xmath18-band extinction , @xmath9 , from the fir emission or the nir absorption ( e.g. , cardelli et al . 1989 ) and to assume or to estimate a dust - to - gas ratio ( dgr ) that linearly relates @xmath9 to the total gas column density @xmath2(h ) = @xmath2(hi ) + 2@xmath2(h@xmath0 ) . the atomic gas column density , @xmath2(hi )
, is then measured from the 21-cm emission and is removed from @xmath2(h ) for an estimate of @xmath2(h@xmath0 ) ( e.g. , israel 1997 ; dame et al .
2001 ; leroy et al .
2007 , 2011 ; lee et al . 2012 ; sandstrom et al . 2013 ) . the derived @xmath2(h@xmath0 ) is finally combined with @xmath3 to estimate @xmath1 .
this procedure has been applied to the milky way and a number of nearby galaxies . for the milky way , dame et al .
( 2001 ) showed that @xmath1 does not change significantly with galactic latitude ( for @xmath19 @xmath4 5@xmath2030@xmath20 ) from the mean value of ( 1.8 @xmath21 0.3 ) @xmath5 10@xmath22 when molecular clouds are averaged over @xmath4kpc scales .
several studies of individual molecular clouds at 3@xmath239@xmath23 angular resolution have estimated similar average @xmath1 values ( e.g. , frerking et al .
1982 for ophiuchus ; lombardi et al . 2006 for pipe ; pineda et al . 2008 for perseus ; pineda et al . 2010 for taurus ; paradis et al . 2012 for aquila - ophiuchus , cepheus - polaris , taurus , and orion ) . at the same time , @xmath1 values different from the galactic mean value have been occasionally found , e.g. , @xmath1 @xmath4 0.5 @xmath5 10@xmath22 for infrared cirrus clouds in ursa major ( de vries et al .
1987 ) and @xmath1 @xmath4 6.1 @xmath5 10@xmath22 for high - latitude clouds ( magnani et al .
1988 ) , suggesting cloud - to - cloud variations in @xmath1 .
rare studies of @xmath1 in spatially resolved molecular clouds have shown some variations as well , e.g. , @xmath1 @xmath4 ( 1.612 ) @xmath5 10@xmath22 for taurus ( pineda et al .
2010 ) and @xmath1 @xmath4 ( 0.91.8 ) @xmath5 10@xmath22 for perseus ( pineda et al .
2008 ) . in studies of nearby galaxies on @xmath4kpc scales ,
@xmath1 values are similar with the galactic mean value and are relatively constant within individual galaxies .
however , systematically smaller and larger @xmath1 values have been found from the central regions of star - forming galaxies ( down to @xmath40.1 @xmath5 10@xmath22 ; e.g. , smith et al .
1991 ; sandstrom et al .
2013 ) and low - metallicity dwarf irregular galaxies ( up to @xmath4130 @xmath5 10@xmath22 ; e.g. , israel 1997 ; leroy et al .
2007 ; gratier et al . 2010 ; leroy et al . 2011 ) , indicating the dependence of the average @xmath1 on interstellar environments . from a theoretical perspective
, @xmath1 has been primarily studied using photodissociation region ( pdr ) models because the majority of the co emission originates from the outskirts of molecular clouds , where the interstellar radiation field ( isrf ) illuminates the cloud ( e.g. , taylor et al .
1993 ; le bourlot et al .
1993 ; wolfire et al .
1993 ; kaufman et al .
1999 ; bell et al . 2006 ; wolfire et al .
for example , bell et al . (
2006 ) used the uclpdr code ( papadopoulos et al .
2002 ) to calculate chemical abundances and emission strengths and showed that @xmath1 changes by more than an order of magnitude with varying depths within molecular clouds .
in addition , they found significant variations in @xmath1 between molecular clouds with a wide range of physical parameters , e.g. , density , metallicity , and cloud age . while the pdr models are limited to simple geometries and density distributions , three - dimensional magnetohydrodynamic ( mhd ) simulations
have been recently performed to investigate @xmath1 in turbulent molecular clouds ( e.g. , glover & mac low 2011 ; shetty et al .
2011a , b ) .
these simulations model chemistry for simple molecules such as h@xmath0 and co as a function of time and show that @xmath1 is not constant within individual clouds .
moreover , @xmath1 in simulations varies over four orders of magnitude between clouds with low densities , low metallicities , and strong radiation fields
. such variability of @xmath1 within resolved clouds and between clouds with different properties predicted by the pdr and mhd models has been rarely found in observations , largely due to the lack of high - resolution observations . in this paper
, we derive @xmath1 for the perseus molecular cloud on sub - pc scales and test two theoretical models of the formation of molecular gas , in an attempt to understand the origins of the variations in @xmath1 and the physical processes of h@xmath0 and co formation . one model is the one - dimensional pdr model originally developed by tielens & hollenbach ( 1985 ) and updated by kaufman et al .
( 2006 ) , wolfire et al .
( 2010 ) , and hollenbach et al .
( 2012 ) . here
we use a further modification of this model , which allows for a two - sided illumination and either a constant density or a simple formulation of the density distribution ( hereafter the modified w10 model ) .
the other model is the three - dimensional mhd model by shetty et al .
( 2011a ) that is based on the modified zeus
mp code described in glover et al .
( 2010 ) ( hereafter the s11 model ) .
there are two primary differences between these two models .
first , the s11 model simulates h@xmath0 and co formation in turbulent molecular clouds by coupling the chemical and dynamical evolution of gas , while the modified w10 model takes into account the impact of turbulence only via a constant supersonic linewidth for spectral line formation and cooling .
second , the s11 model follows the time - dependent evolution of a number of chemical species , including h@xmath0 and co , while the modified w10 model uses a detailed time - independent chemical network that explicitly assumes chemical equilibrium for every atomic and molecular species .
therefore , we consider the modified w10 model and the s11 model as representative `` microturbulent time - independent model '' and `` macroturbulent time - dependent model '' .
our study is one of the first attempts to test the mhd model tracking both the chemical and dynamical evolution of the ism and compare it with a more traditional view of the formation of molecular gas ( pdr model ) .
in addition , considering that small - scale ism models are starting to be implemented in large - scale simulations of galaxy formation and evolution ( e.g. , feldmann et al .
2012a , b ; lagos et al .
2012 ; narayanan et al .
2012 ) , our study will serve as a `` zero point test '' for the models of gas contents in galaxies .
we focus on the perseus molecular cloud because of its proximity and a wealth of multi - wavelength observations .
located at a distance of @xmath4200350 pc ( herbig & jones 1983 ; ernis 1990 ) , perseus has a projected angular size of @xmath46@xmath20 @xmath5 3@xmath20 on the sky . in this paper
, we adopt the distance to perseus of 300 pc . with a mass of @xmath42 @xmath5 10@xmath24 m@xmath25 ( sancisi et al .
1974 ; lada et al .
2010 ) , perseus is considered as a low - mass molecular cloud with an intermediate level of star formation ( bally et al .
the cloud contains a number of dark ( b5 , b1e , b1 , and l1448 ) and star - forming regions ( ic348 and ngc1333 ) shown in figure [ f : av_intro ] . this paper is organized as follows . in section 2 ,
we summarize the results from previous studies highly relevant to our investigation and provide constraints on important physical parameters of perseus . in section 3 , we describe the multi - wavelength observations used in our study . in section 4 , we divide perseus into a number of individual regions and select data points for each region .
we then derive the @xmath1 image ( section 5 ) and investigate the large - scale spatial variations of @xmath1 and their correlations with physical parameters such as the strength of the radiation field and the co velocity dispersion ( section 6 ) .
in addition , we examine how @xmath3 and @xmath1 change with @xmath9 in perseus . in section 7
, we summarize the details of the modified w10 model and the s11 model and compare our observational data with model predictions .
finally , we discuss and summarize our conclusions ( sections 8 and 9 ) .
pineda et al . ( 2008 ) used the @xmath3 and @xmath9 data from the complete survey of star forming regions ( complete ; ridge et al .
2006 ) to investigate @xmath1 in perseus .
they fitted a linear function to @xmath3 vs @xmath9 to estimate @xmath1 and found @xmath1 @xmath4 1.4 @xmath5 10@xmath22 for the whole cloud and a range of @xmath1 @xmath4 ( 0.91.8 ) @xmath5 10@xmath22 for six sub - regions , suggesting a factor of @xmath26 spatial variations of @xmath1 caused by different ism conditions . in the process of performing a linear fit
, they noticed that @xmath1 is heavily affected by the saturation of @xmath3 at @xmath9 @xmath11 4 mag and re - estimated @xmath1 @xmath4 0.7 @xmath5 10@xmath22 from the linear fit only to the unsaturated @xmath3 .
in addition , pineda et al . (
2008 ) compared the observed co and @xmath27co(@xmath13 ) integrated intensities with predictions from the meudon pdr code ( le petit et al .
2006 ) and found that the pdr models reproduce the co and @xmath27co observations reasonably well and the variations among the six sub - regions can be explained by variations in physical parameters , in particular density and non - thermal gas motion .
in our recent study , we derived the @xmath28 and @xmath17 images of perseus on @xmath40.4 pc scales ( section 3.1 ) and investigated how the h@xmath0-to - hi ratio , @xmath29 = @xmath17/@xmath28 = 2@xmath30/@xmath2(hi ) , changes across the cloud ( lee et al .
we found that @xmath28 is relatively uniform with @xmath468 m@xmath25 pc@xmath31 , while @xmath17 significantly varies from @xmath40 m@xmath25 pc@xmath31 to @xmath473 m@xmath25 pc@xmath31 , resulting in @xmath29 @xmath4 010 with a mean of @xmath40.7 . due to the nearly constant @xmath28 , a strong linear relation between @xmath29 and @xmath32 was found .
interestingly , these results are consistent with the time - independent h@xmath0 formation model by krumholz et al .
( 2009 ; hereafter the k09 model ) . in the k09 model , a spherical cloud
is embedded in a uniform and isotropic radiation field and the h@xmath0 abundance is estimated based on the balance between h@xmath0 formation on dust grains and h@xmath0 photodissociation by lyman - werner ( lw ) photons .
the most important prediction of the k09 model is the minimum @xmath28 required to shield h@xmath0 against photodissociation .
this minimum @xmath28 for h@xmath0 formation depends on metallicity ( e.g. , @xmath28 @xmath4 10 m@xmath25 pc@xmath31 for solar metallicity ) but only weakly on the strength of the radiation field .
once the minimum @xmath28 is achieved , additional @xmath28 is fully converted into @xmath17 , resulting in the uniform @xmath28 distribution and the linear increase of @xmath29 with @xmath28+@xmath17 .
we summarize estimates of several important physical parameters of perseus obtained from previous studies .
we will use these parameters in later sections of this paper . * density
* @xmath33 @xmath34 @xmath35 @xmath36 * * : * * young et al . ( 1982 ) estimated @xmath37 @xmath4 ( 1.75 ) @xmath5 10@xmath38 @xmath39 for b5 based on the large velocity gradient ( lvg ) model applied to co and co(@xmath40 ) observations .
bensch ( 2006 ) derived larger @xmath37 @xmath4 ( 330 ) @xmath5 10@xmath38 @xmath39 for the same cloud by comparing pdr models with co , @xmath27co , and [ ci ] observations .
similarly , pineda et al .
( 2008 ) found that pdr models with @xmath37 @xmath4 a few @xmath5 10@xmath41 @xmath39 can reproduce the co and @xmath27co(@xmath13 ) observations of perseus . in summary ,
gas traced by the co emission in perseus is likely to have @xmath42 @xmath39 .
* isrf * @xmath43 @xmath34 * 0.4 * @xmath44 * * : * * lee et al .
( 2012 ) investigated the dust temperature , @xmath45 , across perseus and potential heating sources and concluded that the cloud is embedded in the uniform galactic isrf heating dust grains to @xmath417 k , except for the central parts of ic348 and ngc1333 , where the radiation from internal b - type stars likely dominates . under the assumption that dust grains are in thermal equilibrium
, we can use @xmath45 @xmath4 17 k to estimate the strength of the radiation field by @xmath46 where @xmath47 is the flux at ultraviolet ( uv ) wavelengths and @xmath48 is the size of dust grains ( lequeux 2005 ) . equation ( 2 ) assumes the absorption efficiency @xmath49 = 1 and the dust emissivity index @xmath50 = 2 . for dust grains with @xmath48 @xmath4 0.1 @xmath51 m ,
whose size is comparable to uv wavelengths therefore @xmath49 @xmath52 1 , we estimate @xmath47 @xmath4 1.1 @xmath5 10@xmath53 erg @xmath7 s@xmath8 @xmath4 0.4 @xmath54 ( @xmath54 = the local field measured in the solar neighborhood by draine 1978 @xmath4 2.7 @xmath5 10@xmath53 erg @xmath7 s@xmath8 ) for the uniform isrf incident upon perseus .
the exceptions are the central regions of ic348 and ngc1333 , where the radiation from the b - type stars is dominant .
* cosmic - ray ionization rate * @xmath55 @xmath56 @xmath57 @xmath58 * * : * * there is emerging evidence that @xmath59 likely lies between @xmath410@xmath60 s@xmath8 to @xmath410@xmath61 s@xmath8 with lower values in dense molecular clouds and @xmath410@xmath62 s@xmath8 to @xmath410@xmath61 s@xmath8 in the diffuse ism ( e.g. , dalgarno 2006 ; indriolo & mccall 2012 ; hollenbach et al .
this suggests that @xmath59 could be larger than the canonical @xmath59 @xmath4 10@xmath60 s@xmath8 by a factor of @xmath410100 in the regions where the co emission arises . *
metallicity * @xmath63 @xmath34 @xmath64 @xmath65 * * : * * gonzlez hernndez et al .
( 2009 ) performed a chemical abundance analysis for cernis 52 , a member of ic348 whose spectral type is a3 v , and derived [ fe / h ] = @xmath660.01 @xmath21 0.15 ( corresponding to @xmath67 @xmath4 0.71.4 @xmath68 ) .
in addition , lee et al .
( 2012 ) compared the intensity at 100 @xmath51 m , @xmath69 , with @xmath2(hi ) for perseus and found an overall linear relation .
as @xmath69/@xmath2(hi ) is an approximation of dgr , the fact that a single @xmath69/@xmath2(hi ) fits most of the diffuse regions suggests no significant variation of dgr or @xmath67 across the cloud .
therefore , @xmath67 @xmath4 1 z@xmath25 would be a reasonable estimate for perseus .
note that lee et al .
( 2012 ) derived dgr = @xmath9/@xmath2(h ) @xmath4 1.1 @xmath5 10@xmath70 mag @xmath71 for perseus , which is @xmath42 times larger than the typical galactic dgr @xmath4 5.3 @xmath5 10@xmath72 mag @xmath71 ( bohlin et al . 1978 ) .
* turbulent linewidth * @xmath73 @xmath74 @xmath75@xmath76@xmath77 @xmath78 * * : * * pineda et al .
( 2008 ) compared the co excitation temperature , @xmath79 , with @xmath9 and found that @xmath79 increases from @xmath45 k at @xmath9 @xmath4 2 mag to @xmath420 k at @xmath9 @xmath11 4 mag .
if @xmath37 > @xmath80 @xmath4 10@xmath38 @xmath39 where @xmath80 is the critical density for the co emission , the case likely for the regions with @xmath9 @xmath11 4 mag , we expect that the co emission is in local thermodynamic equilibrium ( lte ) and @xmath79 @xmath4 @xmath81 where @xmath81 is the kinetic temperature .
when we assume @xmath79 @xmath4 @xmath81 @xmath4 20 k for perseus , the mean thermal velocity of co - emitting gas would be @xmath82@xmath83@xmath84 = @xmath85 @xmath4 0.1 km s@xmath8 ( @xmath86 = the boltzmann constant , @xmath51 = the mass of a molecule in amu = 28 for co , @xmath87 = the mass of a hydrogen atom ) .
this @xmath82@xmath83@xmath84 @xmath4 0.1 km s@xmath8 is an order of magnitude smaller than the co velocity dispersion @xmath88 @xmath4 0.92 km s@xmath8 ( corresponding to fwhm = @xmath89 @xmath88 @xmath4 2.14.7 km s@xmath8 ) measured across perseus ( pineda et al .
this suggests that there are most likely contributions from other processes , e.g. , interstellar turbulence , systematic motions such as inflow , outflow , rotation , etc . , and/or multiple components along a line of sight .
for example , b1 and ngc1333 contain a large number of herbig - haro objects that are known to trace currently active shocks in outflows ( e.g. , bally et al .
therefore , not all the observed @xmath88 should be attributed to interstellar turbulence alone . as a result
, we expect @xmath90 to be smaller than the measured fwhm of @xmath425 km s@xmath8 .
* cloud age * @xmath91 @xmath34 @xmath92 @xmath93 * * : * * for ic348 , muench et al .
( 2003 ) derived a mean age of @xmath42 myr with a spread of @xmath43 myr using published spectroscopic observations .
however , there are some indications for the existence of older stars in ic348 . for example , herbig ( 1998 ) found that h@xmath94 emission line stars in ic348 have an age spread from @xmath40.7 myr to @xmath412 myr .
a similar spread in stellar age , from @xmath40.5 myr to @xmath410 myr , has been found by luhman et al .
( 1998 ) from their infrared and optical spectroscopic observations .
considering this duration of star formation in ic348 , @xmath95 @xmath4 10 myr would be a reasonable age estimate for perseus .
we use the @xmath2(h@xmath0 ) image derived in our recent study , lee et al .
we used the 60 @xmath51 m and 100 @xmath51 m data from the improved reprocessing of the _ iras _ survey ( iris ; miville - deschnes & lagache 2005 ) to derive the dust optical depth at 100 @xmath51 m , @xmath97 .
dust grains were assumed to be in thermal equilibrium and the contribution from very small grains ( vsgs ) to the intensity at 60 @xmath51 m was accounted for by calibrating the derived @xmath45 image with the @xmath45 data from schlegel et al .
the @xmath97 image was then converted into the @xmath9 image by finding the conversion factor @xmath14 for @xmath98 that results in the best agreement between the derived @xmath9 and complete @xmath9 .
this calibration of @xmath97 to complete @xmath9 was motivated by goodman et al .
( 2009 ) , who found that dust extinction at nir wavelengths is the best probe of total gas column density .
finally , lee et al . (
2012 ) estimated a local dgr for perseus and derived the @xmath2(h@xmath0 ) image in combination with the hi data from the galactic arecibo l - band feed array hi survey ( galfa - hi ; peek et al .
the hi emission was integrated from @xmath665 km s@xmath8 to @xmath9915 km s@xmath8 , the range that maximizes the spatial correlation between the hi integrated intensity and the dust column density , and @xmath2(hi ) was calculated under the assumption of optically thin hi .
the derived @xmath2(h@xmath0 ) has a mean of @xmath41.3 @xmath5 10@xmath22 @xmath7 and peaks at @xmath44.5 @xmath5 10@xmath100 @xmath7 .
its mean 1@xmath101 uncertainty is @xmath43.6 @xmath5 10@xmath6 @xmath7 .
see section 4 of lee et al .
( 2012 ) for details on the derivation of the @xmath2(h@xmath0 ) image and its 1@xmath101 uncertainty .
the @xmath45 , @xmath9 , @xmath2(hi ) , and @xmath2(h@xmath0 ) images derived by lee et al .
( 2012 ) are all at 4.3@xmath23 angular resolution , corresponding to @xmath40.4 pc at the distance of 300 pc .
we present the @xmath2(h@xmath0 ) image at 4.3@xmath23 angular resolution in figure [ f : h2_cden_display ] .
the blank data points correspond to point sources and regions with possible contamination ( the taurus molecular cloud and a background hii region ) .
see sections 4.2 and 4.3 of lee et al .
( 2012 ) for details .
we use the complete co data cube obtained with the 14-m fcrao telescope ( ridge et al .
this cube covers the main body of perseus with a spatial area of @xmath46@xmath20 @xmath5 3@xmath20 at 46@xmath102 angular resolution .
we correct the co data for the main - beam efficiency of 0.45 , following ridge et al .
( 2006 ) and pineda et al .
the rms noise per channel ranges from @xmath40.3 k to @xmath43.5 k with a mean of @xmath40.8 k. we show the average co spectrum for perseus in figure [ f : co_spectrum_perseus ] . to produce this spectrum
, we average the spectra of all data points where the ratio of the peak main - beam brightness temperature to the rms noise is greater than 3 .
the co emission is clearly contained between @xmath103 km s@xmath8 and @xmath9915 km s@xmath8 and shows multiple velocity components .
to derive @xmath3 , we integrate the co emission from @xmath103 km s@xmath8 to @xmath9915 km s@xmath8 , the range where lee et al .
( 2012 ) found the hi emission associated with perseus , with a spectral resolution of 0.064 km s@xmath8 . at 46@xmath102 angular resolution ,
the derived @xmath3 ranges from @xmath6619.9 k km s@xmath8 to 116.6 k km s@xmath8
. its mean 1@xmath101 uncertainty is @xmath40.9 k km s@xmath8 .
we note that some data points in the co cube are affected by an artificial absorption feature at @xmath104 @xmath4 7.5 km s@xmath8 .
this artifact is due to the contaminated off - position and is responsible for a number of blanked data points in figure [ f : co_vel_dispersion ] that do not correspond to point sources and regions with possible contamination .
we find that this artifact does not affect our estimate of @xmath3 .
as pointed out by pineda et al .
( 2008 ) and lee et al .
( 2012 ) , there are considerable region - to - region variations in physical parameters across perseus .
we therefore divide the cloud into five regions and perform analyses mainly on the individual regions . to define the individual regions ,
we draw the complete @xmath3 contours from 4 k km s@xmath8 ( 5% of the peak ) to 72 k km s@xmath8 ( 90% of the peak ) with 4 k km s@xmath8 intervals and use the contours to determine the boundaries of each region . note that the minimum @xmath3 of 4 k km s@xmath8 for the regional boundaries does not mean that there is no data point with @xmath3 < 4 k km s@xmath8 .
in addition , we select data points that have ( 1 ) @xmath103 km s@xmath8 < co velocity centroid < @xmath9915 km s@xmath8 , ( 2 ) @xmath3 > 0 k km s@xmath8 , and ( 3 ) @xmath2(h@xmath0 ) > 0 @xmath7 .
these criteria are to select data points that are reliable and kinematically associated with perseus . applying these criteria results in 1160 independent data points and
all except three data points have s / n > 1 for both @xmath3 and @xmath2(h@xmath0 ) .
we show the selected data points for each region ( b5 , b1e / b1 , l1448 as dark regions and ic348 , ngc1333 as star - forming regions ) with a different color in figure [ f : complete_five_regions ] .
the individual regions have an average size of @xmath457 pc at the distance of 300 pc ( table 1 ) .
we derive the @xmath1 image at 4.3@xmath23 angular resolution by applying equation ( 1 ) to the @xmath2(h@xmath0 ) image and the complete @xmath3 image ( smoothed to match the angular resolution of the @xmath2(h@xmath106 ) image ) on a pixel - by - pixel basis ( figure [ f : complete_x_co ] ) .
for the five regions defined in section 4 , @xmath1 ranges from @xmath45.7 @xmath5 10@xmath107 to @xmath44.4 @xmath5 10@xmath100 . while @xmath1 shows a substantial range , most data points ( @xmath480% ) have 10@xmath6 < @xmath1 < 10@xmath22 . summing both @xmath2(h@xmath0 ) and @xmath3 over all five regions results in an average @xmath108 = @xmath109@xmath2(h@xmath0)/@xmath109@xmath3 @xmath4 3 @xmath5 10@xmath6 . applying a single criterion of @xmath2(h@xmath0 ) > 0 @xmath7 to the whole cloud to include the regions with h@xmath0 but without co detection results in the same average @xmath108 @xmath4 3 @xmath5 10@xmath6 . the 1@xmath101 uncertainty of @xmath1 is derived based on the propagation of errors ( bevington & robinson 2003 ) and its mean value is @xmath41.6 @xmath5 10@xmath6 .
figure [ f : complete_x_co ] shows interesting spatial variations of @xmath1 across perseus . to quantify these variations
, we estimate the @xmath108 values for the dark and star - forming regions by summing @xmath2(h@xmath0 ) and @xmath3 over each region ( table 1 ) .
we find a factor of @xmath43 decrease in @xmath1 from the northeastern regions ( b5 and ic348 ) where @xmath108 @xmath4 6 @xmath5 10@xmath6 to the southwestern regions ( b1e / b1 , ngc1333 , and l1448 ) where @xmath108 @xmath4 2 @xmath5 10@xmath6 .
our result is consistent with pineda et al .
( 2008 ) in that both studies found regional variations of @xmath1 across perseus . however , while they estimated a single @xmath1 for each sub - region , we derived the spatial distribution of @xmath1 .
based on this distribution , we investigate large - scale trends in several physical parameters and their possible connections with the variations of @xmath1 .
we first derive the @xmath88 image using the complete co data cube .
figure [ f : co_vel_dispersion ] shows that the southwestern part has systematically larger @xmath88 than the northeastern part .
for example , @xmath470% of the data points in the southwestern part have @xmath88 > 1.5 km s@xmath8 , while @xmath440% of the data points in the northeastern part have @xmath88 > 1.5 km s@xmath8 .
in particular , b1e / b1 and ngc1333 have the largest median @xmath88 @xmath4 2 km s@xmath8 compared to other regions where the median @xmath88 is @xmath41.3 km s@xmath8 ( table 1 ) .
the large @xmath88 in the southwestern part could be caused by more complex velocity structure and/or multiple components along a line of sight .
in addition , outflows from embedded protostars could contribute to broaden the co spectra .
for example , b1 and ngc1333 have many herbig - haro objects identified from the surveys of h@xmath94 and [ sii ] emission , which trace currently active shocks in outflows ( e.g. , bally et al .
the @xmath45 image derived by lee et al .
( 2012 ) also shows systematic variations across perseus .
specifically , @xmath45 slightly decreases toward the southwestern part .
this is consistent with pineda et al .
( 2008 ) , who found @xmath45 @xmath4 17 k for b5/ic348 and @xmath45 @xmath4 16 k for b1e / b1/ngc1333 . to investigate the variations of isrf
, we evaluate @xmath47 using equation ( 2 ) and assess its distributions for b5/ic348 ( _ east _ ) and b1e / b1/ngc1333/l1448 ( _ west _ ) .
figure [ f : isrf_histo ] shows the median @xmath47 of _ east _ ( @xmath410@xmath110 erg @xmath7 s@xmath8 ) as a dashed line .
we find that @xmath450% and @xmath42% of the data points have @xmath47 > 10@xmath110
erg @xmath7 s@xmath8 for _ east _ and _ west _ respectively .
this suggests that @xmath47 systematically decreases toward the southwestern part of perseus .
however , the variation of @xmath47 is very mild : the median @xmath47 decreases from _ east _ to _ west _ by only a factor of @xmath41.4 .
this result does not change even when we examine the median @xmath47 for each dark and star - forming region .
finally , lee et al . ( 2012 ) noticed a considerable difference between the northeastern and southwestern parts of perseus regarding the relative spatial distribution of h@xmath0 and co. they estimated the fraction of `` co - dark '' h@xmath0 , which refers to interstellar gas in the form of h@xmath0 along with ci and cii but little or no co , and found a factor of @xmath43 decrease in the fraction toward the southwestern part .
in other words , `` co - free '' h@xmath0 envelopes exist in the northeastern part , while co traces h@xmath0 reasonably well in the southwestern part ( e.g. , figure 14 of lee et al .
this suggests that h@xmath0 takes up a larger volume than co in the northeastern region , which could result in larger @xmath1 .
many theoretical studies have shown that @xmath1 can vary over several orders of magnitude with changes in density , metallicity , turbulent linewidth , isrf , etc .
( e.g. , maloney & black 1988 ; le bourlot et al .
1993 ; wolfire et al .
1993 ; sakamoto 1996 , 1999 ; kaufman et al .
1999 ; bell et al . 2006 ; glover & mac low 2011 ; shetty et al .
2011a , b ) , suggesting that various physical parameters play a role in determining @xmath1 .
this likely applies to perseus as well .
while @xmath88 and @xmath47 show some interesting variations across the cloud , their correlations with @xmath1 are not strong ( spearman s rank correlation coefficient @xmath111 = @xmath112 and 0.6 respectively ; the null hypothesis is rejected at the 99% two - tailed confidence level ) . in addition , as we will show in comparison with the modified w10 model ( section 7.1 ) , changes in density appear to contribute to the observed variations in @xmath1 as well .
it is most likely , therefore , that combinations of changes in density , turbulent linewidth , isrf , and possibly other parameters we do not test in our study result in the variations in @xmath1 across the cloud .
this conclusion is consistent with pineda et al .
( 2008 ) , who suggested that local variations in density , non - thermal gas motion , and isrf can explain the observed scatter of @xmath1 among the sub - regions in perseus . because @xmath1 depends on many properties of the ism , constraining physical conditions by matching models to the observed value of @xmath1 requires a search through a large parameter space .
nevertheless , from a theoretical standpoint , @xmath1 has an interesting characteristic dependance on @xmath9 ( e.g. , taylor et al .
1993 , bell et al . 2006 ; glover & mac low 2011 ; shetty et al .
2011a ; feldmann et al .
we focus on investigating this characteristic dependance over a broad range of @xmath9 by comparing our observations with two different theoretical models , with an aim of understanding the important physical processes of h@xmath0 and co formation .
to do so , we use the models with a simple set of input parameters reasonable for perseus and focus mainly on the general trends of @xmath2(hi ) , @xmath2(h@xmath0 ) , @xmath3 , and @xmath1 with @xmath9 . to understand how @xmath1 varies with @xmath9 , we begin by plotting @xmath3 as a function of @xmath9 in figure [ f : co_int_av_plot_perseus ] for all five regions defined in section 4 .
we use the @xmath9 image at 4.3@xmath23 angular resolution derived by lee et al .
even though there is a large amount of scatter , several important features are noticeable .
first , there appears to be some threshold @xmath115 @xmath4 1 mag below which no co emission is detected .
the sharp increase of @xmath3 with @xmath9 found from the individual regions ( section 6.2.2 ) strongly supports the existence of such threshold .
this may suggest that co becomes shielded against photodissociation at @xmath9 @xmath4 1 mag in perseus .
previous observations of molecular clouds have found the similar @xmath115 @xmath4 1 mag ( e.g. , lombardi et al .
2006 ; pineda et al . 2008
; leroy et al . 2009 ) .
note that a lack of co detection at @xmath9 @xmath10 1 mag is not due to our sensitivity , considering that our mean 1@xmath101 uncertainty of @xmath9 is @xmath40.2 mag .
in addition , the threshold is not the result of the limited spatial coverage of the complete @xmath3 image .
we made a comparison between our @xmath9 image and the @xmath3 image with a large spatial area of @xmath410@xmath20 @xmath5 7@xmath20 from the center for astrophysics co survey ( cfa ; dame et al .
2001 ) at the common angular resolution of 8.4@xmath23 and found essentially the same threshold .
second , @xmath3 significantly increases from @xmath40.1 k km s@xmath8 to @xmath470 k km s@xmath8 for a narrow range of @xmath9 @xmath4 13 mag .
this steep increase of @xmath3 may suggest that the transition from cii / ci to co is sharp once shielding becomes sufficiently strong to prevent photodissociation ( e.g. , taylor et al .
1993 ; bell et al .
2006 ) .
third , @xmath3 gradually increases and saturates to @xmath45080 k km s@xmath8 at @xmath116 @xmath11 3 mag .
this is consistent with pineda et al .
( 2008 ) , who found @xmath117 @xmath4 4 mag for perseus .
similarly , lombardi et al . (
2006 ) found the saturation of @xmath3 @xmath4 30 k km s@xmath8 for the pipe nebula at @xmath116 @xmath4 6 mag ( with their adopted relation @xmath9 = @xmath118/0.112 ) .
the saturation of @xmath3 is expected based on the relation between @xmath3 and @xmath119 , @xmath3 @xmath120 @xmath121 , where @xmath119 @xmath120 @xmath9 .
therefore , @xmath3 does not faithfully trace @xmath9 once it becomes optically thick .
the presence of optically thick co emission in perseus was hinted by pineda et al .
( 2008 ) , who performed the curve of growth analysis for the co and @xmath27co(@xmath122 = 1 @xmath123 0 ) observations . in agreement with pineda et al .
( 2008 ) , we find that the relation between @xmath3 and @xmath9 has significant region - to - region variations across perseus , contributing to the large scatter in figure [ f : co_int_av_plot_perseus ] .
we therefore show @xmath3 vs @xmath9 for each dark and star - forming region in figure [ f : co_int_av_plot ] . to emphasize the steep increase and saturation of @xmath3 , we plot @xmath3 as a function of @xmath9 on a log - log scale . among the five regions ,
b5 and l1448 have the narrowest range of @xmath9 @xmath4 13 mag , simply reflecting their smaller @xmath2(h ) range on average . on the other hand , ic348 has the largest ranges of @xmath9 @xmath4 111 mag and @xmath3 @xmath4 0.250 k km s@xmath8 .
@xmath3 steeply increases from @xmath40.2 k
km s@xmath8 to @xmath435 k km
s@xmath8 at @xmath9 @xmath4 13 mag and then saturates to @xmath450 k km s@xmath8 at @xmath9 @xmath11 3 mag . in the case of b1e / b1 , two components are apparent .
the first component corresponds to the relatively steep increase of @xmath3 from @xmath41 k km s@xmath8 to @xmath420 k km s@xmath8 at @xmath9 @xmath4 1.53 mag
. the second component corresponds to the gradual increase of @xmath3 from @xmath420 k km s@xmath8 to @xmath460 k km s@xmath8 at @xmath9 @xmath4 1.55 mag .
considering the two components together , @xmath3 saturates to @xmath460 k km s@xmath8 at @xmath9 @xmath11 3 mag .
lastly , ngc1333 has the majority of the data points ( @xmath490% ) at @xmath9 @xmath10 3 mag .
@xmath3 increases from @xmath40.5 k
km s@xmath8 to @xmath470 k km s@xmath8 at @xmath9 @xmath4 13 mag and then shows a hint of the saturation to @xmath480 k km s@xmath8 at @xmath9 @xmath11 3 mag .
note that ngc1333 is the region where @xmath3 saturates to the largest value in perseus . in summary ,
the most important properties we find from the individual regions are the abrupt increase of @xmath3 at @xmath9 @xmath10 3 mag and the saturation of @xmath3 at @xmath9 @xmath11 3 mag .
however , @xmath3 saturates to different values , from @xmath450 k km s@xmath8 for ic348 to @xmath480 k km s@xmath8 for ngc1333 .
our derived spatial distribution of @xmath1 allows us to test interesting theoretical predictions such as the dependence of @xmath1 on @xmath9 . in figure
[ f : x_co_av ] , we plot @xmath1 as a function of @xmath9 for each dark and star - forming region .
while b5 and l1448 do not show a clear relation between @xmath1 and @xmath9 due to their narrow range of @xmath9 , ic348 has a distinct trend of @xmath1 decreasing at small @xmath9 and increasing at large @xmath9 .
@xmath1 decreases by a factor of @xmath470 at @xmath9 @xmath4 12.5 mag and increases by only a factor of @xmath44 at @xmath9 @xmath4 2.511 mag . in the case of b1e / b1
, there appears to be two components .
the majority of the data points show a linear increase of @xmath1 from @xmath47 @xmath5 10@xmath124 to @xmath45 @xmath5 10@xmath6 for @xmath9 @xmath4 1.55 mag .
the additional group of the data points is located at @xmath9 @xmath4 23 mag and @xmath1 @xmath4 10@xmath22 with some scatter . finally , ngc1333 has the majority of the data points ( @xmath483% ) at @xmath9 @xmath10 3 mag and @xmath1 @xmath10 5 @xmath5 10@xmath6 with a large degree of scatter ( a factor of @xmath410 ) . at @xmath9 @xmath4 310 mag ,
@xmath1 increases by only a factor of @xmath44 .
overall , we find a factor of up to @xmath4100 variations in @xmath1 for ic348 , b1e / b1 , and ngc1333 with a size of @xmath467 pc .
we notice that the shape of the @xmath1 vs @xmath9 profiles is primarily driven by how @xmath3 changes with @xmath9 . specifically , decreasing @xmath1 with @xmath9 results from the steep increase of @xmath3 at small @xmath9 , while increasing @xmath1 with @xmath9 is due to the saturation of @xmath3 at large @xmath9 .
@xmath1 decreases because @xmath3 increases more steeply than @xmath2(h@xmath0 ) likely due to the sharp transition from cii / ci to co. on the other hand , @xmath1 increases because @xmath3 increases gradually compared to @xmath2(h@xmath0 ) likely due to the saturation of @xmath3 resulted from the large optical depth .
therefore , the transition from decreasing to increasing @xmath1 occurs in the @xmath1 vs @xmath9 profile where the co emission becomes optically thick .
this is particulary prominent for ic348 , where this transition occurs at @xmath9 @xmath4 3 mag .
b1e / b1 is relatively similar with ic348 , while we do not observe a clear indication of this transition for ngc1333 .
several theoretical studies have predicted the similar shape for the @xmath1 vs @xmath9 profile ( e.g. , taylor et al .
1993 ; bell et al .
2006 ; glover & mac low 2011 ; shetty et al .
2011a ; feldmann et al .
2012a ) . in the next sections ,
we compare our @xmath1 data with predictions from two models .
we use a modified form of the model in wolfire et al .
( 2010 ) to calculate h@xmath0 and co abundances and co line emission .
the model in wolfire et al .
( 2010 ) uses a plane - parallel pdr code with one - sided illumination to estimate the distributions of atomic and molecular species as a function of @xmath9 into a cloud .
the density distribution is taken to be the median density as expected from turbulence and the distribution is converted into a spherical geometry . in our modified w10 model ,
a plane - parallel slab of gas is illuminated by uv photons on two sides and has either a uniform density distribution or a distribution described with a simple step function . the gas temperature and the abundances of atomic and molecular species
are calculated as a function of @xmath9 under the assumptions of thermal balance and chemical equilibrium . for details on the chemical and thermal processes ,
we refer the reader to tielens & hollenbach ( 1985 ) , kaufman et al .
( 2006 ) , wolfire et al .
( 2010 ) , and hollenbach et al .
( 2012 ) .
the input parameters for the modified w10 model are @xmath37 , @xmath47 , @xmath59 , @xmath90 , @xmath67 , and dgr . considering the constraints on the physical parameters of perseus ( section 2.2 ) , we use a set of the modified w10 models with the following inputs : @xmath47 = 0.5 @xmath54 , @xmath59 = 10@xmath62 s@xmath8 , @xmath90 = 4 km s@xmath8 , @xmath67 = 1 z@xmath25 , and dgr = 1 @xmath5 10@xmath70 mag @xmath71 . for the density distribution
, we use both a uniform density distribution with @xmath37 = 10@xmath38 , 5 @xmath5 10@xmath38 , and 10@xmath24 @xmath39 and a `` core - halo '' density distribution .
the `` halo '' consists of hi with a fixed density @xmath125 = 40 @xmath39 , comparable to diffuse cold neutral medium ( cnm ) clouds ( wolfire et al .
2003 ) , and has @xmath2(hi ) = 4.5 @xmath5 10@xmath22 @xmath7 on each side of the slab .
in the `` core '' , on the other hand , @xmath37 abruptly increases to a large density @xmath126 = 10@xmath38 , 5 @xmath5 10@xmath38 , or 10@xmath24 @xmath39 .
this `` core - halo '' structure is motivated by observations of molecular clouds that have found hi envelopes with @xmath2(hi ) @xmath4 10@xmath100 @xmath7 ( e.g. , imara & blitz 2011 ; imara , bigiel , & blitz 2011 ; lee et al .
2012 ) . as the minimum density of the densest regions for both the uniform and `` core - halo '' density distributions ( @xmath410@xmath38 @xmath39 )
has already been constrained by previous comparisons between co observations and lvg / pdr models ( section 2.2 ) , we expect that the modified w10 model with density much smaller than 10@xmath38 @xmath39 would not reproduce the observed @xmath3 in perseus and therefore do not demonstrate the effect of @xmath37 , @xmath126 < 10@xmath38 @xmath39 in this paper .
in addition , we note that the modified w10 model is not sensitive to the exact value of @xmath125 , as long as this is small enough to contain a small amount of h@xmath0 and co in the halo ( section 7.1.2 for details ) .
we run the model for @xmath9 = 0.6 , 0.8 , 1 , 1.5 , 2 , 2.8 , 4.8 , 7.2 , 10 mag ( uniform density ) and @xmath9 = 1.25 , 1.3 , 1.5 , 1.7 , 2 , 2.8 , 4.8 , 7.2 , 10 mag ( `` core - halo '' ) and the output quantities are @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath3 for a given @xmath9 .
we summarize the ranges of the output quantities in tables 2 ( `` core - halo '' ) and 3 ( uniform density ) .
note that for both the uniform and `` core - halo '' density distributions an increase in @xmath9 can be thought of as an increase in size of the dense region .
for example , @xmath9 = dgr @xmath5 @xmath2(h ) = dgr@xmath127@xmath128 + @xmath125@xmath129 = 3.1 @xmath5 10@xmath53 @xmath128@xmath126 + 0.9 mag , with @xmath128 in units of pc and @xmath126 in units of @xmath39 for the `` core - halo '' density distribution .
the `` core '' has a typical size @xmath128 @xmath10 1 pc , while the `` halo '' is significantly more extended with @xmath130 @xmath4 7 pc . for the uniform density distribution ,
the size of the slab is generally @xmath131 @xmath10 1 pc .
we note that in the most extreme case the size of the dense region is much smaller than our spatial resolution ( @xmath128 @xmath4 0.01 pc ) , implying a considerably small filling factor of the `` core '' relative to the `` halo '' , but comparable to the size of small - scale clumps observed in the co emission ( e.g. , heithausen et al .
1998 ; kramer et al . 1998 ) .
we compare @xmath1 vs @xmath9 with predictions from the modified w10 model ( `` core - halo '' ) in figure [ f : x_co_av_w10](a ) . while b5 and l1448 probe too narrow ranges of @xmath9 for significant comparisons , the model curves with @xmath126 = 10@xmath41 @xmath39 follow the observed trends for ic348 and b1e / b1 .
the situation is more complicated for ngc1333 , where the model matches the observed @xmath1 only for a partial range of @xmath9 and has difficulties in reproducing the observations at @xmath9 @xmath10 3 mag and @xmath1 @xmath10 10@xmath6 .
in addition , ngc1333 lacks the decreasing portion of the @xmath1 vs @xmath9 profile because of the missing data points with small @xmath3 @xmath10 10 k km s@xmath8 .
here we provide a description of the detailed comparison between our data of ic348 , b1e / b1 , and ngc1333 and the modified w10 model . 1 . for ic348 ,
the model with @xmath126 = 10@xmath38 @xmath39 reproduces well the observed shape of the @xmath1 vs @xmath9 profile ( decreasing @xmath1 at @xmath9 @xmath10 3 mag and increasing @xmath1 at @xmath9 @xmath11 3 mag ) .
2 . for b1e / b1
, the model with @xmath126 varying from 10@xmath38 @xmath39 to 10@xmath24 @xmath39 can reproduce the observed shape of the @xmath1 and @xmath9 profile .
3 . for ngc1333 ,
the observed scatter at small @xmath9 calls for a range of @xmath126 @xmath4 10@xmath41 @xmath39 .
considering that the models with @xmath126 = 5 @xmath5 10@xmath38 @xmath39 and 10@xmath24 @xmath39 are essentially identical , however , the data points at @xmath9 @xmath10 3 mag with @xmath1 @xmath10 10@xmath6 would not be reproduced by the model with @xmath126 > 10@xmath24 @xmath39 .
in addition , our observational data lack the decreasing portion of the @xmath1 vs @xmath9 profile . we suspect that this is due to the limited spatial coverage of the complete @xmath3 image , which does not adequately sample low column density regions for ngc1333 ( only @xmath410% of the data points have @xmath3 < 10 k km s@xmath8 ) . in figure
[ f : x_co_av_w10](b ) , we compare the observed @xmath1 vs @xmath2(h@xmath0 ) profiles with the model and find similar results . in summary , the modified w10 model with the `` core - halo '' structure and
the input parameters appropriate for perseus predicts the ranges of @xmath3 and @xmath2(h@xmath0 ) in good agreement with our data .
ic348 and b1e / b1 are the best cases where the shape of the @xmath1 vs @xmath9 profiles and the location of the minimum @xmath1 are well described by the model .
we note that there are some discrepancies at low column densities in ngc1333 , where the data points with @xmath1 @xmath10 10@xmath6 are not reproduced by the model and at the same time the observed data with @xmath1 @xmath11 10@xmath22 are missing due to the limited observational coverage .
next , we plot @xmath2(hi ) as a function of @xmath2(h ) in figure [ f : x_co_av_w10](c ) and compare the profiles with the modified w10 model . as summarized in section 2.1 , lee et al .
( 2012 ) found a relatively uniform @xmath2(hi ) distribution across perseus with @xmath4(810 ) @xmath5 10@xmath22 @xmath7 .
here we use the same @xmath2(hi ) image as in lee et al .
( 2012 ) and apply the same boundaries for the five regions as in section 4 .
we find that the mean @xmath2(hi ) varies from @xmath47.4 @xmath5 10@xmath22 @xmath7 ( b5 ) to @xmath49.6 @xmath5 10@xmath22 @xmath7 ( ngc1333 and l1448 ) .
the model predicts @xmath2(hi ) @xmath4 ( 99.6 ) @xmath5 10@xmath22 @xmath7 , with essentially no difference between @xmath132 @xmath39 and 10@xmath133 @xmath39 models .
the predicted @xmath2(hi ) distribution with @xmath49 @xmath5 10@xmath22 @xmath7 and its uniformity are consistent with what we observe in perseus .
this agreement will persist even if the @xmath2(hi ) distribution is corrected for high optical depth hi .
our preliminary work on the effect of high optical depth hi that is missing in the hi emission image of perseus shows that @xmath2(hi ) increases by a factor of @xmath41.5 at most due to the optical depth correction ( the corrected @xmath2(hi ) @xmath4 ( 818 ) @xmath5 10@xmath22 @xmath7 ; stanimirovi et al . in prep ) .
the ranges of the predicted @xmath2(hi ) and @xmath2(h@xmath0 ) distributions are comparable to what we find in perseus . in figure
[ f : x_co_av_w10](d ) , we plot @xmath29 against @xmath2(h ) and indeed find that the model matches well our observations . in particular , the linearly increasing @xmath29 with @xmath2(h ) is reproduced well by the model , mainly driven by the uniform @xmath2(hi ) distribution .
so far we made comparisons between the observations of perseus and the modified w10 model with the `` core - halo '' structure . to investigate the role of the diffuse halo in determining h@xmath0 and co distributions , we show our data for ic348 and predictions from the modified w10 model both with the `` core - halo '' structure and the uniform density distribution in figure [ f : w_compare ] .
the uniform density distribution simply assumes a dense core with @xmath37 = 10@xmath38 , 5 @xmath5 10@xmath38 , or 10@xmath24 @xmath39 .
clearly , the `` core - halo '' model describes our data better .
for example , the uniform density model underestimates the @xmath2(hi ) distribution compared to the observed one across the cloud .
in addition , it predicts the decreasing portion of the @xmath1 vs @xmath9 profile shallower than our data , while reproducing the observed range of @xmath1 reasonably well .
we compare the `` core - halo '' model with the uniform density model in detail as follows .
@xmath2(hi ) vs @xmath2(h ) : the uniform density model predicts @xmath2(hi ) significantly smaller than what we measure across perseus , @xmath2(hi ) @xmath4 9 @xmath5 10@xmath22 @xmath7 .
the discrepancy ranges from a factor of @xmath41020 for @xmath37 = 10@xmath38 @xmath39 to a factor of @xmath470160 for @xmath37 = 10@xmath24 @xmath39 .
this large discrepancy results from the fact that h@xmath0 self - shielding is so strong that almost all hydrogen is converted into h@xmath0 . on the other hand ,
the density of the halo is small enough that dust shielding is more important than h@xmath0 self - shielding . to provide the sufficient dust shielding for h@xmath0 formation ,
the entire halo remains atomic with its initial @xmath2(hi ) @xmath4 9 @xmath5 10@xmath22 @xmath7 , resulting in the uniform @xmath2(hi ) distribution .
we expect that if the density of the halo is significantly larger than the current @xmath125 = 40 @xmath39 , the halo will no longer be purely atomic due to the increased h@xmath0 self - shielding .
@xmath2(h@xmath0 ) vs @xmath2(h ) : all models predict the @xmath2(h@xmath0 ) vs @xmath2(h ) profile in good agreement with our data , even though the uniform density model slightly overestimates @xmath2(h@xmath0 ) at small @xmath2(h ) . for example , the uniform density model with @xmath37 = 10@xmath24 @xmath39 predicts @xmath2(h@xmath0 ) = 9.96 @xmath5 10@xmath22 @xmath7 at @xmath2(h ) = 2 @xmath5 10@xmath100 @xmath7 , larger than our data by less than a factor of 2 . however , this discrepancy is significant at such small @xmath2(h ) and results in the small amount of @xmath2(hi ) @xmath10
10@xmath6 @xmath7 .
in addition , models with different densities predict essentially the same @xmath2(h@xmath0 ) for a given @xmath2(h ) .
all these results imply that neither density nor its distribution is critical for the h@xmath0 abundance .
instead , @xmath2(h ) primarily determines @xmath2(h@xmath0 ) . @xmath29 vs @xmath2(h ) : while the `` core - halo '' model reproduces both the range of @xmath29 and the linear increase of @xmath29 with @xmath2(h ) , the uniform density model overestimates @xmath29 for a given @xmath2(h ) by up to a factor of @xmath4300 .
this discrepancy mainly results from the significantly underestimated @xmath2(hi ) in the uniform density model .
@xmath3 vs @xmath2(h@xmath0 ) : all models reproduce the observed @xmath3 vs @xmath2(h@xmath0 ) profile reasonably well . in particular , both the `` core - halo '' and uniform density models with the smallest density show an excellent agreement with our data for ic348 .
while the models with @xmath37 @xmath11 5 @xmath5 10@xmath38 @xmath39 and @xmath126 @xmath11 5 @xmath5 10@xmath38 @xmath39 predict larger @xmath3 at small @xmath2(h@xmath0 ) ( up to a factor of @xmath410 ) , the difference between the models with different densities becomes negligible at @xmath2(h@xmath0 ) @xmath11 1 @xmath5 10@xmath100 @xmath7 , where @xmath3 saturates to @xmath44560 k km s@xmath8 for the uniform density model and @xmath43040 k km s@xmath8 for the `` core - halo '' model .
all these results suggest that @xmath3 depends on density but only at small @xmath2(h@xmath0 ) and changes in physical parameters other than density ( e.g. , @xmath90 ) will be required to produce larger @xmath3 values once @xmath3 becomes optically thick .
@xmath3 vs @xmath9 : while the `` core - halo '' model reproduces the sharp increase of @xmath3 observed at @xmath9 @xmath11 1 mag , the uniform density model predicts the increase of @xmath3 at @xmath9 @xmath11 0.6 mag much more gradually than our data .
this difference comes from the fact that the uniform density model has larger density than the `` core - halo '' model , resulting in the larger @xmath3 for a given @xmath9 @xmath10 3 mag ( @xmath37 @xmath11 10@xmath38 @xmath39 for the uniform density model vs @xmath134 @xmath4 55125 @xmath39 for the `` core - halo '' model ; table 2 ) .
on the other hand , all models predict the saturation of @xmath3 to similar values at @xmath9 @xmath11 3 mag , suggesting that the larger density in the uniform density model no longer has a significant impact on @xmath3 due to the large optical depth of @xmath3 ( @xmath37 @xmath11 10@xmath38 @xmath39 for the uniform density model vs @xmath134 @xmath4 180430 @xmath39 for the `` core - halo '' model ; table 2 ) .
@xmath1 vs @xmath9 : all models reproduce the observed increase of @xmath1 at @xmath9 @xmath11 3 mag , because they predict both the range of @xmath2(h@xmath0 ) and the saturation of @xmath3 comparable to our data . on the other hand , the uniform density model shows the decrease of @xmath1 at @xmath9 @xmath10 3 mag much shallower than our data .
this discrepancy mainly results from the less steep increase of @xmath3 predicted by the model at @xmath9 @xmath10 3 mag .
summary : while we do not perform a full parameter space search , our comparison between the `` core - halo '' and uniform density models is illustrative and demonstrates that the diffuse halo is essential for reproducing the following observed properties : the uniform @xmath2(hi ) distribution , the h@xmath0-to - hi ratio for a given @xmath2(h ) , the sharp increase of @xmath3 and decrease of @xmath1 at 1 mag @xmath10 @xmath9 @xmath10 3 mag . considering that the uniform density model predicts the @xmath3 distribution extended toward smaller @xmath9 , while producing the @xmath2(h@xmath0 ) distribution in reasonably good agreement with our data ( figures [ f : w_compare]d and j ) , we expect that the neglect of the diffuse halo will result in the underestimation of the size of `` co - free '' h@xmath0 envelope . the s11 model is essentially comprised of two parts . the first part is a modified version of the zeus mp mhd code ( stone & norman 1992 ; norman 2000 ) .
gas in a periodic box is set to have a uniform density distribution and is driven by a turbulent velocity field with uniform power 1 @xmath135 @xmath136 @xmath135 2 where @xmath136 is the wavenumber .
in addition , the magnetic field has initially orientation parallel to the @xmath137-axis , with a strength of 1.95 @xmath51 g . to model the chemical evolution of the gas , glover & mac low ( 2007a , b ) , glover et al .
( 2010 ) , and glover & clark ( 2012 ) updated the zeus - mp mhd code with chemical reactions of several atomic and molecular species .
the photodissociation of molecules by a radiation field is treated by the `` six - ray approximation '' method developed by glover & mac low ( 2007a ) .
the effect of self - gravity is not included .
we refer to glover & mac low ( 2007a , b ) , glover et al ( 2010 ) , and glover & clark ( 2012 ) for details on mhd , thermodynamics , and chemistry included in the s11 model .
the second part is a three - dimensional radiative transfer code radmc3d ( dullemond et al . in prep ) .
once the simulated molecular cloud reaches a statistically steady state , radmc3d is executed to model molecular line emission ( e.g. , co ) . to solve the population levels of atomic and molecular species , radmc3d implements the lvg method ( sobolev 1957 ) , which has been shown to be a good approximation for molecular clouds ( e.g. , ossenkopf 1997 ) .
we refer to shetty et al .
( 2011a ) for details on radmc3d
. the mhd simulation follows the evolution of an initially atomic gas in a ( 20 pc)@xmath38 box with a numerical resolution of 512@xmath138 . in this paper
, we use the s11 model with the following input parameters : initial @xmath37 = 100 @xmath39 , @xmath47 = 1 @xmath54 , @xmath59 = 10@xmath60 s@xmath8 , @xmath67 = 1 z@xmath25 , and dgr = 5.3 @xmath5 10@xmath72 mag @xmath71 .
this simulation is essentially the same as the `` n100 model '' in s11 but has a higher numerical resolution and a simpler co formation model based on nelson & langer ( 1999 ) .
we choose this particular simulation because it has a mass of @xmath42 @xmath5 10@xmath24 m@xmath25 , consistent with that of perseus .
the input parameters for the s11 model are reasonably close to what we expect for perseus but not exactly the same as what we used for the modified w10 model .
as it has been shown in s11 and glover & mac low ( 2007b ) that the simulated h@xmath106 and co column densities do not depend on small changes in @xmath47 and @xmath59 , this simulation would be appropriate for the comparison with our observations ( section 8.4.1 for details ) .
compared to the modified w10 model , the final density distribution in the s11 model has a majority of the data points ( @xmath499% ) with @xmath37 < 10@xmath38 @xmath39 , resulting in the small median density of @xmath430 @xmath39 .
another important difference between the modified w10 model and the s11 model is that h@xmath0 formation in the s11 model does not achieve chemical equilibrium until the end of the simulation .
for example , glover et al . (
2010 ) found from their mhd simulations that the h@xmath0 abundance primarily depends on the time available for h@xmath0 formation and shows no indication of chemical equilibrium up to @xmath139 @xmath4 20 myr .
the gas will eventually become fully h@xmath0 unless the molecular cloud is destroyed by stellar feedback such as photoevaporation by hii regions and protostellar outflows . on the other hand ,
the co abundance is controlled by photodissociation and reaches chemical equilibrium within @xmath139 @xmath4 2 myr .
the final products of the s11 model include the @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath3 images obtained at @xmath139 @xmath4 5.7 myr .
we smooth and regrid the simulated @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath3 images so that they have both a spatial resolution of 0.4 pc and a pixel size of 0.4 pc .
recently , beaumont et al . (
2013 ) compared the complete data of perseus with the s11 model and found that the s11 model systematically overestimates @xmath2(h@xmath0 ) ( e.g. , figure 5 of beaumont et al .
one of the possible explanations for this discrepancy is the different size between the simulation box and the individual regions in perseus . because the simulation box is larger than the individual regions in perseus ( 20 pc vs @xmath457 pc ) , the integrated quantities @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath3 would need to be scaled . in the case of @xmath2(hi ) and @xmath2(h@xmath0 ) , the scaling is straightforward under the assumption of isotropic density distribution , which is appropriate for the s11 model , and we simply need to account for the difference between the box and region sizes .
however , estimating a proper scaling for @xmath3 is much more complicated because of the following reasons .
first , the @xmath3 image was produced from the s11 model by integrating the co brightness temperature , which was estimated by three - dimensional radiative transfer calculations , along a full radial velocity range .
second , the co emission is optically thick in some parts of the simulation ( @xmath410% of the volume ) .
re - running the simulation with a smaller box does not solve the problem as molecular cores / clouds form out of initially larger - scale diffuse ism .
we therefore take an approach of determining the optimal line of sight ( los ) depth that minimizes the difference between our observations and the s11 model by investigating the @xmath2(hi ) and @xmath2(h@xmath0 ) images simultaneously . for the simulated @xmath3 image , on the other hand
, we do not apply any scaling . to do this
, we estimate the difference between the observed mean and the simulated mean for each of @xmath2(hi ) and @xmath2(h@xmath0 ) with varying los depths . for example
, we divide the simulated @xmath2(hi ) and @xmath2(h@xmath0 ) images by two to calculate the mean @xmath2(hi ) and @xmath2(h@xmath0 ) for the simulation with the los depth of 10 pc .
we then normalize the difference by the observed mean of each quantity and calculate the sum of the two normalized differences in quadrature .
the results are shown in figure [ f : box_size ] and we find that the los depth that minimizes the difference between our data and the simulation products is 7 pc ( figure [ f : box_size]c ) .
while the final quantity in figure [ f : box_size](c ) has a broad minimum , it is encouraging that the estimated scale length is comparable to both the characteristic size of the five regions in perseus and the total size of the slab for the `` core - halo '' model ( tables 1 and 2 ) . as a double check
that this scale length is appropriate , we use larson s law established for turbulent molecular clouds from both observations and mhd simulations : @xmath140 km s@xmath8 ( heyer & brunt 2004 ) . for a region size of 20 pc
we expect @xmath88 @xmath4 6 km s@xmath8 , while for a region size of 7 pc we expect @xmath88 @xmath4 3 km s@xmath8 .
this level of co velocity dispersion is in agreement with what is shown in figure [ f : co_vel_dispersion ] , confirming that scaling the simulation products to the los depth of 7 pc is reasonable . in summary ,
when we compare our observations with the s11 model , we scale the simulated @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath2(h ) images by multiplying them by 7/20 ( figures [ f : histo_perseus_s11]a , b , c , and figures [ f : hi_tot_s11]a , b ) . on the other hand , because of the uncertainty in @xmath3 scaling , we use the original @xmath3 image produced by the s11 model ( figure [ f : histo_perseus_s11]d and figures [ f : hi_tot_s11]c , d ) . finally , we apply the following thresholds to the simulated data to mimic the sensitivity limits of our observational data : @xmath2(h@xmath0 ) > 3.3 @xmath5 10@xmath6 @xmath7 and @xmath3 > 0.09 k km s@xmath8 ( our mean 1@xmath101 uncertainties calculated for the data points with @xmath2(h@xmath0 ) > 0 @xmath7 and @xmath3 > 0 k km s@xmath8 ) .
this application of the thresholds to the s11 model is reasonable , considering the minimum @xmath2(h@xmath0 ) @xmath4 3.8 @xmath5 10@xmath6 @xmath7 and @xmath3 @xmath4 0.2 k km s@xmath8 for the five regions in perseus .
we first compare our data with the s11 model by constructing normalized histograms of @xmath2(hi ) , @xmath2(h@xmath0 ) , @xmath2(h ) , and @xmath3 in figure [ f : histo_perseus_s11 ] . to construct the histograms
, we use the data points with @xmath2(h@xmath0 ) > 0 @xmath7 in figure [ f : h2_cden_display ] ( `` all '' histograms in black ) , as well as those shown in figure [ f : complete_five_regions ] ( `` subset '' histograms in grey ) .
while the grey histograms are limited to the regions where the co emission is detected , the black histograms represent the whole perseus cloud .
the simulated data from the s11 model ( smoothed , regridded , scaled for 7 pc , and the thresholds applied ) are shown as green histograms .
note that the @xmath3 values from the s11 model are not scaled and therefore the green @xmath3 histogram likely represents the upper limit of actual histogram for sub - regions with a size of @xmath47 pc ( indicated as an arrow ) . because the simulated data ( except for @xmath3 ) are scaled to match the properties of the five regions and the thresholds applied to the s11 model are comparable to the minimum @xmath2(h@xmath0 ) and @xmath3 values of the five regions , the green histograms can be directly compared with the grey histograms . in comparison between our data and the s11 model ,
we find the following .
first , the black and grey @xmath2(hi ) histograms are nearly identical .
this results from the small variation in @xmath2(hi ) across the whole perseus cloud , as discussed in section 2.1 .
the green histogram , on the other hand , has a peak at a factor of @xmath42 smaller @xmath2(hi ) and even more importantly a factor of @xmath46 broader distribution than the observed data ( the black and grey histograms ) .
second , the grey and green @xmath2(h@xmath0 ) histograms agree very well : both peak at a similar @xmath2(h@xmath0 ) , have a similar width , and show a lognormal - like distribution . the black histogram , on the other hand , is broader and has a tail toward small @xmath2(h@xmath0 ) .
the difference between the black and grey histograms results from the existence of h@xmath0 beyond the co spatial coverage ( e.g. , `` co - dark '' h@xmath0 discussed in lee et al .
third , the green @xmath2(h ) histogram peaks at a similar @xmath2(h ) compared to the grey histogram , while showing a broader ( a factor of @xmath42 ) and lognormal - like distribution .
the simulated distribution is broader mainly because the simulated @xmath2(hi ) has a greater range than what is observed .
the black and grey histograms , on the other hand , have a tail toward @xmath2(h ) @xmath11 10@xmath141 @xmath4 2.5 @xmath5 10@xmath100 @xmath7 .
this tail is consistent with kainulainen et al .
( 2009 ) , who found a deviation from the lognormal distribution at @xmath9 @xmath11 3 mag for perseus ( corresponding to @xmath2(h ) @xmath4 2.7 @xmath5 10@xmath100 @xmath7 with dgr = 1.1 @xmath5 10@xmath70 mag @xmath71 ) and interpreted it as a result of self - gravity .
lastly , because the simulated @xmath3 is not scaled for the los depth of 7 pc , we do not compare the exact shapes of the green and grey histograms but emphasize that the simulated @xmath3 becomes comparable to the observed @xmath3 only if we use the whole simulation box of 20 pc . in summary , we find that the scaled s11 model reproduces the observed range of @xmath2(h@xmath0 ) very well . while the predicted @xmath2(hi ) has a relatively similar mean value compared to the observed @xmath2(hi ) , it has a broader distribution and this leads to a broader range of @xmath2(h ) in the simulation .
the @xmath3 values from the s11 model , on the other hand , can not be properly compared with our observations because of the nontrivial scaling of @xmath3 with the los depth .
however , we find that the simulated @xmath3 is similar with the observed @xmath3 only when the co emission is integrated for the full simulation box of 20 pc
. we plot @xmath2(hi ) against @xmath2(h ) for each dark and star - forming region and show predictions from the s11 model ( smoothed , regridded , scaled for 7 pc , and the thresholds applied ) in figure [ f : hi_tot_s11](a ) .
while the observed @xmath2(hi ) @xmath4 9 @xmath5 10@xmath22 @xmath7 is in the range of the predicted @xmath2(hi ) , the relation between @xmath2(hi ) and @xmath2(h ) in the s11 model is different from what we find in perseus : not only does the simulated @xmath2(hi ) have a broader distribution , but the s11 model predicts a factor of @xmath47 increase of @xmath2(hi ) for the range of @xmath2(h ) in perseus , where we observe less than a factor of 2 variation in @xmath2(hi ) .
this suggests that @xmath2(hi ) linearly correlates with @xmath2(h ) in the s11 model and we indeed estimate pearson s linear correlation coefficient @xmath142 @xmath4 0.8 . in addition , the s11 model predicts a factor of @xmath42 smaller @xmath2(hi ) for a given @xmath2(h ) on average . as a result ,
@xmath29 is slightly larger in the s11 model for a given @xmath2(h ) and increases with @xmath2(h ) with a slope smaller than what we observe ( figure [ f : hi_tot_s11]b ) . while our observations show @xmath29 < 1 for the outskirts of the five regions , the simulation has @xmath29 > 1 everywhere , even for the regions with small @xmath37
< 10@xmath143 @xmath39 .
next , we plot the observed @xmath3 as a function of @xmath9 and show the s11 model in figure [ f : hi_tot_s11](c ) .
as discussed in section 7.2.1 , the simulated @xmath2(hi ) and @xmath2(h@xmath0 ) data can be scaled for the five regions in perseus , while the simulated @xmath3 data can not . to properly examine the relation between @xmath3 and @xmath9 in the s11 model , therefore ,
we show the predicted @xmath3 vs @xmath9 profile without applying the scaling and the thresholds and focus on only the general shape of the profile .
we find that the s11 model describes the relation between @xmath3 and @xmath9 reasonably well : a steep increase of @xmath3 at small @xmath9 and a hint of the saturation of @xmath3 at large @xmath9 .
interestingly , the s11 model predicts that @xmath3 increases with a large scatter at small @xmath9 .
finally , we show the @xmath1 vs @xmath9 profile for each dark and star - forming region in figure [ f : hi_tot_s11](d ) with the s11 model . as in figure
[ f : hi_tot_s11](c ) , the unscaled @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath3 data are used for this comparison .
we find that the s11 model predicts a sharp decrease of @xmath1 at small @xmath9 and a gradual increase of @xmath1 at large @xmath9 . while a quantitative comparison is not possible without scaling , the simulated data show the characteristic relation between @xmath1 and @xmath9 in broad agreement with the observational data ( particulary for ic348 and b1e / b1 ) .
this is consistent with shetty et al .
( 2011a ) , who performed a number of mhd simulations ( @xmath37 = 100 , 300 , 1000 @xmath39 and @xmath67 = 0.1 , 0.3 , 1 z@xmath25 ) and found a steep decrease of @xmath1 at @xmath9 @xmath10 7 mag and a steady increase of @xmath1 at @xmath9 @xmath11 7 mag for all simulations probing a large range of interstellar environments .
relative to the observations , we find that the simulated @xmath1 shows a significantly larger scatter at small @xmath9 , while the scatter becomes more comparable to what is found in the observations at the high end of the @xmath9 range .
in their recent review , bolatto et al . ( 2013 ) showed that there is some degree of uniformity among the @xmath1 values in the milky way obtained from a variety of observational methods .
essentially , the typical value for the milky way is @xmath1 @xmath4 2 @xmath5 10@xmath22 and is known within a factor of @xmath42 .
we , on the other hand , found that the dark and star - forming regions in perseus have @xmath108 at least five times smaller than the typical value . in appendix
, we provide a detailed comparison with two previous studies , dame et al .
( 2001 ) and pineda et al .
( 2008 ) , to understand the reasons behind such a significant difference .
we summarize our findings here .
we find three potential sources responsible for the difference : the different resolution of @xmath3 and @xmath2(h@xmath0 ) images used to derive @xmath1 , the application of different dgr , and the treatment of hi in deriving @xmath2(h@xmath0 ) .
for example , dame et al . ( 2001 ) estimated @xmath144 @xmath4 1.2 @xmath5 10@xmath22 for perseus , which is a factor of @xmath44 larger than our @xmath108 @xmath4 3 @xmath5 10@xmath6 , by combining @xmath3 from the cfa survey with @xmath2(h@xmath0 ) derived using the @xmath145 data from schlegel et al .
( 1998 ) and the hi data from the leiden - argentine - bonn ( lab ) survey ( kalberla et al . 2005 ) .
their study , as well as other large - scale studies of @xmath1 in the milky way ( e.g. , abdo et al .
2010 ; paradis et al .
2012 ) , is at 36@xmath23 resolution , mainly limited by the lab hi data . in comparison between our original @xmath1 at 4.3@xmath23 resolution and
our @xmath1 smoothed to 36@xmath23 resolution , we find that spatial smoothing results in a factor of @xmath41.5 increase in @xmath108 . considering a factor of @xmath48 decrease in angular resolution
, the effect of resolution on the estimation of @xmath108 appears to be mild and is within the accepted uncertainties , although this would be likely more significant when comparing extragalactic observations on @xmath4kpc scales .
we then find that the rest of the difference between our @xmath108 and @xmath146 can be explained by the difference in dgr . while both studies measured dgr , dame et al .
( 2001 ) calculated @xmath2(hi ) along a whole line of sight ( while we focused on the velocity range for perseus only ) and estimated dgr using the images smoothed to 10@xmath20 resolution ( while we had 4.3@xmath23 resolution ) .
the dgr effect is slightly larger than the resolution effect ( a factor of @xmath41.8 ) and these two factors together account for most of the difference between our study and dame et al .
( 2001 ) . in the case of pineda
et al . ( 2008 ) , angular resolution is not an issue because essentially the same @xmath9 and @xmath3 data were used .
however , they estimated @xmath147 @xmath4 1.4 @xmath5 10@xmath22 for perseus .
their methodology for deriving @xmath1 is different from our study mainly in two ways .
first , they assumed that the @xmath2(hi ) contribution to @xmath9 is insignificant and therefore did not consider it .
second , they adopted the typical dgr for the milky way = 5.3 @xmath5 10@xmath72 mag @xmath71 ( bohlin et al . 1978 ) .
in contrast , we accounted for the @xmath2(hi ) contribution and estimated dgr = 1.1 @xmath5 10@xmath70 mag @xmath71 ( lee et al . 2012 ) . in appendix
, we show that we estimate @xmath108 @xmath4 1 @xmath5 10@xmath22 , which is comparable to @xmath147 , when we follow the methodology of pineda et al .
in addition , we find that the application of each of the two assumptions made by pineda et al .
( 2008 ) results in a factor of @xmath42 difference in @xmath108 , altogether explaining the difference between our @xmath108 and @xmath147 . our detailed comparison with dame et al .
( 2001 ) and pineda et al .
( 2008 ) shows that different resolutions and methodologies for deriving @xmath1 can result in a difference in @xmath1 by up to a factor of @xmath44 , even for the same method of @xmath1 determination ( @xmath1 based on dust emission / absorption in this case ) .
other methods of @xmath1 determination , e.g. , @xmath1 based on the virial technique and @xmath148-ray observations , have their own assumptions .
this clearly suggests the difficulty in comparing @xmath1 between molecular clouds and/or galaxies when different observational methods are used , as pointed out by bolatto et al .
( 2013 ) as well .
the relatively uniform value of @xmath1 for the milky way found from many studies with various resolutions and methodologies , therefore , appears puzzling . in section 6.3 , we focused on the individual dark and star - forming regions in perseus and found significant spatial variations in @xmath1 .
specifically , @xmath1 varies by up to a factor of @xmath4100 within a single region with a size of @xmath467 pc .
our investigation of the large - scale trends in @xmath47 and @xmath88 ( section 6.1 ) and our comparison with the modified w10 model ( section 7.1 ) suggest that changes in physical parameters are responsible for the variations in @xmath1 observed both within the individual regions and between the different regions .
while @xmath1 shows significant variations across the cloud , we found that there is a characteristic dependence of @xmath1 on @xmath9 ( particularly evident for ic348 and b1e / b1 ) : a steep decrease of @xmath1 at @xmath9 @xmath10 3 mag and a moderate increase of @xmath1 at @xmath9 @xmath11 3 mag .
this relation between @xmath1 and @xmath9 appears to result from the strong dependence of @xmath3 on @xmath9 .
the location at which most carbon is locked in co primarily depends on dust shielding ( e.g. , w10 ; glover & mac low 2011 ) .
once dust shielding becomes sufficiently strong to prevent photodissociation ( @xmath9 @xmath11 1 mag in perseus ) , the co abundance and emission strength sharply rise and this could result in decreasing @xmath1 with @xmath9 .
@xmath3 then saturates to a certain value because the co emission becomes optically thick with increasing depths ( @xmath9 @xmath11 3 mag in perseus ) and this could result in increasing @xmath1 with @xmath9 .
these results suggest that co is a poor tracer of h@xmath0 for those regions where dust shielding is not strong enough to prevent photodissociation , e.g. , low - metallicity environments ( e.g. , leroy et al .
2007 , 2009 , 2011 ; cormier et al . 2014 ) .
in addition , co is unreliable for those regions where the co emission is optically thick , because it provides only a lower limit on @xmath2(h@xmath0 ) .
overall , our study suggests that one can not adopt a single @xmath1 to derive the @xmath2(h@xmath0 ) distribution across a resolved molecular cloud .
the limited dynamic range of co as a tracer of h@xmath0 and the complex dependence of @xmath1 on various physical parameters hamper the derivation of the accurate @xmath2(h@xmath0 ) distribution . on the other hand , calculation of the h@xmath0 mass over the co - observed area , @xmath149(h@xmath0)@xmath150 ,
appears to be less affected by variations in physical parameters .
for example , we estimate @xmath149(h@xmath0)@xmath150 = ( 1799.8 @xmath21 3.2 ) m@xmath25 over the complete co spatial coverage . if we derive @xmath149(h@xmath0)@xmath150 using our @xmath108 , we find @xmath149(h@xmath0,@xmath1)@xmath150 = ( 1814.1 @xmath21 0.2 ) m@xmath25
. these two estimates are comparable for perseus , mainly because a large fraction of the data points ( @xmath460% ) has @xmath1 different from our @xmath108 within a factor of @xmath42 .
the agreement between the observed @xmath1 in perseus and the model predictions ( in particular for the pdr model ) suggests that a theory - based @xmath1 could be used to estimate @xmath149(h@xmath0)@xmath150 for a molecular cloud .
once theoretical models , e.g. , pdr and mhd models , are thoroughly tested against observations of molecular clouds in diverse environments , they will be able to provide predictions over a wide range of physical conditions .
one then can search a large parameter space to select the most appropriate @xmath1 for a target molecular cloud based on reasonable constraints on physical parameters .
note , however , that the total h@xmath0 mass of the cloud would be still uncertain if there is significant `` co - dark '' h@xmath0 outside the co - observed area .
the good agreement between our data and the modified w10 model with the `` core - halo '' structure ( section 7.1.2 ) suggests that the main assumptions of the model , e.g. , h@xmath0/co formation in chemical equilibrium , the microturbulent approximation for co spectral line formation , and the `` core - halo '' density distribution , are valid for perseus on @xmath40.4 pc scales .
this result is consistent with lee et al .
( 2012 ) , who found that @xmath2(hi ) and @xmath2(h@xmath0 ) in perseus conform to the time - independent h@xmath0 formation model by k09 .
we now turn to a couple of interesting aspects of the modified w10 model and discuss their implications . the modified w10 model that is comparable to the observations of perseus uses the `` core - halo '' structure motivated by previous studies of molecular clouds ( section 7.1.1 ) .
we showed that the model with a uniform density distribution predicts @xmath2(hi ) much smaller than the uniform @xmath2(hi ) @xmath4 9 @xmath5 10@xmath22 @xmath7 measured across perseus . the uniform density model with the largest density @xmath37
= 10@xmath24 @xmath39 predicts the smallest @xmath2(hi ) , up to a factor of @xmath4160 smaller than what is observed .
the main reason is that in the uniform density model h@xmath0 self - shielding alone counteracts h@xmath0 photodissociation by lw photons .
traditionally , it has been known that @xmath47/@xmath37 determines whether h@xmath0 self - shielding or dust shielding is more important for h@xmath0 formation and controls the location of the transition from hi to h@xmath0 in a pdr ( e.g. , hollenbach & tielens 1997 ) . with @xmath47 = 0.5 @xmath54 and @xmath37 = 10@xmath38 @xmath39 in the uniform density model , @xmath47/@xmath37
is 5 @xmath5 10@xmath151 cm@xmath38 , small enough that dust shielding is negligible . in this case , most of the hi is converted into h@xmath0 because of the strong h@xmath0 self - shielding . on the other hand , the `` core - halo '' model with @xmath126
= 10@xmath38 @xmath39 and @xmath125 = 40 @xmath39 has @xmath47/@xmath125 @xmath4 0.01 cm@xmath38 in the cloud outskirts .
this increased @xmath47/@xmath37 makes h@xmath0 self - shielding less important for h@xmath0 formation and as a result , the gas remains atomic with @xmath2(hi ) @xmath4 9 @xmath5 10@xmath22 @xmath7 .
the fact that the modified w10 model needs a diffuse hi halo to reproduce the observed @xmath2(hi ) suggests that dust shielding is important for h@xmath0 formation in perseus .
this importance of dust shielding is consistent with what lee et al .
( 2012 ) found from their comparison with the k09 model .
the k09 model investigates the structure of a pdr in a spherical cloud based on h@xmath0 formation in chemical equilibrium and predicts the following variable as one of the key parameters that determine the location of the transition from hi to h@xmath0 : @xmath153 where @xmath154 is the metallicity normalized to the solar neighborhood value and @xmath155 is the ratio of the actual cnm density to the minimum cnm density at which the cnm exists in pressure balance with the warm neutral medium ( wnm ) .
this @xmath156 is the ratio of the rate at which lw photons are absorbed by dust grains ( dust shielding ) to the rate at which they are absorbed by h@xmath0 ( h@xmath0 self - shielding ) and is conceptually similar to @xmath47/@xmath37 .
k09 predicts @xmath156 @xmath4 1 in all galaxies where the pressure balance between the cnm and the wnm is valid , suggesting that dust shielding and h@xmath0 self - shielding are equally important for h@xmath0 formation . by fitting the k09 model to the observed @xmath29 vs @xmath28+@xmath17 profiles , lee et al .
( 2012 ) indeed found @xmath156 @xmath4 1 for perseus . in the modified w10 model , a diffuse hi halo
is also required to reproduce the observed steep increase of @xmath3 at @xmath9 @xmath11 1 mag and sharp decrease of @xmath1 at @xmath9 @xmath10 3 mag ( section 7.1.3 ) .
the uniform density model predicts the shallower increase of @xmath3 at smaller @xmath9 @xmath11 0.6 mag , suggesting a less sharp transition from cii / ci to co located closer to the surface of the gas slab .
the more extended co distribution eventually results in the reduced `` co - free '' h@xmath0 envelope and therefore the uniform density model with @xmath37 = 10@xmath24 @xmath39 would have the smallest amount of `` co - dark '' h@xmath0 . the co distribution deep inside of the gas slab ,
on the other hand , does not appear to be affected by the presence of the diffuse hi halo because of the saturation of @xmath3 . even though the modified w10 model with the `` core - halo '' structure reproduces the observed @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath3 distributions , the agreement is likely to remain only if the halo density is not significantly larger than 40 @xmath39 .
the current density @xmath125 = 40 @xmath39 originates from the theoretical ( e.g. , wolfire et al . 2003 ) and observational ( e.g. , heiles & troland 2003 ) properties of the cnm . while large hi envelopes associated with molecular clouds have been frequently observed ( e.g. , knapp 1974 ; wannier et al .
1983 , 1991 ; reach et al . 1994 ; rogers et al . 1995 ; williams & maddalena 1996 ; imara & blitz 2011 ; lee et al .
2012 ) , a number of fundamental questions still remain to be answered . for example
, what are the physical properties of the hi halos , such as density , temperature , and pressure ?
what is the ratio of the cnm to the wnm in the halos ?
is there any correlation between the ratio and the h@xmath0 abundance / star formation ?
are the halos expanding or infalling ?
the high - resolution hi data from the galfa - hi survey will be valuable for future studies of the extended hi halos around galactic molecular clouds in a wide range of interstellar environments .
finally , further comparisons between observations and theoretical models will be important to fully constrain the parameter space and density structure of the hi halos .
the timescale of h@xmath0 formation on dust grains , @xmath157 , dominates chemical timescales of pdrs ( e.g. , hollenbach & tielens 1997 ) . for the modified w10 model with the `` core - halo '' structure ,
dense regions have @xmath126 @xmath11 10@xmath38 @xmath39 where gas is completely molecular ( @xmath158 @xmath4 0.5@xmath37 ) . in this case , @xmath157 = 0.5/@xmath159@xmath126 @xmath10 0.5 myr , where @xmath159 = 3 @xmath5 10@xmath60 cm@xmath38 s@xmath8 is the rate coefficient for h@xmath0 formation ( wolfire et al .
2008 ) . in diffuse regions with @xmath125 = 40 @xmath39 , on the other hand
, gas is mostly atomic ( @xmath158 @xmath4 0.1@xmath37 ) and therefore @xmath157 = 0.1/@xmath159@xmath125 @xmath4 2.6 myr . because @xmath157 of the model is well within the expected age of perseus , @xmath95 @xmath4 10 myr
, the assumption of chemical equilibrium is valid . in other words , perseus is old enough to reach chemical equilibrium and therefore it is not surprising that the equilibrium chemistry model ( w10 ) fits our observations very well .
however , for steady state chemistry to be valid , @xmath157 @xmath10 @xmath95 is not enough : @xmath157 should be short compared to the dynamical timescale of a molecular cloud , @xmath160 . for perseus , this requires @xmath160 @xmath11 3 myr . as a rough estimate , we calculate a crossing timescale , @xmath161 = @xmath162/@xmath101 @xmath4 10 pc/1.8 km s@xmath8 @xmath4 6 myr , where we choose @xmath162 as the characteristic size of the individual regions in perseus and @xmath101 as the mean co velocity dispersion .
this @xmath161 @xmath4 6 myr satisfies the condition for @xmath160 @xmath11 3 myr .
however , many dynamical processes are involved with the formation and evolution of molecular clouds ( e.g. , cloud - cloud collisions , spiral shocks , stellar feedback ; mac low & klessen 2004 ; mckee & ostriker 2007 ) and therefore it is difficult to pin down the exact process that is most relavant for the formation of molecular gas .
the good agreement between our data and the modified w10 model with the `` core - halo '' structure suggests that the characteristic @xmath160 for the formation of molecular gas in perseus should be @xmath11 3 myr . in section 7.2.2 , we found that the scaled s11 model predicts @xmath2(h@xmath0 ) comparable to the estimated @xmath2(h@xmath0 ) in perseus .
this excellent agreement will likely hold even if some of the input parameters slightly change .
for example , the s11 model was run with @xmath47 = 1 @xmath54 and this is a factor of @xmath42 stronger than what we measure across perseus . considering that s11 found no considerable difference in @xmath2(h@xmath0 ) for their models with @xmath47 = 1 @xmath54 and 10 @xmath54 ( section 3.1 of s11 ) , however ,
decreasing @xmath47 from 1 @xmath54 to 0.5 @xmath54 to match the property of perseus will not make a significant change in @xmath2(h@xmath0 ) .
in addition , increasing @xmath59 from 10@xmath60 s@xmath8 to 10@xmath62 s@xmath8 to be consistent with the modified w10 model will not affect @xmath2(h@xmath0 ) very much , based on the fact that glover & mac low ( 2007b ) found a negligible change in @xmath2(h@xmath0 ) when @xmath59 increased from 10@xmath60 s@xmath8 to 10@xmath61 s@xmath8 in their mhd simulation with initial @xmath37 = 100 @xmath39 ( section 6.3 of glover & mac low 2007b ) .
increasing dgr from 5.3 @xmath5 10@xmath72 mag @xmath71 to 1.1 @xmath5 10@xmath70 mag @xmath71 for perseus will lead to more rapid h@xmath0 formation , but the model with the increased dgr will not be substantially different from the current s11 model since the s11 model becomes h@xmath0-dominated rapidly by @xmath139 @xmath4 3 myr ( figure 7 of glover & mac low 2011 ) .
finally , the extension of the simulation run up to @xmath139 @xmath4 10 myr , comparable to the age of perseus , will not significantly increase @xmath2(h@xmath0 ) , considering that glover & mac low ( 2011 ) found only a factor of @xmath41.3 increase of the mass - weighted mean h@xmath0 abundance from @xmath139 @xmath4 5 myr to @xmath139 @xmath4 10 myr for their mhd simulation with initial @xmath37 = 100 @xmath39 ( section 3.3 of glover & mac low 2011 ) .
similarly , small changes in the model parameters will likely make no substantial difference in @xmath3 .
for example , s11 showed that increasing @xmath47 from 1 @xmath54 to 10 @xmath54 does not change @xmath3 for those regions where co is well shielded against the radiation field ( section 3.1 of s11 ) .
therefore , decreasing @xmath47 from 1 @xmath54 to 0.5 @xmath54 will make only a minor change in @xmath3 at large column densities .
increasing the current dgr of 5.3 @xmath5 10@xmath72 mag @xmath71 by a factor of @xmath42 will cause more rapid co formation , but @xmath3 will not be significantly influenced because co formation in the s11 model reaches chemical equilibrium rapidly by @xmath139 @xmath4 2 myr .
lastly , we do not expect that running the s11 model up to @xmath139 @xmath4 10 myr drastically increases @xmath3 , considering that the mhd simulation with initial @xmath37 = 100 @xmath7 in glover & mac low ( 2011 ) predicts only a factor of @xmath42 increase of the mass - weighted mean co abundance from @xmath139 @xmath4 5 myr to @xmath139 @xmath4 10 myr ( section 3.3 of glover & mac low 2011 ) .
note that changes in co abundance at @xmath139 > 2 myr are stochastic fluctuations after chemical equilibrium is achieved .
we therefore conclude that the input parameters used in the s11 model are reasonable for the comparison with the observations of perseus and small ( a factor of few ) changes in the input parameters will not result in significant changes in @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath3 .
considering that perseus has most likely reached chemical equilibrium , it provides a suitable testbed for investigating whether results from the time - depedent mhd simulation converge to the time - independent pdr model for molecular clouds that are evolved enough .
as shown in section 7.2.2 , the scaled s11 model produces the @xmath2(h@xmath0 ) distribution in excellent agreement with our observations as well as the modified w10 model .
this suggests that the time - dependent h@xmath0 formation model ( s11 ) is consistent with the time - independent h@xmath0 formation model ( w10 ) for a low - mass , old molecular cloud such as perseus .
our result agrees with krumholz & gnedin ( 2011 ) , who found that time - dependent effects on h@xmath0 formation become important only at extremely low metallicities @xmath67 @xmath10 10@xmath31 z@xmath25 .
while the median @xmath2(hi ) in the s11 model is also in reasonably good agreement with the observations , the simulated @xmath2(hi ) distribution is a factor of @xmath46 broader than the observed one and particularly shows a more extended tail toward small @xmath2(hi ) @xmath10 3 @xmath5 10@xmath22 @xmath7 .
this broad @xmath2(hi ) distribution in the mhd simulation likely results from strong compressions and rarefactions by turbulence and the predicted @xmath2(hi ) is on average a factor of @xmath42 smaller than the observed @xmath2(hi ) for a given @xmath2(h ) .
the discrepancy becomes significant at small @xmath2(h ) @xmath4 10@xmath100 @xmath7 , where the s11 model underesimates @xmath2(hi ) by up to a factor of @xmath410 .
finally , the s11 model predicts that @xmath2(hi ) increases with @xmath2(h ) , suggesting no minimum @xmath2(hi ) beyond which the rest of hydrogen is converted into h@xmath0 . in the modified w10 model with the `` core - halo '' structure , on the other hand , the diffuse halo remains atomic with @xmath2(hi ) @xmath4 9 @xmath5 10@xmath22 @xmath7 and the dense core is fully converted into h@xmath0 .
clearly , this discrepancy in @xmath2(hi ) between the simulation and the observations is significant and interesting .
one potential avenue in exploring this in the future is by using a mixture of neutral phases for initial conditions , mimicing in some way the `` core - halo '' structure in the modified w10 model . in the case of @xmath3
, we could not properly compare the s11 model with our observations because of the nontrivial scaling of @xmath3 for different line of sight depths .
instead , we found that the simulated @xmath3 becomes comparable to the observed @xmath3 only if the co emission is integrated for the full simulation box of 20 pc .
this suggests that the s11 model likely underestimates @xmath3 for the conditions relevant to perseus .
interestingly , we estimate @xmath2(co ) @xmath4 1 @xmath5 10@xmath163 @xmath7 for b5 , ic348 , b1 , and ngc1333 by using the @xmath27co(@xmath122 = 1 @xmath123 0 ) excitation temperatures , optical depths , and integrated intensities provided by pineda et al . ( 2008 ) and assuming @xmath2(co ) = 76@xmath2(@xmath27co ) ( lequeux 2005 ) .
this value is in reasonably good agreement with the simulated mean @xmath2(co ) @xmath4 5 @xmath5 10@xmath164 @xmath7 ( calculated from the smoothed , regridded , scaled , and thresholds applied s11 model ) .
this comparison suggests that the potential discrepancy in @xmath3 between the observations and the s11 model would result from the radiative transfer calculations and/or the difference in velocity range .
the velocity range of the co emission , @xmath165@xmath166 , directly affects @xmath3 via @xmath3 = @xmath167 and a smaller @xmath165@xmath166 would result in a smaller @xmath3 for the same @xmath168 .
while we could not compare specific @xmath3 values predicted by the s11 model with our observations , we found that the s11 model reproduces the observed shape of the @xmath3 vs @xmath9 profiles reasonably well .
this suggests that penetration of uv photons into the ism and ( dust and self- ) shielding against the isrf are relatively well captured in the co formation process by s11 .
in addition , we noticed that @xmath3 has a much larger scatter at small @xmath9 .
this could result from large density fluctuations in the turbulent medium .
the gas in the s11 model would be strongly compressed and rarefied by turbulence and the gas density at a given @xmath9 can vary over several orders of magnitude ( e.g. , figure 14 of glover et al .
2010 ) . in this case
, co can form in dense clumps even at small @xmath9 and @xmath3 therefore shows a large scatter .
this scatter is reduced at large @xmath9 where @xmath3 eventually saturates .
finally , turbulent mixing could spread the co distribution , contributing to the large scatter of @xmath3 .
in general , our study shows that the scaled mhd simulation by s11 is successful in reproducing @xmath2(h@xmath0 ) in perseus , which is a low - mass , old molecular cloud most likely in chemical equilibrium . on the other hand ,
future model adjustments are required to better match the observed @xmath2(hi ) and @xmath3 . we have revealed two important areas of future attention : ( 1 ) the role of diffuse halos in the formation of molecular gas and ( 2 ) the effect of density fluctuations and turbulent mixing in the spatial distribution of molecular gas . to investigate these two issues
, we plan to compare observations of several galactic molecular clouds with mhd simulations that explore different fractions of neutral phases and a varying degree of turbulence as initial conditions .
in particular , our future work will include molecular clouds less evolved and/or forming more massive stars ( therefore more turbulent ) than perseus , where the difference between the mhd and pdr models is likely to be more pronounced .
in this paper , we combine high - resolution h@xmath0 and co measurements to investigate @xmath1 across the perseus molecular cloud .
we derive the @xmath1 image at @xmath40.4 pc spatial resolution by using @xmath2(h@xmath0 ) estimated by lee et al .
( 2012 ) and @xmath3 provided by the complete survey .
we examine the large - scale spatial variations in @xmath1 across the cloud and their correlations with local ism conditions .
in addition , we focus on the characteristic dependence of @xmath1 on @xmath9 .
the @xmath2(hi ) , @xmath2(h@xmath0 ) , @xmath3 , and @xmath1 images allow us to test two theoretical models of h@xmath0 and co formation : the modified w10 model ( `` microturbulent time - independent model '' ) and the s11 model ( `` macroturbulent time - dependent model '' ) . for several dark and star - forming regions in perseus ( b5 , b1e / b1 , l1448 , ic348 , and ngc1333 )
, we investigate @xmath2(hi ) vs @xmath2(h ) , @xmath29 vs @xmath2(h ) , @xmath3 vs @xmath9 , and @xmath1 vs @xmath9 and compare the results with model predictions .
we summarize our main results as follows . 1
. we derive @xmath169 3 @xmath5 10@xmath6 for perseus .
this value is a factor of @xmath44 smaller than the previous estimate of @xmath1 @xmath4 1 @xmath5 10@xmath22 for the same cloud ( dame et al .
2001 ; pineda et al . 2008 ) and the discrepancy mainly results from different resolutions , dgrs , and our consideration of @xmath2(hi ) in deriving @xmath2(h@xmath0 ) .
we find a factor of @xmath43 region - to - region variations in @xmath1 . the northeastern part of perseus ( b5 and ic348 ) has on average larger @xmath1 than the southwestern part ( b1e / b1 , ngc1333 , and l1448 ) .
this could be explained by a stronger @xmath47 and/or a smaller @xmath88 in the northeastern part , although the correlations between @xmath1 and @xmath47/@xmath88 are mild .
additionally , variations in @xmath37 and/or @xmath9 could contribute to the observed regional variations in @xmath1 . within the individual dark and star - forming regions with a size of @xmath467 pc , @xmath1 varies up to a factor of @xmath4100 .
the observed @xmath1 vs @xmath9 profiles show two characteristic features : a steep decrease of @xmath1 at small @xmath9 and a gradual increase of @xmath1 at large @xmath9 . among the five dark and star - forming regions , ic348 and
b1e / b1 clearly show the transition from decreasing to increasing @xmath1 at @xmath9 @xmath4 3 mag .
4 . the modified w10 model with the `` core - halo '' density distribution
reproduces the observed @xmath1 vs @xmath9 profiles , particularly well for ic348 and b1e / b1 .
in addition , the model predicts a nearly constant @xmath2(hi ) @xmath4 9 @xmath5 10@xmath22 @xmath7 and a linear increase of @xmath29 with @xmath2(h ) , both consistent with what we find in perseus .
the modified w10 model with the uniform density distribution reproduces the observed @xmath2(h@xmath0 ) reasonably well but underestimates @xmath2(hi ) by a factor of @xmath410160 . as a result ,
the model overestimates @xmath29 for a given @xmath2(h ) by up to a factor of @xmath4300 .
in addition , while matching the observed saturation of @xmath3 at @xmath9 @xmath11 3 mag , the model predicts a more gradual increase of @xmath3 at @xmath9 @xmath10 3 mag .
this results in the @xmath1 vs @xmath9 profile shallower than the observations at @xmath9 @xmath10 3 mag .
the scaled s11 model predicts @xmath2(h@xmath0 ) in excellent agreement with what we estimate in perseus .
however , @xmath2(hi ) increases with @xmath2(h ) by a factor of @xmath47 in the model and this is in contrast with the observed small variation of @xmath2(hi ) with @xmath2(h ) ( less than a factor of 2 ) . while we do not compare specific @xmath3 values in the s11 model with the observed @xmath3 due to a complex issue of scaling @xmath3 for different line of sight depths , we stress that the simulated @xmath3 becomes comparable only when the co emission is integrated along the full simulation box of 20 pc , suggesting that the model likely underestimates @xmath3 for the conditions relevant to perseus . in addition , we find that the s11 model reproduces the observed shapes of @xmath3 vs @xmath9 and @xmath1 vs @xmath9 profiles reasonably well but with a large scatter particulary at small @xmath9 .
our study shows that @xmath1 can vary by up to a factor of @xmath4100 on @xmath40.4 pc scales and depends on local ism conditions such as @xmath47 , @xmath88 , @xmath37 , and @xmath9 .
the characteristic relation of @xmath1 with @xmath9 is mainly driven by how @xmath3 varies with @xmath9 . at small
@xmath9 , @xmath1 steeply decreases with @xmath9 , likely because co becomes sufficiently shielded against photodissociation and @xmath3 sharply increases .
@xmath1 then gradually increases with @xmath9 , likely due to the saturation of @xmath3 .
our results observationally confirm previous theoretical predictions of the @xmath1 vs @xmath9 profile for the first time .
however , the precise details of the @xmath1 vs @xmath9 profile , e.g. , the location where the transition from decreasing to increasing @xmath1 occurs , the slopes of the decreasing and increasing portions , etc . ,
again depend on local environmental parameters ( e.g. , taylor et al . 1993 ;
bell et al . 2006 ; shetty et al .
in general , our results suggest that a single @xmath1 can not be used to derive the spatial distribution of @xmath2(h@xmath0 ) across a molecular cloud .
the detailed comparison between our high - resolution data and theory provides important insights into h@xmath0 and co formation in molecular clouds .
for example , the good agreement we found with the modified w10 model suggests that the steady state and equilibrium chemistry and the microturbulent approximation for co spectral line formation and cooling work well for perseus on @xmath40.4 pc scales .
perseus appears to be old enough to achieve chemical equilibrium and the timescale of the dynamical process(es ) most revelant for the formation of molecular gas is likely @xmath11 3 myr .
however , the good agreement with the model is achieved only if the density distribution has a diffuse halo component . in the modified w10 model ,
the halo provides dust shielding against h@xmath0 and co photodissociation and is essential to reproduce the observed @xmath2(hi ) , @xmath29 , @xmath3 , and @xmath1 distributions .
while our results indicate the importance of the diffuse hi halo for the distributions of two most abundant molecular species , h@xmath0 and co , the properties of the halo have not been observationally well constrained . despite the lack of fine - tuning to match the characteristics of perseus , the s11 model reproduces the observed @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath3 properties reasonably well .
in particular , the predicted range of @xmath2(h@xmath0 ) in the scaled s11 model is in excellent agreement with our data .
these results suggest that the time - dependent chemistry model is generally consistent with the time - independent chemistry model for a low - mass , old molecular cloud such as perseus .
however , there are several interesting discrepancies and they likely result from the nature of turbulence in the s11 model . the strong compressions and rarefactions by turbulence could result in the wider range of @xmath2(hi ) in the s11 model and unlike the modified w10 model , there is no minimum @xmath2(hi ) beyond which the rest of hydrogen is fully converted into h@xmath0 .
in addition , density fluctuations in the s11 model allow the formation of dense clumps even at small @xmath9 and potentially result in a large scatter of @xmath3 .
turbulent motions could mix and spread the co distribution , likely contributing to the scatter of @xmath3 .
our future studies of other galactic molecular clouds , in particular those clouds much less evolved and/or forming more massive stars ( therefore more turbulent ) than perseus , will be important for comprehensive tests of the pdr and mhd models .
we sincerely thank the anonymous referee for suggestions that significantly improved this work .
we also thank chris carilli , paul clark , jay gallagher , miller goss , paul goldsmith , harvey liszt , adam leroy , jrgen ott , josh peek , and jaime pineda for stimulating discussions , tom dame for graciously providing his dame et al .
( 2001 ) data , and the galfa - hi and complete survey teams for making their data publicly available .
y.l . and s.s .
acknowledge support from the nsf grant ast-1056780 , nasa through contract 145727 issued by jpl / caltech , and the university of wisconsin graduate school .
, f.m . , and
acknowledge support from the deutsche forschungsgemeinschaft ( dfg ) via the sfb 881 ( b1 and b2 ) `` the milky way system '' and the spp ( priority program ) 1573 .
the arecibo observatory is operated by sri international under a cooperative agreement with the national science foundation ( ast-1100968 ) , and in alliance with ana g. mndez - universidad metropolitana , and the universities space research association .
we have made use of the karma visualization software ( gooch 1996 ) and nasa s astrophysics data system ( ads ) .
abdo , a. a. , ackermann , m. , ajello , m. , et al .
2010 , apj , 710 , 133 + akritas , m. g. & bershady , m. a. 1996 , apj , 470 , 706 + ballesteros - peredes , j. , vzquez - semadeni , e. , & scalo , j. 1999 , apj , 515 , 286 + ballesteros - peredes , j. , & mac low , m .- m .
2002 , apj , 570 , 734 + bally , j. , walawender , j. , johnstone , d. , kirk , h. , & goodman , a. a. 2008 , handbook of star forming regions , volume i : the northern sky , ed . b. reipurth ( asp monograph publications , vol .
4 ; san francisco , ca : asp ) , 308 + beaumont , c. n. , offner , s. s. r. , shetty , r. , glover , s. c. o. , & goodman , a. a. 2013 , apj , 777 , 173 + bell , t. a. , roueff , e. , viti , s. , & williams , d. a. 2006 , mnras , 371 , 1865 + bensch , f. 2006 , a&a , 448 , 1043 + bevington , p. r. & robinson , d. k. 2003 , data reduction and error analysis for the physical sciences , 3rd ed .
( boston : mcgraw - hill ) + bigiel , f. , leroy , a. k. , walter , f. , et al .
2008 , aj , 136 , 2846 + bohlin , r. c. , savage , b. d. , & drake , j. f. 1978 , apj , 224 , 132 + bolatto , a. d. , wolfire , m. g. , & leroy , a. k. 2013 , ara&a , 51 , 207 + boulanger , f. , abergel , a. , bernard , j .- p . , et al .
1996 , a&a , 312 , 256 + cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , apj , 345 , 245 + ernis , k. 1990 , ap&ss , 166 , 315 + cormier , d. , madden , s. c. , lebouteiller , v. , et al .
2014 , arxiv:1401.0563 + dalgarno , a. 2006 , proc .
, 103 , 12269 + dame , t. m. , hartmann , d. , & thaddeus , p. 2001
, apj , 547 , 792 + draine , b. t. 1978 , apjs , 36 , 595 + de vries , h. w. , thaddeus , p. , & heithausen , a. 1987 , apj , 319 , 723 + feldmann , r. , gnedin , n. y. , & kravtsov , a. v. 2012a , apj , 747 , 124 + feldmann , r. , gnedin , n. y. , & kravtsov , a. v. 2012b , apj , 758 , 127 + frerking , m. a. , langer , w. d. , & wilson , r. w. 1982 , apj , 262 , 590 + gratier , p. , braine , j. , rodriguez - fernandez , n. j. , et al .
2010 , a&a , 512 , 68 + glover , s. c. o. , & mac low , m .-
2007a , apjs , 169 , 239 + glover , s. c. o. , & mac low , m .-
2007b , apj , 659 , 1317 + glover , s. c. o. , federrath , c. , mac low , m .-
m . , & klessen , r. s. 2010 , mnras , 404 , 2 + glover , s. c. o. , & mac low , m .-
2011 , mnras , 412 , 337 + glover , s. c. o. , & clark , p. c. 2012 , mnras , 426 , 377 + gonzlez hernndez , j. , iglesias - groth , s. , rebolo , r. et al .
2009 , apj , 706 , 866 + gooch , r. 1996 , astronomical data analysis software and systems v ( asp conf .
101 ) , ed .
g. h. jacoby & j. barnes ( san francisco , ca : asp ) , 80 igoodman , a. a. , pineda , j. e. , & schnee , s. l. 2009 , apj , 692 , 91 + gratier , p. , braine , j. , rodriguez - fernandez , n. j. , et al .
2010 , a&a , 512 , 68 + heithausen , a. , bensch , f. , stutzki , j. , falgarone , e. , & panis , j. f. 1998 , a&a , 331l , 65 + heiles , c. , & troland , t. 2003 , apj , 586 , 1067 + herbig , g. h. , & jones , b. f. 1983 , aj , 88 , 1040 + herbig , g. h. 1998 , apj , 497 , 736 + heyer , m. h. , & brunt , c. m. 2004 , apj , 615 , 45 + hollenbach , d. j. , kaufman , m. j. , neufeld , d. , wolfire , m. , & goicoechea , j. r. 2012 , apj , 754 , 105 + hollenbach , d. j. , & tielens , a. g. g. m. 1997 , ara&a , 35 , 179 + imara , n. , & blitz , l. 2011 , apj , 732 , 78 + imara , n. , bigiel , f. , & blitz , l. 2011 , apj , 732 , 79 + indriolo , n. , & mccall , b. j. 2012 , apj , 745 , 91 + israel , f. p. 1997
, a&a , 328 , 471 + kainulainen , j. , beuther , h. , henning , t. , & plume , r. 2009 , a&a , 508 , 35 + kalberla , p. m. w. , burton , w. b. , hartmann , d. , et al .
2005 , a&a , 440 , 775 + kramer , c. , alves , j. , lada , c. et al .
1998 , a&a , 329l , 33 + knapp , g. r. 1974 , aj , 79 , 527 + kaufman , m. j. , wolfire , m. g. , hollenbach , d. j. , & luhman , m. l. 1999 , apj , 527 , 795 + kaufman , m. j. , wolfire , m. g. , & hollenbach , d. j. 2006 , apj , 644 , 283 + kennicutt , r. c. 1989 , apj , 344 , 685 + kennicutt , r. c. , & evans , n. j. 2012 , ara&a , 50 , 531 + krumholz , m. r. , mckee , c. f. , & tumlinson , j. 2009 , apj , 693 , 216 ( k09 ) + krumholz , m. r. , & gnedin , n. y. 2011 , apj , 729 , 36 + krumholz , m. r. 2012 , apj , 759 , 9 + lada , c. j. , lombardi , m. , & alves , j. f. 2010 , apj , 724 , 687 + lagos , c. , bayet , e. , baugh , c. m. , et al .
2012 , mnras , 426 , 2142 + lazarian , a. , pogosyan , d. , vzquez - semadeni , e. , & pichardo , b. 2001 , apj , 555 , 130 + lee , m .- y . , stanimirovi , s. , douglas , k. a. , et al .
2012 , apj , 748 , 75 + lequeux , j. 2005 , the interstellar medium ( berlin : springer ) + le bourlot , j. , pineau des forets , g. , roueff , e. , & flower , d. r. 1993 , a&a , 267 , 233 + le petit , f. , nehm , c. , le bourlot , j. , & roueff , e. 2006 , apjs , 164 , 506 + leroy , a. k. , bolatto , a. , stanimirovi , s. , et al .
2007 , apj , 658 , 1027 + leroy , a. k. , bolatto , a. , bot , c. , et al .
2009 , apj , 702 , 352 + leroy , a. k. , bolatto , a. , gordon , k. , et al .
2011 , apj , 737 , 12 + lombardi , m. , alves , j. , & lada , c. j. 2006 , a&a , 454 , 781 + luhman , k. l. , rieke , g. h. , lada , c. j. , & lada , e. a. 1998 , apj , 508 , 347 + mac low , m .-
& klessen , r. 2004 , rvmp , 76 , 125 + maloney , p. , & black , j. h. 1988 , apj , 325 , 389 + mathis , j. s. 1990 , ara&a , 28 , 37 + mckee , c. f. , & ostriker , e. c. 2007 , ara&a , 45 , 565 + magnani , l. , blitz , l. , & wouterloot , j. g. a. 1988 , apj , 326 , 909 + miville - desch@xmath170nes , m .- a . & lagache , g. 2005 , apjs , 157 , 302 + muench , a. a. , lada , e. a. , lada , c. j. , et al .
2003 , aj , 125 , 2029 + narayanan , d. , krumholz , m. r. , ostriker , e. c. , & hernquist , l. 2012 , mnras , 421 , 3127 + nelson , r. p. , & langer , w. d. 1999 , apj , 524 , 923 + norman , m. l. 2000 , rev . mex .
9 , 66 + ossenkopf , v. 1997 , new astron . , 2 , 365 + ostriker , e. c. , stone , j. , & gammie , c. f. 2001 , apj , 546 , 980 + papadopoulos , p. p. , thi , w .- f .
, & viti , s. 2002 , apj , 579 , 270 + paradis , d. , dobashi , k. , shimoikura , t. , et al .
2012 , a&a , 543 , 103 + peek , j. e. g. , heiles , c. , douglas , k. a. , et al .
2011 , apjs , 194 , 20 + pineda , j. e. , caselli , p. , & goodman , a. a. 2008 , apj , 679 , 481 + pineda , j. l. , goldsmith , p. f. , chapman , n. , et al .
2010 , apj , 721 , 686 + rahman , n. , bolatto , a. d. , xue , r. , et al .
2012 , 745 , 183 + reach , w. t. , koo , b .-
c . , & heiles , c. 1994 , apj , 429 , 672 + ridge , n. a. , di francesco , j. , kirk , h. , et al .
2006 , aj , 131 , 2921 + rogers , c. , heyer , m. h. , & dewdney , p. e. 1995 , apj , 442 , 694 + sakamoto , s. 1996 , apj , 462 , 215 + rubio , m. , lequeux , j. , & boulanger , f. 1993 , a&a , 271 , 9 + sakamoto , s. 1999 , apj , 523 , 701 + sancisi , r. goss , w. m. , anderson , c. , johansson , l. e. b. , & winnberg , a. 1974 , a&a , 35 , 445 + sandstrom , k. m. , leroy , a. k. , walter , f. , et al .
2013 , apj , 777 , 5 + schmidt , m. 1959 , apj , 129 , 243 + schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , apj , 500 , 525 + schruba , a. , leroy , a. k. , walter , f. , et al .
2011 , aj , 142 , 37 + shetty , r. , glover , s. c. , dullemond , c. p. , & klessen , r. s. 2011a , mnras , 412 , 1686 ( s11 ) + shetty , r. , glover , s. c. , dullemond , c. p. , et al . 2011b , mnras , 415 , 3253 + shetty , r. , kelly , b. c. , & bigiel , f. 2013 , mnras , 430 , 288 + smith , p. a. , brand , p. w. j. l. , mountain , c. m. , puxley , p. j. , & nakai , n. 1991 , mnras , 252 , 6 + sobolev , v. v. 1957 , sva , 1 , 678 + stone , j. m. , & norman , m. l. 1992 , apjs , 80 , 753 + taylor , s. d. , hartquist , t. w. , & williams , d. a. 1993 , mnras , 264 , 929 + tielens , a. g. g. m. , & hollenbach , d. 1985 , apj , 291 , 722 + wannier , p. g. , lichten , s. m. , & morris , m. 1983 , apj , 268 , 727 + wannier , p. g. , lichten , s. m. , andersson , b .-
g . , & morris , m. 1991 , apjs , 75 , 978 + williams , j. p. , & maddalena , r. j. 1996 , apj , 464 , 247 + wolfire , m. g. , hollenbach , d. , & tielens , a. g. g. m. 1993 , apj , 402 , 195 + wolfire , m. g. , mckee , c. f. , hollenbach , d. , & tielens , a. g. g. m. 2003 , apj , 587 , 278 + wolfire , m. g. , tielens , a. g. g. m. , hollenbach , d. , & kaufman , m. j. 2008 , apj , 680 , 384 + wolfire , m. g. , hollenbach , d. , & mckee , c. 2010 , apj , 716 , 1191 ( w10 ) + young , j. s. , goldsmith , p. f. , langer , w. d. , wilson , r. w. , & carlson , e. r. 1982 , 261 , 513 + * table 1 * + physical properties of the dark and star - forming regions + 0.1 cm [ cols="^,^,^,^,^,^ " , ] @xmath171 the values are provided for the minimum and maximum column densities ( @xmath9 = 0.6 mag and 10 mag ) .
+ @xmath172 the total size of the slab @xmath131 = @xmath2(h)/@xmath37 .
+ complete @xmath9 image of perseus overlaid with the complete @xmath3 contours ( ridge et al .
the contour levels range from 10% to 90% of the peak ( 80 k km s@xmath8 ) with 10% steps .
the angular resolution of the @xmath9 and @xmath3 images here is 5@xmath23 and 4.3@xmath23 respectively .
a number of dark ( b5 , b1e , b1 , and l1448 ) and star - forming regions ( ic348 and ngc1333 ) are labelled . ]
@xmath2(h@xmath0 ) image at 4.3@xmath23 angular resolution derived by lee et al .
the complete @xmath3 contours are overlaid in black and their levels range from 10% to 90% of the peak ( 80 k km s@xmath8 ) with 10% steps .
the blank data points correspond to point sources and regions with possible contamination ( the taurus molecular cloud and a background hii region ) .
see sections 4.2 and 4.3 of lee et al .
( 2012 ) for details . ]
co spectrum obtained by averaging the complete co spectra of all data points where the ratio of the peak main - beam brightness temperature to the rms noise is greater than 3 .
note that the co emission shows multiple velocity components .
] each of the five regions is overlaid on the complete @xmath3 image in different color .
b5 is purple , ic348 is blue , b1e / b1 is green , ngc1333 is orange , and l1448 is red . see section 4 for details on how we determined these regions . ]
@xmath1 image at 4.3@xmath23 angular resolution . ]
@xmath88 image overlaid with the complete @xmath3 contours .
the contour levels range from 10% to 90% of the peak ( 80 k km s@xmath8 ) with 10% steps . ]
histograms of @xmath47 .
b5/ic348 and b1e / b1/ngc1333/l1448 are combined to produce the histograms of _ east _ and _ west_. the median @xmath47 of _ east _ ( @xmath410@xmath110 erg @xmath7 s@xmath8 ) is shown as a dashed line . ]
complete @xmath3 as a function of @xmath9 for all five regions defined in section 4 . ] complete @xmath3 as a function of @xmath9 for each dark and star - forming region .
the mean 1@xmath101 uncertainty of @xmath3 ( @xmath40.09 k km s@xmath8 ) is shown as a dashed line , while that of @xmath9 ( @xmath40.2 mag ) is too small to be shown . ]
@xmath1 as a function of @xmath9 for each dark and star - forming region .
the median 1@xmath101 uncertainty of @xmath1 ( @xmath41.5 @xmath5 10@xmath124 ) is shown as a dashed line , while the mean 1@xmath101 uncertainty of @xmath9 ( @xmath40.2 mag ) is too small to be shown .
note that we show the median 1@xmath101 instead of the mean 1@xmath101 , because it is a better representative of the uncertainty in @xmath1 .
the right lowermost panel shows a normalized histogram of the 1@xmath101 uncertainty in @xmath1 and it is clear that the mean 1@xmath101 shown as a dashed - dot line corresponds to the high end of the distribution , affected by a small fraction of the data points with large uncertainties . ]
comparison with the modified w10 model for ic348 .
the mean 1@xmath101 uncertainties of @xmath2(hi ) ( @xmath43.5 @xmath5 10@xmath6 @xmath7 ) and @xmath3 ( @xmath40.09 k km s@xmath8 ) are shown as dashed lines , while those of @xmath2(h ) ( @xmath41.6 @xmath5 10@xmath22 @xmath7 ) , @xmath2(h@xmath0 ) ( @xmath43.3 @xmath5 10@xmath6 @xmath7 ) , @xmath29 ( @xmath40.09 ) , and @xmath9 ( @xmath40.2 mag ) are too small to be shown .
the median 1@xmath101 uncertainty of @xmath1 ( @xmath41.5 @xmath5 10@xmath124 ) is shown as a dashed line as well .
( left ) the modified w10 model with the `` core - halo '' structure is shown with red , green , and blue stars ( @xmath126 = 10@xmath38 , 5 @xmath5 10@xmath38 , and 10@xmath24 @xmath39 ) .
( right ) the modified w10 model with the uniform density distribution is shown with red , green , and blue squares ( @xmath37 = 10@xmath38 , 5 @xmath5 10@xmath38 , and 10@xmath24 @xmath39 ) . ] 0.3 cm fig .
[ f : w_compare]. ( continued ) ( a ) the difference between the observed mean @xmath2(hi ) and the simulated mean @xmath2(hi ) normalized by the observed mean @xmath2(hi ) , @xmath173 = @xmath165@xmath2(hi)/@xmath2(hi)@xmath174 , is plotted as a function of line of sight depth .
( b ) same as ( a ) but for @xmath2(h@xmath0 ) .
( c ) the sum of the two normalized differences in quadrature , @xmath175 , is plotted as a function of line of sight depth .
note that the line of sight depth of 7 pc results in the minimum discrepancy between our and simulated data . ] normalized histograms of @xmath2(hi ) , @xmath2(h@xmath0 ) , @xmath2(h ) , and @xmath3 .
the histograms in black , grey , and green are constructed using the data points with positive @xmath2(h@xmath0 ) , shown in figure [ f : complete_five_regions ] , and from the s11 model respectively . for the s11 model , the simulated @xmath2(hi ) , @xmath2(h@xmath0 ) , and @xmath2(h ) data are scaled for 7 pc , while the simulated @xmath3 data are not . see section 7.2.2 for details .
the mean 1@xmath101 uncertainties of @xmath2(h@xmath0 ) ( @xmath43.3 @xmath5 10@xmath6 @xmath7 ) and @xmath3 ( @xmath40.09 k km s@xmath8 ) are shown as dashed lines , while those of @xmath2(hi ) ( @xmath43.5 @xmath5 10@xmath6 @xmath7 ) and @xmath2(h ) ( @xmath41.6 @xmath5 10@xmath22 @xmath7 ) are too small to be shown . ]
while we found @xmath108 @xmath4 3 @xmath5 10@xmath6 for perseus , dame et al .
( 2001 ) and pineda et al .
( 2008 ) estimated @xmath41.2 @xmath5 10@xmath22 and @xmath41.4 @xmath5 10@xmath22 respectively .
these two studies are similar with our study in the sense that they utilized dust as a tracer of total gas column density , but applied different methodologies to derive @xmath1 .
we follow their methodologies in order to understand why our result is different .
dame et al .
( 2001 ) used the @xmath145 data from schlegel et al .
( 1998 ) and the hi data from the lab survey .
they estimated dgr on large - scales by smoothing both the @xmath145 and @xmath2(hi ) images to 10@xmath20 resolution and calculating the ratio of the smoothed @xmath145 and @xmath2(hi ) images .
the @xmath145 image was then divided by the large - scale dgr image and @xmath2(hi ) was subtracted to derive @xmath2(h@xmath0 ) .
the derived @xmath2(h@xmath0 ) was finally combined with @xmath3 from the cfa survey to estimate @xmath1 .
the resolution of the hi data was the lowest among all data sets and the estimated @xmath1 values were consequently at 36@xmath23 resolution .
we note that most other large - scale studies of @xmath1 in the milky way are also at 36@xmath23 resolution , limited by the lab hi data ( e.g. , abdo et al .
2010 ; paradis et al .
2012 ) . to show how different resolutions and methodologies affect the estimation of @xmath1
, we first compare our original data at 4.3@xmath23 resolution ( black histograms ; data points for all five regions ) with ( 1 ) our data smoothed to 36@xmath23 resolution ( grey histograms ) and ( 2 ) the data from dame et al .
( 2001 ) ( green histograms ) in figure [ f : compare_dame01 ] .
note that we use the cfa co data here to derive @xmath1 at 36@xmath23 resolution instead of the complete co data we used elsewhere in this paper , because of their larger spatial coverage ( @xmath410@xmath20 @xmath5 7@xmath20 for the cfa co vs @xmath46@xmath20 @xmath5 3@xmath20 for the complete co ) .
this will not cause any complication with our comparison , considering that @xmath483% of the data points are consistent within 1@xmath101 uncertainties when the cfa and complete co data are compared at the common resolution of 8.4@xmath23 . for each histogram in figure
[ f : compare_dame01 ] , we show the mean value of the distribution as a dashed line . in the case of @xmath1 , @xmath108 calculated as @xmath176@xmath2(h@xmath0)/@xmath176@xmath3 is shown instead . in comparison betwen our data at 4.3@xmath23 and 36@xmath23 resolutions , we find that @xmath108 increases from @xmath43 @xmath5 10@xmath6 ( 4.3@xmath23 ) to @xmath44.5 @xmath5 10@xmath6 ( 36@xmath23 ) .
@xmath108 increases because spatial smoothing affects the @xmath3 distribution slightly more than the @xmath2(h@xmath0 ) distribution . to be precise , @xmath3 decreases by a factor of @xmath46 from @xmath423.1 k km s@xmath8 to @xmath43.9 k km s@xmath8 on average , while @xmath2(h@xmath0 ) decreases by a factor of @xmath44 from @xmath46.9 @xmath5 10@xmath22 @xmath7 to @xmath41.7 @xmath5 10@xmath22 @xmath7 on average .
while spatial smoothing to 36@xmath23 resolution results in the slight increase of @xmath1 , there is still a factor of @xmath42.7 discrepancy between our @xmath177 4.5 @xmath5 10@xmath6 and the value derived by dame et al .
( 2001 ) for the same area .
because the same cfa co data were used , as shown from the good agreement between the grey and green histograms in figure [ f : compare_dame01](c ) , the discrepancy in @xmath1 would come from the difference in @xmath2(h@xmath0 ) and we indeed find that the mean @xmath2(h@xmath0 ) @xmath4 5.2 @xmath5 10@xmath22 @xmath7 in dame et al .
( 2001 ) is larger than our mean @xmath2(h@xmath0 ) @xmath4 1.7 @xmath5 10@xmath22 @xmath7 at 36@xmath23 resolution by a factor of @xmath43 .
considering that the equations for deriving @xmath2(h@xmath0 ) in our study and dame et al .
( 2001 ) are essentially the same , @xmath2(h@xmath0 ) = ( @xmath9/dgr @xmath66 @xmath2(hi ) ) @xmath5 0.5 , we compare our @xmath9 and @xmath2(hi ) data smoothed to 36@xmath23 resolution with the data from dame et al .
( 2001 ) in figures [ f : compare_dame01](d , f ) . to convert @xmath178(@xmath179 ) in dame et al .
( 2001 ) into @xmath9 , we use the total - to - selective extinction ratio @xmath180 @xmath4 3.1 for the diffuse ism ( mathis 1990 ) .
in addition , the local dgr @xmath4 1.1 @xmath5 10@xmath70 mag @xmath7 lee et al .
( 2012 ) derived for perseus is compared with the dgr data from dame et al .
( 2001 ) in figure [ f : compare_dame01](e ) . while we find that our @xmath9 at 36@xmath23 resolution is consistent with @xmath9 in dame et al .
( 2001 ) , our @xmath2(hi ) is slightly smaller than theirs by a factor of @xmath41.4 on average .
this difference mainly results from the fact that dame et al .
( 2001 ) integrated the hi emission along a whole line of sight , while our @xmath2(hi ) was derived by integrating the hi emission over the velocity range for perseus , from @xmath103 km s@xmath8 to @xmath181 km s@xmath8 ( section 3.1 ) . the slightly smaller @xmath2(hi ) in our study could affect the estimation of dgr and we indeed find that the local dgr for perseus is larger than the mean dgr in dame et al .
( 2001 ) by a factor of @xmath41.7 on average .
another factor that could affect dgr is spatial smoothing to 10@xmath20 resolution done by dame et al .
( 2001 ) . specifically , they blanked all pixels whose @xmath3 is larger than 1 k km s@xmath8 and replaced the blanked pixels with the gaussian - weighted @xmath178(@xmath179)/@xmath2(hi ) values , the gaussian having a fwhm of 10@xmath20 .
the angular size of 10@xmath20 is comparable to the size of perseus and in this case spatial smoothing could result in the inclusion of the diffuse ism with small dgr in the far outskirts of the cloud .
pineda et al . (
2008 ) used the @xmath9 and @xmath3 images from the complete survey smoothed to 5@xmath23 resolution and derived @xmath1 for perseus by fitting a linear function to @xmath3 vs @xmath9 under the following two assumptions : ( a1 ) all hydrogen traced by @xmath9 is in the form of h@xmath0 and ( a2 ) dgr = 5.3 @xmath5 10@xmath72 mag @xmath71 , the typical galactic value ( bohlin et al . 1978 ) .
considering that our study uses essentially the same data sets , the complete @xmath3 image and the @xmath9 image calibrated with the complete @xmath9 data , any difference in @xmath1 would come from different methodologies for deriving @xmath1 . in figure
[ f : compare_pineda08](a ) , we plot @xmath3 as a function of @xmath9 for perseus and fit a linear function to the data as pineda et al .
( 2008 ) did . to be consistent with pineda et al .
( 2008 ) , we smooth our @xmath3 and @xmath9 images with gaussian kernels to obtain a resolution of 5@xmath23 and regrid the images to a grid of 2.5@xmath23 .
in addition , we use their primary thresholds , i.e. , the co and @xmath27co integrated intensities are positive and the co and @xmath27co peak brightness temperatures are at least 10 and 5 times the rms noises of co and @xmath27co , to select data points .
we do not consider other thresholds adopted in pineda et al .
( 2008 ) , e.g. , exclusion of the data points with a stellar density larger than 10 stars per pixel , and expect that they will not make a significant change in the linear fit , considering that they account only a small fraction of the total number of data points ( @xmath47% ) . as pineda et al .
( 2008 ) performed , we fit the linear function @xmath182 using the bivariate correlated errors and intrinsic scatter estimator ( bces ; akritas & bershady 1996 ) and find @xmath48 = @xmath183 mag and @xmath184 = @xmath185 mag k@xmath8 km@xmath8 s. this result is consistent with pineda et al .
( 2008 ) , once our fitted parameters @xmath48 and @xmath184 are converted into the quantities in equation ( 18 ) of pineda et al .
( 2008 ) , @xmath186 = @xmath48 = @xmath183 mag and @xmath187 = @xmath184 @xmath5 9.4 @xmath5 10@xmath22 = ( @xmath188 ) @xmath5 10@xmath22 , where @xmath186 is the minimum @xmath9 below which there is no co emission and @xmath187 is essentially @xmath1 .
our @xmath187 = ( @xmath188 ) @xmath5 10@xmath22 is very close to @xmath187 = ( @xmath189 ) @xmath5 10@xmath22 in pineda et al .
( 2008 ) . in summary ,
our result is consistent with pineda et al .
( 2008 ) if we use exactly the same methodology for deriving @xmath2(h@xmath0 ) and @xmath1 .
however , instead of fitting a linear function to @xmath3 vs @xmath9 , we derive @xmath1 on a pixel - by - pixel basis in this paper . to investigate whether this could result in a significant difference
, we perform additional tests .
first , we derive @xmath1 by assuming ( a1 ) but with our dgr = 1.1 @xmath5 10@xmath70 mag @xmath71 .
the result is shown in figure [ f : compare_pineda08](b ) ( grey histogram ) , along with our original @xmath1 distribution ( black histogram ) .
second , we assume both ( a1 ) and ( a2 ) and still derive @xmath1 on a pixel - by - pixel basis .
the result is shown in the same panel as a green histogram .
in addition , @xmath108 for each histogram is calculated as @xmath176@xmath2(h@xmath0)/@xmath176@xmath3 and is shown as a dashed line .
we find @xmath108 @xmath4 4.9 @xmath5 10@xmath6 and 1 @xmath5 10@xmath22 for the first and second test respectively .
this suggests that our pixel - by - pixel derivation of @xmath1 is consistent with the linear fit method in pineda et al .
( 2008 ) and therefore the discrepancy between our @xmath108 and @xmath1 in pineda et al .
( 2008 ) results from the assumptions ( a1 ) and ( a2 ) .
specifically , the neglect of @xmath2(hi ) in derivation of @xmath2(h@xmath0 ) ( a1 ) results in a factor of @xmath41.6 difference in @xmath108 , while the use of the galactic dgr ( a2 ) results in an additional factor of @xmath42 difference .
comparison with dame et al .
the black and grey histograms represent the data from our study at 4.3@xmath23 and 36@xmath23 angular resolutions , while the green histograms show the data from dame et al .
all histograms are normalized for easy comparison .
note that the @xmath3 image from the cfa survey was combined with our @xmath2(h@xmath0 ) image to derive @xmath1 at 36@xmath23 angular resolution .
dashed lines show the mean values of individual quantities , except for those shown in ( a ) , which represent @xmath108 = @xmath176@xmath2(h@xmath0)@xmath190@xmath3 . the local dgr = 1.1 @xmath5 10@xmath70 mag @xmath71 lee et al .
( 2012 ) derived for perseus is shown as a grey solid line.,title="fig : " ] comparison with dame et al .
the black and grey histograms represent the data from our study at 4.3@xmath23 and 36@xmath23 angular resolutions , while the green histograms show the data from dame et al .
all histograms are normalized for easy comparison .
note that the @xmath3 image from the cfa survey was combined with our @xmath2(h@xmath0 ) image to derive @xmath1 at 36@xmath23 angular resolution .
dashed lines show the mean values of individual quantities , except for those shown in ( a ) , which represent @xmath108 = @xmath176@xmath2(h@xmath0)@xmath190@xmath3 .
the local dgr = 1.1 @xmath5 10@xmath70 mag @xmath71 lee et al .
( 2012 ) derived for perseus is shown as a grey solid line.,title="fig : " ] comparison with pineda et al .
( left ) @xmath3 vs @xmath9 .
the linear fit obtained from the bces is shown as a dashed line .
( right ) normalized histograms of @xmath1 .
the black histogram is our original @xmath1 at 4.3@xmath23 angular resolution , while the grey and green histograms show @xmath1 estimated by assuming ( a1 ) only and ( a1 ) @xmath99 ( a2 ) respectively .
dashed lines show @xmath191 @xmath176@xmath2(h@xmath0)/@xmath176@xmath3 . ]
|
we derive the co - to - h@xmath0 conversion factor , @xmath1 = @xmath2(h@xmath0)/@xmath3 , across the perseus molecular cloud on sub - parsec scales by combining the dust - based @xmath2(h@xmath0 ) data with the @xmath3 data from the complete survey .
we estimate an average @xmath1 @xmath4 3 @xmath5 10@xmath6 @xmath7 k@xmath8 km@xmath8 s and find a factor of @xmath43 variations in @xmath1 between the five sub - regions in perseus . within the individual regions ,
@xmath1 varies by a factor of @xmath4100 , suggesting that @xmath1 strongly depends on local conditions in the interstellar medium .
we find that @xmath1 sharply decreases at @xmath9 @xmath10 3 mag but gradually increases at @xmath9 @xmath11 3 mag , with the transition occuring at @xmath9 where @xmath3 becomes optically thick .
we compare the @xmath2(hi ) , @xmath2(h@xmath0 ) , @xmath3 , and @xmath1 distributions with two models of the formation of molecular gas , a one - dimensional photodissociation region ( pdr ) model and a three - dimensional magnetohydrodynamic ( mhd ) model tracking both the dynamical and chemical evolution of gas .
the pdr model based on the steady state and equilibrium chemistry reproduces our data very well but requires a diffuse halo to match the observed @xmath2(hi ) and @xmath3 distributions .
the mhd model matches our data reasonably well , suggesting that time - dependent effects on h@xmath0 and co formation are insignificant for an evolved molecular cloud like perseus .
however , we find interesting discrepancies , including a broader range of @xmath2(hi ) , likely underestimated @xmath3 , and a large scatter of @xmath3 at small @xmath9 .
these discrepancies most likely result from strong compressions / rarefactions and density fluctuations in the mhd model .
|
the intrinsic galaxy alignment refers to a cosmic phenomenon that the orientations of the galaxy spin axes ( or galaxy ellipticities ) are not random but locally aligned between neighbors due to the spatial correlations of the initial tidal field .
the topic of the intrinsic galaxy alignments has recently drawn considerable attention in the field of the large - scale structure for three reasons .
firstly , since it is believed to be induced by the spatial correlations of the initial tidal field , it will be possible in principle by measuring the galaxy intrinsic alignments to reconstruct the initial density field on galactic scales @xcite . secondly , the initial tidal correlations are also responsible for the web - like distribution of the large scale structures @xcite .
thus , the galaxy intrinsic alignments will be useful to characterize and quantify the influence of the large - scale structures on the orientations of the galaxies distributed in a cosmic web @xcite .
thirdly and most importantly , the galaxy intrinsic alignments could cause non - negligible contamination on weak lensing signals .
plenty of efforts have been made so far to distinguish and separate the intrinsic contaminations from the galaxy extrinsic alignments caused by the weak gravitational lensing effect @xcite .
for these reasons , it is of fundamental importance to predict the degree of galaxy intrinsic alignments with a valid theory and measure quantitatively their signals in practice .
in fact , @xcite have provided an analytic model for the intrinsic galaxy spin alignments in the frame of the linear tidal torque theory @xcite . according to their model , the galaxy intrinsic alignments
can be expressed in terms of a quadratic scaling of the density correlation function and thus are expected to exist only at small distances of order of a few mpc .
a crucial implication of their analytic model is that due to the quadratic scaling nature of the galaxy intrinsic alignments the cross - correlations between the gravitational lensing shears from large scale density field and the galaxy intrinsic alignments with local tidal shears will be zero .
@xcite tested their analytic predictions against low - redshift samples of spiral galaxies and found that the observed signals are consistent with the analytic predictions . yet , their detections suffered from poor - number statistics and thus were still tentative .
later , several other authors have confirmed the existence of galaxy intrinsic alignments at small scales either in numerical simulations or in low - redshift observational surveys @xcite .
however , @xcite has pointed out that it might be invalid to describe the galaxy intrinsic alignments on large scales as a quadratic function of the density correlation .
their logic is as follows : since the quadratic scaling is based on the linear tidal torque theory which adopts a somewhat oversimplified assumption that the tidal field is gaussian in the subsequent evolutionary stages , it should not be a good approximation to describe the galaxy intrinsic alignments on large scales . in reality
, the density fluctuations will develop non - gaussianity via gravity which would in turn lead to non - negligible contributions of the nonlinear - order of the tidal tensors to the generation of the galaxy angular momentum . due to this nonlinear effect on the galaxy angular momentum ,
their intrinsic spin alignments should be better approximated as linear scaling with the density correlation function . because the linear scaling of the density correlation drops much slowly than the quadratic scaling
, the intrinsic spin alignments would not be completely negligible even on large scales .
if their claims turn out to be true , then it will have a significant impact not only on the weak lensing analysis but also on our fundamental understanding of the evolution of the tidal alignments . in the light of their claims ,
the following questions naturally arise ; does the nonlinear tidal effect on the galaxy intrinsic alignments really exist to a non negligible degree ?
if so , at what epochs and on which scales does its contribution begin to be significant ? does it depend on the intrinsic properties of the galaxies or dark halos ?
our goal here is to answer the above questions using both analytical and numerical methods .
the outline of this paper is as follows : 2 , we overview briefly the previous analytic model for the galaxy spin - spin alignments based on the linear tidal torque theory and propose a new model to account for the nonlinear tidal effect .
3 , we report a detection of the signals of the nonlinear tidal effect on the intrinsic spin alignments of dark matter halos simulated in n - body experiments and show how the signals depend on redshift , scale , and the halo intrinsic property .
4 , we summarize our results and discuss the implications of our work on the weak lensing effect .
the linear tidal torque theory explains that unless a proto - halo has a perfectly spherical shape , it can acquire spin angular momentum at first order through its tidal interaction with the surrounding matter @xcite .
the main prediction of the linear tidal torque theory is that the proto - halo angular momentum vector , ( @xmath4 ) , is proportional to the anti - symmetric product of the two tensors , the inertial momentum tensor @xmath5 and the local tidal shear tensor , @xmath6 , as @xmath7 here the defintions of @xmath8 and @xmath9 are given as @xmath10 where @xmath11 is the lagrangian position of the particles that reside in the proto - halo regions , @xmath12 and @xmath13 are the lagrangian volume and the density of the proto - halo region , respectively , and @xmath14 is the velocity perturbation potential . given the property of the perfectly anti - symmetric tensor @xmath15 in equation ( [ eqn : ang ] ) , an additional condition has to be satisfied for the first - order generation of the proto - halo angular momentum : the principal axes of @xmath8 and @xmath9 have to be misaligned with each other @xcite .
furthermore , a crucial implication of equation ( [ eqn : ang ] ) is that if this additional condition is satisfied and thus the angular momentum of a proto - halo is generated at first order , then the direction of the proto - halo angular momentum is not random but preferentially aligned with the principal axes of the local tidal tensor @xcite .
numerical experiments have revealed that the principal axes of @xmath5 and @xmath6 are correlated strongly but not perfectly @xcite , which indicates that the first - order generation of the proto - halo angular momentum is not so efficient and the degree of the alignments between the proto - halo spin directions and the principal axes of the local tidal shear tensors would not be so high . motivated by this numerical clue , @xcite suggested the following quadratic formula to quantify the expected degree of the alignments between the halo spins and the local tidal shears , generalizing the linear tidal torque theory : @xmath16 where @xmath17 is the unit spin vector of a halo , @xmath18 is the unit traceless tidal tensor smoothed on the halo mass scale , and @xmath19 is a correlation parameter in the range of @xmath20 $ ] which measures the strength of the correlation between @xmath17 and @xmath18 .
it is worth recalling the fact that the ensemble average @xmath21 in the left hand side of equation ( [ eqn : stc ] ) is obtained by taking the average of @xmath22 from the sample halos having all different mass , while the tidal shear field in the right hand side of equation ( [ eqn : stc ] ) is smoothed on the single mass scale which amounts to the mean halo mass ( i.e. , the mean value of the mass of the sample halos).@xcite derived equation ( [ eqn : stc ] ) empirically , relating the angular momenta of the halos with different mass to the tidal shear field smoothed on the mean halo mass .
the correlation parameter @xmath19 in equation ( [ eqn : stc ] ) which characterizes equation ( [ eqn : stc ] ) is introduced to take into account strong correlations between @xmath5 and @xmath6 and any modification in the subsequent evolution after the moment of the turn - around .
the two extreme cases of @xmath23 and @xmath24 represent no correlation and maximum correlation between @xmath17 and @xmath18 , respectively .
it is expected that the true value of @xmath19 would be between these two extremes .
the linear tidal torque theory itself can not make any prediction on the value of @xmath19 .
thus , the true value of this free parameter @xmath19 has to be determined empirically .
so far , several attempts have been made to measure the value of @xmath19 either from numerical simulations or from observations @xcite .
although no general consensus has been reached on the true value of @xmath19 , most of the numerical and observational evidences indicated that the true value of @xmath19 deviates from zero . a difficulty in constraining the value of @xmath19 lies in the fact that it is hard to measure the principal axes of the linear tidal field in real space . an alternative approach to the true value of @xmath19 is to measure the intrinsic spin - spin correlations of dark halos @xcite .
the local alignments between the halo spin vectors and the tidal shear tensors will lead to spatial spin alignments between the neighbor halos . using equation ( [ eqn : stc ] ) , @xcite derived the following analytic model for the halo spin - spin correlations : @xmath25 here , the constant @xmath26 represents the value of @xmath27 for the case of no alignment , and @xmath28 is the linear two - point correlation function defined as @xmath29 where @xmath30 is the dimensionless power spectrum and @xmath31 is the top - hat spherical filter of scale radius , @xmath32 , which is related to the halo mass scale @xmath33 as @xmath34^{1/3}$ ] with the mean density @xmath35 .
note that equation ( [ eqn:1st ] ) is proportional to the square of the two point density correlation @xmath36 , which means that the halo spin - spin correlations would decrease very rapidly .
for instance , on a galactic scale ( @xmath37mpc ) , it is expected that the halo spin - spin correlations exist only at distances of order of a few mpcs and effectively vanishes at larger distances . since the tidal field on different scales
are cross - correlated , equation ( [ eqn : stc ] ) leads to the existence of the cross - correlations of the spin axes of halos on different mass scales . in accordance with equation ( [ eqn:1st ] ) the halo spin cross - correlations on different scales ( @xmath38 and @xmath39 ) are modeled as @xmath40 where @xmath41 is the density cross - correlation function defined as @xmath42 it is worth mentioning that the value of @xmath19 for the cross correlation @xmath43 would not be same as its value for the auto - correlation @xmath44 in equation ( [ eqn:1st ] ) due to the difference in the smoothing scale for the tidal field . note that equation ( [ eqn : crr ] ) is a modified version of equation ( [ eqn:1st ] ) , both of which are based on equation ( [ eqn : stc ] ) .
the correlation parameter @xmath19 in equation ( [ eqn:1st ] ) represents the spin - spin correlations caused by the spatial correlations of the tidal fields smoothed on the same scale ( one mean halo mass scale , @xmath32 ) . while the correlation parameter @xmath19 in equation ( 6 ) represents the spin - spin correlations caused by the spatial correlations of the tidal fields on two different scales ( two mean halo mass scales , @xmath38 and @xmath39 ) .
since the spatial correlations of the two tidal fields smoothed on two different scales , @xmath38 and @xmath39 ( say , @xmath45 ) , are lower than the case that the two tidal fields are smoothed on the same scale of @xmath39 and higher than the case that they are smoothed on the same scale of @xmath38 , it is naturally expected that the value of @xmath19 for the cross - correlation @xmath43 should be different from that for the correlation @xmath44 .
as mentioned in 1 , it was @xcite who first pointed out that equation ( [ eqn:1st ] ) is valid only if the gravitational tidal field is gaussian .
they argued that the non - gaussianity in the tidal field should not be negligible in the nonlinear regime and suggested that the halo spin - spin alignments should be better described as a linear scaling with the density correlation function on sufficiently large scales .
inspired by the work of @xcite , here we suggest a new formula to model the halo spin - spin correlations : @xmath46 where a nonlinear correlation parameter @xmath47 is introduced to measure the strength of the nonlinear tidal effect on the halo spin - spin correlations . in this model
the linear correlation parameter is notated as @xmath48 since its value could be different between the two cases that the nonlinear effect is ignored as in equation ( [ eqn:1st ] ) and taken into account as in ( [ eqn:2nd ] ) .
equation ( [ eqn:2nd ] ) implies that if the nonlinear contribution exists , then the halo spin - spin correlations must exist on larger scale than predicted by the linear model ( eq.[[eqn:1st ] ] ) since it scales linearly with the density correlation function in the nonlinear model .
that is , the nonlinear tidal effect would generate large - scale correlations of the halo spin axes .
of course , the values of the nonlinear correlation parameter @xmath47 as well as the linear correlation parameter have to be determined empirically . the cross - correlation
, @xmath43 can be also modeled in accordance with ( [ eqn:2nd ] ) as @xmath49 it is worth mentioning that the value of @xmath50 for the cross correlation @xmath43 would not be necessarily lower than its value for the auto - correlation @xmath44 in equation ( [ eqn:2nd ] ) , unlike the case of the linear tidal torque model in 2.2 . in the nonlinear regime where the non - gaussianity in the density field grows ,
the occurrence of the halo merging is quite frequent .
this occurrence of the halo merging plays a role of transferring the orbital angular momentum generated by the external tidal field into the spin angular momentum of a merged halo .
in other words , the orbital angular momentum of a system composed of small individual halos become the spin angular momentum of a large halo formed through merging of the small halos .
given that the orbital angular momentum of the system before merging is generated by the external tidal field on larger scale @xcite , the transfer of the orbital angular momentum into the spin angular momentum creates cross - correlations of the tidal fields between different scales , which in turn generates large - scale cross - correlations @xmath51 in the nonlinear regime .
we use the halo catalogs from the millennium run simulation of @xmath52 dark matter particles for a @xmath53cdm universe with the cosmological parameters given as @xmath54 , @xmath55 , @xmath56 and @xmath57 @xcite .
the simulation was performed in a periodic box of linear size of @xmath58mpc to follow the evolution of all particles from @xmath59 to @xmath3 , each of which has a mass of @xmath60 .
we analyze the halo catalogs at four different redshifts @xmath3 , @xmath61 , @xmath62 and @xmath1 which contain information on the halo mass , comoving position , comoving peculiar velocity and the spin vector . at each redshift , we restrict our attention only to those halos which have more than @xmath63 particles for better accuracy in the measurement of the halo spin vector
. it amounts to selecting only halos which are more massive than @xmath64 .
basically , we measure the spatial correlations of the spin axes between neighbor halos as a function of separation distance and investigate how the correlation strength changes with redshift ( z ) , halo mass ( @xmath33 ) , velocity ( @xmath65 ) and specific angular momentum ( @xmath66 , angular momentum per mass ) .
then , we compare the numerical results with the linear and the nonlinear analytic models described in 2 to determine the best - fit values of the correlation parameters , @xmath19 , @xmath67 , and @xmath47 .
figure [ fig : z ] plots the halo spin - spin correlation function @xmath44 at @xmath3,@xmath61,@xmath62 and @xmath1 in the top - left , top - right , bottom - left , and bottom - right panel , respectively . in each panel ,
the solid dots correspond to the numerical results from the millennium data while the solid and the dashed lines represent the nonlinear model ( eq.[[eqn:2nd ] ] ) and the linear model ( eq.[[eqn:1st ] ] ) , respectively .
the case of no alignment is also shown as a dotted line in each panel for comparison . the errors of the numerical results are calculated as one standard deviation for the case of no alignment .
basically , it is computed by the formula of @xmath68 where @xmath69 is the number of halos belonging to each bin @xcite .
note that the sizes of the vertical errors in figure [ fig : z ] depend on the value of @xmath69 which in turn depends on the bin size of the spatial separation @xmath70 .
here we choose the bin size of @xmath70 as @xmath71mpc for all plots .
we have tested whether or not the final results depend on the bin size of @xmath70 and confirmed the robustness of the final results . for the analytic models ,
the correlation parameters are determined through fitting to the numerical results by means of the @xmath72-minimization . for the evaluation of the analytic models , we use the formula for the linear power spectrum given by @xcite using the same values of the cosmological parameters that are used for the millennium run simulations and the shape parameter @xmath73 ( private communication with v. springel ) .
table [ tab : z ] lists the number of halos ( @xmath74 ) , the mean mass ( @xmath75 ) , the best - fit values of @xmath19 , @xmath76 and @xmath47 at @xmath77 . here
, the mean mass @xmath75 for each bin is used to find the lagrangian smoothing scale , @xmath32 which has to be implemented into the analytic models ( eqs.[eqn:1st ] -[eqn:2nd ] ) .
as shown in figure [ fig : z ] , we detect clear signals of the halo spin - spin alignments within a few @xmath78mpc distances at all four redshifts .
note , however , a conspicuous difference between the results at low redshifts ( @xmath3 and @xmath61 ) and at higher redshifts ( @xmath79 and @xmath1 ) . at @xmath79 and @xmath1
, the correlation signals rapidly decrease with distance and disappear at distances around @xmath2mpc .
thus , for the low redshifts , the linear model ( dashed line ) for itself fits the numerical data pretty well .
whereas , at @xmath3 and @xmath61 the correlation signals decrease rather slowly , still existent to a nennegligible level at large separations beyond @xmath2mpc .
thus , at low redshifts the nonlinear model ( solid line ) fits the numerical data much better than the linear model , predicting the existence of the large - scale correlations .
this phenomena can be quantified in terms of the bestfit - values of the correlation parameters listed in table [ tab : z ] . at @xmath1 ,
the best - fit value of @xmath47 is effectively zero within one standard deviation ( @xmath80 ) while the best - fit value of @xmath48 is higher than @xmath81 ( not marginalized level ) .
in contrast , at @xmath3 the value of @xmath47 deviates from zero as significantly as @xmath82 while the value of @xmath48 decreases significantly to a @xmath83 level .
it is , however , worthing mentioning here that the standard deviation , @xmath80 , here is not marginalized .
figure [ fig : cz ] plots the @xmath84 , @xmath85 and @xmath86 contours for @xmath48 and @xmath47 at @xmath77 , showing the degeneracy in putting the constraints on the two parameters .
as can be seen , at @xmath3 and @xmath61 , the values of @xmath47 deviate from zero at higher than @xmath86 level .
our numerical detection implies that the linear model for the halo spin - spin correlations ( eq.[[eqn:1st ] ] ) works well at @xmath87 but the nonlinear effect indeed dominates at low redshifts ( @xmath88 ) , producing large - scale correlations of the halo spins , which can be well described by our nonlinear model ( eq.[[eqn:2nd ] ) .
we define the magnitude of the rescaled specific angular momentum ( angular momentum per unit mass ) as @xmath89 where @xmath90 and @xmath91 are the halo angular momentum magnitude and the halo mass in unit of @xmath92 , respectively . to determine quantitatively how the halo spin - spin correlation changes with @xmath66
, we classify the halos at @xmath3 into four logarithmic bins of @xmath66 and measure the correlations separately for each bin .
figure [ fig : a ] plots @xmath44 at four logarithmic bins of @xmath66 in the top - left , top - right , bottom - left , and bottom - right panels , respectively . table [ tab : a ] also lists the number of halos , the mean mass , the best - fit values of @xmath19 , @xmath48 and @xmath47 for each bin , and figure [ fig : ca ] plots the @xmath84 , @xmath85 and @xmath86 contours for @xmath48 and@xmath47 at the four bins .
as shown in figure [ fig : a ] , in the bins of low-@xmath66 halos ( top two panels ) , the correlations are rather weak and rapidly diminish to zero at separations greater than a few @xmath78mpc .
thus , for this case the linear model with the single parameter @xmath19 fits the data pretty well .
in other words , for the spin - spin correlations of low-@xmath66 halos , the values of @xmath47 are effectively zero as shown in table [ tab : a ] .
while in the bins of high-@xmath66 halos , the correlations are stronger , still existent even at separations as large as @xmath2mpc . therefore for the spin - spin correlations of high-@xmath66 halos , the nonlinear model with non - zero value of @xmath47 fits the data much better than the linear model .
note that for the case of highest-@xmath66 halos , the value of @xmath47 is approximately @xmath93 , which is five times greater than the average value of @xmath94 at @xmath3 .
this phenomena that the halos with higher specific angular momentum tend to have stronger non - linear ( and linear ) spin - spin correlations may be explained as follows .
if a halo is located in a region where the effect of the tidal field is very strong , then the halo will possess high specific angular momentum .
furthermore , its spin direction will be strongly correlated with that of its neighbors since the strong effect of the tidal field tends to diminish any randomization of the spin orientations that can occur in the subsequent stages.therefore , those halos which have higher specific angular momentum are likely to show strong spin - spin correlations under the strong effect of the tidal field .
we have also explored whether this phenomena occur at higher redshifts @xmath95,@xmath62 and @xmath1 and found that this is still the case .
the halos at @xmath3 are classified into four bins according to the value of the rescaled mass @xmath91 and their spin - spin alignments are measured separately for each bin .
figure [ fig : m ] plots @xmath44 , table [ tab : m ] lists the number of halos , the mean mass , the best - fit values of @xmath19 , @xmath48 and @xmath47 , and figure [ fig : cz ] plots the @xmath84 , @xmath85 and @xmath86 contours for @xmath48 and @xmath47 for the four bins .
as shown in figure [ fig : m ] , in the third bin which includes the galactic halos with mass of order of @xmath96 ( bottom - left panel ) , the correlation signal is strongest , existent at separations as large as @xmath2mpc . for this bin , the nonlinear model with nonzero value of @xmath47
gives a better fit the numerical results than the linear with model single parameter @xmath19 .
the best - fit value of @xmath47 deviates from zero by a factor of @xmath83 as shown in table [ tab : m ] . whereas in the first and the fourth bins which include dwarf halos and galaxy - group halos , respectively , the signals are rather weak and the correlations disappear at large distances . for these bins , the linear model for itself
fits the numerical data quite well and the values of @xmath47 are effectively zero .
this result implies that the strength of the nonlinear effect depends on mass scales , being strongest on the galactic mass scale .
note , however , here that the value of the characteristic mass scale changes with @xmath97 .
thus , the results given in 3.2 and 3.4 are related to each other . as mentioned in 2.3 ,
if the nonlinear effect dominates on the halo spin - spin correlations , then it is likely to produce stronger cross - correlations between different mass scales ( eq.[[eqn : crn ] ] ) .
we measure the halo spin cross - correlations between the halos of different mass bins that are constructed in 3.2 ( the first bin corresponds to the lowest mass scales .
see table [ tab : m ] ) , and find the best - fit values of @xmath19 , @xmath76 and @xmath47 by fitting the numerical data to the analytic model ( eq.[[eqn : crn ] ] ) .
figure [ fig : b ] plots the numerical results of @xmath51 for the six different cases : @xmath98 , @xmath99 , @xmath100 , @xmath101 , @xmath102,@xmath103 , where @xmath104 represents the cross - correlations between the @xmath105-th and the @xmath106-th mass bins .
as can be seen , the nonlinear model basically gives better fit to the numerical result for all cases , indicating that the nonlinear effect increases indeed increase spin cross - correlations between different mass scales , as presumed in 2.3 .
figure [ fig : cb ] plots the @xmath84 , @xmath85 and @xmath86 contours for @xmath48 and @xmath47 for these two cases of @xmath107 and @xmath108 where the strong signals of the cross - correlations are detected .
obviously , the values of @xmath47 for each case deviates from zero at @xmath85 confidence level .
we also investigate the dependence of the halo spin - spin correlations on the magnitude of halo peculiar velocity @xmath65 in unit of km / s .
the correlations @xmath44 are plotted in fig.[fig : v ] at four logarithmic bins of @xmath65 , and the @xmath84 , @xmath85 and @xmath86 contours for the correlation parameters of the nonlinear models are shown in fig.[fig : cv ] .
as shown in figure [ fig : v ] , the dependence of the halo spin - spin correlations on the velocity magnitude turns out not to be so strong . the significant nonlinear effect on the halo - halo correlation
is found only in the first bin of low-@xmath65 halos ( top - left panel ) , for this bin , the linear model fails in fitting the numerical data while the nonlinear model gives a better fit , predicting the large - scale correlations . as shown in the top - left panel of fig.[fig : cv ] , the value of @xmath47 for the case of lowest-@xmath65 halos deviates from zero at @xmath85 confidence level .
we summarize our results in the following : * to account for the possible nonlinear tidal effect on the intrinsic halo alignments , we suggest a new model characterized by two free parameters , @xmath48 and @xmath47 .
the two parameters @xmath48 and @xmath47 represents the strength of the linear and the nonlinear tidal effect on the orientations of the halo spins , respectively .
if @xmath109 , it corresponds to the case that the nonlinear tidal effect is negligible and the halo spin - spin correlations is a quadratic scaling of the density correlation function @xmath36 , existent on small scales . while if @xmath47 deviates from zero , the nonlinear tidal effect is nennegligible and the halo spin - spin correlations is a linear scaling of @xmath36 , present on large distance scales
* using the halo catalogs from numerical simulations at @xmath110 and @xmath1 , we have measured the halo spin - spin correlations and determined the best - fit values of @xmath48 and @xmath47 .
it is found that the values of @xmath48 increases with @xmath97 while @xmath47 decreases with @xmath97 . at @xmath79 and @xmath1 ,
the values of @xmath47 are found to be negligibly small , while at @xmath95 and @xmath3 @xmath47 has an order of @xmath111 higher than @xmath112 . especially at @xmath3
, @xmath48 is effectively zero and @xmath47 deviates from zero by a factor more than @xmath81 , which implies that at present epoch the nonlinear tidal effect is dominant inducing halo spin correlations on scales as large as @xmath2mpc .
* we have investigated the dependence of @xmath48 and @xmath47 on the halo intrinsic properties such as mass ( @xmath33 ) , specific angular momentum magnitude ( @xmath66 ) , and velocity magnitude ( @xmath65 ) at @xmath3 .
it is found that the value of @xmath47 depends most sensitively on the value of @xmath66 .
for those halos with highest @xmath66 , the value of @xmath47 reaches as high as @xmath113 and the value of @xmath48 is effectively zero . regarding the dependence on @xmath33 and @xmath65 , it is found that @xmath114 is less sensitive . however , for those halos with @xmath115 @xmath116km / s , the value of @xmath47 tend to be higher .
* we have measured the cross - correlations of the spin axes between halos on different mass scales .
it is found that on two different galactic mass scales ( @xmath117 and @xmath118 , the cross - correlations exist at scales as large as @xmath2mpc , due to the growth of the non - gaussianity increases cross - correlations of the tidal field on different scales .
our results have shown that the galaxy intrinsic alignments can exist not only for the case that the separation distance between the two galaxies is small but also for the case that any galaxy pair separated by distances larger than @xmath2mpc can still have aligned axes .
it also implies preferential e - mode contamination of the galaxy intrinsic alignments in weak lensing signals @xcite .
in addition , we have also found quantitatively that the galaxy intrinsic alignments could be cross - correlated with the larger - scale tidal field .
these results imply that the intrinsic galaxy correlations are indeed correlated with the surrounding density field on large scales , which will in turn lead to correlation between the gravitational lensing shears and the intrinsic galaxy alignments , ( often called , gi correlations ) .
recently , @xcite claimed that a signal of the gi correlations has been detected from the large low - redshift galaxy surveys ( see also , * ? ? ? * ) . according to their claim ,
the gi correlations exist to a nennegligible level for the bright luminous red galaxies ( lrgs ) and it is zero for the faint blue galaxies , mentioning an uncertainty in the estimation of the gi correlation for the faint blue galaxies . according to our numerical detection , the nonlinear tidal effect which causes the large - scale intrinsic alignments is strongest for the halos with high specific angular momentum magnitude and low peculiar velocity , which usually correspond to the galaxies with low surface brightness ( * ? ? ?
* and references therein ) and blue color @xcite .
thus , our results indicate that the gi correlation amplitude for the faint blue galaxies , may exceed the estimation upper limit , as noted by @xcite .
it will be of interest to compare the cross - correlations between the spin axes of the blue galaxies and the shapes of the red galaxies for the comparison of their correlation scales .
the millennium run simulation used in this paper was carried out by the virgo supercomputing consortium at the computing centre of the max - planck society in garching .
we thank v. springel and g. lemson for useful comments .
we also thank an anonymous referee for helpful suggestions .
is very grateful to the warm hospitality of l. kofman and the canadian institute for theoretical astrophysics ( cita ) where this work was initiated and performed .
j.l . acknowledges the financial support from the korea science and engineering foundation ( kosef ) grant funded by the korean government ( most , no . r01 - 2007 - 000 - 10246 - 0 ) .
aragon - calvo , m. a. , van de weygaert , r. , jones , b. j. t. , & van der hulst , j. m. 2007 , , 655 , 5 bardeen , j. m. , bond , j. r. , kaiser , n. , & szalay , a. s. 1986 , , 304 , 15 bond , j. , r. , kofman , l. , & pogosyan , d. 1996 , nature , 380 , 603 brown , m. l. , taylor , a. n. , hambly , n. c. , & dye , s. 2002 , , 333 , 501 catelan , p. , kamionkowski , m. , & blandford , r. d. 2001 , , 320 , l7 catelan , p. , & theuns , t. 1996 , , 282 , 436 crittenden , r. g. , natarajan , p. , pen , u. l & theuns , t. 2001 , , 559 , 552 crittenden , r. g. , natarajan , p. , pen , u. l & theuns , t. 2001 , , 568 , 20 croft , r. a. c. & metzler , c. a. 2000 , , 545 , 561 davis , m. , miller , a. , & white , s.d.m .
1997 , , 490 , 63 doroshkevich , a. g. 1970 , astrofizika , 6 , 581 hahn , o. , porciani , c. , carollo , c. m. , & dekel , a. 2007 , , 375 , 489 heavens , a. , refregier , a. , & heymans , c. 2000 , , 319 , 649 heymans , c. & heavens , a. 2003 , , 339 , 711 heymans , c. , brown , m. , heavens , a. , meisenheimer , k. , taylor , a. , & wolf , c. 2004 , , 361 , 160 hirata , c. m. & seljak , u. 2004 , , 70 , 063526 hirata , c. m. , et al .
2004 , , 353 , 529 hirata , c. m. , et al .
2007 , , 381 , 1197 hui , l. & zhang z. 2002 , preprint [ astro - ph/0205512 ] jimenez , r. , padoan , p. , matteucci , f. & heavens , a. f. 1998 , , 299 , 123 jing , y. 2002 , , 335 , l89 king , l. 2005 , , 441 , 47 lee , j. , kang , x. , & jing , y. 2005 , , 629 , l5 lee , j. & pen , u. l. 2000 , , 532 , l5 lee , j. & pen , u. l. 2001 , , 555 , 106 lee , j. & pen , u. l. 2002 , , 567 , 111 lee , j. & erdogdu , p. 2007 , 671 , 1248 mandelbaum , r. , hirata , c. m. , ishak , m. , seljak , u. , & brinkmann , j. 2006 , , 367 , 611 mo , h. j. , mao , s. & white , s. d. m. 1998 , , 295 , 319 navarro , j.f . , abadi , m.g . , & steinmetz , m. 2004 , , 613 , l41 patiri , s. g. , cuesta , a. j. , prada , f. , betancort - rijo , j. , & klypin , a. 2006 , , 652 , 75 peebles , p. j. e. 1969 , , 155 , 393 pen , u. l. , lee , j. , & seljak , u. 2000 , 543 , l107 porciani , c. , dekel , a. , & hoffman , y. 2002 , , 332 , 339 springel , v. et al .
2005 , , 435 , 629 takada , m. , & white , s. d. m. 2004 , , 601 , l1 trujillo , i. , carretero , c. , & patiri , s. 2006 , , 610 , l111 vitvitska , m. , klypin , a. a. , kravtsov , a. v. , wechsler , r. h. , primack , j. r. , & bullock , j. s. 2002 , , 581 , 799 white , s. d. m. 1984 , , 286 , 38 cccccc @xmath119 & @xmath120 & @xmath121 & @xmath122 & @xmath123 & @xmath124 + @xmath61 & @xmath125 & @xmath126 & @xmath127 & @xmath128 & @xmath129 + @xmath62 & @xmath130 & @xmath131 & @xmath132 & @xmath133 & @xmath134 + @xmath1 & @xmath135 & @xmath136 & @xmath137 & @xmath138 & @xmath139 + [ tab : z ] cccccc @xmath140 & @xmath141 & @xmath142 & @xmath143 & @xmath143 & @xmath144 + @xmath145 & @xmath146 & @xmath147 & @xmath148 & @xmath149 & @xmath150 + @xmath151 & @xmath152 & @xmath153 & @xmath154 & @xmath155 & @xmath156 + @xmath157 & @xmath158 & @xmath159 & @xmath160 & @xmath161 & @xmath162 + [ tab : a ] cccccc @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath167 & @xmath168 + @xmath169 & @xmath170 & @xmath171 & @xmath172 & @xmath173 & @xmath174 + @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath186 + [ tab : m ]
|
the non - gaussian contribution to the intrinsic halo spin alignments is analytically modeled and numerically detected . assuming that the growth of non - gaussianity in the density fluctuations caused the tidal field to have nonlinear - order effect on the orientations of the halo angular momentum , we model the intrinsic halo spin alignments as a linear scaling of the density correlations on large scales , which is different from the previous quadratic - scaling model based on the linear tidal torque theory .
then , we analyze the halo catalogs from the recent high - resolution millennium run simulation at four different redshifts ( @xmath0 and @xmath1 ) and measure quantitatively the degree of the nonlinear effect on the halo spin alignments and its changes with redshifts .
a clear signal of spin correlations is found on scales as large as @xmath2mpc at @xmath3 , which marks a detection of the nonlinear tidal effect on the intrinsic halo alignments .
we also investigate how the nonlinear effect depends on the intrinsic properties of the halos .
it is found that the degree of the nonlinear tidal effect increases as the halo mass scale decreases , the halo specific angular momentum increases , and the halo peculiar velocity decreases .
we discuss implication of our result on the weak gravitational lensing .
|
the primary purpose of the current paper is to show that , as a matter of principle , contrary to discerning lore that is realized in many fascinating systems , e.g. , @xcite , the appearance of a _ topological _ ground state degeneracy _ does not _ imply that these degenerate states are `` topologically ordered '' , in the sense that local perturbations can be detected without destroying the encoded quantum information @xcite . towards this end ,
we introduce various models , including a classical version of kitaev s toric code @xcite , that exhibit robust genus dependent degeneracies but are nonetheless landau ordered .
those models do not harbor long - range entangled ground states that can not be told apart from one another by local measurements .
rather , they ( as well as all other eigenstates ) are trivial classical states . along the way we will discover that these two - dimensional classical models ( including rather mundane clock and @xmath0 gauge like theories with four spin interactions ( specifically , toric clock and @xmath0 theories that we will define ) may not only have genus dependent symmetries and degeneracies but
, for various lattice types , may also exhibit _ holographic _ degeneracies that scale exponentially in the system perimeter .
similar degeneracies also appear in classical systems having two spin interactions .
thus , the classical degeneracies that we find may be viewed as analogs of those in quantum models such as the haah code model on the simple cubic lattice @xcite , a nontrivial theory with eight spin interactions that is topologically quantum ordered , and other quantum systems .
to put our results in a broader context , we first succinctly review current basic notions concerning the different possible types of order .
the celebrated symmetry - breaking paradigm @xcite has seen monumental success across disparate arenas of physics .
its traditional textbook applications include liquid to solid transitions , magnetism , and superconductivity to name only a few examples out of a very vast array . within this paradigm ,
distinct thermodynamic phases are associated with local observables known as order parameter(s ) . in the symmetric phase(s ) , these order parameters must vanish .
however , when symmetries are lifted , the order parameter may become non - zero .
phase transitions occur at these symmetry breaking points at which the order parameter becomes non - zero ( either continuously or discontinuously ) .
landau @xcite turned these ideas into a potent phenomenological prescription .
indeed , long before the microscopic theory of superconductivity @xcite , ginzburg and landau @xcite wrote down a phenomenological free energy form in the hitherto unknown complex order parameter with the aid of which predictions may be made .
albeit its numerous triumphs , the symmetry - breaking paradigm might not directly account for transitions in which symmetry breaking can not occur .
pivotal examples are afforded by gauge theories of the fundamental forces and very insightful abstracted simplified renditions capturing their quintessential character , e.g. , @xcite .
theorem @xcite prohibits symmetry breaking in gauge theories .
another notable example where the symmetry breaking paradigm can not be directly applied is that of the berezinskii - kosterlitz - thouless transition @xcite in two - dimensional systems with a global @xmath0 symmetry . by the mermin - wagner - hohenberg - coleman theorem and its extensions @xcite ,
such continuous symmetries can not be spontaneously broken in very general two - dimensional systems . augmenting these examples , penetrating work illustrated that something intriguing may happen when the quantum nature of the theory is of a defining nature @xcite . in particular ,
strikingly rich behavior was found in fractional quantum hall ( fqh ) systems @xcite , chiral spin liquids @xcite , a plethora of exactly solvable models , e.g. , @xcite , and other systems .
one curious characteristic highlighted in @xcite concerns the number of degenerate ground states in fqh fluids @xcite , chiral spin liquids @xcite , and other systems .
namely , in these theories , the ground state ( g.s . ) degeneracy is set by the topology alone .
for instance , regardless of general perturbations ( including impurities that may break all the symmetries of the hamiltonian ) , when placed on a manifold of genus number @xmath1 ( the determining topological characteristic ) , the fqh liquid at a laughlin type filling of @xmath2 ( with @xmath3 an odd integer ) universally has @xmath4 orthogonal ground states @xcite .
_ equation ( [ top_deg ] ) constitutes one of the best known realization of topological degeneracy . _
exact similarity transformations connect the second quantized fqh systems of equal filling when these are placed on different surfaces sharing the same genus @xcite . making use of the archetypal topological quantum phenomenon , the aharonov - bohm effect @xcite
, it was argued that , when charge is quantized in units of @xmath5 ( as it is for laughlin states ) , the minimal ground state degeneracy is given by the righthand side of eq .
( [ top_deg ] ) @xcite .
this may appear esoteric since realizing fqh states on riemann surfaces is seemingly not feasible in the lab .
recent work @xcite proposed the use of an annular superconductor - insulator - superconductor josephson junction in which the insulator is ( an electron - hole double layer ) in a fqh state ( of an identical filling ) for which this degeneracy is not mathematical fiction but might be experimentally addressed .
associated fractional josephson effects of this type in parafermionic systems were advanced in @xcite . historically , the robust topological degeneracy of eq .
( [ top_deg ] ) for fqh systems and its counterparts in chiral spin liquids suggested that such a degeneracy may imply the existence of a novel sort of order `` topological quantum order '' present in kitaev s toric code model @xcite , haah s code @xcite , and numerous other quantum systems @xcite a quantum order for which no local landau order parameter exists . as we will later review and make precise ( see eq .
( [ tqo ] ) ) , in topologically ordered systems , no local measurement may provide useful information .
as it is of greater pertinence to a model analyzed in the current work , we note that similar to eq .
( [ top_deg ] ) , on a surface of genus @xmath1 the ground state degeneracy of kitaev s toric code model @xcite , an example of an abelian quantum double model representing quantum error correcting codes ( solvable both in the ground state sector @xcite as well as at all temperatures @xcite ) , is @xmath6 thus , for instance , on a torus ( @xmath7 ) , the model exhibits 4 ground states while the system has a unique ground state on a topologically trivial ( @xmath8 ) surface with boundaries . by virtue of a simple mapping @xcite
, it may be readily established that an identical degeneracy appears for all excited states ; that is the degeneracy of each energy level is an integer multiple of @xmath9 .
thus , the minimal degeneracy amongst all energy levels is given by @xmath9 .
same ground state degeneracy @xcite appears in kitaev s honeycomb model @xcite .
as is widely known , an identical situation occurs in the quantum dimer model @xcite . invoking the well - known `` @xmath10ality '' considerations of @xmath11 , leading to a basic spin of 1/2 in @xmath12 and a minimal quark charge of 1/3 in @xmath13 , it was suggested @xcite that in many systems , fractional charges ( quantized in units of @xmath14 ) are a trivial consequence of the @xmath15 phase group center structure of a system endowed with an @xmath11 symmetry , which is associated with the @xmath16 states comprising the ground state manifold .
this @xmath16-ality type phase factors and other considerations , prompted sato @xcite to suggest the use of topological degeneracy ( akin to that of eqs .
( [ top_deg ] ) and ( [ tcd ] ) ) as a theoretical diagnosis delineating the boundary between the confined and the topological deconfined phases of qcd in the presence of dynamical quarks .
other notable examples include , e.g. , the bf action for superconductors ( carefully argued to not support a local order parameter @xcite ) . references @xcite examined the links between various concepts surrounding topological order with a focus on the absence of local order parameters . in particular ,
building on a generalization of elitzur s theorem @xcite it was shown how to construct and classify theories for which no local order parameter exists both at zero and at positive temperatures ; this extension of elitzur s theorem unifies the treatment of classical systems , such as gauge and berezinskii - kosterlitz - thouless type theories in arbitrary number of space ( or spacetime ) dimensions , to topologically ordered systems .
moreover , it was demonstrated that a sufficient condition for the existence of topological quantum order is the explicit presence , or emergence , of symmetries of _ dimension @xmath17 lower _ than the system s dimension @xmath18 , dubbed @xmath17-dimensional gauge - like symmetries , and which lead to the phenomenon of _ dimensional reduction_. the topologically ordered ground states are connected by these low - dimensional operator symmetries @xcite .
all known examples of systems displaying topological quantum order host these low dimensional symmetries , thus providing a unifying framework and organizing principle for such an order .
as underscored by numerous pioneers , features such as fractionalization and quasiparticle statistics , e.g. , @xcite , edge states @xcite , nontrivial entanglement @xcite , and other fascinating properties seem to relate with the absence of local order parameters and permeate topological quantum order . while all of the above features appear and complement the topological degeneracies found in , e.g. , the fqh ( eq .
( [ top_deg ] ) ) , the toric code ( eq . ( [ tcd ] ) ) , and numerous other systems , it is not at all obvious that one property ( say , a topological degeneracy such as those of eqs .
( [ top_deg ] ) and ( [ tcd ] ) ) implies another attribute ( for instance , the absence of meaningful local observables ) .
the current work will indeed precisely establish the absence of such a rigid connection between these two concepts ( viz .
, topological degeneracy is not at odds with the existence of a local order parameter ) .
we will employ the lack of local order parameters ( or , equivalently , an associated robustness to local perturbations ) as the defining feature of topological quantum order @xcite .
this robustness condition implies that local errors can be detected , and thus corrected , without spoiling the potentially encoded quantum information . to set the stage , in what follows
, we consider a set of @xmath19 orthonormal ground states@xmath20 with a spectral gap to all other ( excited ) states .
specifically @xcite , a system will be said to exhibit topological order at zero temperature if and only if for _ any quasi - local operator _
@xmath21 , @xmath22 where @xmath23 is a constant , independent of @xmath24 and @xmath25 , and @xmath26 is a correction that is either zero or vanishes ( typically exponentially in the system size ) in the thermodynamic limit .
the physical content of eq .
( [ tqo ] ) is clear : no possible quantity @xmath21 may serve as an order parameter to differentiate between the different ground states in the `` algebraic language '' @xcite where @xmath21 is local @xcite .
that is , all ground states look identical locally .
similarly , no local operator @xmath21 may link different orthogonal states the ground states are immune to all local perturbations .
notice the importance of the physical , and consequently mathematical , language to establish topological order : a physical system may be topologically ordered in a given language but its dual ( that is isospectral ) is not @xcite
. expressed in terms of the simple equations that we discussed thus far , the goal of this work is to introduce systems for which the ground state sector has a genus dependent degeneracy ( as in eqs .
( [ top_deg ] ) and ( [ tcd ] ) ) while , nevertheless , certain local observables ( or order parameters ) @xmath21 will be able to distinguish between different ground states ( thus violating eq .
( [ tqo ] ) ) .
moreover , they will be connected by global symmetry operators as opposed to low - dimensional ones .
our conclusions are generic and , as shown , they apply to many classical models .
the paradigmatic counterexample that we will introduce is a new classical version of kitaev s toric code model @xcite .
we now turn to the outline of the paper . in section [ general ] ,
we generalize the standard ( quantum ) toric code model .
after a brief review and analysis of the ground states of kitaev s toric code model ( section [ toric_quantum ] ) , we exclusively study our classical systems . in section [ toric_classical ]
, we extensively study the ground states of the classical variant of the model for different square lattices on riemann surfaces of varying genus numbers @xmath27
. a principal result will be that this and many other _ classical systems exhibit a topological degeneracy_. we will demonstrate that an intriguing holographic degeneracy may appear on lattices of a certain type .
as will be explained , topological as well as exponentially large in system linear size ( `` holographic '' ) degeneracies can appear in numerous systems , not only in this new classical version of kitaev s toric code model @xcite .
we further study the effect of lattice defects . the partition function of the classical toric code model is revealed in section [ ctc ] and appendix [ z ] . in section [ sec : clock ] ,
we introduce related classical clock models . generalizing the considerations of section [ toric_classical ]
, we will demonstrate that these clock models may exhibit topological or holographic degeneracies .
the ensuing analysis is richer by comparison to that of the classical toric code model . towards this end
, we will construct a new framework for broadly examining degeneracies .
we then derive lower bounds on the degeneracy that are in agreement with our numerical analysis .
these bounds are _ not confined to the ground state sector_. that is , all levels may exhibit topological degeneracies ( as they do in the classical toric code model ( section [ ctc ] ) ) . in section [ u1sec ] , we will relate our results to @xmath0 models and to @xmath0 lattice gauge theories in particular .
the fact that simple lattice gauge systems , that constitute a limiting case of our more general studied models , such as the conventional classical clock and @xmath0 lattice gauge theories _ on general riemann surfaces _ ( and their toric code extensions ) , _ exhibit topological _ ( or , in some cases , holographic ) _ degeneracies _ seems to have been overlooked until now . in section [ lattice_type ] , we will study honeycomb and triangular lattice systems embedded on surfaces of different genus . in section [ mm ] , we will discuss yet three more regular lattice classical systems that exhibit holographic degeneracies .
we summarize our main message and findings in section [ conclusionss ] . before embarking on the specifics of these various models , we briefly highlight the organizing principle behind the existence of degeneracies in our theories
. irrespective of the magnitude and precise form of the interactions in these theories , the number of independent constraints between the individual interaction terms sets the system degeneracy . as such ,
the degeneracies that we find are , generally , _ not a consequence of any particular fine - tuning_.
we start with a general description of a class of two - dimensional stabilizer models defined on lattices embedded on closed manifolds with arbitrary genus number @xmath1 ( the number of handles or , equivalently , the number of holes ) .
the genus of a closed orientable surface is related to a topological invariant known as euler characteristic @xmath28 which , for a general tessellation of that surface , satisfies the ( euler ) relation @xmath29 in eq .
( [ euler ] ) , @xmath30 is the number of vertices in the closed tessellating polyhedron , or graph , @xmath31 is the number of edges , and @xmath32 the number of polygonal faces .
assume that on each of the @xmath31 edges of the graph there is a spin @xmath33 degree of freedom , defining a local hilbert space of size @xmath34 , and that on each of the @xmath30 vertices and @xmath32 faces we will have a number of conditions to be satisfied by the ground states of a model that we define next .
we now explicitly define , on a general lattice or graph @xmath35 , the `` general toric code model '' . towards this end , we consider the hamiltonian @xmath36 where @xmath37 and @xmath38 are coupling constants ( although it is immaterial , in the remainder of this work we will assume these to be positive ) .
the interaction terms of edges in eq .
( [ h ] ) are so - called `` star '' ( `` @xmath39 '' ) terms ( @xmath40 ) associated with the @xmath30 vertices ( labelled by the letter @xmath41 ) and the @xmath32 `` plaquette '' ( `` @xmath42 '' ) terms ( @xmath43 ) . in the @xmath44 case , these are given by the following products of pauli operators @xmath45 , @xmath46 , @xmath47 the product defining @xmath48 spans the spins on all edges @xmath49 that have vertex @xmath39 as an endpoint , and the plaquette product @xmath50 is over all spins lying on the edges @xmath51 that form the plaquette @xmath42 ( see fig .
[ sym.png ] for an illustration ) . a key feature of this system ( both the well known @xcite quantum variant ( @xmath52 ) as well as , even more trivially , the classical version that we introduce in this paper ( @xmath53 ) )
is that each of the bonds @xmath54 and @xmath50 can assume @xmath55 independent values .
apart from global topological constraints @xcite that we will expand on below , the bonds @xmath56 and @xmath57 are completely independent of one another .
not only , trivially , in the classical but also in the quantum ( @xmath58 ) rendition of the model @xcite all of these operators commute with one another . that is @xmath59 , @xmath60= 0.\end{aligned}\ ] ] in the quantum version of the model , these terms commute as the products defining the star and plaquette
operators must share an even number of spins . as the individual pauli operators @xmath61 and @xmath62 appearing in the product of eq .
( [ ab ] ) anticommute , an even number of such anticommutations trivially gives rise to the commutativity in eq .
( [ triv1 ] ) .
even more simply , one observes that @xmath63= [ b^{\nu}_{p } , b^{\nu}_{p'}]=0.\end{aligned}\ ] ] lastly , from eq .
( [ ab ] ) , it is trivially seen that @xmath64 apart from a number ( @xmath65 ) of constraints , eqs .
( [ triv1 ] ) , ( [ triv2 ] ) , and ( [ triv3 ] ) completely specify all the relations amongst the operators of eq . ( [ ab ] ) . as we will illustrate , @xmath66 is a _ minimal model that embodies all of the elements in eq .
( [ euler ] ) _ such that its minimum degeneracy will only depend on the genus number @xmath1 .
as all terms in the hamiltonian @xmath66 commute with one another , the general toric code model can be related quite trivially to a classical model .
intriguingly , as may be readily established by a unitarity transformation ( a particular case of the bond - algebraic dualities @xcite ) , the quantum version , which includes kitaev s toric code model as a particular example , on a graph having @xmath31 edges spanning the surface of genus @xmath27 is identical @xcite , i.e. _ is isomorphic , to two decoupled * classical * ising chains _ ( with one of these chains having @xmath30 classical ising spins and the other chain composed of @xmath32 ising spins ) augmented by @xmath67 decoupled single ising spins .
perusing eq .
( [ h ] ) , it is clear that , if globally attainable , within the ground state(s ) , @xmath68 , @xmath69 on all vertices @xmath39 and faces @xmath42 and , thus , the ground state energy is @xmath70 .
the algebraic relations above enable the realization of eq .
( [ gsk ] ) for all @xmath39 and @xmath42 .
we now turn to the constraints that augment eqs .
( [ triv1 ] ) , ( [ triv2 ] ) , and ( [ triv3 ] ) . for any lattice @xmath35 on any closed surface of genus @xmath27
, there are @xmath71 universal constraints given by the equalities @xmath72 for the quantum variant @xcite no further constraints appear beyond those of eq .
( [ abc ] ) ( that is , @xmath73 irrespective of the lattice @xmath35 ) .
by contrast , for the classical variant of the theory realized on the relatively uncommon `` commensurate '' lattices , additional constraints will augment those of eq .
( [ abc ] ) ( i.e. , for classical systems , @xmath74 ) .
invoking the @xmath65 constraints as well as the trivial algebra of eqs .
( [ triv1 ] ) and ( [ triv2 ] ) , we may transform from the original variables the spins on each of the @xmath31 edges
@xmath75 to new basic degrees of freedom
all @xmath76 independent _
`` bonds '' _ @xmath77 that appear in the hamiltonian and @xmath78 remaining redundant spins of the original form @xmath75 on which the energy does not depend ( and thus relate to symmetries ) . if the bonds @xmath48 and @xmath79 do not adhere to any constraint apart from that in eq .
( [ abc ] ) then @xmath80 of the @xmath81 bonds in the hamiltonian of eq .
( [ h ] ) will be independent of one another .
correspondingly , @xmath82=2g$ ] . as all bonds must satisfy the constraint of eq .
( [ abc ] ) and thus @xmath83 , the number of redundant spin degrees of freedom @xmath84 . in the general case ,
if there are @xmath85 constraints that augment the two restrictions already present in eq .
( [ abc ] ) , then we may map the original system of @xmath31 spins to @xmath86 independent bonds in eq .
( [ h ] ) and @xmath87 spins that have no impact on the energy .
thus , for genus @xmath27 surfaces , the degeneracy of each energy level is an integer multiple of the _ minimal _ degeneracy possible , @xmath88 with @xmath89 .
equation ( [ mindeg ] ) will lead to a global redundancy factor in the partition function @xmath90 with @xmath25 the inverse temperature .
we now focus on the ground state sector . if there are no constraints apart from eq .
( [ abc ] ) , then to obtain the ground states it suffices to make certain that @xmath91 of the bonds are unity in a given state .
once that occurs , we are guaranteed a ground state in which each bond in the hamiltonian of eq .
( [ h ] ) is maximized ( i.e. , eqs . ( [ gsk ] ) are satisfied ) .
a smaller number of bonds fixed to one will not ensure that only ground states may be obtained .
thus the values of all @xmath91 independent bonds need to be fixed in order to secure a minimal value of the energy .
the lower bound of the degeneracy on each level ( eq .
( [ mindeg ] ) ) is saturated for the ground state sector where it becomes an equality .
that is , very explicitly , the ground state degeneracy is given by @xmath92 the equalities of eqs .
( [ mindeg ] ) and ( [ ngtc ] ) are basic facts that will be exploited in the present article . the degeneracy of eq .
( [ ngtc ] ) is in accord with the general result @xmath93 and differs from that of kitaev s toric code model @xcite ( eq . ( [ tcd ] ) ) by a factor of @xmath94 .
as each of the @xmath95 constraints as well as increase in genus number leads to a degeneracy of the spectrum , a simple `` correspondence maxim '' follows : _ it must be that we may associate a corresponding independent set of symmetries with any individual constraint_. similarly , as eqs .
( [ mindeg ] , [ ngtc ] ) attest , elevating the genus number @xmath1 must introduce further symmetries .
thus , the global degeneracy of eq .
( [ mindeg ] ) is a consequence of all of these symmetries .
( [ h ] ) it is readily seen that the system has a gap of magnitude @xmath96 between the ground state @xmath97 and the lowest lying excited state @xmath98 .
all energy levels @xmath99 , defining the spectrum of @xmath66 , are quantized in integer multiples of @xmath37 and @xmath38 .
in kitaev s toric code model @xcite the symmetries associated with the constraints of eq .
( [ abc ] ) are rather straightforward , and cogently relate to the topology of the surface on which the lattice is embedded .
an illustration for the square lattice is depicted in fig .
[ sym.png ] .
for such a model on a simple torus ( i.e. , one with genus @xmath7 ) , the four canonical symmetry operators are @xmath100 these two sets of non - commuting operators @xcite @xmath101= \ & 0 & \ = \left[z^{q}_1,z^{q}_2\right ] , \nonumber\\ \left[x^{q}_1,z^{q}_2\right]= \ & 0 & \ = \left[x^{q}_2,z^{q}_1\right ] , \end{aligned}\ ] ] realize a @xmath102 symmetry and ensure a four - fold degeneracy ( or , more generally a degeneracy that is an integer multiple of four ) of the whole spectrum . to see this , we may , for instance , seek mutual eigenstates of the hamiltonian @xmath103 along with the two symmetries @xmath104 and @xmath105 with which it commutes .
noting the algebraic relations amongst the above operators , a moment s reflection reveals that a possible candidate for a normalized ground state is given by @xmath106 where @xmath107 , for all @xmath31 edges , and @xmath108 .
this corresponds to @xmath109 .
now , because @xmath110 are symmetries , by the algebraic relations of eq .
( [ cr ] ) , the three additional orthogonal states @xmath111 lattice ( @xmath35 ) independent constraints of the quantum model ( eq .
( [ abc ] ) ) correspond to the @xmath112 sets of symmetry operators associated with the @xmath113 toric cycles ( @xmath114 ) of eq .
( [ zx ] ) .
this correspondence is in agreement with the simple maxim highlighted above .
the symmetry operators @xmath115 and @xmath116 are independent ( and trivially commute ) with the symmetry operators @xmath117 and @xmath118 . notice that in the spin @xmath119 ) language the ground states above are entangled , and they are connected by @xmath120 symmetry operators @xcite .
moreover , the anyonic statistics of its excitations is linked to the entanglement properties of those ground states @xcite .
as mentioned above , the model can be trivially related , by duality , to two decoupled classical ising chains so that in the dual language the mapped ground states are unentangled @xcite . for a riemann surface of genus @xmath1
, we may write down trivial extensions of eqs .
( [ zx ] ) for the @xmath121 cycles circumnavigating the @xmath1 handles of that surface
. that is , instead of the four operators of eq .
( [ zx ] ) , we may construct @xmath122 operators pairs with each of these pairs associated with a particular handle @xmath123 ( where @xmath124 ) , containing the four operators @xmath125 and @xmath126 with @xmath113 .
a generalization of eqs .
( [ cr ] ) leads to an algebra amongst the @xmath122 independent pairs of symmetry operators .
the multiplicity of independent symmetries leads to the first factor in eq .
( [ ngtc ] ) .
the number of constraints is , in the quantum case , lattice independent and given by @xmath127 ( there are no constraints beyond those in eq .
( [ abc ] ) ) .
it is rather straightforward to establish that when @xmath8 ( i.e. , for topologically trivial surfaces ) , the ground state of the quantum model is unique .
putting all of these pieces together , the well known degeneracy of eq .
( [ tcd ] ) follows .
we now finally turn to the examination of the ground states of the classical rendering of eq .
( [ h ] ) in which only a single component @xmath128 of all spins appears .
we will explain how _ the degeneracy of eqs .
( [ mindeg ] ) and ( [ ngtc ] ) emerges_. the upshot of our analysis , already implicitly alluded to above , consists of two main results : @xmath129 in the most frequent lattice realization of this classical model , _ its degeneracy will still be given by eq .
( [ tcd ] ) _ , i.e. , @xmath9 .
that is , in the most common of geometries , the number of ground states will depend on topology alone ( i.e. , the genus number @xmath1 of the embedding manifold ) . for arbitrary square lattice or graph ,
as our considerations universally mandate , the _ minimal _ possible ground state degeneracy will be given by the topological figure of merit of eq .
( [ tcd ] ) .
@xmath129 in the remaining lattice realizations , _ the degeneracy of the system will typically be holographic_. that is , in these slightly rarer lattices , the ground state degeneracy will scale as @xmath130 where @xmath131 is the length of one of the sides of the two - dimensional lattice .
as will be seen , for the square lattice , depending on the parity of the length of the lattice sides , the number of constraints @xmath95 may exceed its typical value of two .
this will then lead to an enhanced degeneracy vis a vis the minimal possible value of @xmath9 . in the next subsection we first broadly sketch the constraints and symmetries of the classical system . as it will be convenient to formulate our main result via the `` correspondence maxim ''
, we will then proceed to explicitly relate the constraints and symmetries to one another .
the @xmath132 consonance , along with eqs .
( [ mindeg ] ) and ( [ ngtc ] ) , will then rationalize all of the degeneracies found for general square lattices embedded on riemann surfaces of arbitrary genus number .
exhaustive calculations for these degeneracies will then be reported in the subsections that follow .
we next list the general symmetries and constraints of the classical toric code model in square lattices of varying sizes .
consider first a lattice @xmath35 of size @xmath133 on a torus ( i.e. , having @xmath134 vertices and @xmath135 edges ) .
we will then examine more general lattices of arbitrary genus @xmath1 .
the square lattice on the torus will be categorized as being one of two types : @xmath136 type i lattices , as defined for the @xmath7 case above and their generalizations for higher genus numbers @xmath137 , only admit two constraints @xmath95 and thus by the correspondence maxim only two symmetries . for these lattices , we will show that the ground state degeneracy is @xmath9 .
by contrast , type ii lattices have a larger wealth of constraints , @xmath138 , and therefore a larger number of symmetries and a degeneracy higher than @xmath9 .
the centers of all nearest neighbor edges on the square lattice ( of lattice constant @xmath139 ) form yet another square lattice @xmath140 ( of lattice constant @xmath141 ) at an angle of @xmath142 relative to the original lattice ( fig .
[ xm.png ] ) .
the spins are located at the vertices of the rotated square lattice @xmath140 . in order to describe the symmetries and constraints of this system ,
let us denote the two ( standard ) sublattices of the square lattice @xmath140 by @xmath143 .
that is , both @xmath144 and @xmath145 are , on their own , square lattices with @xmath146 and @xmath147 .
let us furthermore denote the sites of @xmath143 by @xmath148 , respectively . with these preliminaries ,
it is trivial to verify that @xmath149 are , universally , both symmetries of the classical ( @xmath150 ) version of the hamiltonian of eq .
( [ h ] ) .
most square lattices ( those of type i in eq .
( [ type12 ] ) ) will only exhibit the two symmetries of eq .
( [ t+- ] ) .
the more commensurate type ii lattices admit diagonal contours ( connecting nearest neighbors of sites @xmath151 of @xmath140 ) that close on themselves before threading all of the lattice sites of @xmath140 .
that is , in type ii lattices , it is possible to find diagonal loops @xmath152 at a constant @xmath153 angle ( or a more non - trivial alternating contour ) that contain only a subset of all sites of @xmath140 ( or , equivalently , a subset of all edges @xmath154 of the original square lattice @xmath35 ) . associated with each such independent contour @xmath152 , there is a symmetry operator , @xmath155 augmenting the symmetries of eq .
( [ t+- ] ) .
the form of the symmetries suggests the distinction between type i and type ii lattices on general surfaces . on type
ii lattices , it is possible to find , at least , one diagonal contour @xmath152 that contains a subset of all edges @xmath154 of the lattice @xmath35 .
conversely , due to the lack of the requisite lattice commensurability , on type i lattices , it is impossible to find any such contour .
we now turn to the constraints associated with type i and ii lattices .
these are in one - to - one correspondence with the symmetries of eqs .
( [ t+- ] ) and ( [ tl ] ) . specifically for type i lattices , the only universal constraints present are those of eq .
( [ abc ] ) which we rewrite again for clarity , @xmath156 these two constraints match the two symmetries of eq .
( [ t+- ] ) . in the case of the more commensurate lattices @xmath35 ,
additional constraints appear . in order to underscore the similarities to the symmetries of eq .
( [ tl ] ) , we will now aim to briefly use the same notation concerning the lattice @xmath140 . within the framework highlighted in earlier sections ,
the spin products @xmath157 and @xmath158 of eq .
( [ ab ] ) are associated with geometrical objects that look quite different ( i.e. , `` stars '' and `` plaquettes '' ) , see fig .
[ sym.png ] .
if we now label the plaquettes of @xmath140 by @xmath159 then , we may , of course , trivially express eq .
( [ h ] ) as a sum of local terms , @xmath160 where @xmath161 are the products of all ising spins at sites @xmath151 belonging to plaquette @xmath162 .
this trivial description renders the original star and plaquette terms of eq .
( [ h ] ) on a more symmetric footing , see fig .
[ xm.png ] . associated with each of the symmetries of eq .
( [ tl ] ) there is a corresponding constraint , @xmath163 in accordance with our earlier maxim , insofar as counting is concerned , we have the following correspondence between the symmetries and the associated constraints , @xmath164 in type i systems , wherein only the @xmath165 universal constraints appear , the degeneracy of the spectrum is exactly @xmath9 . in type
ii lattices , @xmath166 ( with the difference of @xmath167 equal to the number of additional independent contours @xmath152 that do not contain all edges of the original lattice @xmath35 ) and , as eq .
( [ ngtc ] ) dictates , the ground state degeneracy exceeds the minimal value of @xmath9 multiplied by two raised to the power of the number of the additional independent loops .
thus far , our discussion has been quite general and , admittedly , somewhat abstract .
we now turn to simple concrete examples .
we first consider the classical toric code model on a simple torus ( i.e. , a surface with genus @xmath7 ) , and examine small specific square lattices of dimension @xmath133 .
we find that for general lattices @xmath35 ( with reference to eq .
( [ type12 ] ) ) , the total number of independent constraints is @xmath168 thus , from eq .
( [ ngtc ] ) , our two earlier stated main results follow : _ while for the more `` incommensurate '' type i lattices , the degeneracy will be * topological * " ( i.e. , given by @xmath9 ) , for type ii lattices , the degeneracy will be * holographic * " ( viz . , the degeneracy will be exponential in the smallest of the edges along the system boundaries)_. as discussed in section [ sac ] , the additional constraints in type ii lattices are of the form of eq .
( [ wl ] ) . expressed in terms of the four spin interaction terms @xmath169 and @xmath170 of eq .
( [ h ] ) , a constraint of the form of eq .
( [ wl ] ) states that there is a subset @xmath171 for which @xmath172 .
an illustration of a constraint of such a type is provided , e.g. in fig . [ 8-1.png ] . here , by virtue of the defining relations of eq .
( [ ab ] ) , the product , @xmath173 similarly , in panel a ) of fig .
[ 8-2.png ] , colored arrows are drawn along the diagonals .
these colors code the constraints on the specific @xmath174 and @xmath175 interaction terms .
for example , along the _ green _ arrows , @xmath176 and the constraints associated with the other diagonals @xmath177 we provide another example in panel b ) of fig . [ 8-2.png ] .
the simplest visually appealing realization of eq .
( [ wl ] ) is that of the subset @xmath152 being a trivial closed diagonal loop .
composites ( i.e. , products ) of independent constraints of the form of eq .
( [ wl ] ) are , of course , also constraints .
we aim to find the largest number @xmath85 of such independent constraints .
non - trivial constraints formed by the product of bonds along real - space diagonal lines may appear .
for example , in fig .
[ 8-1.png ] , the product @xmath178 is precisely such a constraint .
these constraints are more difficult to determine due to the periodic boundary conditions . generally , not all constraints are independent of each other ( e.g. , multiplying any two constraints yield a new constraint ) .
the number of independent constraints , @xmath179 may be generally found by calculating the `` modular rank '' of the linear equations formed by taking the logarithm of all constraints found .
the qualified `` modular '' appears here as the @xmath180 and @xmath181 eigenvalues may only be @xmath182 and thus , correspondingly , their phase is either @xmath183 or @xmath184 .
many , yet generally , not all , of the @xmath179 independent constrains are naturally associated with products along the 45@xmath185 lattice diagonals ( as it appears on the torus ) .
table [ table0 ] lists the numerically computed ground state degeneracies for numerous lattices of genus @xmath7 .
all of these are concomitant with eq .
( [ cdeg ] ) .
@xmath31 & @xmath186 & @xmath187 & + & 3 & 2 & 12 & 2 & 4 + & 5 & 2 & 20 & 2 & 4 + & 4 & 3 & 24 & 2 & 4 + & 5 & 3 & 30 & 2 & 4 + & 2 & 2 & 8 & 4 & @xmath188 + & 4 & 2 & 16 & 4 & @xmath188 + & 6 & 2 & 24 & 4 & @xmath188 + & 3 & 3 & 18 & 6 & @xmath189 + & 4 & 4 & 32 & 8 & @xmath190 + given the symmetry operators of eqs .
( [ t+- ] ) and ( [ tl ] ) , we may rather readily write down all ground states of the system .
denote the ferromagnetic ground state ( i.e. , one with all spins up ( @xmath191 ) on all edges @xmath154 ) by @xmath192 then , the four ground states of type i lattices are @xmath193 where @xmath194 . clearly , since @xmath195 , only the parity of the integers @xmath196 is important . as ( i )
@xmath197=0 $ ] and ( ii ) the ferromagnetic state @xmath198 minimizes the energy in eq .
( [ h ] ) , it follows that all four binary strings @xmath199 in eq .
( [ gn+ ] ) lead to @xmath200 ground states .
the situation for type ii lattices is a trivial extension of the above .
that is , if there are @xmath201 additional independent symmetries @xmath202 of the form of eq .
( [ tl ] ) then , with the convention of eq .
( [ ef ] ) , the ground states will be of the form @xmath203 with @xmath204 binary strings @xmath205 , where @xmath206 .
these strings span all possible @xmath207 orthogonal ground states . given the set of all orthonormal ground states @xmath208 , it is possible to find quasi - local operators @xmath209 composed of @xmath210 `` operators '' on a small number of edges such that @xmath211 assumes different values @xmath212 in , at least , two different ground states . equation ( [ tqo ] ) highlights that the expectation value of @xmath213 is not state independent .
in other words , eq .
( [ tqo ] ) @xcite is violated .
thus , our classical system is , rather trivially , not topologically ordered .
having understood the case of the simple torus ( @xmath7 ) , we will now study lattices on surfaces @xmath215 of genus @xmath216 .
we first explain how to construct a finite size lattice of genus @xmath1 .
such lattices on genus @xmath1 ( @xmath217 ) surfaces may be formed by _
`` stitching together '' _
@xmath1 simple parts @xmath218 , @xmath219 , each of which largely looks like that of a simple torus ( i.e. , each region @xmath220 represents a set of vertices , edges and faces of type i or ii in the notation of eq . ( [ type12 ] ) ) , via @xmath221 `` bridges '' @xmath222 . in figs .
[ g=2.fig ] , and [ g=3.fig ] , the integer number @xmath223 denotes the number of edges that regions @xmath220 and @xmath224 share .
to lucidly illustrate the basic construct , we start first with a @xmath225 lattice . in fig .
[ g=2.fig ] , identical edges are labeled by the same number as a consequence of the periodic boundary conditions . here
, there are @xmath226 edges , @xmath227 vertices , and @xmath228 plaquettes . as in the case of the simple torus ( @xmath7 ) , the typical vertices are endpoints of four edges .
similarly , in fig . [ g=2.fig ] , all plaquettes ( with the exception of two ) are comprised of four edges as in the situation of the simple torus .
the exceptional cases are colored green ( dashed - dotted ) and red ( dashed ) . as seen in the figure ,
the lattice may be splintered into two regions ( labeled by @xmath229 and @xmath230 ) where one end of some of the bonds belonging to @xmath229 are connected to @xmath230 as shown and labeled in the picture under @xmath231 .
each of the regions @xmath229 and @xmath230 looks , by itself , like a square lattice on a genus @xmath7 surface . generally , the regions @xmath229 and @xmath230 may be composed of a different number of edges . employing the taxonomy of eq .
( [ type12 ] ) , we may classify these regions @xmath232 to be of either type i or ii . we remark that the number of edges @xmath231 must be always at least one less than the minimum of the number of bonds of @xmath229 and @xmath230 along the horizontal ( @xmath233 ) axis .
this algorithm trivially generalizes to higher genus number .
the cartoon of fig .
[ g=3.fig ] represents a lattice with @xmath234 .
a synopsis of our numerical results for the ground state degeneracy for surfaces of genus @xmath235 appears in table [ table1 ] .
the ground state degeneracy depends on the type of each @xmath220 and the number of bonds of each @xmath223 .
when all fragments @xmath236 are of type i and are inter - connected by only _ single _
common edges , the degeneracy attains will its minimal possible value ( eq .
( [ ngtc ] ) ) of @xmath9 .
@xmath187 & type & @xmath229 & @xmath231 & @xmath230 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & + & & @xmath243 & 2 i & 2@xmath2441 & 1 & 2@xmath2441 & & & & & & + & & @xmath243 & 2 i & 3@xmath2441 & 1 & 2@xmath2441 & & & & & & + & & @xmath243 & 2 i & 3@xmath2441 & 1 & 3@xmath2441 & & & & & & + & & @xmath243 & 2 i & 3@xmath2442 & 1 & 2@xmath2441 & & & & & & + & & @xmath243 & 2 i & 3@xmath2442 & 1 & 3@xmath2441 & & & & & & + & & @xmath243 & 2 i & 3@xmath2442 & 1 & 3@xmath2442 & & & & & & + & & @xmath243 & 2 i & 5@xmath2442 & 1 & 2@xmath2441 & & & & & & + & & @xmath245 & 2 i & 3@xmath2441 & 2 & 3@xmath2441 & & & & & & + & & @xmath245 & ii+i & 2@xmath2442 & 1 & 2@xmath2441 & & & & & & + & & @xmath245 & ii+i & 2@xmath2442 & 1 & 3@xmath2441 & & & & & & + & & @xmath245 & i+ii & 3@xmath2442 & 1 & 2@xmath2442 & & & & & & + & & @xmath245 & ii+i & 4@xmath2442 & 1 & 2@xmath2441 & & & & & & + & & @xmath245 & ii+i & 4@xmath2442 & 1 & 3@xmath2441 & & & & & & + & & @xmath245 & 2 i & 3@xmath2442 & 2 & 3@xmath2442 & & & & & & + & & @xmath245 & ii+i & 3@xmath2443 & 2 & 3@xmath2441 & & & & & & + & & @xmath246 & 2 ii & 2@xmath2442 & 1 & 2@xmath2442 & & & & & & + & & @xmath246 & ii+i & 3@xmath2443 & 1 & 3@xmath2441 & & & & & & + & & @xmath246 & 2 ii & 4@xmath2442 & 1 & 2@xmath2442 & & & & & & + & & @xmath243 & 3 i & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2441 & 1 & 3@xmath2441 & 1 & 2@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2441 & 2 & 3@xmath2441 & 1 & 2@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2441 & 1 & 3@xmath2441 & 1 & 3@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2441 & 2 & 3@xmath2441 & 1 & 3@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2441 & 1 & 3@xmath2441 & 2 & 3@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2441 & 2 & 3@xmath2441 & 2 & 3@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2442 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2442 & 1 & 3@xmath2441 & 1 & 3@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2442 & 2 & 3@xmath2441 & 2 & 3@xmath2441 & & & & + & & @xmath245 & 2 i+ii & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2442 & & & & + & & @xmath245 & 2 i+ii & 3@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2442 & & & & + & & @xmath245 & 2 i+ii & 3@xmath2441 & 1 & 3@xmath2441 & 1 & 2@xmath2442 & & & & + & & @xmath247 & 2 ii+i & 2@xmath2442 & 1 & 2@xmath2442 & 1 & 2@xmath2441 & & & & + & & @xmath247 & 2 ii+i & 2@xmath2442 & 1 & 2@xmath2442 & 1 & 3@xmath2441 & & & & + & & @xmath248 & 3 ii & 2@xmath2442 & 1 & 2@xmath2442 & 1 & 2@xmath2442 & & & & + & & @xmath243 & 4 i & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & & + & & @xmath243 & 4 i & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 3@xmath2441 & & + & & @xmath243 & 4 i & 3@xmath2442 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & & + & & @xmath245 & ii+3 i & 2@xmath2442 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & & + & & @xmath243 & 5 i & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 + & & @xmath245 & ii + 4 i & 2@xmath2442 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 + if , in eq .
, we set @xmath37 to zero , we will obtain the hamiltonian of the ising gauge model .
as this theory does not have a star term , this hamiltonian involves more symmetries and , therefore , one expects the ground state subspace to have a larger degeneracy
. we numerically verified it to be @xmath249-fold degenerate ( @xmath250 ) @xcite .
@xmath187 & type & @xmath229 & @xmath231 & @xmath230 & @xmath237 & @xmath238 & @xmath239 & @xmath240 + & & @xmath243 & i & 3@xmath2442 @xmath251 & & & & & & + & & @xmath243 & ii & 4@xmath2442 @xmath251 & & & & & & + & & @xmath243 & i & 5@xmath2442 @xmath251 & & & & & & + & & @xmath243 & i & 6@xmath2442 @xmath251 & & & & & & + & & @xmath243 & i & 4@xmath2443 @xmath251 & & & & & & + & & @xmath245 & ii & 3@xmath2443 @xmath252 & & & & & & + & & @xmath245 & ii & 3@xmath2443 @xmath251 & & & & & & + & & @xmath245 & i & 5@xmath2442 @xmath253 & & & & & & + & & @xmath245 & i & 6@xmath2442 @xmath252 & & & & & & + & & @xmath243 & 2 i & 3@xmath2442 @xmath251 & 1 & 2@xmath2441 & & & & + & & @xmath243 & 2 i & 3@xmath2442 @xmath251 & 1 & 3@xmath2441 & & & & + & & @xmath243 & 2 i & 4@xmath2442 @xmath251 & 1 & 3@xmath2441 & & & & + & & @xmath243 & 2 i & 3@xmath2442 @xmath251 & 1 & 3@xmath2442 @xmath251 & & & & + & & @xmath243 & 2 i & 3@xmath2442 @xmath251 & 1 & 3@xmath2442 & & & & + & & @xmath243 & 2 ii & 4@xmath2442 @xmath251 & 1 & 2@xmath2442 & & & & + & & @xmath243 & ii+i & 3@xmath2443 @xmath251 & 2 & 3@xmath2441 & & & & + & & @xmath243 & 2 i & 5@xmath2442 @xmath251 & 1 & 2@xmath2441 & & & & + & & @xmath245 & ii+i & 3@xmath2443 @xmath251 & 1 & 3@xmath2441 & & & & + & & @xmath243 & 3 i & 3@xmath2442 @xmath251 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & & + & & @xmath243 & 3 i & 3@xmath2442 @xmath251 & 1 & 3@xmath2441 & 1 & 3@xmath2441 & & + & & @xmath243 & 3 i & 3@xmath2442 @xmath251 & 2 & 3@xmath2441 & 2 & 3@xmath2441 & & + & & @xmath243 & 4 i & 3@xmath2442 @xmath251 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 + when dislocations and/or any other lattice defects are present in the classical toric code model , the degeneracy is , of course , still bounded by the geometry independent result of @xmath9 . on type i lattice ( and their composites ) , the degeneracy is typically equal to this bound yet it may go up upon the introduction of defects . similarly , in most cases introducing such lattice defects lowers the degeneracy of the more commensurate type ii lattices ( and their composites ) .
table [ table2 ] provides the numerical results for such defective lattices .
for example , in fig .
[ defect.png ] we see the original lattice , panel a ) , along with two types of defects as in panel b ) and c ) .
these are obtained by replacing @xmath254 squares by @xmath112 adjacent or separated pentagons as in panel b ) and c ) , respectively . to avoid confusion
, we will use `` @xmath251 '' sign for the first case and `` @xmath253 '' for the second case . by putting a `` @xmath251 '' ( `` @xmath253 '' ) sign beside a @xmath255 lattice ,
we mean it exhibits a defect of type one ( two ) .
that is , represented as `` @xmath256 '' ( `` @xmath257 '' ) .
previous sections largely focused on the ground states of the classical toric code model . as our earlier considerations make clear , however , a minimal topology ( and general constraint ) dependent degeneracy @xmath258 appears for all levels ( see , e.g. , eq .
( [ mindeg ] ) ) .
this `` global '' degeneracy must manifest itself as a prefactor in the computation of the partition function .
that is , if the whole spectrum has a global degeneracy @xmath259 then the canonical partition function may be expressed as @xmath260 where @xmath261 is the number of states having total energy @xmath99 . in `` incommensurate '' lattices , when no constraints @xmath262 augment those of eq .
( [ abc ] ) , we find that , similar to the partition function of the quantum toric code model @xcite , the partition of the classical toric code model is given by @xmath263 \nonumber \\ & & \times [ ( 2 \cosh \beta j')^{f } + ( 2 \sinh \beta j')^{f}].\end{aligned}\ ] ] the prefactor of @xmath264 embodies the increase in degeneracy by a factor of four as @xmath1 is elevated in increments @xmath265 beyond a value of @xmath7 . on the simple torus ( i.e. , when @xmath7 ) , this partition function ( similar to the partition function of the quantum toric code model @xcite ) is that of two decoupled ising chains with one of these chains having @xmath30 spins and the other composed of @xmath32 spins .
as each such ising chain has a two - fold degeneracy , it thus follows that the degeneracy of the ( more `` incommensurate '' ) type i @xmath7 system is four - fold and that the degeneracy of the classical toric code model on incommensurate lattices on riemann surfaces of genus @xmath1 is @xmath9 for all @xmath27 .
the latter value saturates the lower bound on the degeneracy of eq .
( [ mindeg ] ) . in appendix [ z ] , we list the partition function for several other more commensurate finite size lattice realizations .
in this section , we introduce and study a clock model ( @xmath266 ) extension of the classical toric code model . to that end , we consider what occurs when each spin @xmath33 may assume @xmath267 values .
specifically , on every oriented ( @xmath268 ) edge ( that we will hereafter label as @xmath51 ) , we set @xmath269,\,\,\,\ , ( \alpha_{ij } = 0,1,\cdots,{\sf d}_q-1 ) , \\
\sigma_{ji}&= & \sigma^{*}_{ij}.\end{aligned}\ ] ] the last equality reflects that a change in the orientation ( i.e. , a link in the direction from @xmath270 as opposed to @xmath271 ) is associated with complex conjugation . at each vertex `` @xmath39 ''
, we define @xmath272 as @xmath273 and for each plaquette @xmath42 @xmath274 composed of edges @xmath275 , such that the loop @xmath276 is oriented counter - clockwise around about the plaquette center .
table [ table3 ] provides our numerical results for ground state degeneracy ( @xmath277 ) for different size lattices of varying genus numbers @xmath1 .
the @xmath278 case is that investigated in the earlier sections ( i.e. , that of the classical toric code model with ising variables @xmath279 ) .
it is readily observed that the _ minimal _ ground state degeneracy is set by the genus number , @xmath280 we next introduce a simple framework that rationalizes eq .
( [ min ] ) and enables us to furthermore derive the results of the previous sections ( i.e. , the ising case of @xmath281 ) in a unified way .
furthermore , this approach will allow us to better understand not only the degeneracies in the ground sector but also those of all higher energy states .
in the up and coming , we will study the hamiltonian @xmath282 here , @xmath283 constitute a system of linear equations .
a pair of fixed integers @xmath284 and @xmath285 defines an energy @xmath286 .
there are @xmath287 such pairs . for each fixed pair @xmath288 , @xmath289 , we may express these linear equations as @xmath290 where @xmath291 is a rectangular ( @xmath292 ) matrix
. the matrix elements of @xmath291 are either @xmath183 or @xmath293 .
generally , the form of the matrix @xmath291 depends on both the size and type of lattice .
the dimension of the vector @xmath294 is equal to the number ( @xmath31 ) of edges ; @xmath295 is a @xmath296component vector .
specifically , following eq .
( [ alphaeq ] ) , these two vectors are defined as : @xmath297 , with components @xmath298 , and @xmath299 , for its first @xmath30 components and @xmath300 , for the remaining @xmath32 components . the number of linearly independent equations ( @xmath301 ) is equal to the rank of the matrix @xmath291 . typically , the rank @xmath301 is less than the number of unknown @xmath298 .
therefore , we can not determine all @xmath298 from eq .
( [ matrix ] ) .
we should note that the rank of the matrix @xmath291 is computed modularly , `` @xmath302 '' .
this latter modular rank is of pertinence as the edge variables @xmath298 may only take on particular modular values ( @xmath303 ) .
@xmath31 & type & @xmath229 & @xmath231 & @xmath230 & @xmath237 & @xmath238 & @xmath239 & @xmath240 & @xmath304 & @xmath305 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & @xmath310 & @xmath311 & @xmath312 & @xmath313 & @xmath314 & @xmath315 & @xmath316 & @xmath317 + & & i & 2@xmath2441 & & & & & & & 1 & 2 & 1 & 1 & 1 & 2 & 1 & 1 & 1 & 2 & 1 & 1 & 1 & 2 + & & i & 3@xmath2441 & & & & & & & 3 & 1 & 1 & 3 & 1 & 1 & 3 & 1 & 1 & 3 & 1 & 1 & 3 & 1 + & & i & 4@xmath2441 & & & & & & & 1 & 2 & 1 & 1 & 1 & 4 & 1 & 1 & 1 & 2 & 1 & 1 & & + & & ii & 2@xmath2442 & & & & & & & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & & + & & i & 3@xmath2442 & & & & & & & 3 & 2 & 1 & & & & & & & & & & & + & & ii & 4@xmath2442 & & & & & & & @xmath318 & @xmath330 & & & & & & & & & & & & + & & ii & 3@xmath2443 & & & & & & & @xmath331 & & & & & & & & & & & & & + & & 2 i & 2@xmath2441 & 1 & 2@xmath2441 & & & & & 1 & 2 & 1 & 1 & 1 & 2 & 1 & 1 & 1 & 2 & 1 & 1 & & + & & 2 i & 3@xmath2441 & 1 & 3@xmath2441 & & & & & 3 & 1 & 1 & & & & & & & & & & & + & & 2 i & 3@xmath2441 & 2 & 3@xmath2441 & & & & & 3 & 2 & 1 & & & & & & & & & & & + & & ii+i & 2@xmath2442 & 1 & 2@xmath2441 & & & & & 1 & 4 & 1 & & & & & & & & & & & + & & 2 ii & 2@xmath2442 & 1 & 2@xmath2442 & & & & & @xmath318 & @xmath330 & & & & & & & & & & & & + & & 2 i & 3@xmath2442 & 1 & 3@xmath2441 & & & & & 3 & & & & & & & & & & & & & + & & 3 i & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & & & 1 & 2 & 1 & & & & & & & & & & & + & & 3 i & 3@xmath2441 & 1 & 3@xmath2441 & 1 & 2@xmath2441 & & & 1 & 1 & & & & & & & & & & & & + & & 2 i+ii & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2442 & & & 1 & 4 & & & & & & & & & & & & + & & 2 i+ii & 3@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2442 & & & 1 & & & & & & & & & & & & & + & & 4 i & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2 & & & & & & & & & & & & + & & 4 i & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 2@xmath2441 & 1 & 3@xmath2441 & 1 & & & & & & & & & & & & & + our objective is to calculate the degeneracy @xmath332 of each energy level @xmath333 ( or sector of states that share the same energy of eq .
( [ cgenergy ] ) ) . equation ( [ matrix ] )
imposes @xmath301 constraints on the @xmath334 possible values of @xmath298 .
thus , for each set of integers @xmath284 and @xmath285 , the degeneracy is equal to @xmath335 .
as there are @xmath287 such sets of integers ( see eq .
( [ matrix ] ) ) , the degeneracy of each level @xmath333 is @xmath336 we may recast eq .
( [ dld ] ) to highlight the effect of topology and invoke the euler relation ( eqs .
( [ c22 ] ) and ( [ euler ] ) ) to write the degeneracy as @xmath337 where we define @xmath338 the modular rank of the matrix @xmath291 lies in the interval @xmath339 .
it thus follows that @xmath340 from eqs .
( [ deg - gen ] ) and ( [ clambda2 ] ) , it is readily seen that @xmath341 the degeneracy of eq .
( [ mindg ] ) ( stemming from the spectral redundancy of each level @xmath333 seen in eq .
( [ deg - gen ] ) ) is consistent with an effective composite symmetry @xmath342 i.e. , the product of @xmath343 symmetries of the @xmath344 type .
that is , if each element of such a @xmath344 symmetry gave rise to a @xmath345-fold degeneracy then the result of eq .
( [ deg - gen ] ) will naturally follow . the non - local symmetry of eq .
( [ zdq ] ) compound the standard local symmetries that appear in the gauge theory limit of eq .
( [ cgenergy ] ) in which the @xmath346 terms are absent , i.e. , @xmath347 .
the latter gauge theory enjoys the local symmetries @xmath348 with , at any lattice vertex ( site ) @xmath41 , the angle @xmath349 being an arbitrary integer multiple of @xmath350 . in this case
, we find that the ground state degeneracy ( @xmath351 ) is purely topological ( i.e. , not holographic ) , @xmath352 where , @xmath353 these equations extend the degeneracy @xmath354 found in subsection [ construct_genus ] for the ising ( @xmath355 ) lattice gauge theory @xcite .
we next turn to a simple @xmath0 theory @xmath356 where the `` fluxes '' @xmath357 are , respectively , the sums of the angles on all edges emanating from site @xmath39 and the sum of all angles @xmath358 on edges that belong to a plaquette @xmath42 . in the continuum limit ( in which the lattice constant @xmath139 tends to zero ) , the @xmath359 term may be taylor expanded as the flux is small , @xmath360 in the usual way . then , omitting an irrelevant constant additive term , the hamiltonian becomes in the standard manner @xmath361 where @xmath362 ( with @xmath363 a vector potential ) is the conventional magnetic field along the direction transverse to the plane where the lattice resides . in the @xmath364 limit , the @xmath0 hamiltonian of eq .
( [ fu(1 ) ] ) follows from eqs .
( [ sijk ] ) , ( [ newa ] ) , and ( [ newb ] ) where @xmath365 , and @xmath366 with @xmath303 . in the @xmath367 limit , the discrete clock symmetry becomes a continuous rotational symmetry , @xmath368 .
rather trivially , yet notably , in this limit , the system becomes _ gapless_. repeating _ mutatis mutandis _ the considerations of eqs .
( [ deg - gen ] ) and ( [ mindg ] ) , in the continuous large @xmath345 limit , a genus dependent symmetry is naturally associated with the system degeneracy .
peculiarly , in this limit , similar to eq .
( [ zdq ] ) , a genus dependent @xmath369 symmetry may appear for the toric @xmath0 theory of eq .
( [ fu(1 ) ] ) . in the limiting case in which the star term does not appear in eq .
( [ fu(1 ) ] ) , i.e. , that of @xmath370 , a symmetry of the type of eq . ( [ geq ] ) compounds the known local @xmath0 symmetry , @xmath371 similar to eq . but with an arbitrary real phase @xmath349 at each lattice vertex ( site ) @xmath41 .
these local symmetries are lifted once the @xmath372 term is introduced , as in eq .
( [ fu(1 ) ] ) .
thus , similar to the clock gauge theory ( whose degeneracy was given by eqs .
( [ min_gauge ] ) , and ( [ min_gauge_con ] ) ) , this _
@xmath0 lattice gauge theory exhibits a genus dependent degeneracy_.
thus far , we focused on square lattice realizations of the ising , clock , and @xmath0 theories . for completeness , we now examine other lattice geometries .
specifically , we study the honeycomb lattice ( h ) and triangular lattice ( t ) incarnations of our classical theory and determine their ground state degeneracies . in fig .
[ hex-tri-1.png ] , @xmath180 and @xmath181 are defined for each lattice .
the hamiltonians are given by @xmath373 our numerical results are summarized in table [ table4 ] .
these results are consistent with eqs . and .
@xmath374 & @xmath375 & @xmath376 & @xmath377 & @xmath378 & @xmath379 & @xmath380 & @xmath381 + & & 8 & 27 & 64 & 125 & 216 & 343 & 512 & 729 + & & 16 & 27 & & & & & & + & & 8 & & & & & & & + & & 128 & & & & & & & + & & 128 & & & & & & & + & & 64 & & & & & & & + as is well known , the h and t lattices are dual lattices ( fig . [ hex-tri-2.png ] ) .
this duality implies that the classical toric code models of eq .
( [ hth ] ) yield the same results . from figs .
[ hex-tri-1.png ] and [ hex-tri-2.png ] , as a consequence of duality , what is defined as @xmath180 ( @xmath181 ) in h corresponds to some @xmath382 ( @xmath383 ) in t , and vice versa .
this indicates that @xmath384 after this transformation we can rewrite eqs .
( [ hth ] ) as , @xmath385 and assuming @xmath386 , it is seen that @xmath387 .
this simple analysis does not take into account potential boundary terms that may appear in finite lattices , as a result of the duality transformation .
in this section , we dwell on a few more ising type spin systems , similar to type ii commensurate lattice realizations of the classical toric code model ( eq .
( [ cdeg ] ) ) , in which the degeneracy is holographic , i.e. , exponential in the system s boundary .
we now discuss a discretized version of the compass model @xcite , the `` @xmath388-state potts compass model '' on an @xmath389 square lattice with periodic boundary conditions .
the hamiltonian is given by , @xmath390 where at each site ( vertex ) @xmath41 there are two ising type spins @xmath391 , @xmath392 , while the occupation numbers @xmath393 and @xmath394 .
then , at each site , there is either a @xmath395 or a @xmath396 degree of freedom .
the cartesian unit vectors @xmath397 and @xmath398 link neighboring sites of the square lattice .
spins of the @xmath395 type interact along the @xmath233-direction ( horizontally ) while those of the @xmath396 variety interact along the @xmath399-direction ( vertically ) . minimizing the energy
is equivalent to maximizing the number of products in the summand of eq .
( [ h_p ] ) that are equal to @xmath400 . in a configuration
in which at all sites there is a @xmath395 ( and no @xmath396 ) spin , the system effectively reduces to that of @xmath401 independent ising chains parallel to the @xmath233 direction . for each such chain , there are two ground states : @xmath402 or @xmath403 for all lattice sites .
as these chains are independent , there are @xmath404 ground states .
replacing some sites with @xmath396 spins some bonds turn into @xmath183 and energy increases as a result . repeating the same procedure where all sites are occupied by @xmath396 spins ,
we find out that there are @xmath405 independent vertical ising chains and so @xmath406 states giving the same minimum energy .
the ground state degeneracy of eq .
( [ h_p ] ) is @xmath407 . for a more general case with genus @xmath1 ( composed of regions @xmath408 connected by bridges @xmath409 ( shared by regions @xmath218 and @xmath224 ) ) ,
the degeneracy again depends on the number of independent horizontal ( @xmath410 ) and vertical ( @xmath405 ) ising chains .
if each region @xmath220 is of size @xmath411 ( @xmath219 ) and @xmath223 ( @xmath412 ) is the number of edges connecting @xmath218 and @xmath224 , then , the ground state degeneracy will be @xmath413 where @xmath414 this degeneracy depends on both the geometry and the topology of the lattice .
we briefly highlight the effects of topology in the degeneracy of eqs .
( [ npc ] ) and ( [ lylx+ ] ) .
panel a ) of fig .
[ genus.png ] depicts a genus one lattice for which @xmath415 and @xmath416 . by redefining the way spins
are connected and boundary conditions , as we explained before , we may transform it into , e.g. , @xmath417 lattices as in fig .
[ genus.png ] ( panels b ) and c ) , respectively ) . here
, one may readily verify that although @xmath418 and the total number of spins do not change , @xmath405 varies ( increases ) as a result of increasing the genus number . as discussed earlier , our classical toric code model of eq .
( [ h ] ) is identical to the spin ( defined on vertices ) plaquette model of eq .
( [ hw ] ) .
this latter hamiltonian is , as it turns out , a particular limiting case of the so - called `` xu - moore model '' @xcite , one in which its transverse field is set to zero and the model becomes classical . in its original rendition , this classical limit of the xu - moore model
has a degeneracy exponential in the system s boundary .
this degeneracy appears regardless of the parity of the system sides .
we now discuss how to relate the degeneracy in our system to that of the classical xu - moore model . to achieve this , instead of applying periodic boundary conditions along the cartesian directions as in the classical toric code model ( i.e. along the solid lines of fig .
[ xm.png ] ) , we endow the system with different boundary conditions .
specifically , we examine instances in which periodic boundary conditions are associated with the diagonal @xmath419 and @xmath420 axis ( @xmath142 angle rotation of the original square lattice ) of fig .
[ xm.png ] .
a simple calculation then illustrates that the ground state sector as well as all other energies have a global degeneracy factor , @xmath421 where @xmath422 and @xmath423 are defined as in eq .
( [ lylx+ ] ) but along the diagonal directions ( dotted lines in fig .
[ xm.png ] ) .
a similar ( global ) degeneracy appears in the classical 90@xmath424 orbital compass model @xcite ( having only nearest neighbor two - spin interactions ) to which the xu - moore model is dual .
we conclude our discussion of holographic degeneracy in spin models with a brief review of an ising system even simpler than the ones discussed above . specifically , we may consider an ising spin system on a square lattice with its lattice constant @xmath139 set to unity when it is embedded on a torus ( @xmath7 ) with periodic boundary conditions along the @xmath419 and @xmath420 diagonals with the hamiltonian @xmath425 here interactions are anti - ferromagnetic between next - nearest neighbors ( @xmath426 ) and next - next - nearest neighbors ( @xmath427 ) .
it is straightforward to demonstrate that this system has a ground state degeneracy scales as @xmath428 where @xmath429 are the lattice sizes along the @xmath419 and @xmath420 directions @xcite .
in this work , we demonstrated that a topological ground state degeneracy ( one depending on the genus number of the riemann surface on which the lattice is embedded ) does not imply concurrent topological order ( i.e. , eq .
( [ tqo ] ) is violated and distinct ground states may be told apart by local measurements ) .
we illustrated this by introducing the classical toric code model ( eq .
( [ h ] ) with @xmath430 ) . as we showed in some detail , under rather mild conditions ( those pertaining to `` type i '' lattices in the classification of eq .
( [ cdeg ] ) ) , the ground state degeneracy solely depends on topology . in these classical systems , however , the ground states ( given by , e.g. , eqs .
( [ gn+ ] ) and ( [ ggeneral ] ) on the torus ) are distinguishable by measuring the pattern of @xmath431 on a finite number of nearest neighbor edges ; thus , the ground states do not satisfy eq .
( [ tqo ] ) and are , rather trivially , not topologically ordered .
they are landau ordered instead and , most importantly , illustrate that the ground states are related by @xmath432 ( global ) gauge - like symmetries contrary to the @xmath120 symmetries of kitaev s toric code model @xcite . in the more commensurate type ii lattice realizations of the classical toric code model as well as in a host of other systems ,
the ground state degeneracy is `` holographic''- i.e. , exponential in the linear size of the lattice @xcite .
this classical holographic effect is different from more subtle deeper quantum relations , for entanglement entropies , e.g. , @xcite . in all lattices and topologies , the minimal ground state degeneracy ( and that of all levels in the system ) of the classical model is robust and bounded from below by @xmath9 with @xmath1 the genus number .
we find similar genus dependent minimal degeneracies in clock and @xmath0 theories ( including _ lattice gauge theories _ ) . for completeness
, we remark that a degeneracy of the form @xmath433 with @xmath434 a quantity bounded from above by the linear system size ( viz . , a holographic entropy ) also appears in bona fide topologically ordered systems such as the `` haah code '' @xcite . beyond demonstrating that such degeneracies may arise in classical theories
, we illustrated that these behaviors may arise in rather canonical clock and @xmath0 type theories .
we provided a simple framework for studying and understanding the origin of these ubiquitous topological and holographic degeneracies .
we conclude with one last remark .
our results for classical systems enable the construction of simple _ quantum models _ with ground states that may be told apart locally ( i.e. , violating eq . for topological quantum order ) yet , nevertheless , exhibit a topological ground state degeneracy ) .
we present one , out of a large number of possible , routes to write such models exactly .
consider any one of the different theories studied in our work .
let us denote the classical hamiltonian associated with any of these theories by @xmath435 and corresponding local observables that may differentiate ground states apart by @xmath209 .
one may then apply any product @xmath436 of local unitary transformations to both the hamiltonian and the corresponding `` order parameter '' local observable @xmath209 .
that is , we may consider the `` quantum '' hamiltonian @xmath437 and the corresponding local operator @xmath438 . by virtue of the unitary transformation , both in the ground state sector ( as well as at any finite temperature ) , the expectation value of the local observable @xmath209 in the classical system given by @xmath435 is identical to the expectation value of the @xmath439 in the quantum system governed by @xmath440 . to be concrete
, one may consider , e.g. , the classical toric code ( ctc ) model .
that is , e.g. , one may set @xmath441 that contains only classical ising ( @xmath442 ) spins .
next , consider the unitary operator @xmath443 $ ] that effects a @xmath444 rotation of all spins at sites @xmath445 that belong to the sublattice @xmath144 about the internal @xmath446 axis .
( that is , indeed , @xmath447 . ) thus , trivially , the resulting hamiltonian @xmath440 contains non - commuting @xmath448 and @xmath449 and is `` quantum '' ( just as the kitaev toric code model of section [ toric_quantum ] @xcite that may be mapped to two decoupled classical ising spin chains @xcite ) contains exactly these two quantum spin components and is `` quantum '' ) . by virtue of the local product nature of the mapping operator @xmath436
, the classical local observables @xmath209 that we discussed in our paper become now new local observables @xmath450 in the quantum model .
thus , putting all of the pieces together , we may indeed generate quantum models with a topological degeneracy in which the ground state may be told apart by local measurements .
this work was partially supported by the national science foundation under nsf grant no . dmr-1411229 and the feinberg foundation visiting faculty program at the weizmann institute .
we are very grateful to a discussion with j. haah in which he explained to us the degeneracies found in his model @xcite .
in type i lattices ( and their simplest composites ) , the canonical partition function of the classical toric code model is given by eq .
( [ ctcz ] ) .
the situation is somewhat richer for other lattices .
below , we briefly write the partition functions for several such finite size lattices . for simplicity we set @xmath451 and @xmath278 in the classical rendition of eq .
( [ h ] ) and perform a high temperature ( h - t ) and low temperature ( l - t ) series expansion which is everywhere convergent for these finite size systems .
one can follow a similar procedure and find the partition functions for @xmath452 .
we start with h - t series expansion , @xmath453 where @xmath454 and @xmath455 . in eq .
( [ p1 ] ) after expanding the products , and summing over all configurations , the only surviving terms are those for which the product of a subset of @xmath169 s and @xmath170 s is equal to @xmath456 and this corresponds to one constraint or a product of two or more of them sharing no star or plaquette operators .
thus , @xmath457 where @xmath32 is the number of faces and @xmath30 is the number of vertices .
the factor of @xmath458 ( with @xmath459 the number of spins or lattice edges ) originates from the summation @xmath460 ( each @xmath210 has two values ( @xmath293 ) , with @xmath461 ) .
the sole non - vanishing traces in eq .
( [ p1 ] ) originate from the constraints of eqs .
( [ asbp ] ) and ( [ wl ] ) and their higher genus counterparts . while this procedure trivially gives rise to the partition function of eq .
( [ ctcz ] ) for simple lattices , the additional constraints in other lattices spawn new terms in the partition functions .
in the following we develop the l - t series expansion for @xmath278 . from eq .
( [ ctcz ] ) , @xmath462 where @xmath463 is the ground state energy and @xmath259 is the ground state degeneracy .
numerical results illustrate that the integers @xmath464 are larger than or equal to @xmath456 .
one can generalize this form for @xmath465 @xmath466 where @xmath99 and @xmath332 indicate energy and degeneracy of energy level @xmath333 for a given @xmath345 , respectively .
below is a sample of our numerical results for @xmath467 and @xmath468 of lattices with different sizes , @xmath345 s and genus numbers ( @xmath469 ) . from @xmath468
, we can easily see that exited states have a degeneracy `` higher than or equal to '' the ground state degeneracy ( @xmath470 and @xmath471 ) .
1 . @xmath472 : 1 .
@xmath473 : 1 .
@xmath474 2 .
@xmath475 3 .
@xmath476 4 .
@xmath477 5 .
@xmath478 2 .
@xmath479 : 1 .
@xmath480 2 .
@xmath481 3 .
@xmath482 3 .
@xmath483 : 1 .
@xmath484 2 .
@xmath485 3 .
@xmath486 4 .
@xmath487 : 1 .
@xmath488 2 .
@xmath489 5 .
@xmath490 : 1 .
@xmath491 6 .
@xmath492 : 1 .
@xmath493 2 .
@xmath494 1 .
@xmath495 : 1 .
@xmath496 2 .
@xmath497 3 .
@xmath498 2 .
@xmath499 : 1 .
@xmath500 2 .
@xmath501 3 .
@xmath502 : 1 .
@xmath503 2 .
@xmath504 4 .
@xmath505 : 1 .
@xmath506 2 .
@xmath507 3 .
@xmath508 : 1 .
@xmath509 : 1 .
@xmath510 2 .
@xmath511 for an explicit statement of the perceptive lore , see , e.g. , page 8 of @xcite for a remark concerning the quite deep standard manifestation of these degeneracies in the rich quantum arena : the existence of topologically - degenerate ground states proves the existence of topological order .
topological degeneracy can also be used to perform fault - tolerant quantum computations ( kitaev , 2003 ) . " specifically , on a cube of size @xmath512 endowed with periodic boundary conditions , the degeneracy of haah s code of @xcite is equal to @xmath433 with @xmath434 set by the degree ( in @xmath513 ) of the polynomial that constitutes the greatest common divisor of the below three polynomials , @xmath514 where @xmath515 .
this degeneracy varies between @xmath388 and an exponential in @xmath131 and can exhibit capricious jumps as the lattice size is varied .
for instance ,
( i ) when @xmath516 ( with @xmath16 a natural number ) the degeneracy is @xmath517 while ( ii ) when @xmath518 , the degeneracy is four @xcite .
r. peierls , proc .
32 * , 477 ( 1936 ) ; l. d. landau and e. m. lifshitz , statistical physics - course of theoretical physics vol 5 ( pergamon , london , 1958 ) ; j. frolich , b. simon , and t. spencer , commun . math . phys . * 50 * , 79 ( 1976 ) ; r. peierls , contemporary physics * 33 * , 221 ( 1992 ) .
v. kalmeyer and r. b. laughlin , phys .
* 59 * , 2095 ( 1987 ) ; v. kalmeyer and r. b. laughlin , phys .
b * 39 * , 11879 ( 1989 ) ; x .- g .
wen , f. wilczek , and a. zee , phys .
b * 39 * , 11413 ( 1989 ) ; h. yao and s. a. kivelson , phys . rev . lett . *
99 * , 247203 ( 2007 ) ; l. messio , b. bernu , and c. lhuillier , phys .
* 108 * , 207204 ( 2012 ) ; yin - chen he , d. n. sheng , and yan chen , phys .
lett . * 112 * , 137202 ( 2014 ) .
wen and q. niu , phys .
b * 41 * , 9377 ( 1990 ) .
wen , phys .
b * 40 * , 7387 ( 1989 ) .
b * 4 * , 239 ( 1990 ) .
g. ortiz , z. nussinov , j. dukelsky , and a. seidel , phys .
b * 88 * , 165303 ( 2013 ) . in any particular fixed gauge ,
the degeneracy is @xmath9 ( see , e.g. , eq .
( 187 ) of @xcite for a derivation ) .
the factor of @xmath519 is the number of independent gauge fixes as can be seen as follows .
the operators @xmath520 of eq .
( [ ab ] ) connect one gauge fix to another .
there are @xmath521 such operators , and they are not independent of each other . specifically , the local gauge symmetry operators @xmath520 adhere to the single global constraint of eq .
( [ abc ] ) .
s. a. kivelson , d. s. rokhsar , and j. p. sethna , phys .
b 35 , 8865 ( 1987 ) ; d. s. rokhsar and s. a. kivelson , phys .
61 , 2376 ( 1988 ) ; r. moessner and k. s. raman , arxiv:0809.3051 ( 2008 ) ; f. s. nogueira and z. nussinov , physical review b * 80 * , 104413 ( 2009 ) .
apart from the models that we will introduce in the current paper , holographic degeneracies also appear in short range ising , @xmath522 , non - abelian gauge background , tiling @xcite , and other models .
indeed , holographic entropies have also been reported on fractal structures realized on lattices @xcite .
b. yoshida , phys .
b * 88 * , 125122 ( 2013 ) ; ann . of phys . *
338 * , 134 ( 2013 ) .
e. cobanera , g. ortiz , and z. nussinov , phys . rev
. lett . * 104 * , 020402 ( 2010 ) ; e. cobanera , g. ortiz , and z. nussinov , adv . in phys . *
60 * , 679 ( 2011 ) ; z. nussinov and g. ortiz , europhysics letters * 84 * , 36005 ( 2008 ) ; z. nussinov and g. ortiz , phys .
b * 79 * , 214440 ( 2009 ) ; g. ortiz , e. cobanera , and z. nussinov , nuclear physics b * 854 * , 780 ( 2012 ) .
|
we challenge the hypothesis that the ground states of a physical system whose degeneracy depends on topology must necessarily realize topological quantum order and display non - local entanglement . to this end
, we introduce and study a classical rendition of the toric code model embedded on riemann surfaces of different genus numbers .
we find that the minimal ground state degeneracy ( and those of all levels ) depends on the topology of the embedding surface alone . as the ground states of this classical system
may be distinguished by local measurements , a characteristic of landau orders , this example illustrates that topological degeneracy is not a sufficient condition for topological quantum order .
this conclusion is generic and , as shown , it applies to many other models .
we also demonstrate that certain lattice realizations of these models , and other theories , display a ground state entropy ( and those of all levels ) that is holographic " , i.e. , extensive in the system boundary .
we find that clock and @xmath0 gauge theories display topological ( in addition to gauge ) degeneracies .
|
the goal of this paper is to provide computations of the @xmath0 casson invariant for 3-manifolds obtained by dehn surgery on a two - bridge knot .
our approach is to apply the dehn surgery formula of @xcite and @xcite , and this involves computing the culler shalen seminorms . in general
, the surgery formula applies to dehn surgeries on small knots @xmath4 in homology spheres @xmath5 , and a well - known result of hatcher and thurston @xcite shows that all two - bridge knots are small . the culler
shalen seminorm plays a key role in the surgery formula , and we use the results of ohtsuki @xcite to provide the required computations . as an application , using the classification of exceptional dehn surgeries on two - bridge knots from @xcite ,
we prove that nearly all 3-manifolds given by a nontrivial @xmath1dehn surgery on a hyperbolic two - bridge knot @xmath4 have nontrivial @xmath0 casson invariant . in our previous work ( * ? ? ?
* theorem 2.2 ) , we showed that for any two - bridge knot or torus knot @xmath4 , the homology 3-sphere obtained by @xmath6dehn surgery on @xmath4 has nonzero @xmath0 casson invariant whenever @xmath7 , and theorem [ nontriv ] gives a new proof extending this result to nearly all @xmath8 surgeries .
our second application is to @xmath2-polynomials .
the @xmath2-polynomial as originally defined in @xcite has no repeated factors , but in @xcite , boyer and zhang give an alternative approach by using the degree of the restriction map to assign multiplicities to each one - dimensional component in the character variety .
for a given knot , we denote the original @xmath2-polynomial by @xmath9 and the boyer - zhang polynomial by @xmath10 ( see section [ a - polynomial ] ) .
if @xmath4 is a small knot , then @xmath10 and @xmath9 have the same irreducible factors , only @xmath10 may include factors of higher multiplicity .
we exploit the close relationship between the @xmath0 casson invariant , culler shalen seminorms , and the @xmath3-polynomial to determine closed formulas for the @xmath11-degrees of @xmath10 for all two - bridge knots and for the @xmath12-degrees of @xmath10 for double twist knots .
our techniques also enable computations of the @xmath12-degree of @xmath10 for other two - bridge knots on a case - by - case basis . by comparing our results to known calculations of the @xmath2-polynomial , we are able to identify knots for which @xmath13 . in some cases , we are also able to determine the multiplicities of the factors of @xmath14 and we illustrate this for the examples of the knots @xmath15 and @xmath16 .
we briefly outline the contents of this paper .
we begin with an introduction of the notation for the @xmath0 character varieties , the definition of the @xmath0 casson invariant , culler shalen seminorms , and the surgery formula for the @xmath0 casson invariant .
we then present the two - bridge knots and establish a regularity result for all slopes .
we also determine the multiplicities of the curves in the character variety ; in fact we show they all equal one .
following @xcite , we explain how to calculate the culler
shalen seminorms , and we provide explicit calculations for the family of examples given by the double twist knots @xmath17 ( see figure [ gentwist ] ) .
we use this to compute the @xmath0 casson invariant for 3-manifolds obtained by dehn surgery on @xmath4 .
we prove a nontriviality result for the 3-manifold invariant for most dehn surgeries on hyperbolic two - bridge knots . in the final section
we introduce the @xmath2-polynomial and @xmath3-polynomial , and we apply our results to make some general comments about the @xmath11- and @xmath12-degrees of these polynomials for two - bridge knots , and we show that the corresponding @xmath0 knot invariant is nontrivial for all small knots . in the appendix , we list the information on boundary slopes needed to calculate the @xmath0 casson invariant for surgeries on 2-bridge knots with up to 8 crossings , and we also list there the @xmath11- and @xmath12-degrees of @xmath9 and @xmath10 .
given a finitely generated group @xmath18 , we set @xmath19 to be the space of representations @xmath20 and @xmath21 the subspace of irreducible representations . recall from @xcite that @xmath19 has the structure of a complex affine algebraic set .
the _ character _ of a representation @xmath22 is the function @xmath23 defined by setting @xmath24 for @xmath25 .
the set of characters of @xmath0 representations is denoted @xmath26 and also admits the structure of a complex affine algebraic set .
let @xmath27 denote the subspace of characters of irreducible representations .
define @xmath28 by @xmath29 , and note that @xmath30 is surjective . given a manifold @xmath5 ,
we denote by @xmath31 the variety of @xmath0 representations of @xmath32 and by @xmath33 the associated character variety .
we briefly recall the definition of the @xmath0 casson invariant .
suppose now @xmath5 is a closed , orientable 3-manifold with a heegaard splitting @xmath34 . here , @xmath35 is a closed orientable surface embedded in @xmath5 , and @xmath36 and @xmath37 are handlebodies with boundaries @xmath38 such that @xmath39 .
the inclusion maps @xmath40 and @xmath41 induce surjections of fundamental groups . on the level of character varieties ,
this identifies @xmath33 as the intersection @xmath42 there are natural orientations on all the character varieties determined by their complex structures .
the invariant @xmath43 is defined as an oriented intersection number of @xmath44 and @xmath45 in @xmath46 which counts only compact , zero - dimensional components of the intersection . specifically , there exist a compact neighborhood @xmath47 of the zero - dimensional components of @xmath48 which is disjoint from the higher dimensional components of the intersection and an isotopy @xmath49 supported in @xmath47 such that @xmath50 and @xmath45 intersect transversely in @xmath47 .
given a zero - dimensional component @xmath51 of @xmath52 , we set @xmath53 , depending on whether the orientation of @xmath50 followed by that of @xmath45 agrees with or disagrees with the orientation of @xmath46 at @xmath54 .
let @xmath55 where the sum is over all zero - dimensional components of the intersection @xmath52 . in @xcite ,
a surgery formula was established for @xmath43 for dehn surgeries on small knots in integral homology 3-spheres .
this result is of central importance in this paper , so we introduce the background material on culler shalen seminorms and recall the statement of the theorem .
now suppose @xmath12 is a compact , irreducible , orientable 3-manifold with boundary a torus .
an _ incompressible surface _ in @xmath12 is a properly embedded surface @xmath56 such that @xmath57 is injective and no component of @xmath35 is a 2-sphere bounding a 3-ball .
a surface @xmath35 in @xmath12 is called _ essential _ if it is incompressible and has no boundary parallel components .
the manifold @xmath12 is called _ small _ if it does not contain a closed essential surface , and a knot @xmath4 in @xmath5 is called _ small _ if its complement @xmath58 is a small manifold , where @xmath59 is a tubular neighborhood of @xmath4
. if @xmath60 is a simple closed curve in @xmath61 , the dehn filling of @xmath12 along @xmath60 will be denoted by @xmath62 ; it is the closed 3-manifold obtained by identifying a solid torus with @xmath12 along their boundaries so that @xmath60 bounds a disk .
note that the homeomorphism type of @xmath62 depends only on the _ slope _ of @xmath60 that is , the unoriented isotopy class of @xmath60 .
primitive elements in @xmath63 determine slopes under a two - to - one correspondence .
if @xmath35 is an essential surface in @xmath12 with nonempty boundary , then all of its boundary components are parallel and the slope of one ( hence all ) of these curves is called the _ boundary slope _ of @xmath35 .
a slope is called a _ strict boundary slope _ if it is the boundary slope of an essential surface that is not the fiber of any fibration of @xmath12 over @xmath64 . for each @xmath65
there is a regular map @xmath66 defined by @xmath67 let @xmath68 be the inverse of the hurewicz isomorphism . identifying @xmath69 with its image in @xmath70 under the natural map @xmath71
we obtain a well - defined function @xmath72 on @xmath73 for each @xmath74 let @xmath75 be the regular function defined by @xmath76 for @xmath74 let @xmath77 be the restriction map induced by @xmath78 .
suppose @xmath79 is an algebraic component of @xmath73 with @xmath80 such that @xmath81 is also one - dimensional .
let @xmath82 denote the regular function obtained by restricting @xmath83 to @xmath79 for each @xmath84 .
let @xmath85 denote the smooth , projective curve birationally equivalent to @xmath79 .
regular functions on @xmath79 extend to rational functions on @xmath85 .
we abuse notation and denote the extension of @xmath86 to @xmath85 also by @xmath87 if @xmath12 is hyperbolic , then proposition 3.1.1 of @xcite establishes the existence of a discrete faithful irreducible representation @xmath88 and proposition 2 of @xcite implies that the algebraic component @xmath89 of @xmath73 containing the character @xmath90 is one - dimensional . in section 1.4 of @xcite , the authors construct a norm @xmath91 on the real vector space @xmath92 called the culler shalen norm associated to @xmath89 .
the same construction works for other dimension one components @xmath79 of @xmath73 , and in general , one obtains a seminorm @xmath93 on @xmath92 called the culler
shalen seminorm associated to @xmath79 .
[ cs - seminorm ] for @xmath79 a one - dimensional component of @xmath73 containing an irreducible character and whose restriction @xmath81 is also one - dimensional , define the seminorm @xmath93 on @xmath94 by setting @xmath95 for all @xmath96 in the lattice @xmath63 .
we call a one - dimensional algebraic component @xmath79 of @xmath73 a _ norm curve _ if @xmath97 defines a norm on @xmath94
. the @xmath0 casson invariant of a manifold obtained by dehn filling is closely related to this seminorm ; however we must impose certain restrictions on the slope of the dehn filling .
[ irregular ] the slope of a simple closed curve @xmath60 in @xmath61 is called _ irregular _ if there exists an irreducible representation @xmath98 such that * the character @xmath99 of @xmath22 lies on a one - dimensional component @xmath79 of @xmath73 such that @xmath81 is one - dimensional , * @xmath100 for all @xmath101 in the image of @xmath102 * @xmath103 is the cyclic group generated by @xmath104 \in \pi_1(\partial m)$ ] .
a slope is called _ regular _ if it is not irregular . with these definitions , we are almost ready to state the dehn surgery formula .
but first we recall some useful notation for dehn fillings of knot complements . for any choice of basis @xmath105 for @xmath106
there is a bijective correspondence between unoriented isotopy classes of simple closed curves in @xmath61 and elements in @xmath107 given by @xmath108 where @xmath109 .
if @xmath12 is the complement of a knot @xmath4 in an integral homology 3-sphere @xmath5 , then the meridian @xmath110 and preferred longitude @xmath111 of the knot @xmath4 provide a basis for @xmath112 consider the 3-manifold @xmath62 resulting from dehn filling along the curve @xmath113 in this case , we call @xmath8 the slope of @xmath60 and denote by @xmath114 the 3-manifold obtained by @xmath1dehn surgery along the knot @xmath4 . a slope @xmath8 is called _ admissible _ for @xmath4 if 1 .
@xmath8 is a regular slope which is not a strict boundary slope , and 2 .
no @xmath115-th root of unity is a root of the alexander polynomial of @xmath4 , where @xmath116 if @xmath117 is odd and @xmath118 if @xmath117 is even .
the next result is a restatement of theorem 4.8 of @xcite , as corrected in @xcite .
[ surg - form ] suppose @xmath4 is a small knot in an integral homology 3-sphere @xmath5 with complement @xmath12 .
let @xmath119 be the collection of all one - dimensional components of the character variety @xmath73 such that @xmath81 is one - dimensional and such that @xmath120 is nonempty .
define @xmath121 by @xmath122 if @xmath117 is even and @xmath123 if @xmath117 is odd .
then there exist integral weights @xmath124 depending only on @xmath79 and non - negative numbers @xmath125 depending only on @xmath4 such that for every admissible slope @xmath8 , we have @xmath126 we define the total culler shalen seminorm @xmath127 we note that this is half the norm defined by culler - gordon - luecke - shalen in @xcite and extended to the seminorm case in @xcite and @xcite .
it is twice the norm defined in @xcite .
we briefly recall some useful properties of the @xmath0 casson invariant and we refer to @xcite and @xcite for further details . on closed 3-manifolds @xmath128 the invariant @xmath129 is nonnegative , satisfies @xmath130 under orientation reversal , and is additive under connected sum of @xmath131homology 3-spheres ( cf .
theorem 3.1 , @xcite ) .
if @xmath5 is hyperbolic , then @xmath132 by proposition 3.2 of @xcite .
if @xmath4 is a small knot in an integral homology 3-sphere @xmath5 , then theorem [ surg - form ] implies that the difference @xmath133 is independent of @xmath117 and @xmath134 provided @xmath134 is chosen sufficiently large .
this allows one to define an invariant of small knots @xmath4 in homology 3-spheres by setting @xmath135 for @xmath134 sufficiently large .
for the remainder of this paper , @xmath136 will denote a two - bridge knot in @xmath137 with complement @xmath138 .
recall from ( * ? ? ?
* chapter 12 ) that given relatively prime integers @xmath139 with @xmath140 odd , we can associate a knot with two bridges , denoted @xmath141 whose associated knot group @xmath142 admits the presentation @xmath143 where @xmath144 for @xmath145 .
two such knots @xmath146 and @xmath147 are equivalent if and only if @xmath148 and @xmath149 thus we can choose @xmath150 so that @xmath151 . in this section , we will outline how to apply theorem [ surg - form ] to compute the @xmath0 casson invariant for 3-manifolds obtained by @xmath1dehn surgery on such a knot . recall that by the results of hatcher and thurston @xcite , it follows that all two - bridge knots are small .
further , we know that the @xmath0 casson invariant does not change under orientation reversal ( cf .
theorem 1.2 , @xcite ) , and since the manifolds obtained by @xmath1dehn surgery on a knot @xmath4 are , up to change of orientation , the same as those obtained by @xmath152dehn surgery on its mirror image @xmath153 , for our purposes , we can choose to work with @xmath4 or @xmath154 note further that the mirror image of the two - bridge knot @xmath136 is just @xmath155 ( cf .
p.185 , @xcite ) . in this section ,
we show that every slope @xmath8 on a two - bridge knot is regular
. this amounts to showing that any irreducible @xmath0 representation @xmath22 of a two - bridge knot group with @xmath156 and @xmath157 has @xmath158 as in section 7 of @xcite and in section 1 of @xcite , up to conjugation , we can assume that any irreducible representation @xmath171 sends @xmath172 \quad \text{and } \quad \varrho(y ) = \left [ \begin{array}{cc } \mu & 0\\ t & \mu^{-1 } \end{array } \right],\ ] ] where @xmath30 is chosen so that @xmath173 .
suppose @xmath8 is an irregular slope , where @xmath174 are taken to be relatively prime .
then there exists an irreducible representation @xmath22 of the knot complement satisfying the conditions of definition [ irregular ] , so in particular @xmath175 and @xmath176 we will show that this leads to a contradiction . since @xmath177 , we know that @xmath178 .
each of the four possible terms @xmath179 for @xmath180 has entries that are @xmath181 in the first row and monic linear polynomials ( up to sign ) in the second row .
a simple inductive proof then shows that @xmath182,\ ] ] where @xmath183 are monic polynomials in @xmath30 ( up to sign ) with @xmath184 and @xmath185 then @xmath186 \left [ \begin{array}{cc } a & b \\ c & d \end{array } \right ] = \left [ \begin{array}{cc } \pm a + c & \pm b + d \\
\pm c & \pm d \end{array } \right]\ ] ] and @xmath187 \left [ \begin{array}{cc } \pm 1 & 0 \\ t & \pm 1 \end{array } \right ] = \left [ \begin{array}{cc } \pm a + t b & \pm b \\
\pm c + t d & \pm d \end{array } \right].\ ] ] forming the product for the reversed word @xmath165 gives exactly the same matrix as one gets by flipping the matrix @xmath195 along the anti - diagonal , thus @xmath196.\ ] ] hence @xmath197\left[\begin{array}{cc } 0 & b \\ tb & a \end{array}\right ] = \left[\begin{array}{cc } tb^2 & 2ab \\ 0 & tb^2 \end{array } \right ] = \left[\begin{array}{cc } -1 & 2ab \\ 0 & -1 \end{array } \right]\ ] ] since @xmath198 furthermore , @xmath199^{2n } \left[\begin{array}{cc } -1 & 2 ab \\ 0 & -1 \end{array } \right ] = \left[\begin{array}{cc } -1 & 2 a b \mp 2n \\ 0 & -1 \end{array } \right].\ ] ] in summary , we have @xmath200 \quad \text{and } \quad \varrho({{\mathcal l}})=\left[\begin{array}{cc } -1 & 2ab \mp 2n \\ 0 & -1 \end{array } \right].\ ] ] if @xmath201 , then @xmath202^p \left[\begin{array}{cc } -1 & 2ab - 2n \\ 0 & -1 \end{array } \right]^q \\ & = ( -1)^q \left[\begin{array}{cc } 1 & p - q(2ab -2n ) \\ 0 & 1 \end{array } \right ] . \end{aligned}\ ] ] now @xmath117 and @xmath134 are assumed to be relatively prime , but for this matrix to equal the identity , we must have @xmath134 even and @xmath203 . reducing this equation
mod 2 and noting that @xmath204 are polynomials over @xmath205 in an algebraic integer , we conclude that @xmath117 must also be even , which is a contradiction . if @xmath206 then @xmath207^p \left[\begin{array}{cc } -1 & 2ab + 2n \\ 0 & -1 \end{array } \right]^q
\\ & = ( -1)^{p+q } \left[\begin{array}{cc } 1 & -p - q(2ab + 2n ) \\ 0 & 1 \end{array } \right ] . \end{aligned}\ ] ] for this matrix to equal the identity , we must have @xmath208 even and @xmath209 . again
, reducing mod 2 , this shows that @xmath117 is even , which gives the desired contradiction .
in this section we will show that the weights @xmath210 appearing in theorem [ surg - form ] for surgeries on a two - bridge knot are all equal to 1 .
this is achieved by identifying @xmath210 with the algebraic multiplicity @xmath211 of the associated curve in the character variety @xmath73 of the two - bridge knot complement .. in what follows , we will make use of continued fraction expansions for @xmath212 , which are expressions of the form @xmath213,$ ] where @xmath214 and satisfy @xmath215 note that , by replacing @xmath216 $ ] with @xmath217 $ ] , we can always arrange @xmath218 to be odd . if @xmath146 is a two - bridge knot with continued fraction expansion @xmath216 $ ] , then @xmath4 can be drawn as the 4-plat closure of the braid @xmath219 as shown in figure [ tunnel ] .
following riley @xcite and le @xcite , we explain how to view the character variety of a two - bridge knot group as a plane curve .
if @xmath220 is a group generated by the two elements @xmath221 and @xmath222 , then the cayley - hamilton theorem shows that every representation @xmath223 is determined up to conjugation by the three traces @xmath224 and @xmath225 . in the case of a two - bridge knot group @xmath161
, the presentation implies that @xmath221 and @xmath222 are conjugate , and using the coordinates @xmath226 and @xmath227 , we can view the character variety as a plane algebraic curve in @xmath228 .
any plane curve is defined by a single polynomial @xmath229 , which can be written in factored form as @xmath230 , where @xmath231 are the irreducible components and @xmath232 are their algebraic multiplicities . [ weights ] if @xmath4 is a two - bridge knot with complement @xmath138 , then each of the weights @xmath210 in the surgery formula of theorem [ surg - form ] is equal to the algebraic multiplicity @xmath211 of the associated curve @xmath79 of @xmath73 . in particular , we have @xmath233 .
suppose @xmath234 and let @xmath235 be the result of @xmath1-dehn surgery along @xmath4 . using an unknotting tunnel @xmath236 , we will construct a genus two heegaard splitting of @xmath12 and @xmath235 , which we use to identify the weight @xmath210 with the algebraic multiplicity @xmath211 for each curve @xmath79 of @xmath73 . choose a continued fraction expansion @xmath216 $ ] for @xmath212 with @xmath218 is odd , and draw @xmath4 as the 4-plat knot as shown in figure [ tunnel ] .
the arc @xmath236 connecting the bottom two lobes of @xmath4 is an unknotting tunnel for @xmath4 , and attaching @xmath236 to @xmath4 determines a genus two heegaard splitting @xmath237 for @xmath137 with @xmath36 a regular neighborhood of @xmath238 ( cf .
figure [ tunnel ] ) .
let @xmath239 denote the splitting surface , which has genus @xmath240 and suppose @xmath241 is a closed curve in @xmath35 giving a meridian around @xmath236 as shown .
let @xmath242 be a curve representing @xmath243 , chosen to avoid the attaching disk for the tunnel @xmath236 . if @xmath246 denotes the result of removing a small tubular neighborhood of @xmath4 from @xmath36 , then @xmath247 and @xmath37 form the knot complement @xmath12 .
similarly , if @xmath248 is the result of attaching a disk to @xmath247 along the curve @xmath249 , then @xmath250 is a heegaard splitting for the closed 3-manifold @xmath251 we have the following descriptions of character varieties : @xmath252 the casson invariant @xmath253 is by definition the intersection number of the zero - dimensional part of the intersection of @xmath254 and @xmath45 .
suppose now that @xmath79 is a curve in @xmath255 . in order to compute the weight @xmath210 associated to @xmath79
, we must view @xmath79 as a curve in the intersection of @xmath256 and @xmath45 .
this is the first step in computing the intersection of @xmath254 and @xmath45 .
as explained in the proof of proposition 4.3 of @xcite , the weight @xmath210 given in theorem [ surg - form ] is the intersection multiplicity of @xmath79 as a curve in the intersection @xmath257 in the character variety of the splitting surface .
let @xmath221 and @xmath222 be the two meridians of the knot shown , and choose curves @xmath258 and @xmath30 bounding disks in @xmath37 such that @xmath259 and @xmath30 form a symplectic basis for @xmath260 . thus @xmath261 and
it follows that @xmath262 is homeomorphic to the character variety of the free group on two generators @xmath221 and @xmath222 .
consider the presentation of the knot group @xmath263 from equation , where @xmath144 for @xmath264 .
notice that the generators @xmath161 coincide with the generators of the free group @xmath265 .
the element @xmath266 shown in figure [ tunnel ] is given by @xmath267 , and it follows that @xmath241 represents the element @xmath268 , which is precisely the relation in the knot group @xmath161 .
thus the defining equations for the curve @xmath79 in the intersection @xmath269 are precisely the defining equations of the curve @xmath79 in @xmath73 .
it follows that @xmath270 , and this proves the first statement of the proposition . to make the second conclusion ,
notice that proposition 3.4.1 of @xcite implies that the algebraic multiplicities all satisfy @xmath271 .
clearly , this implies @xmath233 and completes the proof of the proposition . a key ingredient in ohtsuki s computations are the results of hatcher and thurston @xcite determining the boundary slopes for all two - bridge knots .
they showed that two - bridge knots have integral boundary slopes , and they established a one - to - one correspondence between non - closed , incompressible , @xmath273-incompressible surfaces in the knot complement and continued fraction expansions @xmath216 $ ] for @xmath212 and for @xmath274 such that @xmath275 for all @xmath276 given a continued fraction expansion @xmath277 $ ] , we set @xmath278 to be the number of entries in @xmath279 $ ] whose signs agree with the alternating pattern @xmath280 $ ] , and we set @xmath281 to be the number of entries with the opposite sign .
we also denote by @xmath282 and @xmath283 the corresponding numbers for the seifert expansion of @xmath212 , which is the unique continued fraction expansion with only even entries . then by proposition 2 of @xcite , the boundary slope of the incompressible surface associated to the continued fraction expansion @xmath284 $ ] is given by @xmath285.\ ] ] the culler
shalen seminorm of a slope @xmath8 is a weighted sum of the intersections of @xmath8 with the boundary slopes of @xmath4 . in proposition 5.2 of this paper , ohtsuki finds these weights for a function called @xmath286 , with one minor error . in this proposition
he shows that the coefficient of the intersection @xmath287 for @xmath286 is @xmath288 .
however in section 4 he shows that the coefficient of @xmath289 in his formula should rather be @xmath290 , since the continued fraction expansion of @xmath212 for which all terms are even corresponds to the @xmath291boundary slope .
the following result is essentially proposition 5.2 from @xcite , after making this small correction to the statement ( cf .
of mattman s thesis @xcite ) . in this subsection
, we compute the correction terms that appear in the surgery formula ( see theorem [ surg - form ] ) .
these terms compensate for characters @xmath295 of irreducible representations @xmath296 which satisfy @xmath297 such representations satisfy @xmath298 , but they do not extend over @xmath1dehn surgery on @xmath4 .
the correction terms depend only on the parity of @xmath117 and are denoted @xmath299 and @xmath300 .
specifically , by the proof of theorem 4.8 of @xcite , @xmath299 is the sum of the weights of characters in @xmath73 for which @xmath301 , @xmath302 , and @xmath303 but for which there is no corresponding representation @xmath22 of @xmath304 , where @xmath117 is even , and @xmath300 is defined similarly in the case where @xmath117 is odd .
if @xmath117 is odd , we may compute @xmath300 by considering the result of @xmath291surgery on @xmath4 , which yields @xmath137 . in this case @xmath310 , so in fact @xmath311 .
but in section 1 of @xcite , ohtsuki shows that @xmath312 , and this completes the proof .
[ two - b ] let @xmath136 be a two - bridge knot .
suppose @xmath8 is not a strict boundary slope , and suppose no @xmath115-th root of unity is a root of the alexander polynomial of @xmath4 , where @xmath116 if @xmath117 is odd and @xmath118 if @xmath117 is even .
then @xmath313 note that ohtsuki s fomula allows culler shalen seminorms to be computed for any two - bridge knot by hand on a case - by - case basis .
the contributing boundary slopes and their weights are given for all prime knots up to 8 crossings in the appendix .
then theorem [ two - b ] allows the computation of the @xmath0 casson invariant for all admissible surgeries on two - bridge knots on a case - by - case basis .
we now compute general forumlas for the total culler
shalen seminorms for the double twist knots @xmath314 depicted in figure [ gentwist ] . here
, @xmath314 is drawn so that the vertical strands are twisted positively when @xmath315 and negatively when @xmath316 and so that the horizontal strands are twisted positively when @xmath317 and negatively when @xmath318 .
for example , @xmath319 is the figure-8 knot , @xmath320 is the left - handed trefoil , and @xmath321 is the right - handed trefoil .
note that @xmath322 and @xmath323 are unknots , and that @xmath324 and @xmath325 are just @xmath326 torus knots or links . in the general case , by sliding the horizontal crossings up and to the left , we can view @xmath314 as the two - bridge knot ( or link ) with continued fraction expansion @xmath327 $ ] that is obtained as the 4-plat closure of the braid @xmath328 notice that @xmath314 is a knot whenever at least one of @xmath329 and @xmath330 is even .
otherwise , if @xmath329 and @xmath330 are both odd , then @xmath314 is a link with two components . by manipulating the knot diagram
, one can easily show that @xmath331 and that the @xmath314 has mirror image given by @xmath332 . noting that for any two - bridge knot @xmath4 , the continued fraction expansions of the mirror image @xmath153 are precisely the negatives of the continued fraction expansions of @xmath4 , we see from the boundary slope formula of hatcher
thurston @xcite that the boundary slopes of the mirror image @xmath153 are the negatives of the boundary slopes of @xmath4 and from the weight formula of ohtsuki @xcite that the weights of the corresponding boundary slopes for @xmath4 and its mirror image @xmath153 are identical .
we therefore restrict our attention to knots @xmath333 with @xmath329 and @xmath330 positive .
we consider separately the two cases @xmath334 and @xmath314 , where @xmath335 .
using the continued fraction expansion @xmath337 $ ] , we determine the rational number @xmath338 setting @xmath339 and @xmath340 , we see that the double twist knot @xmath341 coincides with the two - bridge knot @xmath146 . in the special case @xmath342 , the knots @xmath343 and @xmath344 specialize to the @xmath345-torus knots and the @xmath330-twist knots , respectively .
when both @xmath346 , the continued fraction expansion @xmath347 $ ] corresponds to an incompressible , @xmath273-incompressible surface in the complement of the knot .
an arbitrary two - bridge knot @xmath136 is a double twist knot of this form if @xmath348 for some positive integer @xmath349 and its mirror image , which is the two - bridge knot @xmath350 , is a double twist knot of this form if @xmath351 for some positive integer @xmath352 to calculate the culler shalen seminorms of @xmath334 , we first determine all possible continued fraction expansions @xmath353 $ ] for @xmath212 and for @xmath354 with @xmath355 . for @xmath212 , because @xmath356 , we see that @xmath357 must be either @xmath329 or @xmath358 . if @xmath359 the only possibility is @xmath347 $ ] . otherwise , if @xmath360 then the condition that @xmath361 implies that the only possibility is the alternating expansion @xmath362 $ ] of length @xmath330 . because @xmath363 , any continued fraction expansion @xmath353 $ ] for @xmath364 with @xmath355 must begin with @xmath365 and the only possibility is the alternating expansion @xmath366 $ ] of length @xmath329 .
let @xmath2 denote the continued fraction expansion @xmath347 $ ] , @xmath367 the continued fraction expansion @xmath368 $ ] of length @xmath330 , and @xmath369 the continued fraction expansion @xmath370 $ ] of length @xmath329 . for @xmath371 or @xmath372
, because the expansion @xmath373 $ ] has entries with absolute value less than 2 , it does not correspond to a boundary slope .
recall that the seifert expansion is the unique continued fraction expansion @xmath353 $ ] with each @xmath374 even for @xmath375 which of @xmath376 or @xmath369 is the seifert expansion depends on the parities of @xmath329 and @xmath330 .
since @xmath4 is a knot , at most one of @xmath329 or @xmath330 is odd , and one sees that the seifert expansion is @xmath2 if both @xmath329 and @xmath330 are even , @xmath367 if @xmath329 is odd , and @xmath369 if @xmath330 is odd
. we now use equation to calculate the boundary slopes for @xmath377 and @xmath369 . for the continued fraction expansions obtained above , each of @xmath278 and @xmath281 can be computed directly , and we obtain that @xmath378 for @xmath2 , @xmath379 for @xmath367 , and @xmath380 for @xmath369 . the other two terms @xmath381 and @xmath382 depend on the parities of @xmath329 and @xmath330 and the difference @xmath383 will equal @xmath384 for @xmath385 , or @xmath369 depending on whether @xmath329 and @xmath330 are even , or @xmath329 is odd , or @xmath330 is odd , respectively .
the resulting values of the boundary slopes @xmath386 for each of @xmath385 and @xmath369 in these three separate cases are summarized in table [ boundaryslope ] .
_ acknowledgements .
_ both authors would like to thank steve boyer for several helpful discussions .
we also appreciate the careful reading and thoughtful suggestions of the referee .
the first author would also like to thank the max planck institute for mathematics in bonn for its support .
|
we investigate the behavior of the @xmath0 casson invariant for 3-manifolds obtained by dehn surgery along two - bridge knots . using the results of hatcher and thurston , and also results of ohtsuki
, we outline how to compute the culler
shalen seminorms , and we illustrate this approach by providing explicit computations for double twist knots .
we then apply the surgery formula to deduce the @xmath0 casson invariant for the 3-manifolds obtained by @xmath1dehn surgery on such knots .
these results are applied to prove nontriviality of the @xmath0 casson invariant for nearly all 3-manifolds obtained by nontrivial dehn surgery on a hyperbolic two - bridge knot .
we relate the formulas derived to degrees of @xmath2-polynomials and use this information to identify factors of higher multiplicity in the @xmath3-polynomial , which is the @xmath2-polynomial with multiplicities as defined by boyer - zhang .
[ multiblock footnote omitted ]
|
ever since their discovery , chiral molecular systems puzzled and inspired researches .
the mirror symmetry , characterizing two molecular enatiomers , is possibly the simplest broken symmetry to think of .
the complexity and variety of interactions and occurrences in nature of different enatiomers , flowing from such a small difference in their structure , is amazing .
chiral nature of living matter poses both fundamental and practical questions , from the origins of homochirality in biomolecules @xcite to detection and manipulation @xcite of chiral properties .
the oldest method used to detect chirality in such media is optical rotation : the polarization plane of a linearly polarized light propagating through left and right chiral media rotates in opposite directions ( see e.g. optical rotation measurement in epoxypropane molecule ; we shall use this molecule to illustrate our concepts here @xcite ) , microscopically , optical rotation is governed by the interplay of dipole transitions in a chiral medium caused by electric and magnetic fields of a light wave .
while the electric transition dipoles are identical in the two enantiomers in their respective molecular frames , the magnetic dipoles are pointing in opposite directions leading to opposite dynamical electronic response in the two enantiomers .
the same mechanism underlies other linear chiroptical techniques which use circularly polarized light to detect differences in absorption ( circular dichroism ) , raman scattering , or circular fluorescence @xcite between the two enantiomers .
it is useful to note that while chirality is a topological property of molecular structure , chiral response to light is caused by chiral electronic dynamics , which thus effectively measures this structural property .
weak magnetic effects lead to weak chiroptical response .
for example , circular dichroism is often three to six orders of magnitude smaller than light absorbance at the same wavelength .
weakness of chiral response poses challenges for time - resolved measurements of chiral dynamics .
one way of increasing chiral response is to employ techniques that do not rely on weak interaction with the magnetic field component of the electromagnetic wave , such as e.g. photoelectron spectroscopy @xcite , microwave detection @xcite , or coulomb explosion imaging @xcite .
two of these techniques , one - photon @xcite and multiphoton @xcite photoelectron circular dichroism ( pecd ) are pertinent to the approaches to chiral discrimination involving high harmonic generation ( chhg method @xcite ) .
in particular , one - photon ionization is the inverse of photorecombination , which is the key step in high harmonic generation .
the pecd detection requires resolving the direction of the final electron momentum : the photoelectron angular distribution obtained from randomly oriented chiral molecules shows asymmetry with respect to the direction of light propagation .
this asymmetry originates from the electron interaction with the chiral potential of the core @xcite , i.e. is the dynamical consequence of the chiral structure .
the effect is most prominent for low energy electrons @xcite , since interaction with the chiral core potential is maximized in this case .
the chiral dichroism signal achievable with pecd , up to some 10% , sets the golden standard for the chhg method @xcite . with the advent of ultrafast light sources
, there has been an increasing interest in developing ultrafast chiroptical spectroscopic techniques to investigate chiral response in the time domain . in the condensed phase ,
several breakthroughs have been achieved @xcite using non - linear chiroptical spectroscopy @xcite . in the gas phase , where isolated compounds can be studied ,
such measurements are a lot more challenging due to low density of samples , so that no ultrafast measurement at @xmath2 fsec time scale has yet been reported
. high sensitivity of pecd may open a path towards such measurements , provided that ultrashort circularly polarized laser @xcite or xuv pulses @xcite are used .
importantly , chhg naturally offers sub - femtosecond temporal resolution and allows one to detect multielectron chiral dynamics on their natural attosecond timescale .
the chiral response can be monitored with @xmath3 0.1 fsec resolution @xcite , orders of magnitude better than achieved so far ( see e.g. @xcite ) . moreover ,
already the first experiments @xcite demonstrated very high enantio - sensitivity of the chhg approach , leading to about 2 - 3% chiral signal for nearly linearly polarized laser fields , with ellipticity @xmath4
. thus , chhg may represent an exciting alternative to pecd for ultrafast studies in the gas phase .
here we discuss the opportunities for extending the first chhg measurements @xcite to increase their sensitivity and versatility , striving to turn chhg into a standard ultrafast chiroptical detection technique .
high - harmonics generation ( hhg ) can be understood as a sequence of three steps @xcite : ionization in a strong laser field , laser - induced acceleration of the liberated electron , and its recombination with the parent ion , all within the same laser cycle .
high harmonic generation spectroscopy can be viewed as a pump - probe technique .
ionization acts as a pump , simultaneously launching the continuum electron and the correlated multielectron dynamics in the cation .
recombination acts as a probe of such dynamics @xcite .
the pump - probe delay is controlled by light oscillation with attosecond precision ; the ionization and emission times can be accurately characterised @xcite .
effectively , the emitted light records a movie of the recombining system , with each harmonic representing a single frame @xcite .
the ability to record multielectron dynamics on attosecond time - scale @xcite is one of the most exciting features of high harmonic spectroscopy .
this multielectron dynamics can be visualised as the motion of a hole created in a molecule upon ionization .
the shape , location , and momentum of the hole during recombination are mapped onto the properties of the emitted light - its amplitudes , phases , and polarizations @xcite .
this link is crucial for chhg .
indeed , one of the effective mechanisms of chiral sensitivity of hhg is directly linked to attosecond laser - driven dynamics of the hole between ionization and recombination .
equivalently , one can think about laser - driven multi - electron dynamics in the molecular ion , i.e. light - induced transitions between its electronic states .
chiral sensitivity arises due to the interference of two types of transitions between the electronic states of the cation @xcite : the electric dipole transitions driven by the minor component of the electric field , and the magnetic dipole transitions driven by the major component of the magnetic field .
this inteplay of the magnetic dipole and electric dipole transitions occurs on the sub - cycle time - scale .
the enantio - sensitive sub - cycle dynamics reflects the molecular structure , which dictates the strength of the strong - field driven chiral response .
note that this dynamics occurs with the nuclei still located at their positions in the neutral molecule : there is very little time for the nuclei to move during the fraction of the laser cycle between ionization and recombination .
optimal regime for the interference of the magentic dipole transition driven by the major component of the magnetic field and the electric dipole transition driven by the minor component of the electric field requires balance between the magnitudes of the respective interactions , @xmath5 . here
@xmath6 are the amplitudes of the electric and magnetic fields of the laser pulse , @xmath7 is the matrix element of the magnetic dipole transition , @xmath8 is the matrix element of the electric dipole transition , @xmath9 is the ellipticity of the laser field , @xmath10 is the amplitude of the minor component of the electric field . since @xmath11 ,
the balance is achieved for @xmath12 , i.e. for nearly linear fields .
since @xmath13 changes direction by @xmath14 between the two enantiomers of the chiral molecule , constructive interference requires opposite values of helicity for right and left enantiomers , giving rise to chiral dichroism .
chiral dichroism at a few percent level for such almost linear fields observed in ref .
@xcite demonstrates high potential for chiral discrimination of this new technique .
however , as with any new technique , its first implementation may not be optimal . the observed chhg signal @xcite maximises for very low ellipticities @xmath4 .
on the one hand , the chiral response typically increases with increasing ellipticity , likely maximizing for circular rather than nearly linear fields . on the other hand ,
the harmonic signal drops quickly with increasing ellipticity .
these conflicting requirements seem to preclude exploration of what could be a more favourable regime .
resolving this conflict should enhance sensitivity and flexibility of the chhg technique .
tailoring driving fields to maximize the chiral response and optimize the chhg is a clear direction forward .
specifically , one should aim to ( i ) maximise effects of the chiral hole dynamics in the molecular ion between ionization and recombination , without sacrificing high harmonic generation efficiency ; ( ii ) explore application of other tools of high harmonic spectroscopy to chiral discrimination , complementing measurements of the harmonic spectra with measurements of the harmonic phase @xcite .
below , we show how both aims can be achieved using tailored multicolor fields .
the use of multicolor , shaped fields for chhg should allow one to achieve two important goals .
first , additional fields can couple ionic states populated during ionization to the ionic states not directly populated during the ionization step , but characterized by strong magnetic dipole transitions , for example due to the proximity of their respective orbitals to a chiral center .
this would make the chhg method truly general , applicable to many chiral molecules , even if the ionization step does not directly populate electronic states with strong magneto - dipole transitions .
second , circular light has long been used to detect chiral response because such light has high chirality .
however , the high harmonic yield maximises for linear fields and quickly drops with increasing @xmath9 .
application of chiral multi - color fields resolves this problem .
the family of counter - rotating two - color fields , introduced in @xcite , can combine circular or elliptic polarization with strong high harmonic response @xcite . here
we consider the application of elliptic counter - rotating two - color laser fields , @xmath15 to induce , enhance and manipulate the chiral hhg response from randomly oriented gas of chiral molecules . in eq.(1 )
@xmath16 is the field strength , @xmath17 , @xmath18 are the ellipticities of the fundamental field and its second harmonic , @xmath19 are the unit vectors in the poarization ( x , y ) plane of the laser field .
below , we theoretically show that application of such fields should allow one ( i ) to improve the sensitivity of the chhg method to rival the sensitivity of the pecd measurements , without relying on using the driving light with very small ellipticity , ( ii ) to increase the range of harmonics for which the chiral signal is strong .
we also show how phase measurements of the chiral high harmonic signal allow one to determine the amount of specific enatiomers ( enantiomeric excess ) in a mixture , in a way that does not require one to know the medium concentration
. finally , we present the theoretical background for the reconstruction of time - dependent chiral response for arbitrary ellipticities and configurations of the driving laser fields .
in this section , we apply the approach we have already used in @xcite to estimate the high harmonic signal vs ellipticity of the field given by eq .
( [ field ] ) .
we set the ellipticities of both fields to be equal , @xmath20 .
we will consider an epoxypropane molecule ch@xmath0chch@xmath21o ( also known as 1,2-propylene oxide ) , which has been used in the first chhg experiments @xcite .
there are two key chhg channels in strong field ionization of epoxypropane .
these channels are associated with leaving the ion either in the ground electronic state @xmath22 upon ionization , or in the first excited state electronic state @xmath23 . in the former case
, the returning electron can recombine with the ion in state @xmath22 , in the latter case , with the ion in state @xmath23 .
higher excited electronic states are well separated in energy from these two , and do not contribute to ionization in strong mid - infrared laser fields .
recombination of the continuum electron with the parent ion in these states gives raise to direct hhg channels , @xmath24 and @xmath25 . here
the first letter marks the ionization channel and the last marks the recombination channel .
laser - induced transitions between the states @xmath22 and @xmath23 of the ion open additional cross - channels @xcite : @xmath26 and @xmath27 . for example , strong field ionization can leave the ion in the state @xmath22 , but laser - induced transitions can excite the ion to the state @xmath23 .
the returning electron now has to recombine with this state : the hole created in the molecule during ionization is different from the hole before recombination .
this results in the @xmath26 channel in hhg .
importantly , @xmath26 and @xmath27 are chiral sensitive since they involve the interference of the laser - driven electric dipole and magnetic dipole transitions @xcite .
our calculations @xcite show that ionization and recombination create strongly preferred direction for the @xmath28 channel , while the ionization - recombination angular dependence for the channel @xmath22 has no such preferred direction . also , the cross - channel @xmath26 associated with the ionization to the channel @xmath29 followed by recombination to the channel @xmath28 is suppressed relative to the channel @xmath27 . along the direction of maximal ionization for the channel @xmath28 , the strength of the channel @xmath26
is @xmath30 compared to the channel @xmath27 .
these considerations allow us to develop a very simple model of the overall process @xcite : @xmath31e^{-ie _ { a x}\tau(n)}+r _ { x x}\right ] \label{sfa : model0}\end{aligned}\ ] ] the symbol @xmath32 denotes n - th harmonic component of the laser - induced polarization associated with channel @xmath33 , i.e. ionization that leaves the ion in the state @xmath34 and recombination with the ion in the state @xmath35 .
the overline denotes orientation - averaged quantities .
the complex - valued relative amplitude of the cross - channel , @xmath36 , is given by orientation - averaged amplitude of the laser - induced transition from the state @xmath23 to the state @xmath22 .
this amplitude takes into account all laser - driven transitions , including the transitions driven by the magnetic field component due to the strong magnetic dipole which couples these states .
the cross - channel @xmath37 is maximized in the direction given by the solid angle @xmath38 that optimizes ionization into @xmath23 and recombination into @xmath22 in the molecular frame . the contribution of the cross - channel is still averaged over all angles in the plane orthogonal to this direction . in the last equality in eq.([sfa : model0 ] ) , we have written down explicitly the relative phase evolution between the channels @xmath24 and @xmath25 , with @xmath39 being the time - delay between ionization and recombination and @xmath40 the difference in the energies between the ionic states @xmath23 and @xmath22 .
relevant molecular properties including transition matrix elements are described in detail in @xcite , where we have performed calculations for the single - color elliptical field .
here we use the same input and strategy and only change the single - color elliptical field to the field given by eq .
according to the calculations of the ionization and recombination amplitudes @xcite , the relative contribution of the @xmath41 channel , @xmath42 , is close to unity .
we have established @xcite that in the single color elliptical field , due to the weakness of the chiral sensitive cross - channels ( @xmath26 , @xmath27 ) , these channels can only leave their fingerprint in the high harmonic spectrum near a specific harmonic order , for which the main channels @xmath24 and @xmath25 interfere destructively . the phase term @xmath40 in eq.([sfa : model0 ] ) controls this interference .
following the model in eq.([sfa : model0 ] ) , to estimate chiral dichroism in the two - color elliptically polarized counter - rotating fields , we need to know the ionization @xmath43 and recombination @xmath44 times for each harmonic order @xmath45 .
this allows one to calculate the chiral cross - channel : the amplitude of populating the state @xmath22 at the time @xmath44 after the ion was created in the state @xmath23 at the time @xmath43 .
these times are the well - known solutions of the saddle point equations for high harmonic generation in such two - color fields @xcite . using eq.([sfa : model0 ] )
we can now evaluate the normalized high harmonic signal vs ellipticity @xmath20 ( see eq.(1 ) ) .
1 ( left panel ) shows the calculated chiral dichroism as a function of @xmath46 , for @xmath16=0.0239 a.u .
( i=@xmath47w/@xmath48 ) and the carrier frequency @xmath49=0.0224 a.u .
( @xmath50 nm ) .
high values of chiral dichroism are observed in a wide range of harmonic numbers , not only in the region of the destructive interference of the @xmath24 and @xmath25 channels , which is around harmonic 43 .
stronger chiral dichroism in this region ( about 80 @xmath51 ) as compared to results obtained in weakly elliptical fields ( about 3 @xmath52 ) @xcite is due to the order of magnitude higher ellipticity ( @xmath53 ) .
there are two factors leading to this increase for relatively small ellipticities .
firstly , in the range @xmath54 , chiral dichroism is proportional to ellipticity : increasing ellipticity 14 times will respectively increase the chiral dichroism .
secondly , for the field used here , the amplitude of the transition between the states @xmath23 and @xmath22 is almost two times higher for @xmath53 then for the single - color field of the same ellipticity . for higher ellipticities , the disparity between the laser - induced transitions between left and right enantiomers increases and the destructive interference is no longer the decisive mechanism in revealing chiral effects . for example ,
chiral dichroism at the level of about 20% is found for harmonics 35 - 51 and ellipticity @xmath55 .
such high values of chiral dichroism are on par with values achievable with pecd method .
the specific value of chiral dichroism for large ellipticities depends on many additional factors .
one such factor is the achiral background .
achiral contributions to the harmonic signal are associated @xcite with the diagonal hhg channels , involving the cation staying in the same electronic state between ionization and recombination .
we found that importance of these channels in chhg can be controlled via the stark shift in the two - color field eq.(1 ) .
specifically , most chiral molecules are polar and have permanent dipoles in the cationic states .
laser interaction with the permanent dipole induces the linear stark shift . for intense laser fields , these shifts easily approach the @xmath56 scale .
such shifts play important role in laser - induced dynamics in the cation between ionization and recombination .
therefore , they can be used to control chhg by shaping the laser field in the plane of polarization .
this is precisely what is achieved by varying @xmath20 in eq.(1 ) .
indeed , for a given electronic state of the cation @xmath35 with a permanent dipole @xmath57 , the additional phase accumulated in this state due to its interaction with the laser electric field is : @xmath58 here @xmath59 describe the sub - cycle temporal structure of the electric field in the @xmath60 polarization plane , @xmath61 and @xmath62 are the corresponding components of the permanent dipole .
the amplitude @xmath63 of the achiral contribution to the high harmonic field , associated with this state , acquires the corresponding additional phase factor , @xmath64 . in strong mid - ir fields @xmath65
thus , the phase term @xmath66 will affect the outcome of coherent averaging of the harmonic response over all molecular orientations , in every channel . changing @xmath9 in eq.(3 )
changes @xmath67 $ ] , which affects coherent averaging over molecular orientations .
1 ( right panel ) demonstrates the linear stark - based ellipticity control of the amplitude of the achiral channel @xmath25 in epoxypropane , using the same theoretical approach as in ref.@xcite .
clearly , we can manipulate the amplitude of this channel within large dynamic range , approaching one order of magnitude in amplitude for the highest harmonics .
the channel @xmath24 in epoxypropane shows similar ( but not identical ) behaviour .
/[y_s(n,\epsilon)+y_r(n,\epsilon)]$ ] , where @xmath68 is the harmonic yield for left ( s ) and right ( r ) molecules .
the color scale is in percentage .
right panel : the amplitude of @xmath25 channel , coherently averaged in the direction orthogonal to the direction of maximal ionization of the epoxypropane .
bottom panel : relative phase ( between the left and right enantiomer ) in radians for the chiral sensitive channel ax in epoxypropane , without the contribution of the achiral channels .
all calculations were done for electric field amplitude @xmath16=0.0239 a.u .
( @xmath69v / cm ) , fundamental frequency @xmath49=0.0224 a. u. ( @xmath50 nm ) , @xmath20 . , title="fig : " ] /[y_s(n,\epsilon)+y_r(n,\epsilon)]$ ] , where @xmath68 is the harmonic yield for left ( s ) and right ( r ) molecules .
the color scale is in percentage .
right panel : the amplitude of @xmath25 channel , coherently averaged in the direction orthogonal to the direction of maximal ionization of the epoxypropane .
bottom panel : relative phase ( between the left and right enantiomer ) in radians for the chiral sensitive channel ax in epoxypropane , without the contribution of the achiral channels .
all calculations were done for electric field amplitude @xmath16=0.0239 a.u .
( @xmath69v / cm ) , fundamental frequency @xmath49=0.0224 a. u. ( @xmath50 nm ) , @xmath20 . , title="fig : " ] /[y_s(n,\epsilon)+y_r(n,\epsilon)]$ ] , where @xmath68 is the harmonic yield for left ( s ) and right ( r ) molecules .
the color scale is in percentage .
right panel : the amplitude of @xmath25 channel , coherently averaged in the direction orthogonal to the direction of maximal ionization of the epoxypropane .
bottom panel : relative phase ( between the left and right enantiomer ) in radians for the chiral sensitive channel ax in epoxypropane , without the contribution of the achiral channels .
all calculations were done for electric field amplitude @xmath16=0.0239 a.u .
( @xmath69v / cm ) , fundamental frequency @xmath49=0.0224 a. u. ( @xmath50 nm ) , @xmath20 .
, title="fig : " ]
high harmonic spectroscopy gives access to many observables , including harmonic phases @xcite .
chiral sensitivity of the harmonic phase has not been explored so far .
1 ( bottom panel ) shows the relative phase between the chiral - sensitive transition amplitudes in the cation , for left and right enantiomers .
this relative phase is substantial , up to one radian , well within the accuracy of the experimental phase measurements @xcite .
thus , chhg can ( and should ) also aim at using phase measurements to detect enantiomers . in general
, achiral channels may hide this phase .
however , here we are assisted by the suppression of achiral background shown in fig.1 ( right panel ) , which indicates that the enantio - senitive harmonic phase could be observable for large ellipticities of the driving two - color field , where suppression of the achiral background is strong .
the combination of the harmonic phase measurement with the harmonic spectral measurement brings the new opportunity : _ concentration - independent _ detection of enantiomeric excess in a mixture @xmath70 of left @xmath71 and right @xmath72 chiral molecules ( see fig.2(a ) ) .
right - handed and @xmath73 left - handed molecules .
the goal is to find @xmath74 ( b ) schematic representation of the harmonic response from three gas cells with pure left - handed molecules , pure right - handed molecules , and the unknown mixture .
the overall number density of the mixture is @xmath75 , the relative fractions of the left - handed and right - handed molecules are @xmath76 and @xmath77 , @xmath78 are the polarizations of the media , @xmath79 are the harmonic dipoles .
( c ) in a mixture , the harmonic dipoles from the left - handed and right - handed molecules add coherently , therefore the relative phase between them is crucial for determining @xmath80 and the enatiomeric excess . , scaledwidth=100.0% ] polarization of the medium is proportional to number density . therefore , accurate determination of enantiomeric excess in chiroptical spectroscopies requires precise determination of number density , which can be very challenging . in the following
we show how one can perform the measurement of enantiomeric excess without any knowledge about number densities of the mixture and of pure samples .
suppose we have a mixture of chiral molecules with an overall number density @xmath75 and the relative fractions of the left - handed and right - handed molecules @xmath76 and @xmath77 . our goal is to find @xmath80 , which maps into the enantiomeric excess ( @xmath81 ) as @xmath82 .
the first step is to measure the harmonic yield from the unknown mixture , @xmath83 .
ideally , this should be done for all @xmath9 , changing it from @xmath84 to @xmath85 and then reversing the helicity ( denoted as negative @xmath9 ) . at the very least
, one should measure the yield for , say , @xmath85 and @xmath86 , i.e. perform two spectral measurements .
next , we use these two measurements to obtain the number density - independent quantity such as the elliptic dichroism , @xmath87 or the ratio of the two harmonic intensities @xmath88 we can then compare these measurements to the harmonic yields @xmath89 from the samples of pure s ( left - handed ) or r ( right - handed ) molecules , also made number density - independent by measuring the corresponding elliptic dichroisms @xmath90 and/or finding the number density - independent ratios @xmath91 for definitiveness , we assume that @xmath92 , i.e. for the r molecules the harmonic yield for @xmath9 is higher then for @xmath93 .
suppose we find that for the pure r - sample and , say , @xmath94 , the corresponding elliptic dichroism is @xmath95 , and for our mixture we have measured @xmath96 .
we also know that for the racemic mixture the dichroism is zero .
can we conclude that our mixture is exactly half - way between pure and racemic ?
fig.2 shows why we can not make such conclusion : the harmonic dipoles from s and r molecules add coherently in the mixture .
thus , the total signal also depends on the relative phase @xmath97 of the harmonic emission between the left - handed and right - handed molecules the angle between the two harmonic dipoles in fig.2 ( c ) .
this phase can be measured using the two - slit harmonic phase measurement technique developed in @xcite . in this case
, the measurement should use far - field interference of the harmonic light from two gas cells with enantiomerically pure @xmath71 and @xmath72 molecules .
the shift of the intereference fringes relative to the case of two identical cells determines the relative phase @xmath97 in a concentration - independent way , with relative concentrations determining fringe contrast but not fringe positions .
once we have measured @xmath97 , we can use it together with the measured @xmath98 and @xmath99 to find @xmath80 .
simple algebra following addition of the harmonic dipoles shows that @xmath100 here we have also taken into account that latexmath:[$|d_s(n,-\epsilon)|= varying @xmath80 between @xmath102 ( pure r - sample ) and @xmath103 ( pure s - sample ) , we see that @xmath104 is squeezed between @xmath92 and @xmath105 , i.e. , @xmath106 . next , we re - write the above equation as a quadratic equation for @xmath107 and , taking into account that @xmath106 , find its positive solution : @xmath108 where all quantities depend on the harmonic number , @xmath45 , and the ellipticity @xmath9 .
clearly , the main challenge in this approach stems from the experimental signal to noise ratio , which may result in using small quantities with large error bars .
this challenge can be addressed by taking advantage of the huge redundancy of the two - dimensional chhg measurement , performed for various harmonics n as a function of @xmath9 .
indeed , every observable used in eq .
( [ alpha5 ] ) depends on two parameters : @xmath109 .
therefore , the application of eq .
( [ alpha5 ] ) to every harmonic and every ellipticity should yield exactly the same value of @xmath110 .
the above scheme includes a set of reference measurements performed on pure samples .
their outcome can be used to characterize multiple samples containing unknown mixtures of left and right molecules , under the same experimental conditions .
the additional measurements , specific to the mixture , are the _ hhg intensity _ measurements yielding elliptical dichroism or @xmath111 for the unknown sample . crucially ,
identification of @xmath107 using eq .
( [ alpha5 ] ) is concentration - independent : one does not need to know the overall number density of molecules in the pure r - sample , and one does not need to know the overall number density of molecules in the mixture . the above measurement scheme for detecting enatiomeric
excess is not unique .
one can employ two additional phase measurements , specific for the mixture , measuring the relative phase between harmonics emitted by the left - handed molecules and the mixture and the relative phase between the harmonics from the right - handed molecules and the mixture .
we now consider such a scheme .
again , we prepare 3 gas cells : with enantiomerically pure left molecules @xmath71 , right @xmath72 molecules , and the unknown mixture @xmath70 . using the two - slit technique ,
we now measure three relative phases : ( i ) the relative phase between the harmonics generated from the left - handed and the right - handed samples @xmath97 , ( ii ) between the left - handed molecules and the mixture @xmath112 , and ( iii ) between the right - handed molecules and the mixture , @xmath113 .
we now define number density - independent signals by normalizing the harmonic fields at @xmath114 to their values at @xmath84 , @xmath115 next , we note that @xmath116 since @xmath117 and @xmath118 . here
@xmath119 are the normalized values of harmonic fields as defined in eqs .
( [ alpha00],[alpha01 ] ) .
we now show how @xmath120 can be extracted from the measurements described above .
writing @xmath121=\phi_{mr}$ ] in two steps .
first , @xmath122d^{*}_r = m_1d_sd^{*}_r+m_2|d_r|^2 . \label{alpha1}\end{aligned}\ ] ] second , for the quantity @xmath123 in the above equation we write @xmath124+m_2|d_r|^2 , \label{alpha2}\end{aligned}\ ] ] using these two expressions , we finally obtain @xmath125 analogously , we obtain for @xmath112 : @xmath126 introducing the ratio of the spectral amplitudes of the harmonic signals from left and right molecules @xmath127 and @xmath128 , which can be obtained from the phase measurements , we can express @xmath120 from eqs .
( [ alpha3],[alpha4 ] ) : @xmath129 again , we note that every observable in eq . ( [ alpha05 ] )
depends on @xmath45 and @xmath9 . therefore , the application of eq .
( [ alpha05 ] ) to every harmonic and every ellipticity should yield exactly the same value of @xmath110 , thus reducing the signal to noise ratio .
we remind the reader that in this scheme , the independence of the measurement on concentration ( number density ) arises from normalizing the harmonic response to its value at @xmath84 , see eqs.([alpha00],[alpha01 ] ) . from now on , all signals will be understood as normalized to their values at @xmath84 , thus explicitly removing their dependence on molecular concentration ( number density ) . note that in the second approach we perform redundant spectral measurements by measuring harmonic response for both s- and r - molecules while varying ellipticity in the full range , @xmath130 , increasing the redundancy of the first scheme by factor two .
as established in ref.@xcite , the main difference between the chhg signals coming from the two enantiomers is due to the small magnetic - dipole response in the cation .
thus , the multi - electron wavepacket in the cation acquires small correction , which is linear in @xmath131 ( a straightforward result of the first order perturbation theory ) : @xmath132 here @xmath133 is the time - dependent wavefunction describing the dynamics in the cation in the dipole approximation , @xmath134 is the chiral correction of the time - dependent response , opposite for s and r molecules .
this small correction to the multi - electron dynamics in the cation yields chiral correction to the harmonic field @xmath135 .
\label{eq : dysonhhg22}\end{aligned}\ ] ] here we have used the number density - independent signals , which can be obtained by normalizing the harmonic fields at @xmath114 to their values for @xmath84 , @xmath136 here @xmath137 is the normalized harmonic field in the dipole approximation .
the first - order in @xmath7 expression is general .
its only assumption is the linear response of the cation to the weak effect of the magnetic field .
the dipole @xmath137 is not chiral sensitive ( once averaged over randomly oriented ensemble ) , @xmath138 .
the chiral - sensitive component is @xmath139 .
the harmonic yield for left and right enantiomers is then : @xmath140 .
\label{eq : y}\ ] ] using eq.([eq : y ] ) , one obtains for the chiral dichroism @xmath141 ( the difference between the harmonic yields @xmath142 for left and right molecules divided by the signal from the racemic mixture ) : @xmath143 here @xmath144 is the relative phase between the chiral and achiral contributions to the emission amplitude .
thus , the standard measure of the chiral signal , when applied to high harmonic yield , is proportional to @xmath145 the chiral part of hhg response .
an exciting goal is to separately reconstruct the * amplitude * @xmath146 and the relative * phase * @xmath147 of the chiral response .
this can be done by augmenting the measurement of @xmath148 with two more measurements : ( i ) of the relative harmonic phase @xmath149 between the @xmath71 and @xmath72 molecules , and ( ii ) of the harmonic amplitude in the racemic mixture @xmath150 .
eq.([eq : dysonhhg22 ] ) allows us to explicitly calculate @xmath151 : @xmath152\big [ d^{(0)}+\mu \delta d_{\rm ch } \big]^{*}\approx|d^{(0)}|^2+\mu\delta d_{\rm ch}^{*}d^{(0)}-\mu\delta d_{\rm ch}d^{(0 ) * } , \label{eq : dysonhhg2}\end{aligned}\ ] ] where we have neglected the second order terms in @xmath7 .
thus , @xmath153 since the sign of @xmath154 becomes important , we specify that @xmath155 $ ] . adding @xmath156 and @xmath157 we obtain @xmath158 where @xmath159 gives access to the relative amplitude of the chiral component :
@xmath160 the dependence of @xmath159 on @xmath150 can be further eliminated by performing the harmonic measurements in the racemic mixture ( rm ) .
suppose this measurement yields the harmonic spectrum @xmath161 ( also normalized to the signal for @xmath84 ) .
then , the amplitude of the chiral component can be obtained as follows : @xmath162 on the other hand , combining eqs.([eq : q]),([eq : w ] ) we obtain the phase between the chiral and achiral components : @xmath163 thus , both the amplitude and the relative phase of the chiral signal can be obtained from the combination of spectral and amplitude measurements .
the result described by eqs.([eq : w],[eq : ach ] , [ eq : phase ] ) is in fact more general then may appear at first glance : eq.([eq : dysonhhg22 ] ) may equally apply to chiral dynamics initiated in a neutral molecule and then probed with hhg .
this scheme is sketched in fig.3 .
such pump - probe scheme could allow one to probe the evolution of chiral component @xmath164 in neutral species during structural changes on a longer , femtosecond time scale .
this time scale is determined by the delay between the pump pulse , which excites chiral dynamics in the neutral molecule , and the probe pulse which generates the chiral - sensitive high harmonic signal .
up to now , no pump - probe time - resolved measurement of ultrafast chirality has been done in the gas phase due to the absence of sufficiently sensitive technique .
chhg promises to open this direction .
and therefore changes the high harmonic response induced by the hhg probe .
thus , hhg allows one to follow chiral response from sub fs to fs time scale .
, scaledwidth=100.0% ]
using the example of epoxypropane molecule , we have shown theoretically that chhg method should strongly benefit from the application of the two - color , counter - rotating elliptically polarized fields .
application of such fields should lead to much higher values of circular dichroism compared to the original scheme @xcite , which relied on weakly elliptical fields .
we predict that the combination of the harmonic amplitude and phase measurements should allow one to directly measure enantiomeric excess in a concentration - independent way and monitor the evolution of the chiral dynamics with sub - femtosecond time resolution .
we have shown how chhg method can be used to extract both the amplitude and phase of the chiral signal , for arbitrary configurations of laser fields used to initiate high harmonic process .
indeed , the analysis in section 3 does not assume any specific field configuration .
we predict that eqs.([eq : w],[eq : ach],[eq : phase ] ) can also be used to extract the amplitude and the phase of the chiral signal from the hhg observables in the pump - probe set - up . in this setup
chiral dynamics is first excited in a neutral molecule and then probed with chiral - sensitive high harmonic signal . from single shot to pump - probe , chhg driven by tailored laser fields thus enables imaging of a wide range of chiral dynamics , with time resolution extended from sub - femtosecond to many tens of femtoseconds .
we thank prof .
m. ivanov for fruitful discussions .
o. s. and s.p .
acknowledge the support of the dfg grant sm 292/3 - 1 .
o.s . gratefully acknowledges the support of the dfg grant 292/5 - 1 .
all authors acknowledge the support of the european cost action cm1204 xlic .
10 url # 1#1urlprefix[2][]#2 bonner w a 2000 _ chirality _ * 12 * 114126 cireasa r , boguslavskiy a e , b p , wong m c h , descamps d , petit s , ruf h , thir n , ferr a , suarez j , higuet j , schmidt b e , alharbi a f , lgar f , blanchet v , fabre b , patchkovskii s , smirnova o , mairesse y and bhardwaj v r 2015 _ nat .
phys . _ * 11 * 654 - 658 http://www.nature.com/nphys/journal/v11/n8/full/nphys3369.html mairesse y , higuet j , dudovich n , shafir d , fabre b , mvel e , constant e , patchkovskii s , walters z , ivanov m y and smirnova o 2010 _ phys . rev . lett . _
* 104 * 213601 issn 0031 - 9007 , 1079 - 7114 http://link.aps.org/doi/10.1103/physrevlett.104.213601
|
chiral discrimination with high harmonic generation ( chhg method ) has been introduced in the recent work by r. cireasa et al ( _ nat
. phys . _ * 11 * , 654 - 658 , 2015 ) . in its original implementation
, the chhg method works by detecting high harmonic emission from randomly oriented ensemble of chiral molecules driven by elliptically polarized field , as a function of ellipticity .
here we discuss future perspectives in the development of this novel method , the ways of increasing chiral dichroism using tailored laser pulses , new detection schemes involving high harmonic phase measurements , and concentration - independent approaches . using the example of the epoxypropane molecule c@xmath0h@xmath1o (
also known as 1,2-propylene oxide ) , we show theoretically that application of two - color counter - rotating elliptically polarized laser fields yields an order of magnitude enhancement of chiral dichroism compared to single color elliptical fields .
we also describe how one can introduce a new functionality to chhg : concentration - independent measurement of the enatiomeric excess in a mixture of randomly oriented left - handed and right - handed molecules .
finally , for arbitrary configurations of laser fields , we connect the observables of the chhg method to the amplitude and phase of chiral response , providing a basis for reconstructing wide range of chiral dynamics from chhg measurements , with femtosecond to sub - femtosecond temporal resolution .
|
in cite @xcite hooft and veltman developed the method of dimensional regularisation to deal with divergent integrals in quantum field theory .
the idea they had was to evaluated the corresponding integrals in @xmath4 dimensions for @xmath5 instead of the original @xmath6 dimensions .
this approach plays a key role in modern quantum field theory computations .
it is therefore a natural question whether it is possible to mathematically construct geometric spaces which have dimension @xmath7 .
as described by connes and marcolli in @xcite this is indeed possible in the framework of noncommutative geometry .
they found a spectral triple for which the ` dirac operator ' @xmath8 satisfies the following generalization of the gaussian integral in @xmath2 dimensions : @xmath9 such spectral triples can be found in the generalisation of spectral triples to semifinite spectral triples , introduced by benameur and fack in @xcite .
+ in the first section of this paper we will focus on these semifinite spectral triples , in particular we will consider the dimension spectrum @xcite of semifinite spectral triples and we will deal with products of semifinite spectral triples . combining this , one can obtain a semifinite spectral triple of ` dimension ' @xmath10 from the @xmath2-dimensional triple and a spin - manifold of dimension @xmath11
. we will construct the product of semifinite spectral triples , this is a natural generalisation of the product of ordinary spectral triples and then we will show that the product preserves finite summability and regularity .
in the second part of this paper we will give the construction of the @xmath2-dimensional semifinite spectral triples .
we will elaborate on why they can be considered as a generalisation of a geometric space of dimension @xmath2 and we will compare our construction with the definition of connes and marcolli . in particular we will pay attention to the ` dimension spectrum ' , this is a subset of @xmath12 which is a generalisation of the dimension of manifolds to spectral triples .
connes and moscovici @xcite use the operator @xmath13 in their definition of the dimension spectrum .
this causes some problems , because the dirac operator is not invertible in the triple we will consider .
an alternative definition of the ` dimension spectrum ' uses the operator @xmath14 which in this case is well - defined .
however this definition has as a disadvantage that computations are a lot more complicated .
in the last part of this paper we will show that these semifinite spectral triples can indeed be used to define integrals in @xmath2 dimensions so that we obtain a little more concrete picture of dimensional regularisation .
we will not build a full theory , but we will work out an example in details which illustrates the main ideas .
also we will show that it is possible to combine these @xmath2-dimensional spaces and the product of semifinite spectral triples to construct a tool for zeta - function regularisation .
an appendix is added in which the required results about traces on semifinite von neumann algebras are stated .
* acknowledgements .
* this paper is based on the master s thesis i wrote at the radboud university nijmegen for the completion of my master in mathematics .
i would like to thank walter van suijlekom for supervising me , the discussions we had and valuable comments he gave me during the work on my thesis and while writing down this paper .
the objective of section [ real_spaces ] is to construct spectral triples which satisfy a specific requirement so that they can be considered to be @xmath2-dimensional for @xmath15 .
the construction which is given makes use of semifinite traces and not the ordinary trace @xmath16 .
this naturally leads to the notion of semifinite spectral triples .
in this section we will derive some general results about these semifinite spectral triples , in particular we will be concerned with products of such triples .
the difference between an ordinary spectral triple and a semifinite one is that we no longer require that the resolvent of the dirac operator is compact , but we want it to be compact relative to a trace on a semifinite von neumann algebra ( c.f .
definition [ tau_compact ] ) .
[ semifinite_def]@xcite a _ semifinite spectral triple _ @xmath17 consists of a hilbert space @xmath18 a semifinite von neumann algebra @xmath19 acting on @xmath18 with a faithful normal semifinite trace @xmath20 , an involutive algebra @xmath21 and a self - adjoint operator @xmath8 affiliated to @xmath19 .
furthermore we require that for all @xmath22 the operator @xmath23 $ ] is densely defined and extends to a bounded operator on @xmath18 and that the operator @xmath8 is @xmath20-discrete .
this operator @xmath8 is called the _ dirac operator _ of the triple .
+ if in addition there exists a @xmath24-grading @xmath25 such that @xmath26 and @xmath27 for all @xmath22 .
then the tuple @xmath28 is called an _ even semifinite spectral triple_. the reason for the requirement @xmath25 is that we want that the trace @xmath29 is defined for @xmath22 .
almost all definitions of the classical case copy to the semifinite setting , in most cases we only have to deal with the substitution of @xmath16 by a trace @xmath20 .
[ summable ] suppose @xmath17 is a semifinite spectral triple . for @xmath30
we say that the triple is _
@xmath31-@xmath20-summable _ if @xmath32 .
the triple is _
@xmath20-finitely summable _ if it is @xmath31-@xmath20-summable for some @xmath30 .
the triple is _
@xmath33-@xmath20-summable _ if @xmath34 for all @xmath35 .
the triple is said to be _
@xmath36-@xmath20-summable _ if @xmath37 for any @xmath38 .
note that since @xmath8 is self - adjoint , @xmath39 and thus @xmath40 is well defined .
these different notions of @xmath20-summability are related to one another in the following way .
[ notions_summability ] suppose @xmath17 is a semifinite spectral triple and @xmath41 .
if the triple is @xmath31-@xmath20-summable , then it is @xmath33-@xmath20-summable , @xmath42-summable and @xmath36-@xmath20-summable .
suppose @xmath43 . because @xmath8 is affiliated to @xmath19 and @xmath44 is bounded , the operator @xmath45 .
thus by the following inequality holds @xmath46 and thus @xmath8 is @xmath42-@xmath20summable .
if @xmath35 , put @xmath47 from which @xmath33-@xmath20-summability follows .
+ we know that for @xmath38 and @xmath48 fixed , the function @xmath49 , @xmath50 is bounded , say by @xmath51 .
then @xmath52 so the operator is @xmath36-@xmath20-summable and the last assertion follows .
@xcite for a semifinite spectral triple @xmath53 we define for @xmath22 the operator @xmath54 $ ] , it is the unbounded derivation of @xmath55 .
we denote @xmath56 if for all @xmath22 the operators @xmath57\in\operatorname{dom}(\delta^k)$ ] , we call the triple a @xmath58_-triple _ , or @xmath58 for short . if the triple is a @xmath58-triple for all @xmath59 , we call it @xmath60 or _
regular_. an operator @xmath61 with @xmath62 for all @xmath63 is called _ smooth_.
suppose @xmath17 is a regular @xmath20-finitely summable semifinite spectral triple .
let @xmath64 be the algebra generated by the elements @xmath65 and @xmath66)$ ] for @xmath63 and @xmath22 .
this triple is said to have _ dimension spectrum _
@xmath67 if for every @xmath68 the zeta function @xmath69 ( for @xmath70 large ) , extends analytically to @xmath71 .
if the set @xmath72 is discrete , then @xmath17 is said to have a _ discrete dimension spectrum_. if all zeta functions @xmath73 have at most simple poles , the dimension spectrum is called _
there are two commonly used definitions of the dimension spectrum , namely the one introduced in @xcite and the one stated above . in @xcite the dimension spectrum is defined as the set of singularities of the the zeta functions @xmath74 however this definition gives problems if @xmath8 is not invertible , therefore we choose for the definition which uses @xmath14 instead .
the two different definitions are compatible with each other , see for instance @xcite . in @xcite the product of spectral triples is introduced . a detailed proof that this construction indeed yields a spectral triple
can be found in e.g. @xcite .
we extend connes construction to a product of semifinite spectral triples .
it is clear that we can not expect this product to be a classical spectral triple , but we do obtain a semifinite spectral triple . this is the content of theorem [ semifinite_product ] .
such a theorem extends the classical product , because a classical spectral triple is a semifinite spectral triple with the type @xmath3 von neumann algebra @xmath75 and trace @xmath16 .
we start with the preparations for the proof of this theorem .
[ takesaki_defiv1.2 ] ( * ? ? ?
* iv.2 ) let @xmath76 and @xmath77 be two hilbert spaces .
suppose @xmath77 has an orthonormal basis @xmath78 .
denote @xmath79 for a copy of @xmath76 .
then the map @xmath80 is an isometry .
note that this sum is well - defined , because if @xmath81 at most countably many elements @xmath82 can be non - zero . now clearly u is bijective and it preserves the norm .
hence it is an isometry .
[ product_semifinite_and_semifinite ] let @xmath83 and @xmath84 be semifinite von neumann algebras , then @xmath85 is a semifinite von neumann algebra .
see ( * ? ? ?
iii.2.5.27 ) . since a von neumann algebra is semifinite if and only if it admits a semifinite faithful normal trace ( * ? ? ?
v.2.15 ) , proposition [ product_semifinite_and_semifinite ] shows that given two semifinite von neumann algebras @xmath83 and @xmath84 with semifinite faithful normal traces @xmath86 and @xmath87 there exists a semifinite normal trace on @xmath85 . the trace @xmath88 indeed works .
[ trace_on_product ] suppose for @xmath89 @xmath90 is a semifinite von neumann algebra with faithful semifinite normal trace @xmath91 , then @xmath92 $ ] is a faithful semifinite normal trace .
in particular this trace factors , thus @xmath93 . from ( * ? ? ? * viii.4 ) it follows that the map @xmath88 is a faithful semifinite normal weight .
so it remains to show that @xmath20 has the trace property , i.e. @xmath94 for all @xmath95 .
let @xmath96 , then @xmath97 since @xmath91 is a trace we have @xmath98 for all @xmath99 with @xmath100 .
therefore @xmath91 extends to a linear functional on @xmath101 . and
if @xmath102 it holds hat @xmath98 . observe if @xmath103 then also @xmath104 .
so assume @xmath105 .
in that case @xmath96 with @xmath106 and @xmath107 for all @xmath11 .
hence @xmath108 so indeed @xmath88 is a trace .
[ tensor_compact_operators ] in the notation of lemma [ trace_on_product ] suppose for @xmath89 the operators @xmath109 are @xmath91-compact .
then @xmath110 is a @xmath20-compact operator .
let @xmath35 .
select for @xmath89 operators @xmath111 such that @xmath112 , with the property that for the projection @xmath113 on the range of @xmath114 the trace @xmath115 .
then @xmath116 is the projection on the range of @xmath117 . by the factorisation of @xmath20 ( lemma [ trace_on_product ] ) we have @xmath118 . by the cross - norm property of the norm on the tensor product
@xmath119 [ semifinite_product ] suppose @xmath120 is an even semifinite spectral triple and @xmath121 is a semifinite spectral triple .
then @xmath122 is a semifinite spectral triple . if in addition also @xmath123 is even with a grading @xmath124 , then @xmath125 is even with grading @xmath126 .
the triple @xmath125 is called the _ product _ of the triples @xmath127 and @xmath123 and will be denoted by @xmath128 .
if we start with two even spectral triples @xmath127 and @xmath123 , the triples @xmath128 and @xmath129 are related in the following way .
the algebras , hilbert spaces and von neumann algebras of these two triples are isomorphic and the dirac operators @xmath130 and @xmath131 are unitarily equivalent . namely @xcite for @xmath132
it holds that @xmath133 we will now prove theorem [ semifinite_product ] .
the proof of this theorem is quite lengthy since we have to check several things .
it is similar to the proof in @xcite but it involves at lot more technical difficulties in particular the self - adjointness of the dirac operator and compactness of its resolvent is difficult since we no longer have a basis of the hilbert space consisting of eigenvectors of the dirac operator .
+ for the ease of notion we introduce the following objects @xmath134 where @xmath135 , the algebraic tensor product of vector spaces .
if we have a second grading @xmath124 on @xmath123 we denote @xmath136 .
note that we are in the special situation that @xmath137 and @xmath138 anti - commute and therefore that @xmath139 .
+ it is clear that @xmath19 is a von neumann algebra acting on @xmath18 . by lemma [ product_semifinite_and_semifinite ]
@xmath19 is a semifinite von neumann algebra and by proposition [ trace_on_product ] @xmath20 is a semifinite faithful normal trace .
concerning the algebra , the tensor product of two involutive algebras is again an involutive algebra .
since @xmath140 , the inclusion @xmath21 is trivial .
if we have a grading @xmath124 on @xmath141 , then obviously @xmath142 . before we can prove self - adjointness of the dirac operator we will prove the next lemma .
[ dom(d_1)tensor1 ] the operators @xmath137 and @xmath138 with domains respectively @xmath143 and @xmath144 are self - adjoint .
note that @xmath145 and @xmath146 are algebraic tensor products of vector spaces and not tensor products of hilbert spaces because @xmath147 are not hilbert spaces . as in lemma [ takesaki_defiv1.2 ]
let @xmath78 be an orthonormal basis of @xmath77 and let @xmath148 be the isometry @xmath149 .
then @xmath150 observe @xmath151 .
hence @xmath152 suppose @xmath153 and there exists an @xmath154 such that @xmath155 .
then by self - adjointness of @xmath156 the map @xmath157 is unbounded . for an element @xmath158 with @xmath159 if @xmath160 , we have @xmath161 .
hence shows that for @xmath162 the map @xmath163 is unbounded .
thus @xmath164 .
so @xmath165 . for the converse inclusion let @xmath166 , say @xmath167 .
then for @xmath168 we have by self - adjointness of @xmath156 @xmath169 so @xmath170 , @xmath171 , is bounded by @xmath172 . hence @xmath173 .
+ similarly we have @xmath174 .
the map @xmath175 is bounded and maps @xmath144 in itself , thus @xmath176 the computation in shows @xmath177 is symmetric , and we have @xmath178 .
so @xmath137 is self - adjoint on the domain @xmath145 .
a similar argument applies to @xmath179 .
now we are able to prove that @xmath8 is self - adjoint on the domain @xmath180 .
the idea is to apply nelson s theorem ( c.f .
proposition [ nelson ] ) to the operator @xmath8 , from which we will obtain that @xmath8 is essentially self - adjoint and then we will show that @xmath8 is closed .
+ from proposition [ nelson ] and the fact that @xmath156 and @xmath181 are self - adjoint we obtain that @xmath182 dense for @xmath89 .
let @xmath183 and @xmath184 .
we will show that @xmath185 .
select @xmath186 such that for all @xmath11 @xmath187 observe that @xmath188 and that the operators @xmath189 and @xmath190 commute . hence @xmath191 similarly using @xmath192 gives @xmath193 hence @xmath194 .
since @xmath195 is a linear subspace and @xmath196 dense , we have @xmath197 dense .
but since @xmath198 then clearly @xmath199 dense .
it is also clear that @xmath8 is symmetric , so @xmath8 is essentially self - adjoint .
+ it remains to show that @xmath8 is closed on @xmath200 . to that end , let @xmath201 be a sequence in @xmath200 converging to @xmath202 such that @xmath203 is a cauchy sequence in @xmath204 .
for the moment fix @xmath205 and write @xmath206 . then @xmath207 and @xmath208 .
for ease of notation write @xmath209 and @xmath210
. then @xmath211 and thus @xmath212 by lemma [ dom(d_1)tensor1 ] the operators @xmath137 and @xmath138 are self - adjoint and hence closed on the domains @xmath145 respectively @xmath144 . the fact that @xmath203 is cauchy combined with gives that @xmath213 and @xmath214 are cauchy .
by closedness of the operators @xmath215 and @xmath216 we have @xmath217 . also the sequences @xmath213 and @xmath214 both have a limit in @xmath18 , say @xmath218 respectively @xmath219 and @xmath220 and @xmath221 .
thus @xmath222 and @xmath8 is closed . to prove that the operator @xmath8 is affiliated to the von neumann algebra @xmath19 , observe that by self - adjointness of @xmath8 it is sufficient to show that @xmath223 . by assumption @xmath224 .
also @xmath225 hence @xmath226 . combination with the commutation theorem for tensor products ( c.f .
iii.4.5.8 ) ) gives the inclusions @xmath227 and hence @xmath8 is affiliated to @xmath19 .
+ suppose @xmath228 and assume @xmath229 . then @xmath230= [ d_1,a_1]\otimes a_2 - a_1\gamma_1\otimes[d_2,a_2].\ ] ] because @xmath231 $ ] and @xmath232 $ ] are densely defined and bounded and the operators @xmath233 and @xmath234 are bounded , the commutator @xmath235 $ ] is densely defined and extends to a bounded operator .
now for an arbitrary element @xmath236 we can show boundedness of the commutant @xmath23 $ ] using linearity of @xmath237 and bilinearity of @xmath238 $ ] .
hence for all @xmath22 the operator @xmath23 $ ] is densely defined and extends to a bounded operator on @xmath18 . +
before we will prove @xmath20-discreteness of @xmath8 we will need the results proven in the following two lemmas .
[ resolvent_of_sum ] the identity @xmath239 holds as operators on @xmath18
. moreover @xmath240 is bounded .
first we make a formal manipulation of the symbols
. then we have to check that the operators involved indeed extend to the whole hilbert space @xmath18 . +
again write @xmath209 and @xmath210 .
we have @xmath241 and hence @xmath242 the operators @xmath243 , @xmath244 and @xmath245 are self - adjoint , thus @xmath246 . and hence @xmath247 , @xmath248 and @xmath249 are bounded operators on @xmath18 . furthermore @xmath250
since @xmath251 anticommutes with @xmath156 @xmath252 observe that @xmath253 .
hence @xmath254 is densely defined and thus @xmath255 is closable .
we have previously shown that @xmath256 and @xmath8 is self - adjoint on this domain .
hence the operator @xmath247 maps @xmath18 into @xmath200 and is bounded .
thus @xmath240 is a closed operator defined on @xmath18 and hence by the closed graph theorem it is a bounded operator .
and equality holds on @xmath18 .
[ product_compact ] the operator @xmath257 is @xmath20-compact .
the grading @xmath251 satisfies , by definition , @xmath258 .
this grading therefore induces a direct sum decomposition @xmath259 , where @xmath260 is the eigenspace of the eigenvalue @xmath261 of @xmath251
. then also @xmath262 now @xmath263 then the inverse of @xmath264 is given in the matrix representation with respect to @xmath259 as @xmath265 since @xmath156 anti - commutes with @xmath251 and @xmath156 is self - adjoint , in this decomposition the operator @xmath156 can be written as @xmath266 where @xmath267 .
one can check that @xmath268 is given by the matrix @xmath269 multiplication of and gives that @xmath270 is represented as @xmath271 because @xmath272 is @xmath86-compact , there exists a sequence @xmath273 of @xmath86-finite rank operators with @xmath274 .
decompose @xmath275 as a matrix with respect to @xmath259 by @xmath276 thus from the estimate @xmath277 and a similar one for the lower entries of the matrices we obtain @xmath278 since the range of each of the operators @xmath279 is contained in the range of @xmath275 ( @xmath280 ) the operators @xmath279 are @xmath86 compact .
therefore the operators represented by @xmath281 are @xmath86 compact . by assumption @xmath181
is @xmath87-discrete , thus @xmath282 and @xmath283 are @xmath87-compact . using lemma [ tensor_compact_operators ]
the operator given by is @xmath20-compact , thus @xmath268 is @xmath20-compact .
now it has become easy to prove compactness of the resolvent . since @xmath8 is self - adjoint , @xmath284 . according to theorem [ thm - tau - discrete ] it is therefore sufficient to show that @xmath247 is @xmath20-compact .
recall that the @xmath20-compact operators are an ideal in @xmath285 .
combination of this fact with lemmas [ resolvent_of_sum ] and [ product_compact ] imply that @xmath247 is @xmath20-compact .
this concludes the proof that @xmath286 is a semifinite spectral triple .
+ in the case that we have a grading @xmath124 on the second triple , for @xmath287 we have @xmath288 and @xmath289 hence @xmath290 is a grading on @xmath125 , thus the product of two even triples is again an even semifinite spectral triple . if one takes the product of two manifolds say of dimension @xmath291 and @xmath11 , the product has dimension @xmath292 .
therefore we might expect that the product of two finitely summable semifinite spectral triples is again finitely summable .
this is indeed true .
[ summability_product ] suppose for @xmath89 the tuples @xmath293 are semifinite spectral triples and the first triple is even with grading @xmath251 .
if the triples are @xmath294-@xmath91-summable , then the semifinite product spectral triple is @xmath295-@xmath20-summable .
if both spectral triples are @xmath36-@xmath91-summable , then the product spectral triple is @xmath36-@xmath20-summable . to prove the first statement ,
suppose that both triples are @xmath294-@xmath91-summable .
note that it holds that @xmath296 and similarly @xmath297 , thus @xmath298 factorisation of the trace now gives @xmath299 so the triple is @xmath295-@xmath20-summable .
if both triples are @xmath36-@xmath91-summable , then @xmath300 thus the triple is @xmath36-@xmath20-summable . in the case of spectral triples
it has been proved @xcite that the product of two regular spectral triples is again regular .
we will now prove it in a different way ( not involving pseudo - differential operators ) for semifinite spectral triples .
we will use some results due to connes : lemma [ domain_d_preserved ] and lemma [ regularity - d ] below .
those results are about spectral triples and not semifinite ones , but since there is no trace or compactness of the resolvent involved , we can use them . in the following
we will use the notation @xmath301 .
[ domain_d_preserved ] suppose @xmath8 is a self - adjoint operator and the operators @xmath302 are bounded
. then @xmath55 preserves @xmath303 and on @xmath304 the bounded extension of @xmath305 $ ] and the commutator @xmath306 coincide
. see ( * ? ? ?
* lemma 13.1 ) .
[ regularity - d ] suppose @xmath307 is a spectral triple and @xmath308 is a bounded linear operator .
denote @xmath309(1+d^2)^{-1/2}$ ] . then the following holds : 1 .
if @xmath310 and @xmath311 are bounded , then @xmath312 is bounded ; 2 .
the operators @xmath313 are bounded for all @xmath314 if and only if @xmath315 are bounded for all @xmath314 .
see ( * ? ? ?
* lemma 13.2 ) .
[ regularity_product ] suppose that for @xmath89 the tupels @xmath316 are regular semifinite spectral triples and @xmath127 is even with grading @xmath251 , then the product @xmath125 of these spectral triples is again regular .
we start by showing that @xmath317 .
recall @xmath318 furthermore using lemma [ dom(d_1)tensor1 ] one can easily show that @xmath319 , a similar statement holds for @xmath181 .
thus by commutativity of @xmath320 and @xmath321 we have @xmath322 if we now use the even case , we obtain the following result for odd powers of @xmath8 @xmath323 as a result we obtain @xmath317 .
+ recall @xmath324 . since @xmath325=[a_1,d_1]\otimes a_2 + \gamma_1a_1\otimes[a_2,d_2]$ ] , for regularity it is sufficient to show that @xmath326 is smooth
if both @xmath327 and @xmath233 are smooth ( thus @xmath328 is not necessarily an element from the algebra @xmath329 ) .
so suppose that @xmath327 and @xmath233 are smooth .
we have to prove that for all @xmath314 it holds that @xmath330 and @xmath331 are bounded . to prove the second claim we invoke lemma [ regularity - d ] , @xmath332(1+d^2)^{-1/2 } \notag\\ & = \big([d_1 ^ 2,a_1]\otimes a_2 + a_1\otimes[d_2
^ 2,a_2]\big)(1+d^2)^{-1/2}\notag\\ & = \big([d_1,a_1](d_1 ^ 2 + 1)^{-1/2}\otimes a_2\big)\ ; \big((d_1 ^ 2 + 1)^{1/2}\otimes1\big)\,(1+d^2)^{-1/2 } \notag\\ & \qquad + \big(a_1\otimes[d_2,a_2](d_2 ^ 2 + 1)^{-1/2}\big)\ ; \big(1\otimes ( d_2 ^ 2 + 1)^{1/2}\big)\,(1+d^2)^{-1/2}\notag\\ & = ( \delta_1(a_1)\otimes a_2 ) \big((d_1 ^ 2 + 1)^{1/2}\otimes1\big)\,(1+d^2)^{-1/2 } + ( a_1\otimes \delta_1(a_2))\big(1\otimes ( d_2 ^ 2 + 1)^{1/2}\big)\,(1+d^2)^{-1/2}.\label{eq - regularity_product1}\end{aligned}\ ] ] we will show that is bounded , for this we will only show that the first summand is bounded , the other one is similar .
note @xmath333 so @xmath334 to show that higher powers @xmath335 are bounded one can do the same as in .
if one expands @xmath335 , one gets a sum of products of elements of the form @xmath336 and they are all bounded .
hence by lemma [ regularity - d ] the operators @xmath335 are bounded for all @xmath314 . that the operators @xmath331 preserve the domain of @xmath8 is now a direct consequence of lemma [ domain_d_preserved ] .
we conclude that @xmath337 for all @xmath338 .
so the triple is regular .
this section is based on the work of connes and marcolli in @xcite . in this paragraph connes and marcolli
propose a definition of a class of semifinite spectral triples which can be considered as geometric spaces of dimension @xmath2 for @xmath0
. we will give a slightly different construction , derive properties of these triples and compare it to the triples of connes and marcolli .
we start by giving the definition and we will derive some general properties . in the second subsection we will show that the dimension spectrum consists of the singleton @xmath339 such that in combination with the generalised gaussian integral the triples can indeed be considered as geometric spaces of dimension @xmath2 .
we want to construct for each @xmath7 a spectral triple @xmath341 which has the following property @xmath342 this requirement comes from the gaussian integral @xmath343 the requirement may look like an simple one , but it appears to be a great constraint ( c.f . proposition [ restriction_dz ] ) .
connes and marcolli construct such a spectral triple , but in their construction one is are naturally led to the class of semifinite spectral triples , classical spectral triples are not sufficient .
namely in the construction an operator @xmath344 is needed which has the property @xmath345 which can not be satisfied by the ordinary trace @xmath16 , details on this construction can be found in @xcite .
[ def_tz ] for @xmath346 consider the following tuple @xmath347 here @xmath348 acts on @xmath349 by pointwise multiplication of functions . denote @xmath350 where @xmath351 is a normalisation constant .
now @xmath352 is given by @xmath353 on the domain @xmath354 we will use @xmath355 , @xmath352 etcetera in the rest of this paper to denote this triple and elements thereof .
[ triple_tz ] for @xmath346 the tuple @xmath355 is a semifinite spectral triple , @xmath352 has spectrum @xmath356 and the triple satisfies @xmath357 it is clear that @xmath349 is a hilbert space .
the type i von neumann algebra @xmath348 acts on @xmath349 by left - multiplication , because @xmath358 .
clearly the trace @xmath359 is faithful and semifinite . from the monotone convergence theorem of measure
theory it immediately follows that @xmath359 is normal . since @xmath360 it is obvious that @xmath361 and @xmath362=0 $ ] , which thus extends to a bounded operator on @xmath363 .
+ we will now show that @xmath364 is a self - adjoint operator .
since @xmath365 is real - valued @xmath352 is a symmetric operator , thus @xmath366 .
it remains to show the converse inclusion .
suppose @xmath367 but @xmath368
. then @xmath369 .
observe that @xmath365 is a continuous function , so @xmath365 is bounded on compact sets . in particular for each @xmath338 there exists a constant @xmath370 such that @xmath371}|\leq c_n$ ] .
put @xmath372}$ ] . then @xmath373 and @xmath374}|^2\,dx \leq c_n^4 \int_{{\mathbb{r } } } |g|^2\,dx < \infty.\ ] ] so @xmath375 .
but @xmath376}|^2\,dx \big)^{-1/2}\int_{{\mathbb{r } } }
f_z g f_z1_{[-n , n ] } \bar{g}\,dx\\ & = \big(\int_{-n}^n |f_z g|^2\,dx \big)^{1/2 } { \rightarrow}\infty \qquad \textrm{as } n{\rightarrow}\infty.\end{aligned}\ ] ] thus @xmath377 and @xmath352 is self - adjoint .
+ since @xmath352 is self - adjoint , @xmath378 .
we will show that the converse inclusion also holds .
observe that for @xmath346 the function @xmath379 is continuous , bijective and strictly increasing .
since @xmath380 , the only possible candidate for @xmath381 is given by @xmath382 .
but if @xmath383 the function @xmath384 has a zero . by continuity of @xmath365
the function @xmath385 is therefore not essentially bounded .
hence @xmath382 is an unbounded map and thus @xmath386 is not invertible , which implies that @xmath356 .
+ since @xmath352 is self - adjoint we can give a spectral decomposition of @xmath352 .
observe that the inverse of @xmath365 is given by @xmath387 define @xmath388 it is clear that @xmath389 is a spectral measure .
for an interval @xmath390 $ ] we have @xmath391 ) = f_z(a)1_{[a , b]}.\ ] ] if we approximate the identity map @xmath392 and use the above identity , we see that @xmath393 , thus @xmath389 is the spectral measure for @xmath352 . for each @xmath394 , the borel subsets of @xmath395 , it holds that @xmath396 , thus @xmath352 is affiliated with @xmath397 . also @xmath398 ) ) = \frac{1}{2}\int_{{\mathbb{r}}}1_{f_z^{-1}([-\lambda,\lambda])}\ , dx = \frac{1}{2}\int_{f_z^{-1}(-\lambda)}^{f_z^{-1}(\lambda ) } 1\,dx < \infty,\ ] ] hence by theorem [ thm - tau - discrete ] , @xmath352 is compact relative to @xmath397 .
thus @xmath355 is a semifinite spectral triple .
+ we will now show that @xmath355 satisfies the property .
for @xmath399 we have @xmath400 use the substitution @xmath401 , then @xmath402 so we obtain @xmath403 in the last line we inserted the definition of @xmath404 . hence holds .
of course the triple constructed by connes and marcolli satisfies the requirement . since the dirac operator @xmath405 is given as a function of @xmath344 and the spectral measure @xmath389 of @xmath344 is known ( c.f . ) , one can calculate @xmath406 via theorem [ integration_trace ] .
namely @xmath407 now the calculation is the same as in proposition [ triple_tz ] .
this reduction of connes and marcolli s spectral triple to the triple @xmath355 can be applied every time .
so all results we obtain also hold for their triple .
for @xmath408 the operator @xmath352 is not self - adjoint .
this is easy to see because for @xmath408 the lebesgue measure of @xmath409 is strictly positive .
hence it does not hold that @xmath410 almost everywhere .
but then @xmath411 .
so for @xmath408 the tuple @xmath355 is not a spectral triple .
in fact , as the next result shows , we can not expect that any self - adjoint operator satisfies for @xmath412 .
[ restriction_dz ] suppose @xmath413 is a semifinite von neumann algebra , with a faithful semifinite normal trace @xmath20 .
suppose @xmath414 .
if @xmath8 is a self - adjoint ( unbounded ) operator on @xmath18 affiliated with @xmath19 , then there exists a scalar @xmath399 such that @xmath415 .
let @xmath19 , @xmath20 , @xmath2 and @xmath8 be as in the assumptions of the proposition .
let @xmath399 and consider the function @xmath416 then @xmath417 $ ] , in particular @xmath418 is real valued and bounded .
thus @xmath419 is bounded and self - adjoint .
since @xmath8 is affiliated with @xmath19 it holds that @xmath420 therefore @xmath421 is well - defined and it holds that @xmath422 $ ] for all @xmath399 .
now suppose @xmath408 , then it is impossible that @xmath423 \qquad\textrm{for all } \lambda>0.\ ] ] this proves the statement for @xmath424 .
if we have @xmath425 $ ] , let @xmath426 .
then clearly @xmath427 but also @xmath428 and thus @xmath429 . if @xmath430 for all @xmath431 this would give @xmath432 which is a contradiction with . a last observation about this spectral triple .
[ d_z_grading ] the map @xmath433 , @xmath434 induces an operator on the triple @xmath355 by @xmath435 since @xmath360 , the operator @xmath436 clearly commutes with @xmath437 .
also it holds that @xmath438 clearly @xmath436 is bounded and that @xmath439 if and only if @xmath440 .
but since @xmath436 is not given by a function , it is no element of @xmath441 .
so @xmath436 is not a grading as in definition [ semifinite_def ] , but it is very similar . in this subsection
we will establish some facts about the dimension spectrum of the triple @xmath355 .
we will calculate in various ways the dimension spectrum because the function @xmath365 has a zero , which causes some problems .
[ cutoff ] connes and marcolli use the set of poles of the meromorphic extension of @xmath442 for @xmath68 as definition of the dimension spectrum . since for this triple @xmath443 it is enough to compute @xmath444 .
the operator @xmath445 is given by multiplication with the function @xmath446 .
so @xmath447 but this yields @xmath448 for all values of @xmath449 . to solve this problem an infrared cutoff is imposed , that is the integral is computed on the subset @xmath450\cup[1,\infty)$ ] instead of on @xmath395 .
then for @xmath451 we have @xmath452 this function has a meromorphic continuation to @xmath12 with a simple pole for @xmath453 .
the residue of at @xmath453 is given by @xmath454 connes and marcolli also gave an alternative to such a cutoff , namely to smoothly change the function @xmath365 such that the new function does not attain @xmath455 .
we will work out their suggestion .
the function @xmath456 $ ] given by @xmath457 is smooth and strictly decreasing .
the interval @xmath458 $ ] is compact , @xmath459 is continuous , so @xmath459 is bounded on @xmath458 $ ] .
let @xmath460 be such that @xmath461 for all @xmath462 $ ] .
define @xmath463 , @xmath464 , then * @xmath465 if @xmath466 ; * @xmath467 for all @xmath468 , thus @xmath418 is strictly increasing ; * @xmath469 .
define an operator @xmath470 on @xmath18 by @xmath471 , it is a smooth modification of @xmath352 near @xmath472 .
this operator @xmath473 is very closely related to @xmath352 , therefore we expect that the meromorphic continuation of @xmath474 has the same poles as the meromorphic continuation of , the infra - red cutoff of @xmath475 .
this is indeed the case .
[ dim_spec_ez ] the meromorphic continuation of @xmath476 is holomorphic on @xmath477 and has a simple pole at @xmath453 with residue @xmath478 from the construction of @xmath473 it follows that @xmath479 observe that @xmath418 is continuous and non - zero , so the integral @xmath480 exists for all @xmath449 and therefore does not create any singularities .
the second term @xmath481 has precisely a simple pole at @xmath453 .
therefore @xmath482 as desired .
a modification of the dirac operator is not necessary if one uses @xmath483 in the definition of the dimension spectrum instead of @xmath13 .
we will compute the dimension spectrum of @xmath355 with @xmath483 .
for this we need the machinery of hypergeometric functions . before we will compute the dimension spectrum we state the properties of the hypergeometric functions which we will use .
we will use the shorthand notation @xmath484 to denote the hypergeometric function .
the definition and properties of this function can be found in several books , for example in ( * ? ? ?
the following two identities can be found in ( * ? ? ?
@xmath485 equality holds if the following requirements are satisfied @xmath486 and @xmath487 .
+ the last property we will use is that the indefinite integral of the function @xmath488 is given by @xmath489 this will need a proof . by the binomial theorem we have @xmath490 here @xmath491 for @xmath492 is defined by @xmath493 , with @xmath494 .
taking primitives on both sides yields @xmath495 as desired .
[ dim_spec_dz ] assume @xmath0 .
then the triple @xmath355 is @xmath340-summable , regular , the dimension spectrum is simple and consists of @xmath339 and @xmath496 note that we have a factor @xmath497 difference between the residue at @xmath453 of @xmath498 and . since @xmath360 , the commutators @xmath362=[|{\slashed{d}_z}|,a]=0 $ ] for all @xmath499 ,
thus it is obvious that the triple is regular . for the dimension spectrum we have to compute the poles of @xmath500 .
we can take @xmath501 , since @xmath360 .
thus we consider the meromorphic function @xmath502 the last equality follows from the substitution @xmath503 .
note that the constant @xmath504 does not affect the location of the poles . by equation we have @xmath505 since @xmath506 , for every @xmath507 we have @xmath508 .
so @xmath509 .
therefore if @xmath510 , i.e. if @xmath511 we can apply and to obtain @xmath512 now suppose @xmath513 such that @xmath514 , then @xmath515 . on the disk @xmath516
the function @xmath517 is holomorphic .
since @xmath518 we have @xmath519 inserting in gives for @xmath514 @xmath520 by analytic continuation this holds for any @xmath521 . recall that @xmath522 , so @xmath523 observe that the function @xmath524 has simple poles for @xmath525 . note that @xmath526 is holomorphic .
thus @xmath527 has a simple pole at @xmath453 . for @xmath528
the residues of the gamma function are given by @xmath529 .
thus it follows that @xmath530 for every pole @xmath449 we have @xmath531 .
therefore @xmath532 for all @xmath533 , hence the triple is @xmath340-summable .
+ it remains to show that @xmath2 is the only pole .
recall that the transformation was only valid for @xmath510 .
we have to deal with these points in a different way .
fix @xmath534 and @xmath535 .
say @xmath536 .
let @xmath449 be close to @xmath537 , then @xmath538 we will now consider the limit @xmath539 . to prove that does not have a pole at @xmath540 we compute ( still for @xmath541 fixed ) the residue of and show that it equals @xmath455 . since we singled out @xmath291 in the infinite sum , near @xmath537
the function @xmath542 is holomorphic . to conclude that has a removable singularity at @xmath537 we compute the following residue @xmath543 indeed , this residue is independent of @xmath541 .
so the dimension spectrum of @xmath352 consists of @xmath544 .
it is interesting to compare the dimension spectrum of @xmath352 and @xmath473 , the latter operator is the operator introduced in remark [ cutoff ] . with the previous theorem
it has become easy to compute the poles of @xmath545 .
suppose @xmath0 , then the dimension spectrum of @xmath473 equals the dimension spectrum of @xmath352 . from the definition of the trace @xmath359 and the operator @xmath473 it immediately follows that @xmath546 it follows from the proof of theorem [ dim_spec_dz ] that it suffices to show that the functions @xmath547 are holomorphic on @xmath12 .
we start with the first one .
@xmath548 note @xmath549 , thus @xcite implies that @xmath550 is holomorphic on @xmath12 .
now the second one .
the function @xmath418 is smooth , strictly increasing , @xmath551 and @xmath552 . thus there
exist @xmath553 such that @xmath554 for all @xmath555 $ ] .
but from these bounds it is clear that for each @xmath513 the function @xmath556{\rightarrow}{\mathbb{c } } , \qquad x\mapsto \big(1+\rho(z)^2f(|x|)^{2/z}\big)^{-s/2}\ ] ] is bounded .
thus @xmath557 does not have any poles in @xmath12 .
in theorem [ regularity_product ] it was proved that the product of two regular semifinite spectral triples is again regular . since @xmath360 and thus @xmath355 is regular , the following corollary is immediate .
suppose @xmath125 is an even regular semifinite spectral triple .
then the product triple @xmath558 is also regular . since the product of a spectral triple with @xmath355 is regular
, one can compute its dimension spectrum .
we expect that the whole spectrum shifts over the vector @xmath2 , but we can not prove this fact in its full generality .
the difficulty lies in computing the algebra @xmath64 of the product triple , because the operator @xmath559 is difficult to work with .
however we can prove the result for the pole with the largest real part .
we will use fubini s theorem for traces c.f .
proposition [ interchange_trace_int ] .
[ dimspec_product ] suppose @xmath560 is an even finitely summable regular semifinite spectral triple with @xmath561 .
denote @xmath562 .
suppose @xmath563 , the dimension dimension spectrum of @xmath125 , such that @xmath564 and for all @xmath565 it holds @xmath566 .
then if @xmath567 , the function @xmath568 has a pole at @xmath569 and all other singularities @xmath570 of the zeta functions @xmath571 satisfy @xmath572 .
the main idea of the proof of this proposition is to write the operator @xmath573 as an elementary tensor and then use the factorisation of the trace @xmath574 .
this can be done by writing this operator as an integral of exponential functions and we will use lemma [ interchange_trace_int ] to interchange the integral and the trace .
we start with the identity @xmath575 in the following calculation we will interchange two times an integral with a trace .
we will justify those manipulations later .
for @xmath513 such that @xmath576 it holds that @xmath577 by assumption the function @xmath578 has a pole at @xmath579 , thus @xmath580 has a pole at @xmath569 .
since @xmath581 , this is the largest pole of @xmath582 , because @xmath583 only has poles at the non - positive integers .
+ suppose @xmath584 .
the following estimate shows that one does not obtain any poles in the half - plane @xmath585 .
@xmath586 which converges for @xmath449 with @xmath587 .
+ it remains to show why one can interchange the integral and trace in and we use proposition [ interchange_trace_int ] .
suppose @xmath389 is an unbounded self - adjoint operator on @xmath588 . by the uniqueness of the analytic continuation it is sufficient to prove that the switch is allowed for @xmath589 with @xmath590 .
fix such an @xmath449 .
consider @xmath591 we check the conditions of proposition [ interchange_trace_int ] .
we know that if @xmath592 are positive , then the map @xmath593 , @xmath594 is uniformly bounded .
using the functional calculus @xmath595 is norm - bounded .
+ to prove the second requirement note that @xmath596 corresponds to @xmath597 .
thus it is sufficient to show that @xmath598 is measurable .
we will prove continuity of @xmath598 .
@xmath599 which tends to @xmath455 as @xmath600 . for the last requirement note that @xmath601 then
because @xmath590 , it follows that @xmath602 and also @xmath603 .
therefore @xmath604 is uniformly bounded in @xmath605 . if we now let @xmath606 and @xmath607 then it is obvious that as a function of @xmath605 , is uniformly bounded on @xmath608 , in particular on the interval @xmath609 $ ] .
thus using the fact that @xmath610=0 $ ] and proposition [ interchange_trace_int ] we obtain that for all @xmath611 : @xmath612 taking the limit @xmath613 gives the desired equality . for , we can do the same trick , but we have to replace @xmath389 by @xmath8 and @xmath614 by @xmath20 .
in section [ real_spaces ] we described a set of spectral triples which can be considered as being @xmath2-dimensional . in this section
we will apply these spectral triples to dimensional regularisation and zeta function regularisation in quantum field theory .
we will not develop a general theory , but we will describe an example which will give a good idea how one can apply these triples in more general computations . the basic idea of regularisation and renormalisation is the following .
suppose we are given a divergent expression @xmath615 . for _ regularisation
_ one modifies @xmath615 by inserting in some way an extra parameter @xmath449 to a obtain a different expression @xmath616 .
we want that @xmath616 satisfies the following two properties : @xmath616 must be well - defined on @xmath617 ( where @xmath618 is some neighbourhood of @xmath455 ) and @xmath619 .
the parameter @xmath449 is called the _
regulator_. since @xmath619 , the expression @xmath616 will then have a pole for @xmath620 .
_ renormalisation _ of @xmath615 is to subtract this pole at @xmath620 .
so @xmath621 is then a finite quantity . in quantum field theory ,
this quantity @xmath622 is what one is interested in . in quantum field theory ,
the integrals considered are typically of the form @xmath623 this specific integral corresponds to the following feynman diagram : graph1 ( 200,60 ) it represents a particle with mass @xmath291 which propagates and self - interacts . since these integrals are divergent , one needs to regularize these integrals before one is able make sense of it . to regularize such integrals
t hooft and veltman @xcite proposed dimensional regularisation .
the idea is that if an integral @xmath615 is divergent in @xmath6 dimensions one tries to compute @xmath624 , the same integral but then in @xmath625 dimensions .
the method of t hooft and veltman is based on the following formula : @xmath626 a priori this equality is valid for @xmath627 . for non - integer values of @xmath8 this can not be proved , because their is no such thing as a ( lebesgue ) integral in @xmath8-dimensions .
but instead of proving it , the right hand side is used as a definition for the left hand side if @xmath628 . in this way one
obtains a method to integrate in @xmath8 dimensions .
we will work out an example .
[ exam_dimreg1 ] we will compute in dimension @xmath629 for @xmath564 . to start , note that @xmath630 so @xmath631 where we used the substitution @xmath632 and @xmath633 .
we therefore obtain @xmath634 although we do not have a theorem of fubini at our disposal , we just change the order of `` integration '' .
we do this , because this is a way to give a meaning to the integral in @xmath629-dimensions .
after this we insert the essential identity to obtain @xmath635 the general theory can for example be found in ( * ? ? ?
* chapter 7 ) .
what one usually does is introduce new variables ( in our example @xmath636 and @xmath431 ) , rewrite the integrand as an exponential function and then use identity .
+ we however do not need to use as a definition , but using our previous developed machinery we can explicitly compute @xmath637 and use this as a definition of an integral in @xmath2 dimensions instead .
so if we want to calculate an integral in @xmath2 dimensions , we have to replace the variable over which we integrate by the operator @xmath352 and the integral by the trace @xmath359 . via this method
we have a genuine calculation and not just a formal manipulation
. we will illustrate this with an example , we compute again but now with @xmath352 and @xmath359 .
we have @xmath638 as before we use identity , interchange integral and trace and finally we use to obtain @xmath639 note that it is valid to interchange trace and integral , because by the definition of @xmath359 and @xmath352 we have @xmath640 since @xmath346 it holds that @xmath641 for all @xmath642 .
thus by fubini interchanging the integrals is allowed and therefore we can interchange @xmath20 and the integral . +
to finish the calculation , we can copy the end of example [ exam_dimreg1 ] .
so @xmath643 which has precisely a simple pole at @xmath644 .
note that this expression is well defined for all @xmath346 and has a meromorphic extension to @xmath12 . in this subsection we will show that it is also possible to use semifinite spectral triples for zeta function regularisation @xcite . in zeta function regularisation on typically has to deal with divergent integrals of the form @xmath645 for zeta function regularisation one inserts an extra power @xmath646 to make the integral convergent and then computes the behaviour as @xmath647 . in the above case one
gets @xmath648 which has a simple pole at @xmath620 with residue @xmath649 .
note that this result is very similar to dimensional regularisation : the introduced regulator @xmath449 appears in a gamma function which has a pole at @xmath455 and this pole describes exactly the divergence of the original integral .
assume that we have a space given by a spectral triple @xmath650 .
in this subsection we will consider the tensor product of the triple @xmath125 with the semifinite triple @xmath355 .
denote as before @xmath651 for the product of the spectral triple @xmath125 with the semifinite spectral triple @xmath355 as described in theorem [ semifinite_product ] .
the use of @xmath437 instead of @xmath652 is no typo , because @xmath653 .
define the _ gauge potentials _ of a spectral triple @xmath307 by @xmath654\,:\ , a_j , b_j\in{\mathcal{a}}\big\}.\ ] ] in quantum field theory one is interested in computing @xmath655 , where @xmath8 is the dirac operator of a spectral triple @xmath307 and @xmath615 a gauge potential .
however @xmath656 is not defined for ( un)bounded operators on infinite dimensional hilbert spaces .
recall from linear algebra the identity @xmath657 for any square matrix @xmath215 . thus using functional calculus we have for positive square matrices @xmath658 .
a formal manipulation gives @xmath659 therefore we are interested in computing @xmath660 instead of @xmath655 .
however , in general this quantity is not finite , as the operators considered are in general not trace class .
therefore it is necessary to regularize and renormalise this quantity .
we will describe a method which makes use of the semifinite spectral triple @xmath355 and the product of semifinite spectral triples .
motivated by zeta - function regularization we insert an extra factor @xmath661 and examine in what way the result diverges as a function of @xmath449 .
so we consider the regularised quantity @xmath662 to renormalise this , one needs to compute the residue at @xmath620 .
for this we use the noncommutative integral @xmath663 this functional defines a trace on the algebra generated by @xmath664 $ ] and @xmath665 with @xmath513 , if the triple has a simple discrete dimension spectrum @xcite .
the following theorem by connes and chamseddine about spectral actions will be useful later on .
[ variation_zeta ] suppose @xmath307 is a finitely summable regular spectral triple , @xmath8 is invertible and @xmath666 is a self - adjoint gauge potential .
write @xmath667 .
the following holds : 1 .
the function @xmath668 extends to a meromorphic function with a most simple poles ; 2 .
the function @xmath668 is regular at @xmath620 ; 3 .
the following equality holds @xmath669 see ( * ? ? ?
the above use of @xmath670 needs some comments .
namely the operator @xmath671 does not need to be positive , hence @xmath670 is not defined by the functional calculus .
but for the logarithm we do have the power series expansion @xmath672 note , that due to summability of the spectral triple , for @xmath11 large the residue @xmath673 .
so the infinite series @xmath674 is in fact a finite sum .
this sum is well defined for all gauge potentials @xmath615 .
so we use @xmath674 as a definition for @xmath675 .
+ we know that if @xmath8 is invertible and if @xmath666 is a gauge potential , then @xmath615 is bounded .
so there exists a constant @xmath676 such that @xmath677 .
for such an @xmath55 the operator @xmath678 is well - defined , in which case of course both interpretations ( the functional calculus and the finite sum ) in theorem [ variation_zeta ] are equal .
the following proposition relates the residues of the triples @xmath125 and @xmath679 .
it can also be used to compute anomalous graphs .
for example the result for @xmath680 of the proposition corresponds to the following graph .
graph2 ( 120,120 ) [ connes_1.242 ] suppose the dirac operator @xmath8 is invertible , let @xmath681 and @xmath682 .
then the function @xmath683 has at most a simple pole at @xmath684 , with residue given by @xmath685 see ( * ? ? ?
. [ fluctuation ] suppose @xmath686 is an even @xmath33-summable regular spectral triple , @xmath666 is a self - adjoint gauge potential and @xmath8 is invertible .
then @xmath687 the proof is a simple combination of the results in this section .
the first equality of is given by theorem [ variation_zeta ] and the second equality follows from proposition [ connes_1.242 ] .
applying the above theorem to zeta - function regularisation shows that we have two equivalent ways to renormalise the quantity @xmath660 .
namely write @xmath688 , then @xmath689
in this appendix we will recall some results , most of them related to traces and measures , which are being used in this paper . recall that if @xmath215 is densely defined and symmetric linear operator , then @xmath690 is the closure of @xmath215 .
so an equivalent definition of essentially self - adjointness is that @xmath215 has a unique self - adjoint extension .
[ bounded_vector ] suppose @xmath691 is an ( unbounded ) operator .
denote @xmath692 .
if @xmath693 and there exists a constant @xmath694 ( dependent on @xmath636 ) such that @xmath695 for all @xmath11 , then @xmath636 is called a _ bounded vector_. if there exists a constant @xmath186 such that @xmath696 for all @xmath11 , then @xmath636 is called an _ analytic vector_.
we denote all bounded vectors of @xmath215 by @xmath697 and all analytic vectors by @xmath698 .
it is clear that @xmath697 and @xmath698 are linear subspaces and @xmath699 . 1 .
if @xmath215 is self - adjoint , then @xmath701 dense ; 2 .
( nelson s theorem ) if @xmath215 is symmetric and @xmath702 is dense , then @xmath215 is essentially self - adjoint ; 3 . if @xmath215 is closed and symmetric , then @xmath215 is self - adjoint if and only if @xmath703 dense .
a trace on a von neumann algebra can be considered as a noncommutative integral . for integrals
we have the hlder inequality and fubini s theorem .
we therefore expect that such results generalise to traces .
such a generalisation is true if one assumes some regularity conditions on the trace .
e.g. ( * ? ?
ix.2.13 ) if @xmath704 is a semifinite von neumann algebra with a faithful , semifinite , normal trace @xmath20 , for all @xmath705 it holds that @xmath706 the analogue of fubini is more complicated , we need the following definition .
suppose @xmath707 is a finite measure space , @xmath708 a semifinite von neumann algebra and @xmath614 a semifinite faithful normal trace on @xmath704 .
denote @xmath709 a bounded function @xmath710 is _
@xmath711-measurable _ if for all @xmath712 the functions @xmath713 and @xmath714 are measurable .
define @xmath715 [ interchange_trace_int ] ( * ? ? ?
* lemma 3.10 ) let @xmath708 be a semifinite von neumann algebra with semifinite faithful normal trace @xmath614 and @xmath707 a finite measure space .
assume that the function @xmath716 and @xmath418 is uniformly @xmath717-bounded ( i.e. there exists @xmath186 such that @xmath718 for all @xmath719 ) , then @xmath720 , the function @xmath721 is measurable with respect to the @xmath722-algebra @xmath723 and @xmath724 let @xmath723 be a @xmath722-algebra on a set @xmath725 .
let @xmath19 be a von neumann algebra and @xmath726 $ ] be a normal trace .
suppose @xmath727 is a spectral measure such that @xmath728 for all @xmath729 .
then @xmath730 defines a measure on @xmath723 .
if @xmath20 is a finite trace , then @xmath731 is a finite measure .
in general from a measure we can construct an integral . in the case the spectral measure is given as the spectral decomposition of a self - adjoint operator we obtain the following link between the functional calculus and integration with respect to the measure @xmath731 .
[ integration_trace ] suppose @xmath215 is a self - adjoint operator on a hilbert space @xmath18 with spectral decomposition @xmath732 .
let @xmath733 be a von neumann algebra , with a normal trace @xmath734 $ ] and assume that @xmath215 is affiliated with @xmath19 .
if @xmath735 is a borel - measurable function such that @xmath736 or @xmath737 then , @xmath738 to prove this theorem we will apply the standard machine of measure theory .
so first suppose @xmath739 for some @xmath740 . then since @xmath215 is affiliated with @xmath19 each of its spectral projections @xmath741 . in particular since @xmath742 is a projection @xmath743 .
then we obtain @xmath744 now for a simple function @xmath745 equation holds , because the integral , the trace and the functional calculus are linear functionals . + if @xmath418 is a positive measurable function , then there exists a sequence of simple functions @xmath746 with @xmath747 pointwise .
the functional calculus implies that @xmath748 which tends to @xmath455 as @xmath613 because of the monotone convergence theorem ( here @xmath749 is the measure given by @xmath750 ) . by construction
the operators @xmath751 .
since @xmath19 is a von neumann algebra it is wot - closed , hence @xmath752 .
clearly @xmath753 , because @xmath418 is positive .
using the fact that @xmath20 is a normal trace gives @xmath754 .
again an application of the monotone convergence theorem yields @xmath755 this proves the proposition in the case that @xmath418 is positive .
+ if @xmath418 is not necessarily positive , but @xmath737 , then @xmath756 . we split up @xmath418 in four parts , @xmath757 and apply the previous equality to each of the four summands .
it follows that @xmath758 and similar statements for @xmath759 . by linearity of the integral and trace we obtain @xmath760 , as desired .
let @xmath704 be a von neumann algebra acting on @xmath18 and @xmath20 be a trace on @xmath704 .
we denote by @xmath763 , the projections .
the @xmath764 , are called the _
@xmath20-finite rank operators_. the _ @xmath20-compact operators _ are defined as @xmath765 , where the closure is taken in the norm topology .
it is immediate that if @xmath704 is a von neumann algebra acting on @xmath18 and @xmath20 a trace on @xmath704 , then @xmath766 is a closed ideal in @xmath75 in the norm - topology .
furthermore @xmath766 equals the closed ideal generated by the @xmath20-finite projections .
[ thm - tau - discrete ] let @xmath704 be an infinite , semifinite von neumann algebra acting on @xmath18 , equipped with a normal , faithful trace @xmath20 .
suppose @xmath215 is a self - adjoint @xmath704-affiliated ( unbounded ) operator on @xmath18 with spectral decomposition @xmath767
. then the following are equivalent : clearly ( i ) implies ( ii ) .
the converse implication is immediate from the first resolvent formula ( * ? ? ?
viii.2 ) and the fact that @xmath766 is an ideal in @xmath75 .
+ now suppose ( iii ) holds .
let @xmath773 .
define the bounded measurable function @xmath774 then @xmath775 is a bounded operator . put @xmath776}$ ] and define @xmath777 . since @xmath778 is bounded and measurable and the operator @xmath215 is affiliated to @xmath704 we have @xmath779 .
let @xmath780 be the orthogonal projection on @xmath781 .
then @xmath782)$ ] , thus by assumption @xmath783))<\infty$ ] .
so the operators @xmath784 have @xmath20-finite rank .
now let @xmath35 .
select @xmath785 such that @xmath786 . for @xmath787 and @xmath788
, we have @xmath789}\frac{1}{\lambda-\lambda_0}\,de\big)\,x - \big(\int_{{\mathbb{r}}}\frac{1}{\lambda-\lambda_0}\,de\big)\,x\big\|\notag\\ & = \big\|\big(\int_{(-\infty ,- n)\cup(n,\infty)}\frac{1}{\lambda-\lambda_0}\,de\big)\,x\big\|\notag\\ & \leq { \varepsilon}\|x\|.\end{aligned}\ ] ] this establishes ( i ) .
+ now suppose ( i ) holds we will show ( iii ) is true . because @xmath790 , the operator @xmath791 is @xmath20-compact .
so there exists a sequence of @xmath20-finite rank operators @xmath792 such that @xmath793 .
since @xmath794 , for all @xmath35 there exist @xmath785 and @xmath795 such that for all @xmath787 and @xmath796 it holds that @xmath797 } \frac{1}{\lambda - i}\,de\big\|\leq \big\|s_n - ( t - i)^{-1}\big\| + \big\| ( t - i)^{-1 } - \int_{[-\nu,\nu ] } \frac{1}{\lambda - i}\,de\big\| \leq{\varepsilon}/2+{\varepsilon}/2 = { \varepsilon}.\ ] ] now suppose that @xmath798))=\infty$ ] .
let @xmath799 and select corresponding @xmath800 and @xmath801 .
denote by @xmath780 again the projection on the range of @xmath784 . by orthogonality of @xmath780 and @xmath802
we obtain @xmath803 ) ) & = \tau(e([-\mu_0,\mu_0]\wedge q_n + e([-\mu_0,\mu_0])\wedge ( 1-q_n))\\ & \leq \tau(q_n ) + \tau(e([-\mu_0,\mu_0])\wedge ( 1-q_n)).\end{aligned}\ ] ] since @xmath804 for all @xmath11 @xmath805)\wedge ( 1-q_n))\geq \tau(e([-\mu_0,\mu_0 ] ) ) - \tau(q_n ) = \infty.\ ] ] hence @xmath806)\wedge ( 1-q_n)){\mathcal{h}}\neq\{0\}$ ] .
select an element @xmath807){\mathcal{h}}\cap\ker(s_n)$ ] , with @xmath808 . for this @xmath636 and for any @xmath809
the following identities hold @xmath810}\frac{1}{\lambda - i}\,de\big ) x = \big(\int_{[-\mu_0,\mu_0]}\frac{1}{\lambda - i}\,de\big ) x , \qquad s_nx=0.\ ] ] combining these equalities and using the properties of @xmath389 gives @xmath811 } \frac{1}{\lambda - i}\,de\big)\ , x\big\| & = \big\|\big(\int_{[-\mu_0,\mu_0 ] } \frac{1}{\lambda - i}\,de\big)\ , x\big\|\\ & \geq \big|\big\langle\big(\int_{[-\mu_0,\mu_0 ] } \frac{\lambda+i}{\lambda^2 + 1}\,de\big)\ , x , x\big\rangle\big|\\ & \geq \big|\operatorname{im}\big(\int_{[-\mu_0,\mu_0 ] } \frac{\lambda+i}{\lambda^2 + 1}\,de_{x , x}\big)\big|\\ & = \big|\int_{[-\mu_0,\mu_0 ] } \frac{1}{\lambda^2 + 1}\,de_{x , x}\big|\\ & > { \varepsilon}\ , |e_{x , x}([-\mu_0,\mu_0])|= { \varepsilon}.\end{aligned}\ ] ] but this is a contradiction with equation . hence we established ( iii ) .
[ tau_compact ] if @xmath704 is a semifinite von neumann algebra acting on @xmath18 , with a normal , faithful trace @xmath20 and if @xmath215 is a self - adjoint @xmath704-affiliated operator on @xmath18 which satisfies one of the equivalent conditions of theorem [ thm - tau - discrete ] , then @xmath215 is called a _
@xmath20-discrete operator_. n.a .
azamov , a.l .
carey , p.g .
dodds and f.a .
sukochev , _ operator integrals , spectral shift and spectral flow _ , canad . j. math . 61 ( 2009 ) , 241 - 263 .
m. benameur and t. fack _ type ii non - commutative geometry i. dixmier trace in von neumann algebras _ , adv . math . 199
( 2006 ) 29 - 87 .
b. blackadar , _ operator algebras : theory of c * -algebras and von neumann algebras _ , springer , berlin , 2006 . a.l .
carey , j. phillips , a. rennie and f.a .
sukochev , _ the local index formula in semifinite von neumann algebras . i. spectral flow _
, advances in math .
202 ( 2006 ) , 451 - 516 .
a. connes , _ gravity coupled with matter and the foundation of noncommutative geometry _ ,
182 ( 1996 ) , 155 - 176 .
a. connes , _ noncommutative geometry _ , academic press , san diego , 1994 .
a. connes , _ on the spectral characterization of manifolds _ ,
a. connes and a.h .
chamseddine , _ inner fluctuations of the spectral action _ , j. geom .
57 ( 2006 ) , n.1 , 1 - 21 .
a. connes and m. marcolli , _ noncommutative geometry , quantum fields and motives _ , american mathematical society , providence , 2008 .
a. connes and h. moscovici , _ the local index formula in noncommutative geometry _ , geom .
5 no . 2 ( 1995 ) ,
174 - 243 .
l. dabrowski and g. dossena , _ product of real spectral triples _
methods mod .
phys . - vol . 8 , no . 8
( 2011 ) , 1833 - 1848 .
folland , _ quantum field theory , a tourist guide for mathematicians _ , american mathematical society , providence , 2008 .
gracia - bondia , j.c .
varilly and h. figueroa , _ elements of noncommutative geometry _ , birkhuser , boston , 2001 .
s.w . hawking , _ zeta function regularization of path integrals in curved spacetime _ , communications in mathematical physics 55 ( 2 ) ( 1977 ) , 133 - 148 . g. t hooft and m. veltman , _ regularization and renormalization of gauge fields _ , nucl .
b44 ( 1972 ) , 189 - 213 .
olver , _ nist handbook of mathematical functions _ , national institute of standards and technology , u.s . and cambridge university press , cambridge , 2010 .
o. uuye , _ pseudo - differential operators and regularity of spectral triples _ , fields communications series , volume 61 - `` perspectives on noncommutative geometry '' , 2011 , 153 - 163 .
m. reed and b. simon , _ methods of modern mathematical physics , i : functional analysis _ , academic press , new york , 1972 .
k. schmdgen , _ unbounded self - adjoint operators on hilbert space _ , springer , dordrecht , 2012 .
m. takesaki , _ theory of operator algebras .
i _ , springer , berlin , 2002 .
m. takesaki , _ theory of operator algebras .
ii _ , springer , berlin , 2003 . f. j. vanhecke , _ on the product of real spectral triples _ , lett .
50 ( 1999 ) , 157162 .
|
in this paper we will extend the product of spectral triples to a product of semifinite spectral triples .
we will prove that finite summability and regularity are preserved under taking products .
connes and marcolli constructed for each @xmath0 a type @xmath1-semifinite spectral triple which can be considered as a geometric space of dimension @xmath2 . a small adaption of their construction yields a type @xmath3-semifinite spectral triple .
we will investigate the properties of these semifinite spectral triples . at the same time
we will also avoid the need for an infra - red cutoff to compute the dimension spectrum .
using this collection of semifinite spectral triples and the product of semifinite spectral triples one can construct a mathematical tool for dimensional and zeta - function regularisation in quantum field theory .
real dimensional spaces in noncommutative geometry
|
in previous publications @xcite , methods to compute the modified bessel function @xmath0 for positive @xmath3 were described .
we complete here this analysis by describing analogous methods for the computation of the function @xmath1 . with this
, methods for the reliable computation of a pair of linearly independent numerically satisfactory solutions become available which find their implementation in the accompanying paper @xcite . methods to compute their derivatives are also provided .
the functions @xmath0 and @xmath1 are solutions of the modified bessel equation for imaginary orders @xmath5 the function @xmath0 finds application in a number of problems of physics and applied mathematics @xcite .
the function @xmath1 is a real valued numerically satisfactory companion to @xmath0 in the sense described in @xcite , pp .
154155 . in terms of the modified bessel function of the first kind @xmath6 ,
the solutions are defined as : @xmath7\,,\,\ , l_{ia}(x)=\frac{1}{2}\left[i_{-ia}(x)+i_{ia}(x)\right ] , \label{definition}\ ] ] with wronskian @xmath8=1/x$ ] .
both @xmath0 and @xmath1 are real solutions for real @xmath9 and @xmath10 . because they are even functions of @xmath2 , in the sequel we will consider @xmath11 , although this restriction is not present in the code . in section 2
, we describe the different methods of computation considered , namely : series expansions , asymptotic expansions for large @xmath3 , airy - type uniform asymptotic expansions , non - oscillating integral representations ( including a discussion on the quadrature rule ) and a continued fraction method .
we avoid duplicating information already given in previous papers ; in particular , the references @xcite provide information required for building the algorithms of the accompanying paper @xcite .
a few misprints in @xcite are corrected . in section 3 , we include a discussion on the dominant asymptotic behaviour of the functions . these exponential dominant factors can be taken out , leading to scaled functions which can be computed in a much wider range .
the algorithm described in the accompanying paper @xcite offers the possibility of computing scaled and unscaled functions .
in @xcite methods are described to compute the @xmath0 for different regions in the @xmath12 plane . in particular
, we considered series expansions @xcite , asymptotic expansions for large @xmath3 ( @xcite , eq .
9.7.2 ) , uniform asymptotic expansions for @xmath13 ( @xcite and @xcite , pg .
425 ) . also , non - oscillating integral representations @xcite are available .
similar techniques are available for the computation of @xmath1 and the derivatives @xmath14 , @xmath15 . in addition , a continued fraction method can be applied for the computation of @xmath16 and @xmath14 .
those techniques generally give at least two alternative methods for computing the functions in the @xmath12 plane for moderate values of @xmath3 and @xmath2 ; therefore , we can always compare different methods of computation for testing their accuracy .
the selection of one or another method of computation in a given region will depend on the range of applicability of each method and its efficiency .
we now describe the different methods of computation which are used in the programs .
series expansions can be built which properly describe the solutions near the singular point ( @xmath17 ) of the defining differential equation ( [ ode ] ) .
the idea , as described in @xcite , is to substitute the maclaurin series for @xmath6 ( @xcite , eq .
9.6.10 ) in eqs . ( [ definition ] ) .
the following expansions are obtained @xmath18c_k\\ l_{ia}(x)=n(a ) \displaystyle\sum_{k=0}^{\infty}r_k c_k & , & l'_{ia}(x)= n(a ) \frac{2}{x}\displaystyle\sum_{k=0}^{\infty } \left[k r_k+a^2 \frac{f_k}{2}\right]c_k , \end{array } \label{series}\ ] ] where @xmath19 and @xcite @xmath20 the coefficients @xmath21 and @xmath22 differ from those in @xcite by a constant factor ( for fixed @xmath2 ) .
the new normalization shows explicitly ( eqs .
( [ series ] ) ) the dominant exponential behaviour @xmath23 and @xmath24 as @xmath25 ( @xmath26 ) .
an efficient method to compute the coefficients was described in @xcite ; this method is based on the fact that both @xmath21 and @xmath22 satisfy the three - term recurrence relation @xmath27 perron s theorem is inconclusive with respect to the existence of minimal solutions for this recurrence relation ; anyhow , the second equation in ( [ explicit ] ) confirms that neither @xmath21 nor @xmath22 are minimal solutions .
therefore , forward recursion will be numerically stable .
starting values can be computed taking into account that @xmath28 , where @xmath29 is the coulomb phase shift , for which chebyshev expansions are available for double precision @xcite .
namely , we have : @xmath30,\\ \\ r_1=\frac{1}{1+a^2}\left\{\cos [ \sigma_0 ( a)-a\ln ( x/2)]- a \sin[\sigma_0 ( a)-a\ln ( x/2 ) ] \right\ } , \end{array}\ ] ] and @xmath31,\\ \\ f_1=\frac{1}{a(1+a^2)}\left\{\sin [ \sigma_0 ( a)-a\ln ( x/2 ) ] + a \cos[\sigma_0 ( a)-a\ln ( x/2)]\right\}\ , . \end{array}\ ] ] these formulas correct two misprints in @xcite ( eqs .
12 and 13 ) .
series can be used for @xmath32 small .
see @xcite , section 2 [ canv ] .
asymptotic expansions for large @xmath3 are available from the known asymptotic expansion of @xmath6 ( @xcite , eq .
9.7.1 ) : @xmath33 where @xmath34 is the hankel symbol , which satisfies @xmath35 bounds for the error terms ( @xmath36 , @xmath37 ) can be found in @xcite , pg .
269 , ex . 13.2 . as discussed in @xcite the numerical performance of
the asymptotic expansion for @xmath0 is of more restricted applicability than for the case of the evaluation of @xmath38 for real @xmath39 .
furthermore , the continued fraction method described in @xcite covers the region where this expansion is of numerical interest . for this reason ,
the continued fraction method is the preferred algorithm for the evaluation of @xmath0 and @xmath40 for moderate values of @xmath2 . on the other hand
the asymptotic expansion for @xmath1 turns out to be accurate in a wider region , which is a fortunate situation given that the continued fraction method is not available in this case .
see @xcite , section 2 , for further details .
asymptotic expansions for the derivatives are also available by differentiating eqs .
( [ largex ] ) . the airy - type asymptotic expansions for @xmath0 can be found in @xcite and @xcite pg .
425 ; the analogous expansions for @xmath1 @xcite and @xmath40 are also available @xcite , while the expansion for @xmath41 can be derived in the same way .
we summarize here the main features needed for the computation through these expansions , neglecting the error terms .
further details can be found in @xcite and @xcite . the expansion for @xmath0 and @xmath1 in terms of airy functions ( @xmath42 , @xmath43 and their derivatives ) reads @xmath44 , \\ \\ l_{ia}(a z)=\frac{e^{a\pi/2 } \phi(\zeta)}{2a^{1/3 } } \left[{\mbox{bi}}(-a^{2/3}\zeta ) f_a ( \zeta ) + \frac{1}{a^{4/3 } } { \mbox{bi}}^{\prime}(-a^{2/3}\zeta ) g_a ( \zeta)\right ] , \end{array } \label{unifun}\ ] ] where @xmath45 as @xmath46 uniformly with respect to @xmath47 .
error bounds for the asymptotic expansions of the @xmath0 and @xmath1 are given in @xcite .
the quantity @xmath48 is given by @xmath49 and @xmath50 of course , it is crucial to compute accurately eqs .
( [ zeta ] ) for small @xmath48 . for this
, series expansions around @xmath51 can be used .
the evaluation of the coefficients near the turning point @xmath51 ( which is our region of interest ) can be performed via maclaurin series expansions of the quantities @xmath52 , @xmath53 and @xmath54(@xcite ) in terms of the variable @xmath55 ( see @xcite and @xcite for further details ) .
asymptotic expansions for the derivatives can be found by differentiating eqs .
( [ unifun ] ) . in this way : @xmath56 , \\ \\
l'_{ia}(a z)=\frac{e^{a\pi/2}}{z a^{2/3}\phi(\zeta ) } \left[{\mbox{bi}}^{\prime}(-a^{2/3}\zeta ) p_a ( \zeta)+\frac{1}{a^{2/3}}{\mbox{bi}}(-a^{2/3}\zeta ) q_a ( \zeta)\right ] , \end{array } \label{unifun2}\ ] ] where @xmath57 and @xmath58 can be written in terms of @xmath59 , @xmath60 and their derivatives and they have asymptotic expansions @xmath61 whose coefficients can be obtained from the computed coefficients @xmath53 and @xmath54 ( in taylor series around @xmath62 ) through the relations : @xmath63 where @xmath64 the prime in eqs .
( [ ces ] ) and ( [ chi ] ) denotes the derivative with respect to @xmath48 . using eqs .
( [ ces ] ) the coefficients @xmath65 and @xmath66 can be computed from the coefficients @xmath53 and @xmath54 .
details on the evaluation of @xmath53 and @xmath54 are given in @xcite , where an explicit maple algorithm is given for the computation of @xmath53 and @xmath54 for @xmath67 . by computing the wronskian relation for the modified bessel functions and using the wronskian for bessel functions , it is easy to derive the relation @xmath68 which is a useful relation for checking the correctness of the approximations for the coefficients in the asymptotic expansions . an algorithm to compute airy functions of a real variable
is needed for the computation of these asymptotic expansions . in the routines
@xcite we use algorithm 819 @xcite .
these airy - type asymptotic expansion are applied in @xcite in a broad region around the turning point line @xmath69 .
paths of steepest descent for integral representations of the modified bessel functions of imaginary orders and their derivatives are given in @xcite .
apart from their application in asymptotics @xcite , these integrals are useful for building numerically stable ( non - oscillating ) integral representations for @xmath0 and @xmath70 , as described in @xcite .
we complete here the analysis in @xcite by providing analogous expressions for the computation of @xmath1 and @xmath15 .
additionally , we study further transformations of the integrals which enable us to obtain integral expressions suitable for computation by means of the trapezoidal rule .
we have the following integral representations in the monotonic region @xcite @xmath72\ , e^{-x \phi(\tau)}\ , d \tau , \end{array } \label{kmon}\ ] ] where @xmath73 and @xmath74 , @xmath75 $ ] . the dominant exponential term ( @xmath76 )
has been factored out .
the argument of the exponential in the integrand is @xmath77 this formula corrects a misprint in @xcite ( eq .
33 ) . the difference @xmath78 can be computed in a stable way for small values of @xmath79 by using the expression .
@xmath80\,,\ ] ] together with the definition of @xmath81 ( [ definiciones ] ) and specific algorithms to compute @xmath82 and @xmath83 for small @xmath79 .
the non - oscillating integral representations for @xmath1 and its derivative can be written after factoring the dominant exponential contribution as : @xmath84 \label{lmon}\ ] ] where @xmath85=2x\left(\sqrt{1-(a / x)^2}-\frac{a}{x}\arccos(\frac{a}{x})\right ) \end{array}\ ] ] and using eq .
( [ definiciones ] ) @xmath86 .\ ] ] the first integral is dominant over the second one for large values of the parameters and @xmath87 not too close to @xmath69 . as @xmath88 both integrals become of the same order .
similarly , we have the following representation for @xmath15 : @xmath89 \label{lpmon}\ ] ] where @xmath90\ ] ] these integral representations for @xmath1 and @xmath15 can be used for checking the computation of these functions in the monotonic region .
they are not used by our algorithms @xcite because the airy - type asymptotic expansion ( section [ atuae ] ) and the expansion for large @xmath3 ( section [ asyx ] ) are sufficiently accurate for this functions and they are faster to compute ( see @xcite , section 2 ) .
the non - oscillating integral representations for the oscillatory region are more difficult to evaluate numerically than those for the monotonic case .
indeed , as it was discussed in @xcite , the steepest descent method leads to three integrals , which have to be computed separately .
however , as we later discuss , for moderately large @xmath2 it will be enough to compute a single integral . in @xcite ,
the following formula was obtained : @xmath92\ , , \end{array } \label{oscil}\ ] ] where @xmath93 , @xmath94 , @xmath95 and @xmath96 notice that each of the three integrals in eq .
( [ oscil ] ) can in principle be integrated with respect to any of the two variables @xmath81 and @xmath79 , taking into account eq .
( [ sigma ] ) together with the fact that the integration path @xmath97 is such that @xmath98 , @xmath99 , @xmath100 ; however , as discussed in @xcite there are strong numerical reasons for the selections made .
in particular , the third integral is performed with respect to @xmath101 ( which requires numerical inversion of ( [ sigma ] ) ) to avoid the singularity of @xmath102 at @xmath103 . as explained in @xcite the numerical inversion of ( [ sigma ] ) in the interval @xmath104 $ ]
can be efficiently performed in parallel with the numerical integration .
similar integral representations exist for @xmath105 , @xmath1 and @xmath15 .
we have : @xmath106 \end{array } \label{oscil2}\ ] ] where @xmath107 in addition , integral representations for @xmath1 and its derivative are : @xmath108 \label{oscil3 } \end{array}\ ] ] and @xmath109 \label{oscil4 } \end{array}\ ] ] notice that the dominant exponential behaviour has been factored for both the functions @xmath0 and @xmath1 and their derivatives , which coincides with the exponential behaviour of the uniform asymptotic expansion .
this is an interesting feature when computing scaled functions in order to avoid overflows and/or underflows in the computation .
after factoring the dominant exponential terms ( @xmath110 ) , the overflow and/or underflow problems are eliminated ; notice , however , that when computing the integrals over finite intervals we should evaluate @xmath111 , @xmath112 for @xmath1 and its derivative and @xmath113 , @xmath114 for @xmath0 ( and @xmath105 ) instead of computing the hyperbolic and the exponential separately ( otherwise , overflows will take place for moderately large @xmath2 ) .
for this reason it is convenient to use the expressions @xmath115 in the evaluation of eqs .
( [ oscil ] ) and ( [ oscil2 ] ) and to proceed in the same way for @xmath116 and @xmath117 in eqs .
( [ oscil3 ] ) and ( [ oscil4 ] ) .
notice that in the oscillatory region @xmath118 and that for large @xmath2 and @xmath3 both @xmath119 and @xmath120 will underflow .
these underflow problems can be easily avoided by neglecting these exponential terms for large parameters .
in addition , when both exponentials become negligible , the integral over sigma becomes negligible and the remaining two integrals can be approximated by only one integral .
we can write @xmath121 \\ \\ k'_{ia}(x)\approx e^{-\pi a/2}\left [ \displaystyle\int_{\tau_0}^{\infty } e^{-\psi(\tau ) } \left(\cos \chi\ , a(\tau)+\sin\chi\ , c(\tau ) \right)d\tau + { \cal o}(e^{-\pi a/2 } ) \right ] \\
\\ l_{ia}(x)\approx\frac{e^{\pi a/2}}{2\pi}\left [ \displaystyle\int_{\tau_0}^{\infty } e^{-\psi(\tau ) } \left ( \sin\chi -\cos\chi\frac{d\sigma}{d\tau } \right)d\tau + { \cal o}(e^{-\pi a/2 } ) \right ] \\ \\
l^{\prime}_{ia}(x)\approx \frac{e^{\pi a/2}}{2\pi } \left [ \displaystyle\int_{\tau_0}^{\infty } e^{-\psi(\tau ) } \left ( \sin\chi\ , a(\tau ) - \cos\chi\ , c(\tau)\right)d\tau + { \cal o}(e^{-\pi a/2 } ) \right ] \end{array } \label{simplificado}\ ] ] where @xmath122 these approximations can be used for moderately large @xmath2 , which is the region where integrals for the oscillatory case are employed in the code @xcite .
it is however useful to have the complete expressions for testing the rest of the methods .
the computation through quadrature using eqs .
( [ kmon ] ) , ( [ lmon ] ) , ( [ lpmon ] ) , ( [ oscil ] ) , ( [ oscil2 ] ) , ( [ oscil3 ] ) and ( [ oscil4 ] ) , provides a way for computing the functions in the whole @xmath12 plane , except close to @xmath69 , where the integrands become non - smooth . for this reason ,
they have been used for checking the algorithm , although in the oscillatory region only eq .
( [ simplificado ] ) is necessary when building the numerical algorithm @xcite .
as reported in goodwin @xcite , the trapezoidal rule is a very efficient method of computation of integrals @xmath123 for rapidly decaying integrands @xmath124 ; in particular , it is know that the error decays as @xmath125 for integrals of the type @xmath126 with @xmath127 analytic in @xmath128 . after appropriate changes of variable , similar arguments
follow for integrals over finite intervals with a smooth integrand @xcite .
the semi - infinite integrals in this section are appropriate for their computation by using the trapezoidal rule , because they decay as a double exponential as @xmath129 . on the other hand ,
the integrals over finite intervals show abrupt variations as @xmath88 , particularly in the oscillatory case , but under an adequate change of variables they can be also computed efficiently by means of the trapezoidal rule . for finite integrals , we consider a change of variable in order to map the finite interval @xmath130 $ ] into @xmath131 and a successive change to improve the convergence of the trapezoidal rule @xcite ; namely , we consider the following transformation : @xmath132 and the integral is discretized by means of the trapezoidal rule with equal mesh size : @xmath133 where the error @xmath134 is expected to decay very fast as the mesh size is decreased because the integrand is analytic and its decay is doubly exponential .
we use a trapezoidal rule which halves the mesh size until the prescribed precision is reached ; the same rule controls that the truncation of the infinite series ( [ dis ] ) gives an error well below the accuracy claim . regarding the semi - infinite integrals , we use a change of variable to transform the integration interval @xmath135 to @xmath136 .
we consider the following change of variables to perform this map . @xmath137
the additional change @xmath138 improves the convergence of the trapezoidal rule .
it is observed that , typically , no more than @xmath139 iterations of the trapezoidal rule are needed , which means that the integrands are evaluated at @xmath140 points in the worst cases .
this is the typical number of iterations for the evaluation of @xmath141 by means of a recursive trapezoidal rule when double precision accuracy is demanded .
this fact confirms that the above mentioned changes of variable are adequate for the computation of the integrals for the modified bessel functions . as discussed in @xcite
both @xmath0 and @xmath142 can be computed for moderate @xmath2 by means of a continued fraction method , similar to the corresponding method for bessel functions of real orders ( see @xcite and @xcite , pp .
239 - 240 ) .
we refer to @xcite for a full description of this scheme .
as numerical experiments show , this method is competitive in speed with asymptotic expansions for large @xmath3 ( section [ asyx ] ) and the range of application is larger .
therefore , the continued fraction method substitutes the use of asymptotic expansions for large @xmath3 .
as described in previous sections , the integral representations which were developed indicate that the dominant behaviour for the functions when the parameters are large is of exponential type .
this means that the computations can only be carried for not too large values of @xmath2 and @xmath3 in order to avoid overflows / underflows in the computation .
for instance , from eqs .
( [ oscil ] ) and ( [ oscil2 ] ) it is seen that for large @xmath2 ( @xmath143 ) , we have @xmath144 and similarly for the derivative , while for @xmath1 and its derivative ( eqs .
( [ oscil3 ] ) and ( [ oscil4 ] ) ) the asymptotic behaviour is @xmath145 .
this means that to avoid overflow / underflows in the computation , we must restrict the range of @xmath2 in the oscillatory region to @xmath146 where @xmath147 is either the inverse of the underflow number ( when computing @xmath0 or its derivative ) or the overflow number ( for @xmath1 and its derivative ) ; @xmath148 is a safety factor ( in the program , we take @xmath149 ) . for processors using the ieee standard in double precision
this will approximately limit @xmath2 to @xmath150 . on the other hand , for the monotonic region ( @xmath71 ) the integral representations show that the dominant exponential behaviour is @xmath151 , @xmath152 where @xmath153 , @xmath154 ( @xmath155 $ ] ) , and similarly for the derivatives .
this means that , in order to avoid overflows / underflows , the range of computation must be restricted to : @xmath156 given that all our expressions have the dominant exponential contributions factored out , exponentially scaled functions can be defined which are computable in wider ranges .
namely , we define : @xmath158 and @xmath159 note that , as in the rest of the article , we are considering positive @xmath2 because @xmath0 and @xmath1 are even functions of @xmath2 .
of course , when applying the scaling factors for negative @xmath2 , we should replace @xmath2 by @xmath160 in the exponential scaling factors of eqs .
( [ escaladas ] ) and ( [ escaladas2 ] ) . in this way ,
the scaled functions are also even functions of @xmath2 .
the definitions in ( [ escaladas ] ) and ( [ escaladas2 ] ) eliminate exactly the front exponential factor in the oscillatory region @xmath161 ) from the series and the airy type asymptotic expansion and in all the @xmath12 plane for the the integral representations .
in other cases , there remains an exponential factor with soft variation .
for example , when using airy - type expansions in the monotonic region ( neglecting non - exponential factors ) , we have @xmath162 where @xmath163 which is small for @xmath164 ( @xmath165 ) ; loss of accuracy in the computation of @xmath166 for @xmath164 can be reduced by expanding @xmath166 in powers of @xmath167 . 99 m. abramowitz , i. stegun ( eds ) . _ handbook of mathematical functions . _ national bureau of standards .
applied mathematics series , no .
government printing office , washington dc ( 1964 ) .
asymptotic expansions of the modified bessel function of the third kind of imaginary order .
siam j. appl .
15(5 ) ( 1967 ) 13151323 .
cody , k.e .
chebyshev approximations for the coulomb phase shift .
math . comp . 24
( 1970 ) 671 - 677 .
bessel functions of purely imaginary order , with an application to second order linear differential equations having a large parameter .
siam j. math .
21(4 ) ( 1990 ) 951018
. b. r. fabijonas .
laplace s method on a computer algebra system with an application to the real valued modified bessel functions .
j. comput .
( 2002 ) 323342 .
a. gil , j. segura , n. m. temme .
algorithm xxx : modified bessel functions of imaginary order and positive argument .
acm trans . math . soft .
( submitted for publication in this same issue ) .
a. gil , j. segura , n. m. temme .
evaluation of the modified bessel function of the third kind for imaginary orders .
j. comput .
( 2002 ) 398 - 411 a. gil , j. segura , n.m .
computation of the modified bessel function of the third kind of imaginary orders : uniform airy - type asymptotic expansion .
j. comput .
( 2003 ) 225 - 234 .
a. gil , j. segura , n.m .
algorithm 819 : aiz , biz : two fortran 77 routines for the computation of complex airy functions .
acm trans . math .
28 ( 2002 ) 325 - 336
. e. t. goodwin .
the evaluation of integrals of the form @xmath175 .
cambridge philos .
( 1949 ) 241245 .
_ asymptotics and special functions_. reprinted by a.k .
peters ltd . , 1997 .
w.h.press , s.a .
teukolsky , w.t .
vetterling , b.p .
numerical recipes in fortran . the art of scientific computing .
second edition .
cambridge university press , cambridge , 1992 .
. numerical integration of analytic functions .
j. comput .
phys . 4 ( 1969 ) 1929 .
h. takahasi , m. mori .
quadrature formulas obtained by variable transformation .
math . 21 ( 1973 ) 206219 .
h. takahasi , m. mori .
double exponential formulas for numerical integration .
inst . math .
( 1974 ) 721741 . n. m. temme . on the numerical evaluation of the modified bessel function of the third kind .
j. comput .
19(3 ) ( 1975 ) , 324337 . n.m .
steepest descent paths for integrals defining the modified bessel functions of imaginary order .
methods appl .
1(1 ) ( 1994 ) , 1424 . n.m .
numerical algorithms for uniform airy - type asymptotic expansions .
algorithms 15(2 ) ( 1997 ) 207225 .
|
we describe a variety of methods to compute the functions @xmath0 , @xmath1 and their derivatives for real @xmath2 and positive @xmath3 .
these functions are numerically satisfactory independent solutions of the differential equation @xmath4 . in the accompanying paper @xcite
, we describe the implementation of these methods in fortran 77 codes .
|
electron energy - loss spectroscopy ( eels ) is an efficient and widely used experimental method to study excitation processes on clean and adsorbates - covered surfaces of solids , and in thin ( including atomically thin ) films @xcite .
this method utilizes the inelastic scattering of electrons , resulting in both the energy and momentum transfer from the projectiles to diverse kinds of excitations in the samples . reflected or transmitted electrons
are analyzed with respect to the energy and momentum loss they have experienced in the interaction with a target , revealing a wealth of information about the properties of the latter .
much efforts have been exerted over years to complement eels experimental techniques with comprehensive theoretical pictures @xcite . in this way , a clear understanding of elementary excitations ( such as electron - hole pairs generation , collective electronic excitations
plasmons , atomic vibrational modes , _ etc . _ ) , including their momentum dispersion , for solid surfaces , interfaces , and in thin films have been achieved .
presently , the main approach to interpret eels data theoretically is to use _ energy - loss functions_. a clear example is the surface energy - loss function of a semi - infinite solid , which , with the neglect of the momentum dispersion , can be written as @xcite @xmath0 where @xmath1 is the frequency - dependent dielectric function ( df ) of the bulk solid .
this example exhibits an important feature common also to other , much more sophisticated , loss - functions : @xmath2 of eq .
( [ ls ] ) is a property of the target only . indeed ,
it is not concerned with the setup of the eels experiment , such as the angles of incidence and reflection ( or transmission ) , the energy of the electrons in the incident beam , and , which is subtler , the detailed , desirably quantum - mechanical , motion of the probe electrons both outside and inside the target . as a clear reason why such an approach may not be adequate
, we note that it can not , in principle , determine the relative intensities of the surface and the bulk plasmons in a given eels setup , the bulk response being given in the same approximation by another energy - loss function @xmath3 . for systems where the bulk and the surface excitations overlap ,
as is the case , e.g. , of silver , this constitutes a serious limitation .
meanwhile , a theoretical approach to eels taking full account of the incident electron kinematics had been introduced two decades ago @xcite .
this is based on the solution to the problem of the energy - loss by an electron traveling in the lattice potential of a target , utilizing the method known in the scattering theory as the _ distorted - wave approximation _
@xcite ( see eq .
( [ dsdwdw ] ) of the next section ) .
that formal theory of the response of the target system coupled to the quantum - mechanical motion of the projectile electron has , however , never been implemented to the full extent in calculations for specific systems .
indeed , the formalism importantly stipulates that the density - functional theory@xcite ( dft ) potential used in the calculations of the ground - state and of the response of the target , on one hand , and the potential which determines the motion of the projectile electron , on the other , should be the same crystalline potential .
only two specific applications of the theory have been made so far . in the first , the theory of ref . has been implemented for jellium within a model of the incident electron reflected from an infinite barrier at a given position above / below the surface @xcite . in the other , which is an application to the inelastic low - energy electron diffraction ( leed ) of simple metals ,
a severe approximation of the kinematic diffraction theory was used @xcite . at the same time , detailed measurements in the high - resolution eels ( hreels ) of silver surface in the wide energy range have become available recently@xcite , calling for the implementation of refined theoretical methods .
[ h ] the purpose of this paper is , therefore , two - fold .
first , we aim at the implementation of the theory of eels of ref . in its original form ,
i.e. , that would treat the incident electron and the electrons of the target system on the same footing .
secondly , we apply this theory to the eels of ag surface , which is exactly the case when the interplay of the response of the target with the details of the probe s motion is especially important , due to the overlap of the bulk and the surface features in the excitation spectrum of this material , as is illustrated in fig .
thereby we both further advance the theory of eels and achieve an improvement in the understanding of the experimental spectra of the ag surface .
since the fully _ ab initio _ solution to the problem of the dielectric response of @xmath4-metals still remains a computationally formidable task , we have to resort to some model considerations .
first , we substitute the three - dimensional ( 3d ) problem with a one - dimensional ( 1d ) one , neglecting the system s non - uniformity in the surface plane .
second , the @xmath4-electrons are included in a phenomenological way , using the model of liebsch @xcite of the background df .
this work should , therefore , be considered as a step forward toward the full - featured 3d implementation of the same method , treating also @xmath4-electrons _
ab initio_. however , the main ingredients of the theory ( among them , importantly , the necessary inclusion of the optical potential ) are presented and discussed in this work , facilitating the future implementation of the method in full .
the paper is organized as follows . in sec .
[ form ] we remind and further work out details of including the motion of the scattered electrons in the theory of eels . in sec .
[ res ] , results of the calculations conducted with the use of our theory are presented and discussed .
conclusions are collected in sec .
[ concl ] . in the appendix we detail on some important properties of the model utilized in the calculations .
we use atomic units ( @xmath5 ) throughout unless otherwise indicated .
a formal solution to the problem of the inelastic scattering of an electron in the eels setup , which incorporates the detailed quantum - mechanical motion of the projectile , can be quite generally written as @xcite @xmath6 where the left - hand side of eq .
( [ dsdwdw ] ) is the differential cross - section of the scattering from the state of the momentum @xmath7 to a state of the momentum within the solid angle @xmath8 around @xmath9 , and to lose the energy within @xmath10 around @xmath11 . in the right - hand side of eq .
( [ dsdwdw ] ) , @xmath12 is the interacting - electrons , @xmath13 includes electron - electron interactions , and the two response functions are related as @xmath14 , where @xmath15 is the xc kernel @xcite .
] density - response function of the target , the complex - valued `` external charge density '' @xmath16 is determined by the _ elastic _ scattering _ incoming _ and _ outgoing _ wave - functions , @xmath17 and @xmath18 , respectively , which are the solutions to the lippmann
schwinger equation @xcite @xmath19 where @xmath20 is the non - interacting green s function , @xmath21 is the single - particle static lattice potential , and @xmath22 are plane - waves . )
is the fourrier transform . ]
the structure of eq .
( [ dsdwdw ] ) has a transparent physical interpretation : the external charge @xmath23 creates an external coulomb potential , which , through the density - response function @xmath13 , induces the charge fluctuation in the target . finally , the coulomb potential of that fluctuation couples to the external charge itself , causing the inelastic scattering of the latter
. it must be , however , emphasized that the above picture is no more than a convenient verbal description of the strict mathematical formalism presented in ref . : the derivation of eq .
( [ dsdwdw ] ) does not rely on the substitution of the true quantum - mechanical scattering problem for an electron with an artificial charge - density .
it rather solves the problem of the combined elastic and inelastic scattering of a charge at an arbitrary many- ( or a few ) body system , which , with a mild assumption that the impinging electron can be considered distinguishable from those of the target , can be put into the terms of the density - response function of the target and the elastic scattering states of the projectile .
obtained within the distorted - wave approximation@xcite , eq .
( [ dsdwdw ] ) is exact to the first order in the inelastic processes ( the first born approximation ) and it is exact to _ all _ orders in the elastic scattering .
it includes both the long- and the short - range interaction of the probe electron with the target , i.e. , the dipole and impact scattering@xcite , respectively , within , most importantly , the quantum - mechanical treatment of the probe itself .
of course , practically , the quality of specific calculations by eq .
( [ dsdwdw ] ) depends on the accuracy of the approximations used to calculate its ingredients , i.e. , the density - response function of the target @xmath13 and the wave - functions of the incoming and outgoing electron utilized in the construction of @xmath23 .
we now turn to the use of specific models . in this work we will use a simplification of the potential @xmath21=@xmath24 averaged in the plane parallel to the surface ( which is chosen as the @xmath25-plane , with the @xmath26-axis normal to the surface and directed into vacuum ) . in this case
the wave - functions are plane - waves in the direction parallel to the surface @xmath27 where the subscript ` parallel ' denotes the @xmath25-projection of a vector . to take advantage of the scattering theory framework , in the following we represent the target with a sufficiently thick slab , with vacuum both above and below , which is also consistent with our numerical implementation of the method .
then @xmath28 can be conveniently found as a solution to the schrdinger equation with the following asymptotic boundary conditions @xmath29 @xmath30 the asymptotic of @xmath31 is easily obtained from the relation @xmath32 therefore , using eqs .
( [ eq6 ] ) and ( [ eq61 ] ) , we have @xmath33 @xmath34 equation ( [ eq6 ] ) together with the lower line of eq .
( [ eq61 ] ) describes the electron incident on the surface , as in the low energy electron diffraction ( leed ) experiment .
interestingly , the wave function of eq .
( [ eq9 ] ) together with the upper line of eq .
( [ eq91 ] ) is a time - reversed leed state .
this kind of function describes the photoelectron ( pe ) final state in the one - step theory of photoemission @xcite .
note that in vacuum it contains both outgoing and incoming beam .
thus , while in leed and pe setups each of these kinds of the wave - functions enters separately , in eels they are present together . for @xmath35 we can write @xmath36 where @xmath37 and @xmath38
. then , finally , eq . ( [ dsdwdw ] ) takes the convenient form @xmath39 where @xmath40 is the surface normalization area . according to eq .
( [ dsdwdw1 ] ) , the external potential applied to our system is @xmath41 therefore , eq .
( [ dsdwdw1 ] ) can be rewritten as @xmath42 where @xmath43 is the potential induced in the system in response to the external charge - density @xmath44 .
to determine @xmath45 , a simplified model of ag surface , introduced by liebsch @xcite , is used , in which only @xmath46-electrons are treated quantum - mechanically through the calculation of their response function , while the influence of @xmath4-electrons is included effectively by the use of a background df @xmath47 comprising the half - space @xmath48 , as schematized in fig .
[ geometry ] .
[ h ! ] then , for the total scalar potential @xmath49 we can write separately in the regions i of @xmath50 and ii of @xmath48 @xmath51 where @xmath52 @xmath53 is the potential of the response of @xmath46-electrons only , and @xmath40 and @xmath54 are constants to be determined from the boundary conditions of the continuity of the tangential component of the electric field and the normal component of the electric displacement vector , which give , respectively , @xmath55 for @xmath56 we can write @xmath57 d z ' d z '' , \end{split } \label{phip}\ ] ] where @xmath58 is the density - response function of @xmath46-electrons only . by further rewriting eq .
( [ phi ] ) as @xmath59 substituting eq .
( [ eq20 ] ) into the right - hand side of eq .
( [ phip ] ) and eq .
( [ phit ] ) into eqs .
( [ eq18 ] ) , we arrive at a closed system of equations for @xmath53 , @xmath40 , and @xmath54 , which is numerically solved on a grid of @xmath26
. then from @xmath53 we obtain @xmath60 by eq .
( [ phit ] ) , @xmath61 by eq .
( [ phi ] ) , and @xmath45 as @xmath62 .
the latter is finally used in eq .
( [ dsdwdw2 ] ) to calculate the eel spectrum .
our calculation of the ground - state of the @xmath46-electrons of the ag ( 111 ) uses the 1d interpolation of the surface and the bulk potential of ref . . a super - cell with the period @xmath63 a.u
was used , which included 31 layers of the model @xmath46-subsystem of ag , the rest occupied with vacuum . the time - dependent density - functional theory ( tddft ) calculation of the density - response function @xmath64 is performed on the level of the random - phase approximation ( rpa ) , i.e. , setting the exchange - correlation kernel @xcite @xmath15 to zero .
then we apply the procedures of secs .
[ rss ] and [ flat ] to account for the response of @xmath4-electrons and to finally obtain the eel spectra .
the edge of the @xmath4-electrons was set at @xmath65 a.u . above the upper atomic layer .
for @xmath66 we take @xmath67= \varepsilon_{\text{ag}}(\omega)+\frac{\omega_{p}^2}{\omega^2 } , \label{epsd}\ ] ] where @xmath68 is the experimental optical df of silver@xcite and @xmath69 is the drude df of @xmath46-electrons with the plasma frequency @xmath70 ev . to construct @xmath44 by eq .
( [ rhoz ] ) , we were solving the schrdinger equation with the asymptotic boundary conditions of eqs .
( [ eq6 ] ) and ( [ eq9 ] ) .
the scattering wave functions were obtained by solving the inverse band - structure problem as explained in ref . and matching the bloch solutions in the crystal to the linear combination of the incident and reflected wave in the vacuum .
the same crystal potential as for the evaluation of @xmath13 was used , with the addition of the absorbing imaginary potential @xmath71 as explained below .
importantly , similar to the low - energy electron diffraction ( leed ) theory @xcite , the inclusion of the optical potential ( op ) into the hamiltonian is necessary for eels theory as well .
this can be understood considering that , without op , electrons having gone an arbitrarily long round - trip into the depth of the sample , would contribute to the spectrum .
since , in the first born approximation , the probability of the bulk energy - loss is proportional to the path - length travelled , this would make the intensity of the bulk losses infinitely high .
the influence of the deep interior of the sample is , however , suppressed , in leed by all the inelastic processes , and in eels by the inelastic processes beyond the first born approximation . in the present calculation @xmath72
was taken to be spatially constant in the solid and zero in vacuum . at @xmath73
ev it was @xmath74 ev for the angle with the normal to the surface of @xmath75 , @xmath76 , @xmath77 , @xmath78 ev for @xmath79 , and @xmath80 ev for @xmath81 . in fig .
[ ext ] @xmath44 and the corresponding @xmath82 are shown for representative values of the parameters of the eels experiment .
[ h ! ] [ h ! ] [ h ! ] in figs .
[ sp1 ] and [ sp2 ] results of calculations of the eel reflection spectra are presented .
they are compared to experimental hreels of the system of 10 monolayers of ag on ( 111 ) surface of ni substrate @xcite . in fig .
[ sp1 ] the theoretical and experimental eel spectra are shown for the primary energy of electrons @xmath73 ev , the angle of incidence @xmath83 , and three values of the angle of scattering @xmath84 of @xmath79 , @xmath76 , and @xmath77 . in fig .
[ sp2 ] the results for the specular geometry with @xmath85 and the same primary energy are presented .
the comparison of the theory with the experiment is reasonably good .
most importantly , the sharp bulk and surface plasmon peaks near 3.7 ev , separately present in the plots of the corresponding energy - loss functions ( fig .
[ sp0 ] ) , are never resolved from each other in our calculations , but they form a joint broadened peak with a contribution from the both types of excitations .
this is in full agreement with the hreels experiments @xcite .
a notable exception from the agreement between the theory and the experiment is the case of @xmath83 and @xmath86 , upper spectrum in fig .
[ sp1 ] , when , surprisingly , the lower - energy plasmon peak , which is strong in the experimental spectrum , is absent in the theoretical one . to examine the latter discrepancy closer , in fig .
[ sp3 ] we plot theoretical spectra for the angle of scattering gradually changing from @xmath76 , when the peak in question is pronounced , to @xmath77 , when this peak disappears .
these results show that the strength of the peak near 3.7 ev decreases systematically when the scattering angle @xmath87 increases .
we note that a similar effect of the disappearance of the 3.7 ev peak with the growing momentum can be observed in the results of the calculations of ref . , performed in the dipole - scattering mode within the same model of @xmath4-electrons .
[ h ] we analyze this tendency in detail in appendix [ prop ] to the conclusion that , although , the background df model is applicable in the higher energy range around 8 ev to account for the bulk , surface , and the multipole plasmons in ag@xcite , it fails for the lower - energy plasmons at larger values of the wave - vector .
obviously , the future theory , which will include @xmath4-electrons from the first principles , will be free from this deficiency . since our calculations based on eq .
( [ dsdwdw ] ) are linear with respect to the interaction between the probe electron and the electronic subsystem of the target ( distorted - wave@xcite with the first born approximation for the inelastic processes ) , the multiple energy losses , e.g. , multiple plasmon excitations , are beyond the capacity of this approach .
nonetheless , especially at higher primary energies , multiple losses can be expected in the experimental spectra . in fig .
[ 8080 ] we plot the theoretical spectra together with the corresponding experimental ones for three different primary energies of 170 , 100 , and 70 ev at the specular geometry of @xmath88 , where we now focus on the higher energy range . while the theoretical lines are rather smooth in this range , the experimental spectra at @xmath89 170 and 100 ev have prominent peaks around 7.6 ev .
considering that ( i ) the positions of these peaks are very close to the twice the energy of the strong single - plasmon peaks around 3.7 ev , ( ii ) their intensities change with @xmath90 consistently with those of the corresponding single - plasmon peaks , and ( iii ) these peaks are present in the experiment while absent in the linear - response based calculations , we are led to the conclusion that these peaks are due to the double - plasmon excitations .
[ h ] to reproduce the multiple plasmons theoretically , a theory of eels beyond the first born approximation is required .
we note that the formal theory of inelastic scattering of a quantum - mechanical particle to the _ second _ born approximation , expressed in terms of the _ quadratic _ density - response function of a target , was constructed in ref . .
the practical implementations of this theory have not , however , been yet developed .
we have revisited the problem of the energy losses by electrons in reflection electron energy - loss spectroscopy with the focus on the role of the probing electrons kinematics .
the inadequacy of the description of eels in terms of the energy - loss functions has been emphasized for the materials where the bulk and the surface features overlap or are close in energy in the excitation spectra .
we have implemented the theory including the effect of the detailed quantum - mechanical motion of the probe electrons during their energy losses , using ag surface as a representative example of a system where the kinematic aspect of the problem is particularly important . since
our primary interest lies in the role of the kinematic effects , for a sophisticated problem of the @xmath4-electrons response we have used a simplified model of the background dielectric function , which has immensely simplified the numerical implementation . as a side effect , although we have found a reasonably good overall agreement with hreels , at strongly off - specular geometries of the experiment , the theory and the measurements disagree .
we have tracked this discrepancy down to the failure of the substitution of the @xmath4-electrons with the background dielectric function to describe the dispersion of the main loss feature of 3.7 ev in ag at larger values of the momentum .
thus , the limits of the applicability of that , otherwise very useful model , have been set .
this difficulty is anticipated to be overcome in the future theory , with all the ingredients included within the _ ab initio _ approach .
another deviation of the theory from experiment we have found can be qualified as an evidence of the consistency of the former rather than its deficiency .
namely , the experiment shows peaks at eel spectra that , by all the evidence , can be attributed to the double - plasmon excitations .
the linear - response theory , which our approach is based on , fundamentally can not account for such losses , and we do not , accordingly , obtain the double - excitation peaks in the calculations .
future implementations of the theory of the quadratic and higher - order response will be able to account for these processes .
vun acknowledges support from the ministry of science and technology , taiwan , grant .
this work was supported by the spanish ministry of economy and competitiveness mineco ( project no . fis2013 - 48286-c2 - 1-p ) .
in this appendix we scrutinize the model of the background df used in this paper to account for the @xmath4-electrons in the dielectric response of silver@xcite . for the sake of maximal clarity , we do this analytically by considering the bulk response . in this case the same model is determined by the df @xmath91 + \varepsilon_{ls}(q,\omega ) \\ & = \varepsilon_d(\omega)+\left [ \varepsilon_{ls}(q,\omega)-1 \right ] , \end{split } \label{epsb}\ ] ] where @xmath68 is the optical experimental df of ag , @xmath66 is given by eq .
( [ epsd ] ) , and @xmath92 is the lindhard s df of the homogeneous electron gas@xcite , taken at the density of the @xmath46-electrons of ag . in fig .
[ bulk ] we plot the energy - loss function with the use of the df of eq .
( [ epsb ] ) for several values of the wave - vector @xmath93 .
the plasmon peak near 3.7 ev weakens with the increase of @xmath93 , until it disappears at @xmath94 0.2 a.u .
this is consistent with the behavior of the df itself plotted in fig .
[ bulke ] . indeed , for @xmath95 0 and 0.1 a.u . , re @xmath96 crosses zero in the corresponding energy range , thereby producing the plasmon peak in the loss - function .
this is not the case any more for @xmath95 0.15 , 0.175 , and 0.2 a.u . , although , for the former two wave - vectors , the peaks in question still persist in the loss - function due to the re @xmath97 _ approaching _ the zero axis ( cf . , ref . ) .
lastly , at @xmath95 0.2 a.u . , re @xmath97 is very far from zero in this @xmath98-range , and no peak in the loss - function can be discerned any more .
the above results are consistent with those of sec .
[ res ] of this paper .
indeed , the main contribution to the @xmath26-component of the wave - vector of the external perturbation can be estimated from eqs .
( [ rhoz ] ) , ( [ eq6 ] ) , and ( [ eq9 ] ) as @xmath99 ( note that @xmath100 and @xmath101 ) .
when @xmath89 40 ev and @xmath102 80@xmath103 , this is equal to 0.123 and 0.259 a.u .
, for @xmath104 75@xmath103 and 70@xmath103 , respectively .
the corresponding values of @xmath105 are 0.117 and 0.160 a.u .
, respectively . then @xmath106^{1/2}$ ] are 0.170 and 0.304 a.u . , respectively , explaining the presence of the plasmon near 3.7 ev in the theoretical spectra in the former and its absence in the latter case .
for the specular geometry in fig .
[ sp2 ] , the corresponding value is @xmath95 0.09 a.u . ,
consistent with the strong theoretical lower - energy plasmon peak in this figure .
the same argument holds for the spectra in fig .
[ 8080 ] . experimentally , however , this prediction of the model is not supported , as can be seen in fig .
[ sp1 ] , upper experimental spectrum : the plasmon near 3.7 ev persists in the measurements on ag at larger wave - vectors .
we , therefore , can conclude that , at larger wave - vectors , the response of @xmath4-electrons can not be realistically substituted with a wave - vector independent dielectric function , when it concerns the lower - energy plasmon in ag , while this model is quite successful in the higher energy range , where the spectra are dominated by the response of @xmath46-electrons outside the @xmath4-electrons background@xcite .
32ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty \doibase http://dx.doi.org/10.1063/1.1707491 [ * * , ( ) ] @noop _ _ ( , , ) link:\doibase 10.1103/physrevb.77.233406 [ * * , ( ) ] @noop * * , ( ) @noop _ _ ( , , ) link:\doibase 10.1103/physrev.106.874 [ * * , ( ) ] link:\doibase 10.1103/physrevb.1.203 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.64.44 [ * * , ( ) ] link:\doibase 10.1103/physrevb.57.3803 [ * * , ( ) ] @noop _ _ ( , , ) @noop * * , ( ) \doibase http://dx.doi.org/10.1016/0039-6028(95)00090-9 [ * * , ( ) ] @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1103/physrevb.59.9866 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0039-6028(01)00784-1 [ * * , ( ) ] link:\doibase 10.1007/s11468 - 013 - 9587-x [ * * , ( ) ] , ed .
, @noop _ _ ( , , ) link:\doibase 10.1103/physrevb.52.14219 [ * * , ( ) ] link:\doibase 10.1103/physrevb.10.4932 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0039-6028(99)00668-8 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.56.12874 [ * * , ( ) ] link:\doibase 10.1103/physrevb.66.235403 [ * * , ( ) ] link:\doibase 10.1103/physrevb.52.14947 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/0167-5729(95)00004-6 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0039-6028(97)00278-1 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.155405 [ * * , ( ) ] link:\doibase 10.1103/physrevb.65.094303 [ * * , ( ) ] @noop * * , ( )
|
inelastic scattering of electrons incident on a solid surface is determined by the two properties : ( i ) electronic response of the target system and ( ii ) the detailed quantum - mechanical motion of the projectile electron inside and in the vicinity of the target .
we emphasize the equal importance of the second ingredient , pointing out the fundamental limitations of the conventionally used theoretical description of the electron energy - loss spectroscopy ( eels ) in terms of the `` energy - loss functions '' .
our approach encompasses the dipole and impact scattering as specific cases , with the emphasis on the quantum - mechanical treatment of the probe electron .
applied to the high - resolution eels of ag surface , our theory largely agrees with recent experiments , while some instructive exceptions are rationalized .
|
there has been much interest lately in the behavior and control of quantum linear systems , particularly as these are amenable to transfer matrix function techniques . in this note
, we wish to exploit the structural features of quantum markovian models to construct an analogue of the transfer matrix function for non - linear systems .
coming from the classical direction there has been fruitful application of operator techniques to control systems in recent years @xcite employing for instance characteristic functions techniques , multi - analytic operators and commutant lifting methods .
here we introduce a natural characteristic operator function associated with a quantum markov ( or slh ) model . as in standard quantum mechanics ,
the model is formulated by representing physical quantities ( observables ) as self - adjoint operators on a hilbert space . the quantum mechanical system ( plant )
will have underlying hilbert space @xmath0 while the input will be a continuous quantum field with hilbert space @xmath1 .
the coupled model will have joint hilbert space @xmath2 , which is also the space on which the output observables act .
the input - plant - output model can be summarized as @xmath3 where @xmath4 is an arbitrary plant observable , @xmath5 is a component of the input field , and @xmath6 is the unitary entangling the plant with the portion of the bath that has interacted with it over the time period @xmath7 $ ] . in the following
we shall specify to a category of model where @xmath8 is a unitary family of operators on @xmath9 , satisfying a differential equation of the form @xcite @xmath10 formally , we can introduce input process @xmath11 for @xmath12 satisfying singular commutation relations of the form @xmath13=\delta _ { ij}\delta ( t - t^{\prime } ) $ ] , so that the processes appearing in ( [ hp_qsde ] ) are @xmath14 more exactly , the are rigorously defined as creation and annihilation field operators on the boson fock space @xmath1 over @xmath15 . the increments in ( [ hp_qsde ] ) are understood to be future pointing in the ito sense .
we have the following table of non - vanishing products @xmath16 necessary and sufficient conditions for unitarity @xcite are that we can collect the coefficients of ( [ hp_qsde ] ) to form a triple @xmath17 , which we call the _ hudson - parthasarathy ( hp ) parameters _ , consisting of a unitary matrix @xmath18 , a column vector @xmath19 , and a self - adjoint operator @xmath20 , @xmath21 , \quad l=\left [ \begin{array}{c } l_{1 } \\ \vdots \\
l_{n } \end{array } \right ] , \end{aligned}\ ] ] with @xmath22 are all operators on @xmath0 , and where @xmath23 it has become fashionable to refer to this , plus the related feedback network models @xcite , as the slh formalism .
we shall refer to @xmath6 as the unitary determined by the coupling parameters @xmath24 .
in differential form , the input - plant - output model then becomes @xcite * plant dynamical ( heisenberg ) equation : * @xmath25 here @xmath26 + \frac{1}{2}\sum_{i}\left [ l_{i}^{\ast } , x\right ] l_{i}-i\left [ x , h\right ] , \quad \text{(the lindbladian ! ) } , \\ \mathscr{m}_i x & = & s_{ji}^\ast [ x , l_j ] , \\ \mathscr{n}_i x & = & [ l_k^\ast , x ] s_{ki } , \\ \mathscr{s}_{ik } x & = & s_{ji}^\ast xs_{jk } - \delta_{ik } x .\end{aligned}\ ] ] * input - output relations : * @xmath27 if we specify to a system of quantum mechanical oscillators with modes @xmath28 satisfying canonical commutation relations
@xmath29 = \delta _ { \alpha \beta } \end{aligned}\ ] ] then we obtain a linear dynamical model with the prescription @xmath30 specifically , the plant dynamics and input - output relations are affine linear in the mode variables @xmath31 : @xmath32 where , setting @xmath33 \in \mathbb{c}^{n\times n}$ ] , @xmath34 \in \mathbb{c}^{n\times m}$ ] and @xmath35 \in \mathbb{c}^{m\times m}$ ] , we have @xmath36 in turn , a model having this specific structure is said to be physically realizable . the transfer matrix associated with the linear dynamics
is then defined to be @xcite @xmath37 ( s)\triangleq d+c\left ( si - a\right ) ^{-1}b , \label{eq : t}\end{aligned}\ ] ] and there exists a well - established literature developing control theory from analysis of these functions .
the definition here leads to transfer functions that are positive real functions of the complex variable @xmath38 , and they model passive systems .
the generalization to active linear models , which we do not need here , is given in @xcite .
in the mathematical formulation of open quantum markov systems , a natural role is played by the model matrix , introduced in @xcite , @xmath39 .
\label{model_matrix}\ ] ] we now use it as the basis for the definition of an operator - valued generalization of the characteristic function . for given @xmath24
we define the corresponding characteristic operator by @xmath40 ( s)=s - l(si+\frac{1}{2}l^{\ast } l+ih)^{-1}l^{\ast } s. \label{characteristic op}\ ] ] we shall often write @xmath41 for emphasis .
the characteristic operator @xmath42 is a bounded operator for re@xmath43 .
for all @xmath44 , such that @xmath45 lies in the resolvent set of @xmath46 ( that is , whenever @xmath47 is invertible ) , we have @xmath48i@xmath49 well - defined and unitary : @xmath50 the proof follows mutatis mutandis of the proof of an analogous result in @xcite .
suppose that we have no coupling @xmath51 then the characteristic operator is @xmath52 , constant .
this is true even if @xmath20 is non - zero . without coupling ,
we can not infer anything about the system hamiltonian . for the model considered in subsection [ sec : lqm ]
we have @xmath53 where @xmath54 $ ] .
for the @xmath55 case we have explicitly @xmath56 where @xmath57 is the number operator for the single mode .
in fact , we see that @xmath58 where @xmath59 is the transfer function ( [ eq : t ] ) .
the same vacuum expectation is obtained for the cases @xmath60 greater than one .
a simple example is a qubit system with master equation @xmath61\end{aligned}\ ] ] where @xmath62 and we set @xmath63 , @xmath64 .
this models a qubit in a thermal bath with @xmath65 being the equilibrium occupancy of the state @xmath66 in the presence of the oscillation @xmath67 .
we shall take the scattering to be by a polarization - dependent phase @xmath68 in the @xmath69-basis @xmath70 , |\downarrow
\rangle = \left [ \begin{array}{c } 0 \\ 1 \end{array } \right ] , \,k=\left [ \begin{array}{cc } -\frac{1}{2 } \gamma \left ( n+1\right ) -i\omega & 0 \\ 0 & -\frac{1}{2 } \gamma n + i\omega \end{array } \right ] $ ] and the characteristic operator explicitly is @xmath71 .\end{aligned}\ ] ] the characteristic operator is diagonal in the basis @xmath72 , but this would no longer be true if @xmath73 \neq 0 $ ] .
we consider a model of a cavity mode @xmath74 between a fixed leaky mirror and a perfect mirror with quantum mechanical position @xmath75 , see fig .
[ fig : optomech ] .
the slh model takes the form @xmath76 where @xmath77 is the damping to the input field at the leaky mirror , @xmath78 is the cavity detuning , @xmath79 is the harmonic frequency of the mirror , and @xmath80 is the coupling strength associated with mirror - mode interaction .
note that the interaction @xmath81 couples the position of the mirror to the cavity mode photon number in accordance with the notion of radiation pressure .
this is a standard opto - mechanical model and we obtain the langevin equations @xmath82
a simplifying assumption is that the mechanical processes are much slower than the optical ones , in which case we set @xmath83 .
the characteristic operator in this case is @xmath84 this is recognizable as the characteristic operator of a quantum linear passive system as in ( [ eq : qlps ] ) , but with the operator @xmath85 taking the form @xmath86 .
that is , @xmath85 is no longer scalar valued , but depends explicitly on the position observable @xmath4 of the mirror .
note that @xmath85 is still strictly hurwitz since @xmath4 is self - adjoint .
we remark that position dependent transfer functions have been proposed for single photon input - output models for this type of model with one - particle fields related by @xcite @xmath87 and here the transfer function corresponds to the partial trace of @xmath88 over the vacuum state of the cavity .
the characteristic operator admits the following `` all - pass '' representation : @xmath89 where @xmath90 .
this is proved in @xcite , and we recall briefly the proof . * proof * an application of the woodbury matrix identity @xcite @xmath91 shows that @xmath92 substituting into @xmath93 then gives the above relation after some straightforward algebra . @xmath94
note that @xmath95 for real @xmath96 , so that we could alternatively have deduced unitarity by a cayley transformation argument .
suppose that the model parameters satisfy the condition @xmath97 \equiv 0
$ ] , then the characteristic operator takes the form @xmath98 the condition @xmath97 \equiv 0 $ ] arises as the qnd condition for measurement disturbance in the sense of braginsky @xcite . [
ref : passive_equiv ] for a finite - dimensional system , say with hilbert space @xmath99 , we may fix an orthonormal basis of @xmath100 vectors for @xmath101 . in this representation
, we may describe @xmath20 as an @xmath102 matrix which we denote as @xmath103 .
the coupling operator @xmath19 is then a column vector of @xmath104 operators , each represented as an @xmath105 matrix , so that @xmath19 may be represented as an @xmath106 matrix which we denotes as @xmath107 . in this manner
, @xmath18 becomes a complex valued matrix @xmath108 .
we then have the equivalence @xmath109 ( s ) \equiv \left [ \begin{tabular}{l|l } $ a$ & $ b$ \\
\hline $ c$ & $ d$\end{tabular } \right ] ( s ) , \label{eq : equiv}\ ] ] where @xmath110 and @xmath111 . in this
was we realise the characteristic function as the transfer operator of a linear passive system @xmath112 , structurally similar to those considered in subsection [ sec : lqm ] , with a state space of @xmath100 dimensions and @xmath113 inputs .
we show now that the characteristic operator function can be described in terms of the coefficient operators in the stratonovich qsde .
the stratonovich differential is defined using the midpoint rule convention which leads to the algebraic rule @xcite @xmath114 it can then be shown that the stratonovich form of the qsde ( [ hp_qsde ] ) takes the form @xmath115 with @xmath116 , @xmath117 and @xmath118 .
it is convenient to collect all the coefficients into a ( hermitean ) matrix @xmath119 \label{eq : model e}\ ] ] the components of @xmath120 are related to the @xmath17 by the transformation @xcite : @xmath121_{1\leq i , j\leq n}$ ] is the cayley transform @xmath122_{1\leq i , j\leq n}$ ] , @xmath123 and therefore @xmath18 is unitary , while @xmath124 with @xmath20 self - adjoint .
note that the operator @xmath125 is then given by @xmath126 we may write the characteristic operator in terms of the coefficients making up the stratonovich matrix @xmath120 ( [ eq : model e ] ) as @xmath127 * proof * we have explicitly that @xmath128 the woodbury matrix identity @xcite with @xmath129 shows that @xmath130 rearranging for then gives the desired result . @xmath94 note that we have the correct limit @xmath131 .
suppose that we have @xmath132 , @xmath133 and @xmath134 independent of @xmath135 , then the associated transfer operator @xmath136 has the well - defined limit @xmath137 provided that @xmath138 is invertible . here
@xmath139 with @xmath140 .
this limit , which corresponds physically to high - energy and strong damping , leads to a purely scattering model but with a shifted scattering matrix @xmath141 .
we shall study more general examples of this type of scaling leading to slh models with nontrivial couplings @xmath142 and and hamiltonians @xmath143 .
as we have seen , the characteristic operator for a system with underlying hilbert space @xmath0 with @xmath104 inputs is a function taking values in @xmath144 , the set of @xmath145 matrices with entries in @xmath146 , the bounded operators on @xmath0 .
let @xmath85 and @xmath147 be models with the same input dimension @xmath104 and having coefficient parameters @xmath148 and @xmath149 respectively .
we may _ cascade _ the systems by feeding the output of @xmath85 and input to @xmath147 and in the instantaneous feedforward limit we get the model @xmath150 on @xmath151 with parameters given by the series product , see @xcite and @xcite , @xmath152 . in this case
we typically have @xmath153 ( here we employ the shorthand @xmath154 for the matrix with @xmath155-entries @xmath156 _ { jl}\otimes \left [ s_{a}\right ] _ { lk}$ ] , etc . )
thus characteristic function for cascaded systems is not naturally the product of their characteristic operators . for cascaded classical systems ,
the state spaces take the form @xmath157 and @xmath158 so that the combined state space is the direct sum @xmath159 .
the rule in quantum theory is that the combined hilbert state space for the cascaded systems is the tensor product and not the direct sum .
( note that for quantum linear systems , the hilbert space is the fock space @xmath160 over @xmath161 , and for combined linear systems we have @xmath162 , which is the usual rule for fock spaces @xcite . in this way
the usually cascade rule re - emerges for the corresponding transfer functions @xcite . ) with this observation , we see that model reduction techniques based around the characteristic operator should involve direct sum decompositions , say @xmath163 into orthogonal subspaces .
each of the coefficients @xmath164 , etc .
, can be represented as @xmath165\end{aligned}\ ] ] where @xmath166 maps from @xmath167 to @xmath168 .
the characteristic operator may similarly be decomposed as @xmath169\end{aligned}\ ] ] with @xmath170 _ { \mathsf{de}}l_{\mathsf{ce}}^{\ast } \right\ } s_{\mathsf{cb}}.\end{aligned}\ ] ] ( here we have the convention that repeated indices
are summed over the range and .
we also adopt the notation that @xmath171 is the @xmath172-valued output @xmath173 , input @xmath135 entry of @xmath18 , while @xmath174 is the component of @xmath18 mapping from @xmath167 to @xmath168 , which is an @xmath145 matrix of maps from @xmath167 to @xmath168 .
similarly @xmath175 is the @xmath104-column vector of maps from @xmath167 to @xmath176 . ) using the schur - feshbach identity we may write the resolvent @xmath177 as @xmath178 ^{-1}=\left [ \begin{array}{ll } \widehat{\delta}_{\mathsf{11}}\left ( s\right ) & \widehat{\delta}_{\mathsf{12}}\left ( s\right ) \\ \widehat{\delta}_{\mathsf{21}}\left ( s\right ) & \widehat{\delta}_{\mathsf{22}}\left ( s\right ) \end{array } \right]\end{aligned}\ ] ] where , introducing @xmath179 and @xmath180 , we have @xmath181 the blocks of the characteristic operator partitioned with respect to the direct sum @xmath182 are then @xmath183 given the direct sum @xmath182 , we say that orthogonal subspaces @xmath184 and @xmath185 are decoupled if the characteristic operator takes the block diagonal form @xmath186 , \end{aligned}\ ] ] that is @xmath187 and @xmath188 . we note that if @xmath189 is a unitary on the system space , then the basic unitary rotation behaviour for characteristic operators is @xmath190 however we note following result , which is easily derived . for any unitary @xmath189 on the plant hilbert space , the hp parameters @xmath191 generate the same characteristic operator as @xmath24 .
more generally we have the following invariance property of the characteristic function : @xmath192 = \left [ \begin{tabular}{l||l } $ v^{\ast } av$ & $ v^{\ast } b$ \\
\hline\hline $ cv$ & $ d$\end{tabular } \right ] .\end{aligned}\ ] ] therefore , while the characteristic operator is a quantum object - for @xmath104 inputs , it is an @xmath145 matrix with entries that are operators on the plant space - its dependence on the plant operators is only up to a unitary equivalence as outlined in the proposition .
let @xmath24 be given hp parameters for a fixed plant hilbert space @xmath0 .
if @xmath193 are hp parameters for a proper subspace @xmath194 of the plant space then we say that @xmath195 is a reduced model of @xmath24 if we have @xmath196 , \end{aligned}\ ] ] with respect to the decomposition @xmath197 .
a reduced model is minimal if it allows no further model reduction . as a simple toy model ,
let us consider a two - level atom with ground and excited states states @xmath198 and @xmath199 .
we fix the open system as being a single input model with @xmath200 , @xmath201 and hamiltonian @xmath202 where @xmath203 , etc . here
@xmath204 is interpreted as a detuning parameter and @xmath205 as the amplitude of a drive .
both the detuning and amplitude are assume to be large , which corresponds to the limit @xmath206 .
the characteristic operator for the two - level system is then given by @xmath207 - \left [ \begin{array}{cc } \sqrt{\gamma } & 0 \\ \sqrt{\kappa } & -\sqrt{\gamma } \end{array } \right ] \left [ \begin{array}{cc } s+\frac{1}{2}(\gamma + \kappa ) + ik^{2}\delta + i \omega_0 & - \frac{1}{2 } \sqrt{\kappa \gamma } + ik\beta \\ - \frac{1}{2}\sqrt{\kappa \gamma } + ik\beta ^{\ast } & s+\frac{1}{2}\gamma + i\omega _ { 0 } \end{array } \right ] ^{-1 } \left [ \begin{array}{cc } \sqrt{\gamma } & \sqrt{\kappa } \\
0 & -\sqrt{\gamma } \end{array } \right ] , \nonumber\end{aligned}\ ] ] which can be calculated explicitly as a @xmath208 matrix whose entries are rational polynomials in @xmath38 of degree 2 .
what is of interest here is that for large @xmath135 the characteristic operator takes the limit form @xmath209\ ] ] where @xmath210 , where we have the shifted frequency @xmath211 .
the limit model corresponds to the transfer function of a linear system with a single degree of freedom having the damping @xmath77 and frequency @xmath212 .
what is happening in this limit is that the excited state plays an increasingly negligible role in the model as its decay rate starts to increase : the limit is a reduced model , however , with a shift of the frequency . as a next example
, we consider a qubit driven by three input fields , with @xmath213 , h=\delta \sigma ^{\ast } \sigma -i\sqrt{\kappa _ { 1}}\left ( \alpha \sigma ^{\ast } -\alpha ^{\ast } \sigma \right)\ ] ] where @xmath214 are the lowering and raising operators , @xmath78 is a fixed detuning and @xmath215 the amplitude of a drive field .
the characteristic operator now takes the form @xmath216 $ ] where we have the components @xmath217 for @xmath218 and where @xmath219 . in the special case
where @xmath220 , there is a zero - pole cancellation .
we begin by considering the description of perturbations to open system models in terms of their characteristic operators .
we discuss regular perturbations first for completeness : suppose we have a model @xmath221 which is a perturbation of solvable model @xmath222 with @xmath223 so that @xmath224 .
the resolvents @xmath225 and @xmath226 are then related by @xmath227 . for bounded perturbation
@xmath189 we have the neumann series @xmath228 so that the characteristic operators are related by @xmath229 this formula will be valid for suitably small constants @xmath230 . in principle
this formula may be useful for perturbative approaches to system modelling .
our main focus , however , will be singular perturbations corresponding to adiabatic elimination . there exist a large body of results under the name of _ adiabatic elimination _ applicable to open quantum models .
a universal mathematical approach has been developed by bouten , silberfarb and van handel @xcite .
we formulate their presentation in a slightly different language . essentially , the common element in adiabatic elimination problems is that the system space can be decomposed into a fast space , which is viewed as increasingly strongly coupled to the bath , and a slow space .
specifically we assume a decomposition of the system space as @xmath231 a recent example of this is the approximate qubit regime for nonlinear optical cavities @xcite .
the coupling parameters are then taken as @xmath232 where @xmath135 is a strength parameter which we eventually take to be large . for a given operator @xmath4 on @xmath0 ,
we write @xmath233 .\end{aligned}\ ] ] more generally we use this notation when @xmath4 is an array of operators on @xmath0 . the projections onto @xmath234 and @xmath235
are denoted by @xmath236 $ ] and @xmath237 $ ] respectively . 1 .
the coupling operator takes the form @xmath238 where @xmath239 , that is , @xmath240 ; \label{eq : l(k)}\end{aligned}\ ] ] 2 .
the hamiltonian takes the form @xmath241 where @xmath242 and @xmath243 , that is , @xmath244 ; \end{aligned}\ ] ] 3 . in the expansion @xmath245
we require that the operator @xmath246 be invertible on @xmath247 . employing a repeated index summation convention over the index range @xmath248 from now on , we find that the operator @xmath249 has components @xmath250 with respect to the slow - fast block decomposition . likewise @xmath251 , \\ z & \equiv & \left [ \begin{array}{ll } 0 & -\frac{1}{2}l_{\mathsf{cs}}^{\left ( 0\right ) \ast } l_{\mathsf{cf}}^{\left ( 1\right ) } -ih_{\mathsf{sf}}^{\left ( 1\right ) } \\
-\frac{1}{2}l_{\mathsf{cf}}^{\left ( 1\right ) \ast } l_{\mathsf{cs}}^{\left ( 0\right ) } -ih_{\mathsf{fs}}^{\left ( 1\right ) } & -\frac{1}{2}l_{\mathsf{cf}}^{\left ( 0\right ) \ast } l_{\mathsf{cf}}^{\left ( 1\right ) } -\frac{1}{2}l_{\mathsf{cf}}^{\left ( 1\right ) \ast } l_{\mathsf{cf}}^{\left ( 0\right ) } -ih_{\mathsf{ff}}^{\left ( 1\right ) } \end{array } \right ] . \label{k relations}\end{aligned}\ ] ] in particular , we note the identities @xmath252 in an adiabatic elimination problem , the coupling parameters @xmath253 lead to the associated characteristic operator @xmath254 ^{-1}l\left ( k\right ) ^{\ast } s .\end{aligned}\ ] ] [ lem_limit ] let @xmath255 be a matrix parametrized by scalar @xmath135 of the form @xmath256\ ] ] with @xmath257 invertible .
then we have the limit @xmath258 \left [ s+m\left ( k\right ) \right ] ^{-1}\left [ \begin{array}{cc } 1 & 0 \\ 0 & k \end{array } \right ] = \left [ \begin{array}{ll } \dfrac{1}{s+\widehat{m}_{11 } } & -\dfrac{1}{s+\widehat{m}_{11}}m_{12}\dfrac{1}{m_{22 } } \\
-\dfrac{1}{m_{22}}m_{21}\dfrac{1}{s+\widehat{m}_{11 } } & \quad \dfrac{1}{m_{22}}+\dfrac{1}{m_{22}}m_{21}\dfrac{1}{s+\widehat{m}_{11}}m_{12}\dfrac{1}{m_{22 } } \end{array } \right ] .\ ] ] * proof * see appendix [ app : t_lim ] . @xmath94 [ prop : limit ] in the situation where the @xmath259 and @xmath260 are bounded operators for each @xmath135 fixed , the characteristic operator has the strong limit @xmath261 for re @xmath262 , where we have @xmath263 \frac{1}{s-\widehat{k}_{\mathsf{ss}}}\left [ l_{\mathsf{cs}}^{\left ( 0\right ) \ast } -z_{\mathsf{sf}}\frac{1}{a_{\mathsf{ff}}}l_{\mathsf{cf}}^{\left ( 1\right ) \ast } \right ] \right\ } s_{\mathsf{cb}}.\ ] ] where @xmath264 * proof * this is a corollary to lemma [ lem_limit ] . in this case
we have the limit @xmath258 \left [ s - k\left ( k\right ) \right ] ^{-1}\left [ \begin{array}{cc } 1 & 0 \\ 0 & k \end{array } \right ] = \left [ \begin{array}{ll } \frac{1}{s-\widehat{k}_{\mathsf{ss } } } , & -\frac{1}{s-\widehat{k}_{\mathsf{ss}}}z_{\mathsf{sf}}\frac{1}{a_{\mathsf{ff } } } \\ -\frac{1}{a_{\mathsf{ff}}}z_{\mathsf{fs}}\frac{1}{s-\widehat{k}_{\mathsf{ss } } } , & \quad -\frac{1}{a_{\mathsf{ff}}}+\frac{1}{a_{\mathsf{ff}}}z_{\mathsf{fs}}\frac{1}{s-\widehat{k}_{\mathsf{ss}}}z_{\mathsf{sf}}\frac{1}{a_{\mathsf{ff } } } \end{array } \right ] .\ ] ] @xmath94 [ prop : limit top ] the limit characteristic operator is given by @xmath265 = \widehat{s}-\widehat{l}\left ( s+\frac{1}{2}\widehat{l}^{\ast } \widehat{l}+i\widehat{h}\right ) ^{-1}\widehat{l}^{\ast } \widehat{s}.\end{aligned}\ ] ] where the parameters @xmath266 are defined by @xmath267 , \;\widehat{l}=\left [ \begin{array}{cc } \widehat{l}_{\mathsf{s } } & 0 \\ \widehat{l}_{\mathsf{f } } & 0 \end{array } \right ] , \;\widehat{h}=\left [ \begin{array}{cc } \widehat{h}_{\mathsf{ss } } & 0 \\ 0 & 0 \end{array } \right ] , \end{aligned}\ ] ] with @xmath268 * proof * see appendix [ app : t_form ] .
@xmath94 we may impose additional constraints @xmath269 to ensure that limit dynamics excludes the possibility of transitions that terminate in any of the fast states . in this case @xmath270 is unitary .
if additionally ( [ eqn_extra_conditions ] ) holds , then the slow and fast subspaces are decoupled : @xmath271\end{aligned}\ ] ] where @xmath272 .\end{aligned}\ ] ] * proof * this follows directly from @xmath273 $ & $ -\left [ \begin{array}{cc } \widehat{l}_{\mathsf{s}}^{\ast } \widehat{s}_{\mathsf{ss } } & 0 \\ 0 & 0 \end{array } \right ] $ \\ \hline\hline $ \left [ \begin{array}{cc } \widehat{l}_{\mathsf{s } } & 0 \\ 0 & 0 \end{array } \right ] $ & $ \left [ \begin{array}{cc } \widehat{s}_{\mathsf{ss } } & 0 \\ 0 & \widehat{s}_{\mathsf{ff } } \end{array } \right ] $ \end{tabular } \right ] .\end{aligned}\ ] ] @xmath274 the convergence of the characteristic operator is not sufficient to guarantee the convergence of the corresponding unitary processes . in paper @xcite the extra condition ( [ eqn_extra_conditions ] ) is required .
suppose we are given a sequence of bounded operator parameters @xmath253 satisfying the assumptions in equation ( [ eqn_extra_conditions ] ) .
then @xmath275 converges strongly to @xmath276 , that is @xmath277 for all @xmath278 with @xmath279 .
the restriction to bounded coefficients was lifted in a subsequent publication @xcite .
it is possible to consider more specific limits which may exist in favourable circumstances .
foe instance , the all - pass form will lead to the scaled @xmath280-function @xmath281 which will converge provided @xmath282 is invertible on the slow space . in this case
it happens the limit is well - defined and given by @xmath283 \left [ \begin{array}{ll } \dfrac{1}{s+i\tilde{h}_{\mathsf{ss } } } & -\dfrac{1}{s+i\tilde{h}_{\mathsf{ss}}}h_{\mathsf{sf}}^{\left ( 1\right ) } \dfrac{1}{h_{\mathsf{ff}}^{\left ( 2\right ) } } \\ -\dfrac{1}{h_{\mathsf{ff}}^{\left ( 2\right ) } } h_{\mathsf{fs}}^{\left ( 1\right ) } \dfrac{1}{s+i\tilde{h}_{\mathsf{ss } } } & \quad -i\dfrac{1}{h_{\mathsf{ff}}^{\left ( 2\right ) } } + \dfrac{1}{h_{\mathsf{ff}}^{\left ( 2\right ) } } h_{\mathsf{fs}}^{\left ( 1\right ) } \dfrac{1}{s+i\tilde{h}_{\mathsf{ss}}}h_{\mathsf{sf}}^{\left ( 1\right ) } \dfrac{1}{h_{\mathsf{ff}}^{\left ( 2\right ) } } \end{array } \right ] \left [ \begin{array}{ll } 0 & l_{\mathsf{sf}}^{\left ( 1\right ) } \\ 0 & l_{\mathsf{ff}}^{\left ( 1\right ) } \end{array } \right ] ^{\ast } \\ & = & \left [ \begin{array}{l } l_{\mathsf{sf}}^{\left ( 1\right ) } \\ l_{\mathsf{ff}}^{\left ( 1\right ) } \end{array } \right ] \left ( -i\dfrac{1}{h_{\mathsf{ff}}^{\left ( 2\right ) } } + \dfrac{1}{h_{\mathsf{ff}}^{\left ( 2\right ) } } h_{\mathsf{fs}}^{\left ( 1\right ) } \dfrac{1}{s+i\tilde{h}_{\mathsf{ss}}}h_{\mathsf{sf}}^{\left ( 1\right ) } \dfrac{1}{h_{\mathsf{ff}}^{\left ( 2\right ) } } \right ) \left [ \begin{array}{ll } l_{\mathsf{sf}}^{\left ( 1\right ) \ast } & l_{\mathsf{ff}}^{\left ( 1\right ) \ast } \end{array } \right ] , \end{aligned}\ ] ] with @xmath284 .
we shall refer to this a the existence of a limit in all pass .
as we have seen , however , the general limit may exists even when the hamiltonian is zero .
more robust however , is the limit formulated in terms of the stratonovich form , where we have suitably - scaled coefficients @xmath285 and we use the stratonovich form ( [ eq : t_strat ] ) along with lemma [ lem_limit ] .
we note the inverse relations @xmath286 as @xmath18 is required to be @xmath135-independent , the same must be true for @xmath287 .
for convenience , we will fix the decompositions as @xmath288 and assume that @xmath289 is block diagonal : @xmath290 .\end{aligned}\ ] ] taking the form ( [ eq : l(k ) ] ) for @xmath259 , it follows that @xmath291 , \end{aligned}\ ] ] with @xmath292 ( no summation ! ) , for @xmath293 or @xmath294 .
it follows that in this case @xmath295 \equiv \left [ \begin{array}{cc } e_{00}^{\left ( \mathsf{ss}\right ) } & ke_{00}^{\left ( \mathsf{sf}\right ) } \\ ke_{00}^{\left ( \mathsf{fs}\right ) } & k^{2}e_{00}^{\left ( \mathsf{ff}\right ) } \end{array } \right ] \end{aligned}\ ] ] which is again of the same form of the general matrix appearing in lemma [ lem_limit ] .
provided that the self - adjoint term @xmath296 is invertible on @xmath247 , the limit for the stratonovich expression exists and will agree with the previous limits .
we omit the more general situation where @xmath289 is not block diagonal as it is more complicated and not very enlightening .
in this section we describe how the unitary process @xmath6 can alternatively be viewed as dirac picture unitaries relating a ( singularly ) perturbed hamiltonian dynamics to a free hamiltonian dynamics .
let @xmath297 and @xmath298 be strongly continuous one - parameter groups , that is @xmath299 and @xmath300 , then we may view @xmath189 as a perturbed dynamics with respect to the free dynamics of @xmath301 by transforming to the interaction picture via the wave operator @xmath302 physically @xmath6 transforms to the dirac picture @xcite .
it inherits unitarity and strong continuity , but does not form a group . instead
we have the so - called _ cocycle property _ @xmath303 where @xmath304 . by stone
s theorem , both @xmath301 and @xmath189 possess self - adjoint ( hamiltonian ) infinitesimal generators @xmath305 and
@xmath143 respectively : @xmath306 , and @xmath307we say that @xmath143 is a regular perturbation of @xmath305 if @xmath308 defines an operator with dense domain . in this case , @xmath6 will be strongly differentiable and @xmath309 where the time - dependent hamiltonian is @xmath310 . in situations
where @xmath311 is not densely defined , we will have a singular perturbation and @xmath6 will not generally be strongly differentiable . the quantum input processes
@xmath312 may be view these processes a as singular operators acting formally on the hilbert space with the fock space @xmath1 over @xmath313 .
for @xmath314 , we have a well - defined amplitude @xmath315 which is completely symmetric under interchange of the @xmath100 pairs of labels @xmath316 , and this represent the amplitude to have @xmath100 quanta with a particle of type @xmath317 at @xmath318 , particle of type @xmath319 at @xmath320 , etc .
we have the following resolution of identity on @xmath1@xmath321 the annihilator input process @xmath312 is then defined almost everywhere as @xmath322 the annihilation operators , together with their formal adjoints the creator operators @xmath323 satisfy the singular canonical commutation relations @xmath324=\delta _ { ij}\delta ( t - s).$ ] let us introduce the following operator on the fock space @xmath325 which is the second quantization of the one - particle operator @xmath326 .
this is clearly a self - adjoint operator and the unitary group @xmath327 it generates is just the time shift : @xmath328 the free evolution @xmath329 will translate the input processes in time : @xmath330 .
the stochastic process @xmath6 is strongly continuous , but due to the presence of the noise fields @xmath331 and @xmath332 is not typically strongly differentiable .
here we see that the local interaction @xmath311 is a singular perturbation of the generator of the time - shift @xmath333 .
we remark that nevertheless @xmath6 is a @xmath334-cocycle and that if we now define @xmath298 by @xmath335 then @xmath298 is a strongly continuous unitary group and therefore admits an infinitesimal generator @xmath143 .
surprising as it may seem , the quantum stochastic process @xmath6 may be considered as the wave - operator for a quantum dynamics with hamiltonian @xmath143 with respect to the free dynamics of the time shift generated by @xmath305 .
the relation @xmath336 however has only a formal meaning as the @xmath311 is singular with respect to @xmath305 .
it has been a long standing problem to characterize the associated hamiltonian @xmath143 for slh models @xcite .
the major breakthrough came in 1997 when a.n .
chebotarev solved this problem for the class of quantum stochastic evolutions satisfying hudson - parthasarathy differential equations with bounded commuting system coefficients @xcite .
his insight was based on scattering theory of a one - dimensional system with a dirac potential , say , with formal hamiltonian @xmath337 describing a one - dimensional particle propagating along the negative @xmath338-axis with a delta potential of strength @xmath339 at the origin .
( in chebotarev s analysis the @xmath340-function is approximated by a sequence of regular functions , and a strong resolvent limit is performed . )
the mathematical techniques used in this approach were subsequently generalized by gregoratti @xcite to relax the commutativity condition .
more recently , the analysis has been further extended to treat unbounded coefficients @xcite .
independently , several authors have been engaged in the program of describing the hamiltonian nature of quantum stochastic evolutions by interpreting the time - dependent function @xmath341 as being an expression involving quantum white noises satisfying a singular ccr @xcite .
this would naturally suggest that @xmath311 should be interpreted as a sesquilinear expression in these noises at time @xmath342 .
the generator of the free dynamics @xmath343 is not semi - bounded and the @xmath340-perturbation is viewed as a singular rank - one perturbation
. here methods introduced by albeverio and kurasov @xcite may be employed to construct self - adjoint extensions of such models , which we show in the next section for a wave on a 1-d wire .
the form of the hamiltonian @xmath143 is known to be @xcite @xmath344 on the domain of suitable functions satisfying the boundary condition @xmath345 here the suitable functions in question are those on the joint system and fock space that are in the domain of the free translation along the edges ( excluding the vertex at the origin ) and in the domain of the one - sided annihilators @xmath346 .
this agrees with the expression found in @xcite and @xcite .
the global hamiltonian form is essential for building up arbitrary quantum feedback networks @xcite .
we now specify to the case where the plant has finite dimensional hilbert space , say @xmath348 . in this case
the operators @xmath221 are naturally represented as complex - valued matrices with dimensions @xmath349 that is , we have the matrix representations @xmath350 for a fixed orthonormal basis of @xmath351 . in terms of the @xmath352
we then have @xmath353\in \mathbb{c}^{m\times nm } , \\ c & = & \mathsf{l}=\left [ \begin{array}{c } \mathsf{l}_{1 } \\ \vdots \\ \mathsf{l}_{n } \end{array } \right ] \in \mathbb{c}^{nm\times m } , \\ d & = & \mathsf{s}=\left [ \begin{array}{ccc } \mathsf{s}_{11 } & \cdots & \mathsf{s}_{1n } \\ \vdots & \ddots & \vdots \\ \mathsf{s}_{n1 } & \cdots & \mathsf{s}_{nn } \end{array } \right ] \in \mathbb{c}^{nm\times nm}.\end{aligned}\ ] ] this is essentially the equivalent linear passive model considered in remark [ ref : passive_equiv ] .
explicitly , the input - state - output equations behind this will be @xmath354 where @xmath338 is a @xmath355-values state variable and @xmath356 and @xmath357 should be @xmath358-valued functions .
let @xmath359 be a solution to the global hamiltonian problem ( [ eq : ham_def ] ) and satisfying the correct boundary conditions ( [ eq : ham_bc ] ) .
this system may be rewritten as @xmath360 where now the input and output functions are @xmath361 absorbing the relatively unimportant free dynamics due to @xmath362 , we see that ( [ eq : q1],[eq : q2 ] ) is linear system with `` input signal '' @xmath356 and `` output signal '' @xmath357 .
the functions @xmath356 and @xmath357 are boundary terms related by ( [ eq : uy_bc ] ) and not to be interpreted literally as control functions which we can assign .
we now discuss some well - known examples from the perspective of control theory .
let us set @xmath200 , @xmath363 , and @xmath364 . in this case
the only damping of significance is that of the slow component .
then we have @xmath365 and we require that @xmath366 is invertible on @xmath367 .
it is easy to see that the decoupling conditions now apply and we obtain the open dynamics with @xmath368 where the reduced hamiltonian is @xmath369 now @xmath370 is the shorted version ( schur complement ) of @xmath371 $ ] .
equivalently , @xmath370 is the the limit @xmath372 of shorted version of @xmath373 .
the detuned two level atom model considered in subsection [ subsubsec : detuned ] is a special case .
let us consider a cavity consisting of a single photon mode with annihilator @xmath74 , so that @xmath374 = i$ ] .
the number states @xmath375 , @xmath376 , span an infinite dimensional hilbert space .
mabuchi @xcite shows how a large kerr non - linearity leads to a reduced dynamics where we are restricted to the ground and first excited state of the mode , and so have an effective qubit dynamics .
we consider the @xmath377 input model with @xmath378 _ { jk } & = & \delta _ { jk}i , \\
\left [ l\left ( k\right ) \right ] _ { j } & = & \sqrt{\kappa _ { j}}e^{i\omega t}a,\quad ( j=1,2 ) \\ h\left ( k\right ) & = & k^{2}\chi _ { 0}a^{\ast 2}a^{2}+\delta a^{\ast } a -i \sqrt{\kappa _ { 1 } } \left ( \alpha \left ( t\right ) a^{\ast } -\alpha ^{\ast } \left ( t\right ) a\right ) .\end{aligned}\ ] ] in the model we are in a rotating frame with frequency @xmath96 and the cavity is detuned from this frequency by a fixed amount @xmath78 .
there is a kerr non - linearity of strength @xmath379 which will be the large parameter .
we have two input fields with damping rate @xmath380 , and the first input introduces a coherent driving field @xmath381 .
we now have @xmath382 where @xmath383 is the number operator .
the kernel space of @xmath85 is therefore @xmath384 for this situation we have @xmath385 , and we find @xmath386 since @xmath387 .
the bouten - silberfarb conditions are then satisfied and we have @xmath388 where @xmath389 .
we then have that @xmath390 _ { jk } & = & \delta _ { jk}\,i_{\mathsf{s } } , \\
\left [ \widehat{l}_{\mathsf{s}}\right ] _ { j } & = & \sqrt{\kappa _ { j}}e^{i\omega t}\sigma , \\ \widehat{h } & = & \delta \sigma ^{\ast } \sigma -i\sqrt{\kappa _ { 1}}\left ( \alpha \left ( t\right ) \sigma ^{\ast } -\alpha ^{\ast } \left ( t\right ) \sigma \right ) .\end{aligned}\ ] ] the system is then completely controllable through the control policy @xmath215 , and observable through quadrature measurement ( homodyning with @xmath391 , and @xmath392 ) and by photon counting .
the characteristic operator is as computed in subsection [ subsubsec : qubit ] .
the limit characteristic operator is then ( @xmath393 ) @xmath394 .\ ] ] we consider the case where @xmath200 , @xmath386 and @xmath395 , but @xmath396 .
the decoupling conditions are automatically satisfied , so all that is further required is that @xmath397 , which is now given by @xmath398 is invertible .
if so the reduced @xmath347 takes the simplified form @xmath399 where now @xmath400 consider a three level atom with ground states @xmath402,@xmath403 and an excited state @xmath404 with hilbert space @xmath405 .
the atom is contained in a cavity with quantum mode @xmath74 with hilbert space @xmath406 where @xmath374 = 1 $ ] and @xmath74 annihilates a photon of the cavity mode .
the combined system and cavity has hilbert space @xmath407 , and consider the following @xcite , @xmath408 here the cavity is lossy and leaks photons with decay rate @xmath77 , we also have a transition from @xmath404 to @xmath402 with the emission of a photon into the cavity , and a scalar field @xmath215 driving the transition from @xmath404 to @xmath403 . we see that @xmath409 and that @xmath85 has a 2-dimensional kernel space spanned by the pair of states @xmath410 the reduced subspace is then the span of @xmath411 and @xmath412 , and the resulting @xmath347 operators are @xmath413 where @xmath414 . here
the dynamics has a vanishing hamiltonian , but is partially observable through filtering as @xmath415 .
the limit characteristic operator is then @xmath416 .\ ] ] further examples of adiabatic elimination , particularly where the fast degrees of freedom are oscillators , can be found in @xcite .
the characteristic operator is introduced here as a mathematical object containing information about quantum input - output relations when processed by a quantum mechanical system .
the concept allows us to characterise quantum systems , and many of the features associated with classical transfer functions carry over .
we have shown that it picks out the particular scaling introduced by bouten and silberfarb for adiabatic elimination for quantum open systems as being the one which leads to the convergence of characteristic operators using schur - feshbach type resolvent expansions .
it is useful to note that strong coupling that restricts the degrees of freedom adiabatically may also be interpreted as a projection onto a zeno subspace , though generally of an open systems character @xcite .
we expect that the concept will play an important role in studying features of quantum control systems such as model reduction , controllability and observability .
again , by the schur - feshbach identity , we may write the resolvent @xmath417 as @xmath418 ^{-1 } = \left [ \begin{array}{ll } \delta_{11}\left ( s , k\right ) & \delta_{12}\left ( s , k\right ) \\ \delta_{21}\left ( s , k\right ) & \delta_{22}\left ( s , k\right ) \end{array } \right]\ ] ] where , setting @xmath419 we have @xmath420 using the fact that @xmath421 and @xmath422 , we note @xmath423 and the following scaled limit @xmath424 so that @xmath425 is a schur complement of @xmath426 $ ] . similarly it follows that @xmath258 \left [ s+m\left ( k\right ) \right ] ^{-1}\left [ \begin{array}{cc } 1 & 0 \\ 0 & k \end{array } \right ] = \left [ \begin{array}{ll } \dfrac{1}{s+\widehat{m}_{11 } } & -\dfrac{1}{s+\widehat{m}_{11}}m_{12}\dfrac{1}{m_{22 } } \\ -\dfrac{1}{m_{22}}m_{21}\dfrac{1}{s+\widehat{m}_{11 } } & \quad \dfrac{1}{m_{22}}+\dfrac{1}{m_{22}}m_{21}\dfrac{1}{s+\widehat{m}_{11}}m_{12}\dfrac{1}{m_{22 } } \end{array } \right ] .\ ] ]
let us first note that we may define @xmath427 by @xmath428 in which case @xmath429 .\end{aligned}\ ] ] we note that @xmath430 can be written as @xmath431 where we use @xmath432 .
therefore , with @xmath433 is as defined in @xmath434 , we have @xmath435 .\end{aligned}\ ] ] moreover , we see that @xmath141 is unitary . to see this ,
set @xmath436 then @xmath437[\delta _ { \mathsf{cb}}+l_{\mathsf{cf}}^{\left ( 1\right ) } \frac{1}{a_{\mathsf{ff}}}l_{\mathsf{bf}}^{\left ( 1\right ) \ast } ] = \delta _ { \mathsf{ab}}+l_{\mathsf{af}}^{\left ( 1\right ) } \frac{1}{a_{\mathsf{ff}}^{\ast } } \left\ { a_{\mathsf{ff}}+a_{\mathsf{ff}}^{\ast } + l_{\mathsf{cf}}^{\left ( 1\right ) \ast } l_{\mathsf{cf}}^{\left ( 1\right ) } \right\ } \frac{1}{a_{\mathsf{ff}}}l_{\mathsf{bf}}^{\left ( 1\right ) \ast } \end{aligned}\ ] ] however the expression in braces vanishes identically leaving @xmath438 .
the proof of the co - isometric property of @xmath439 is similar .
we note that @xmath440 it remains to show that the limit characteristic function @xmath441 has the stated form . substituting in form @xmath442 ,
we have @xmath443 = \widehat{l}_{\mathsf{a}}\frac{1}{s-\widehat{k}_{\mathsf{ss}}}\left\ { -(l_{\mathsf{ds}}^{\left ( 0\right ) } -z_{\mathsf{sf}}\frac{1}{a_{\mathsf{ff}}}l_{\mathsf{df}}^{\left ( 1\right ) \ast } ) + \widehat{l}_{\mathsf{c}}^{\ast } ( \delta _ { \mathsf{cd}}+l_{\mathsf{cf}}^{\left ( 1\right ) } \frac{1}{a_{\mathsf{ff}}}l_{\mathsf{df}}^{\left ( 1\right ) \ast } ) \right\ } s_{\mathsf{db}}\end{aligned}\ ] ] and the term in braces equals @xmath444 l_{\mathsf{df}}^{\left ( 1\right ) \ast } \end{aligned}\ ] ] and using @xmath432 again we see that the term in square brackets is @xmath445 which vanishes identically .
we note that we have the alternative form @xmath446
this research was supported in part by the national science foundation under grant no .
nsf phy11 - 25915 , and the author is grateful for the support of the kavli institute for theoretical physics , u.c .
santa barbara , where this work begun during the control of complex quantum systems programme in january 2013 .
it was also supported by epsrc grant ep / l006111/1 , and the author is particularly grateful to dr hendra nurdin for several valuable technical comments while writing this paper .
finally he as the pleasant duty to thank the isaac newton institute for mathematical sciences , cambridge , for support and hospitality during the programme _ quantum control engineering _ august 2014 where work on this paper was completed .
he acknowledges several fruitful discussions with luc bouten , jake taylor , matthew james and gerard milburn . c. gardiner and p. zoller , quantum noise : a handbook of markovian and non - markovian quantum stochastic methods with applications to quantum optics , 2nd ed .
springer series in synergetics .
springer , ( 2000 ) .
gough , h.i .
nurdin , s. wildfeuer , commutativity of the adiabatic elimination limit of fast oscillatory components and the instantaneous feedback limit in quantum feedback networks , _ journ . of math .
_ , 51(12 ) , pp .
123518 - 1123518 - 25 ( 2010 ) j.e .
gough , h.i .
nurdin , on structure - preserving transformations of the ito generator matrix for model reduction of quantum feedback networks , _
* 28 * , vol
1979 , 5422 - 5436 , november ( 2012 ) j. gough , theor .
phys . * 111 * , no . 2 , 218 - 233 , may ( 1997 ) ; j. gough , theor .
phys . * 113 * , no . 2 , 276 - 284 , november ( 1997 )
; j. gough , comptes rendus acad . ,
paris * 326 * , serie i , 981 - 985 , ( 1998 ) ; j. gough , potential analysis * 11 * , 213 - 233 , ( 1999 ) ; j. gough , rep .
. vol . * 44 * , 313 - 338 , ( 1999 ) ; j. gough , doklady mathematics , * 64 * , no.1 , 112 - 116 , ( 2001 )
|
we introduce the characteristic operator as the generalization of the usual concept of a transfer function of linear input - plant - output systems to arbitrary quantum nonlinear markovian input - output models .
this is intended as a tool in the characterization of quantum feedback control systems that fits in with the general theory of networks .
the definition exploits the linearity of noise differentials in both the plant heisenberg equations of motion and the differential form of the input - output relations .
mathematically , the characteristic operator is a matrix of dimension equal to the number of outputs times the number of inputs ( which must coincide ) , but with entries that are operators of the plant system . in this sense
the characteristic operator retains details of the effective plant dynamical structure and is an essentially quantum object .
we illustrate the relevance to model reduction and simplification definition by showing that the convergence of the characteristic operator in adiabatic elimination limit models requires the same conditions and assumptions appearing in the work on limit quantum stochastic differential theorems of bouten and silberfarb @xcite .
this approach also shows in a natural way that the limit coefficients of the quantum stochastic differential equations in adiabatic elimination problems arise algebraically as schur complements , and amounts to a model reduction where the fast degrees of freedom are decoupled from the slow ones , and eliminated .
|
it may come as a pleasant surprise to the dynamical systems community , to learn that the work of yasha sinai and david ruelle has had a major impact in experiments on microwave geometries .
the connections arise from recent developments which demonstrate that wave mechanics experiments using microwaves are an ideal laboratory for studying the so - called quantum - classical correspondence , a central issue in quantum chaos .
these microwave experiments have shown that several theoretical results on the mathematics of chaotic dynamical systems have manifestations in their corresponding quantum or wave mechanics .
the term chaos often means hyperbolicity which guarantees the decomposition of the tangent space at each phase space point into expansion and contraction subspaces .
a systematic study of hyperbolic systems from the geometric theory perspective was initiated by smale @xcite .
another viewpoint , namely the ergodic theory or the probability approach , was pioneered by sinai and ruelle . of great interest and possible experimental realization
are billiard systems which are two - dimensional , bounded or open .
not all closed billiards have hyperbolic properties . for a dispersing billiard which is closed and
has finite concave boundaries , sinai @xcite proved that the system is ergodic .
but certain billiards with convex boundaries @xcite such as the bunimovich stadium @xcite , were also found to be hyperbolic .
another important billiard system is that of @xmath0 disks on a plane , which is hyperbolic and a paradigm model of axiom a systems considered by ruelle @xcite and pollicott @xcite . experiments utilizing thin 2-d microwave structures have been shown recently to provide an ideal laboratory system to explore issues in hyperbolic dynamics and quantum chaos .
these experiments utilize thin geometries in which maxwell s equations reduce to the time - independent schrdinger equation @xmath1 , where @xmath2 is @xmath3 the @xmath4-component of the electric field .
this mapping enables precision room temperature experiments exploring wave mechanics on laboratory length scales that can be easily manipulated .
the experiments typically probe the two - point green s function of geometries which may be _ closed _ , such as the sinai billiard , or _
, such as @xmath0 disks on a plane .
in addition to yielding eigenvalue or resonance spectra , a unique advantage of the microwave experiments is the ability to directly map eigenfunctions and standing wave patterns .
a major theme of work in this area has focussed on the quantum - classical boundary .
some of the principal results that have emerged are the observation of scars @xcite , precision tests of random matrix theory ( rmt ) and non - linear sigma models in chaotic and disordered cavities @xcite , observation of localization in disordered billiards @xcite , and classical and quantum properties of open systems @xcite .
these microwave experiments may also be classified as _ experimental mathematics _ !
a noteworthy example is the work on `` not ` hearing the shape ' of drums '' @xcite , which demonstrated experimentally the exact equivalence of eigenvalues of certain isospectral geometries .
the isospectrality is purely a consequence of geometry or combinatorics , and is independent of the dynamics ( whether chaotic or regular ) .
this paper summarizes recent results exemplifying the impact of sinai and ruelle s work on the microwave quantum chaos experiments .
we first discuss the properties of the eigenvalues and eigenfunctions of sinai billiard - shaped microwave cavities .
we then discuss , in the context of 2-d @xmath5-disk billiards , the major role of the ruelle dynamical zeta - functions in the classical and semiclassical theory of chaotic repellers , and the experimental observation of ruelle - pollicott resonances .
following the pioneering work of sinai , the sinai billiard has become a model geometry exemplifying classical and quantum chaos in a closed system .
an experimental realization of the sinai billiard was first achieved in microwave cavities @xcite .
the microwave experiments on sinai billiard cavities yield the eigenvalue spectra @xmath6 directly as peaks at frequencies @xmath7 in the experimental 2-point microwave transmission function @xmath8 of the cavity ( fig.1(top ) ) .
typically @xmath9 eigenvalues can be measured in the experimentally accessible frequency window . the corresponding stick spectrum is then analyzed to yield information on spectral statistics .
the approach here closely parallels work in spectra of atomic and nuclear systems @xcite , which have been analyzed in terms of rmt .
here we summarize the principal results for the sinai billiard .
the main conclusion is that the eigenvalue and eigenfunction statistics are in good agreement with universal predictions of rmt expected to apply to quantum systems with classically chaotic dynamics @xcite .
however , sufficiently precise experiments clearly reveal deviations from the universal results due to the presence of marginal unstable `` bouncing ball '' ( b - b ) orbits .
the nearest neighbor spacing statistics @xmath10 for the sinai billiard spectrum shows the well - known level repulsion @xmath11 as the spacing @xmath12 , a characteristic feature of quantum chaos ( fig.1(middle ) ) .
fits to the brody function @xmath13 , yield a brody parameter @xmath14 , where @xmath15 corresponds to exact agreement with the famous wigner - dyson statistic , @xmath16 .
this deviation is a small but definite signature of the b - b orbits .
more striking manifestations of these orbits can be found in longer range correlations , such as the spectral rigidity @xmath17 ( fig.1(bottom ) ) .
this quantity for the sinai billiard shows initial agreement with rmt for small @xmath18 followed by an essentially linear rise from the weak logarithmic dependence expected .
incorporation of the b - b orbits contributions gives excellent agreement between experimental data and theory .
the eigenfunctions corresponding to each eigenvalue peak are obtained using cavity perturbation techniques described in ref .
representative eigenfunctions of the sinai billiard are shown in fig.2 , and show a fascinating variety of patterns .
this is in contrast with the rectangular cavity , obtained from the sinai billiard by removing the central circular disk , for which each eigenfunction can be labeled by @xmath19 , which arise from the integrability of @xmath20 .
non - integrability in the sinai billiard is immediately apparent from the fact that the only quantum number that can be assigned to each eigenfunction is @xmath21 - the lack of integrals of motion precludes any further quantum number assignments .
the question , what classical phase space structures characterize the eigenfunctions , has occupied physicists for several years .
heller pointed out that the eigenfunctions of chaotic billiards should organize themselves around periodic orbits , even the unstable ones , and called these structures `` scars''@xcite .
scars were seen in numerical simulations , and were implicated indirectly in atomic phenomena .
the first direct experimental observation of scars was in eigenfunctions of a sinai billiard microwave cavity@xcite .
an example is shown in fig.2 at @xmath22 .
this eigenfunction is scarred along the diagonal periodic orbit , which is unstable .
the presence of stable b - b orbits is also visible in some eigenfunctions such as the one at @xmath23 .
however most of the eigenfunctions are not _ visibly _ scarred by either stable or unstable po .
a favorite is the one at @xmath24 .
it shows the circular symmetry of the central disk with almost no visible influence of the rectangular boundary .
a more complete characterization of the eigenfunctions in terms of classical phase space structures is a major challenge of research in this area .
another approach to analyzing eigenfunctions is in terms of their statistical properties .
the leading statistic is the density distribution @xmath25 which is shown in fig.3(top ) . for the quarter sinai billiard , in which the geometric symmetries are absent ,
the finite probability of finding large densities is evident from the figure .
the dependence of @xmath25 on the density @xmath26 follows closely , but not exactly , the porter - thomas ( p - t ) distribution obtained from rmt .
this again is due to the presence of the b - b orbits . we have confirmed this by showing almost exact agreement with the p - t distribution by removing the b - b orbits by constructing a so - called sinai stadium , which is a hybrid of the sinai and stadium billiards .
spatial correlations @xmath27 can also be examined and are shown in fig.3(bottom ) .
the experimental data follow closely the dependence @xmath28 , with @xmath29 , expected from rmt for 2-d .
the data for the sinai billiard lie just below this functional form , this can be seen from the inverse participation ratio @xmath30 . the deviation from the universal value of @xmath31 on the lower side is due to the b - b orbits . here
again better agreement with the rmt universal value and function is obtained by eliminating the b - b orbits such as in the sinai stadium . in summary
the sinai billiard has been a model system exemplifying the fundamental properties of a quantum chaotic system , with near perfect agreement with rmt predictions for universal distributions of spectral and eigenfunction statistics , and where the deviations from universality are clearly identifiable as arising from b - b orbits .
the system of @xmath0 disks on a plane @xcite provides us with a paradigm model system for classical and quantum chaos in open systems .
the system is hyperbolic and not ergodic .
the microwave realization of the @xmath0-disk system is similar to that of the sinai billiard . to make the system open , microwave absorbers
were placed around and far from the disks to have experimentally ignorable reflection , and to provide a good realization of escape to infinity .
because of the @xmath32 symmetry of the @xmath0-disk system , one can probe the system in many different ways . periodic orbit
based treatments have been successfully used both classically and semi - classically .
the ruelle dynamical zeta - function , in both the classical and semiclassical versions , has played a major role in studying the properties of this system .
@xmath33 and the classical ruelle zeta - function is @xmath34 the product is over all classical primitive periodic orbits without repetition .
here @xmath35 and @xmath36 are the length and period of periodic orbit @xmath37 with the maslov index @xmath38 .
@xmath39 is the bigger eigenvalue of the monodromy stability matrix for billiard systems .
the semiclassical ruelle zeta - function can be derived from the gutzwiller trace formula @xcite of the green s function in the semiclassical limit @xmath40 which is summed over all classical periodic orbits .
the finite widths of the quantum resonances arise in the @xmath0-disk system because the system is open and the strange repeller made of classical orbits can trap waves only for finite time .
the classical ruelle zeta - function can be derived from the trace of the resolvent @xmath41 with @xmath42 for hamiltonian system . for integrable system ,
the poles of the resolvent operator are purely imaginary , this gives the periodic motion of the classical dynamics . for open or ergodic system
, these poles may have nonzero real part and some of them may be real .
this gives rise to the correlation decay of the classical dynamics @xcite .
these complex poles are the ruelle - pollicott ( r - p ) resonances @xcite .
if the poles are isolated , the decay is exponential . in the case of a branch cut
, the decay may be algebraic .
contrary to the trace formula of the classical resolvent @xcite which is exact , the semiclassical trace formula of gutzwiller is not , and it is not always guaranteed that the poles of the semiclassical ruelle zeta - function are also the poles of the green s function .
but for hard wall systems such as the billiards , the discrepancy was found to be very small @xcite .
the index @xmath43 and @xmath44 comes in from the decomposition of the trace formulas into ruelle zeta - functions with @xmath45 and @xmath46 .
one may extrapolate @xmath44 from integers to real numbers to get more information of the classical dynamics @xcite .
for the calculation of sharp resonances , only @xmath47 and @xmath48 will be needed in the ruelle zeta - functions . for the @xmath0-disk system ,
since the velocity of the particle is constant , @xmath49 , one can simply set @xmath50 , and work in the @xmath51-space for both the classical and semiclassical ruelle zeta - functions .
if the separation of the disks are big enough to have no tangent orbits , the dynamics is continuously hyperbolic and stable .
a symbolic dynamics of @xmath52 symbols with finite grammar exists for the system .
the euler product of the ruelle zeta - functions can be expanded as @xmath53 . here
@xmath54 is the weight of fundamental orbits @xmath55 , @xmath56 is the curvature which is the weight difference between long orbit and combination of shorter ones
. only a few primitive pos will be needed in the ruelle zeta - function to give quite accurate calculation of the poles because the curvature in the cycle expansion will decay exponentially with increasing po length @xcite .
this further simplifies the task of calculating resonances . for the 3-disk system ,
the number of fundamental orbits is just 2 .
only 8 periodic orbits with period up to 4 were be needed in the ruelle zeta - functions to give very accurate estimate of the quantum and classical resonances .
a typical trace of the transmission @xmath57 of the quantum resonances is shown in fig.4 for the 3-disk system in the fundamental domain .
the experimental trace was fitted well by the sum of lorentzians @xmath58 here @xmath59 are the semiclassical resonances obtained through the semiclassical ruelle zeta - function .
the coupling @xmath60 depend on the probes location and were chosen to optimize the fitting .
the similarity between the classical ruelle zeta - function and the semiclassical ones is not coincidental .
this implies a deeper relation between the quantum and the classical dynamics besides the correspondence principle .
there is increasing interest in understanding the nature of the classical - quantum correspondence .
the quantum dynamics is found to approach the classical dynamics with increasing coarse - graining in the phase space for a mixed phase space system @xcite .
the autocovariance of the two - level spacing and the correlation of the density of states are also found to be related to the classical zeta - function in the study of the riemann zeta - function zeroes @xcite .
for an open hyperbolic system , we have shown that the correlation decay of the quantum system is determined by the classical dynamics @xcite .
that means the classical r - p resonances display their fingerprints in the quantum dynamics .
we find that the autocorrelation of the transmission @xmath61 are well - described as a sum of lorentzians @xmath62 here @xmath63 are the _ classical _ r - p resonances .
the coefficients @xmath64 are related to the eigenfunctions of the perron - frobenius operator .
the experimental observation of the r - p resonances is illustrated in fig.5 .
we are thus able to observe both the quantum and classical resonances in the same microwave experiment !
this exactly parallels the periodic orbit treatment , in which the poles of the semiclassical ruelle zeta - function yields the quantum resonances while the classical ruelle zeta - function provides the classical resonances for the hyperbolic @xmath0-disk system .
the convergence of experiment and theory have been made possible by extensive developments in the machinery of both approaches .
the advantages of the microwave experiments are the well - specified nature of the hard wall potentials , and the efficient ensemble averaging that can be achieved .
the microwave experiments thus have emerged as an ideal laboratory system to study both quantum and classical chaos , and some of the mysteries of the quantum - classical correspondence . in closing , we express our appreciation of the influential role played by the statistical physical workshops at rutgers university organized by professor joel lebowitz .
our observation of scars in sinai billiard cavities was first reported at the 1991 meeting , and the observation of r - p resonances was reported in the 2000 meeting .
|
we discuss the impact of recent developments in the theory of chaotic dynamical systems , particularly the results of sinai and ruelle , on microwave experiments designed to study quantum chaos .
the properties of closed sinai billiard microwave cavities are discussed in terms of universal predictions from random matrix theory , as well as periodic orbit contributions which manifest as ` scars ' in eigenfunctions .
the semiclassical and classical ruelle zeta - functions lead to quantum and classical resonances , both of which are observed in microwave experiments on @xmath0-disk hyperbolic billiards .
* key words * : microwave , hyperbolic , sinai billiard , correlation , ruelle zeta - function , resonances
|
most galactic black hole candidates exhibit at least two distinct spectral states ( see liang & narayan 1997 , liang 1998 , poutanen 1998 for reviews ) . in the hard state
(= soft x - ray low state ) the spectrum from @xmath2 kev to a few hundred kev is a hard power law ( photon index @xmath3 ) with an exponential cutoff .
this can be interpreted as inverse comptonization of soft photons . in the soft state
( often , but not always , accompanied by the soft x - ray high state ) , the spectrum above @xmath4 kev is a steep power law ( photon index @xmath5 ) with no detectable cutoff out to @xmath2 mev .
this multi - state behavior is seen in both persistent sources ( e.g. cygnus x-1 ) and transient black hole x - ray novae ( bhxrn , e.g. grs 100945 ) .
gx 3394 is unusual in that it is a persistent source , being detected by x - ray telescopes most of the time , but it also has nova - like flaring states . in 1996
, we performed a series of multiwavelength observations of gx 3394 when it was in a very low state .
this paper is one of a series that describe the results of this campaign . in paper
i ( @xcite ) we discuss the radio , x - ray , and gamma - ray daily light curves and spectra obtained in 1996 july . in paper ii ( @xcite )
we discuss the x - ray timing analysis from 1996 july . in bttcher ,
liang , & smith ( 1998 ) we use the gx 3394 spectral data to test our detailed self - consistent accretion disk corona models .
these papers expand significantly on our preliminary analyses ( @xcite ) .
in this paper we focus on keck spectroscopy performed on 1996 may 12 ut . beginning with the discovery observations in the 1970s ( @xcite ) it was clear that the optical counterpart to gx 3394 is highly variable ( e.g. @xcite ) .
for example , one month prior to our keck observations , when the source was in a similar low state , it was found to have 16 second optical quasi - periodic oscillations ( @xcite ) . in previous spectroscopic observations ,
the optical emission has always been dominated by the highly variable accretion disk , preventing a direct observation of the spectrum of the companion star .
this is also true of our keck observations . for the first time we are able to resolve the emission line ,
showing it is double peaked .
this is similar to the emission lines from the quiescent optical counterparts of many bhxrn .
however , we show that the peak separation in gx 3394 is much smaller than in the bhxrn . instead , the emission line may be more akin to the one in cygnus x-1 .
our keck observations were performed on 1996 may 12 ut ( mjd 50215 ) . at this time
, the asm on the _ rossi x - ray timing explorer _ did not detect the source , with count rates of @xmath6 cps in the 1.3 3.0 kev band , @xmath7 cps in the 3.0 5.0 kev band , and @xmath8 cps in the 5.0 12.1 kev band .
similarly , using the earth occultation method , batse on the _ compton gamma - ray observatory _ also did not detect it with flux @xmath9 in the 20 100 kev band .
the source was very faint or not detected for weeks on either side of this time : see figure 1 of smith et al .
( 1997a ) for the rxte light curve .
this is as close to `` quiescence '' as the high - energy emission comes in gx 3394 .
our observations of gx 3394 were made during one of the same runs in which we studied the bhxrn nova ophiuchi 1977 in quiescence and found its orbital period and mass function ; see filippenko et al .
( 1997 ) and harlaftis et al .
( 1997 ) for full instrumental and reduction details .
the low - resolution imaging spectrometer was used at the cassegrain focus of the keck i telescope .
two consecutive 400 second exposures were taken , starting at 11:34 and 11:41 ut . a 1 slit was used , but the seeing was poor ( 2.2 ) and variable , probably due to the high airmass ( 2.8 ) and possibly affected by thin clouds .
the original goal of our observations was to study the source in optical quiescence to determine the nature of the companion star ; it was hoped that the lack of detectable high - energy emission would make this possible .
however , we found that the optical emission was still dominated by the accretion disk .
figure 1 shows the two consecutive 400 s spectra .
the data are not fully photometric but give @xmath10 mag .
this is typical for the source in the low state , @xmath11 : here we include both the results reported for the traditional `` soft x - ray low '' state ( @xmath12 ) and the sometimes discussed `` off '' state ( @xmath13 ) , since it is now generally believed that the off state is a lower luminosity low hard state ( @xcite ) .
there is no artificial offset in the two curves in figure 1 .
the ratio of the spectra is consistent with being constant across the whole range , including the emission lines .
such a variability and constant spectral shape agrees with previous observations .
for example , in the high state in 1986 june / july it was found that there were large changes in @xmath14 , but no accompanying variations in the ( ) index ( @xcite ) .
however , since the slit was narrow relative to the seeing conditions at the time of the observation , we can make no accurate statements regarding the true variability of the source .
the keck continua can be approximately fit by a linear relationship . extrapolating this , we find that the flux density goes to zero at @xmath15 , easily consistent with the lack of observed x - ray emission ( 5 kev corresponds to 2.45 ) .
however , given the relatively narrow wavelength range of the keck spectrum , and a misalignment between the slit position angle ( 60 ) and the parallactic angle ( 175 ) which can distort the continuum shape at the edges of the spectrum ( @xcite ) , we can not make conclusive statements about the spectrum at higher energies .
the dominant emission line in our spectra is .
its equivalent width ( ew ) of @xmath16 is similar to the lines seen in previous gx 3394 observations ( @xcite ) .
however , this is small compared to other bhxrn ; for example , the line in nova oph 1977 had ew @xmath17 during the same observing run ( @xcite ) .
an expanded view of the line in gx 3394 is shown in figure 2 .
for the first time we are able to resolve a double peaked line about the rest wavelength of 6562.80 in both spectra .
we used a linear fit to the continuum around the line to determine the rms noise . adding a single gaussian line ( with unconstrained centroid ) does not give a good fit in either of our spectra .
for the upper spectrum , the reduced = 1.57 for @xmath18 degrees of freedom .
the probability that a random set of data points would give a value of as large or larger than this is @xmath19 .
the lower spectrum gives very similar results : = 1.54 , @xmath20 , @xmath21 .
using two unconstrained gaussian lines greatly improves the fits .
figure 2 shows the best fit results .
the fit to the upper spectrum now has = 1.23 , @xmath22 , @xmath23 , while the lower spectrum now has = 0.95 , @xmath24 , @xmath25 .
it therefore appears that the line in gx 3394 is double peaked , as is the case for the quiescent optical counterparts of many bhxrn .
the best two gaussian fits have different fit parameters for the two spectra .
however , if we multiply the fit to the upper spectrum by a constant 1/1.062 given by the ratio of the continua shown in figure 1 , we find that this gives an adequate fit to the lower spectrum : = 1.11 , @xmath26 .
thus we are unable to claim there is significant variability in the lines in these data .
searching for this variability is the subject of an ongoing project , at which point it will also be possible to use more realistic emission - line profiles ( e.g. orosz et al .
1994 , and references therein ) .
the separation of the peaks is @xmath27 , implying a @xmath1 .
if we assume this peak emission comes from a circular keplerian orbit , it would be at a distance @xmath28 cm , where @xmath29 is the mass of the black hole .
this would be in the outer regions of the accretion disk .
this distance can be contrasted with the suggestion by motch et al .
( 1982 ) that the 20 second optical qpo they detected came from a ring orbiting with a keplerian period of 20 seconds at a radius of @xmath30 cm
. the only other emission lines in the spectra are from . for @xmath315875 ,
the combined equivalent width is @xmath32 , although this is difficult to fit accurately because of the adjacent nad interstellar absorption lines . for @xmath316678 ,
the combined equivalent width is @xmath33 .
we do not see any evidence for an absorption line from at @xmath316708 .
this result agrees with previous observations of gx 3394 ( @xcite ) .
a double peaked emission line has been seen in several bhxrn in quiescence , showing that there is optical emission from their accretion disks long after the x - ray nova event .
table 1 lists representative values for their peak separations .
the remarkable feature that has already been noted in previous studies is that the peak separations in these different sources are all surprisingly similar .
the peak separation in gx 3394 clearly shows that this source is different from the usual bhxrn
. this may not be too surprising .
although gx 3394 had no detectable high - energy emission around the time of our keck observation , the source did not exhibit a very long period of `` quiescence '' in the same way as the other bhxrn .
the narrower peak separation in gx 3394 implies that the optical emission comes from a larger keplerian radius than in the bhxrn , which may be a clue to its different behavior .
gx 3394 is most often compared to cygnus x-1 .
unfortunately , for cygnus x-1 the absorption and emission lines from the companion star dominate , and only the wings of the emission line are detected ( @xcite ) .
it is therefore very difficult to determine if the emission line in cygnus x-1 is double peaked .
however , the width of the base of the emission line in cygnus x-1 is quite similar to that in gx 3394 , suggesting that the optical emission from their accretion disks may be similar .
it is also interesting to note that the peak separation was @xmath34 in the emission lines from gro j165540 _ during outbursts _ ( @xcite ) .
this is a superluminal jet source , and may indicate a connection to the suggested radio jet in gx 3394 ( @xcite ) .
we thank philippe durouchoux for suggesting we check for a line .
we also thank the referee for carefully reading the manuscript and providing useful suggestions and clarifications .
this work was supported by nasa grants nag 5 - 1547 and 5 - 3824 at rice university , and nsf grant ast-9417213 at uc berkeley .
this work made use of the rxte asm data products provided by the asm / rxte teams at mit and at the rxte sof and gof at nasa s goddard space flight center .
the batse daily light curves were provided by the compton observatory science support center at nasa s goddard space flight center .
lrl gro j0422 + 32 & 900 & orosz & bailyn ( 1995 ) , filippenko , matheson , & ho ( 1995 ) nova muscae 1991 & 900 & orosz et al .
( 1994)nova oph 1977 & 900 & remillard et al .
( 1996 ) , filippenko et al .
( 1997 ) grs 100945 & 1000 & shahbaz et al .
( 1996 ) a062000 & 1100 & orosz et al .
( 1994)gs 2000 + 25 & 1400 & casares et al .
( 1995 ) , filippenko , matheson , & barth ( 1995 ) gx 3394 & 370 & this paper
|
as part of our multiwavelength campaign of observations of gx 3394 in 1996 we present our keck spectroscopy performed on may 12 ut . at this time , neither the asm on the _ rossi x - ray timing explorer _ nor batse on the _ compton gamma - ray observatory _ detected the source .
the optical emission was still dominated by the accretion disk with @xmath0 mag .
the dominant emission line is , and for the first time we are able to resolve a double peaked profile . the peak separation @xmath1 .
double peaked emission lines have been seen in the quiescent optical counterparts of many black hole x - ray novae .
however , we find that the peak separation is significantly smaller in gx 3394 , implying that the optical emission comes from a larger radius than in the novae
. the emission line may be more akin to the one in cygnus x-1 , where it is very difficult to determine if the line is intrinsically double peaked because absorption and emission lines from the companion star dominate .
|
the metallicity of hii regions in small isolated dwarf galaxies is key to investigating the physical processes that govern star formation in undisturbed stellar systems .
the small isolated gas rich irregular dwarf galaxy ( sigrid ) sample of small isolated gas rich irregular dwarf galaxies @xcite was selected with the aim of exploring the behavior of the mass or luminosity metallicity relation at the low end of the mass scale .
this is based on the observation that nebular metallicity decreases with galaxy stellar mass / luminosity ( see , for example , * ? ? ?
* ; * ? ? ?
however , the low end of the mass scale shows significantly more scatter in metallicity than the high end in the tremonti sdss data . by selecting isolated dwarf galaxies
, it was our intention to see if this scatter persisted , and whether it was an intrinsic property of small galaxies .
the sigrid study is complementary to the `` choirs '' study which looks for tidal dwarf emission line galaxies in group environments @xcite .
it is distinct from the spitzer local volume legacy survey used by @xcite and the sdss data of @xcite in using targets specifically chosen for their isolation .
it is most similar in concept to the study by @xcite of galaxies in the lynx - cancer void , but is limited to small very isolated dwarf objects .
other questions that the sigrid observations are intended to address are the existence or otherwise of a primary nitrogen `` plateau '' at metallicities below z=8.45 @xcite , and the relationship between oxygen abundances determined using `` direct method '' , based on the measurement of the electron temperature and the estimation of the ionization correction factors to account for unseen ionization stages , and `` strong line '' technique , based on a calibration of the bright emission lines and emission line ratios .
there has not been good agreement to date between the two methods , attributed to the empirical nature of the strong line methods .
they have been calibrated in terms of the direct method , and have not until recently had an analytical basis .
the direct method has been used as a standard for temperature and metallicity measurement , against which the strong line methods have been calibrated .
@xcite subsequently presented a set of strong line diagnostic grids derived from the mappings photoionization modelling code , based on the latest atomic data ( see * ? ? ?
we use both the new atomic data and the new diagnostic grids in our analysis .
one might expect there to be greater scatter in the mass metallicity relation at low masses , due to ( 1 ) measurement noise in nebular spectra in fainter galaxies , and ( 2 ) different star formation histories in the galaxies .
@xcite suggest that the apparent scatter diminishes in observations at longer wavelengths ( 4.5@xmath7 m ) , and we present additional optical spectral evidence on this question .
the behavior of the ratio of nitrogen to oxygen abundances at low metallicities also shows increased scatter at lower metallicity .
the current consensus appears to be that there is a low metallicity plateau in log(n / o ) , indicating the existence of primary nitrogen ( see , for example , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , these previous works were not confined to small isolated dwarf galaxies .
results in our earlier paper on two isolated local void dwarf galaxies @xcite , indicated that log(n / o ) did not plateau at low metallicity , suggesting no evidence for primary nitrogen . in this paper
we present additional evidence for this .
the paper is structured as follows : in section 2 we detail the sample selection , the spectroscopic observations , and the data reduction details .
h@xmath8 images of each observed target , spectra , and de - reddened nebular emission line fluxes are presented in section 3 . in section 4
we present the principal results of the analyses : electron temperatures , gas - phase nebular metallicities with the diagnostic grids , the nitrogen to oxygen flux ratios , the [ s ii ] line ratios and electron densities , and the luminosity
metallicity results . in section 5
we discuss these results , including the anomalies , and in section 6 we present our conclusions .
a discussion of methods used to estimate errors in the emission line fluxes is given in the appendix .
the sigrid sample was selected to identify small isolated gas - rich irregular dwarf galaxies using the criteria described in @xcite .
all objects are members of the survey for ionization in neutral gas galaxies ( singg ) catalog @xcite , selected from their hi signatures in the hi parkes all - sky survey ( hipass ) @xcite and the presence of h@xmath8 emission from star forming regions . from this sample
we have , to date , observed 34 objects using ifu optical spectroscopy , as detailed below . from these observations we report here on 12 galaxies where the [ o iii ] auroral line is evident , allowing us to calculate the electron temperature t@xmath0 , and the gas - phase oxygen abundance ; and an additional galaxy , j1118 - 17 , with an unusual spectrum with no observed [ n ii ] lines . in four objects , two separate hii regions
were observed which exhibited the auroral line , resulting in 18 separate hii region observations .
three objects ( j1152 - 02 , j1225 - 06 , j1328 + 02 ) are not members of the final sigrid sample , but had been observed during the refining of that sample .
they were later excluded due to possible influence by regional galaxy groups and clusters , although they are clearly isolated objects , as evidenced by their isolation @xmath9 index values @xcite .
five objects qualify as blue compact dwarf ( bcd ) galaxies , using the definition by @xcite and discussed by @xcite , though they have not previously been identified as such .
the targets were observed using the wifes ifu spectrograph on the australian national university ( anu ) 2.3 m telescope at siding spring .
the wifes instrument is a double - beam image - slicing ifs , designed specifically to maximise throughput .
it covers the spectral range 35009000 , at resolutions of 3000 ( full spectral range ) and 7000 ( long wavelength limit 7000 ) .
it has a science field of view ( fov ) of 25@xmath1038 arc sec .
as most of the sigrid objects subtend angles less than its fov , wifes is an ideal instrument to measure nebular metallicities in the ionized hydrogen star - forming regions .
the instrument generates a data cube that allows exploration of nebular and continuum spectra in different regions of the target objects .
typically , even in poor seeing wifes resolves sigrid object star formation regions easily , making possible exploration of excitation and abundances in different regions of each object . in these observations ,
resolutions used were r=3000 for the blue camera and r=7000 for the red , spanning a usable wavelength range of @xmath43600 to 7000 .
short period ( 150 second on object , 75 second on sky ) nod - and - shuffle observations were used for all objects , to allow near - complete removal of the sky lines .
the exposure times recorded in column 5 of table [ t1 ] are the on - object integration times .
details of the observations are given in table [ t1 ] .
the sample is described in detail in @xcite : essentially , the objects lie between redshifts of 300 and 1650 km / s , have neutral hydrogen masses less than and r - band magnitudes fainter than the small magellanic cloud , low rotation velocities , show evidence of current active star formation , and are isolated , away from galaxy clusters and the tidal effects of other galaxies . all objects exhibit low ( gas - phase ) oxygen abundance ( log(o / h ) @xmath11 z@xmath12 ) , as we describe below .
the seeing listed in column 7 shows that in all but one case , ( j1403 - 27 ) , the seeing was average for siding spring . even in that case , the seeing was better than the spaxel sample size , resulting in little if any flux loss .
several classes of object were identified in @xcite , including `` bloaters '' , which are objects considerably more spatially extended than their masses would suggest .
one of these is j1118 - 17 .
it is very faint , but as we show below , appears to have a very low metallicity . in the light of the results obtained for the relatively faint object , j1118 - 17 ( s1 and s2 targets ) , it would have been desirable to undertake significantly longer integration times , but observing conditions ( weather and moonlight ) did not permit this . we intend to undertake further longer integration time observations for this unusual object .
the spectrum of j1118 - 17s1 is very noisy with few usable spectral lines , so we have analysed only s2the two objects appear similar apart from luminosity .
we reported results for j1609 - 04 , a very isolated galaxy at the edge of the local void , in a previous paper @xcite , and the results are included again here for completeness .
the data were reduced using the revised wifes python `` pywifes '' pipeline .
this involves steps generally similar to those described in for the older pipeline : bias modelling and subtraction , flat fielding , arc line identification and wavelength calibration , cosmic ray removal , sky - line subtraction using nod - and - shuffle , initial data cube construction and atmospheric dispersion correction , standard flux star calibration , telluric correction , assembly into the final data cubes and , where necessary , combination of multiple cubes into a final object data cube .
the standard stars used were taken from @xcite .
spectral sampling was undertaken using a 6 arc sec diameter circular spatial area centered on each hii region , through the full wavelength range of the data cube , to obtain spectra for each region .
line fluxes were measured from these spectra using iraf / splot .
particular care was taken to account for any stellar absorption features underlying the balmer emission lines , although in all cases , this was minor or absent , due to low stellar continuum .
in fact , the stellar continuum was extremely faint , with the exception of the object j0005 - 28 ( see figure [ fig_7 ] , displayed on a log - intensity scale ) . unlike single slit spectra , with ifu data cubes , we are able to select the entire area of the hii region from which to extract the spectrum , and exclude the majority of the galaxy stellar background , resulting in better signal - to - noise .
test sample sizes showed that all the detectable h - alpha and [ oiii ] in each hii region lay within the sample aperture , except where there are closely adjacent hii regions ( e.g. , j1609 - 04 ) , where limiting the sample size to 6 arc sec diameter avoids sampling a different region .
ideally , single spaxel - based analysis would be preferable to multi - spaxel sampling , but these objects are so faint that the resultant noise is prohibitive .
flux de - reddening was performed on the raw flux data using two methods .
first , for consistency with other work , we used the dust reddening formulae from @xcite with a@xmath13=3.1 , using the resultant balmer line flux ratios as a check . to confirm these results , we used the dust models from @xcite , using a relative extinction curve with r@xmath14=4.3 , where r@xmath14 = a@xmath13/(e@xmath15 ) and a@xmath13 is the v - band extinction .
this is discussed in more detail in ( * ? ? ?
* appendix 1 ) .
we used an initial balmer decrement ratio of 2.82 for h@xmath8/h@xmath16 , corresponding to an electron temperature of 12 500k , adjusted the electron temperature using the direct method derived from the [ o iii ] line ratios , then adjusted the apparent balmer ratios by varying the value of a@xmath13 for the best fit to the h@xmath17/h@xmath16 ratio , using the ratio h@xmath18/h@xmath16 as a check , fitting to the @xcite case b balmer ratios .
the de - reddened flux values reported in table [ t2 ] are those using the cardelli method . in all cases ,
the two approaches gave similar results ( to within @xmath43% in the de - reddened balmer line ratios ) . in only one case , j2234 - 04 , object a , did the de - reddening fail to provide a plausible result , and this has been excluded from the results reported here .
it appears likely that two or more incompletely removed cosmic ray artefacts were the cause of the problem .
images of the objects listed in table [ t1 ] are shown in figure [ fig_1 ] .
these are 38@xmath1025 arc sec h@xmath8 slices from the wifes data cubes .
spectra were extracted from these cubes using 6 arc sec diameter samples , centered on each ( bright ) hii region .
note that the seeing during the observations of object 7 , j1403 - 27 , was poor3 to 5 arc sec so the dimensions of the image do not indicate the true size of the hii region .
the image scaling does not reflect the true brightness , but has been adjusted to illustrate the extent of the fainter parts of the hii regions .
the sample size is larger than the worst seeing so avoids any sample size flux losses . in all but the one case the sample size is much larger than the object ( see table [ t1 ] ) .
slices from wifes 35@xmath1028 arc sec image cubes .
the size of the areas sampled to extract spectra is shown in panel 13 .
north is indicated by the long arrow and east by the bar ( @xmath19 ) .
the red crosses mark the center of the sampled areas .
note that the images have been stretched to show the fainter areas . ]
spectra extracted from the wifes data cubes are shown in figures [ fig_2][fig_6 ] .
the current data reduction pipeline creates a `` sag '' artefact for wavelengths shorter than @xmath44000 , in the absence of a strong stellar continuum .
stellar continua are weak or non - existent in most objects except for j0005 - 28 , j2242 - 06 , and j2349 - 42 .
defective ccd chip amplification at the time of the observations caused two high noise regions in the spectrum of j2242 - 06 , which have been replaced in figure [ fig_3 ] with straight lines .
incipient noise from these two chip amplifiers is apparent in other spectra , but does not impact on any of the important diagnostic emission lines .
de - reddened fluxes , equivalent widths , and logarithmic extinction coefficients ( c(h@xmath16 ) ) for the observed optical nebular lines are shown in table [ t2 ] .
as noted in the appendix , the extinction coefficients were calculated using the cardelli reddening law with a@xmath13=3.1 similar results ( @xmath203% ) were obtained using methods derived from @xcite .
the equivalent widths are large for for some objects , indicating the very low continua , because the host galaxies are very small and faint , and the spectra were measured from an area sampling only the immediate area of the hii region .
figure [ fig_4b ] is a close up of the spectra in figures [ fig_2 ] to [ fig_4 ] , from 4200 to 4500 , illustrating the h@xmath17 and auroral [ o iii ] 4363 lines .
the signal to noise is mainly very good , but for three of the 14 objects in figures [ fig_2 ] to [ fig_4 ] , the detections are real but noisy ( see table [ t2 ] ) . and
[ o iii ] 4363 lines for eight of the sigrid objects . ]
the spectrum from * j1118 - 17s2 * is shown in figure [ fig_5 ] .
an enlarged section of the spectrum of j1118 - 17s2 near h@xmath8 is shown in figure [ fig_6 ] , illustrating the apparent absence ( to within the noise ) of [ n ii ] , although both [ s ii ] lines are apparent .
this lack of any evidence for nitrogen suggests a particularly low metallicity , which we have estimated below using strong line diagnostic measurements .
this object ( and the associated j1118 - 17s1 ) warrants further observation to reduce the noise and establish the [ n ii ] flux .
we have not presented the spectrum of j1118 - 17s1 as the signal - to - noise ratio was very poor and did not permit reliable measurement of any fluxes other than h@xmath8 , h@xmath16 and [ o iii ] 5007 . emission lines . ]
figure [ fig_7 ] shows the emission line rich spectrum of the bright bcd * j0005 - 28 * with flux on a logarithmic scale .
twelve balmer lines can be seen , allowing particularly accurate de - reddening to be calculated .
the de - reddening process based on the h@xmath8 to h@xmath16 ratio gave ratios to within 0.3% for h@xmath18 and 12% for h@xmath21 of the expected values for case b. 4363 line is strong . ]
the electron temperature , t@xmath0 , can be derived from collisionally excited line fluxes , for a variety of ionic species , provided the auroral line is observed ( in the case of o iii , the 4363 line ) .
the method most frequently used makes use of the ratio of fluxes of the bright [ o iii ] lines to the auroral line .
this is a well - established technique ( see discussions in * ? ? ?
* ; * ? ? ?
* ) , but it calculates only the o@xmath22 abundance , not the total gas - phase abundance of oxygen . in most hii regions , the contributions to total oxygen from neutral o and o@xmath23
are minor , so in addition to o@xmath22 we only need to calculate the contribution from o@xmath24 . if the equivalent auroral lines for [ o ii ] , [ n ii ] and [ s iii ] are present in the spectra , the electron temperatures can be calculated using these lines too , and since they peak at different regions in the hii region , the auroral lines collectively sample the complete volume . when these auroral lines are not observed , empirical methods can be used to estimate the o@xmath24 abundance , for example in .
however , as those authors note , the methods depend on having reliable atomic data ( energy levels , transition probabilities , and collision strengths ) .
consequently we have approached the problem again from first principles , using the latest atomic data , to derive the total oxygen abundance .
@xmath17320,30 lines , so the izotov method provides a useful comparison . ] the rate of collisional excitation for o@xmath22 from the @xmath25p ground state(s ) to the @xmath26d@xmath27 level is given , for the thermal equilibrium case @xcite , by , @xmath28 where @xmath29 is the planck constant , @xmath30 is the electron mass , @xmath31 is the statistical weight of the ground state (= 9 for o@xmath22 ) , @xmath32 is the boltzmann constant , @xmath33 is the net effective collision strength for collisional excitations from the ground @xmath25p states to the @xmath26d@xmath27 state and e@xmath34 is the energy level of the @xmath26d@xmath27 state . ignoring the small contribution to the population of the @xmath26d@xmath27 level from radiative cascade from higher energy levels , the rate of emission of photons from that level is equal to the rate of excitation , i.e. , @xmath35 .
the emissivity of [ o iii ] from transitions from the @xmath26d@xmath27 ( @xmath1 5007 , 4959 and 4931 ) level is proportional to @xmath36 .
here we have used the total effective collision strengths for the forbidden @xmath26d@xmath27 to @xmath25p transitions , so we use the flux - weighted photon energy , corresponding to a wavelength of 4997 , for @xmath37 .
the emissivity of h@xmath16 is proportional to @xmath38 @xcite , where @xmath39 is the electron density , @xmath40 is the ionized hydrogen density , and @xmath41 is the effective emissivity for h@xmath16 , which takes into account photon energies and branching ratios , and for which values have been computed by @xcite . given that the ratio of the flux of [ o iii ] to that of h@xmath16 is equal to the ratio of their emissivities multiplied by their photon energies , for a given geometry , one may reorganise the above equations to derive an expression for the ratio of the number density of o@xmath22 ions to hydrogen ions ( i.e. , the o@xmath22 abundance ) in terms of the flux ratio of [ o iii](@xmath26d@xmath27 ) to h@xmath16 , @xmath42 where @xmath43 is the electron temperature derived from the [ o iii ] line ratio , for which there is a simple expression from @xcite , @xmath44 where , for [ o iii ] , @xmath45 and a= 13229 , b= 0.92350 , and c=0.98196 . in an identical fashion
, one may derive an expression for the abundance of o ii using the observed fluxes from the [ o ii ] 3726,3729@xmath1 lines , @xmath46 where , in this case , @xmath43 is the electron temperature derived from the [ o ii ] ratio ( see * ? ? ?
* ) using the ratio of the 7320,30@xmath1 lines to the 3726,3729@xmath1 lines . if , as in the case of these observations , the nir lines are not available , it is possible to derive an expression for the [ o ii ] electron temperature from the mappings photoionization models as a polynomial in terms of _ total _ gas - phase oxygen abundance , @xmath47)=t_e([\text{o~{\sc iii } } ] ) \times ( 3.0794 - 0.086924~z - 0.1053~z^2 + 0.010225~z^3)\ ] ] where z=12+log(o / h )
. this does not provide the final answer , and it is necessary to iterate to a final value for the abundance of o@xmath24 , starting by using the o@xmath22 abundance as the total oxygen abundance .
the process converges in less that five iterations . @xcite and @xcite have used a simpler approach , expressing the low ionization zone temperature ( effectively the [ o ii ] temperature ) in terms of the [ o iii ] temperature , which does not require iteration .
@xcite used a linear relation , whereas @xcite used a more complex fit to photoionization model data .
equation [ eq6a ] shows the expression used by @xcite : @xmath48\ ] ] equation [ eq6a ] gives total oxygen abundance values close to those from iterating equation [ eq6 ] .
values determined for oxygen abundances are not exact , because of the nature of the approximations used , the calculated values for oxygen abundances depend on the photoionization models used to build the models , and the use of a model derived from a single value of the ionization parameter , @xmath49 . testing the two methods ( equations [ eq6 ] and [ eq6a ] ) against artificial data indicates that they generate total oxygen abundances within 1% of the input values .
the iterative approach ( equation [ eq5 ] ) is marginally the more consistent of the two over a range of ionization parameter values .
the above equations may be simplified for computation by using accurate expansions in terms of the [ o iii ] electron temperature to @xmath41 , @xmath50 , and @xmath51 .
the case b emissivity data for h@xmath16 as a function of temperature , from @xcite may be fit with a simple power law , @xmath52 the effective collision strengths for the o@xmath22 @xmath26d@xmath27 level @xcite , from which the @xmath14959 , 5007 doublet originates , can be fit with a simple exponential function of temperature , @xmath53 and the effective collision strengths for o@xmath22 from @xcite for the composite upper state , @xmath54d@xmath55 , from which the @xmath13726,9 doublet originates , can similarly be fit by a linear function of temperature , @xmath56 applying these methods to the present observations , we obtain the electron temperatures and total gas - phase oxygen abundances shown in table [ tx ] .
the values for j1118 - 16s2 are not listed as the @xmath574363 line was not observed .
the uncertainties in t@xmath0 were calculated from the flux error values , in equation [ eq3 ] , and propagated through to the abundance values .
see the appendix for a discussion of the error estimation .
there are two principal methods for determining oxygen abundances from hii region optical spectra , the direct or electron temperature ( t@xmath0 ) method and the strong line methods .
the t@xmath0 method is possible if one of the auroral lines is observed at adequate signal - to - noise ( @xmath58 ) , usually [ o iii ] 4363 .
this is the case in all but one of the galaxies discussed here and is detailed below .
the so - called strong line methods use flux ratios of the prominent nebular lines to determine abundances ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
conventionally , the strong line methods were empirical , calibrated against results using the direct method .
however , recently , @xcite have extensively revised the strong line techniques , developing diagnostic grids based on the mappings photoionization modelling code , using the latest atomic data , and the possibility that the electrons exhibit a non - equilibrium @xmath59 energy distribution @xcite .
the grids are new , and the ratios used have been selected to maximise the orthogonality of the parameters , avoiding to a large extent the degeneracy of older diagnostics , and to solve for both metallicity and ionization parameter .
these new diagnostics generate values for both the oxygen abundance and the ionization parameter , @xmath49 , and give substantially more consistent abundance values than the older methods .
this can be seen by comparing the metallicity results for the different diagnostics from table [ t3 ] and the older diagnostic results listed in table [ tx ] . in table
[ t3 ] , the diagnostics involving the ratio [ n ii]/[s ii ] and [ n ii]/[o ii ] are particularly consistent , differing by typically @xmath20.03 dex . the ionization parameter @xmath49 ( sometimes expressed as @xmath60 , where @xmath61 is the speed of light ) is the ratio of the number of ionizing source photons passing through a unit volume to the neutral hydrogen density .
the photon flux matches the number of new ions it produces , and as @xmath49 has the dimensions of velocity , it can be understood as the maximum speed at which the boundary of the ionized region can move outwards @xcite .
@xmath49 is at its maximum at the inner edge of the ionized region of an hii region , and falls to zero at the outer edge of the ionized nebula , where the ionizing flux is fully depleted .
a problem with older diagnostics such as r@xmath62 is that measured metallicities depend on the ionization parameter .
there have been previous attempts to solve for the ionization parameter @xcite but these new diagnostics solve for its value independently of the metallicity , and consequently , take into account the ionization gradients present in hii regions .
the diagnostics chosen here are also relatively insensitive to non - equilibrium ( @xmath59 ) electron energy distributions , especially for values of @xmath63 20 . in addition , @xmath59 distributions have a smaller effect on the excitation of lower metallicity hii regions than in higher metallicity objects .
for this reason , and in the interests of clarity , we present here only the equilibrium ( maxwell - boltzmann , or @xmath64 ) results
. figures [ fig_8 ] and [ fig_9 ] plot the log flux ratios for the observed objects on the diagnostic grids from @xcite .
not all of the objects can be accommodated within these grids .
there are several possible reasons , and these will be discussed in detail in the second paper in this series .
one likely cause relates to electron densities .
the grids shown here are calculated for an electron density , n@xmath0 @xmath4 5 @xmath5 . for the majority of the observed objects ,
this is accurate , but for some the densities are somewhat higher .
below we analyse the ratios of the two [ s ii ] lines ( 6716 and 6731 ) which are a useful diagnostic of electron density @xcite , and it is clear that some of the objects exhibit higher densities . some of the `` misfit '' points can be accommodated on grids calculated for higher electron densities ( see below and figure [ fig_11 ] ) .
table [ t3 ] lists the oxygen abundances and ionization coefficients computed from the new diagnostic grids using the `` pyqz '' interpolation described in @xcite .
while the interpolation scheme does not always work reliably for near - vertical grid lines , leading to null results , it is clear that different diagnostics yield somewhat different results .
however , the consistency is far better than earlier methods permitted .
we have found that for low metallicity objects ( @xmath2 0.5 z@xmath3 ) , diagnostics listed in table [ t3 ] involving log(n ii / s ii ) give values for the metallicity that differ by typically less than 0.02 dex and diagnostic using the log(n ii / o ii ) ratios are similar .
it is also evident from table [ t3 ] that the direct methods are nearly all lower by @xmath40.13 dex than the best log(n ii / s
ii ) strong line diagnostic values .
this is in agreement with the findings of @xcite that the direct method abundances are generally lower than strong line estimates .
however , with the newer atomic data , the recalculated direct method abundances , and the revised mappings photoionization code , these differences are smaller . for comparison we also present the results of older strong line diagnostics , in table [ tx ] .
perhaps the most variable result is that for j1118 - 17s2 .
this is not surprising , as the [ n ii ] flux is poorly defined .
it appears likely that an oxygen abundance figure of @xmath47.2 ( @xmath4z@xmath3/30 ) is a reasonable estimate .
it is worth noting that the mappings photoionization modelling grids used here take into account the total oxygen abundance , i.e. , both the gas - phase oxygen and that incorporated in dust grains .
when comparing the electron temperature and strong line abundances , it is necessary to increase the electron temperature oxygen abundance values by @xmath40.07dex , to allow for the oxygen in dust grains that the direct method does not account for .
this , of course , assumes a particular level of dust in the ism . in the mappings strong line diagnostic grids
, we assume a 1.0 solar dust depletion . found that the gas - to - dust ratio varies considerably between objects , depending on their individual star formation histories , particularly at the low end of the galaxy mass scale .
this is a further complication for any attempt to estimate the total oxygen abundance in an hii region .
has shown a correlation between the reddening coefficient c(h@xmath16 ) and the gas - to - dust ratio in wolf - rayet galaxies , and it is likely that a similar relation holds for smaller dwarf galaxies .
( we do not have the necessary fir data to allow us to calculate the dust mass here ) .
in the meantime , the direct method oxygen abundance measurements provide a lower limit to the total oxygen .
in this section we use the approach from our previous paper , @xcite .
one of the more important parameters in understanding galactic evolution is the nitrogen abundance , and in particular , the ratio of nitrogen to oxygen .
the observations reported here include relatively low noise measurements of both [ n ii ] and [ o ii ] , allowing us to explore the values of log(n / o ) for each hii region .
to calculate the value of n / o from [ n ii ] and [ o ii ] line fluxes , we use empirical formulae from , equations ( 3 ) and ( 6 ) .
( this approach was chosen because it accounts for the temperature dependencies of the [ n ii ] and [ o ii ] fluxes ) .
these equations reduce to : @xmath65 where @xmath66 is the [ o iii ] electron temperature in units of 10,000k .
this equation differs only by a small constant offset ( 0.033 ) from that quoted by ( * ? ? ?
* equation ( 9 ) ) , most probably due to the latter using older atomic data .. we assume the same electron temperature for o ii and n ii ( reasonable , as they both arise primarily from the outer parts of the hii region ) , and further , that n@xmath24/o@xmath24 = n / o , following @xcite and others .
the errors from these assumptions are likely to be of the same order as the measurement uncertainties .
the results are shown in table [ ty ] . using the abundance values listed in table [ ty ] , we can plot log(n / o ) versus oxygen abundance .
figure [ fig_10 ] shows the data from this work ( yellow triangles ) , data from other sigrid objects from @xcite ( brown triangles ) and data from @xcite ( black circles ) .
the sigrid data are consistent with the van zee results , without any obvious evidence of a floor .
however , @xcite state that the nitrogen floor does not become apparent until the oxygen abundance falls below z=7.7 , so the sigrid data do not resolve the question of whether the floor exists .
while the data for j1118 - 17s2 are not plotted because the value of the [ n ii ] flux is not well defined , the best estimate values for this object ( log(n / o ) @xmath11 2.481 and z @xmath11 7.2 ) extend the trend considerably further in the same direction , off the graph , below and to the left .
if correct , this suggests a very low level for any primary nitrogen , but further observations are necessary to confirm this the red curve is the fit to the van zee data used in the mappings model grids , from ( * ? ? ?
* figure 3 ) , making allowance for the oxygen depletion into dust grains .
there is an increased scatter in the distribution with decreasing oxygen abundance ( z ) .
two possibilities may contribute to this .
first , as there are populations of older stars in these galaxies ( for example , kk246 , * ? ? ?
* ) , so intermediate mass agb stars will contribute nitrogen to the interstellar medium ( ism ) through hot - bottom burning processes .
second , the amounts contributed by such processes will depend on the ( unknown ) star formation histories of different galaxies .
as the abundances derived using the strong line diagnostics depend at least in part on the log(n ii / o ii ) and log(n ii / s ii ) ratios , the results are sensitive to deviations in the nitrogen fit from the theoretical fit used in the mappings models , which was derived from van zee s 1998 data ( see * ? ? ?
* ) , so any error here affects the model outcomes .
figure [ fig_10 ] shows that , at lower metallicities , the data exhibit increasing scatter and may even have started to fork into two branches .
the upper region of the scatter may indicate nitrogen enrichment by wolf - rayet wn stars , as suggested by .
smaller galaxies may divide into two classes , those with ( or that have had ) wn stars , and those without , depending on the stochastic nature of individual star formation events . @xcite
have suggested the [ n / o ] ratio of hii regions in a galaxy arise from nitrogen that is significantly primary in origin , and are a measure of the early star formation history .
while this may be correct for larger galaxies , at least in the case of the very isolated dwarf galaxy kk246 , it is not the case , as the log(n / o ) ratio is low but there is evidence of older stellar populations @xcite . the presence or absence of wn stars in a dwarf galaxy s hii regions is a plausible explanation for this scatter or bifurcation .
this would be consistent with the observations of the blue compact dwarf galaxy , hs0837 + 4717 .
the object appears to harbor over 100 wolf - rayet stars and has both a very low oxygen abundance and a high nitrogen abundance .
the flux ratios of the two [ s ii ] lines at 6716 and 6731 are good indicators of electron density ( section 5.6 , * ? ? ?
table [ t8 ] shows the variation of this ratio calculated for electron densities n@xmath0 of @xmath4 5 and @xmath450 @xmath5 , for an ionization parameter @xmath67 = 7.5 , typical for hii regions .
the trends in the [ s ii 6716]/[s ii 6731 ] line ratio are due to two factors : ( 1 ) the relatively small dependence of the line ratio on n@xmath68 resulting from the collisional excitation rates of the s ii line upper states , and ( 2 ) the use of the isobaric setting in the mappings photoionization modelling code , such that the density structure of the s ii region is a function of the ( varying ) temperature within it , which depends on metallicity .
table [ ty ] shows the measured [ s ii ] line ratios and electron densities , calculated using pyneb @xcite , for all objects except j1118 - 17s2 , for which we have no electron temperature . comparing the observed s ii line flux ratios to table [ t8 ] , it is reasonable to conclude that the mappings values show j0005 - 28 , j1152 - 02a&b , j1403 - 27 , j2039 - 63a ,
j2234 - 04b , j2254 - 26 have electron densities n@xmath69 5 @xmath5 , while the remainder have n@xmath70 5 @xmath5 .
this is confirmed quantitatively using pyneb to estimate the actual electron densities .
figure [ fig_11 ] shows the diagnostic grids for o iii / h@xmath16 versus n ii / s ii at the two electron densities the blue ( upper ) grid is for n@xmath71 50 @xmath5 , the green ( lower ) is for @xmath45 @xmath5 .
it is clear that all but j2254 - 26 can be accommodated even on the higher electron density grid .
the abundances for each object are very similar on both grids , but the estimated ionization parameter @xmath67 changes .
similar results apply for the other diagnostic grids .
it is interesting to note that j2254 - 26 has the highest calculated electron density of the observed objects . ) versus log(nii / oii ) for electron densities n@xmath715 and @xmath450 @xmath5 ( isobaric case , @xmath72 = 5 and 6 , respectively , where @xmath73 is the pressure and @xmath74 is the boltzmann constant . ) ] figure [ fig_12 ] shows electron temperature , plotted versus gas - phase oxygen abundance , z , from table [ tx ] .
z ( = 12+log(o / h ) ) is derived using the formulae in equations [ eq2 ] and [ eq5 ] .
the quadratic fit to the data with 66% confidence errors is : @xmath75 while figure [ fig_12 ] may be used to calculate the total gas - phase oxygen abundance from the [ o iii ] electron temperature , it applies only to the data presented here , and we will provide a substantially more accurate fit to the model - derived curve in paper 2 . , versus gas - phase oxygen abundance for the sigrid objects , calculated using equations [ eq2 ] and [ eq5 ] . ]
mass ( or luminosity ) versus metallicity behavior is one of the important evolutionary diagnostics for galaxies .
it has been extensively mapped for larger galaxies ( e.g. , * ? ? ?
* ) , but it is less well known for dwarf galaxies
. it has been studied by several authors @xcite . exploring it was one of the initial motivations for the sigrid sample @xcite .
figure [ fig_13 ] shows the gas - phase oxygen abundance versus neutral hydrogen mass ( left panel , data from @xcite and table [ t1 ] ) and gas - phase oxygen abundance versus absolute b - band magnitude ( right panel ) .
there is no clear trend in the first graph , suggesting the neutral hydrogen mass is not strongly correlated with metallicity , at least in this sample . in the right panel
, we compare the sigrid data against data from the @xcite survey of galaxies in the lynx - cancer void .
the sigrid data are consistent with the pustilnik et al .
data , and both samples are selected for isolation .
the four blue points in the right panel are objects from pustilnik et al . , but which meet the more stringent selection criteria for the sigrid sample , for luminosity , galaxy type and isolation ( distance from nearest neighbor ) .
the trend line is from @xcite for field di galaxies , but which were not otherwise selected for isolation .
both the pustilnik and sigrid data tend to fall below the line , indicating that more isolated objects have slightly lower metallicities than similar objects in more congested regions , as suggested by @xcite .
there is insufficient data in our observations to confirm the increasing spread of metallicity values at low mass , as implied by ( * ? ? ?
* figure 6 ) . however , the log(n / o ) data ( figure [ fig_10 ] ) are consistent with such a spread . ) .
right panel : oxygen abundance versus absolute b - band magnitude , comparing data from figure 3 from @xcite with sigrid data .
eleven sigrid objects have measured b - band magnitudes .
the four blue points are those from the pustilnik sample which meet the luminosity , galaxy type , and and isolation selection criteria for sigrid .
the straight line fit is for field di galaxies from @xcite .
the b - band magnitudes for the sigrid sample have been taken from the nasa extragalactic database ( ned ) . ] in this work , we have calculated the gas - phase oxygen abundance using the electron temperature direct method routine developed here , and the diagnostic grids .
table [ tx ] compares the electron temperature abundances using the methods described here with those using the iterative method from , with the same input temperatures ; and the most reliable values from the diagnostic grids , those using the log(n ii / s ii ) and log(n ii / o ii)diagnostics .
the values derived using the two t@xmath0 methods are similar , suggesting that the method developed here is reliable . see also a discussion of this in the forthcoming paper 2 .
it is interesting to note that the diagnostic grid abundances , with two notable exceptions , are consistently a little higher than the direct method values , consistent with the findings of @xcite .
the average value of the difference @xmath18z is [email protected] .
the complete explanation of this difference is unclear , but in part it can be explained by the nature of the diagnostic grids derived from the mappings photoionization modelling code . in the grids , we have assumed a dust depletion of 1.0 solar @xcite , to account for the elements locked up in dust grains .
this leads to an overestimate of 0.07 dex in the abundance values derived from the diagnostic grids , compared to the gas - phase - only oxygen abundances from the direct method , explaining about half of the discrepancy .
it is useful to examine two of the `` outliers '' in table [ tx ] , where the diagnostic grid oxygen abundances differ substantially from the electron temperature oxygen abundances .
j1225 - 06s2 has a very low oxygen abundance , @xmath47.45 , from the direct method , and @xmath47.9 from the grids .
this could be explained if there is more n ii than implied by the mappings models parameters , although this is not obvious from table [ ty ] .
there may also be increased nitrogen due to enrichment by wn stars , as in the case of hs 0837 + 4717 .
the oxygen abundance discrepancy for j1609 - 04(2 ) is very likely a result of uncertainty in the flux of the [ o iii ] 4363 auroral line , which is weak in this object .
the diagnostic grids themselves ( table [ t3 ] ) are somewhat discrepant in the values yielded for oxygen abundances . in particular , two trends are clear .
first , diagnostics involving log(n ii / s ii ) are particularly consistent , and the closest to the oxygen abundances derived using the direct method .
diagnostics involving [ o ii ] fluxes are nearly as consistent .
this concordance and consistency lead us to believe that these diagnostics are the most reliable , and we have used the means of the log(n ii / s ii ) and log(n ii / o ii)diagnostics in table [ tx ] .
second , diagnostics involving log(n ii / h@xmath8 ) give somewhat higher oxygen abundances than both the other diagnostics and the direct method values .
the source of these discrepancies is unclear , but may be related to the abundance fit for nitrogen used in mappings .
they do not materially affect the results reported here , provided we rely on the log(n ii / s ii ) and log(n ii / o ii ) diagnostics and direct method oxygen abundances . in the second paper examining these observations
, we will explore the effect of three - dimensional diagnostic charts .
these use three independent diagnostic ratios plotted and explored in three dimensions , whose purpose is to investigate whether the observations lie on a diagnostic plane , along the lines of vogt at al .
( 2014 , in prep . ) .
we will examine the effects of using diagnostic grids calculated using higher electron densities .
we will investigate the effects of optically thin hii regions , and show that they can have considerable effects on the diagnostics , and that there is evidence of optical thinness in some of the observed objects .
we will re - examine the electron temperature versus oxygen abundance plots , for both these observations and for 124 sdss objects from . using the mappings photoionization modelling code
, we will demonstrate that with reasonable assumptions about the star clusters exciting hii regions , there is an effective upper limit to the temperature that can be reached , even in optically thin regions .
the implications appear to be that some of the high temperatures reported in low metallicity hii regions may be somewhat in error .
we will demonstrate the effect of taking into account the additional contribution to total oxygen abundance of the oxygen in dust grains .
we will also suggest that the apparent spread in metallicities at the low end of the mass - metallicity relation are due to stochastic effects in stellar mass distributions in the small star clusters exciting hii regions in small irregular galaxies .
in this paper we have presented the results of observations of seventeen hii regions in thirteen small isolated dwarf irregular galaxies , most from the sigrid sample , all but one exhibiting the [ o iii ] auroral line .
all have measured oxygen abundances @xmath28.2 ( @xmath20.3 z@xmath76 ) , one has an apparent abundance of 7.44 and another very low metallicity object with [email protected] .
we have derived a method for calculating total gas - phase oxygen abundances using only the optical spectra between 3700 and 7000 .
this method gives very similar results to previous empirical fit methods . from an analysis of abundances and ionization coefficients using the diagnostic grids developed by @xcite
, we find the direct method oxygen abundances are consistently within 0.07 dex of the strong line diagnostic results , making allowance for the oxygen locked up in dust grains . from the line ratio of the two red [ s ii ] lines we find that the electron densities occurring in the objects observed are between @xmath45 and 100 @xmath5 .
the nitrogen abundance , as expressed in log(n / o ) , continues the trend evident in @xcite , but from this sample we find no clear evidence for a nitrogen floor .
there is increased scatter at lower oxygen abundances , and some evidence for a bifurcation in the trend , possibly due to the presence of wn stars in some of the hii regions .
the slope of the luminosity
metallicity relation for these observations is very close to that for void galaxies in @xcite .
the spectra from an apparently very low metallicity galaxy , j1118 - 17s2 , show no nitrogen lines : we intend to undertake follow up observations on this galaxy to estimate the metallicity more accurately .
mike dopita acknowledges the support of the australian research council ( arc ) through discovery project dp0984657 .
this work was funded in part by the deanship of scientific research ( dsr ) , king abdulaziz university , under grant no .
( 5 - 130/1433 hici ) .
the authors acknowledge this financial support from kau . 55 natexlab#1#1 , b. h. & martini , p. 2013
, , 765 , 140 , d. a. , skillman , e. d. , garnett , d. r. , et al .
2013 , , 775 , 128 , d. a. , skillman , e. d. , marble , a. r. , et al .
2012 , , 754 , 98 , m. s. 1999 , , 111 , 1426 , j. a. , clayton , g. c. , & mathis , j. s. 1989 , , 345 , 245 , m. j. , vogt , f. p. a. , nielsen , j. , & sharp , r. g. 2013 , , t. , treyer , m. a. , sullivan , m. , & ellis , r. s. 2002 , , 330 , 75 , m. a. , hart , j. , mcgregor , p. , et al .
2007 , , 310 , 255 , m. a. , rhee , j. , farage , c. , et al .
2010 , , 327 , 245 , m. a. & sutherland , r. s. 2003 , astrophysics of the diffuse universe ( springer verlag ) , m. a. , sutherland , r. s. , nicholls , d. c. , kewley , l. j. , & vogt , f. p. a. 2013 , , 208 , 10 , m. g. & pagel , b. e. j. 1978 , , 185 , 77p , j. & dopita , m. 2005 , , 619 , 340 , h. u. , mende , s. b. , arens , j. f. , mccullough , p. r. , & swenson , g. r. 2000 , , 27 , 41 , d. r. 1992 , , 103 , 1330 , m. , walker , a. r. , suntzeff , n. b. , et al . 1992 , , 104 , 533 , y. i. , stasiska , g. , meynet , g. , guseva , n. g. , & thuan , t. x. 2006 , , 448 , 955 , l. j. & dopita , m. a. 2002 , , 142 , 35 , l. j. & ellison , s. l. 2008 , , 681 , 1183 , h. , mccall , m. l. , kingsburgh , r. l. , ross , r. , & stevenson , c. c. 2003 , , 125 , 146 , h. , skillman , e. d. , cannon , j. m. , et al .
2006 , , 647 , 970 , .
r. 2010 , , 521 , a63 , .
r. , dopita , m. a. , kewley , l. j. , et al .
2012 , , 426 , 2630 , .
r. & esteban , c. 2010 , , 517 , a85 , v. , morisset , c. , & shaw , r. a. 2012 , in iau symposium , vol .
283 , iau symposium , 422423 , s. s. 1994 , , 426 , 135 , g. r. , hanish , d. j. , ferguson , h. c. , et al .
2006 , , 165 , 307 , m. j. , zwaan , m. a. , webster , r. l. , et al .
2004 , , 350 , 1195 , d. c. , dopita , m. a. , jerjen , h. , & meurer , g. r. 2011 , , 142 , 83 , d. c. , dopita , m. a. , & sutherland , r. s. 2012 , , 752 , 148 , d. c. , dopita , m. a. , sutherland , r. s. , kewley , l. j. , & palay , e. 2013 , , 207 , 21 , d. c. , jerjen , h. , dopita , m. a. , & basurah , h. 2014 , , 780 , 88 , j. b. 1990 , , 99 , 1621 , d. e. & ferland , g. j. 2006 , astrophysics of gaseous nebulae and active galactic nuclei , 2nd edn .
( university science books ) , b. e. j. , simonson , e. a. , terlevich , r. j. , & edmunds , m. g. 1992 , , 255 , 325 , e. , nahar , s. n. , pradhan , a. k. , & eissner , w. 2012 , , 423 , l35 , e. & contini , t. 2009 , , 398 , 949 , e. , vlchez , j. m. , cedrs , b. , et al .
2011 , , 532 , a141 , l. s. & thuan , t. x. 2005 , , 631 , 231 , l. s. , vlchez , j. m. , & thuan , t. x. 2010 , , 720 , 1738 , s. , kniazev , a. , pramskij , a. , et al .
2004 , , 419 , 469 , s. a. , tepliakova , a. l. , & kniazev , a. y. 2011 , astrophysical bulletin , 66 , 255 , a. , madden , s. c. , galliano , f. , et al .
2013 , , 536 , a31 , e. d. , televich , r. j. , kennicutt , jr . , r. c. , garnett , d. r. , & terlevich , e. 1994 , , 431 , 172 , p. j. & hummer , d. g. 1995 , , 272 , 41 , e. , chun , m. , freeman , k. c. , & chaboyer , b. 2002 , in astronomical society of the pacific conference series , vol .
273 , the dynamics , structure & history of galaxies : a workshop in honour of professor ken freeman , ed .
g. s. da costa & h. jerjen , 341347 , s. m. , drinkwater , m. j. , meurer , g. , et al .
2014 , , 782 , 35 , s. s. 2007 , , 171 , 331 , c. a. , heckman , t. m. , kauffmann , g. , et al .
2004 , , 613 , 898 , p. g. 2001 , , 113 , 1420 , l. & haynes , m. p. 2006
, , 636 , 214 , l. , salzer , j. j. , & haynes , m. p. 1998 , , 497
, l1 , l. , salzer , j. j. , haynes , m. p. , odonoghue , a. a. , & balonek , t. j. 1998 , , 116 , 2805 , m. b. & edmunds , m. g. 1993 , , 265 , 199 , f. p. a. , dopita , m. a. , & kewley , l. j. 2013 , , 768 , 151
this appendix describes the methods used to estimate emission line flux uncertainties for spectra extracted from wifes data cubes of objects in the sigrid sample .
the data reduction process is described in detail in , and in this paper in section 2.3 .
briefly , the steps where noise is involved or systematic errors are incurred are bias subtraction , flat fielding , cosmic ray removal , sky - line subtraction using nod - and - shuffle , and standard flux star calibration . the principle sources of uncertainty are the ccd detector and amplifier readout noise , and the amplification of this noise through the data reduction chain ; the effects of cosmic rays and sky lines , and their removal ( partial or complete ) ; the calibration of the emission line fluxes using standard star flux data ; the de - reddening process ; and the measurement of the line fluxes from the flux - calibrated spectra .
as the ifu data frame is convolved into a data cube in the pipeline , the process of error calculation is more complex than for single slit or echelle spectroscopy . for the spectral noise uncertainties ,
there are two approaches we could take .
one is to estimate the errors accumulating from each step , such as described for echelle spectroscopy by @xcite .
the other approach , used here , is to measure the statistical noise from spectra extracted from the reduced data cube , and to estimate the systematic errors arising from the de - reddening and flux calibration , which are independent of the statistical noise . unlike single slit spectra , with ifu data cubes , we are able to select the entire area of the hii region from which to extract the spectrum , and exclude the majority of the galaxy stellar background , resulting in better signal - to - noise .
note that the statistical noise varies with the size of the sampled spaxel area , due to averaging .
for the objects in this study , sampling using a 6 arc sec diameter circular spatial area maximises the amount of flux from the hii region and minimises both the statistical noise , though averaging , and the stellar continuum from the area outside the hii region . in every case , the galaxies were so faint that the stellar extent was at best barely detectable .
however , images from the dss survey and from the singg data ( as illustrated in * ? ? ?
* ) suggest the individual galaxies are less than or approximately equal to the fov of the wifes spectrograph , 25@xmath1038 arc sec .
the benefit of the ifu is that the sample was centered on the hii region , and excluded virtually all areas of the galaxies without hii emission .
line fluxes and noise were measured from the extracted spectra using iraf / splot .
the standard splot ` k - k ' method was used to fit a gaussian to each emission line , to measure the equivalent width ( where possible ) , the gaussian full width at half maximum , and the integrated flux .
noise was measured on both sides of the emission line using the splot ` m - m ' method .
these results were checked using the deblend ` d - d ' method , but using a single line , which automatically generates values of the same parameters .
particular care was taken to account for any stellar absorption features underlying the balmer emission lines , although in all cases , this was minor or absent , due to low stellar continuum .
in fact , the stellar continuum was extremely faint , with the exception of the object j0005 - 28 ( see figure [ fig_7 ] , displayed on a log - intensity scale ) .
test sample sizes showed that all the detectable h@xmath8 and [ o iii ] in each hii region lay within the sample aperture .
the observed fluxes mostly peak at or less than a radius of 2.5 arc sec , except where there are closely adjacent hii regions ( e.g. , j1609 - 04 ) . for these , limiting the sample size to 6 arc sec diameter avoided sampling a different hii region . ideally , single spaxel - based analysis would be preferable to multi - spaxel sampling , but these objects are so faint that the resultant noise is prohibitive . detector noise is added to the data frame during bias subtraction and flat fielding , as the bias and flat field frames used also incur readout noise .
the sky subtraction process using the nod and shuffle process or the sky frame method adds additional noise during the subtraction process .
nod - and - shuffle sky subtraction was used for all hii region observations , with sub - exposure times chose to be shorter than the shortest observed fluctuation in the oh airglow lines @xcite .
the removal of the critical oh lines is effectively complete in all observations .
the [ o i ] airglow lines are at wavelengths that did not interfere with any of the observed hii region spectral lines .
cosmic ray removal is reasonably efficient , using the laplacian kernel technique described by @xcite .
the process is not perfect , but virtually all the remaining cosmic ray artefacts are removed using the process . in isolated cases ,
separate cosmic rays occur on all object data frames at the same location , and this can lead to erroneous results , but this can be detected by the labor intensive process of inspecting all the lines on all the ifu slitlets ( 25 ) on all the data frames ( usually 3 ) .
figure [ cosray ] shows part of a raw wifes blue data frame including segments of 7 slitlets ( of a total 25 ) , centered on the h@xmath17 and [ o iii ] @xmath57 4363 auroral line , for the galaxy j0005 - 28 .
the auroral line is very prominent in this frame , to the left of the h@xmath17 line .
[ cosray ] line , showing sections of 7 slitlets .
each frame was inspected to check for cosmic ray contamination of key lines .
note that , due to the optical paths in the wifes ifu blue camera , the right side of the image corresponds to shorter wavelengths.,title="fig : " ] in addition to the intrinsic statistical noise amplified through the data reduction pipeline , when measuring the emission line fluxes , it is necessary to take into account any broad absorption lines in the stellar continuum , on which the nebular spectra are superimposed .
there are three approaches here .
the first is to correct for an assumed 2 ew absorption in each line , as described in @xcite .
the second is to use an automated method such as the lzifu idl program developed by several workers at the university of hawaii , which fits model stellar continua to observed spectra and then calculates the emission line fluxes ( a paper on this application is planned ) .
the third approach , which we use here , works better when the stellar continuum is weak , as with the objects reported here .
it involves manual fitting of gaussian profiles to the emission lines using standard iraf / splot methods .
the technique described by @xcite is very similar in detail to the method used here .
the errors arising from the de - reddening process are due to uncertainties in the nature and amount of dust between the nebular emission and the observer . in the case of the sigrid objects considered here , all are further than 10@xmath77 from the galactic equator , to avoid significant reddening by milky way dust .
we calculated the de - reddening using two independent methods and used the differences between the results as an estimate of the de - reddening errors .
we used the dust reddening formulae from @xcite with a@xmath13=3.1 , adjusting the de - reddening to set the resultant balmer h@xmath8/h@xmath16 flux ratios to the @xcite case b balmer ratios for the calculated [ o iii ] electron temperature .
we used the ratios of h@xmath17 and h@xmath18 to h@xmath16 as confirmation . to confirm these results , we employed the dust models from @xcite , using a relative extinction curve with r@xmath14=4.3 , where r@xmath14 = a@xmath13/(e@xmath15 ) and a@xmath13 is the v - band extinction .
this is discussed in more detail in ( * ? ? ?
* appendix 1 ) .
we used an initial balmer decrement ratio of 2.82 for h@xmath8/h@xmath16 , corresponding to an electron temperature of 12 500k , adjusted the electron temperature using the direct method derived from the [ o iii ] line ratios , then adjusted the apparent balmer ratios by varying the value of a@xmath13 for the best fit to the h@xmath17/h@xmath16 ratio , using the ratio h@xmath18/h@xmath16 as a check , again fitting to the storey and hummer case b balmer ratios .
the de - reddened flux values reported in table [ t2 ] are those using the cardelli method .
in all cases , the two approaches gave similar results : the average difference between the two methods for the important diagnostic lines varies between 0.1% and 0.7% . as a consequence
, we have adopted a figure of 1% for the de - reddening error . in only one case , j2234 - 04 , object a , did the de - reddening fail to provide a plausible result , and this has been excluded from the results reported here .
it appears likely that two or more incompletely removed sets of cosmic ray artefacts were the cause of the problem , in this particular case .
flux calibration errors depend on how well one can fit flux - calibrated vales to the standard stars .
the standard stars used were taken from @xcite , using bessell s recalibration of hamuy s southern spectrophotometric standards @xcite .
calibrating to these standards is likely to be more accurate than the older oke standards @xcite , but we have retained an estimated 2% error for calibrating the flux values , as per @xcite . we have recently corrected a problem with the measured [ o ii ] line fluxes .
the source of the problem is a sharp absorption edge at 3850 , due to the adhesive used in the beam splitter , and the lack of any significant output below about 3900 from the lamps used for flat exposures to identify and correct this problem .
the lamps are now being replaced , and future measurements using wifes will no longer require this compensation .
the problem was understood in the testing phase of the construction of wifes , and was identified in the observations from poor matches to the diagnostic grids that involve [ o ii ] in the ratios . to compensate requires boosting the [ o ii ] flux by a factor of @xmath101.5 .
it affects only the [ o ii ] @xmath1 3726,9 lines .
it does not make a substantial difference to the calculated oxygen abundances or other results .
|
in this paper we present the results of observations of seventeen hii regions in thirteen galaxies from the sigrid sample of isolated gas rich irregular dwarf galaxies .
the spectra of all but one of the galaxies exhibit the auroral [ o iii ] 4363 line , from which we calculate the electron temperature , t@xmath0 , and gas - phase oxygen abundance .
five of the objects are blue compact dwarf ( bcd ) galaxies , of which four have not previously been analysed spectroscopically .
we include one unusual galaxy which exhibits no evidence of the [ n ii ] @xmath1 6548,6584 lines , suggesting a particularly low metallicity ( @xmath2 z@xmath3/30 ) .
we compare the electron temperature based abundances with those derived using eight of the new strong line diagnostics presented by @xcite . using a method derived from first principles for calculating total oxygen abundance ,
we show that the discrepancy between the t@xmath0-based and strong line gas - phase abundances have now been reduced to within @xmath40.07 dex .
the chemical abundances are consistent with what is expected from the luminosity metallicity relation .
we derive estimates of the electron densities and find them to be between @xmath45 and @xmath4100 @xmath5 .
we find no evidence for a nitrogen plateau for objects in this sample with metallicities 0.5 @xmath6 z@xmath3 @xmath6 0.15 .
|
imaging atmospheric cherenkov telescopes use large mirror areas to reflect the cherenkov photons from cosmic - ray and gamma - ray air showers onto a photo - detector camera , usually comprised of photo - multiplier tubes ( pmts ) .
the number of photoelectrons generated at each pmt is directly proportional to the charge under the pmt output pulse , and this is easily measured using adcs with fixed - length integration gates .
an alternative method is to use `` flash '' adcs to rapidly sample the output pulse and record the complete pulse shape .
this allows to maximize the signal - to - noise ratio for individual pulses at the analysis stage , lowering the effective energy threshold of the telescope @xcite .
a number of authors have suggested that the timing and pulse shape information could also be used to improve sensitivity through improved gamma - hadron separation or better reconstruction of shower parameters ( core location , primary energy , etc . )
@xcite . [ cols="^,^ " , ]
a preliminary examination of the timing information provided by the veritas fadcs verifies that the effective charge integration gate can be reduced to @xmath0 .
while the muon background currently limits the analysis threshold of the first veritas telescope , the reduced charge integration gate should result in a correspondingly reduced energy threshold for analysis when stereo data becomes available .
this preliminary study does not indicate a significant improvement in gamma - hadron separation using the timing information ; however , this situation may also change when the muon background is removed .
refinements to the simulations , the application of advanced digital signal processing methods and an increased dataset of gamma - ray source observations should all help future studies in this area .
|
the 499 pixel photomultiplier cameras of the veritas gamma ray telescopes are instrumented with 500mhz sampling flash adcs .
this paper describes a preliminary investigation of the best methods by which to exploit this information so as to optimize the signal - to - noise ratio for the detection of cherenkov light pulses .
the fadcs also provide unprecedented resolution for the study of the timing characteristics of cherenkov images of cosmic - ray and gamma - ray air showers .
this capability is discussed , together with the implications for gamma - hadron separation .
|
standard cold dark matter ( cdm ) and its variants are the current leading models for the origin of large - scale structure . the canonical cdm model hypothesizes a primordial scale - free spectrum of primordial adiabatic perturbations and a critical density of cold dark matter .
its variants include low - density ( e.g. , @xmath1 ) models , either in an open universe or in a flat universe with a cosmological constant , tilted models in which the power - law index @xmath2 of the primordial power spectrum differs slightly from scale - free ( @xmath3 ) , low
hubble - constant models , or mixed - dark - matter models in which the universe has a critical density but roughly 30% is in the form of hot dark matter .
although most of the matter in these models does not undergo gravitational collapse until relatively late in the history of the universe , some small fraction of the mass is expected to collapse at early times .
ultraviolet photons released by this early generation of star and/or galaxy formation will partially reionize the universe , and these ionized electrons will re - scatter at least some cosmic microwave background ( cmb ) photons after recombination at a redshift of @xmath4 .
theoretical uncertainties in the process of star formation and the resulting ionization make precise predictions of the ionization history difficult .
constraints to the shape of the cmb blackbody spectrum and detection of cmb anisotropy at degree angular scales suggest that if reionization occurred , the fraction of cmb photons that re - scattered is small .
still , estimates show that even if small , at least some reionization is expected in cdm models @xcite : for example , the most careful recent calculations suggest a fraction @xmath5 of cmb photons were re - scattered @xcite . scattering of cmb photons from ionized clouds will lead to anisotropies at arcminute separations below the silk - damping scale ( the ostriker - vishniac effect ) @xcite .
these anisotropies arise at a higher order in perturbation theory and are therefore not included in the usual boltzmann calculations of cmb anisotropy spectra .
the level of anisotropy is expected to be small and it has so far eluded detection .
however , these anisotropies may be observable with forthcoming cmb interferometry experiments @xcite that probe the cmb power spectrum at arcminute scales . in this paper
, we present a new derivation of the anisotropy spectrum from the ostriker - vishniac effect .
we then calculate the anisotropy expected in cdm models .
we provide numerical results for the anisotropy predicted by a variety of cdm models with several plausible ionization histories .
we also provide simple semi - analytic fitting formulas for the ostriker - vishniac power spectrum that can be used to evaluate the expected anisotropy in cdm models for any arbitrary ionization history .
we estimate the anisotropy amplitude that may be detectable with forthcoming interferometry experiments .
we find that the signal should be detectable in a number of cdm models with plausible ionization histories . to a good approximation ,
the damping of the acoustic peaks in the primary ( linear - theory ) cmb spectrum from reionization is the same for any ionization history that produces the same optical depth to the surface of last scatter ( see , e.g. , ref .
@xcite ) . therefore , although map and planck @xcite should provide a good estimate of this optical depth @xcite , they will not strongly constrain the ionization history , that is , whether reionization occurred earlier or later . on the other hand ,
our results indicate that the amplitude of the ostriker - vishniac anisotropy depends primarily on the ionization fraction and only secondarily on the optical depth .
therefore , when combined with the determination of the optical depth from map and planck , the ostriker - vishniac effect can determine the epoch of reionization . this will be important for precise determination of the baryon density , hubble constant , and cosmological constant from the higher acoustic peaks in the cmb spectrum .
it will also provide a unique window to the epoch at which structures first undergo gravitational collapse in the universe .
the scale factor of the universe , @xmath6 , satisfies the friedmann equations , @xmath7 @xmath8,\ ] ] where @xmath9 km sec@xmath10 mpc@xmath10 is the hubble constant , @xmath11 is the current nonrelativistic - matter density in units of the critical density , @xmath12 is the current contribution of the cosmological constant to the critical density , and the overdot denotes a derivative with respect to time .
if @xmath13 , the universe is flat ; if it is greater ( less ) than unity the universe is closed ( open ) . an einstein - de sitter universe has @xmath14 and @xmath15 , so @xmath16 .
we choose the scale factor such that @xmath17 .
if we are located at the origin , @xmath18 , then an object at redshift @xmath19 is at a comoving distance , @xmath20 we define a conformal time by @xmath21 , so the comoving distance to the horizon is the conformal time today , @xmath22 . in an einstein - de sitter universe , @xmath23 and @xmath24 with our conventions .
if the density contrast at comoving position @xmath25 at time @xmath26 is @xmath27 , with fourier transform @xmath28 then the power spectrum @xmath29 is defined by the expectation value @xmath30 where the angle brackets denote an average over all realizations . in linear theory ,
the spatial and time dependence of @xmath31 can be factorized , so @xmath32 , where @xmath33 is the age of the universe , @xmath34 , and the growth factor ( written as a function of redshift ) is @xcite @xmath35 ^ 3 } \ , dz'.\ ] ] the time @xmath26 is related to the redshift by @xmath36 therefore , the power spectrum is @xmath37 , where @xmath38 . in an einstein - de sitter universe , @xmath39 , and @xmath40 . in linear theory ,
the fourier components of the velocity field are related to those of the density field by @xmath41 it can also be shown @xcite that @xmath42 ^ 2\,d(z)}.\ ] ]
the ostriker - vishniac effect is the cmb anisotropy produced by scattering from ionized regions or clouds with bulk peculiar velocities .
the fractional temperature perturbation in the direction @xmath43 induced by bulk motions is @xmath44 a(w ) dw , \label{eq : startingpoint}\ ] ] where @xmath45 is the electron density along the line of sight , @xmath46 is the bulk velocity at position @xmath25 at a conformal time @xmath47 , @xmath48 is the thomson cross section , and @xmath49 is the optical depth from us to @xmath50 .
the factor of @xmath51 arises because the physical time is @xmath52 .
[ note that @xmath43 represents a three - dimensional unit vector along the line of sight , whereas @xmath53 will refer to a dimensionless two - dimensional vector in the plane perpendicular to the line of sight
i.e . , for directions near @xmath54 , @xmath55 , whereas @xmath56 . ] the visibility function , @xmath57 is the probability distribution for scattering from reionized electrons . here
, @xmath58 is the mean electron density , where @xmath59 is the baryon density today in units of the critical density ( actually , this number should be multiplied by 7/8 to account for the neutrons bound in helium nuclei ) ; @xmath60 is the critical density ; @xmath61 is the ionization fraction ; and @xmath62 is the proton mass .
the visibility function is normalized so that @xmath63 where @xmath64 is the optical depth to the standard - recombination surface of last scatter , the fraction of cmb photons which scattered after standard recombination at @xmath4 . with these developments ,
the fractional temperature perturbation becomes @xmath65 where @xmath66 \vecv(\vecw , w)$ ] . according to eq .
( [ eq : intermediateone ] ) , only the component of @xmath67 along the line of sight contributes to the anisotropy . in linear theory , we neglect @xmath68 and the total anisotropy is the sum of the contributions of each fourier component of the velocity field
. however , in linear theory ( and in fact to all orders ) , @xmath69 , so only @xmath70 modes along the line of sight can contribute to the anisotropy .
the contributions of troughs and crests of each fourier mode cancel approximately when projected along the line of sight .
therefore , the contribution of bulk velocities to anisotropies vanishes ( or is very small ) to lowest order in perturbation theory on small angular scales . on larger angular scales ,
there are fewer crests and troughs , so the cancellation is not as complete . in other words , angular correlations ( on small angular scales )
can be due only to @xmath70 modes perpendicular to the line of sight , and these contribute nothing to the anisotropy from bulk velocities alone .
furthermore , no anisotropy can be produced by @xmath71 modes parallel to @xmath70 , and therefore the anisotropy can be due only to the components @xmath72 perpendicular to @xmath70 @xcite . in the appendix , we make this argument more precise and show that the projection , eq .
( [ eq : intermediateone ] ) , has fourier coefficients that satisfy @xmath73 with @xmath74 in a flat universe [ making the replacement eq .
( [ eq : replacement ] ) for a closed or open universe ] where @xmath75 is the three - dimensional power spectrum for @xmath72 .
we now need to determine this three - dimensional power spectrum .
again , @xmath69 , so @xmath72 can only come from the fourier transform of the nonlinear term @xmath76 .
the fourier transform of @xmath77 [ not the full @xmath78 is @xmath79 \nonumber \\ & = & { i a \dot d d \over 2 d_0 ^ 2 } \int\ { d^3 \veck ' \over ( 2 \pi)^3}\ , \tilde \delta_0(\veck ' ) \tilde\delta_0(\veck-\veck ' ) \left ( { \veck -\veck ' \over |\veck -\veck'|^2 } + { \veck ' \over k'^2 } \right).\end{aligned}\ ] ] the components of @xmath80 perpendicular to @xmath70 are @xmath81 for a gaussian density field , @xmath82.\end{aligned}\ ] ] we combine these expressions to find ( after some extensive although straightforward algebra ) @xmath83 with @xmath84 the vishniac power spectrum is @xcite @xmath85 ( the differences with eq .
( 2.13 ) in @xcite are the result of differing fourier conventions . )
assembling these results , in a flat universe , the ostriker - vishniac effect produces a power spectrum , @xmath86 ^ 2 \left ( { \dot d d \over d_0 ^ 2 } \right)^2 s(\kappa / w)\ , dw , \label{eq : vishniacresult}\ ] ] with the vishniac power spectrum @xmath87 given by eq .
( [ eq : sofk ] ) . in an open or closed universe ,
one makes the replacement eq .
( [ eq : replacement ] ) in the argument of @xmath88 and in the two factors of @xmath50 in the denominator of the integrand.the numbers for the open models in the table and the open - model curves need to be fixed .
if the visibility function is approximated as a gaussian in conformal time , @xmath89,\ ] ] where @xmath90 is the comoving distance to the re - scattering surface , @xmath91 is the width of the re - scattering surface , and @xmath64 is the fraction of cmb photons re - scattered , then the cmb power spectrum is @xmath92 ^ 2 s(\kappa / w_r).\ ] ] in practice , the visibility functions for the ionization histories usually considered are very poorly approximated by a gaussian .
still , the gaussian approximation provides a simple analytic description of how the anisotropy spectrum depends on the redshift of re - scattering , optical depth , and @xmath87 .
given the power spectrum @xmath93 and our fourier conventions , it is straightforward to write the temperature autocorrelation function in terms of the power spectrum : @xmath94 the correlation function can also be written in terms of the more commonly used cmb multipole moments @xmath95 as @xmath96 where @xmath97 are legendre polynomials . since @xmath98 as @xmath99 , it follows that @xmath100 this identification allows us to illustrate our results in the multipole moments that have become familiar and to compare our results with those of previous authors .
first , we compare with the results of ref.@xcite ( and therefore ref.@xcite ) .
to do so , we consider only the einstein - de sitter universe as they do . in an einstein - de sitter universe ,
( [ eq : vishniacresult ] ) becomes ( after changing the integration variable to @xmath101 ) , @xmath102 we then note that they approximate the power spectrum by evaluating @xmath88 at @xmath103 and taking it outside the integral in eq .
( [ eq : edsresult ] ) .
now we compare our result with eq .
( 6.15 ) in ref.@xcite and their @xmath104 in their eq . ( 6.18 ) .
if we note that their @xmath105 is equal to our @xmath106 and assume that @xmath107 in the integrand , then our results reproduce theirs ( although it appears that their result is twice ours ; we have not been able to locate the source of the discrepancy ) .
now we want to compare our results with @xcite .
although their derivation , unlike ours , relies on a solution to the boltzmann equation , our results should agree . specifically , our eq .
( [ eq : vishniacresult ] ) should agree with their eq .
( 24 ) . to check that this is so , we change our integration variable in eq .
( [ eq : vishniacresult ] ) to @xmath108 and find @xmath109 where @xmath110 ^ 2 \left ( { a \dot d d \over d_0 ^ 2 } \right)^2.\ ] ] if we put their eq .
( 24 ) into this form [ to do this , keep in mind that their overdot denotes differentiation with respect to @xmath111 instead of @xmath26 , so our @xmath112 is equal to their @xmath113also , they normalize @xmath114 such that @xmath115 , we find agreement with their results . the variance of the temperature distribution induced by the ostriker - vishniac effect is @xmath116 in the gaussian approximation for @xmath117 and in a flat universe , @xmath118 ^ 2 \ , \int \ , k \ , dk\ , s(k ) , \label{eq : gaussianvariance}\ ] ] and for an einstein - de sitter universe this simplifies to @xmath119
for the power spectrum , we use @xmath120 where @xmath121 is the usual cdm transfer function , @xmath122 is the physical wave number with our conventions , and @xmath123 is given more accurately in terms of @xmath124 and the baryon fraction @xmath59 by eqs . ( d-28 ) and ( e-12 ) in ref.@xcite .
the factor of 8 in the denominator in eq .
( [ eq : powerspectrum ] ) arises because we are using @xmath125 . for the transfer function ,
we use @xcite @xmath126^{1/4}}. \label{eq : transferfunction}\ ] ] for @xmath127 , we use the fits to the _ cobe _ anisotropy given in @xcite . for the tilted model ,
we use the _ cobe _ normalization from ref .
@xcite obtained including tensor modes with the amplitude predicted by power - law inflation .
( note that the tensor contribution in other inflationary models may be different . )
alternatively , the power spectrum may be normalized at small distance scales through the cluster abundance which fixes @xmath128 , the variance in the mass enclosed in spheres of radius @xmath129 mpc , to @xmath130 @xcite . in terms of the power spectrum , @xmath131 ^ 2,\ ] ] where @xmath132 mpc , and @xmath133 is a spherical bessel function . since we are using @xmath134 , @xmath135 ( rather than @xmath136 ) enters into the argument of the spherical bessel function
in this section we present numerical results for the anisotropy spectrum and rms temperature fluctuation for a variety of cdm models and ionization histories .
we consider a family of ionization histories parameterized by a constant ionization fraction @xmath137 after a redshift @xmath138 . in a cosmological - constant universe
, this induces an optical depth , @xmath139 in a flat universe , this becomes @xmath140 .
\label{eq : flattauequation}\ ] ] the prefactor 0.069 ( or 0.046 in the flat - universe expression ) is obtained if all the baryons are in the form of protons , an assumption often made when calculating ionization histories .
if one takes into account the fact that one quarter of the baryonic mass is helium , then the prefactor should be multiplied by 7/8 . [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ] [ tab : predictions ] table [ tab : predictions ]
lists the predicted rms temperature anisotropy @xmath141 due to the ostriker - vishniac effect for a variety of cdm models , which are parameterized by @xmath11 , @xmath12 , @xmath142 , @xmath2 , and @xmath143 , and ionization histories , which are parameterized by @xmath137 and @xmath138 .
the table also lists @xmath64 , @xmath128 , and @xmath144 .
we present a variety of models and ionization histories to illustrate the effects of variation of various parameters on the predictions .
it should be kept in mind that not all of the _
cobe_-normalized models satisfy the aforementioned observational constraint to @xmath128 from cluster abundances @xcite .
we therefore list in the table the predicted rms temperature anisotropy expected if the _ cobe _ normalization is disregarded and the amplitudes of the power spectrum is fixed instead to the cluster abundance .
similarly , the @xmath145 model is meant to be illustrative ; this value of the hubble constant is in disagreement with most measurements .
figure [ fig : clsplot ] shows the predicted anisotropy for the canonical standard - cdm model ( @xmath14 , @xmath15 , @xmath146 , @xmath3 , @xmath147 ) , for a variety of ionization histories , with the linear power spectrum , also shown , normalized to _
cobe_. we know from quasar absorption spectra that the universe has been significantly reionized at least since a redshift of @xmath148 .
therefore , we include predictions for this minimal level of anisotropy .
the effect of reionization on the primary anisotropies can be quantified primarily by the optical depth @xmath64@xcite ; any two ionization histories that have the same @xmath64 produce roughly the same primary anisotropies . however , as fig .
[ fig : clsplot ] and table [ tab : predictions ] show , the secondary anisotropies are more dependent on the ionization fraction than on the optical depth .
the growth of density perturbations leads to a growth of the peculiar velocities that induce the secondary anisotropy . therefore ,
if reionization takes place later , the anisotropy is significantly larger , even for the same optical depth .
furthermore , if reionization takes place later , the peak of the anisotropy spectrum , which is fixed in comoving distance scale by the peak of the power spectrum , subtends a larger angle . in open models ,
the peak of the power spectrum moves to larger physical scales , while the effects of geometery shift these physical scales back to smaller angular scales ; the result is that the spectrum peaks at only slightly lower @xmath149 than in a flat universe .
the weakness of the dependence of the anisotropy amplitude on @xmath64 indicates that a good fraction of the anisotropy is produced at _ later _ times when the amplitude of the power spectrum has grown .
the convergence of the predictions at small @xmath149 for ionization histories with the same @xmath137 indicates that , in particular , the anisotropy at larger scales comes primarily from lower redshifts @xcite , and therefore that assuming that re - scattering occurs primarily at the epoch of reionization@xcite does not provide a good approximation , especially at larger scales .
moreover , in open models the evolution of the growth factor differs from a flat universe by larger amounts at later times ; therefore the slope of @xmath95 at larger scales is greater than in a flat universe .
the growth of density perturbations induces an initially gaussian distribution of perturbations to become non - gaussian .
therefore , the distribution of anisotropies produced by the ostriker - vishniac effect will be non - gaussian even for gaussian initial conditions , and the higher @xmath2-pt temperature correlation functions may be calculated in perturbation theory @xcite .
the anisotropy is due primarily to density perturbations on distance scales closer to those probed by galaxy surveys than to those probed by _
cobe_. furthermore , the anisotropy is due to a large extent to scattering at later times .
therefore , when the power spectra are normalized to the cluster abundance , the spread in predicted values for the rms anisotropy is smaller than it is if the models are normalized to _
cobe_. since the anisotropy is a second - order effect , the rms anisotropy is proportional to @xmath150 . in a flat universe ,
models normalized to cluster abundances produce rms temperature anisotropies listed in table [ tab : predictions ] of @xmath151@xmath152 @xmath0k for a constant ionization fraction of unity , whereas an ionization fraction of 0.2 yields rms anisotropies of @xmath153@xmath151 @xmath0k .
our numerical results can also be easily scaled for a different value of the ionization fraction @xmath137 if the ionization epoch is held fixed and if @xmath154 .
the rms temperature anisotropy should be roughly proportional to @xmath137 if @xmath138 is held fixed .
also , if we neglect the effect of @xmath59 and @xmath142 on the power spectrum ( in particular , @xmath59 should have only a relatively weak effect on the power spectrum as long as @xmath155 ) , then eq .
( [ eq : tauequation ] ) shows that the rms anisotropy should be proportional to the combination @xmath156 , itself proportional to the column density of electrons to the last scattering surface .
this also suggests that the ostriker - vishniac anisotropy in models with a high baryon density should be larger . to illustrate ( and to illustrate the effects of geometry ) , we also include in table [ tab : predictions ] and in fig .
[ fig : clsplot ] predictions for an open model with @xmath157 , as suggested possibly by a peak in the observed power spectrum at 100 @xmath158 mpc @xcite . high - baryon - density models will be explored more carefully in ref .
here we provide simple approximate analytic fits to the vishniac power spectrum @xmath87 for standard - cdm power spectra . these can be used to quickly estimate the ostriker - vishniac anisotropy for cdm models with an arbitrary ionization history . with the standard - cdm power spectrum given in the form of eq .
( [ eq : powerspectrum ] ) , @xmath87 may be written , @xmath159 where @xmath160 for the standard - cdm transfer function , eq .
( [ eq : transferfunction ] ) , the functions @xmath161 may be approximated , for @xmath162 , by an asymmetric gaussian : @xmath163-\log_{10}[q_0(n ) ] \over \sigma(n ) } \right]^2 \right\ } , \label{eq : iapprox}\ ] ] with @xmath164=-4.716 - 0.636(n-1)+0.832(n-1)^2 $ ] , @xmath165 = 0.477 + 1.36(n-1)+0.975(n-1)^2 $ ] , and @xmath166 for @xmath167 , and @xmath168 for @xmath169 .
[ fig : approxclsplot ] shows the multipole moments obtained from numerical calculation of @xmath87 and the analytic approximation given here for a variety of models .
although this analytic approximation does not accurately reproduce the asymptotic large- and small-@xmath136 behavior of @xmath87 , it accurately approximates the peak of @xmath170 . as indicated by eq .
( [ eq : gaussianvariance ] ) , this should provide reliable estimates of the mean - square anisotropy .
although precise limits depend on imprecisely determined cosmological parameters , the existence of significant cmb anisotropy at the degree scale and current limits to arcminute - scale anisotropy and distortions to the cmb blackbody spectrum seem to indicate that @xmath171 .
however , some reionization is generically expected in cdm models .
again , the detailed ionization history will depend on the model and on details of star formation that are still difficult to quantify reliably .
still , optical depths @xmath172 are observationally consistent and expected in cdm models .
such a level of reionization will lead to secondary anisotropies at arcminute scales like those shown in section v. furthermore , the bulk of the reionization in cdm models generally occurs later , so the secondary anisotropy will be larger for a given @xmath64 than if reionization occurred earlier .
the predicted level of anisotropy in virtually all of the models we considered here is smaller than the current limit of about 50 @xmath0k at arcminute angular separations @xcite .
however , there has been much recent progress in the design of interferometer experiments dedicated to cmb measurements , and there should be much more in forthcoming years . to estimate the signal that might be detectable ,
we suppose that a detector with a sensitivity @xmath173 ( in @xmath0k @xmath174 ) maps a fraction @xmath175 of the sky with an angular resolution @xmath176 sufficient to detect these anisotropies for a time @xmath177 months .
such a detector should be sensitive to an rms temperature anisotropy as small as @xmath178 although the interferometry experiments that will likely reach these angular scales are not really parameterized by a sensitivity @xmath173 , the effective value of @xmath173 for these experiments is already better than 1000 @xmath0k @xmath179 .
furthermore , with a full power - spectrum analysis of the data , the sensitivity could be improved perhaps by an order of magnitude over this simple estimate .
comparing eq .
( [ eq : sensitivity ] ) for a reasonable @xmath177 with the predictions in section v , it seems quite plausible even though the experiments may be challenging that the ostriker - vishniac anisotropy expected in many leading cdm models will be detectable .
the results of ref.@xcite imply that the anisotropy generated by some of the more optimistic models considered here may already be detectable by upcoming experiments such as cbi , especially if binned over a wide range of @xmath149 .
observation of the ostriker - vishniac effect will be important for understanding the ionization history of the universe and for determining cosmological parameters with the cmb .
the linear - theory anisotropies depend primarily on the optical depth @xmath64 and only secondarily on the details of the ionization history ( e.g. , the values of @xmath138 and @xmath137 ) .
therefore , degree - scale anisotropies may determine @xmath64 , but they will not strongly constrain the epoch of reionization . when combined with sensitive measurements of the degree - scale anisotropy , measurement of the ostriker - vishniac effect will therefore provide the epoch of reionization .
the shape of the cmb power spectrum for these anisotropies is relatively featureless .
therefore , even detection of the anisotropy , without a precise determination of the shape of the power spectrum , will provide the most essential information on the ionization history .
of course , more precise constraints to the ionization history can be obtained if the power spectrum can be mapped and , some shape information along with multi - frequency observations will be necessary to distinguish the ostriker - vishniac anisotropy from foreground contamination such as that induced by extragalactic point sources .
determination of the geometry of the universe relies primarily on the first acoustic peak in the primary cmb spectrum @xcite
. however , precise and reliable determination of other cosmological parameters , such as the baryon density , hubble constant , cosmological constant , and spectral index , depends on the structure of the smaller - scale acoustic peaks .
the details of the ionization history ( e.g. , the epoch of reionization , which we have explored here ) will , in fact , have some effect on these higher peaks .
this will in turn affect the reconstruction of some cosmological parameters , although this uncertainty has not yet been taken into account in the parameter - estimation analyses of refs.@xcite .
therefore , the information on the epoch of reionization provided by mapping the secondary anisotropies will be important for precise determination of these other cosmological parameters with map and planck .
it seems unlikely that the current firas constraints to compton-@xmath180 distortions of the cmb blackbody spectrum will be improved significantly .
therefore , the ostriker - vishniac effect seems to be the most promising probe of the ionization history .
in addition to its significance for cosmological - parameter determination , the ostriker - vishniac effect can provide a window to the epoch of the earliest collapsed objects that is inaccessible with any other observations .
we thank m. white and w. hu for useful comments . this work was supported at columbia by d.o.e .
contract defg02 - 92-er 40699 , nasa nag5 - 3091 , and the alfred p. sloan foundation and at berkeley by nag5 - 6552 .
mk would like to acknowledge the hospitality of the center for particle astrophysics where part of this work was completed .
if an isotropic random field @xmath181 is decomposed into fourier components , @xmath182 then the fourier components may be written as the sum of divergence - free and curl - free parts parts : @xmath183 with @xmath184 and @xmath185 ; in particular , @xmath186 .
suppose we observe a projection @xmath187 of @xmath188 , the component of @xmath181 along the line of sight , with a visibility function @xmath117 , and we define the fourier components , @xmath189 the contribution of the parallel components @xmath190 to @xmath191 will be small because crests and troughs of each fourier component projected along the line of sight will tend to cancel .
the purpose of this appendix is to show that if the perpendicular components satisfy @xmath192 , then the fourier components of the projection satisfy @xmath193 with @xmath194 to do so , we follow the steps in the appendix of ref .
@xcite that lead to the fourier - space analog of limber s equation . rather than reconstruct the entire argument
, we refer the reader to kaiser s paper @xcite for more details .
the components of an isotropic divergence - free field satisfy @xcite , @xmath195 if the contribution of a narrow shell of width @xmath91 centered at @xmath196 is @xmath197 then with eq .
( [ eq : moninequation ] ) , we can show that its fourier coefficients @xmath198 satisfy @xmath199 with @xmath200 [ j_0(k_3 \delta w/2)]^2
p_\perp \left ( \sqrt { { \kappa^2 \over w_0 ^ 2 } + k_3 ^ 2 } \right ) , \label{eq : kaiserlimbertwo}\ ] ] and @xmath201 . as argued by kaiser , only modes with @xmath202 ( those very nearly perpendicular to the line of sight ) will contribute appreciably to the integral , so @xmath203 .
we then add the contributions @xmath204 along the line of sight to obtain eq .
( [ eq : kaiserlimber ] ) . in an open or closed universe
one replaces @xmath205 in the first argument of @xmath206 in eq .
( [ eq : kaiserlimber ] ) , and in the two factors of @xmath50 that appear in the denominator , where @xmath207 in an open universe and @xmath208 in a closed universe @xcite . note that this derivation also explains why the contribution from @xmath209 is suppressed .
this term satisfies @xcite @xmath210 this would therefore have a contribution to the projected power spectrum of @xmath211 proportional to @xmath212 .
the approximation used here is valid for small angular separations , @xmath213 .
since the ostriker - vishniac effect produces anisotropies at multipole moments @xmath214 , this flat - sky approximation should lead to errors no greater than @xmath215 . at large angular scales
, we can also expect the anisotropy from the parallel term to be significant@xcite .
also , the visibility function @xmath117 should be smoothly varying over wavelengths at which there is significant power .
again , the visibility function for re - scattered cmb photons will be very broad , so this approximation should be fine for our calculation .
g. jungman , m. kamionkowski , a. kosowsky , and d. n. spergel , phys .
d 54 , 1332 ( 1996 ) ; j. r. bond , g. efstathiou , and m. tegmark , mnras , to appear , astro - ph/9702100 ; m. zaldarriaga , d. n. spergel , and u. seljak , astrophys .
j. , to appear , astro - ph/9702157 .
|
we present a new derivation of the cosmic microwave background anisotropy spectrum from the ostriker - vishniac effect for an open , flat , or closed universe , and calculate the anisotropy expected in cold dark - matter ( cdm ) models .
we provide simple semi - analytic fitting formulas for the vishniac power spectrum that can be used to evaluate the expected anisotropy in cdm models for any arbitrary ionization history . in a flat universe ,
cdm models normalized to cluster abundances produce rms temperature anisotropies of 0.82.4 @xmath0k on arcminute angular scales for a constant ionization fraction of unity , whereas an ionization fraction of 0.2 yields rms anisotropies of 0.30.8 @xmath0k . in an open and/or high - baryon - density universe ,
the level of anisotropy is somewhat higher .
the signal in some of these models may be detectable with planned interferometry experiments .
the damping of the acoustic peaks in the primary - anisotropy spectrum at degree angular scales depends primarily on the optical depth and only secondarily on the epoch of reionization . on the other hand ,
the amplitude of ostriker - vishniac anisotropies depends sensitively on the epoch of reionization .
therefore , when combined with the estimate of the reionization optical depth provided by maps of degree - scale anisotropies , the ostriker - vishniac effect can provide a unique probe of the epoch of reionization .
# 1#1 # 1
|
in this paper i discuss the importance of the upcoming _ ` space infrared telescope for cosmology & astrophysics ' _ ( hereafter , _ spica _ ) for the study of the icy small bodies of the outer solar system . here , i focus on the study of kuiper belt objects ( kbos ) but the same ideas can be applied to any atmosphereless bodies .
i will mainly discuss how _ spica _ can help us to measure the sizes of kbos , and how that leads to more accurate estimates of the size distribution and total mass of kbos .
i will also mention how this new infrared space telescope might probe the rotational properties , chemical composition and thermophysical parameters of those icy bodies .
other uses of a space - based infrared telescope for solar system studies are detailed elsewhere , in the context of _ spica _
s precursor observatory , _ herschel _ @xcite . in 2 to 4
i begin by summarising what we know about kbos and why their study is interesting and important .
although many accounts exist elsewhere i believe these proceedings should include a broad overview of the subject . in 5 and 6
i discuss how _ spica _ may contribute to the study of kbos .
just a couple of decades ago our understanding of the solar system was quite different from what it is today .
scientific interest lay mainly with the nine planets : six ( mercury to saturn ) already known to the greeks , another two ( uranus and neptune ) discovered in the 18th and 19th centuries and a very peculiar ninth planet ( pluto ) discovered in 1930 .
the small rocky planets , all within 1.5 astronomical units ( au ) from the sun , stood in sharp contrast with the outer gaseous giants extending out to 30 au .
more intriguing was pluto , a moon - sized world on an eccentric and very inclined orbit beyond neptune .
pluto did not seem to fit in with the rest and was oddly isolated given its small size .
comets , asteroids , and even some planetary moons had a mysterious character to them and were for the most part not well understood .
this view of the solar system has changed dramatically in the last 20 years or so , mainly following the discovery of the kuiper belt and of discs and planetary systems around other stars .
the focus of planetary science has moved to the smaller bodies of the solar system as most of the interesting results and paradigm - shifting discoveries have come from the study of their properties .
we begin to understand how the planets formed and evolved , how different families of small bodies relate to one another , and how our solar system fits in the larger picture of what we are finding in extrasolar planetary systems .
the kuiper belt was identified in 1992 with the discovery of 1992 qb@xmath1 @xcite . since
then more than 1000 kbos have been discovered in the region roughly from 30 to 50 au .
the _ known _ kbos range from about 25 to 2500 km in diameter @xcite but numerous smaller objects are believed to exist down to the micrometer - sized dust grains that have been detected by voyager 1 and 2 @xcite .
larger bodies could also exist and remain undetected .
the kuiper belt is the solar system analogue to the debris discs found around other stars .
kuiper belt objects provide us with probably the best picture of what the planetesimals that formed the planets might have looked like .
dynamically , most kbo orbits are stable against gravitational perturbations from the giant planets on gyr timescales .
chemically , the low temperatures found at such large heliocentric distance ensure that kbos preserve significant volatile content from the protosolar nebula .
kbos display the largest spread in surface colours of all objects in the solar system @xcite .
the origin of the diversity is unknown but could reflect dynamical mixing resulting from planetary migration @xcite .
physically , as the largest ( @xmath2 km ) remnants of the planetesimals that formed the planets , kbos retain valuable information about the size , density and angular momentum distributions at the poorly understood epoch of accretion .
the study of kbos can thus provide unique clues about the formation of solar system bodies . .
the thick black lines indicate the orbits of jupiter , saturn , uranus and neptune .
centaurs are not shown for clarity.,width=8 ] kuiper belt orbits form a thick disc ( 7% of known kbos have inclinations @xmath3 ) beyond neptune ( fig .
[ xxx20091029b_fig : fig1 ] ) but the orbital distribution within that disc is not random .
most orbits fall in one of four dynamical classes ( fig .
[ xxx20091029b_fig : fig2 ] ) . _ resonant kbos _ lie in mean motion resonances with neptune .
because of their resonant character they avoid close encounters with the massive planet and are stable over the age of the solar system .
pluto lies in the 3:2 resonance at roughly 39 au from the sun and so other kbos in the same resonance are often called plutinos .
_ classical kbos _ are the dynamically quintessential objects .
they have relatively low eccentricity and low inclination orbits between the 3:2 and the 2:1 resonances at @xmath4 au and @xmath5 au .
they are called classical because their orbits most closely match what would be expected for a cold dynamically stable disc .
the classical belt turned out to have a broader inclination distribution than first expected .
1992 qb@xmath1 , the first object discovered in the kuiper belt , is a classical kbo .
_ scattered kbos _ are objects that interact strongly with neptune near perihelion and are thus being scattered by the giant planet .
it is generally believed that scattered kbos originate in slightly unstable regions within the resonants and classicals .
once in the scattered population these kbos can be thrown into planet crossing , _
centaur_-type orbits which eventually feed the jupiter family comet population @xcite .
this scenario is not unique , though @xcite .
centaurs are a population closely related to kbos .
they have giant - planet - crossing orbits unstable on myr timescales .
possible end - states for centaurs are collision with a giant planet , or ejection from the solar system .
some return to the scattered kuiper belt and a few end up ( or spend part of their lifetime ) as jupiter family comets @xcite .
_ detached kbos _ have perihelia beyond 40 au indicating that neptune had little influence in their orbital evolution .
they may have been emplaced at an earlier epoch by the pull of a star passing close to the sun or by past gravitational perturbations by an unseen ( or since ejected ) planet beyond @xcite .
lccrrcccccr + object & d [ km ] & @xmath6 & @xmath7 [ hr ] & @xmath8 [ mag ] & @xmath9 [ kg m@xmath10 & albedo & surface comp . &
@xmath11 [ au ] & @xmath12 & @xmath13 [ @xmath14 + + eris & 2400 & 1.0 & ? & @xmath15 & 2300 & 0.85 & ch@xmath16 & 67.8 & 0.44 & 44.0 + pluto & 2290 & 1.0 & 153.2 & 0.33 & 2000 & 0.60 & ch@xmath16 , co , n@xmath17 & 39.6 & 0.24 & 17.1 + haumea & 1500 & 1.2 & 3.9 & 0.29 & 2600 & 0.70 & h@xmath17o & 43.2 & 0.19 & 28.2 + quaoar & 1250 & 1.1 & 17.7 & 0.13 & ?
& 0.09 & h@xmath17o & 43.1 & 0.04 & 8.0 + varuna & 1000 & 1.3 & 6.3 & 0.42 & 1000 & 0.07 & h@xmath17o ? & 42.8 & 0.06 & 17.2 + huya & 530 & 1.0 & ? & @xmath18 & ? & 0.06 & h@xmath17o ? & 39.8 & 0.28 & 15.5 + 2000 gn@xmath19 & 320 & 1.6 & 8.3 & 0.61 & 600 & 0.06 & ? & 39.7 & 0.29 & 10.8 + 2001 qg@xmath20 & 230 & 2.9 & 13.8 & 1.14 & 600 & ? & ? & 39.6 & 0.20 & 6.5 + + the orbital architecture of the kuiper belt has provided many clues about the past dynamical evolution of the solar system .
for instance , we now know that the giant planets did not form in their current locations but instead started out in a much more compact configuration , from roughly 5 to 15 au and then migrated to their current positions . the stable resonances in the kuiper belt were populated as neptune migrated outwards into a cold disc of planetesimals by sweeping bodies as they moved along with the planet . indeed , the most encompassing model of the dynamical evolution of the solar system the _ nice _ model ( named after the french city ; * ? ? ?
* ) is largely designed to fit the observed properties of kbos and small solar system bodies in general .
it fits , among other things , the global architecture of solar system orbits , the migration of the planets by interaction with the planetesimals , the enhanced lunar cratering record @xmath03.8 gyr ago known as late heavy bombardment , the formation of the oort cloud , the origin of trojan asteroids and the existence of hot ( high-@xmath13 ) and cold ( low-@xmath13 ) classical kbo populations .
the strength and weakness of the _ nice _ model lies in its adaptability to new observational discoveries .
new and more stringent observational constraints should be sought to truly test the model .
the cumulative luminosity function of kbos is well described by a power law , @xmath21 , with @xmath22 and @xmath23 @xcite . to translate luminosity into size
requires knowing the surface albedo . assuming a uniform albedo for all kbos and no relation between heliocentric distance and size
, the slope of the size distribution can be inferred from that of the luminosity function as @xmath24 @xcite .
however , the few known albedos ( see examples in table [ xxx20091029b_tab : tab1 ] ) show that albedo is probably a strong function of size and surface properties . measuring and understanding the albedo distribution is crucial if we are to constrain the size distribution and total mass of the kuiper belt ; _ spica _ is expected to play a major role in achieving this goal .
kuiper belt objects exhibit a tremendous diversity in surface colours , unparalleled in the solar system @xcite .
this probably reflects significant chemical diversity , which is a puzzle given the small range of temperatures in the 30 to 50 au region .
spectroscopy , only possible for the few brightest objects , shows that the largest kbos are predominantly coated in either methane ice ( e.g. eris , pluto ) or water ice ( e.g. haumea , quaoar ) .
smaller objects are generally too faint for spectroscopic studies but the few observed at sufficient s / n show mostly featureless spectra . in summary , the kuiper belt is now understood as a significant component of the sun s debris disc , as the most likely source of centaurs and jupiter family comets , and possibly even of the trojans , unusual planetary moons ( e.g. triton , phoebe ) and the irregular satellites of the giant planets .
the belt has also provided a context for global models of the evolution of the solar system .
_ 1992 qb@xmath1 _ was the first identified kbo .
it is roughly 250 km in diameter assuming a cometary albedo of 4% .
its orbit is nearly circular and has low inclination .
_ varuna _ was discovered on 28 nov .
2000 by r. s. mcmillan .
being one of the brightest known kbos at the time , varuna was intensely observed and studied . combined sub - millimetre and optical observations were used to solve the degeneracy between size and albedo @xcite and estimate varuna s diameter ( @xmath25 km ) and albedo ( @xmath26 . light curve observations revealed a rotationally deformed , fast - spinning object ( @xmath27 hr ; * ? ? ?
varuna is too large to support significant topography and its overall shape is set by the balance between gravitational and rotational accelerations .
this property allows its bulk density to be estimated ( @xmath28 kg m@xmath29 ; * ? ? ?
1998 ww@xmath30 _ was the first binary kbo to be discovered after pluto / charon @xcite .
like most kbo binaries , this system has nearly equal sized components .
binaries are important because their total mass can be estimated using kepler s 3rd law and , if the size of the components is known , their densities can be estimated . _
2001 qg@xmath20 _ was found to display extremely large photometric variability , @xmath31 mag @xcite .
the large @xmath8 combined with a relatively slow rotation , @xmath32 hr , suggest this object is a contact binary .
the shape of the components , as inferred from the light curve , and the spin period imply a low bulk density of about @xmath33 kg m@xmath29 @xcite .
statistically , about 20% of kbos could display the extreme properties of qg@xmath20 , meaning that many more await discovery .
the prospect of measuring densities from contact - binary - type light curves is compelling . _
eris _ was the first truly pluto - sized kbo found @xcite ; those are the only kbos than can currently be resolved by _
the possibility that eris is larger than pluto intensified the planethood controversy .
eris is covered in methane ice and has a high albedo ( @xmath34 ) surface .
its density is about 2300 kg m@xmath29 , comparable to that of pluto . _ haumea _ is one of the strangest known kbos .
it spins extremely rapidly , @xmath35 hr and , like varuna , it is rotationally distorted into a triaxial shape @xcite . unlike pluto and eris
, haumea is covered in almost pure water ice but its high bulk density ( @xmath36 kg m@xmath29 , estimated in the same way as for varuna ; * ? ? ?
* ) indicates that it must have a rocky core and thus be differentiated .
a violent collision has been proposed to explain haumea s fast rotation , the fact that it has two small , water - ice - rich moonlets , and the presence of half a dozen water - ice - rich kbos in its orbital vicinity .
the collision would have happened more than 1 gyr ago , onto a proto - haumea that was differentiated and had a thick water ice mantle @xcite.=-1
without knowing the surface albedo of a kbo it is not possible to tell large and dark from small and bright objects . ideally , the albedo can be measured by combining visible and thermal infrared observations .
the reflected , visible light is proportional to the kbo cross - section , @xmath37 , and albedo , @xmath38 , while the thermal emission is proportional to @xmath39 , i.e. to the fraction of light absorbed by the object which contributes to heating its surface .
visible and thermal observations can thus be used to solve for @xmath37 and @xmath38 . in practice ,
other unknown parameters describing the spin state and orientation of the kbo , and its surface roughness , emissivity , and thermal diffusivity ( or inertia ) complicate the process and generally require more detailed observations at different wavelengths .
m , is indicated by a horizontal dotted line.,width=8 ] due to their great heliocentric distances , kbos have very cold surfaces with equilibrium temperatures around 40 to 50 k. consequently , their thermal emission peaks at 50 to 80 @xmath40 m , well within _ spica _ s wavelength coverage ( fig .
[ xxx20091029b_fig : fig3 ] ) .
but by far the most important feature of _ spica _ will be its enhanced sensitivity .
active cooling of the mirror should bring the sensitivity down to @xmath050 @xmath40jy , nearly two orders of magnitude better than _ herschel_. _
spica _ will be able to detect objects almost as small as 10 km across , one order of magnitude smaller than _ herschel _ , meaning that thousands more bodies will be accessible to the new telescope .
_ herschel _ will carry out a key project to observe @xmath0140 large kbos and centaurs .
_ spica _ should vastly increase this number and extend the sample to smaller bodies , approaching the sizes of comet nuclei which could then be studied in their pristine state before they reach the inner solar system.=-1 .,width=8 ] another advantage of looking at the smaller bodies is to understand how albedo varies with size .
understanding the albedo distribution of kbos is important in many ways .
for instance , estimating the size distribution and total mass of the kuiper belt from the observable luminosity function relies strongly on assumptions about the albedo .
when derived from its brightness , the diameter of a kbo varies as @xmath41 , where @xmath38 is the albedo , and its mass varies as @xmath42 .
the mass estimate has the further complication that the densities are also unknown .
plausible solar system densities range from @xmath43 kg m@xmath29 for comet nuclei , to @xmath44 kg m@xmath29 for terrestrial planets .
figure [ xxx20091029b_fig : fig4 ] illustrates how uncertainties in albedo and density can lead to differences of a few orders of magnitude in the derived masses .
the current estimated total mass of kbos is 0.01 to 0.1 @xmath45 .
accretion models in the outer solar system require 10 @xmath45 of material in the proto - kuiper belt to explain the formation of pluto - sized objects @xcite in reasonable 10 to 100 myr timescales , and the _ nice _ model assumes 35 @xmath45 to explain the current orbital architecture of the giant planets @xcite .
nice _ model can also explain how @xmath099% of the initial mass has been lost but the uncertainty in the current mass prevents it from offering a solid constraint to the model . for binary kbos it becomes even more interesting to be able to measure their physical size ( cross - section ) .
binaries offer the opportunity of measuring mass which can be translated into bulk density if we know the size .
density is hard to measure remotely but is very useful as first indicator of inner structure and composition .
exactly how density depends on size may reveal whether the former is more strongly influenced by composition or porosity . by the time _
spica _ begins operations , the next - generation space telescope _ jwst _
will presumably have identified many more binary kbos suitable for mass determination .
accurate albedos are also important for spectral modelling
. models by hapke and shkuratov to investigate the composition and relative abundances of surface materials can only be applied if the absolute reflectance is known .
_ spica _ s increased sensitivity will , for the first time , open the far - infrared domain to spectroscopic studies of kbo surfaces .
the far - infrared is rich in diagnostic features diagnostic of silicates and ices expected to incorporate kbos .
amorphous and crystalline water ice can also be identified through far - infrared spectral features at 44 , 45 and 62 @xmath40 m @xcite . at low temperatures ice
will form in the amorphous state and at 40 to 50 k it should remain so for the age of the solar system . however , whenever identified on the surfaces of the largest kbos water ice appears consistently crystalline @xcite .
amorphous ice could exist in subsurface layers though , or in smaller kbos , as it is thought to drive cometary activity of centaurs beyond 5 au by converting into the crystalline phase as these objects move in from the cold kuiper belt @xcite .
the thermal emission of kbos depends not only on their size and albedo , but also on thermophysical properties of the surface such as emissivity and thermal conductivity ( or inertia ) , and on surface roughness .
multi - wavelength measurements can be used to constrain these unknown parameters and provide extra information about the surface of the distant kbos . however , to accurately model the temperature distribution across the surface of the kbo we need to know its spin state ( rotation period and spin orientation ; see below ) .
but for an elongated object with a uniform surface . here , the brightness modulation in the reflected light ( dashed grey ) is due to the varying apparent cross - section as the object rotates .
the thermal light curve ( points with error bars and solid line ) follows the same pattern because the emitting cross - section is larger when the reflecting cross - section is larger.,width=8 ] six of the brightest kbos to be observed by _
herschel _ will be observed repeatedly as they spin to obtain thermal emission curves . given _
spica _ s high sensitivity we will be able to extend this type of observation to many more objects . by comparing the optical and thermal light curves we will be able to establish with certainty whether the photometric variability is due to shape or albedo spottiness .
if the variability is due to albedo markings then optically bright , high albedo regions will be fainter at thermal wavelengths because they reflect more sunlight and remain cooler the optical and thermal light curves appear uncorrelated ( see fig .
[ xxx20091029b_fig : fig5 ] ) . if due to shape
, the body will be brighter at optical and thermal wavelengths at roughly the same time ( fig .
[ xxx20091029b_fig : fig6 ] ) . breaking this degeneracy between albedo and shape is important for studies of the shape distribution and angular momentum content of kbos @xcite .
time - resolved thermal observations may also constrain the orientation of kbo spin axes w.r.t . the line of sight .
the closer the spin vector lies to the line of sight the warmer the object will be , on average , as a large fraction of its surface close to the pole is continuously exposed to sunlight .
modulations due to shape will also be smaller than if the object is seen equator - on .
_ spica _ will extend these techniques to a much larger sample and for those six objects previously studies by _ herschel _ the observations about a decade apart may help constrain the longitude of the spin pole as well .
the alignment of kbo spin poles can test models of planetesimal formation and collisional evolution @xcite .
the enhanced sensitivity of the upcoming _ ` space infrared telescope for cosmology & astrophysics ' _ ( _ spica _ ) will play a key role in the study of kuiper belt objects .
multi - wavelength thermal observations will enable us to measure the albedos and sizes of several hundreds of bodies and to investigate their surface composition and thermophysical properties . the high sensitivity two orders of magnitude better than _ herschel _ will allow the detection of kbos as small as a few tens of kilometres across , very close to the sizes of cometary nuclei .
sizes measurements of binary kuiper belt objects will lead to estimates of their bulk density , a first step into the interior structure of these bodies .
time resolved thermal observations of a statistical ensemble of kbos will help break the degeneracy between albedo variability and shape as the cause for observed light curves , which is important for assessing the shape and angular momentum distributions of kuiper belt objects . by extending the far - infrared domain to potentially hundreds of kuiper belt objects and associated families , _ spica
_ should allow us to identify meaningful correlations and patterns between orbital , physical and chemical properties and thus help us probe the formation and evolutionary processes that operated at the early stages of our solar system .
|
one of the most fundamental problems in the study of kuiper belt objects ( kbos ) is to know their true physical size . without knowledge of their albedos
we are not able to distinguish large and dark from small and bright kbos . _
spitzer _ produced rough estimates of the sizes and albedos of about 20 kbos , and the _ herschel _ space telescope will improve on those initial measurements by extending the sample to the @xmath0150 brightest kbos .
_ spica _ s higher sensitivity instruments should allow us not only to broaden the sample to smaller kbos but also to achieve a statistically significant sample of kbo thermal light curves ( _ herschel _ will measure only six objects ) .
a large sample covering a broad range of sizes will be key to identify meaningful correlations between size and other physical and surface properties that constrain the processes of formation and evolution of the solar system .
|
the use of quantum mechanical systems for information processing and computation has been studied extensively over the last decades .
models have been well established for using the principles of quantum mechanics to achieve computational advantages in the form of a speedup over classical methods , and to allow efficient simulation of physical systems . however , a central challenge on the way to fulfilling these goals continues to lie in the realisation of such systems - how to construct a large - scale system , which entails the required quantum properties and at the same time allows efficient manipulations necessary for performing these tasks .
some of the main techniques for the realisation problem feature the use of a _ double - well potential _ system to store and manipulate the qubit .
realisations of a double - well - based qubit have been demonstrated using superconducting circuits ( _ squid _ ) , which provide efficient control over the qubit transformations and feasible integration with electronic circuitry @xcite . as an alternative to the qubit
, work has been done to study the analogue 3-state register , the _ qutrit _ , and more generally the @xmath0-state _ qudit_. several advantages of using qutrits and qudits rather than qubits have been discussed , e.g. in the context of improved fault - tolerance @xcite and advantages in cryptography @xcite . motivated by these results we suggest a novel _ periodic triple - well qutrit _ system . on the one hand
, it is based on the known control and manipulation principles of the double - well qubit , while on the other hand it presents special symmetries beneficial to the higher dimensional analysis of its evolution and control using the fundamental representation of @xmath2 .
we further opt for a generalization of these concepts to a general dimension periodic @xmath0-well qudit , with a central role played by the respective representation of @xmath1 .
the paper is organized as follows : sec .
1 reviews the dynamics of a particle in a double - well potential while emphasizing its close relations with a spin-1/2 system .
we mostly discuss in this introductory section the methods for control and manipulation of the qubit state with the underlying goal of efficient execution of general transformations .
this simple analysis will be hopefully useful in the rest of the paper which manifests its main contribution .
2 proposes the extension to a periodic triple - well qutrit system , discussing its dynamics and manipulation with regard to a set of transformations in @xmath2 .
3 aims to further extend the above ideas in a system of @xmath3 wells , and in particular the corresponding evolution of a qudit in the scope of @xmath1 .
two different topologies are considered for the problem : ( i ) a fully - connected well system featuring a highly symmetrical solution ; ( ii ) a cyclic linearly - connected well chain , featuring a bloch - state based solution , and essentially of higher feasibility .
the double - well potential system is a widely considered tool for producing a 2-level system , due to its energy separation featuring a pair of lower states well - separated from the rest of the spectrum : @xmath4 eq .
[ energies_low_states ] can be derived by solving the schrdinger equation for the specific double - well potential form .
a solution for the simple square - well case is summarized in _ appendix a _ and further detailed in @xcite .
+ the energy gap @xmath5 is related to the tunneling amplitude , denoted by @xmath6 . as denoted for example in @xcite
, this tunneling amplitude can be related to the wavefunction using the wkb approximation @xmath7 where @xmath8 is the localized wavefunction in the right or left well .
the solution can generally be written as @xcite @xmath9 for a barrier of width @xmath10 around the origin . ] .
for example , in the square potential ( see _ appendix a _ ) we obtain for @xmath6 @xcite , @xmath11 the proportion between @xmath6 and @xmath5 can be simply found in a two - state system analysis , performed in the next section . solving the time - dependent schrdinger equation
, it is apparent that the tunneling implies oscillatory dynamics of the 2-level state .
shall we define @xmath12 and @xmath13 as projections on the left and right wells , and take initial conditions of @xmath14 for example , the right - well probability amplitude @xmath15 then oscillates : @xmath16 with a frequency @xmath17 in general we use the following criteria to define a double - well system , in either one or higher dimensions : 1 .
parity symmetry : the two wells are identical under a discrete symmetry denoted by @xmath18 , e.g. reflection about @xmath19 for a 1d well @xmath20 or about a center plane in 3d , @xmath21 .. ] the commutation relation is therefore satisfied : @xmath22 = 0.\ ] ] 2 . the coupling between the two wells is weak enough , i.e. the barrier is high enough , so that eq .
[ energies_low_states ] applies .
these criteria suffice to produce a symmetric and anti - symmetric pair of lower states , @xmath23 and @xmath24 , which serve as a 2-level system in the low - energy regime ( with amplitudes for higher levels negligible , @xmath25 ; @xmath26 ) . setting a reference basis of _ right and left projections _
, we may write a general two - state symmetric hamiltonian for the system : @xmath27 with @xmath6 the tunneling amplitude ( eq . [ tunneling amplitude eq ] ) , and the basis vectors @xmath28 are localized states in the right / left well , namely @xmath29 .
+ the hamiltonian s invariance to a @xmath30 swap manifests the parity symmetry defined in eq .
[ parity_commutation ] . solving for the energy eigenstates relates the tunneling amplitude with the energy gap : @xmath31 thus , remaining in the low energy regime with significant amplitudes held only for the states @xmath32 , we may treat the system as equivalent to a spin-1/2 particle , under the generating group @xmath33 with @xmath34 being the pauli matrices : @xmath35 the eigenvectors of the simplified hamiltonian @xmath36 are the @xmath37 eigenstates : @xmath38 satisfying the @xmath18 symmetry , the energy states @xmath32 are _ symmetric _ and _ antisymmetric _ , respectively , in context of the @xmath18 operator .
+ mapping the @xmath39 and @xmath37 operators to location and energy , respectively , the third generator @xmath40 remains to be addressed .
we shall demonstrate that the corresponding physical quantity for this spin operator is the _ probability current _ between the two wells .
+ the positive / negative probability current values account for the change of total density from the left to the right well and vice versa , @xmath41 in our spin states notation , the flow operator hence acts as @xmath42 in the heisenberg picture . using our previous identifications of @xmath43 and @xmath44
we derive in the heisenberg picture : @xmath45 = -i\frac{\nu}{\hbar}[\sigma_x,\sigma_z ] = i\frac{\nu}{\hbar}\sigma_y \end{array}.\ ] ] hence , the probability current between wells is proportional to the @xmath40 in the 2-level notation , with the coefficient @xmath46 corresponding to the maximal current magnitude .
+ the current eigenstates are therefore the @xmath47 vectors : @xmath48 it is straightforward to show that these states also produce maximal ( positive ) and minimal ( negative ) values of probability current at @xmath19 : @xmath49 where @xmath50 , while using the continuity equation in the rightmost equality . the relation of energy , current and position to the qubit operators @xmath37 , @xmath40 , @xmath39 provides physical meaning to the different qubit transformations , explicitly connecting each one with a corresponding observable . with the hamiltonian @xmath36 , the time evolution of
the double - well qubit corresponds to precession about the @xmath51-axis , equivalent to a spin-1/2 particle in a magnetic field @xmath52 . in general , the evolution of a state @xmath53 is : @xmath54 = e^{-\frac{i}{\hbar}e_0t}(\alpha{\mbox{$|0\protect\rangle$ } } + e^{-i\omega t}\beta{\mbox{$|1\protect\rangle$}}).\\ \end{array}\ ] ] hence the evolution of a qubit state is periodic : @xmath55 with a global phase @xmath56 , and a period @xmath57 determined by the rabi frequency : @xmath58 if we intend to perform a transformation on the double - well qubit , let us first ask what set of qubit transformations would be sufficient . in order to carry out a general unitary transformation , one would alter the hamiltonian to a modified form @xmath59 for a specific time period , generating a precession about an axis defined by the @xmath60 direction ( using the notation of a spin-1/2 particle in a magnetic field ) , its magnitude reflecting the precession frequency .
the original system hamiltonian ( eq . [ qubit_hamiltonian ] ) refers to the case @xmath61 . as we note later on , to perform a general qubit transformation it is sufficient to control only two of the field components , for instance : @xmath62 the choice of @xmath51 and @xmath63 is not incidental ; as we show in the following , this form can be realized by inducing asymmetry in the well system .
we examine the effect of the asymmetry on the eigenstates and energy eigenvalues . using the notation of rastelli
@xcite , the asymmetry is quantified by a dimensionless parameter @xmath64 , i.e. @xmath65 in the symmetric case and @xmath66 otherwise .
our two basis vectors remain the localized states @xmath67 with _ local _ energies @xmath68 , respectively , found by solving the schrdinger equation locally in the left and right wells .. the difference between the energies @xmath69 is related to the difference between the well depths .
( denoting by @xmath70,@xmath71 the @xmath72 , @xmath73 sides , respectively , and taking the local minima as benchmarks ) . ] as derived by rastelli et al . in @xcite
, the value of @xmath74 in the presence of asymmetry can be derived using the wkb approximation to be @xmath75 where @xmath76,\ ] ] and @xmath77 are obtained by two _ symmetric _ double - well potentials @xmath78 and @xmath79 , defined by a symmetric reflection of the left or right side of the asymmetric well , as demonstrated in fig . [ asymmetric_well_reflections ] .
.*,height=188 ] using the basis states @xmath80 in the regime of eq .
[ energies_low_states ] , the ground state and first excited state are + @xmath81 + with the angle @xmath82 related to the hamiltonian parameters by : @xmath83 to what degree of asymmetry is this result valid ? if the potential is tilted to a certain level , the local ground state of one well begins to interact with the excited state of the other , lower well .
thus , our results hold as long as the energy difference @xmath84 is not comparable to the one - side energy gap between ground and first excited state , @xmath85 where @xmath86 stands for the side of low potential . + moving on to examine the effects in the 2-level system , the modified hamiltonian depends on the degree of asymmetry @xmath64 , and reflects the different energies of the @xmath87 states with a tunneling amplitude @xmath74 : @xmath88 removing a @xmath89 factor as a physically irrelevant global phase , we are left with the hamiltonian @xmath90 where @xmath91 .
the implication of this asymmetry on the direction of @xmath60 in the hamiltonian ( eq . [ hamiltonian as magnetic field ] ) , is simply setting @xmath60 in an angle @xmath82 ( eq . [ angle and params relation ] ) from the @xmath92 direction in the @xmath93 plane . in the case of symmetric energies ,
@xmath94 , the resulting angle is @xmath95 and @xmath60 is aligned with @xmath37 , resorting to the symmetric and antisymmetric states of the symmetric double - well in eq . [ two symmetric well states ] .
the two - state eigenvectors of the asymmetric system are thus @xmath96 \cos\frac{\theta}{2 } \end{pmatrix } , \quad { \mbox{$|1(\theta)\protect\rangle$ } } = \begin{pmatrix } \cos\frac{\theta}{2 } \\[1 mm ] -\sin\frac{\theta}{2 } \end{pmatrix}. \end{array}\ ] ] the two energy eigenvalues are straightforwardly found from eq .
[ perturbed hamiltonian 2 ] to be @xmath97 reducing to eq .
[ symmetric eigen energies ] for @xmath98 .
the asymmetric eigenstates @xmath99 hence imply a modified dynamics of the qubit state .
the state now ` oscillates ' around @xmath100 , rather than @xmath37 . the rabi frequency @xmath101 is now increased relative to the symmetric case of eq .
[ symmetric eigen energies ] .
an instructive example for a double - well implementation , extensively studied and tested , is via the use of superconductor devices , based on the josephson effect @xcite , and specifically , the rf - squid circuit @xcite . a full analysis of this device , namely the _
double squid _ , is presented in @xcite .
we shall briefly describe the system and introduce its resulting hamiltonian .
a scheme of the double squid is given in fig .
[ double_squid ] .
the two sides of the junction with capacitance @xmath102 and critical current @xmath103 are connected by a superconducting loop with inductance @xmath70 ( the inner loop inductance is assumed negligible ) .
an external flux @xmath104 is imposed through the loop by an auxiliary coil .
taking the flux through the superconducting loop @xmath105 as the quantum variable ( @xmath106 with @xmath107 the flux quantum , @xmath108 ) , and its conjugate variable the charge on the capacitance , @xmath109 , the resulting hamiltonian is @xmath110,\ ] ] + where @xmath111 is the conjugate momentum of @xmath112 , with @xmath113 the charge on the capacitor ; @xmath114 is the effective mass ; @xmath115 ; @xmath116 and @xmath117 are the control fluxes in reduced units @xmath118 . the resulting potential is a continuous double - well as shown in fig .
[ squid_well]a .
the potential barrier height is modified by changes in @xmath119 ( [ squid_well]b ) .
an asymmetry in the wells energies may be induced by tuning @xmath120 ( [ squid_well]c ) .
the resulting two - state system is also referred to as a _ flux qubit_. note that the state readout is done via coupling of the flux with a second squid , activated when the measurement is needed .
however , other readout methods exist .
@xcite on the potential barrier .
( c ) the effect of the control flux @xmath120 on the symmetry . , height=158 ] note that in practice , to maintain quantum coherence in a given two - state system , it has to be cooled down to very low temperatures . for the system to be robust to thermal excitations , the energy gap has to significantly exceed the thermal energy : @xmath121 this condition sets a lower bound to the rabi frequency in the high barrier state of @xmath119 .
a temperature of @xmath122 corresponds to a frequency of @xmath123 , while the typical range of the squid qubit is around @xmath124 ( see [ sec : qubit manipulation ] ) ; thus it is required to cool the system to about @xmath125 or below @xcite .
+ other noise sources contribute to the decoherence of the qubit state , which vary in dominance per system conditions .
these factors normally divide into two classes - energy relaxation ( drifting towards the lower energy state ) and dephasing ( diffusing of the vector in the longitudinal direction ) . for further details and analysis of decoherence factors , see @xcite .
we also note that other realisations have been made for the double - well potential , e.g. double quantum dots @xcite or interacting bose - einstein condensates trap @xcite . nevertheless , the superconducting realisation is of central influence , considering its well - established design in a scalable size and possibility of on - chip integration with electronic circuits @xcite .
we may now use the above derivations to describe a procedure of double - well qubit transformation : a desired operation on the ` spin ' is translated to a sequence of changes of the potential features - barrier height and asymmetry - which in turn translate in the double squid to appropriate pulses in the fluxes @xmath126 @xcite .
the common method of control is via rf - pulses throughout the coils ; however , in their work , castelli et al .
introduce a short pulse method for ` ultra - fast ' manipulation of the qubit @xcite .
further implementation related questions of operation frequency and coherence time are addressed later in this section .
we list and show the qubit operations in the system ; first rotations about the x- and z- axes , then a tilted xz axis , and finally the generation of rotations about any axis in the bloch sphere using the former steps .
rotation about the @xmath51-axis corresponds to @xmath127 in the hamiltonian ( [ perturbed hamiltonian 2 ] ) , or simply the unperturbed hamiltonian ( [ qubit_hamiltonian ] ) .
the rabi oscillation frequency may be changed by adjusting the tunneling amplitude ( @xmath119 control ) . in @xcite , a rapid lowering of the barrier
was used for reaching a single - well condition , before resetting the high barrier after some time @xmath128 . the rotation angle
@xmath56 then depends on the precession time by @xmath129
. the potential change can not be too fast though , or it would excite upper energy levels . for details
see @xcite . ] .
thus the rabi frequency varies and the rotation angle is @xmath130 setting @xmath131 as the gate initialization moment .
the resulting qubit gate is an x - rotation , @xmath132 : @xmath133 -i\sin \frac{\alpha}{2 } & & \cos \frac{\alpha}{2 } \end{pmatrix}. \end{array}\ ] ] for the special case of @xmath134 we get the not gate . a rotation about the @xmath63-axis , or a _ phase gate _
, can be realized by inducing an amplitude - modulated rf pulse to @xmath120 in the resonance frequency @xmath135 .
the addition to the hamiltonian is then @xmath136 in the rotating frame or reference , the rf - pulse produces a @xmath137-rotation with a frequency proportional to @xmath138 .
the area of the pulse envelope determines the angle of rotation . as the pulse ends , the state returns to its @xmath51-precession , and the phase gate is complete as the rotating frame syncs with the original axes @xcite .
the shaping of the pulse is an important issue .
the operations are in general desired to be fast , in order to perform as many operations as possible within the qubit coherence time .
on the other hand , if the pulse is too fast , it will excite higher energy levels which we intend to avoid .
+ a solution to this problem was offered by mcdermott et al .
@xcite by using sfq ( single flux quantum ) technique : the control flux is changed not in a single pulse but in a series of short and weak pulses , with pulse - to - pulse spacing equal to the period of the oscillation , i.e. @xmath139,\ ] ] where @xmath140 is the flux quantum equals an integral of the pulse shape ; the shape itself is not significant since the time of a single pulse is short enough relative to @xmath57 .
] , @xmath57 is the oscillation period and @xmath141 is the number of pulses .
this approach is analogous to pumping up a swing by giving a short pulse once per cycle rather than forcing the swing throughout its entire movement .
since the pulses add up coherently , the deposited energy scales as @xmath142 and an operation can be carried out in the order of @xmath143 pulses , with a typical total time of several @xmath144 @xcite . using the asymmetry control ( @xmath120 ) we may tilt the rotation axis by an angle @xmath82 ( [ angle and params relation ] ) from the @xmath63-axis , and get the general form of @xmath145 where @xmath146 .
the rotation axis @xmath147 is in the @xmath93 plane of the bloch sphere , as illustrated in fig .
[ bloch_sphere_xz_plane ] .
axis in the bloch sphere.*,height=188 ] this hamiltonian generates a rabi oscillation about the tilted @xmath60 axis , in an angle @xmath56 determined again by eq .
[ rabi rotation angle ] .
however , we note that similarly to the rf - pulse case , the transition of the axis from its previous position is carried out in a finite time in order to not excite high energy levels . the rotation of the state during the time of transition may not be neglected .
this issue may be resolved again by using a pulse train approach : the hamiltonian is changed not in a single pulse but in a series of short and weak pulses .
their spacing is synchronized in each step with the current precession frequency @xmath148 so that for step @xmath149 : @xmath150 and the pulse train has the form @xmath151.\ ] ] this process ends with the @xmath60 axis tilted in the desired angle , without disorienting the state from its initial alignment .
taking @xmath152 as the @xmath60 direction , we may write the gate matrix : @xmath153 = \begin{pmatrix } \cos \frac{\alpha}{2 } + i\sin\frac{\alpha}{2 } \cos\theta & & -i\sin \frac{\alpha}{2 } \sin \theta \\[1 mm ] -i\sin \frac{\alpha}{2 } \sin \theta & & \cos \frac{\alpha}{2 } - i\sin\frac{\alpha}{2 } \cos\theta \end{pmatrix}. \end{array}\ ] ] the special case @xmath95 coincides with ( [ qubit x rotation ] ) as we would expect . as an example , shall we set the parameters to @xmath154 for an angle @xmath155 , and a @xmath128 which gives @xmath156 ( where @xmath157 ) , we receive the _ hadamard gate _ , namely @xmath158 1 & & -1 \end{pmatrix}. \end{array}\ ] ] how far can @xmath159 be tilted from @xmath160 ?
[ angle and params relation ] holds in the regime of eq .
[ regime for theta ] . for a given @xmath84 within this regime
, we may raise the barrier to make @xmath6 small to the extent in which thermal noise @xmath161 is comparable to the energy gap and decoheres the state .
thus the angle from @xmath92 is roughly lower - bounded by @xmath162 and @xmath159 may be practically set in the range @xmath163 .
this allows precessions close to the @xmath92 axis but not exactly aligned with it . for an exact @xmath137-rotation ,
the rf method above is available instead .
any single - qubit gate can be associated with a certain rotation in the bloch sphere about an axis @xmath164 , which we may denote as : @xmath165 this rotation can be achieved in 5 steps of @xmath166- and @xmath137-rotations as follows : @xmath167 with @xmath168 . the first steps @xmath169 transform @xmath164 to the @xmath92 axis , @xmath170 performs the desired rotation , and finally @xmath171 rotate @xmath164 back to its original orientation . + by applying the symmetric point @xmath166-rotation and rf pulse @xmath137-rotations described above
, any general transformation of the qubit may thus be tailored .
although this approach is quite intuitive , it is not the most efficient in the number of required steps .
it can be shown @xcite that a rotation about a general axis in the sphere , @xmath172 , may be performed by a two - step rotation about two axes @xmath173 set in one plane , e.g. the @xmath93 plane : @xmath174 these axes may be set within one half of the plane without loss of generality , i.e. with angles @xmath175 $ ] .
it is thus straightforward to use the above result for execution of a general single - qubit transformation , using two tilted rotations about proper axes in the @xmath93 plane .
we may thus write a general rotation in the bloch sphere as @xmath176 where @xmath64 stands for a global phase .
the rotation @xmath172 is characterized by three parameters - the spherical coordinates @xmath177 of @xmath164 and the rotation angle @xmath112 .
the right - hand - side however involves five parameters - @xmath178 .
the equation is therefore under - constrained , and many solutions exist for the rotations @xmath179 .
solutions may be analytically constructed in a simple procedure - see @xcite for details and examples .
it is therefore shown that the scheme of inducing asymmetry to carry out @xmath93 rotations can be used for a general single - qubit operation in only 2 steps . as far as we know
, this operational method has not been proposed so far , whereas the asymmetry was mainly used for preparation in @xmath12 or @xmath13 , while followed by the @xmath166 and @xmath137 operations described above .
the time scale for this sfq train would in fact be shorter than in the rf pulse case since the @xmath180 get smaller with @xmath149 ; eventually the arbitrary gate can thus be carried out in the time scale of few @xmath144 . the ability to prepare
the qubit state is essential for performing any experiment or computation . in general , preparation of the state
can be performed by simply waiting for the system to relax in its ground state .
for instance , preparation of the qubit in a @xmath181 or @xmath182 state ( left / right well ) is done by creating an extreme potential tilt , with a @xmath120 pulse , to the left or right , so that only one well is allowed , and letting the system relax there .
this prepares a localization of the system in one side , and following computational operations ( changes of @xmath126 ) would continue from this point . as mentioned above , this preparation scheme is currently the common use of the @xmath120 bias flux , rather than @xmath93 axis - tilting .
a preparation of the symmetric @xmath183 state can be made similarly by relaxing the system with no @xmath120 tilt , i.e. in the symmetric double - well configuration , where the ground state @xmath23 coincides with @xmath183 .
+ for a general initialization along a @xmath184 axis , an appropriate @xmath120 bias is applied for the state to relax there .
the most promising implementation of a double - well qubit nowadays , as discussed above , is using superconductors and the josephson effect .
several experiments have realized such systems in attempt to observe high coherence times ( limited by unwanted couplings in the system ) , while maintaining control of the qubit and connectivity in a circuit . by coupling the qubit to a stabilizing harmonic oscillator ibm s system ( 2006 )
reached coherence times of 25 - 35@xmath144 @xcite .
castellano et al . (
2010 ) have demonstrated oscillations of the flux qubit for several cases of lowering the barrier ( @xmath119 ) , thus controlling the oscillation frequency in a range of 10 - 25@xmath185 with coherence times of order 10@xmath144 @xcite . in recent years ( 2011 - 2013 )
several groups have begun using a technique of embedding qubits in 3d cavities . this method has managed to bring the coherence time scale up to 20 - 100 @xmath186 @xcite .
the use of such qubits in scalable integrated circuits was demonstrated in 2013 by barends et al .
these recent experiments therefore show promising implementations for the double - well josephson qubit , reaching increasing achievements in the balance of coherence time , control and connectivity @xcite .
it is apparent that progress in the coherence times has an exponential fashion , resembling that of moore s law @xcite .
the oscillation frequency @xmath187 differs among systems in the approximate range of @xmath188 , or periods of @xmath189@xcite . a gate such as @xmath137-rotation
is claimed to be carried out as fast as @xmath190 by huang et al .
@xcite . roughly estimating @xmath191 as a sufficient upper limit for arbitrary transformations , recent coherence times of @xmath192 already fit hundreds , if not thousands , of consecutive single - qubit gates . the above analysis describes implementation and control of a 2-level register , or qubit , using a double - well . in the following section a generalization of these principles
is presented to a three - level system in a triple - well .
in a manner resembling the double - well case , we present a realisation of a qutrit system using a particle in a spatial potential . for this case
, we henceforth define and use the one - dimensional _ triple - well _ potential , with periodic boundary conditions .
the qutrit analysis will be based on the @xmath2 group similarly to @xcite .
the periodic triple - well potential consists of a line ( with total length @xmath70 ) of three cavities , separated by narrow potential barriers ( the identification @xmath193 applies ) .
an example for such a physical system would be a particle confined within a ring ; taking the angle @xmath82 as the position coordinate , we recognize @xmath194 , with a potential @xmath195 . for simplicity , we again illustrate our discussion using a square potential model shown in fig .
[ triple_well_potential ] . ; the ground state ( red ) and a state in the first excited subspace ( green ) are illustrated.,height=188 ] nevertheless , this analysis applies for a more general potential as well , should it fulfill the following set of requirements : + ( i ) translational symmetry : the three wells are equally shaped , namely : @xmath196 . in other words ,
the potential remains unchanged under the translation operation @xmath197 .
+ ( ii ) parity symmetry : the wells and barriers are symmetric about their center : @xmath198 , where @xmath199 is the center of a well or a barrier . given ( i ) , it is then sufficient to require symmetry under the parity operation : @xmath200 given that @xmath19 is the center of a well .
+ ( iii ) the barriers are high enough compared to the first energy eigenstates , so that energy separations among the three lowest eigenstates are much smaller than the distance from higher levels : @xmath201 similarly to the double - well case , a state of the triple - well system would in effect consist of a combination of three one - well ground states .
although the square triple - well is not realistic , much like the square double - well , realisations of a smooth form are implementable .
for example , a coherent triple - well system has been formed using superconductor circuits analogous to the rf - squid , namely a _
superconducting qutrit _ as described in @xcite , although it so far lacks the full symmetry required in our discussion .
an alternative implementation may be considered using quantum dots for a controllable charge qudit @xcite , as further discussed later in [ sec : dwells ] , via fabrication of the dots in a triangular topology . in whichever implementation
is used , controllability of the potential features is indeed a crucial component , as it sets the ground for executing qutrit transformations ( similar to the qubit manipulation discussion in [ sec : qubit manipulation ] ) .
the eigenfunctions and evolution of the three - well system in the low energy regime may be obtained by the schrdinger equation , in a similar manner to those of the double - well .
however , we shall utilize the mathematical simplicity of the 3-state formalism and perform the derivations directly within its frame , in analogy to the two - state analysis of [ chap : double - well ] .
we first define basis vectors for our @xmath2 representation to be the _
localized _ states of the wavefunction in each of the three wells - similarly to the double - well case . denoting individual wells by @xmath202 ,
our basis vectors are hence the localized states : @xmath203 in analogy to the @xmath34 matrices in the 2-level case , we shall make use of the 8-generator set in the fundamental representation of @xmath2 , the @xmath204 matrices : @xmath205 , where : @xmath206 \ \lambda_3= \begin{pmatrix } 1 & 0 & 0 \\ 0 & -1 & 0 \\ 0 & 0 & 0 \end{pmatrix } , \lambda_4= \begin{pmatrix } 0 & 0 & 1 \\ 0 & 0 & 0 \\ 1 & 0 & 0 \end{pmatrix } , \\[7 mm ] \
\lambda_5= \begin{pmatrix } 0 & 0 & -i \\ 0 & 0 & 0 \\ i & 0 & 0 \end{pmatrix } , \
\lambda_6= \begin{pmatrix } 0 & 0 & 0 \\ 0 & 0 & 1 \\ 0 & 1 & 0 \end{pmatrix } , \\[7 mm ] \lambda_7= \begin{pmatrix } 0 & 0 & 0 \\ 0 & 0 & -i \\ 0 & i & 0 \end{pmatrix } , \
\lambda_8= \frac{1}{\sqrt{3 } } \begin{pmatrix } 1 & 0 & 0 \\ 0 & 1 & 0 \\ 0 & 0 & -2 \end{pmatrix}. \end{array}\ ] ] these lie algebra matrices obey the commutation relations @xmath207
= if_{ijk}\lambda_k \end{array}\ ] ] with structure constants @xmath208 , completely antisymmetric in their three indices : @xmath209 the hamiltonian matrix is constructed by combining the three generators corresponding to two - well transitions : @xmath210 where @xmath211 . in the gell - mann notation
this is equivalent to : @xmath212 the hamiltonian matrix is hence @xmath213 where again , @xmath6 corresponds to the tunneling amplitudes .
it can be easily shown that this is the most general matrix to comply with the system s symmetries . using the matrix hamiltonian we may derive the qutrit eigenvalues and corresponding eigenstates .
two eigenvalues emerge : + the lower energy @xmath214 corresponds to a _
symmetric _ eigenvector : @xmath215 demonstrated by the lower wavefunction in fig .
[ triple_well_potential ] .
+ the energy of the excited level @xmath216 corresponds to a _ degenerate subspace _ of ` antisymmetric ' states satisfying : @xmath217 for @xmath218 being the ( complex ) amplitudes of @xmath219 .
+ + for example , a basis for the @xmath220 subspace may be comprised by the states : @xmath221 the upper wavefunction illustrated in fig .
[ triple_well_potential ] exhibits the form of @xmath222 , or equally @xmath223 shifted one cell to the left .
thus a general state of the qutrit may be expanded in terms of these three states : @xmath224 the time evolution of the general state will be described as : @xmath225 a key result thus emerges for the symmetric triple - well qutrit system : any state evolves _ periodically _ in time , as in the double - well case - even when the hilbert space is three - dimensional ; i.e. the wavefunction undergoes periodic _
revival_. + + the periodicity in eq .
[ qubit periodicity ] applies for the qutrit evolution as well .
the revival frequency is @xmath226 in analogy with the double - well qubit , the evolution of a qutrit state may be viewed as precession about the hamiltonian axis ( eq .
[ eq : h_direction ] ) in the @xmath2 ` space ' with frequency @xmath135 . continuing the double - well analogy ,
let us define the _ cyclic current _ operator : @xmath227 unlike the current operator @xmath40 in the qubit case , the cyclic current commutes with @xmath228 .
the eigenstates of @xmath229 , namely the _ current states _ , are listed by the corresponding current eigenvalues : @xmath230 + a perturbation @xmath231 would split the @xmath220 degeneracy to different energies for the current states @xmath232 .
the three energy levels after the perturbation are : @xmath233 + i.e. the current states evolve differently in a manner determined by the strength of perturbation @xmath234 .
we have noted that ` rotation ' of the qutrit state is produced by the 8 generator matrices @xmath205 . by generalizing the controllable double - well
, we shall assume in this triple - well system we have control over : 1 .
the height of each barrier , and thus the tunneling between each well couple , @xmath235 with the well indices @xmath236 .
2 . the energy difference ( asymmetry ) between the adjacent two wells @xmath237 .
the three differences are subject to the periodicity constraint @xmath238 a general transformation on a qutrit takes 8 parameters . at a given instant ,
the control fluxes may be set to control 5 degrees of freedom - the tunneling amplitudes @xmath235 and energy differences @xmath237 .
this corresponds to control of the amplitude of 5 generators : the three off - diagonals @xmath239 and two diagonals @xmath240 had we had freedom to shape the control flux pulses as we liked - i.e. instant and high pulses - the control of these components could have allowed any @xmath2 transformation we desire : a transformation @xmath241 may be decomposed into three @xmath242 transformations , operating on each pair of indices at a time : @xmath243 with a proper choice of the right - hand - side parameters @xmath244 .
this decomposition may be achieved by fast operation on one pair at a time , by suppressing the other tunneling amplitudes ( high barrier ) so that the rotation outside the ` active ' pair is negligible .
a transformation @xmath245 could be made in as little as 6 steps - two for each double - well rotation , as noted in [ sec : qubit manipulation ] .
however , as we know by now , the strong instant pulses necessary for this scheme cause unwanted excitations and thus do not fulfill the purpose .
changing the fluxes to a certain bias point is spread over a time interval , in which the original state would undergo an unwanted transformation .
a complex ad - hoc calculation may be done for a certain transformation to be carried out via the gradual change of fluxes .
nonetheless , we propose a method to perform a certain subset of transformations on the system .
a general perturbation @xmath246 to the hamiltonian can be written as @xmath247 for @xmath248 .
thus , perturbing the hamiltonian towards @xmath249 , sets the time evolution to be : @xmath250 an interesting subset of the possible @xmath246 is all the perturbations that _ commute with h _ , namely @xmath251 = 0 $ ] .
in this case the evolution turns out to be : @xmath252 by applying the perturbation @xmath246 for a time @xmath253 , we find @xmath254 namely , that the state is effectively transformed only by the perturbation term : @xmath255 any such operation is part of a subspace of rotations that _ commute with h _ in @xmath2 .
+ the basis for this subspace consists of the following four matrices : @xmath256 or spanned by the gell - mann matrices : @xmath257 therefore , any transformation generator given by @xmath258 will commute with @xmath228 , and thus satisfy eq .
[ eq : commuting_perturbation ] .
+ we divide the above generators into two groups : \(i ) the antisymmetric generator @xmath259 is simply the cyclic current operator @xmath260 presented above .
\(ii ) the symmetric generators @xmath261 : + this trio corresponds to certain actions of the control fluxes , i.e. changes of @xmath235 and @xmath237 . hence , these matrices define a basis for transformations we can induce on the qutrit , based on eq .
[ eq : perturbation_operator ] . to execute such a transformation we activate the flux transition in a low magnitude pulses of @xmath262 , e.g. @xmath263 , over several cycles of @xmath57 as required for the transformation .
consequently , we may gradually lower it back to @xmath264 to restore the revival of @xmath228 .
the cumulative energy of the pulses and the time span determine the degree of rotation generated by @xmath246 .
+ + applying any such perturbation splits the state degeneracy of @xmath220 , defining three ( generally ) non - degenerate eigenstates for @xmath145 .
taking @xmath265 for instance , the emerging eigenstates , apart from @xmath23 , are the @xmath266 defined in eq .
[ eq : example_eigenstates_of_e1 ] above .
the corresponding energy levels are : @xmath267 e_{1,a } = h - \frac{4}{3}\epsilon,\\[1 mm ] e_{1,b } = h + \frac{2}{3}\epsilon . \end{array}\ ] ] note that @xmath268 and @xmath269 produce similar eigenvectors with respective permutation of their values . to simplify the action of these generators we transform into a non - traceless form , moving into @xmath270 : @xmath271 effectively adding a global phase factor to the transformation ( i.e. adding a constant @xmath272 to the energy ) .
we receive the following matrices for @xmath273 : @xmath274 with unaltered eigenstates and constantly raised energies @xmath275 e_{1,a } = h - \epsilon,\\[1 mm ] e_{1,b } = h + \epsilon . \end{array}\ ] ] we may calculate the qutrit operator ( eq . [ eq : perturbation_operator ] ) with @xmath276 for each of the @xmath273 : @xmath277 trying to interpret the above operations we conclude that each @xmath273 generates a @xmath37 rotation for the corresponding pair of wells , with a simultaneous @xmath39 rotation adding a relative phase to the amplitude of the third well . the operation angle is given by @xmath278 .
+ we note a specific choice of setting @xmath279 to produce @xmath95 .
this results in an action of the @xmath273 generators directly on the qutrit state , namely , the elementary @xmath166-gate qutrit operations @xcite : @xmath280 the set of @xmath166-gates provides , for instance , implementation of a set of _ ternary shift gates _ ( see @xcite for further details ) .
another method is proposed to perform a certain important qutrit gate - the _ quantum fourier transform _ ( qft ) : @xmath281 in a sense , it is a generalization of the hadamard gate ( eq .
[ hadamard gate ] ) for more than 2 levels .
the quantum fourier transform is of main interest in quantum computing , being a key step in shor s factoring algorithm .
the advantageous use of @xmath282 to carry out the quantum fourier transform has been discussed @xcite . the three - state fourier transformation ( eq . [ discrete fourier transform 3 level ] ) may be achieved by a _ change of basis _ to the conjugate variable of the flux - the charge in the junctions capacitance : @xmath283 as conjugate variables , the eigenstates of the charge are fourier combinations of the flux eigenstates . in our three - state regime ,
the eigenvalues of the charge operator @xmath284 are @xmath285 ( note that @xmath286 ) .
therefore , measuring the charge @xmath284 corresponds to applying the dft matrix on the qutrit and performing a measurement on the result .
note that the flux states which produce a certain charge eigenvalue with certainty are the eigenstates of the current operator @xmath229 .
+ other methods may be considered to utilize the charge - flux relation for a computation , i.e. coupling of the charge to other qutrits in the system .
in this section we turn to generalize the three - well qutrit notion to the case of @xmath0 cyclically coupled wells , acting as a @xmath0-level register , i.e. a _
qudit_. it turns out that two different directions may be considered as a canonical generalization of the double- and triple - well . in what follows we present both and
then focus on the second which appears to be more practical .
one proposed generalization is a fully connected network of wells , with equal tunneling amplitudes among each of the pairs .
this corresponds to the following @xmath287 hamiltonian : @xmath288 the system is symmetric under permutations of the wells , and thus is highly symmetric in @xmath1 .
furthermore , a dynamical analysis reveals but two energy levels : @xmath289 for the symmetric ground state , @xmath290 and the ( @xmath291 ) - degenerate @xmath292 corresponding to the orthogonal subspace , i.e. @xmath293 , where @xmath294 are the eigenvector components . as a system of purely two energies , a _ state revival _ mechanism occurs , much like in the previously discussed qutrit case .
however well - behaved the above generalization is , we also note that a physical realisation of this system may be a challenge , whereas an equal tunneling between all pairs should me maintained .
we thus proceed to discuss a different generalization , of a seemingly more feasible nature .
a cyclic @xmath0-well system consists of linearly connected @xmath0 potential wells , each coupled to its two adjacent wells , once again defined in a compact space , @xmath193 .
the system is illustrated in fig .
[ d_well_potential ] .
-well potential in a compact space .
* edge points are identified : @xmath295.,height=188 ] the system satisfies the analogue of conditions ( i ) - ( iii ) defined above for the three - well system , with the appropriate changes for applying to the @xmath0 wells case : * the system is symmetric under translations by @xmath296 , or : @xmath297 .. * the resulting spectrum consists of @xmath0 energy states with separations much smaller than the rest of the states : @xmath298 for @xmath299 , as shown in the next subsection to result of the bloch theorem . as in the previous sections , based on these conditions we turn to the @xmath0-level formalism to derive our results , rather than using the explicit spatial eigenstates .
the @xmath0-dimensional basis of the qudit hilbert space is defined , in a straightforward generalization of the @xmath300 cases , as the states of maximally localized wavefunction in each of the @xmath0 wells , or : @xmath301 transformations of the qudit state are matrices in the fundamental representation of the @xmath1 group .
the generator set consists of @xmath302 matrices , spanning all the possible qudit transformations .
the symmetric hamiltonian is of the form @xmath303 with an identical tunneling amplitude @xmath6 for all neighbouring transitions .
solving for the system eigenstates , we note that the resulting vectors are _ bloch states _ ; this is evident from the symmetry and neighbour interactions of @xmath228 , which corresponds to a periodic lattice in a tight - binding regime @xcite .
the states are eigenstates of the particle s _ modular momentum _ @xcite , with its @xmath0 corresponding eigenvalues @xmath304
the corresponding energies are given by @xmath305 we note that apart from the ground state @xmath306 , and the @xmath307 state in case of odd @xmath0 ( where @xmath308 ) , the spectrum is divided to _ degenerate pairs _
@xmath309 , in agreement with the symmetry of @xmath310 about @xmath311 .
we also note that eq .
[ ddifferences ] is justified by the model s band structure , featuring a gap between the bands with order of magnitude of the one - well system energy separation @xmath312 , much higher than the in - band separations , which are of order @xmath6 .
the manipulation of the @xmath0-well qudit may be similarly performed via control of each potential barrier and each adjacent energy gap .
we do not go into a detailed discussion on qudit evolution and operations , as it would follow similar guidelines of the previous @xmath300 discussion .
we note , however , the following points : * in general , the qudit state does _ not _ undergo periodic revival , as occurred in the double- or triple - well cases .
this is due to the fact that more than two energy values exist among the state basis vectors .
nevertheless , certain cases do feature periodic revival , such as @xmath313 and @xmath314 , where the energy gaps satisfy proper rationally numeric ratios .
* the @xmath0 modular momentum states are in fact a generalization of the three current states defined in eqs .
[ current_states_triple ] for the @xmath315 case , and are hence eigenvectors of the cyclic current operator in eq .
[ cyclic_current_operator ] , straightforwardly extended to @xmath287 matrices .
as mentioned above , multi - qudit gates are necessary in order to execute general quantum algorithms , when two - qudit gates are the simplest choice .
methods of magnetically coupling two flux qubits have been shown using mutual inductive coupling or coupling with transformer loops , and two - qubit operations have been demonstrated such as controlled - z and controlled - not @xcite .
furthermore , execution of shor s factoring algorithm have been demonstrated using five flux qubits with bi- and tripartite entanglements , factoring the number 15 with 48% success @xcite .
quantum fast fourier transform using multilevel atoms was also suggested @xcite . the broad subject of flux qudit couplings produced gates is out of scope for this work .
for further details on the entangling process and construction of the gates , one is referred to @xcite .
we have discussed above the computational advantage of a higher @xmath0 qudit .
what then is the downside of higher @xmath0 , and what range of @xmath0 values would be optimal ? first , as we increase @xmath0 we may have to increase the total length @xmath70 of the device , which might increase the duration of time needed for completing a computation , and moreover , may be longer than the achievable decoherence times .
second , higher @xmath0 values make the energy differences between the levels smaller , as can be seen from eq .
[ energiesen ]
. it might hence challenge our basic assumption in eq .
[ ddifferences ] .
these considerations may indeed suggest that there is a non - trivial @xmath0 value for which the system of wells is optimal , given a specific implementation and a computation to perform . in the double- and triple - well sections we have referred to the main implementation using superconducting devices , which form the controllable potential well forms about the flux quantum variable .
however , a similar implementation for a superconducting @xmath0-well system has yet to be proposed .
nonetheless , an implementation of interacting @xmath0-well qudits using quantum dots have been demonstrated by schirmer et al .
their proposed solution features a charged particle in a quantum dot array forming a potential well chain . to agree with the above qudit analysis ,
however , the fabrication of the quantum dot should be modified to an enclosed circular form .
in addition , to reach a fault tolerant operation in a real setup , a scheme for error correction , e.g. the one implemented in @xcite would probably be needed .
the use of double - well systems for computation , particularly via superconductor systems , has been an active research area during the last years .
its characteristics were studied via multiple experiments , a selection of which were mentioned throughout the work @xcite .
while many studies of certain operations and gates on the flux qubit were performed , the feasibility of arbitrary operations in the context of high computational efficiency has not been much discussed to our knowledge .
furthermore , it appears that the possibly advantageous extension of the ideas towards large @xmath0 system with periodic boundary conditions has not been thoroughly considered .
this work has examined the foundations of the potential well system in the two - state regime : from the evolution of its ` spin ' operators and the corresponding physical observables , to its manipulation within the common control scheme of squid systems .
it was shown how these changes of the potential s shape translate to qubit operations , and eventually , how they may be utilized to carry out arbitrary @xmath242 operations in minimal steps for efficient usage of the coherence time .
an extension of the sfq method @xcite used for rf - pulsing ( @xmath137-rotation ) was proposed for rapid execution of hamiltonian axis tilting ( rotation about an arbitrary axis in @xmath93 ) .
we proceeded to propose an analogous setup for realisation of a qutrit using a controllable cyclic triple - well potential .
analysis of the three - state dynamics was shown , and schemes for execution of a subset of @xmath2 operators , among which the ternary x - gates @xcite and the 3-state quantum fourier transform .
we eventually generalized the ideas to @xmath0 well systems in two proposed directions - the fully connected wells and the periodic well chain - noting their characteristics and higher dimensional dynamics in its @xmath1 context .
the results of our work may hopefully strengthen the motivation for using qutrits or qudits as alternatives for the traditional qubit , in the context of potential - wells realisation . achieving a fully controllable @xmath0-well system - either based upon flux qubits in superconducting circuitry ,
charge qubits via quantum dots , or another suitable mechanism - may indeed prove as a significant step towards the goal of a general - purpose quantum computation system .
further research of implementation methods is encouraged , with the open questions of optimal @xmath0 and system scalability to be tackled within the scope of each considered method . + * acknowledgments * we thank boaz tamir for many helpful discussions .
y.a . acknowledges support from the israel science foundation ( grant no .
1311/14 ) , the icore excellence center ` circle of light ' and the german - israeli project cooperation ( dip ) for support . e.c . was supported by erc adg nlst .
we analyze the example of an ideal one - dimensional square potential double - well . in the 1d hamiltonian @xmath316 we have for @xmath317 : @xmath318
our regime of interest refers to a confined particle at energies less ( particularly much less ) than @xmath319 .
namely , @xmath320 the wavefunction @xmath321 thus splits into three parts , i.e. @xmath322 in the left well ( region i ) , @xmath323 in the right well ( region iii ) and @xmath324 for the barrier ( region ii ) . note that based on eq . [ potential_relations ]
we may take the amplitude in the outer regions to be negligible : @xmath325 for @xmath326 .
the energy spectrum consists of pairs of states , @xmath327 and @xmath328 , nearly degenerate with respect to the energy difference between pairs .
our interest is focused on the lowest states @xmath23 , @xmath24 ( illustrated in fig .
[ double_well_states ] ) , for which the energy difference @xmath329 satisfies eq .
[ energies_low_states ] .
g. wendin and v. shumeiko , superconducting quantum circuits , qubits , and computing , handbook of theoretical and computational nanoscience 1.ch .
chalmers university of technology , gothenburg , sweden ( 2005 ) , arxiv preprint cond - mat/0508729 .
c. checkley , interferometry of flux qubits driven by radio frequency field , diss .
royal holloway , university of london ( 2009 ) .
devoret , a. wallraff , and j.m .
martinis , superconducting qubits : a short review , arxiv preprint cond - mat/0411174 ( 2004 ) .
f. yoshihara et al .
, decoherence of flux qubits due to 1/f flux noise , phys .
97 , 167001 ( 2006 ) .
j. gorman , d.g .
hasko , and d. a. williams , charge - qubit operation of an isolated double quantum dot , phys .
95 , 090502 ( 2005 ) .
j. esteve et al . , realizing a stable magnetic double - well potential on an atom chip , eur .
j. d 35 , 141 - 146 ( 2005 ) .
shim et al .
, single - qubit gates in two steps with rotation axes in a single plane , arxiv preprint arxiv:1303.0297 ( 2013 ) .
h. paik et al .
, observation of high coherence in josephson junction qubits measured in a three - dimensional circuit qed architecture , phys .
107 , 240501 ( 2011 ) .
b. sanzone , d - wave systems announces the general availability of the 1000 + qubit d - wave 2x quantum computer | d - wave systems ( 2015 ) .
e. cohen and b. tamir , int .
j. quantum inf .
12 , 1430002 ( 2014 ) .
khan , m.a .
perkowski , and p. kerntopf , multi - output galois field sum of products synthesis with new quantum cascades , ieee conf .
proc . of the 33rd international symposium on multiple - valued logic , 146 - 153 ( 2003 ) .
j. ghosh and m.r .
geller , controlled - not logic gate for phase qubits based on conditional spectroscopy , quantum inf .
11 , 1349 - 1357 ( 2012 ) . t. yamamoto et al .
, quantum process tomography of two - qubit controlled - z and controlled - not gates using superconducting phase qubits , phys . rev .
b 82 , 184515 ( 2010 ) .
e. lucero et al . , computing prime factors with a josephson phase qubit quantum processor , nat .
phys . 8 , 719 - 723 ( 2012 ) .
|
quantum computation strongly relies on the realisation , manipulation and control of qubits . a central method for realizing qubits is by creating a double - well potential system with a significant gap between the first two eigenvalues and the rest . in this work
we first revisit the theoretical grounds underlying the double - well qubit dynamics , then proceed to suggest novel extensions of these principles to a triple - well qutrit with periodic boundary conditions , followed by a general @xmath0-well analysis of qudits .
these analyses are based on representations of the special unitary groups @xmath1 which expose the systems symmetry and employ them for performing computations .
we conclude with a few notes on coherence and scalability of @xmath0-well systems .
|
as we initiate the era of the large hadron collider ( lhc ) , we await with anticipation the expected discovery of supersymmetry and the higgs states required to break electroweak symmetry and stabilize the electroweak scale . on the other hand
, there is thus far no concrete model that can explain all observed physics in a comprehensive mathematical framework .
unique predictions that can be tested at the lhc , future international linear collider ( ilc ) , and other forthcoming experiments are necessary if string theory is to be substantiated as the correct fundamental description of nature . following a top - down approach , it may be feasible to derive all known observable physics from a fundamental theory such as string theory .
in contrast , the bottom - up approach offers the possibility to infer the framework of the fundamental theory at high - energy from a low - energy signal at the experiments . in the spirit of this bottom - up approach ,
our goal here is to study grand unified theories ( guts ) from f - theory , which have seen exciting progress the past two years , and present f - theory gut low - energy physics observable at current and future experiments .
f - theory can be considered as the strongly coupled formulation of ten - dimensional type iib string theory with a varying axion ( @xmath2)-dilaton ( @xmath3 ) field @xmath4 @xcite .
guts were first locally constructed in f - theory two years ago @xcite , and subsequently , model building and phenomenological consequences have been studied extensively @xcite . in f - theory model building , the gauge fields are on the observable seven - branes that wrap a del pezzo @xmath5 ( @xmath6 ) surface for the extra four space dimensions , the sm fermions and higgs fields are localized on the complex codimension - one curves ( two - dimensional real subspaces ) in @xmath6 , and the standard model ( sm ) fermion yukawa couplings arise from the triple intersections of the sm fermion and higgs curves .
a brand new feature is that the @xmath1 gauge symmetry can be broken down to the sm gauge symmetry by turning on the @xmath7 flux @xcite , and the @xmath8 gauge symmetry can be broken down to the @xmath9 and @xmath10 gauge symmetries by turning on the @xmath11 and @xmath12 fluxes , respectively @xcite .
intriguingly , the @xmath8 models have both gauge interaction unification and sm fermion unification , and consequently we believe the @xmath8 models are more interesting to study than the @xmath1 models .
moreover , we can break the @xmath8 gauge symmetry down to the flipped @xmath9 gauge symmetry by turning on the @xmath11 flux , wherein the doublet - triplet splitting problem can be solved elegantly @xcite .
the flipped @xmath9 models with tev scale vector - like particles @xcite , which have been dubbed `` @xmath0-@xmath1 '' , have been constructed systematically in f - theory @xcite .
models of this nature can be also obtained in the free fermionic string constructions @xcite . in @xmath0-@xmath1 ,
the @xmath13 gauge symmetries are unified at about @xmath14 gev , and the @xmath9 gauge symmetries are unified above @xmath15 gev .
thus , we can solve the monopole problem .
on top of this , the tev scale vector - like particles are potentially observable at the large hadron collider ( lhc ) , the lightest cp - even higgs boson mass can be lifted @xcite , and the predicted proton decay @xcite is within the reach of the future hyper - kamiokande @xcite and deep underground science and engineering laboratory ( dusel ) @xcite experiments .
recently , the cryogenic dark matter search ( cdms ) collaboration observed two candidate dark matter events in the cdms ii experiment @xcite .
the recoil energies for these two events were 12.3 kev and 15.5 kev , respectively , and the data set an upper limit on the dark matter - nucleon elastic - scattering spin independent cross section around @xmath16 pb .
the probability of observing two or more background events is @xmath17 , so the cdms ii results can not be considered statistically significant evidence for dark matter interactions , although there is some strong possibility that they do in fact represent an authentic signal .
in particular , the favored dark matter mass from the cdms ii data is about 100 gev .
interestingly , the cdms ii experiment can be explained in the supersymmetric standard model with @xmath18 parity where the lightest supersymmetric particle ( lsp ) neutralino is the dark matter candidate @xcite . in this paper , we first briefly review @xmath0-@xmath1 , and then subsequently study the supersymmetry breaking soft terms in the framework of gravity mediation .
although the gaugino masses are not unified at the traditional gut scale , we still have the same gaugino mass relation at the electroweak scale as the minimal supergravity ( msugra ) scenario at one loop @xcite .
we also introduce two parameters testable at the colliders to measure the small two - loop deviations from the msugra gaugino mass relations at the electroweak scale .
next , we incorporate tev scale vector particles and generate regions of the parameter space that can satisfy all current experimental constraints and are consistent with the cdms ii experiment @xcite . in addition , we compute the annihilation cross - section of two neutralinos into two gamma - rays and evaluate the results in light of the first published fermi - lat measurement .
then we compute the new parameters to measure the two - loop deviations for both @xmath0-@xmath1 and msugra .
we find that the deviations from the msugra gaugino mass relation in msugra are smaller than 5% , while the deviations in the @xmath0-@xmath1 are larger as a result of the tev scale vector - like particles .
an analytical comparison of the deviation between msugra and @xmath0-@xmath1 is illustrated .
next , the expected observable final states at the lhc are given for all viable regions of the parameter space .
lastly , the proton lifetime is calculated for the experimentally allowed regions of the @xmath0-@xmath1 parameter space , and we show that the rate of predicted decay is indeed within the reach of the future hyper - kamiokande and dusel experiments .
first , we briefly review the minimal flipped @xmath9 model @xcite .
the gauge group of the flipped @xmath1 model is @xmath19 , which can be embedded into @xmath8 .
we define the generator @xmath20 in @xmath1 as t_u(1)_y=diag ( - , - , - , , ) .
[ u1yp ] the hypercharge is given by q_y = ( q_x - q_y ) .
[ ycharge ] there are three families of sm fermions whose quantum numbers under @xmath19 are f_i= , |f_i= , |l_i= , [ smfermions ] where @xmath21 . to break the gut and electroweak gauge symmetries , we introduce two pairs of higgs fields h= , = , h= , h= , [ higgse1 ] where particle assignments of the higgs fields are h=(q_h , d_h^c , n_h^c ) , = ( _ , ^c _ ,
^c_h ) , [ higgse2 ] + h=(d_h , d_h , d_h , h_d ) , h=(_h , _ h , _ h , h_u ) , [ higgse3 ] and @xmath22 and @xmath23 are one pair of higgs doublets in the supersymmetric sm .
we also add a sm singlet field @xmath24 .
to break the @xmath19 gauge symmetry , we introduce the following higgs superpotential = _ 1 h h h + _ 2 h h h + ( h h - m_h^2 ) . [ spgut ] there is only one f - flat and d - flat direction , which can always be rotated such that @xmath25 .
in addition , the superfields @xmath26 and @xmath27 are absorbed , acquiring large masses via the supersymmetric higgs mechanism , except for @xmath28 and @xmath29 .
the superpotential terms @xmath30 and @xmath31 couple the @xmath28 and @xmath29 with the @xmath32 and @xmath33 , respectively , to form heavy eigenstates with masses @xmath34 and @xmath35 . in consequence
, we naturally achieve doublet - triplet splitting due to the missing partner mechanism @xcite .
the triplets in @xmath36 and @xmath37 only have small mixing through the @xmath38 term with @xmath39 around the tev scale , so we also solve the dimension five proton decay problem from the colored higgsino exchange . in flipped @xmath9 models ,
the @xmath13 gauge couplings are first joined at the scale @xmath40 , and the @xmath1 and @xmath11 gauge couplings are subsequently unified at the higher scale @xmath41 . to separate the @xmath40 and @xmath41 scales and obtain true string - scale gauge coupling unification in free fermionic string models @xcite or the decoupling scenario in f - theory models @xcite
, we introduce vector - like particles which form complete flipped @xmath9 multiplets . in order to avoid the landau pole problem for the strong coupling constant
, we can only introduce the following two sets of vector - like particles around the tev scale @xcite @xmath42 for notational simplicity , we define the flipped @xmath9 models with @xmath43 and @xmath44 sets of vector - like particles as type i and type ii flipped @xmath9 models , respectively .
we choose to focus in this paper on the type ii model , although results are similar in most regards for type i. we emphasize that the type i and ii flipped @xmath9 models have been constructed consistently from f - theory in refs .
@xcite . next ,
let us consider gravity mediated supersymmetry breaking . with the @xmath11 flux ,
the gauge kinetic functions @xmath45 and @xmath46 , respectively for @xmath1 and @xmath11 gauge symmetry , are @xcite @xmath47 where @xmath2 is an integer .
the first term is the original gauge kinetic function for @xmath8 , and the second term arises from the @xmath11 flux effects .
after the @xmath9 gauge symmetries are broken down to the sm gauge symmetry , the gauge kinetic function @xmath48 for @xmath7 is @xmath49 in the gravity mediated supersymmetry breaking scenario where the f - terms of @xmath50 and @xmath51 break supersymmetry , we obtain the gaugino mass relation from the scale @xmath41 down to the scale @xmath40 @xcite @xmath52 where @xmath53 and @xmath54 are gaugino masses for @xmath1 and @xmath11 respectively , and @xmath55 and @xmath56 are the @xmath1 and @xmath11 gauge couplings .
this gaugino mass relation is renormalization scale invariant under one - loop renormalization group equation ( rge ) running . using eq .
( [ u(1)y - gkf ] ) , we obtain the gaugino mass @xmath57 for @xmath7 at the scale @xmath40 @xmath58 where @xmath59 is the @xmath7 gauge coupling . with eq .
( [ m51-gmr ] ) , we obtain at the scale @xmath40 @xmath60 note that the gauge couplings @xmath61 and @xmath55 are split , and thus we do not have universal gaugino masses at the usual gut scale . applying rge running , we obtain the gaugino mass relation which is valid from the scale @xmath40 down to the electroweak scale at one loop @xcite @xmath62 where @xmath63 and @xmath64 are couplings of the @xmath65 and @xmath66 gauge symmetries respectively , and @xmath67 and @xmath68 are masses of the @xmath65 and @xmath66 gauginos .
this gaugino mass relation is the same as that in msugra @xcite .
the gaugino masses can be measured at the lhc and future ilc @xcite , hence , we can test this relation at future experiments . considering two - loop rge running
, we will have a small deviation for the gaugino mass relation given by eq .
( [ msugra ] ) . to quantify the deviations
, we first define @xmath69 ~,~\ , \nonumber \\ \left({\frac{m_i}{\alpha_i}}\right)_{\rm m } & \equiv & { \rm median } \left[{\frac{m_3}{\alpha_3}},~ { \frac{m_2}{\alpha_2 } } , ~{\frac{m_1}{\alpha_1}}\right ] ~,~\ , \nonumber \\ \left({\frac{m_i}{\alpha_i}}\right)_{\rm s } & \equiv & { \rm minimum } \left[{\frac{m_3}{\alpha_3}},~ { \frac{m_2}{\alpha_2 } } , ~{\frac{m_1}{\alpha_1}}\right ] ~.~\,\end{aligned}\ ] ] then , we define the small deviations as follows @xmath70 we emphasize that @xmath71 and @xmath72 . in other words , we have @xmath73 for example , in @xmath74-@xmath1 we have @xmath75
the @xmath0-@xmath1 type models have been constructed locally in f - theory @xcite , and thus we do not know the khler potential for the sm fermions and higgs fields and can not calculate the supersymmetry breaking scalar masses and trilinear soft terms .
interestingly , as long as the scalar masses and trilinear soft terms are around the tev scale , the gaugino mass relation can be preserved very well at low energy to two loops . for simplicity , in this paper , we consider the universal scalar mass @xmath76 and the universal trilinear soft term @xmath77
. essentially speaking , we study the f - theory inspired low energy supersymmetry phenomenology , so that we have four free parameters in our model : @xmath78 , @xmath76 , @xmath77 , and @xmath79 , where @xmath80 , the ratio of the vacuum expectation values of the higgs scalar fields .
in addition , we must choose the sign of @xmath39 .
we take @xmath81 , as suggested by the results of @xmath82 for the muon .
we also assume that the masses for the vector - like particles are universal , at 1 tev .
the contributions to the beta functions of the sm gauge couplings from the vector - like particles must be accounted for in the one - loop and two - loop gauge coupling rges when running the soft terms down to the electroweak scale .
we alter the one - loop and two - loop gauge coupling rge code in suspect 2.34 @xcite to realize the beta function corrections . in an effort to minimize the complexity of revising the code , we use only the one - loop rges for the sm fermion yukawa couplings and supersymmetry breaking soft terms .
the yukawa coupling rges contribute to the gauge coupling rges at second order , so it is consistent to run only the yukawa coupling rges at one - loop .
however , for future investigations of guts with vector - like particles , we intend to extend our analysis to include two - loop rges for all supersymmetry breaking terms .
next , we modify the one - loop gaugino mass rges to include the strong coupling effects of the vector - like particles .
finally , the suspect code must be revised to require gauge coupling unification at @xmath83 , rather than the default unification configuration of @xmath84 .
this necessitates lowering the ceiling on the running of the gauge couplings at high scale due to large effects on @xmath85 at scales above @xmath86 gev .
we see in fig .
[ fig : running_plot ] the split in unification at the @xmath40 scale , as explained in the previous section .
below 1 tev , there is only the running of the sm gauge couplings , which are shown as the red dotted lines in fig .
[ fig : running_plot ] for an unflipped @xmath1 point with universal gaugino masses , however , at 1 tev and higher energies , we see the effects of the vector - like particles on the running of the couplings and masses .
the relic lsp neutralino density and wimp - nucleon direct detection cross - sections are computed with micromegas 2.1 @xcite , although the micromegas code must also be reworked to include the revised rges .
we use suspect as the rge code in our relic density and wimp - nucleon cross - section computations , and hence we implement the revised suspect rges in micromegas as well .
supersymmetry breaking soft terms for the present model are generated using the gaugino mass relation in eq .
( [ gauginomr ] ) at the scale @xmath40 .
the gaugino mass @xmath87 is dependent upon the ratio of the unification scale gauge couplings @xmath88 and @xmath89 , therefore , an iterative procedure must first be used to determine the final value of @xmath87 before any low energy phenomenology can be investigated .
we use a top quark mass of @xmath90 = 173.1 gev @xcite and employ the following experimental constraints : -@xmath1 gaugino masses and gauge couplings from the @xmath91 scale to the @xmath40 partial unification scale .
the red dotted line represents the running of the gauge couplings for unflipped @xmath1 with universal gaugino masses at @xmath92 .
we suppress labeling of the vertical axes to preserve general heuristic applicability to the full parameter space considered.,scaledwidth=75.0% ] 1 .
the wmap 2@xmath93 measurements of the cold dark matter density @xcite , 0.095 @xmath94 0.129 .
in addition , we look at the supercritical string cosmology ( ssc ) model @xcite , in which a dilution factor of @xmath95(10 ) is allowed @xcite , where @xmath96 1.1 . for a discussion of the ssc model within the context of msugra , see @xcite . we also investigate another case where a neutralino lsp makes up a subdominant component and employ this possibility by removing the lower bound .
2 . the experimental limits on the flavor changing neutral current ( fcnc ) process , @xmath97 .
the results from the heavy flavor averaging group ( hfag ) @xcite , in addition to the babar , belle , and cleo results , are : @xmath98 .
there is also a more recent estimate @xcite of @xmath99 . for our analysis
, we use the limits @xmath100 , where experimental and theoretical errors are added in quadrature .
the anomalous magnetic moment of the muon , @xmath101 .
for this analysis we use the 2@xmath93 level boundaries , @xmath102 @xcite .
the process @xmath103 where the decay has a @xmath104 dependence .
we take the upper bound to be @xmath105 @xcite .
5 . the lep limit on the lightest cp - even higgs boson mass , @xmath106 gev @xcite .
we commence with @xmath78 , @xmath107 , @xmath77 , and @xmath108 as free parameters , however , for simplicity we take @xmath109 .
a comprehensive scan of the entire parameter space uncovers @xmath110 to @xmath111 as consistent with all the experimental constraints and latest cdms ii results @xcite , thus we choose tan@xmath112 = 51 for our benchmark point to study in this work .
as shown in fig .
[ fig : movsm5_plot ] , the experimentally allowed parameter space for @xmath113 after applying all these constraints consists of two segregated small regions .
we look to examine more closely those regions of the experimentally allowed parameter space that comply with the recent cdms ii data and the wmap relic density observations .
this encompasses the narrow wmap strip in both experimentally allowed regions , so we choose points from each allowed region within these narrow wmap strips as our benchmark points for analysis .
in fact , with relic densities of @xmath114 = 0.1093 and @xmath114 = 0.1151 , these benchmark points additionally satisfy the very constrained wmap 7-year @xcite results . inspecting the sparticle and higgs spectrum for our benchmark point in table [ tab : masses1 ] with relic density @xmath114 = 0.1093
reveals that the additional contribution of the 1 tev vector - like particles lower the gluino mass quite dramatically .
the gaugino mass @xmath115 runs flat from the @xmath40 unification scale to the electroweak scale as shown in fig .
[ fig : running_plot ] , though , due to susy radiative corrections , the physical gluino mass is larger than @xmath115 at the @xmath40 scale .
this is true for all regions of the @xmath0-@xmath1 parameter space .
in msugra , the focus point region consists of large @xmath107 where the wmap observed relic density can be satisfied with a mixed bino - higgsino state in the lsp due to a small @xmath116 , leading to enhanced @xmath117 annihilation .
however , even though @xmath107 is reasonably large for the table [ tab : masses1 ] benchmark point , here the lsp is 99% bino for @xmath0-@xmath1 .
in contrast , the sparticle and higgs spectrum for our benchmark point in table [ tab : masses2 ] illustrates that the wmap relic density @xmath118 is generated through stau - neutralino coannihilation .
this is demonstrated by the near degenerate mass between the @xmath119 neutralino lsp and @xmath120 stau next - to - lightest supersymmetric particle ( nlsp ) .
accordingly , the lsp for the benchmark point in table [ tab : masses2 ] is 99.9% bino .
-@xmath1 experimentally allowed parameter space for @xmath113 , @xmath109 .
the benchmark points detailed in table [ tab : masses1 ] and table [ tab : masses2 ] are annotated by the black points .
the solid black lines demarcate the wimp - nucleon spin - independent cross - sections .
the dotted line indicates the lep boundary for the light higgs .
the two dashed lines represent an lsp mass of 100 gev and 200 gev .
the dark green line denotes the region satisfying the wmap 2@xmath93 relic density , however , only the bright green region within the designated allowed parameter space can fulfill all the experimental constraints . the dark red contours label @xmath121 proton lifetime predictions , in units of @xmath122 years.,scaledwidth=75.0% ] .sparticle and higgs spectrum for the @xmath78 = 670 gev and @xmath107 = 1215 gev benchmark point illustrated in fig . [
fig : movsm5_plot ] . here , @xmath113 , @xmath114 = 0.1093 , @xmath123 pb , and @xmath124 .
the gut - scale mass parameters for this point are ( in gev ) @xmath115 = 670 , @xmath125 = 670 , @xmath87 = 296 , @xmath109 = 1215 .
the central prediction for the @xmath121 proton lifetime is @xmath126 years .
[ cols="^,^,^,^",options="header " , ] [ tab : masses2 ]
the wimp - nucleon direct - detection cross - sections @xmath127 depicted in fig . [ fig : sigma_plot ] underscore the fact that the case of @xmath113 produces wimps with @xmath127 that comply with the cdms ii upper limits , with our benchmark point in table [ tab : masses1 ] at @xmath123 pb and @xmath128 = 107 gev , and the benchmark point in table [ tab : masses2 ] at @xmath129 pb and @xmath128 = 152 gev . the constraints from previous zeplin @xcite , xenon @xcite , and cdms @xcite experiments are also delineated on the plot , in addition to the forthcoming lux experiment @xcite . despite the fact the wimp - nucleon cross - section of our table [ tab : masses1 ] benchmark point in fig . [ fig : sigma_plot ] is above the cdms ii upper limit , experimental uncertainties and qcd corrections can account for this variation . on the contrary , the entire region of the experimentally allowed parameter space with large @xmath78 and small @xmath107 has wimp - nucleon cross - sections less than the recent cdms ii upper limits .
the present model also possesses regions of the parameter space where the neutralino can account for only a small portion of the overall composition of the total observed dark matter .
the remaining fraction of the observed relic density in this situation would be composed of other particles , such as axions or cryptons , or additional astrophysical matter . in order to enable a direct comparison of fig .
[ fig : sigma_plot ] to the data from cdms ii and other direct - detection experiments , for those points in fig .
[ fig : sigma_plot ] with a relic density less than the observed wmap 2@xmath93 data , we plot a modified wimp - nucleon cross - section @xmath130 . -@xmath1 spin - independent wimp - nucleon cross - sections for @xmath113 , @xmath109 , overlaid with direct - detection limits from recent and forthcoming experiments .
the shaded regions satisfy all experimental constraints .
the benchmark point in table [ tab : masses1 ] is annotated by the blue point , while the benchmark point in table [ tab : masses2 ] is annotated by the red point.,scaledwidth=75.0% ] indirect detection experiments search for high energy neutrinos , gamma - rays , positrons , and anti - protons emanating from neutralino annihilation in the galactic halo and core , or in the case of neutrinos , in the core of the sun or the earth . here , we focus on the annihilation cross - section @xmath131 of two neutralinos into two gamma - rays in the galactic core or halo
. two possible decay channels where wimps can produce gamma - rays in the galactic core and halo are @xmath132 and @xmath133 .
one such current experiment to measure the debris from wimp annihilations is fermi - lat ( formerly glast ) @xcite , with new constraints on cross - sections of neutralino annihilations into two gamma - rays @xcite .
figure [ fig : fermi_plot ] shows that the fermi - lat sensitivity has reached the @xmath0-@xmath1 parameter space .
the @xmath0-@xmath1 cross - sections @xmath131 in figure [ fig : fermi_plot ] are calculated using the modified micromegas 2.1 code to include the effects of the 1 tev vector multiplet in suspect 2.34 .
the fermi - lat collaboration applies four dark matter structure evolution scenarios ( msii - res , msii - sub1 , msii - sub2 and bulsub ) , which we overlay onto the @xmath0-@xmath1 parameter space in figure [ fig : fermi_plot ] , as well as conservative and stringent upper limits on @xmath134 . for thorough descriptions of these four dark matter scenarios and upper limits ,
we refer the reader to @xcite .
additionally , the upper 95% confidence limits for each of the four dark matter scenarios as determined by the fermi - lat collaboration are identified in figure [ fig : fermi_plot ] , including the 90% and 99.999% confidence limits for the fermi - lat reference model msii - sub1 . for the benchmark point in table
[ tab : masses1 ] , @xmath124 , in the very near proximity of the fermi - lat reference model msii - sub1 upper limits for both the conservative and stringent cases , whilst for the benchmark point in table [ tab : masses2 ] , @xmath135 , well below the fermi - lat reference model msii - sub1 upper limits .
-@xmath1 annihilation cross - section @xmath131 of two neutralinos into two gamma - rays for @xmath113 , @xmath109 , overlaid with the most recent fermi - lat constraints .
the shaded regions satisfy all experimental constraints .
the benchmark point in table [ tab : masses1 ] is annotated by the blue points , while the benchmark point in table [ tab : masses2 ] is annotated by the red points . for detailed explanations of the four dark matter scenarios ( msii - res , msii - sub1 , msii - sub2 , bulsub ) , upper limits ( conservative , stringent ) , and confidence limits ( 90% , 95% , 99.999% ) , we refer the reader to @xcite.,title="fig:",scaledwidth=49.0% ] -@xmath1 annihilation cross - section @xmath131 of two neutralinos into two gamma - rays for @xmath113 , @xmath109 , overlaid with the most recent fermi - lat constraints .
the shaded regions satisfy all experimental constraints .
the benchmark point in table [ tab : masses1 ] is annotated by the blue points , while the benchmark point in table [ tab : masses2 ] is annotated by the red points . for detailed explanations of the four dark matter scenarios ( msii - res , msii - sub1 , msii - sub2 , bulsub ) , upper limits ( conservative , stringent ) , and confidence limits ( 90% , 95% , 99.999% ) , we refer the reader to @xcite.,title="fig:",scaledwidth=49.0% ] in the preceding section , we introduced the @xmath136 and @xmath137 parameters of eq .
( [ delta ] ) in order to quantify the small two - loop deviations from the msugra gaugino mass relations in eq .
( [ msugra ] ) at the electroweak scale .
the gaugino masses can be measured at the lhc and ilc @xcite , allowing for a test of these gaugino mass relations . in fig .
[ fsu(5):delta_plot ] we present the deviations @xmath136 and @xmath137 in @xmath0-@xmath1 .
[ fsu(5):delta_plot ] demonstrates that @xmath136 and @xmath137 are indeed small , with @xmath138 and @xmath139 for @xmath0-@xmath1 .
however , these deviations are larger than the f-@xmath1 second loop deviations in ref . @xcite .
thus , it is imperative we understand the reason for the larger deviations in @xmath0-@xmath1 . and @xmath140 parameters for @xmath113 , @xmath109 for @xmath0-@xmath1 .
the shaded regions satisfy all experimental constraints .
the @xmath0-@xmath1 benchmark point in table [ tab : masses1 ] is annotated by the blue points , while the benchmark point in table [ tab : masses2 ] is annotated by the red points . for the benchmark point in table
[ tab : masses1 ] , we find @xmath136 = 0.0615 and @xmath137 = -0.1113 , and likewise , for the benchmark point in table [ tab : masses2 ] , @xmath136 = 0.0607 and @xmath137 = -0.1077.,scaledwidth=75.0% ] in an effort to compare the deviations in @xmath0-@xmath1 , we calculate the deviations in msugra , _
i.e. _ , the minimal @xmath1 model with gravity mediated supersymmetry breaking .
we consider two cases : ( i ) two - loop rge running for the sm gauge couplings and one - loop rge running for the gaugino masses ; ( ii ) two - loop rge running for both the gauge couplings and the gaugino masses .
we find that in case ( i ) , the order of @xmath141 is the same as that in @xmath0-@xmath1 given in eq .
( [ fsuv - gm ] ) , while in case ( ii ) we have @xmath142 we present the small deviations @xmath136 and @xmath137 in msugra for both case ( i ) and case ( ii ) in fig .
[ msugra : delta_plot ] .
we find that for case ( i ) we have @xmath143 and @xmath144 , and for case ( ii ) we have @xmath145 and @xmath146 .
these deviations are similar to the f-@xmath1 second loop deviations in ref .
@xcite , and are indeed smaller than these in @xmath0-@xmath1 .
therefore , we conclude that the existence of tev scale vector - like particles in @xmath0-@xmath1 enlarge the deviations , thereby presenting a potential opportunity to experimentally infer the underlying theory at high energies since there is a rather compelling need for tev scale vector - like particles in @xmath0-@xmath1 . on the contrary ,
unflipped guts require no such tev scale vector - like particles .
the vital point here is that an experimental measurement of the two - loop deviations through the @xmath147 parameters can certainly assist us in determining whether the underlying theory at the string scale is a flipped or unflipped gut . and @xmath140 parameters for @xmath113 , @xmath109 for case ( i ) ( left ) and case ( ii ) ( right ) .
the shaded regions satisfy all experimental constraints.,title="fig:",scaledwidth=49.0% ] and @xmath140 parameters for @xmath113 , @xmath109 for case ( i ) ( left ) and case ( ii ) ( right ) .
the shaded regions satisfy all experimental constraints.,title="fig:",scaledwidth=49.0% ] there is a noticeable resemblance of the experimentally allowed parameter space in fig .
[ fig : movsm5_plot ] with that of msugra . for comparison ,
we study a recent analysis of the msugra parameter space in @xcite . upon closer examination of the two models
, we see that it appears the vector - like particles and partial gaugino mass universality are shifting the upper narrow wmap strip to a larger gaugino mass and smaller @xmath107 .
the essential aspect is that the gaugino mass rge running is changed significantly due to the additional vector - like particles .
the consequences of this include a heavier lsp for large @xmath107 when compared to msugra , with the lsp mass increasing to a limited range centered around 100 gev .
recall , our benchmark point from the wmap strip at large @xmath107 is @xmath128 = 107 gev . in contrast , the lsp mass in the experimentally allowed region of msugra for large @xmath107 and small @xmath148 is well below 100 gev .
moreover , the experimentally allowed region of a larger gaugino mass and smaller @xmath107 gets shifted to a heavier gaugino mass . in these regions of small @xmath107 ,
as already noted , the wmap relic density is generated through stau - neutralino coannihilation .
however , as we described , in @xmath0-@xmath1 this region of large @xmath78 and small @xmath107 possesses wimp - nucleon cross - sections well below the cdms ii upper limit .
it is interesting to consider what we could see at the lhc for the @xmath0-@xmath1 model framework .
we calculate the differential cross - sections and branching ratios with pythia 6.411 @xcite , using the revised suspect 2.34 code that includes the effects of the vector - like particle contributions to compute the sparticle masses for input into pythia 6.411 .
the three lightest sparticles for the @xmath113 table [ tab : masses1 ] benchmark point are @xmath149 .
the production of gluinos @xmath150 and squarks @xmath151 have the largest differential cross - sections at the lhc .
the @xmath151 decays to a gluino and hadronic jet due to the noticeably lighter mass of the gluino in comparison to the squarks , while the @xmath150 will produce a neutralino or chargino along with a hadronic jet .
the @xmath152 produce @xmath153 through @xmath154 with a branching ratio of 64% .
we will get @xmath155 quarks from the remainder of the @xmath152 through light higgs in the process @xmath156 with a branching ratio of 36% .
these light higgs will in turn decay to @xmath157 with a 75% branching ratio .
leptons and hadronic jets will result from the decay @xmath158 , where this is the only kinematically allowed @xmath159 process .
thus , this benchmark point will produce mainly @xmath160 and @xmath161 bosons , and some @xmath155 quarks through light higgs @xmath162 at lhc .
additionally , the vector - like particles can couple to sm fermions and higgs fields via yukawa interactions , and they can then decay to sm fermions and higgs fields , as well as the lsp neutralino .
we see quite different lhc states for the benchmark point in table [ tab : masses2 ] .
this point resides in the region of the experimentally allowed parameter space that generates the wmap relic density through stau - neutralino coannihilation .
hence , the four lightest sparticles for this benchmark point are @xmath163 . here , as in the benchmark point in table [ tab : masses1 ] , the gluino is lighter than the squarks , so all squarks will predominantly decay to a gluino and hadronic jet , with a small percentage of squarks producing a neutralino or chargino , plus a jet , and the gluinos will then decay to neutralinos or charginos as well .
the result is low - energy tau through the processes @xmath164 and @xmath165 .
the lhc final states of low - energy tau in the @xmath0-@xmath1 stau - neutralino coannihilation region are similar to those same low - energy lhc final states in msugra , however , in the stau - neutralino coannihilation region of msugra , the gluino is typically heavier than the squarks .
again , the strong coupling effects from the additional vector - like particles on the gaugino mass rge running reduce the physical gluino mass below the squark masses in @xmath0-@xmath1 . as a consequence
, the lhc final low - energy tau states in the stau - neutralino coannihilation regions of @xmath0-@xmath1 and msugra will differ in that in @xmath0-@xmath1 , the low - energy tau states will result largely from neutralinos and charginos produced by gluinos , as opposed to the low - energy tau states in msugra resulting primarily from neutralinos and charginos produced from squarks .
the first year lhc run in 2011 at 7 tev and anticipated luminosity of 1 - 2 @xmath166 is expected to reach gluino production possibly up to 1 tev , which makes the light gluino in our study here tantalizing for discovery in the early lhc run .
however , for our tan@xmath112 = 51 benchmark study in this work , the heavy squarks are most likely too heavy for production in this early lhc phase .
nonetheless , a closer examination of smaller values of tan@xmath112 shows correspondingly lighter spectra , indicating that signals for tan@xmath167 are within the first year lhc discovery reach .
additionally , in contrast to our tan@xmath112 = 51 points studied here where the gluino is lighter than all the squarks , for tan@xmath168 spectra , the lightest stop is the lone squark lighter than the gluino , providing the potential for a very unique signal .
this same @xmath169 mass pattern has also been uncovered in follow - up studies where this model has been extended to include the @xmath9 unification @xcite .
we shall study this signal much more in - depth in our future work .
instability of the proton is an essential signature of guts , the merger of the sm forces necessarily linking quarks to leptons , and providing a narrow channel @xmath121 of dimension six decay via heavy gauge boson exchange .
we have undertaken a comprehensive analysis of the proton decay rate for the available parameter space of the @xmath0-@xmath1 model of the current report , employing the methods established in refs .
@xcite and the further citations therein .
the 50-kiloton ( kt ) water erenkov detector of the super - kamiokande facility has set lower bounds of @xmath170 and @xmath171 years at the @xmath172 confidence level for the partial lifetimes in the @xmath173 and @xmath174 modes @xcite .
hyper - kamiokande is a proposed 1-megaton detector , about 20 times larger volumetrically than super - k @xcite , which we can expect to explore partial lifetimes up to a level near @xmath175 years for @xmath121 across a decade long run .
the proposal for the dusel experiment @xcite features both water erenkov and liquid argon ( which is around five times more sensitive per kilogram to @xmath176 than water ) detectors , in the neighborhood of 500 and 100 kt respectively , with the stated goal of probing partial lifetimes into the order of @xmath122 years for both the neutral pion and @xmath177 channels . flipped @xmath1 evades the dangerously rapid @xmath178 dimension five triplet higgsino mediated decay by way of the missing - partner mechanism @xcite , which realizes higgs doublet - triplet splitting naturally , without the side effect of strong triplet mixing .
proton lifetime in the @xmath121 dimension six mode is proportional in the fourth power to the gut mass scale , and inversely proportional in the fourth power to the gut coupling .
this extreme sensitivity argues for great care in the selection and study of a unification scenario .
the f - theory derived gut models which feature additional vector - like multiplets at the tev scale modify the renormalization group to yield a substantial increase in the @xmath13 unified coupling , which translates in turn into something like a seven - fold enhancement of the proton decay rate .
moreover , inclusion of tev scale vector multiplets in the renormalization significantly magnifies the separation of the flipped scale @xmath40 , at which @xmath1 breaks and proton decay is established , from the point @xmath179 of true grand unification , which can be extended to the order of the reduced planck mass . in our calculations ,
we take the sparticle and higgs spectrum generated by suspect 2.34 , modified for flipped @xmath180 unification , as input for the light threshold corrections .
these are applied individually to each of the three gauge couplings , taking into account the distinct quantum numbers of each field .
the second loop contribution is likewise individually numerically determined for each gauge coupling , including the top and bottom quark yukawa couplings from the third generation , taken themselves in the first loop .
all three gauge couplings are integrated recursively with the second loop into the yukawa coupling renormalization , with the boundary conditions at @xmath181 set by the value of @xmath182 .
we take this value , reduced modestly from that of @xmath183 employed elsewhere in this report , to avoid pathologies in convergence of the renormalization .
we emphasize that this slightly modified value of tan@xmath112 = 44 is employed only for the proton lifetime calculations , whereas the value of tan@xmath112 = 51 is used in all other calculations for the benchmark points . finally , recognizing that the second loop itself influences the upper limit @xmath40 of its own integrated contribution
, this feedback is accounted for in the dynamic calculation of the unification scale .
the benchmark spectra of the selected points featured in tables [ tab : masses1 ] and [ tab : masses2 ] yield central lifetimes of @xmath184 and @xmath185 years , respectively .
in addition to the light @xmath186-scale threshold corrections from the superpartner s entry into the rges , there may also be shifts occurring near the @xmath187 unification point due to heavy higgs fields and the broken gauge generators of @xmath1 .
the light fields carry strong correlations to cosmology and low energy phenomenology , as is indeed the central theme of this work , so that we are guided toward plausible estimates of their mass distribution .
the heavy thresholds are somewhat more difficult to pin down , although we may conclusively suggest that they act only to lengthen , rather than reduce , the dimension six proton lifetime @xcite . including uncertainties for parameter input and light and heavy thresholds , we conclude that the central values for proton decay , and also a large fraction of the plausible variation , are indeed within the reach of proposed next - generation experiments such as hyper - kamiokande and dusel .
additionally , detectability of tev scale vector supermultiplets at the lhc presents an opportunity for cross correlation of results between the most exciting particle physics experiments of the coming decade .
we emphasize that the cumulative effect of flipped @xmath1 grand unification in the f - theory model building context , applying the freshly detailed methods of analysis recently described , and focusing on the phenomenologically preferred parameter space , is surprisingly fast proton decay .
we have considered gravity mediated supersymmetry breaking in @xmath0-@xmath1 . the gaugino masses are not unified at the traditional grand unification scale , though we do indeed obtain the msugra one - loop gaugino mass relation at the electroweak scale .
however , the gaugino mass relation will have a small two - loop deviation , and this deviation may be measurable at the lhc and ilc . there is a considerable need for tev scale vector - like particles in @xmath0-@xmath1 , while unflipped guts , such as msugra or f-@xmath1 , require no such vector - like particles . in light of this key distinction between flipped and unflipped guts
, we introduced a parameter to measure the small two - loop deviation from the msugra gaugino mass relation at the electroweak scale . to implement a numerical analysis , we modified a popular and well - established public rge code to incorporate the effects of tev scale vector - like particles . in this work , we employed two - loop rges for the gauge couplings , but only considered one - loop rges for the yukawa couplings and soft - supersymmetry breaking terms , though we look to extend this to all two - loop supersymmetry breaking rges in our future work .
the results lead us to conclude there is a clear disparity in the extent of the deviation between guts with vector - like particles , such as @xmath0-@xmath1 , and guts without vector - like particles , such as msugra or f-@xmath1 . the predicted correlation between largeness of the deviation from the msugra gaugino mass relation and the existence of vector - like particles
can be tested at the colliders .
furthermore , we have determined the viable parameter space of this model which simultaneously satisfies all the current experimental constraints and is consistent with the findings of cdms ii .
the cross - section of two neutralinos into two gamma - rays for the experimentally allowed regions of the parameter space was computed and assessed against the first published fermi - lat measurement .
the results showed the @xmath0-@xmath1 parameter space is consistent with the recent fermi - lat findings .
finally , we have calculated the proton lifetime for these experimentally allowed regions , and found it to be within the reach of the future hyper - kamiokande and dusel experiments .
a wealth of experimental data is on the horizon , so it is imperative that phenomenologically appealing guts , for instance @xmath0-@xmath1 , be researched so that unambiguous experimental predictions may be presented .
these predictions will be key milestones in deducing the underlying theory at high energies as we progress through the next few exciting years of lhc , direct dark matter detection , fermi - lat , and proton decay experiments .
we have supplemented the conventional bottom - up analysis of traditional guts to include tev scale vector - like particles , and our results feature encouraging prospects for the experimental determination of whether high - energy theory indeed admits these proposed multiplets .
we believe that in the next few years experiment will certainly have something key to say about @xmath0-@xmath1 in particular and string theory in general .
this research was supported in part by the doe grant de - fg03 - 95-er-40917 , by the natural science foundation of china under grant no .
10821504 ( tl ) , and by the mitchell - heep chair in high energy physics .
h. hayashi , t. kawano , r. tatar and t. watari , nucl .
b * 823 * , 47 ( 2009 ) . c. m. chen and y. c. chung , nucl .
b * 824 * , 273 ( 2010 ) .
j. j. heckman , g. l. kane , j. shao and c. vafa , jhep * 0910 * , 039 ( 2009 ) .
s. cecotti , m. c. n. cheng , j. j. heckman and c. vafa , arxiv:0910.0477 [ hep - th ] .
t. li , d. v. nanopoulos and j. w. walker , arxiv:0910.0860 [ hep - ph ] .
t. li , d. v. nanopoulos and j. w. walker , arxiv:1003.2570 [ hep - ph ] .
j. marsano , n. saulina and s. schafer - nameki , arxiv:0912.0272 [ hep - th ] . g. k. leontaris and n. d. tracas , arxiv:0912.1557 [ hep - ph ] .
j. j. heckman , j. shao and c. vafa , arxiv:1001.4084 [ hep - ph ] .
i. antoniadis , j. r. ellis , j. s. hagelin and d. v. nanopoulos , phys .
b * 194 * , 231 ( 1987 ) .
j. jiang , t. li and d. v. nanopoulos , nucl .
b * 772 * , 49 ( 2007 ) .
j. l. lopez , d. v. nanopoulos and k. j. yuan , nucl .
b * 399 * , 654 ( 1993 ) .
y. j. huo , t. li and d. v. nanopoulos , arxiv:1011.0964 [ hep - ph ] .
j. r. ellis , j. l. lopez and d. v. nanopoulos , phys .
b * 371 * , 65 ( 1996 ) .
j. r. ellis , d. v. nanopoulos and j. walker , phys .
b * 550 * , 99 ( 2002 ) .
m. endo , s. shirai and k. yonekura , arxiv:0912.4484 [ hep - ph ] .
m. holmes and b. d. nelson , arxiv:0912.4507 [ hep - ph ] .
j. hisano , k. nakayama and m. yamanaka , phys .
b * 684 * , 246 ( 2010 ) [ arxiv:0912.4701 [ hep - ph ] ] .
i. gogoladze , r. khalid , s. raza and q. shafi , arxiv:0912.5411 [ hep - ph ] .
j. a. maxin , v. e. mayes and d. v. nanopoulos , arxiv:0912.5372 [ hep - ph ] .
x. gao , z. kang and t. li , arxiv:1001.3278 [ hep - ph ] .
a. h. chamseddine , r. l. arnowitt and p. nath , phys .
lett . * 49 * , 970 ( 1982 ) ; h. p. nilles , phys .
b * 115 * , 193 ( 1982 ) ; l. e. ibanez , phys .
b * 118 * , 73 ( 1982 ) ; r. barbieri , s. ferrara and c. a. savoy , phys .
b * 119 * , 343 ( 1982 ) ; h. p. nilles , m. srednicki and d. wyler , phys . lett .
b * 120 * , 346 ( 1983 ) ; j. r. ellis , d. v. nanopoulos and k. tamvakis , phys .
b * 121 * , 123 ( 1983 ) ; j. r. ellis , j. s. hagelin , d. v. nanopoulos and k. tamvakis , phys .
b * 125 * , 275 ( 1983 ) ; l. j. hall , j. d. lykken and s. weinberg , phys .
d * 27 * , 2359 ( 1983 ) .
a. djouadi , j. l. kneur and g. moultaka , comput .
commun . * 176 * , 426 ( 2007 ) [ arxiv : hep - ph/0211331 ] .
g. belanger , f. boudjema , a. pukhov and a. semenov , comput .
commun . * 180 * , 747 ( 2009 ) [ arxiv:0803.2360 [ hep - ph ] ] .
g. belanger , f. boudjema , a. pukhov and a. semenov , comput .
commun .
* 176 * , 367 ( 2007 ) [ arxiv : hep - ph/0607059 ] ; comput .
commun .
* 174 * , 577 ( 2006 ) [ arxiv : hep - ph/0405253 ] ; comput .
commun .
* 149 * , 103 ( 2002 ) [ arxiv : hep - ph/0112278 ] .
[ tevatron electroweak working group and cdf collaboration and d0 collab ] , arxiv:0903.2503 [ hep - ex ] .
d. n. spergel
_ et al . _
[ wmap collaboration ] , astrophys .
j. suppl .
* 148 * , 175 ( 2003 ) [ arxiv : astro - ph/0302209 ] ; astrophys .
j. suppl .
* 170 * , 377 ( 2007 ) [ arxiv : astro - ph/0603449 ] .
i. antoniadis , c. bachas , j. r. ellis and d. v. nanopoulos , phys .
b * 211 * , 393 ( 1988 ) ; nucl .
b * 328 * , 117 ( 1989 ) ; phys . lett .
b * 257 * , 278 ( 1991 ) ; j. r. ellis , n. e. mavromatos and d. v. nanopoulos , lectures given at international workshop on recent advances in the superworld , woodlands , tx , 13 - 16 apr 1993 , published in woodlands superworld 1993:3 - 26 ( qcd161:i966:1993 ) [ arxiv : hep - th/9311148 ] ; j. r. ellis , n. e. mavromatos and d. v. nanopoulos , mod . phys .
lett . a * 10 * , 1685 ( 1995 ) [ arxiv : hep - th/9503162 ] ; phys .
b * 619 * , 17 ( 2005 ) [ arxiv : hep - th/0412240 ] ; d. v. nanopoulos and d. xie , phys . rev .
d * 78 * , 044038 ( 2008 ) [ arxiv:0710.2312 [ hep - th ] ] .
a. b. lahanas , n. e. mavromatos and d. v. nanopoulos , pmc phys .
a * 1 * , 2 ( 2007 ) [ arxiv : hep - ph/0608153 ] ; phys .
b * 649 * , 83 ( 2007 ) [ arxiv : hep - ph/0612152 ] .
b. dutta , a. gurrola , t. kamon , a. krislock , a. b. lahanas , n. e. mavromatos and d. v. nanopoulos , phys .
d * 79 * , 055002 ( 2009 ) [ arxiv:0808.1372 [ hep - ph ] ] .
e. barberio _ et al .
_ [ heavy flavor averaging group ( hfag ) collaboration ] , [ arxiv:0704.3575 [ hep - ex ] ] .
m. misiak _ et al .
_ , phys .
* 98 * , 022002 ( 2007 ) [ arxiv : hep - ph/0609232 ] . g. w. bennett _ et al . _
[ muon g-2 collaboration ] , phys .
lett . * 92 * , 161802 ( 2004 ) [ arxiv : hep - ex/0401008 ] .
t. aaltonen _ et al . _
[ cdf collaboration ] , phys .
lett . * 100 * , 101802 ( 2008 ) [ arxiv:0712.1708 [ hep - ex ] ] .
r. barate _
[ lep working group for higgs boson searches and aleph collaboration ] , phys .
b * 565 * , 61 ( 2003 ) [ arxiv : hep - ex/0306033 ] ; w. m. yao _ et al _ ( particle data group ) , j. phys .
g * 33 * , 1 ( 2006 ) .
e. komatsu _ et al .
_ , arxiv:1001.4538 [ astro-ph.co ] . v. n. lebedenko _ et al .
_ , phys .
d * 80 * , 052010 ( 2009 ) [ arxiv:0812.1150 [ astro - ph ] ] .
j. angle _ et al . _
[ xenon collaboration ] , phys .
lett . * 100 * , 021303 ( 2008 ) [ arxiv:0706.0039 [ astro - ph ] ] .
z. ahmed _ et al . _
[ cdms collaboration ] , phys .
lett . * 102 * , 011301 ( 2009 ) [ arxiv:0802.3530 [ astro - ph ] ] .
p. sorenson , lux collaboration , direct detection of dark matter with lux , dnp october meeting 2009 .
a. morselli , a. lionetto , a. cesarini , f. fucito and p. ullio [ glast collaboration ] , nucl .
suppl . * 113 * , 213 ( 2002 ) [ arxiv : astro - ph/0211327 ]
. t. abdo _ et al .
_ [ lat collaboration ] , arxiv:1002.4415 [ astro-ph.co ] .
t. sjostrand , s. mrenna and p. skands , jhep * 0605 * , 026 ( 2006 ) [ arxiv : hep - ph/0603175 ] .
t. li , j. a. maxin , d. v. nanopoulos and j. w. walker , arxiv:1007.5100 [ hep - ph ] .
t. li , j. a. maxin , d. v. nanopoulos and j. w. walker , arxiv:1009.2981 [ hep - ph ] .
t. li , j. a. maxin , d. v. nanopoulos and j. w. walker , arxiv:1101.2197 [ hep - ph ]
. h. nishino _ et al . _
[ super - kamiokande collaboration ] , phys .
lett . * 102 * , 141801 ( 2009 ) [ arxiv:0903.0676 [ hep - ex ] ] .
s. raby _ et al .
_ , arxiv:0810.4551 [ hep - ph ] .
d. b. cline , private communication ; e. kearns , `` next generation proton decay '' ( 2008 ) , the dusel planning meeting at fermilab .
|
we study gravity mediated supersymmetry breaking in @xmath0-@xmath1 and its low - energy supersymmetric phenomenology .
the gaugino masses are not unified at the traditional grand unification scale , but we nonetheless have the same one - loop gaugino mass relation at the electroweak scale as minimal supergravity ( msugra ) .
we introduce parameters testable at the colliders to measure the small second loop deviation from the msugra gaugino mass relation at the electroweak scale . in the minimal @xmath1 model with gravity mediated supersymmetry breaking , we show that the deviations from the msugra gaugino mass relations are within 5% .
however , in @xmath0-@xmath1 , we predict the deviations from the msugra gaugino mass relations to be larger due to the presence of vector - like particles , which can be tested at the colliders .
we determine the viable parameter space that satisfies all the latest experimental constraints and find it is consistent with the cdms ii experiment .
further , we compute the cross - sections of neutralino annihilations into gamma - rays and compare to the first published fermi - lat measurement . finally , the corresponding range of proton lifetime predictions is calculated and found to be within reach of the future hyper - kamiokande and dusel experiments .
|
the run ii at tevatron will define a new level of precision for qcd studies in hadron collisions .
both collider experiments , cdf and d0 , expect to collect up to @xmath0 of data in this new run period .
the increase in instantaneous luminosity , center - of - mass energy ( from 1.8 tev to 1.96 tev ) and the improved acceptance of the detectors will allow stringent tests of the standard model ( sm ) predictions in extended regions of jet transverse momentum , @xmath1 , and jet rapidity , @xmath2 . in this contribution , a review of some of the most important qcd results from run ii is presented , together with future prospects as the integrated luminosity increases .
the measurement of the inclusive jet production cross section for central jets constitutes one of the pillars of the jet physics program at the tevatron .
it provides a stringent test of perturbative qcd predictions over almost nine orders of magnitude and probes distances up to @xmath3 .
thanks to the increase in the center - of - mass energy in run ii the jet production rate has been multiplied ( by a factor of five for jets with @xmath4 gev ) and the first measurements have already extended the @xmath5 coverage by 150 gev compared to run i. in addition , both cdf and d0 experiments explore new jet algorithms , following the theoretical work that indicates that the cone - based jet algorithm employed in run i is not infrared safe and compromises a future meaningful comparison with pqcd calculations at nnlo .
figure [ fig1 ] shows the measured inclusive jet cross section by cdf using the longitudinally invariant @xmath6 algorithm @xcite and based on the first 145 @xmath7 of run ii data .
measurements have been performed using values for the d parameter in the @xmath6 expression , @xmath8 equal to 0.5 , 0.7 and 1.0 .
the measurements are compared to pqcd nlo calculations @xcite using cteq6 @xcite parton density functions in the proton and antiproton and the renormalization and factorization scales set to @xmath9 .
the measured cross section is reasonably well described by the predictions for @xmath10 within the present uncertainties .
the systematic errors on the data are dominated by the uncertainty on the jet energy scale determination while the theoretical predictions suffer from our limited knowledge of the gluon distribution at high @xmath11 . at lower @xmath5 ,
the data is systematically above the predictions and the effect increases as d increases ( see figure [ fig2 ] ) .
this indicates the presence of soft - gluon contributions and fragmentation effects that have not been taken into account yet .
figure [ fig3 ] shows the measured inclusive jet cross section by d0 based on the first 143@xmath12 of run ii data .
the new midpoint@xcite jet algorithm has been used with a cone size r=0.7 .
this algorithm constitutes an improved version of the cone - based algorithm used in run i and it is shown to be infrared safe in fixed - order parton - level calculations .
the data is in good agreement with the pqcd nlo predictions using cteq6 parton density functions and @xmath13 .
however , the measurement is dominated by a relatively large uncertainty on the absolute jet energy scale .
+ figure [ fig4 ] shows the measured cross section by d0 as a function of the dijet invariant mass in dijet production of central jets .
this measurement is particularly sensitive to the presence of narrow resonances decaying into jets of hadrons up to masses of 1.3 tev .
the data is well described by pqcd nlo predictions .
nowadays , the tevatron high-@xmath1 jet data is used , together with prompt - photon data from fixed target experiments , to constrain the gluon distribution at high @xmath11 .
jet measurements at large rapidities are important because they constrain the gluon density in a region in @xmath5 where no effect from new physics is expected .
the d0 experiment has already extended the jet cross section measurements to the forward region for jets with @xmath14 ( see figure [ fig5 ] ) . at the moment , the results are affected by large systematic errors . in the near future
the experiments will highly reduce their uncertainties and precise cross section measurements will allow to further constrain the gluon distribution , and enhance their sensitivity to new physics at very high @xmath5 .
as mentioned in the previous section , the hadronic final states at 2 tev are characterized by the presence of soft underlying emissions , usually denoted as _ underlying event _ , in addition to highly energetic jets coming from the hard interaction .
the underlying event contains contributions from initial- and final - state soft gluon radiation , secondary semi - hard partonic interactions and interactions between the proton and anti - proton remnants that can not be described by perturbation theory .
these processes must be approximately modeled using monte carlo ( mc ) programs tuned to describe the data .
the jet energies measured in the detector contain an underlying event contribution that has to be taken into account in order to compare the measurements to pqcd predictions .
hence , a proper understanding of this underlying contribution is crucial to reach the desired precision in the measured cross sections . in the analysis presented here ,
the underlying event in dijet production has been studied by looking at regions well separated from the leading jets , where the underlying event contribution is expected to dominate the observed hadronic activity .
jets have been reconstructed using tracks with @xmath15 gev and @xmath16 and a cone algorithm with r=0.7 .
the @xmath17 space around the leading jet is divided in three regions : _ toward _ , _ away _ and _ transverse _ ( see figure [ fig6]-left ) , and the transverse region is assumed to reflect the underlying event contribution .
figure [ fig6]-right shows the average track density in the transverse region as a function of @xmath18 of the leading jet .
measurements are presented for the dijet inclusive sample and for events where the two leading jets are forced to be back - to - back in @xmath17 , in order to further reduce extra hard - gluon radiation .
the observed plateau indicates that the underlying event activity is , to a large extend , independent from the hard interaction .
the measurements have been compared to the predictions from pythia @xcite and herwig @xcite mc programs .
the pythia samples have been created using a special tuned set of parameters , denoted as pythia - tune a , which includes an enhanced contribution from initial - state soft gluon radiation and secondary parton interactions .
it was determined as a result of similar studies of the underlying event performed using cdf run i data @xcite .
pythia - tune a describes the hadronic activity in the transverse region while herwig underestimates the radiation at low @xmath18 .
similar measurements in z+jet(s ) events at high integrated luminosity will allow to explore the univesality of the underlying event contribution , in events with a very different color configuration in the final state .
the internal structure of jets is dominated by multi - gluon emissions from the primary final - state parton .
it is sensitive to the relative quark- and gluon - jet fraction and receives contributions from soft - gluon initial - state radiation and beam remnant - remnant interactions .
the study of jet shapes at the tevatron provides a stringent test of qcd predictions and tests the validity of the models for parton cascades and soft - gluon emissions in hadron - hadron collisions .
the cdf experiment has presented results on jet shapes for central jets with transverse momentum in the region
@xmath19 gev @xmath20 gev , where jets are searched for using the midpoint merging fraction has been used instead of the default @xmath21 . ]
algorithm and a cone size @xmath22 .
the integrated jet shape , @xmath23 , is defined as the average fraction of the jet transverse momentum that lies inside a cone of radius @xmath24 concentric to the jet cone : @xmath25 where @xmath26 denotes the number of jets .
the measured jet shapes have been compared to the predictions from pythia - tune a and herwig mc programs .
in addition , two different pythia samples have been used with default parameters and with and without the contribution from multiple parton interactions ( mpi ) between proton and antiproton remnants , the latter denoted as pythia-(no mpi ) , to illustrate the importance of a proper modeling of soft - gluon radiation in describing the measured jet shapes .
figure [ fig7](left ) presents the measured integrated jet shapes , @xmath27 , for jets with @xmath28 gev , compared to herwig , pythia - tune a , pythia and pythia-(no mpi ) predictions .
figure [ fig7](right ) shows , for a fixed radius @xmath29 , the average fraction of the jet transverse momentum outside @xmath30 , @xmath31 , as a function of @xmath32 .
the measurements show that the fraction of jet transverse momentum at a given fixed @xmath33 increases ( @xmath31 decreases ) with @xmath32 , indicating that the jets become narrower as @xmath32 increases .
pythia with default parameters produces jets systematically narrower than the data in the whole region in @xmath32 .
the contribution from secondary parton interactions between remnants to the predicted jet shapes ( shown by the difference between pythia and pythia-(no mpi ) predictions ) is relatively small and decreases as @xmath32 increases .
pythia - tune a predictions describe all of the data well .
herwig predictions describe the measured jet shapes well for @xmath34 gev but produces jets that are too narrow at lower @xmath32 .
figure [ fig8](left ) shows the measured integrated jet shapes , @xmath27 , for jets with @xmath28 gev , compared to pythia - tune a and the predictions for quark- and gluon - jets separately .
figure [ fig8](right ) shows the measured @xmath31 , @xmath29 , as a function of @xmath32 .
the mc predictions indicate that the measured jet shapes are dominated by contributions from gluon - initiated jets at low @xmath32 while contributions from quark - initiated jets become important at high @xmath32 .
this can be explained in terms of the different partonic contents in the proton and antiproton in the low- and high-@xmath32 regions , since the mixture of gluon- and quark - jet in the final state partially reflects the nature of the incoming partons that participate in the hard interaction . for a given type of parton - jet in the mc ( quark- or gluon - jet ) ,
the observed trend with @xmath32 shows the running of the strong coupling , @xmath35 .
+ jet shape measurements thus introduce strong constrains on phenomenological models describing soft - gluon radiation and the underlying event in hadron - hadron interactions . similar studies with b - tagged jets will be necessary to test our knowledge of the b - quark jet fragmentation processes in hadron interactions , which is essential for future precise top and higgs measurements at the tevatron and the lhc .
the d0 experiment has employed the dijet sample to study azimuthal decorrelations , @xmath36 , between the two leading jets . the normalized cross section :
@xmath37 is sensitive to the spectrum of the gluon radiation in the event .
the measurements have been performed in different regions of the leading jet @xmath5 starting at @xmath38 gev , and where the second jet is required to have at least @xmath39 gev .
figure [ fig9 ] shows the measured cross section compared to lo and nlo predictions from nlojet++ program @xcite .
the lo predictions , with at most three partons in the final state , is limited to @xmath40 , for which the three partons define a _ mercedes - star _ topology .
it presents a prominent peak at @xmath41 corresponding to the soft limit for which the third parton is collinear to the direction of the two leading partons .
the nlo predictions , with four partons in the final state , describes the measured @xmath36 distribution except at very high and very low values of @xmath36 where additional soft contributions , corresponding to a resummed calculation , are necessary . a reasonable approximation to such calculations
is provided by parton shower mc programs .
figure [ fig10 ] presents the measured cross section compared to pythia - tune a , pythia and herwig predictions in different regions of @xmath5 .
pythia with default parameters underestimates the gluon radiation at large angles .
pythia - tune a predictions , which include an enhanced contribution from initial - state soft gluon radiation and secondary parton interactions , describe the azimuthal distribution .
herwig also describes the data although it tends to produce less radiation than pythia - tune a close to the direction of the leading jets .
a detailed study of hard processes involving the associated production of a w boson and a given number of jets in the final state is a main goal of the cdf physics program in run ii .
these processes constitute the biggest background to top and higgs production in hadron colliders .
therefore , precise measurements of w+@xmath42 cross sections will be essential to test the lo and nlo qcd calculations used in order to estimate qcd - related backgrounds to top and higgs signals . during the last years a number of new boson+@xmath26 lo programs have become available @xcite which include larger jet multiplicities in the final state , in addition to nlo calculations for the w+dijet case .
these different programs are being interfaced to parton - shower models using different matching procedures to avoid double counting in the gluon radiation .
figure [ fig11 ] shows the measured inclusive cross section for @xmath43 production by cdf based on 127@xmath44 of run ii data .
jets with @xmath45 gev and @xmath46 have been considered , where jets are searched for using the run i cone algorithm with @xmath47 .
the measurements are compared to similar results from run i @xcite and pqcd lo predictions for @xmath48 as implemented in alpgen interfaced to the parton cascades and fragmentation from herwig .
the measured cross section in run ii is about @xmath49 larger than that in run i thanks to the new centre - of - mass energy .
the relative rates as a function of jet multiplicity are similar to those observed in run i which indicates that the @xmath50 cross section is reduced by @xmath51 per each jet required .
the pqcd lo predictions describe the data well but suffer from large uncertainties due to the strong dependence on the hard scale used in the calculation .
+ spectrum for the @xmath52 jet in @xmath43 production compared to pqcd lo predictions as implemented in alpgen interfaced to herwig .
[ fig12 ] _ , width=377 ] figure [ fig12 ] present the measured @xmath18 spectrum for the @xmath53 jet in @xmath43 production .
the spectrum for the least energetic jet is sensitive to the details of the interface between the pqcd lo calculation and the parton shower evolution .
the measured spectrum is in agreement with the predictions from alpgen+herwig within the present uncertainties . in the data
the systematic errors are dominated by the jet energy scale determination while the lo theoretical predictions present a strong dependence on the selected renormalization and factorization scales .
+ as the recorded luminosity of the experiments increases , precise measurements for @xmath54 production will be possible .
those measurements will be compared to pqcd nlo calculations with much reduced theoretical uncertainties .
the cdf experiment has presented measurements of cross sections for the production of high-@xmath56 photons in association with heavy quark jets , based on the first 66@xmath57 of run ii data .
those processes can signal the presence of new physics like , for example , stop production when they are produced together with large missing transverse energy in the detector .
in addition , these cross sections directly probe the heavy quark contents in the proton .
displaced secondary vertices , coming from the decay of long - lived charm and bottom hadrons and reconstructed in the cdf silicon tracker , are used in order to identified jets coming from heavy quarks .
the mass of the observed secondary vertex is employed to separate the different contributions from charm , bottom and light quarks ( the latter producing fake tags ) to the sample of heavy - quark jet candidates .
monte carlo templates are fitted to the measured secondary vertex mass distribution in the data ( see figure [ fig13 ] ) to extract the different contributions and use them to compute the cross sections .
figure [ fig14 ] shows the measured cross sections for @xmath58 and @xmath59 production compared to lo qcd predictions as implemented in pythia .
the measurements are in agreement with the predicted cross section within the still limited statistics in the data .
i thank pic04 organizers for their invitation , help and kidness during the conference in boston and for the exciting program of talks and discussions they made possible .
99 stephen d. ellis , davision e. soper , phys.rev .
d48 ( 1993 ) 3160 - 3166 .
giele , e.w.n .
glover , d. a. kosower , nucl .
b403 ( 1993 ) 633 - 670 .
j. pumplin et al . , jhep 0207 ( 2002 ) 012 .
g. c. blazey , et al . , hep - ex/0005012 .
ellis , j. huston and m. toennesmann , hep - ph/0111434 .
s. d. ellis , z. kunszt , d. e. soper , phys.rev.lett .
69 ( 1992 ) 3615 - 3618 .
bengtsson and t. sjstrand , comp .
46 ( 1987 ) 43
. g. marchesini et al . , comp .
67 ( 1992 ) 465 .
d. acosta et al . , cdf collaboration , phys .
d65 ( 2002 ) 092002 .
zoltn nagy , zoltn trcsnyi , phys.rev.lett .
79 ( 1997 ) 3604 - 3607 .
m. mangano et al . : ` alpgen ' , preprint hep - ph/0206293 .
+ e. boss et al . : ` comphep ' , preprint hep - ph/9503280 .
+ s. tsuno et al . : ` gr@appa ' , preprint hep - ph/0204222 .
+ f. maltoni and t. stelzer : ` madgraph ' , preprint hep - ph/0208156 .
+ k. ellis and j. campbell : ` mcfm ' , preprint hep - ph/0202176 .
t. affolder et al . ,
the cdf collaboration , phys .
d63 ( 2001 ) 072003 .
|
in this contribution , a review of some of the most important qcd jet preliminary results from both cdf and d0 experiments in run ii is presented , together with future prospects as the integrated luminosity increases .
= 14.5pt
|
the obstructed transport in disordered and inhomogeneous systems may lead to anomalous diffusion .
transport of a dense fluid in a porous host structure , ballistic tracer motion in a spatially heterogeneous medium , a probe or solute diffusing through fixed scatterers within liquid crystalline phases and protein motion in cellular membranes or in cytoplasm , all exhibit characteristics typical for hindered diffusion @xcite .
models of such systems usually include variants of lorentz approach @xcite in which static scatterers are fixed in space and tracer particles are allowed to move freely through the resulting mesh .
the obstacle density is crucial for the anisotropy of motion and excluded volume effects , since only unoccupied , void volume is accessible to the probe particle traveling in the system . in the original lorentz model proposed as a description of the electric conductivity of metals , the motion of a ballistic particle elastically scattered off
randomly distributed spherical obstacles has been analyzed .
microscopic theory predicts for this model appearance of a localization transition @xcite where at a critical obstacle density the long - range transport ceases . in accordance , as the critical density of scatterers is approached the tracer particle experiences a sublinear in time increase of its mean - square displacement and the diffusion constant vanishes at a critical concentration of obstacles .
such a signature of anomalous diffusive motion is a common observation in experimental studies of various transport processes in cellular fluids @xcite .
dependent on density and shapes of crowding agents , meandering subdiffusive motion occurs on larger time windows and its characteristics close to the localization transition exhibit scaling behaviour which can be rationalized in terms of dynamic critical phenomena @xcite .
effects of the entrapment within the mesh composed of the obstacles and elucidation of dynamical behavior of spherical tracers diffusing through the collective structure of scatterers have been investigated so far both experimentally @xcite and theoretically @xcite . in particular , diffusion of silica tracer beads with large radii through colloidal rods have been observed experimentally to depend on the rod density , exhibiting domination of caging effects at high densities of rods .
similar experiments with apoferritin , quantum dots and lysozyme @xcite diffusing in poly(ethylene glycol ) solutions have been analyzed addressing a connection between the motion of nanoscopic probes in polymer solutions and their macroviscosity .
diffusion of spherical tracers through a liquid crystal consisting of ellipsoidal particles near the isotropic - nematic phase transitions has been investigated theoretically @xcite pointing to the increasing anisotropy of the probe dynamics with increasing long - range order of nematogens . in silico
, studies performed by hfling et al.@xcite analyzed dynamic scaling picture of anomalous transport in two - dimensional models , in which a brownian tracers of different sizes were moving between uncorrelated and overlapping circular obstacles .
these studies confirmed the universal values of the dynamic exponents of the mean - square displacement ( msd ) at leading and next - to - leading order and the subdiffusive motion of tracers was shown to experience the crossover to normal diffusion away from the localization transition , when tracers become immobilized .
typically , anomalous and complex transport is conveniently described by means of msd which , by means of the central limit theorem , is expected to grow linearly in time for time scales much larger than microscopic .
on the other hand , observed non - linear growth of msd is a benchmark of unusual , anomalous transport indicating subdiffusive behavior for msd scaling according to a power law , @xmath0 , with an exponent @xmath1 .
apart from this most common characteristics of subdiffusive motion , other phenomena like strongly reduced and time - dependent diffusion coefficients , persistent correlations in time , non - gaussian distributions of spatial displacements and heterogeneous diffusion of particles can be also observed @xcite . in order to study the effects of varying geometric and spatial anisotropy in the structure of scatterers , in this article we investigate the dynamics of a single tracer particle performing brownian motion in a two - dimensional mesh of
randomly distributed elongated hard obstacles . at a certain critical obstacle density ,
the motion of the tracer becomes anomalous over many decades in time , which is rationalized in terms of an underlying percolation transition of the void space .
we analyze the scaling behavior of the time - dependent msd and describe various facets of entrapped motion related to the tracer size and organization of the mesh .
spherocylindrical objects forming scatterers are assumed to have a form of a fibrinogen molecule ( cf .
fig.[fig : fibrynogenmodel ] ) approximated by a linear chain of stiff beads .
the side beads have diameter @xmath2 nm and are connected with a central , middle - size bead of diameter @xmath3 nm by chains of ten smaller @xmath4 nm balls .
nm each , the middle bead has the diameter @xmath3 nm and small spheres forming the chains are of the size @xmath4 nm . the total length of the model fibrinogen molecule is @xmath5 nm . ] presented model had successfully explained the main properties of fibrinogen layers created during irreversible adsorption , like the saturated random coverage ratio and kinetics of the process @xcite .
the layers of different concentrations of fibrinogen are generated numerically by using the random sequential adsorption ( rsa ) algorithm @xcite .
the procedure is based on a stochastic adsorption process in which objects are placed consecutively on a surface in such a way that they do not overlap any previously adsorbed particles and adsorb only after making a contact with an uncovered surface area . as a consequence of prohibited penetration of objects ,
the adsorption is completed when there is no available ( free ) area .
the pattern of coverage of the surface obtained in such a way reflects the interplay between entropic packing of molecules and their concentration and has been displayed in fig.[fig : examples ] .
accordingly , the concentration of the fibrinogen molecules @xmath6 depends on the orientation of the molecules , i.e. @xmath7 where @xmath8 is the orientational parameter described below .
the single simulation run lasts for @xmath9 steps , where @xmath10 nm@xmath11 is an area covered by a single fibrinogen molecule and @xmath12 nm@xmath11 is a circular surface of diameter @xmath13 mn . the saturated random coverage of the surface is determined by the dimensionless density @xmath14 where @xmath15 represents the number of adsorbed molecules .
the maximal coverage grows weakly @xcite with increasing orientational order @xmath8 ( the overall impact is only about 10% ) and resembles isotropic - nematic transition in a liquid crystal , where typically no significant density change is observed . and @xmath16 . the mean direction @xmath17 in all cases .
the orientational order parameters derived from relation ( [ order ] ) are @xmath18 and @xmath19 respectively . ]
orientational order of a fibrinogen layer is controlled here by the parameter @xmath20 - a width of a gaussian probability density function , according to which the molecule s orientations are drawn . however , in order to compare results with other works and experiments it is worthy to introduce yet another order measure , independent of the mechanism that forces molecules to align in parallel .
for that purpose we define a global order parameter dependent only on geometrical properties of the coverage @xcite : @xmath21 . \label{order}\ ] ] here @xmath15 is a number of molecules in a layer , @xmath22 $ ] describes a unit vector parallel to the orientation of @xmath23-th fibrinogen and @xmath24 denotes the order s direction .
parameter @xmath8 is @xmath25 for totally disordered phase and rises with growing orientational order up to @xmath26 . in consequence
, the maximum value of @xmath8 is reached when all molecules are parallel .
the main properties of ordered fibrinogen coverages were discussed in @xcite .
here we focus on diffusion processes in such layers .
therefore , an appropriate model of random walk process has to be introduced .
diffusion inside fibrinogen layer has been studied using gaussian random walk ( grw ) process .
the process starts at a randomly chosen , uncovered point .
single step is given by a random vector @xmath27 $ ] with coordinates drawn ( as two independent random variables ) according to a normal gaussian @xmath28 probability distribution .
the step is accepted if it follows to an uncovered place .
otherwise , it is treated as the @xmath29 $ ] step .
the most typical property of such a random walk process is the mean square displacement ( msd ) of the walker @xmath30 and its characteristic exponent @xmath31 which reflects the dependence of the asymptotic diffusion process on time @xmath32 : @xmath33 here @xmath34 is a diffusion coefficient and @xmath31 describes the character of the motion : for a standard ( normal ) diffusion msd is expected to grow linearly in time , i.e. @xmath35 .
in contrast , non - linear growth of the msd is usually taken as an indicator of anomalous diffusive transport with slower than for the normal diffusion increase of msd in time ( @xmath36 ) typical for the subdiffusion and the enhanced superballistic ( superdiffusive ) transport corresponding to @xmath37 .
this classification requires however some additional care , when diffusive motion is defined as the asymptotic property of random walks , as e.g. charge transport in disordered semiconductors or diffusion in chemical space for random walks over polymer chains . in such realms
, the combination of long steps and long breaks between them can be responsible for scaling properties of moments which are in line with ordinary brownian motion msd ( @xmath35 ) despite the asymptotic process is significantly non - markov and different from gaussian @xcite .
characteristics of diffusive motion , namely coefficients @xmath34 and @xmath31 , can be obtained directly from numerical simulations by fitting the power law ( [ diffusion ] ) to measured dependence of @xmath38 on @xmath32 .
clearly , in such case the coefficient @xmath34 carries different units for different exponents @xmath31s , @xmath39=[m^2/s^{\alpha}]$ ] .
otherwise , the time - scale dependent diffusion coefficient can be introduced @xmath40 and , by analogy , the apparent exponent @xmath31 can be defined as a logarithmic derivative of the msd : @xmath41 here @xmath42 and @xmath43 are arbitrary units of length and time , respectively .
note that @xmath31 obtained from ( [ diffusion ] ) is the mean of above @xmath44 over a finite time window used for fitting purposes . in our simulations
@xmath43 represents a single step of the grw trajectory with a typical jump length @xmath45 ( see below ) . in order to check whether obtained results depend on a particular simulation model of the diffusion process
, two additional models of diffusive transport have been analyzed . in the first one the step size of a tracer
is not drawn from the gaussian pdf but instead , is preset to a constant @xmath46 nm with the director of the motion sampled from a uniform distribution . the second model , similarly to case ( a ) uses the gaussian pdf of steps , however with twice bigger standard deviation ( instead of @xmath47 nm , the value @xmath48 nm has been used ) .
the total msd @xmath49 and its dependence on time for all three models are displayed in fig.[fig:3models ] .
dependence on time for three different models of a tracer diffusion : ( a ) - the original model , ( b ) - a constant step size model , ( c ) - a `` twice standard deviation '' model .
the power law fits for msd in the log - log scale are : ( a ) - @xmath50 , ( b ) - @xmath51 and ( c ) -@xmath52 .
the tracer diameter has been set to @xmath53 nm and the order parameter @xmath54 . ]
derived results indicate that subdiffusion inside ordered fibrinogen layers ( @xmath54 ) is not essentially model - dependent .
although diffusion coefficient @xmath34 differs , the slopes of ( a ) and ( b ) lines corresponding to subdiffusive @xmath31 exponents do not differ significantly for both models .
slightly larger values of @xmath34 and @xmath31 for ( c ) line , when a step size of the grw was ( on average ) twice larger than in cases ( a ) and ( b ) can be explained by nonzero probability of jumping over the obstacles , which effectively speeds up the diffusion process .
for grw trajectories of length equal to @xmath55 subsequent steps , the msd has been estimated as an average over @xmath56 independent realizations of the walk starting from different , randomly chosen points distributed in the void space of surface with saturated random coverage of fibrinogen molecules ( cf . section ii a ) .
we have observed that derived values of @xmath34 and @xmath31 differ significantly for various fibrinogen layers generated for the same value of @xmath20 .
therefore , those parameters have been further averaged over ensembles of @xmath57 up to @xmath58 different fibrinogen layers .
fig.[fig : example_diffusion ] displays exemplary diffusive trajectories of small , spherical tracer molecules wandering inside the accessible space of the fibrinogen mesh and starting their motion from randomly chosen location . steps .
the orientational order measured for this layer is @xmath59 . ] to relate character of the diffusion process inside fibrinogen layers to the degree of order , we have first concentrated on evaluation of the effective exponent @xmath31 in relation ( [ diffusion ] ) .
it s dependence on orientational order parameter @xmath8 is displayed in fig.[fig : diffusion1 ] .
on orientational order parameter @xmath8 for diffusive motion of tracers of different dimension .
left ( right ) panel : a tracer is modeled as a spherical molecule of diameter @xmath53 nm and @xmath2 nm , accordingly .
analysis has been performed on trajectories of @xmath55 steps of `` time '' duration .
insets present index @xmath31 analyzed separately for parallel and perpendicular components of the diffusive motion .
dots represent simulation data , lines have been fitted to guide the eye.,title="fig : " ] on orientational order parameter @xmath8 for diffusive motion of tracers of different dimension .
left ( right ) panel : a tracer is modeled as a spherical molecule of diameter @xmath53 nm and @xmath2 nm , accordingly . analysis has been performed on trajectories of @xmath55 steps of `` time '' duration .
insets present index @xmath31 analyzed separately for parallel and perpendicular components of the diffusive motion .
dots represent simulation data , lines have been fitted to guide the eye.,title="fig : " ] in the short time domain subdiffusion is observed in the whole range of the orientational order parameter . moreover , the slowdown of the diffusion process ( as expressed by the value of the effective exponent @xmath31 ) is stronger in more ordered phases , cf .
fig.[fig : diffusion1 ] .
this observation seems to be counterintuitive , as the motion in ideally ordered two dimensional space often reduces to a one dimensional problem ( e.g. when formation of channels favors unrestricted motion in a preferential direction ) where standard diffusion should be observed .
however , in the case analyzed here only nematic - like order is observed and additionally , ordering favors a denser fibrinogen packing @xcite , which results in more obstacles encountered by a diffusing molecule . to analyze this phenomenon more carefully we treated diffusion as being composed of two separate grw : parallel and perpendicular to the direction of the global orientational order : @xmath60 results obtained for those processes are presented in insets to fig.[fig : diffusion1 ] .
the parallel diffusion speeds up with increasing order ( @xmath61 ) , in contrast to the perpendicular transport which becomes more obstructed and slows down .
both motions converge to the same exponent value @xmath62 at @xmath63 , as in disordered phase there is no intrinsic preferred direction and the diffusion homogenizes .
the effective exponent @xmath31 depends on a tracer s size and the overall diffusive motion is slower for larger molecules , as expected ( see figs.[fig : diffusion1],[fig : diffusion22 ] ) . for small tracers ( @xmath64 ) , parallel motion approaches normal diffusion ( @xmath65 ) and this behaviour is stronger in ordered layers , see fig.[fig : diffusion22 ] .
in contrast , low ordering implies higher degree of structural homogeneity of layers and results in almost equal values @xmath66 . on size of the tracer analyzed for three different orientational order parameters : @xmath54 , @xmath67 and @xmath68 .
analysis has been performed on trajectories of @xmath55 steps of `` time '' duration .
insets display @xmath31 characterizing parallel and perpendicular diffusion in function of the mean global orientational order direction .
dots represent simulation data , full lines have been drawn to guide the eye.,title="fig : " ] on size of the tracer analyzed for three different orientational order parameters : @xmath54 , @xmath67 and @xmath68 .
analysis has been performed on trajectories of @xmath55 steps of `` time '' duration .
insets display @xmath31 characterizing parallel and perpendicular diffusion in function of the mean global orientational order direction .
dots represent simulation data , full lines have been drawn to guide the eye.,title="fig : " ] on size of the tracer analyzed for three different orientational order parameters : @xmath54 , @xmath67 and @xmath68 .
analysis has been performed on trajectories of @xmath55 steps of `` time '' duration .
insets display @xmath31 characterizing parallel and perpendicular diffusion in function of the mean global orientational order direction .
dots represent simulation data , full lines have been drawn to guide the eye . ] to get a deeper insight into the global diffusion process , longer trajectories of tracer movement have been analyzed .
additionally , structures of fibrinogen layers have been formed out of a fixed , constant concentration of molecules .
samples of surfaces created in this way are characterized by various spatial configurations of the same number of molecules distributed over the target area . in simulations
the maximum length of a single grw trajectory has been preset to include @xmath69 steps and the msd has been evaluated as an average over @xmath55 independent grw trajectories starting from randomly chosen , uncovered points .
the dependence of @xmath38 on number of steps for two different sizes of a spherical tracer is presented in fig.[fig : msd_time ] .
nm ( top ) and @xmath2 nm ( bottom ) , respectively .
black lines indicate the total msd , whereas red and green ones correspond to perpendicular and parallel diffusion , see eq.([xperppar ] ) .
simulations have been performed for a constant concentration of fibrinogen molecules on a sample surface , c=2140@xmath70m@xmath71 .
, title="fig : " ] nm ( top ) and @xmath2 nm ( bottom ) , respectively .
black lines indicate the total msd , whereas red and green ones correspond to perpendicular and parallel diffusion , see eq.([xperppar ] ) .
simulations have been performed for a constant concentration of fibrinogen molecules on a sample surface , c=2140@xmath70m@xmath71 .
, title="fig : " ] nm ( top ) and @xmath2 nm ( bottom ) , respectively .
black lines indicate the total msd , whereas red and green ones correspond to perpendicular and parallel diffusion , see eq.([xperppar ] ) .
simulations have been performed for a constant concentration of fibrinogen molecules on a sample surface , c=2140@xmath70m@xmath71 .
, title="fig : " ] nm ( top ) and @xmath2 nm ( bottom ) , respectively .
black lines indicate the total msd , whereas red and green ones correspond to perpendicular and parallel diffusion , see eq.([xperppar ] ) .
simulations have been performed for a constant concentration of fibrinogen molecules on a sample surface , c=2140@xmath70m@xmath71 .
, title="fig : " ] unlike for small tracer molecules , diffusive transport of large particles can be seen only in relatively short time intervals .
the motion is then characterized by a sublinear msd behavior ( see fig .
[ fig : msd_time ] ) and gradually becomes effectively confined within the network of fibrinogen molecules adsorbed on the surface . at longer times ( @xmath72 ) msd saturates pointing to the immobilization of particles within the mesh .
consequently , in order to avoid the entrapment , further analysis was performed using only small tracers .
their grw motion has been analyzed by inspecting trajectories of duration @xmath73 .
those longer trajectories exhibit sometimes the slowdown of the diffusion process related to a finite size of the layers .
such cases , whenever observed in simulations , have been excluded from statistical analysis .
the parallel and perpendicular msd , defined by eq.([xperppar ] ) are presented in fig.[fig : perppar_time ] . .
insets show the same data but in linear scale .
the tracer diameter was @xmath53 nm and the concentration of fibrinogen molecules c= 2140@xmath70m@xmath71.,title="fig : " ] .
insets show the same data but in linear scale .
the tracer diameter was @xmath53 nm and the concentration of fibrinogen molecules c= 2140@xmath70m@xmath71.,title="fig : " ] inspection of fig.[fig : perppar_time ] indicates that in ordered layers of fibrinogens the parallel component of msd is promoted , whereas the perpendicular motion becomes hindered , in line with observations drawn from the upper panel of fig.[fig : msd_time ] . in order to further elucidate effects due to varying spatial anisotropy and geometry of the fibrinogen mesh
, we focused our analysis on the effective exponent @xmath31 and coefficient @xmath34 in relation ( [ diffusion ] ) , ( [ altd ] ) and ( [ alta ] ) .
it s dependence on time and orientational order parameter is summarized in fig.[fig : diffusion2 ] .
derived from the fit to eq.(2 ) as a function of orientational order @xmath8 inside a layer .
the inset presents time - scale dependent diffusion coefficient @xmath74 evaluated from eq.([altd ] ) .
parameters @xmath31 , @xmath34 were averaged over @xmath75 $ ] time window .
+ right panel : the temporary value ( cf .
eq.([alta ] ) ) of the exponent @xmath44 is presented for the order parameter @xmath54 .
the inset displays the time - scale dependent diffusion coefficient @xmath74 evaluated for the same time window . in both cases the tracer diameter was @xmath53 nm and the concentration of obstacles @xmath76 .,title="fig : " ] derived from the fit to eq.(2 ) as a function of orientational order @xmath8 inside a layer .
the inset presents time - scale dependent diffusion coefficient @xmath74 evaluated from eq.([altd ] ) .
parameters @xmath31 , @xmath34 were averaged over @xmath75 $ ] time window .
+ right panel : the temporary value ( cf .
eq.([alta ] ) ) of the exponent @xmath44 is presented for the order parameter @xmath54 .
the inset displays the time - scale dependent diffusion coefficient @xmath74 evaluated for the same time window . in both cases the tracer diameter was @xmath53 nm and the concentration of obstacles @xmath76 .,title="fig : " ] similarly to the case of transport observed in the short time domain ( see section [ short ] ) , subdiffusion is observed in the whole range of the orientational order parameter and the exponent @xmath31 is highest for parallel msd component in almost all range of @xmath77 . on the other hand ,
the diffusion coefficient is significantly lower than for the whole msd , which reflects variability of @xmath34 and @xmath31 parameters in perpendicular and parallel directions of motion : @xmath78 and @xmath79 .
moreover , for ordered phases ( @xmath54 ) the perpendicular exponent @xmath80 has a broad minimum for intermediate times ( see fig.9 , right panel ) which corroborates with observation of the plateau in the perpendicular component of msd ( see fig.8 , left panel ) .
this minimum corresponds to the confinement of a tracer between long , parallel filaments made of fibrinogen molecules .
similar effect has been also detected in models of the anisotropic diffusion of a strongly hindered rod represented by a thin needle moving in an array of hard point obstacles @xcite .
the model we have to do with is essentially a continuous percolation model @xcite , but not a standard one . to see this ,
let us consider on the motion of a single tracer of radius @xmath81 and concentrate on the motion of its center .
since both the tracers and the fibrinogen molecules are modeled as solid , impenetrable objects , the motion of center of the tracer is geometrically restricted to the domain of the plane outside the @xmath82-vicinities of all fibrinogen molecules .
although the areas of the molecules themselves can not intersect , their @xmath82-vicinities can , thus leading to creation of effectively impenetrable domains , see fig.[fig : effectivelayer ] .
( left ) , @xmath4 nm ( middle ) and @xmath2 nm ( right ) . the fibrinogen concentration is c=@xmath83 and @xmath84 . ]
the percolation threshold is defined by a critical concentration of connected accessible points at which the unbounded motion through the lattice becomes possible . in our system
the percolation threshold depends however in a complex manner on three control parameters : on the orientational parameter @xmath8 , on the concentration of fibrinogen molecules @xmath6 which itself depends on @xmath8 via our generation algorithm , and on the tracer s diameter @xmath85 .
figs.[fig : msd_time ] , [ fig : alpha_q_conc1 ] show essentially the behavior of the msd in the percolating and non - percolating cases . in the first case
the msds continuously grow , in the second case they saturate at the value corresponding to the size of typical finite clusters ( accessible areas ) .
the transition here takes place due to both : increasing size of the tracer @xmath85 ( fig.[fig : msd_time ] ) and increasing concentration of obstacles ( fig.[fig : alpha_q_conc1 ] ) for ordered and disordered situations . + + on fibrinogen concentration ( left ) and tracer size ( right ) for three different orientational orders in a layer .
the tracer diameter in the left plot is @xmath2 nm , and the concentration in the right plot is @xmath86 .
evaluation of the exponent @xmath31 has been performed on trajectories of duration of @xmath87.,title="fig : " ] on fibrinogen concentration ( left ) and tracer size ( right ) for three different orientational orders in a layer .
the tracer diameter in the left plot is @xmath2 nm , and the concentration in the right plot is @xmath86 .
evaluation of the exponent @xmath31 has been performed on trajectories of duration of @xmath87.,title="fig : " ] in the percolating regime the system homogenizes , and the diffusion in it is normal , corresponding to the long - time value of @xmath88 , cf .
upper panel of fig.[fig : msd_time ] and fig.[fig : perppar_time ] .
however , this normal diffusion can be strongly anisotropic , as demonstrated in figs.[fig : msd_time ] , [ fig : perppar_time ] , [ fig : alpha_q_conc ] .
the exact calculation of anisotropy is a hard task , and the results are known only either for small concentrations or for very simple geometric forms of inclusions ( ellipses or ellipsoids ) .
nevertheless , several general remarks can be made based on what is known for such simple cases @xcite . if the obstacles are not oriented , no anisotropy can be observed .
if they are oriented , the final diffusion coefficient is anisotropic . as a measure of anisotropy the quotient of the values of diffusivity ( or , otherwise of the msd components ) parallel and perpendicular to the orientation of obstacles
@xcite can be taken : @xmath89 , which changes in the interval @xmath90 . ) on fibrinogen concentration and orientational order in a layer .
the tracer diameter is @xmath53 nm and @xmath2 nm for left / right panel , respectively.,title="fig : " ] ) on fibrinogen concentration and orientational order in a layer .
the tracer diameter is @xmath53 nm and @xmath2 nm for left / right panel , respectively.,title="fig : " ] the problem of finding the anisotropy of the diffusion coefficient in a system with obstacles is similar to the problem of finding the heat conductivity or the electric conductivity in similar settings .
such problems were investigated for several different geometries of inclusions . for small concentration of obstacles the diffusivity @xmath34 of the system
is well - described by the perturbative approaches @xcite , and its anisotropy vanishes ( which corresponds to @xmath91 ) , as also seen in fig.[fig : alpha_q_conc ] . for the case of intersecting obstacles with centers randomly placed on a plain the anisotropy close to the percolation threshold can grow quite large .
thus , close to percolation transition the anisotropy in a system of highly oriented elongated defects might be approximated by those of a system with ellipsoidal inclusions of the same aspect ratio @xmath92 with @xmath93 being the largest and @xmath94 the smallest dimension of the inclusion , a problem discussed in @xcite .
the discussion in this work is essentially for 3d , but can be easily adapted to a two - dimensional case . in this case close to the criticality @xmath95 , and by taking geometrical parameters of the fibrinogen model , may get to be as large as @xmath96 in our case . note that the possibility of intersection and the random position of the centers of inclusions are both crucial for the discussion , and the departures from these assumptions may lead to strong deviations from this prediction in what concerns the anisotropy s concentration dependence .
thus , in the case of parallel non - intersecting inclusions with large aspect ratio ( " scratches ) of length @xmath93 thrown onto a plane with concentration @xmath6 one gets @xmath97 , see @xcite .
our situation here is more complex than the both ones considered theoretically due to the fact that the domains are not fully oriented , can intersect but possess hard cores , so that their centers are not randomly distributed .
this leads to highly non - monotonous concentration dependencies for the highly oriented case , as seen in the lower panel of fig.[fig : alpha_q_conc ] . the overall accessible volume of the mesh close to the percolation threshold forms a fractal - like percolation cluster . the effective exponent @xmath31 is connected with a critical exponent of conductivity in percolation theory .
the last one is either determined by the lattice percolation exponent @xmath70 or by the distribution of the resistances of the bottleneck ( narrow paths between the almost contacting obstacles ) @xcite . despite much more complicated geometry of our model , the bottlenecks in it appear close to the contact points of essentially circular areas , and
therefore should show the same resistivity singularity behavior as the swiss cheese model @xcite which has the same conductivity exponent as the lattice percolation .
this one is translated into the standard value of the anomalous effective diffusion exponent being @xmath98 as defined through the known values of spectral and fractal dimensions of the percolation clusters @xcite .
this value of the final exponent of anomalous diffusion has to be the same for both principal values of the displacement ( i.e. in the directions parallel and perpendicular to the obstacles orientation ) , and is observed to happen for orientational order @xmath99 in our case ( cf .
fig.[fig : perppar_time],[fig : diffusion2 ] ) .
the behavior of each of the components for the values of parameters corresponding to the percolating phase is also quite typical : the normal short - time diffusion with the seed diffusivity coefficient ( corresponding to practically non - constrained particle motion ) , a crossover to a more or less pronounced intermediate asymptotic subdiffusive behavior typical for percolation , and another crossover back to normal diffusion with the effective diffusion coefficient considerably smaller than the seed one ( cf .
left panels of figs.7,8 ) . in the non - percolating domain
the last crossover is the one to a saturating msd .
however , fig.[fig : perppar_time ] also shows that the positions of the crossover regions and the widths of the crossover domains for the two components are quite different .
for example , for @xmath100 the perpendicular component already underwent both crossovers at @xmath101 , while the parallel component has just entered the first crossover domain at the same time .
summing up two functions with so strongly different behaviors to obtain the total diffusivity leads to the effective behavior which is seemingly far from typical percolation predictions .
thus , in the figs.[fig : diffusion1 ] , [ fig : diffusion22 ] the exponents @xmath31 of single components of msd are either larger than or equal to ( within the error bars ) the percolation value of @xmath102 , as it has to be the case for the typical double crossover situation .
the exponent characterizing the total displacement can however fall below the value of 0.55 , thus absolutely obscuring the percolative nature of the behavior observed .
in this paper properties of diffusive motion of a tracer particle in crowded , anisotropic environment has been analyzed . in general ,
transport slowdown with the characteristic effective exponent @xmath103 has been observed .
however , details of the process depend in a complex way on three control parameters : concentration of obstacles , their orientational ( nematic - like ) ordering , as well as on the size of a tracer
. moreover , due to environment anisotropy , the perpendicular and parallel motion to the intrinsic direction represented by the director orientation , cf . eq.([order ] ) has to be studied separately , as both components of the motion can exhibit a totally different behaviour .
although our situation is definitely a variant of a continuous percolation model , it shows a very rich and interesting behavior due to its novel aspect , namely the anisotropy
. the characteristic crossover times for the components of motion parallel and perpendicular to the main axes can be very different , so that the total diffusion coefficient can show effective time dependence which does not resemble the one for lattice percolation , and can even exhibit much slower effective subdiffusive behavior on intermediate times .
analysis of the effective exponent @xmath31 averaged over the time window @xmath75 $ ] shows that subdiffusive transport of small tracers is observed in the whole range of the orientational order parameter and the expected crossover to percolation limit is detected for @xmath104 , see fig.9 . for high orientational order , closer inspection of time dependent @xmath44 allows to detect another crossover @xmath105 for very long times @xmath106 .
yet , this `` almost normal '' diffusion remains strongly anisotropic , ( @xmath107 ) as demonstrated in the right panel of fig.9 and left panel of fig.12 .
in general , positions of crossover regions and duration of the crossover domains are different for two orientational components of motion and can be rationalized in terms of transport on an anisotropic percolating space .
we expect our findings to be relevant in understanding sources of anomalous transport in crowded , heterogeneous environments
. active research in this field is beneficial for control of anomalous subdiffusion in biological media , calibration of measurements by use of fluorescence correlation spectroscopy ( fcs ) and analysis of single particle tracking within the dense environment of the cell @xcite .
it is expected also to contribute to further progress in a general theory @xcite of the anisotropic motion of objects in entangled suspensions responsible for steric effects .
this work was supported in part by european science foundation epsd grant .
francesc sagus acknowledges financial support from micinn ( fis2010 - 21924-c02 - 01 ) and from dursi generalitat de catalunya ( 2009sgr 1055 ) .
the authors acknowledge many fruitful discussions with jakub barbasz .
99 f. hfling , t. franosch , rep .
. phys . * 76 * 046602 ( 2013 ) .
saxton , biophys .
j. * 66 * 394 ( 1994 ) .
i. bronstein , y. israel , e. kepten , s. mai , y. shav - tal , e. barkai , and y. garini , phys .
lett . * 103 * 018102 ( 2009 ) .
e. barkai , y. garini , and r. metzler , phys . today * 65 * ( 8) , 29 ( 2012 )
sokolov , soft matter , * 8 * 9043 ( 2012 ) .
w. gtze , _ complex dynamics of glass - forming liquids : a mode - coupling theory _ , oxford university press ( 2009 ) .
ellis , a.p .
minton , nature * 425 * 27 ( 2003 ) .
n. ziebacz , s.a .
wieczorek , t. kalwarczyk , m. fialkowski , r. holyst , soft matter * 7 * ( 16 ) 7181 ( 2011 ) .
k. kang , a. wilk , j. buitenhuis , a. patkowski , j.k.g .
dhont , j. chem . phys . * 124 * 044907 ( 2006 ) .
tucker , r. hernandez , phys .
b. , * 116 * 1328 ( 2012 ) .
a. ochab - marcinek , r. holyst , soft matter * 7 * ( 16 ) 7366 ( 2011 ) .
r. vasanthi , s. ravichandran , b. bagchi , j. phys . chem .
* 115 * 10022 ( 2001 ) .
m. majka , p.f .
gra , acta phys .
b * 43 * 1133 ( 2012 ) .
m. majka , p.f .
gra , phys .
e * 86 * 051122 ( 2012 ) .
b. wang , j. kuo , s.c .
bae , s. granick , nature materials , * 11 * 481 ( 2012 ) .
f. mondiot , j - c .
loudet , o. mondain - monval , p. snabre , phys .
e. * 86 * 010401(r ) ( 2012 ) .
m. khoury , a.m. lacasta , j.m .
sancho , k. lindenberg , phys .
lett . * 106 * 090602 ( 2011 ) .
b. dybiec , e. gudowska - nowak , phys .
e * 80 * 061122 ( 2009 ) .
z. adamczyk , j. barbasz , m. ciesla , langmuir , * 26 * 11934 ( 2010 ) .
z. adamczyk , j. barbasz , m. ciesla , langmuir , * 27 * 6868 ( 2011 ) .
j. feder , j. theor .
biol . * 87 * , 237 ( 1980 ) .
m. ciesla , j. barbasz , j. stat .
mech . p03015 ( 2012 ) .
j. barbasz , m. ciesla acta phys .
b * 44 * 937 ( 2013 )
. m. ciesla , j. barbasz , colloids and surf .
b * 110 * 178 ( 2013 )
. j. klafter , i.m .
sokolov , _ first steps in random walks : from tools to applications _
, oxford university press ( 2011 ) . t. munk , f. hfling , e. frey and t. franosch , epl * 85 * 30003 ( 2009 ) .
s. havlin , d. ben - avraham , adv .
phys . * 51 * 187 ( 2002 ) .
klemens , _ thermal conductivity of inhomogeneous media _ , high temp . - high press .
* 23 * 241 ( 1991 ) b.ya balagurov , j. exp .
* 113 * 849 ( 2011 ) .
shklovskii , phys .
solidi * 85 * k111 ( 1978 ) .
b.ya balagurov , russ .
j. of phys .
b * 7 * 71 ( 2013 ) .
s. feng , b. i. halperin and p. n. sen , phys .
b * 35 * 197 ( 1987 ) .
i.l . novak , f. gao , p. karikivski , b.m .
slepchenko , j. chem . phys .
bf 134 154104 ( 2011 ) .
|
we investigate the obstructed motion of tracer ( test ) particles in crowded environments by carrying simulations of two - dimensional gaussian random walk in model fibrinogen monolayers of different orientational ordering .
the fibrinogen molecules are significantly anisotropic and therefore they can form structures where orientational ordering , similar to the one observed in nematic liquid crystals , appears .
the work focuses on the dependence between level of the orientational order ( degree of environmental crowding ) of fibrinogen molecules inside a layer and non - fickian character of the diffusion process of spherical tracer particles moving within the domain .
it is shown that in general particles motion is subdiffusive and strongly anisotropic , and its characteristic features significantly change with the orientational order parameter , concentration of fibrinogens and radius of a diffusing probe .
|
understanding the origin and evolution of the fuel supply that feeds quasars and other active galactic nuclei ( agn ) is a key mystery in the astrophysics of these systems . to account for luminosities as high as @xmath2 erg / s the most efficient conceivable process - accretion onto a black hole - requires a mass supply of order one solar mass per year . that matter ( which over the lifetime of the active phase may amount to @xmath3 or more )
must be transported from galactic scales down to the schwarzschild radius of the hole at just a few a.u .
evidently this transport mandates an efficient mechanism for disposing of the inflowing material s angular momentum . a partial solution to the problem is provided by invoking the presence of an accretion disk in the immediate vicinity of the hole . although the details are uncertain , viscosity in the disk transports angular momentum outward , and thereby permits mass to flow inward .
recent observations by _ asca _ of relativistically broadened iron @xmath4 lines lend strong support to the idea of the existence of both massive black holes and disks in agn ( tanaka et al .
1995 ) . for the purpose of fuelling the hole , however , the disk is unlikely to provide more than a short term reservoir of mass , as models suggest that the disk mass is only a small fraction ( perhaps @xmath5 for a @xmath6 black hole ) of the mass of the hole ( e.g. clarke 1988 ) . even allowing for generous uncertainties in our models of accretion disks
, therefore , it seems inevitable that the disk itself must be continually replenished from some further , external , source of mass .
two possibilities suggest themselves for the source of this external fuel supply . in the first
, the disk could be fed by gas funnelled in from large ( kpc ) galactic scales , perhaps as a consequence of interactions between the agn host galaxy and its neighbours ( hernquist & mihos 1995 ) .
this scenario is attractive insomuch as there is undoubtably sufficient mass in gas at those large scales , though whether it can be transported to the sub - parsec scales of the accretion disk efficiently and without forming stars in the process is less clear . as a second possibility
, the disk could be resupplied as a consequence of interactions occurring in a dense nuclear star cluster .
such clusters are observed in many well - studied nearby galaxies , with the core of m32 exhibiting a central stellar density exceeding @xmath7 ( lauer et al 1992 ; crane et al 1993 ) .
clusters of similar or modestly greater ( @xmath8 ) richness then have the potential to be a significant source of agn fuel_if _ a mechanism exists that allows the agn to tap into that supply .
disruption of stars in the tidal field of the black hole was suggested by hills ( 1975 ) as a possible way to liberate the gas .
disruption occurs when the pericentre @xmath9 of the orbit becomes comparable or smaller than the ` tidal ' radius @xmath10 where @xmath11 is the hole mass ( rees 1988 )
. an attractive prediction of this scenario is that activity should turn off when the central black hole has grown to the point that @xmath12 is comparable to the schwarzschild radius , @xmath13 after which stars are swallowed whole without significant release of radiation . for solar type stars
this occurs when the black hole has reached a mass of a few @xmath14 . for the purpose of long - term fuelling of the black hole
, however , this mechanism proves inadequate
. rapid disruption of stars passing within @xmath12 depletes a region in phase space which is refilled via relaxation of the stars in the central cluster on a timescale @xmath15 yr .
this rate is insufficient to provide enough fuel for the most luminous agn ( shields & wheeler 1978 ) .
other possibilities that have been assayed include mass loss from stellar winds , which might be enhanced as a result of the agn s ultraviolet or high energy particle flux .
a generic problem with such ideas is that gas released throughout the volume of the nuclear star cluster is vulnerable to being expelled by the radiation pressure from the central source .
if only a fraction of the liberated mass reaches the disk or black hole then the demands placed on the supply of mass are appropriately aggravated .
the interaction between a nuclear star cluster and the accretion flow has been explored for the case of main sequence stars by a number of authors . for
such stars collisions with the disk will tend to expel matter out of the disk plane while leaving the star relatively unscathed .
this mechanism has been proposed as an origin for the gas producing the broad emission lines seen in agn spectra ( zurek , siemiginowska & colgate 1991 , 1994 ; see addendum in apj xxxx ) .
repeated collisions will gradually dissipate the star s orbital energy and in time drag it bodily into the plane of the disk ( syer , clarke & rees 1991 ) .
the consequences of stars captured in this fashion , for example for the metallicity of quasar accretion disks , have been extensively investigated ( artymowicz 1993 ; artymowicz , lin & wampler 1993 ) . here , we extend previous studies on star - disk interactions to consider the case of giant stars , with radii typically @xmath16 .
the greatly increased geometric area and reduced binding energy suggest that such stars should be susceptible to large - scale stripping of mass on collision with the disk ( zurek , siemiginowska & colgate 1994 ) . within some critical radius @xmath17 collisions
will then destroy the star and unbind the envelope mass within the red giant lifetime .
the value of @xmath17 will depend on the extent and properties of the disk as well as the velocity and structure of the star , but as we show later it is invariably larger than the radius within which stars are trapped while still on the main sequence .
if @xmath17 and the central stellar density are sufficiently large , this mechanism could then provide an important source of fuel for the central black hole .
the plan of this paper is as follows . in section 2
we discuss the numerical methods and initial conditions of our calculations . in section 3 we present results of hydrodynamic simulations of the star - disk interaction for a range of stellar velocities @xmath18 , and disk surface densities @xmath19 .
our aim is to delineate that portion of the @xmath20 parameter space in which collisions lead to significant envelope loss . in section 4
we then combine the results of our calculations with simple models for the accretion disk and stellar cluster , in order to estimate how much fuel could be provided to the agn by this mechanism .
section 5 summarises our results and conclusions .
the calculations described here were performed using a three - dimensional smoothed particle hydrodynamics code ( sph ) , minimally modified from that described by benz ( 1990 ) .
this code has been extensively used in other simulations of stellar collisions ( eg davies , benz & hills 1991 ) , and has been shown to give the same results on such problems as codes based on alternative numeric techniques ( eg ppm davies et al . 1993 ) . in constructing our model star ,
we commence with a red giant structure calculated using the latest version of the eggleton stellar evolution code ( pols et al . 1995 and references therein ) .
the model is computed for a mass of @xmath21 , with a core mass of @xmath22 and a radius @xmath23 .
we take this model as being reasonably representative of red giant structures in galactic nuclei , while acknowledging that a more complete investigation than is attempted here should consider a range of stellar masses and radii .
the model also ignores the possible consequences of the star s proximity to the agn , though we note at this stage that the main effect of irradiation on giants is to _ increase _ their radii ( tout et al .
1989 ) , which would tend to accentuate envelope loss on collision with the disk .
the structure of the giant model is shown in fig .
1 , which plots the column density through the star ( @xmath24 ) as a function of the fractional radius @xmath25 . for comparison
we also show the structure of a main sequence star of the same mass , modelled as an @xmath26 polytrope . for the red giant the typical column densities near the outer edge of the envelope are comparable to those seen in models of agn accretion disks ( @xmath27 ) , while for the main sequence star
the densities are several orders of magnitude greater .
thus if each collision strips off an annulus of mass with column density below some critical threshold ( and we show later that this is indeed a reasonable estimate ) , then the red giant is vastly more susceptible to mass loss .
this is quantified in fig .
2 , which plots the mass exterior to a given column density for the two models . at the relatively low @xmath19 values found in agn disks , the giant has a vulnerable mass fraction more than four orders of magnitude in excess of a main sequence star of the same mass .
this implies that a star may survive repeated disk transits on the main sequence yet be destroyed relatively swiftly on reaching the red giant phase of its evolution . to represent the stellar structure in sph
, we replace the core of the giant with a single point mass that interacts with the remaining particles via gravitational forces only .
the envelope is modelled using 2274 sph particles that are set down on a close - packed grid with masses proportional to the local density in the model star .
as the density in an sph calculation is not a local quantity this is a crude procedure , but it gives a density run that matches the stellar model to about 20% over 2 orders of magnitude in density ( fig .
3 ) , which is is adequate for our purposes here . the internal energy and pressure of the particles
are then set to ensure hydrostatic equilibrium , using an equation of state that includes a mixture of gas and radiation pressure , @xmath28 where @xmath29 is the mean molecular weight and @xmath30 the radiation constant .
we have verified that stars modelled using this procedure are stable for at least several sound crossing times , which is an order of magnitude longer than the duration of the star - disk simulations . for the disk
we adopt a simple approach .
the timescale for the star to transit the disk is only @xmath31 s , which is negligible compared to any timescale in the disk .
we therefore represent the disk as a column of initially cold gas , with a gaussian density profile in the vertical direction ( corresponding to an isothermal temperature distribution ) .
disk models suggest that the thickness of the disk is at least an order of magnitude greater than the size of even our giant star , so we pick a scale height @xmath32 such that the density gradient across the star is small . for our main simulations
we take @xmath33 , and truncate the disk in the vertical direction at 1.5 scale heights .
we also experiment with other choices for @xmath32 . with these parameters ,
we achieve comparable resolution in the star and disk with around 19,000 disk particles .
table 1 lists the main parameters for the sph simulations of star - disk encounters , the stellar velocity @xmath18 , disk surface density @xmath19 , and scale height @xmath32 .
we also tabulate the impacting disk mass @xmath34 , the incident momentum @xmath35 , and kinetic energy @xmath36 . as
the binding energy of the model stellar envelope is only @xmath37 erg , _ all _ of these collisions have the potential in principle to unbind a large fraction of the stellar material .
the final column in the table shows the main result of these calculations the mass loss from the red giant evaluated at a time 0.5 disk transit times after the star exits the disk . for all except the most violent collisions ( which have anyway
almost totally destroyed the stellar envelope by this stage ) the induced mass loss has ceased by this time .
figure 4 ( _ omitted from this draft _ ) shows a montage of frames from simulation e , where at each snapshot we have plotted particles within 2 sph smoothing lengths of the @xmath38 plane .
this is the most energetic encounter , and as such demonstrates most clearly the two effects leading to mass loss .
firstly , momentum transfer between the disk gas and the star can bodily strip material from the sides of the star .
this reduces the cross - sectional area the star presents to the disk , and creates a wake composed primarily of stellar material behind the star .
second , the disk gas impinging on the front surface of the star shock heats and compresses the material there .
when the star emerges from the disk the heated gas then blows off as an almost spherical expanding cloud .
these general hydrodynamic properties of the collision are very similar to those seen in simulations of stars impacted by shells of supernova ejecta ( fryxell & arnett 1981 ; taam & fryxell 1984 ; livne , tuchman & wheeler 1992 ) . for supernovae that explode with red giant companions , indeed ,
the energy and momentum of the encounter are similar to those considered here .
the main difference is then the relative thickness of the disk compared to the star unlike shells of supernova ejecta typical agn disks are always many stellar radii thick .
figure 5 shows for runs a - e the bound mass in the stellar envelope as the star crosses the disk .
we compute the bound mass at regular intervals by summing the masses of particles whose total energy ( gravitational plus internal and kinetic ) is less than zero .
these simulations all have the same disk surface density ( @xmath39 ) and scale height ( @xmath33 ) , and differ only in the stellar velocity . to facilitate comparison between the runs ,
we scale the time axis so that the disk crossing takes place between ( in arbitrary time units ) @xmath40 and @xmath41 for all the simulations . from the figure
, it can be seen that the behaviour divides into two regimes .
the two most energetic collisions ( d and e ) disrupt the stellar structure entirely .
mass loss during the disk crossing is rapid , and most of the gas in the remnant that emerges from the disk is itself formally unbound from the white dwarf core .
this prompt destruction of the star as a consequence of a single disk passage occurs at this disk surface density for velocities of around 4000 km / s or greater .
for agn , that implies radii in the disk of less than 6000 @xmath42 , which as we shall show in section 4.2 encompasses all the non - self - gravitating disk in a supermassive system of @xmath3 .
less violent impacts ( a - c ) produce smaller fractions of stripped and ablated mass , and leave the inner regions of the star relatively unaffected . in these cases the rate of mass loss declines rapidly once the star has exited the disc , and so the final bound mass in our simulations is likely to give a good estimate of the total mass loss from the disk encounter .
we find mass loss that ranges from @xmath43 for the @xmath44 km / s run down to approximately 1% of a solar mass for the @xmath45 km / s simulation . to further interpret our simulations , we compare our numerical results to the semi - analytic theory developed by wheeler , lecar & mckee ( 1975 , henceforth wlm ) for the analogous situation of a star impacted by a supernova blast wave .
although we have already noted that there are important differences between that scenario and the star - disk encounter , the wlm theory has the merit of providing a simple baseline against which to compare results . in the case of supernovae the theory has been shown to agree closely with detailed calculations ( fryxell & arnett 1981 ) .
following wlm , we identify two components to the total mass loss .
first , direct momentum transfer strips matter from the outside of the star down to some fraction of the stellar radius @xmath46 .
this critical radius is that where the momentum transfered to a cylindrical shell is just sufficient to accelerate it to the stellar escape velocity @xmath47 at that radius .
as the impacts we are considering are highly supersonic ( @xmath48 ) this condition implies that mass is lost exterior to the cylinder where @xmath49 the quantity @xmath50 is the column density through the star as plotted in fig .
the stripped mass fraction by this direct mechanism , @xmath51 , is then the total mass in the envelope exterior to a cylinder of radius @xmath52 . from our red giant model @xmath53 and
hence @xmath51 are straightforward to evaluate for a given incident momentum @xmath35 .
calculating the second , ablation component of the mass loss requires computing the heating generated by the shock that is driven into the star as a consequence of the collision .
as the star exits the ejecta shell or disk , this thermal energy will become available as kinetic energy as the heated material ` boils off ' the stellar surface .
wlm give the result , @xmath54 where their parameter @xmath55 is given in our notation by , @xmath56 for highly supersonic encounters , @xmath55 is thus a measure of the incident momentum .
figure 6 compares the wlm predictions with our numerically derived mass loss estimates for runs a
e . at high velocities of 4000 km
/ s or greater there is reasonable agreement between the mass loss seen in our simulations and the total predicted by the wlm analysis @xmath57 .
both agree that at these velocities essentially all of the stellar envelope becomes unbound as a consequence of the collision . at lower impact energies and momenta ,
we find that @xmath51 alone provides a good estimate of the mass loss seen in the simulations , suggesting that for these lower velocity ( @xmath58 ) encounters mass loss is driven primarily by momentum transfer rather than ablation .
this transition as the velocity decreases is plausible on physical grounds the large thickness of the disk relative to the stellar radius means that the collision is less like a single impulsive impact and more closely resembles a steady flow , albeit one in which the star undergoes steady stripping .
thus compared to the supernova case the strength of shock heating and ablation is liable to be reduced , leaving the directly stripped component dominant at low velocity . in support of this interpretation
we note that repeating run c with the disk scale height halved ( run i ) leads to a doubling of the mass loss .
whereas run c shows mass loss intermediate between @xmath51 and @xmath59 , run i exhibits mass loss given closely by @xmath59 .
table 2 shows the results for all the runs described here . from our simulations
we can then identify two regimes .
for high velocities and/or small disk scale heights the mass loss is large and can be approximated well by the total mass loss predicted by wlm , @xmath59 .
this result agrees with the supernova studies referred to above .
conversely for lower velocity encounters with disks that are thick compared to the stellar size the mass loss is substantially reduced below the wlm prediction . in this case
we can take @xmath51 alone as a reasonable estimate of the mass loss that will occur in the collision .
for the purpose of supplying fuel to the disk , the most important collisions are those that occur at relatively low velocity at a large radius in the disk .
although the mass loss per collision is small , the cumulative effect of many such impacts dwarfs that of the violent impacts that unbind the envelope in a single disk transit .
we quantify this statement in the following section , but note here that it is only true if the equilibrium structure of the perturbed star is _ at least _ as vulnerable to stripping as the original model .
if this is the case we can regard the mass loss predicted for a given @xmath18 and @xmath19 by our simulations as a _ lower _ limit on the mass loss in subsequent disk encounters . to calculate the post - encounter stellar structure we first note that radiation pressure is negligible in the bound stellar envelope of the remnant , so that the polytropic exponent @xmath60 is everywhere close to @xmath61 . from our sph simulations
we can then extract the polytropic constant @xmath62 at the location of each particle , @xmath63 following the impact @xmath62 displays considerable scatter as a function of radius .
this is a result of the non - symmetric injection of entropy from the shock - heating of the envelope .
we therefore average over angles to obtain a smooth approximation to @xmath64 for use in calculating the structure .
we also compute the mass interior to radius @xmath65 to express @xmath62 as a function of enclosed mass @xmath66 .
physically we expect the star to settle to a new equilibrium with a spherically symmetric @xmath62 distribution on of order the convective timescale much shorter than the interval between collisions with the disk .
provided that this is the case our averaging procedure and assumption of spherical symmetry should be valid . combining the equation representing hydrostatic equilibrium , @xmath67 with equation ( [ 3.3.1 ] ) above
then gives an expression for the stable density gradient in terms of the function @xmath68 , @xmath69 this equation together with the known total mass of the remnant constitute an eigenvalue problem for the central density , which we solve using a shooting method to yield the equilibrium density profile .
the results of this analysis are shown in fig .
7(a ) for run a the lowest velocity encounter considered here .
the density profile of the inner regions ( @xmath70 ) is predicted to be almost unaltered by the impact , as expected since the influence of the collision on the star is confined to the outer layers for the less violent impacts .
the main effect of the increased entropy generated by the encounter is to puff up the outer envelope , leading to an increase in the radius of about 30% .
7 illustrates how the recomputed stellar model differs from the initial red giant in its vulnerable mass fraction , expressed as the mass @xmath71 exterior to a cylinder of column density @xmath19 ( cf fig .
the post - encounter star is _ more _ vulnerable to stripping than the initial model , by factors of around 2 for the typical values of column density considered here .
we are therefore confident that repeated star
disk encounters will act cumulatively to disrupt the stellar envelope , and regard the estimates given in section 4 ( where we assume that the total mass loss is linear in the number of impacts ) as conservative predictions of the rate at which destruction occurs .
in the previous section we computed the mass stripping associated with a variety of star - disk collision parameters . the most violent collisions ( @xmath72 km / s , @xmath73 ) lead to essentially complete stripping of the stellar envelope on a single disk transit . for less violent
encounters the mass loss and disruption of the stellar structure is greatly reduced . in this regime
we have used our results to validate a simple estimation procedure for the mass loss , based on the amount of momentum imparted to the stellar envelope by the impact . in this section
we utilise the above results to investigate the significance of this mechanism for the fuelling of agn . in section 4.1
we calculate the radius out to which a disk could strip stars of their envelopes within the giant lifetime
this determines the volume of the cluster that is vulnerable to disk stripping .
we compare this to the radii at which the competing processes of tidal disruptions ; star - star collisions ; and stellar trapping would operate . in section 4.2
we calculate thin disk models to determine the ` target area ' of the disk which has surface density sufficient to cause stripping . finally in section 4.3
we estimate the total mass that this mechanism might provide to fuel agn activity for a variety of central stellar densities , and use the results of stellar evolution calculations to predict the evolution of the mass supply with time .
on each disk transit , the star will be stripped down to a column density @xmath74 given by equation ( [ 3.15 ] ) .
the corresponding mass loss can be directly read off from fig .
2 , which is well approximated for small @xmath75 by @xmath76 the timescale between encounters for stars orbiting at a radius @xmath65 is @xmath77 where @xmath78 is the total enclosed mass at that radius including the contribution from the cluster itself . imposing the condition that at @xmath17 the star will just be stripped during the red giant lifetime @xmath79 then yields , @xmath80 where we have inserted reasonable values for all the parameters .
a @xmath3 black hole is therefore capable of stripping red giants out to a distance of several tenths of a parsec , _ if _ the disk matter itself extends to that radius .
the red giant stripping process discussed here will evidently be important only if stars survive to become giants without first being destroyed by other processes , such as tidal disruption , stellar trapping , or star
star collisions .
the radius at which stars are disrupted by the tidal field of the black hole is given by equation ( [ 1 ] ) ; we now estimate the critical radii for the other two competing processes . in evaluating the timescale for stars
to be ground into the plane of the disk we follow closely the approach of syer , clarke & rees ( 1991 ) . if the only mechanism leading to decay of the orbit is direct hydrodynamic drag on passing through the disk , then to order of magnitude the trapping timescale will be the time required for the star to impact its own mass of disk material . if on each passage the star collides with a mass @xmath81 , @xmath82 where @xmath83 is a dimensionless scaling factor that depends on the initial orbital parameters ( syer , clarke & rees 1991 ) .
stars will be brought into the disk plane provided that @xmath84 is shorter than the relaxation time @xmath85 , after which stars will typically have diffused onto less radial orbits .
this criteria gives a radius interior to which stars will be trapped , @xmath86 which is typically intermediate between the tidal disruption and giant stripping radii .
for the dense stellar systems we are considering , physical collisions may destroy stars before they even reach the giant stage of their evolution .
we note at this stage that physical collisions might themselves be important in fuelling agn , however they are undoubtably a different type of mechanism to that considered here and require detailed study to determine , for example , the efficiency of mass loss . for stars of a single representative radius @xmath87 , a simple expression for the collision rate
@xmath88 is , @xmath89 where @xmath90 gives the run of the stellar number density with radius . for simplicity , we assume a stellar profile that is a power - law in radius , with a stellar number density @xmath91 . for an isothermal sphere @xmath92 , while more detailed cusp solutions for stellar clusters containing a central black hole have @xmath93 ( binney & tremaine 1987 and references therein ) .
a condition for collisions to be important at a radius @xmath65 is that @xmath94in which case stars within that radius will be destroyed before completing their main sequence life @xmath95 . for @xmath92
this radius is @xmath96 where @xmath97 is the mean stellar mass , @xmath98 is the cluster mass enclosed within a radius of 1 pc , and @xmath11 is the total mass within radius @xmath65 .
8 shows these radii as a function of black hole mass . for this plot
we have taken parameters @xmath99 yr , @xmath100 yr , and @xmath101 yr .
we assume a cluster with a mass @xmath102 within 1 pc , and a disk with a surface density @xmath19 of @xmath103 ( ignoring for now the complication of having @xmath19 itself a function of radius ) .
there is a large parameter space here , but from the figure it is evident that the red giant stripping radius @xmath17 exceeds the other radii plotted by at least an order of magnitude for all black hole masses in the illustrated range .
this implies , in particular , that most stars that are vulnerable to stripping as giants _ will _ survive to reach the red giant phase without suffering prior destruction by physical collisions .
we note here that our estimate for the collision time on the main sequence is consistent with that of begelman & sikora ( 1991 ) , who concluded that models in which red giants generated broad emission lines from their stellar winds could be ruled out , because the stars would collide before reaching the giant phase .
the apparent difference with our result is caused by the requirement for such models , ( in which the stellar line - producing envelopes must provide a covering fraction towards the central source of @xmath1 ) , to invoke a vastly greater stellar density in red giants than is required here .
experimentation with the simple models presented above shows that , for most values of the parameters , we obtain plots qualitatively similar to fig . 8 , except that for lower central stellar densities the collision radius is reduced and lies interior to the trapping radius .
we therefore conclude that for a disk that extends to large radius stellar stripping of giants will provide a greater source of fuel to the disk than the other stellar mechanisms considered here . to determine the typical extent and properties of agn accretion disks , we need the run of surface density with radius in the disk . to calculate this ,
we follow the approach of clarke ( 1988 ) to construct vertically averaged thin disk solutions , which give the radial dependence of the surface density , sound speed , and other disk quantities .
although it is possible to go beyond vertically averaged models by solving for the details of the disk vertical structure , such calculations are more ` accurate ' only if the site of energy dissipation in the disk is known . as this in turn depends on the unknown processes generating the viscosity , we can not in practice expect to do better than the simple vertically averaged models presented here .
our models use the ` alpha - prescription ' for the viscosity ( shakura & sunyaev 1973 ) ; for completeness we set out the thin disk equation solved here in an appendix .
figure 9 shows contour plots of the disk surface density @xmath19 as a function of radius and mass transfer rate @xmath104 in the disk , for black hole masses that cover the range generally considered for agn .
these models assume constant @xmath104 with radius , and an @xmath30 of 0.1 , similar to the inferred values in dwarf nova disk systems .
the lower ( @xmath105 ) values favoured by some authors from analyses of uv variability ( siemiginowska & czerny 1989 ) would increase the surface density in the disk , and thus somewhat enhance the efficacy of the stripping process . at each radius in our solutions , we also evaluate the value of the toomre @xmath106 parameter , which measures the importance of self - gravity , @xmath107 in this equation @xmath108 is the mid - plane sound speed and @xmath109 the keplerian angular velocity in the disk . if @xmath110 the disk becomes unstable to local gravitational instabilities , and although the structure in this regime is uncertain , the disk may develop spiral density waves or break up into clumps . as in such a situation
the angular momentum is _ not _ transported via an @xmath30 viscosity , consistency requires that we truncate our solutions when @xmath106 falls below unity .
9 shows that for accretion rates between @xmath5 and @xmath111 @xmath112 the surface density in the disk will be of order @xmath103 or greater .
this is the case for all the black hole masses considered here .
however we also find that the disk would be expected to become self - gravitating at relatively small radii , ranging from @xmath113 cm ( @xmath114 pc ) for the @xmath115 hole to around @xmath116 cm for the @xmath3 case .
the properties of the disk beyond these radii are unclear , though calculations by laughlin & bodenheimer ( 1994 ) for self - gravitating protostellar disks suggest that gravitational torques transport angular momentum with an efficiency equivalent to an @xmath30 of around 0.03 .
if this applies also to agn disks , then it suggests that the average @xmath19 in the self - gravitating regime should be at least as large as that calculated from the shakura - sunyaev analysis above . from fig .
8 we find that supermassive black holes should be able to strip mass from red giants out to radii of @xmath1 pc provided only that their accretion disks extend to that radius
. a 0.1 pc disk with average @xmath117 has a mass of @xmath118 , which is still only a small fraction of the hole mass for the larger mass black holes .
if the cluster mass @xmath102 , then the mass of stars intersecting the disk will be of order @xmath115 , with the precise number dependent on the distribution of stellar orbits in the cluster ( with more radial orbits leading to a greater fraction of vulnerable stars ) . if this mass of stars on orbits susceptible to stripping is fixed , then the rate at which this mass is liberated to the disk is controlled purely by stellar evolutionary processes .
once a given star reaches the red giant phase of its evolution its envelope mass will rapidly be deposited into the disk . to investigate the evolution predicted by this model
, we shall assume that the stellar population in the nuclear star cluster can be described as a coeval population with a power - law initial mass function ( imf ) @xmath119 typically we take @xmath120 , corresponding to a salpeter imf , and impose a lower mass cut - off at the hydrogen - burning limit ( @xmath121 ) .
if after a time @xmath122 ( where @xmath123 is the time of the initial star formation ) stars of mass @xmath124 are leaving the main sequence and becoming giants , then the rate at which mass is being deposited into the disk is @xmath125 in this expression @xmath126 is the core mass which is not captured by the disk , while the envelope mass has been assumed to be instantaneously deposited into the disk once the star has evolved off the main sequence .
this should be a reasonable approximation given that the red giant lifetime is much shorter than that on the main sequence .
once the hydrogen burning stellar lifetime @xmath127 is specified , equation ( [ 4.3 ] ) gives the mass deposition rate as a function of time . to gain some insight into the prediction of this model for the fuelling rate , we first assume that @xmath127 is a power - law in the stellar mass . for masses of a few @xmath128 ,
a crude estimate is , @xmath129 at early times , when @xmath130 , the rate of deposition of agn fuel is then given by @xmath131 where the final step uses for definiteness @xmath120 .
the fuelling history is dominated by the steep dependence of the stellar lifetime on the mass , so that the mass deposition rate declines roughly inversely with time .
if all other parameters of the system remained fixed ( for example disk radius and surface density ) , we would then expect the supply of fuel today to be only a thousandth of the rate at an early ( @xmath132 yr ) epoch .
rapid evolution with cosmic time is , of course , a well - known property of agn .
figure 10 shows the results of detailed calculations of the hydrogen burning lifetime @xmath127 by fagotto et al .
( 1994 ) , for stars in the mass range 0.6 @xmath133 120 @xmath128 .
these calculations assume a low metallicity ( @xmath134 ) and employ modern ( opal ) opacities . combining this function with an assumed imf yields curves for the mass deposition rate , and the fraction of total mass liberated by a given time , that are shown in fig .
we take @xmath135 , and vary the cluster mass and slope of the imf . for a cluster mass that intersects the disk of @xmath136
, we obtain a deposition rate at @xmath132 yr of nearly @xmath137 , which declines by approximately 2 orders of magnitude by @xmath138 yr . for imfs that are close to the salpeter value , we find that about 20 - 30% of the initial cluster mass has been released to the disk within that time , though this is fairly sensitive to the assumed imf . a steep imf locks up a large fraction of the total mass in long - lived low - mass stars , and thus lowers the fraction of mass that reaches the giant phase and is supplied to the disk
. we also note that the effect of an extended period of star formation would be to lower and ` smear out ' the period of peak mass deposition .
this would occur if the high mass star formation epoch exceeded approximately @xmath132 yr
. typical agn luminosities are of the order of @xmath139 erg / s , which corresponds to an accretion rate of @xmath140 ( assuming a 10% efficiency of converting rest mass into radiation ) .
11 then shows that mass deposition rates of this magnitude can be supplied from the early evolution of a cluster that has a few @xmath141 of stars on vulnerable , disk - intersecting orbits .
for the accretion disk radii of @xmath1 pc postulated above , this requires a central stellar density within 1pc of the hole of order @xmath136 .
an immediate prediction of this model is that the rate of mass deposition into the disk should evolve rapidly .
11 is calculated assuming that the target area of the disk remains fixed , and so strictly represents the time - evolution of the mass supply if the disk is primarily replenished from some source other than star - disk collisions .
in such a scenario , the main effect of the stripping would be to enrich the disk with the products of stellar nucleosynthesis .
such enrichment would be most significant at early times , shortly after the cluster was formed , and would subsequently drop off rapidly as shown in the figure .
even faster evolution is predicted if star - disk collisions _
dominate _ the mass supply to the disk , as in this instance the radius to which the disk can strip stars would itself decrease with time .
the evolutionary track would then move toward the lower curves in the figure , as fewer and fewer stars became vulnerable to destruction via disk impacts . although the details would depend on the ( poorly known ) disk physics ,
the general scenario is then one in which a burst of star formation in the cluster leads to a short , ( perhaps @xmath142 yr , as by then the accretion rate would have fallen to the values where @xmath19 starts dropping rapidly - fig .
9 ) , but intense flare in the agn activity . as the mass deposition drops , so the disk weakens until it can no longer strip stars , and the activity then ceases altogether .
in this paper we have presented simulations of encounters of red giant stars with agn accretion disks .
we find that for encounters within around 0.1 pc of the black hole , and disk surface density exceeding @xmath103 , the fractional mass loss on collision is sufficiently large that the star would not survive the red giant phase .
rather , the envelope mass would be deposited into the disk from the cumulative effect of numerous disk transits , leaving only a stripped core orbiting the black hole .
the most violent collisions achieve this end in only a single passage .
although we have considered only one type of giant , we believe these conclusions are likely to apply equally to a range of red giant masses and structures . we have compared the radius out to which red giants
can be stripped via encounters with the disk with the radii at which stellar collisions and trapping by the disk become important .
the stripping radius is generally one or two orders of magnitude greater than @xmath143 and @xmath144 , and indeed is most likely to be limited by the actual physical extent of the accretion disk . even if the stellar density in the nucleus is relatively low
, this mechanism could provide an important source of gas that has been enriched as a result of stellar nucleosynthesis .
artymowicz ( 1993 ) has suggested that heavy elements might be generated _ in situ _ by trapped stars accreting and then exploding as supernovae within the disk . the larger radius at which stripping can occur means that a similar quantity of enriched material could be provided to the disk by this process .
an immediate consequence of our models is that stars in the innermost regions of galaxies harbouring accretion disks will have their lifetime as red giants severely curtailed by star - disk collisions .
the resultant modification of the stellar population is probably on too small a scale to be observable , though the depletion of the red giant population close to the black hole does cast further doubt on models that seek to invoke giants as important components in the central regions of agn .
conversely , the rapid destruction of giants will _ enhance _ the rate of production of white dwarfs , which will remain as a dense cluster on disk - intersecting orbits . as suggested in a recent paper by shields ( 1996 ) , these single white dwarfs may accrete sufficient gas on passage through the disk to greatly enhance the rate of novae , and this provides yet a further mechanism for enriching the disk gas .
we note that in principle better observations of the chemical mix in the centers of agn could constrain which , if any , of these mechanisms giant stripping , novae or supernovae were important in supplying heavy elements to the gas
. how important might this mechanism be in fulfilling the _ overall _ fuelling requirements of quasars and other agn ?
since the stellar densities present in the cores of agn remain unknown , it is impossible to answer this question with any great certainty .
we find that typical agn luminosities could be achieved from this mechanism if the central stellar densities are as high as @xmath145 or greater , and the accretion disk extends to a few tenths of a pc .
these are _ lower _ densities than are required for collisions between stars to liberate equivalent quantities of fuel .
the activity of agn powered by this fuel source would decline rapidly as the mass of stars becoming giants dropped , and would eventually switch off entirely as the disk became too tenuous to strip further stars .
such a scenario is not inconsistent with the evolutionary properties of quasars and agn .
although we have concentrated here on feeding the disk via a single population of stars in a central cluster , we also note that higher total masses of fuel could be provided by this mechanism if the depleted phase space could be replenished perhaps by stars driven inward from the host galaxy by some means .
this is of course essentially the ` interaction - driven ' model of fuelling that we have previously mentioned .
in such a scenario , the results given here imply that it is irrelevant whether the mass arrives at the nucleus in the form of gas or stars even in the latter case a reasonable fraction of the total mass will be readily made available to feed the agn activity .
we thank jose font - roda , chris fryer and cathie clarke for commenting on earlier versions of this work , and chris tout for supplying the stellar model .
pja thanks t6 , lanl , for their hospitality over the summer .
mbd acknowledges support from a royal society urf .
this work was partially supported by iggp , lanl .
in section 4.2 we have presented solutions to the vertically averaged thin disk equations .
these are set out in , for example , faulker , lin & papaloizou ( 1983 ) , and follow the approach of shakura & sunyaev ( 1973 ) .
explicitly we use the following forms .
local thermal equilibrium of an isolated annulus of the disk implies , @xmath146 where @xmath147 is the disk effective temperature , @xmath148 the stefan - boltzmann constant , @xmath149 the kinematic viscosity , and @xmath109 the keplerian angular velocity .
a vertically averaged treatment of the disk radiative transport gives a relation between the effective temperature and the central temperature @xmath150 , @xmath151 where @xmath152 is the midplane opacity .
the local mass flux @xmath104 is @xmath153 for a disk with a zero - torque ( @xmath154 ) inner boundary condition at a radius @xmath155 .
the alpha prescription for the viscosity is conventionally expressed as , @xmath156 with @xmath30 a free parameter and @xmath157 the central disk density . in lieu of greater knowledge of the form of the viscosity in agn disks , we take @xmath30 to be a constant , typically 0.1 .
the equation of state is as previously used , @xmath158 finally we have the equation representing hydrostatic equilibrium in the vertical direction , @xmath159 where @xmath108 is the midplane sound speed and @xmath32 is the disk scale height .
the set of equations ( [ a1 ] ) - ( [ a6 ] ) can be readily solved once the opacity @xmath152 is specified .
we utilise the analytic fits given by bell & lin ( 1994)these are identical to those used by lin & papaloizou ( 1985 ) except for modifications to the low ( below 3000k ) temperature regime suggested by alexander et al .
detailed expressions are as given in bell & lin ( 1994 ) .
|
we explore the role that red giants might play in the fuelling and evolution of active galactic nuclei . due to their large radii and the low binding energy of the stellar envelope
, giants are vulnerable to envelope stripping from collisions with the accretion disk . using hydrodynamic simulations
we show that such collisions will typically deposit a substantial fraction of the envelope mass into the disk on each passage .
repeated encounters will then lead to the complete destruction of the star save for the dense core .
we estimate the rate of fuel supply by this mechanism using simple models for the agn disk and central stellar cluster .
if the central stellar density is @xmath0 , then stripping of giants could account for the activity of typical agn provided that the accretion disk extends out to @xmath1 pc . for agn with smaller disks , or clusters of lower central density , giant stripping could supply gas enriched via stellar nucleosynthesis to a disk replenished from some other source .
we find that , for typical parameters , this mechanism is able to supply important quantities of gas to the disk at lower stellar densities than previously proposed stellar fuelling models for agn .
|
liv@xmath9o@xmath10 is a rare @xmath11-electron heavy fermion system at low temperatures @xmath12 k.@xcite the low temperature linear electronic specific heat coefficient @xmath13 ( 0.42 j / mol k@xmath14 ) and pauli magnetic susceptibility @xmath15 ( @xmath16 @xmath17/mol ) are 180 and 310 times those of a free electron gas , respectively , assuming each vanadium atom contributes 1.5 free electrons .
the wilson ratio @xmath18 , which is the ratio of the enhancement factors of @xmath15 and @xmath13 , is equal to 1.7 , typical for a heavy fermion system.@xcite heavy fermion behavior was further confirmed by electrical resistivity measurements which show a @xmath19 dependence below 2 k with a large coefficient @xmath20 @xmath21 cm / k@xmath14.@xcite the @xmath22 and @xmath13 values approximately follow the kadowaki - woods relation , @xmath23 cm(mol k / j)@xmath14 , which holds for a variety of heavy fermion systems.@xcite despite continuous theoretical work , a detailed explanation of the heavy fermion behaviors in liv@xmath9o@xmath10 remains a challenge.@xcite @xmath6li nuclear magnetic resonance ( nmr ) was an important local probe in establishing the low temperature heavy fermion behavior in magnetically pure samples of liv@xmath9o@xmath10.@xcite the low temperature @xmath6li nuclear spin - lattice relaxation rate @xmath24 follows a korringa relation @xmath25 , with a coefficient @xmath26 s@xmath27k@xmath27 , which is 6000 times larger than in the non - heavy fermion isostructural superconducting@xcite compound liti@xmath9o@xmath10.@xcite the korringa ratio @xmath28k^2t_1t$ ] , where @xmath29 is the knight shift , @xmath30 and @xmath31 are the gyromagnetic ratios of the @xmath6li nuclear spin and the conduction electron spin , respectively , is equal to 0.7 , which is close to the value of unity expected for a free electron gas .
recently , we found that the low temperature @xmath6li nmr properties of polycrystalline liv@xmath9o@xmath10 are very sensitive to the presence of a small concentration of magnetic defects ( @xmath32 mol% ) within the spinel structure.@xcite in a sample containing a negligible concentration of magnetic defects , the longitudinal component of the global bulk @xmath6li nuclear magnetization @xmath33 after time delay @xmath34 following a sequence of saturation pulses showed a single exponential recovery @xmath35 $ ] , where @xmath24 was proportional to @xmath36 as noted above .
however , in the sample with @xmath32 mol% , the @xmath33 showed a stretched exponential recovery @xmath37 $ ] , with the characteristic relaxation rate @xmath38 showing a peak at @xmath39 k. here @xmath40 is the stretching exponent with , in general , @xmath41 .
there was also a clear difference in the @xmath6li nmr spectrum in these two samples . at low temperatures @xmath42
k , the magnetically pure sample had a narrow spectrum with an almost temperature independent width ( full width at half maximum peak intensity fwhm @xmath43 khz ) .
in contrast , a strong temperature dependent inhomogeneous broadening ( fwhm @xmath44 khz at @xmath42 k ) was observed in the sample with @xmath32 mol% . in order to further clarify the nature of the magnetic defects and their effect on the heavy fermion properties of liv@xmath9o@xmath10
, we report herein @xmath6li nmr studies on liv@xmath9o@xmath10 versus magnetic defect concentration .
three polycrystalline samples and a collection of single crystals are studied .
the powder samples are labeled as 6b , 7a , and 6a , with @xmath45 mol% , 0.49 mol% , and 0.83 mol% , respectively .
the single crystal sample is labeled as sample 1 with @xmath3 mol% .
we determined the magnetic defect concentrations from static magnetization measurements in the temperature range 1.85 k and applied magnetic field range 05.5 t.@xcite furthermore , to study the possible spin freezing of the magnetic defects at low temperatures , we measured the ac magnetic susceptibility at 14 mhz from 0.5 to 6 k of the single crystals and of the powder sample 6a with @xmath46 mol% using the tunnel - diode resonator technique.@xcite the temperature dependences of the @xmath6li nuclear spin - lattice relaxation rates in our polycrystalline samples are similar to that of sample 3 - 3-a2 ( @xmath32
mol% ) that we studied in ref . , which showed a peak in @xmath47 at about 1 k. however , we find a qualitative difference in the temperature dependence of @xmath48 in the collection of single crystals , which instead decreases monotonically with decreasing temperature from 4.2 k down to 0.5 k. we include two important aspects into the analysis of the nmr data .
first , we consider the effect of a distribution of local fields due to different positions of the @xmath0li nuclei relative to their nearby magnetic defects . for the polycrystalline samples ,
this approach is quantitatively consistent with the inhomogeneous broadening of the spectrum and the nonexponential relaxation behavior . in the single crystals ,
this purely geometric origin for the nuclear relaxation fails to explain the observed behavior at @xmath49 k. we then extend our analysis to take into account a possible size distribution of postulated magnetic defects of finite size ( magnetic droplets ) .
we speculate that the differences between the natures and interactions of the magnetic defects / droplets in the powders versus the crystals arise from the very different synthesis conditions and procedures of the powders and crystals .
the paper is organized as follows .
experimental details are given in sec.[scn : exp ] . in sec .
[ scn : results ] , we report the experimental results of the magnetization , ac susceptibility , @xmath0li nmr spectra , and @xmath0li nuclear spin - lattice relaxation rate measurements . in sec .
[ scn : models ] , we analyze the nmr results . in sec . [
scn : conclusions ] , we summarize the main conclusions of the paper .
polycrystalline liv@xmath4o@xmath5 samples were prepared using conventional solid state reaction at temperatures up to 700 @xmath50c .
the starting materials were v@xmath9o@xmath51 ( 99.99% , mv labs ) , v@xmath9o@xmath52 ( 99.99% , mv labs ) , and li@xmath9co@xmath51 ( 99.999% , alfa aeser ) .
details of the sample synthesis procedure can be found in ref . .
the typical size of the polycrystalline grains is in the range of 110 @xmath53m,@xcite as determined from scanning electron microscope ( sem ) micrographs .
single crystals were grown at 9501040 @xmath50c using a self - flux technique.@xcite the flux consisted of a mixture of li@xmath51vo@xmath10 and liv@xmath4o@xmath5 .
the typical size of the crystals used in the present work is 0.2 mm .
static magnetization measurements were performed using a quantum design squid magnetometer in the temperature range 1.8350 k and applied magnetic field range 05.5 t. the ac magnetic susceptibility was measured using a highly sensitive self - resonating @xmath54 circuit where losses are compensated by a tunnel diode that has a region of negative differential resistance in its @xmath55-@xmath56 characteristic .
the resonant frequency of an empty coil @xmath57 changes when a sample is placed in the coil .
the shift of the resonant frequency , @xmath58 is directly related to the dimensionless volume ac susceptibility @xmath59 of the sample via@xcite @xmath60 where @xmath61 is the sample volume and @xmath62 is the coil volume .
the volume magnetization is the magnetic moment per unit volume of the sample , with gaussian units g @xmath63 .
the volume susceptibility is the volume magnetization divided by field , which is then dimensionless .
the optimized and thermally stabilized circuit resonates at 14 mhz with a stability of @xmath64 over hours.@xcite the resonator was mounted in a @xmath65he cryostat with a temperature range 0.5150 k. a static external field up to 90 koe can be applied to study field - dependent properties .
@xmath0li nmr measurements were performed utilizing a phase - coherent pulse spectrometer at applied magnetic fields @xmath66 , 1.68 and 3.0 t and in the temperature range 0.54.2 k. measurements above 1.5 k were performed with a @xmath67he bath cryostat and measurements below 1.5 k with a janis @xmath65he cryostat .
the typical @xmath68 pulse length was @xmath69 .
the @xmath6li nmr spectra for narrow lines ( fwhm @xmath70 khz ) were measured by fourier transform of half the hahn echo signals , while for wider lines , the spectra were measured by integrating the echo area as a function of the applied magnetic field at a fixed frequency of rf pulses .
nuclear spin - lattice relaxation rates were measured by monitoring the recovery of the spin echo height using the standard saturation - recovery pulse sequence .
versus applied magnetic field @xmath71 isotherms at different temperatures @xmath36 for powder and crystals samples of liv@xmath9o@xmath10 .
( e ) the magnetic defect contributions @xmath72 to the data in panels ( a)(d ) versus @xmath73 .
the @xmath74 and @xmath75 values are listed in table [ tbl : mfit ] .
the solid lines are plots of the second term in eq .
( [ eqn : m ] ) , @xmath76 , versus @xmath73 with values of @xmath1 , @xmath75 , and @xmath77 given in table [ tbl : mfit].,width=316 ] the magnetic defect concentrations of the samples were determined from the low temperature ( @xmath78 k ) magnetization @xmath79 versus applied magnetic field @xmath71 isotherms.@xcite figures [ fig : mvsh](a ) , [ fig : mvsh](b ) , [ fig : mvsh](c ) , and [ fig : mvsh](d ) show the @xmath80 isotherms at different temperatures for samples 6b , 1 , 7a , and 6a , respectively .
the magnetic defect concentration @xmath1 and spin value @xmath77 of the magnetic defects in each sample are determined by fitting the equation @xmath81 to all the @xmath82 isotherm data for each sample at @xmath83 k.@xcite in eq .
( [ eqn : m ] ) , @xmath1 is the magnetic defect concentration in dimensionless mole fraction units , @xmath15 is a field- and temperature - independent contribution to the molar susceptibility at low temperatures @xmath83 k , @xmath84 is avogadro s number , @xmath85 is the ( powder - averaged ) spectroscopic splitting factor ( @xmath85-factor ) for the defect spins , @xmath86 is the brillouin function for spin @xmath77 , and @xmath87 $ ] where @xmath88 is boltzmann s constant .
we have replaced @xmath36 in the usual brillouin function by @xmath89 in order to take into account weak interactions between the magnetic defects .
@xmath15 , @xmath77 , @xmath90 , and @xmath1 are free parameters in the fit , whereas the @xmath85 value is fixed to be equal to 2 during the fit.@xcite the best fit parameters are listed in table [ tbl : mfit ] .
figure [ fig : mvsh](e ) shows the magnetic defect magnetization contributions @xmath91 versus @xmath92 for the four samples .
all the data points in figs .
[ fig : mvsh](a ) , [ fig : mvsh](b ) , [ fig : mvsh](c ) , and [ fig : mvsh](d ) fall onto a universal curve in fig .
[ fig : mvsh](e ) for each sample , respectively , as described by the second term in eq .
( [ eqn : m ] ) , thus confirming the consistency of the fits . the fitted functions @xmath93 in eq .
( [ eqn : m ] ) for the four samples are plotted versus @xmath94 as the solid curves in fig .
[ fig : mvsh](e ) and show excellent agreement with the data . .
best fit values of the magnetic defect concentration @xmath1 , the spin value @xmath77 , the intrinsic susceptibility @xmath95 , and the effective weiss temperature @xmath75 for powder samples 6b , 7a , and 6a and crystal sample 1 of liv@xmath9o@xmath10 obtained by fitting eq .
( [ eqn : m ] ) to the low temperature ( 1.8 k @xmath96 k ) magnetization versus field isotherms in figs .
[ fig : mvsh](a)[fig : mvsh](d ) with @xmath97 t. [ cols="^,^,^,^,^",options="header " , ] we fit the @xmath98 data in fig .
[ fig : t1d ] on all three powder samples simultaneously by the combination of eqs .
( [ eqn : acttau ] ) and ( [ eqn : lorentz2 ] ) . possible field and temperature dependences of the parameter @xmath99
are ignored in the fit .
there are seven free parameters in the fit , @xmath99 and @xmath100 for each sample and @xmath101 which is assumed to be sample independent .
the fitting results are displayed in fig .
[ fig : t1d ] by the solid curves .
the best fit value of @xmath101 is @xmath102 s and the best fit values of @xmath99 and @xmath100 for each sample are listed in table [ tbl : t1fit ] .
the energy barrier @xmath100 increases with increasing concentration of magnetic defects , which indicates that the dynamic slowing down with decreasing temperature originates from the interaction between the magnetic defects .
interaction between magnetic defects should increase with increasing concentration of the magnetic defects since the average nearest - neighbor distance decreases .
the values of @xmath99 in all three samples are much less than unity , a fact which can not be explained by the presence of other nuclear spin - lattice relaxation mechanisms since additional relaxation mechanisms would increase @xmath99 .
such small values of @xmath99 might be related to the spin - glass like freezing as observed in the ac magnetic susceptibility measurements . in spin glass systems ,
the spin autocorrelation functions are highly nonexponential,@xcite which reduces the spectral density of the magnetic defect spin fluctuations at @xmath103 as compared to the lorentzian in eq .
( [ eqn : lorentz2 ] ) .
the reduction in spectral density thus results in a reduction in the fitted value of @xmath99 in eq .
( [ eqn : lorentz2 ] ) .
our previous discussion demonstrated that the non - exponential time dependence of the @xmath6li nuclear spin - lattice relaxation in the sample of single crystals below 1.3 k can not be understood in terms of the interaction of the nuclear spins with separable contributions from a heavy fermi liquid and a random distribution of point - like magnetic defects .
this problem is exemplified by the unphysical @xmath104 data at long times in fig .
[ fig : ptxtal](a ) where the data move away from equilibrium rather than towards equilibrium with increasing time @xmath34 .
an appealing approach that refines our previous analysis is based on the hypothesis that the magnetic defects are not point - like , but are rather magnetic droplets with an average size significantly larger than an atomic size .
this hypothesis is supported by the large average spins of the magnetic defects inferred above in sec .
[ sec : ndefect ] and previously@xcite to be @xmath1054 .
we envision that the magnetic droplets have a variable size and spin and that the corresponding microscopic internal relaxation time @xmath106 varies with droplet size .
our arguments parallel those for the unusual spin dynamics due to statistically rare fluctuations that becomes crucial in the context of griffiths singularities close to phase transitions.@xcite in most three - dimensional systems a high temperature curie - weiss behavior of the susceptibility with weiss temperature @xmath90 is indicative for magnetic ordering at a lower temperature @xmath107 .
however , the geometric frustration for antiferromagnetic ordering within the vanadium sublattice of the spinel structure is likely the reason why long - range antiferromagnetic order is suppressed in pure @xmath4@xmath5 and a heavy electron state emerges instead .
defects in the crystal structure can locally lift the frustration and easily cause magnetic order in a finite region of volume @xmath108 around the crystal defect where the linear size @xmath109 is , due to the proximity to an ordered state , expected to be larger than the interatomic spacing . here
@xmath11 is the dimensionality of the system .
these finite regions of magnetic order are what we are calling magnetic droplets .
fluctuations in the local tendency towards order usually lead to a probability density for the linear droplet size @xmath109 that decreases exponentially with the volume of the droplet . in three dimensions one obtains@xcite @xmath110,\ ] ] where @xmath111 is the mean droplet size .
it is then natural to assume that the typical internal excitation energy of a droplet decreases with increasing droplet size .
these excitations can change the magnitude and/or direction of the magnetic moment of the droplet .
often the excitation energy @xmath112 varies with @xmath109 according to a power law @xmath113 where we have set @xmath114 and @xmath106 is the relaxation time of the excitation .
the arguments of ref .
yield a result @xmath115 for classical heisenberg spins , where @xmath116 for the present problem .
quantum effects can then yield deviations from this behavior and typically yield exponents @xmath117 that are larger than the above classical result.@xcite in some cases , such as ising spins in a magnetic field or heisenberg spins in a metallic host,@xcite the quantum dynamics can lead to a droplet dynamics in three dimensions where @xmath118 finally , in case of magnetic defects with ising anisotropy inside a metallic host , the quantum dynamics of the defect leads to a freezing of all droplets beyond a certain size , typically of order @xmath119 ( see refs . and ) . to illustrate the effects of such droplet dynamics we concentrate first on the case eq . ( [ power ] ) . in the absence of a microscopic model of the droplet spin dynamics
we leave @xmath120 an open parameter of the model .
we start from eqs .
( [ eqn : rootexp ] ) and ( [ eqn : t1d ] ) and write @xmath121 , \label{eqpave}\ ] ] where @xmath122 is given in eq . ( [ eqn : lorentz2 ] ) with @xmath106 replaced by @xmath123 . in the evaluation of this average
we have two very distinct limits .
if @xmath124 the long time behavior of the system is dominated by small clusters since @xmath125 and slow nuclear relaxation is caused by fast droplet dynamics . in this limit , the average over droplet sizes will not cause any changes in the stretched exponential behavior as compared to our previous , purely geometric , considerations for point - like magnetic defects .
the inequality @xmath124 is expected to be valid at higher temperatures , before a freezing or dramatic slowing down of the droplet moments sets in with decreasing temperature .
this is fully consistent with our findings that we obtain @xmath126 in this regime .
the situation changes dramatically in the limit where @xmath127 , relevant at lower temperature , i.e. most likely for @xmath128 .
now @xmath129 and slow nuclear relaxation is tied to slow droplet relaxation . in this regime
we find @xmath130 in the long time limit , the average over the droplet sizes in eq .
( [ eqpave ] ) can be performed via saddle point integration and yields @xmath131\ ] ] with @xmath132 i.e. , the additional inclusion of droplet size variations yields a stretched exponent @xmath133 .
the characteristic relaxation rate of the droplets is @xmath134 we emphasize that the static size distribution of magnetic droplets only becomes apparent by nmr at low temperatures such that @xmath135 , where @xmath136 is the nuclear larmor frequency .
we also emphasize that the analyses in this section do not include interactions between droplets and a possible resultant spin glass phase . above @xmath137 size variations of the droplets
will not affect the long time nuclear relaxation , even if they are present . in case of an exponential dependence of the droplet excitation energy , eq .
( [ exp ] ) , the long time dynamics of the nuclear spins at low @xmath36 is even more dramatically affected and changes to a power law decay @xmath138 , with nonuniversal exponent @xmath139 . at the same time heat capacity and
susceptibility measurements also experience power law behavior.@xcite in this regime droplet quantum fluctuations will not only dominate the long time dynamics of the nuclear decay , but also thermodynamics quantities
. this might be responsible for the fact that it is not possible any longer to clearly separate the response of the underlying fermi liquid from that of the droplet at low temperatures in our single crystals , as indicated by fig .
[ fig : ptxtal](a ) . our data for
single crystals do not allow at present to distinguish between the different scenarios outlined in this section .
they do however suggest that dynamics of the magnetic droplets plays an important role .
the modeling approach in sec .
[ sec : geometry ] gives a good description of our @xmath6li nmr results from 0.5 to 4.2 k for our liv@xmath4o@xmath5 powder samples containing magnetic defect concentrations up to 0.83 mol% .
this approach assumes that ( i ) there is a random distribution of magnetic point defects , ( ii ) the heavy fermi liquid in magnetically pure liv@xmath4o@xmath5 survives in samples containing up to @xmath140 mol% magnetic defects , and that ( iii ) the influences of the magnetic defects and of the fermi liquid on the magnetization and nmr properties are separable .
this description explains very well the defect concentration - independent @xmath15 value from our low - temperature magnetization measurements , the inhomogeneous broadening of the @xmath0li nmr spectrum , the nonexponential @xmath6li nuclear spin - lattice relaxation versus time behavior , and the lack of spectral diffusion in the @xmath6li nmr hole burning experiments .
it also explains the smaller @xmath6li nuclear spin - lattice relaxation rate at the peak of the spectrum as compared to that at the wings .
however , it is hard to reconcile the picture of magnetic point defects with the high magnetic moments for the defects ( spins of 24 ) deduced here ( see table [ tbl : mfit ] ) and in refs . and from magnetization measurements .
these large defect spin values suggest that the magnetic defects may not behave like point - like magnetic moments under all circumstances . in sec . [
sec : droplets ] we discussed magnetic defects that are more extended entities that we have called magnetic droplets with a distribution of sizes , and likely a distribution of spin values .
we showed that such a size distribution can affect the nmr magnetization recovery at long times at low temperatures in the regime where interactions between the magnetic droplets can be neglected .
our study shows that there can be different kinds of magnetic defects in the liv@xmath9o@xmath10 system . as revealed by the nuclear spin lattice relaxation rate data and ac magnetic susceptibility measurements at 14 mhz , it is amazing that the magnetic defects in the powder samples undergo a spin glass - like freezing below 1 k , while the magnetic defects in the single crystals with a similar magnetic defect concentration exhibit a very different behavior at such low temperatures , with no evidence for spin freezing .
the different kinds of magnetic defects and/or interactions in the crystals and powders must be associated with different types of structural defects in the system , which might be expected because the crystals are grown at about 1000 @xmath50c whereas the powders are synthesized at 700@xmath50c .
different types of magnetic defects were even found in an annealing study of different single crystals,@xcite where heat treatment at 700 @xmath50c was found to remove the magnetic defects in one but not in other single crystals .
further experiments on the single crystals are urgently needed at @xmath141 k and below .
these experiments should test whether the fermi liquid is modified by quantum fluctuations of large magnetic droplets in the single crystals at @xmath142 k , whether the magnetic properties of the crystals contain separable contributions from the fermi liquid and the magnetic defects , and whether the conduction electrons in the crystals even form a fermi liquid .
in addition , the origin of the distinct kinks at about 1.4 k in the temperature dependences of @xmath38 and @xmath40 in fig . [ fig : it1beta ] and of @xmath143 in fig . [ fig : it1beta - xstal ] for the single crystals remains to be explained .
we are grateful to b. j. suh for many beneficial suggestions .
work at the ames laboratory was supported by the department of energy - basic energy sciences under contract no .
de - ac02 - 07ch11358 .
here we discuss the stretched exponential @xmath6li nuclear relaxation versus time following saturation that arises from interactions between the @xmath6li nuclear spins and a low concentration of magnetic defect spins , which is @xmath104 in eq .
( [ eqn : rootexp ] ) .
we demonstrate that instead of theoretically solving for the relaxation curve and showing that it is a stretched exponential with @xmath145 as in sec .
[ sec : nslr ] , we can understand the occurrence of this root exponential relaxation as arising from the probability distribution of nuclear @xmath24 values .
we assume that @xmath104 in eq .
( [ eqn : rootexp ] ) is due to a continuous sum of exponential decays with a distribution of relaxation rates @xmath24
. then one can write the stretched exponential relaxation function in eq .
( [ eqn : rootexp ] ) as @xmath146 where @xmath147 equals @xmath24 normalized by @xmath148 , i.e. , @xmath149 @xmath150 , and @xmath151 is the probability density for occurrence of @xmath147 for a fixed exponent @xmath40 with @xmath152 .
thus the stretched exponential function is the laplace transform of @xmath151 .
closed analytic expressions for @xmath151 with rational values of @xmath40 can be obtained from the inverse laplace transform of the stretched exponential function , and physical interpretations of the parameters @xmath98 and @xmath40 have been determined.@xcite we show below that the probability distribution of @xmath24 due to dipolar interaction of nuclear spins with dilute magnetic defects corresponds very well to the @xmath24 distribution leading to the stretched exponential relaxation in eqs .
( [ eqn : rootexp ] ) and ( [ eqlaplace ] ) with @xmath153
. this probability density is @xcite @xmath154 a plot of this distribution function is given below as the solid curve in fig .
[ fig : t1distb0p5 ] .
@xmath155 is proportional to @xmath156 for large @xmath147 [ in general @xmath151 at large @xmath147 is proportional to @xmath157 , and has a low-@xmath147 cutoff since @xmath158 exponentially approaches zero at small @xmath147 values ( also true in general for arbitrary @xmath40).@xcite a qualitative @xmath156 dependence of the @xmath24 probability distribution arises due to a @xmath159 dependence of @xmath24 as in eq .
( [ eqn : c ] ) as follows .
ignoring the angular dependence in eq .
( [ eqn : c ] ) and assuming the same dynamics for all the magnetic defect spins such that @xmath160 , in the single paramagnetic center limit the distribution of @xmath147 arising from a _ continuum _ distribution of nuclear spins around a magnetic defect is@xcite @xmath161 this @xmath147 dependence is the same as the large-@xmath147 limit of the result in eq .
( [ eqn : pbeta12 ] ) for stretching exponent @xmath145 noted above .
the distribution ( [ eqn : psnglm ] ) diverges as @xmath24 approaches zero .
this divergence is caused by the single impurity approximation .
nuclei with @xmath162 approaching zero correspond to those far away from the paramagnetic center . due to the finite distance between different paramagnetic centers in a system with a finite concentration of them ,
the actual probability of finding a nuclear spin with @xmath163 should instead vanish , so a low @xmath24 cutoff has to be applied .
this results in a distribution function with a shape ( see the dashed curve in fig . [
fig : t1distb0p5 ] below ) roughly similar to that in eq .
( [ eqn : pbeta12 ] ) ( the solid curve in fig .
[ fig : t1distb0p5 ] below ) .
li nuclear spin - lattice relaxation rate @xmath24 probability distributions @xmath164 , normalized by @xmath165 , due to dilute point - like paramagnetic defects .
the filled circles are results obtained from computer simulation of liv@xmath4o@xmath5 where the magnetic defects randomly occupy vanadium sites with probability 0.25% .
the solid curve is the best fit of these data by eq .
( [ eqn : pbeta12 ] ) with @xmath166 s@xmath27 .
the dotted curve is a plot of @xmath167 with a small @xmath168 cutoff at @xmath169 , where @xmath170 .
the lower cutoff @xmath171 is chosen so that the normalization condition @xmath172 is satisfied.,width=288 ] we have carried out a numerical simulation of the @xmath24 probability distribution that turns out to be in very good agreement with the exact result for the probability distribution in eq .
( [ eqn : pbeta12 ] ) for the stretched exponential function with @xmath145 . in the simulation , we calculated the @xmath24 distribution of @xmath6li nuclei due to a random distribution of dilute point - like paramagnetic centers ( defects ) in the liv@xmath9o@xmath10 spinel structure .
the paramagnetic defects randomly occupy the vanadium sites with a probability of @xmath173 ( @xmath174 mol% ) and the configuration of the random defects repeats every 80 unit cells in all three crystallographic axis directions .
the @xmath24 of each @xmath6li nucleus is calculated using@xcite @xmath175 where @xmath176 is a constant , @xmath177 is the distance between paramagnetic center @xmath178 and the @xmath6li nucleus , and @xmath179 is the angle between the applied magnetic field and the vector from paramagnetic center @xmath178 to the @xmath6li nucleus . the applied magnetic field was arbitrarily chosen to be along the @xmath180 direction . equation ( [ eqn : t1sim ] ) is the nuclear spin - lattice relaxation due to the dipolar magnetic field fluctuations from the _ longitudinal _ spin component of the paramagnetic defects.@xcite in the presence of a strong applied magnetic field , the _ transverse _ spin fluctuation is often modulated by the larmor frequency of the _ electronic _ spins and thus has negligible contribution to @xmath24 for which magnetic field fluctuations at the _ nuclear _ larmor frequency are most important.@xcite the summation over @xmath178 in eq .
( [ eqn : t1sim ] ) includes all magnetic defects with @xmath181 , where @xmath182 is the lattice constant of cubic liv@xmath9o@xmath10 .
we checked that changing the summation range to @xmath183 resulted in a negligible change in the @xmath24 distribution .
the distribution of @xmath24 resulting from the above simulation is displayed as the filled circles in fig .
[ fig : t1distb0p5 ] , where the distribution is normalized by @xmath165 . the maximum relaxation rate plotted in fig .
[ fig : t1distb0p5 ] at @xmath184 is not a large relaxation rate cutoff to the probability distribution .
data at larger relaxation rates are not plotted due to insufficient statistics . the simulated @xmath24 distribution in fig . [ fig : t1distb0p5 ]
can be fitted very well by eq .
( [ eqn : pbeta12 ] ) with @xmath185 , as shown by the solid curve in fig .
[ fig : t1distb0p5 ] .
@xmath98 calculated from eq .
( [ eqn : t1d ] ) is equal to @xmath186 , close to the simulated result .
the difference may be due to the neglected angular dependence in deriving eq .
( [ eqn : t1d ] ) . for comparison
, a simple power law distribution @xmath187 with a small @xmath168 cutoff of @xmath188 is also displayed in fig .
[ fig : t1distb0p5 ] .
this @xmath168 dependence is the same as in eq .
( [ eqn : psnglm ] ) and is the same as the asymptotic large-@xmath147 dependence of eq .
( [ eqn : pbeta12 ] ) .
the prefactor 0.08 is chosen to make the distribution overlap with the simulated result at large @xmath168 , and the low-@xmath168 cutoff @xmath188 is determined from the normalization condition @xmath189 .
s. kondo , d. c. johnston , c. a. swenson , f. borsa , a. v. mahajan , l. l. miller , t. gu , a. i. goldman , m. b. maple , d. a. gajewski , e. j. freeman , n. r. dilley , r. p. dickey , j. merrin , k. kojima , g. m. luke , y. j. uemura , o. chmaissem , and j. d. jorgensen , phys .
* 78 * , 3729 ( 1997 ) .
|
we present a systematic study of the variations of the @xmath0li nmr properties versus magnetic defect concentration @xmath1 within the spinel structure of polycrystalline powder samples ( @xmath2 , and 0.83 mol% ) and a collection of small single crystals ( @xmath3 mol% ) of liv@xmath4o@xmath5 in the temperature range from 0.5 to 4.2 k. we also report static magnetization measurements and ac magnetic susceptibility measurements at 14 mhz on the samples at low temperatures .
both the @xmath6li nmr spectrum and nuclear spin - lattice relaxation rate are inhomogeneous in the presence of the magnetic defects .
the @xmath6li nmr data for the _ powders _ are well explained by assuming that ( i ) there is a random distribution of magnetic point defects , ( ii ) the same heavy fermi liquid is present in the samples containing the magnetic defects as in magnetically pure liv@xmath4o@xmath5 , and ( iii ) the influences of the magnetic defects and of the fermi liquid on the magnetization and nmr properties are separable . in the _ single crystals _
, somewhat different behaviors are observed , possibly due to a modification of the heavy fermi liquid , to a lack of separability of the relaxation effects due to the fermi liquid and the magnetic defects , to non - fermi liquid behavior of the conduction electrons , and/or to quantum fluctuations of finite - size magnetic defects ( magnetic droplets ) .
remarkably , the magnetic defects in the powder samples show evidence of spin freezing below @xmath7 k , whereas in the single crystals with similar magnetic defect concentration no spin freezing was found down to @xmath8 k. thus different types of magnetic defects and/or interactions between them appear to arise in the powders versus the crystals , possibly due to the substantially different synthesis conditions of the powders and crystals .
|
since onsager s solution @xcite in the transfer matrix approach @xcite of the two - dimensional ising model @xcite with vanishing magnetic field and its subsequent simplification by kaufman @xcite , there have been a number of related as well as alternative solutions , see e.g. baxter s book on exactly solved models in statistical mechanics @xcite and references therein . among the transfer matrix solutions
are the ones by schultz , mattis and lieb , by thompson , by baxter and by stephen and mittag @xcite .
nevertheless , the author of this work feels that there is still room for a nice and straightforward solution .
a much abbreviated version of this work , sketching only the new aspects may be found in @xcite , while here we give a self - contained account of our approach with all relevant details .
we study the two - dimensional ising model with zero magnetic field on a square lattice with @xmath2 rows and @xmath3 columns subject to toroidal boundary conditions .
the transfer matrix is expressed in terms of the generators of two commuting representations of the complex rotation group so@xmath1 .
these representations naturally arise from projected bilinears of @xmath4 spin matrices . conservatively speaking , we reduce kaufman s approach to its essential steps , avoiding in particular the doubling of the number of eigenvalues of the transfer matrix and subsequent rather involved arguments for the choice of the correct ones . we also need not investigate the transformation properties of the spin matrices . in our opinion , the value of this work is not that it gives the most concise solution , but that it provides a clear line of arguments and requires only rather basic tools from mathematics and mathematical physics .
we proceed in a step - by - step fashion , exhibiting also those steps that closely parallel the corresponding steps in kaufman s treatment @xcite or huang s corresponding textbook write up @xcite . on the one hand
this is necessary because of our slightly different ( but , in our opinion , more natural ) conventions , in particular for @xmath5 . on the other hand
, we would like to keep this work as self - contained as possible . finally , this work may help sharpen the view on the difficulties encountered when attempting to solve the two - dimensional model with magnetic field or the three - dimensional model and might be useful for simplifying other problems in statistical mechanics for which an exact solution is attempted .
our notation is in the spirit of @xcite , with sans serif capitals reserved for @xmath4 matrices .
the structure of this work is as follows : in section [ def ] the model and its transfer matrix @xmath6 are defined . in section [ xyz ]
we express the transfer matrix in terms of @xmath4 spin matrices @xmath7 , @xmath8 , @xmath9 .
a rescaled transfer matrix @xmath10 is defined whose eigenvalues are , up to a trivial factor , the eigenvalues of @xmath6 . in section [ gammaandj ] , we define further spin matrices @xmath11 and two commuting projection classes @xmath12 and @xmath13 of their bilinears . after investigating the relevant properties of the @xmath14 , we express @xmath10 in terms of them . in section [ vpmdiag ] ,
we introduce @xmath15 matrices @xmath16 whose algebra is identical to that of @xmath14 and define matrices @xmath17 in terms of the @xmath16 such that the relation between @xmath17 and @xmath16 is closely related to that between @xmath10 and @xmath14 .
the @xmath17 are subsequently diagonalized . in section [ vdiag ] ,
the analogy between @xmath17 and @xmath10 is exploited for the diagonalization of @xmath10 .
the eigenvalues of @xmath10 and the partition function are found explicitly . in section [ extensions ]
we comment on the technical difficulties encountered when attempting to solve the two - dimensional model with magnetic field or the three - dimensional model .
we work with a square lattice with @xmath2 rows and @xmath3 columns and consequently @xmath0 sites .
@xmath18 the energy is given by e = e_a+e_b , -e_a & = & a_ij_as_i s_j = a_=1^m_=1^ns_s_+1 , , -e_b & = & b_ij_bs_i s_j = b_=1^m_=1^ns_s_,+1 , with @xmath19 , @xmath20 and @xmath21 , where @xmath22 and @xmath23 are temperature - independent interaction energy parameters which by convention are negative ( other sign combinations do not change the partition function , as can be seen by appropriate redefinitions of half of the signs of the spin variables ) .
consequently , @xmath24 and @xmath25 are real and positive . @xmath26 and @xmath27 indicate summation over neighboring spins in the respective directions . we identify rows @xmath28 and @xmath29 and columns @xmath28 and @xmath30 , i.e.the lattice is wrapped on a torus ,
see also fig .
[ twodimlattice ] .
the @xmath31 can take the values @xmath32 . the partition function
is then given by [ psum ] + & = & _ s_11_s_mn + & = & _ s_11_s_mn_=1^m_=1^n ( as_s_+1,+bs_s_,+1 ) .
it can be expressed with the help of a @xmath4 transfer matrix @xmath6 , + & = & _ \{_i } _
1||_m_m||_m-1_3||_2_2||_1 + & = & ^m , where @xmath33 for @xmath34 and @xmath6 is defined by its elements ( @xmath35 ) , ||=_=1^n(as_s_+bs_s_+1 ) .
we can split @xmath6 into a product of two matrices [ t ] = , defining @xmath36 and @xmath37 by their elements
with the help of the pauli matrices _
x=(&1 + 1 ) , _ y=(&-i +
i ) , _ z=(1 + & -1 ) and the @xmath38 unit matrix @xmath39 , define hermitian @xmath4 spin matrices by the direct products _ = _
-1_x_n- , and analogously for @xmath8 and @xmath9 . for any @xmath40 and @xmath41
, they obey the commutation relations = [ _ , _ ] = [ _ , _
]=0 , while only for @xmath42 holds = [ _ , _ ] = [ _ , _ ]=0 . for any @xmath40 we have _
^2=_^2=_^2= , \{_,_}=\{_,_}=\{_,_}=0 , _ _
= i _ , _ _ =
i _ , _ _ =
i _ , where @xmath39 is now the @xmath4 unit matrix .
define @xmath43 by ( 2)(2a)=1 , so that @xmath44 , = ( -2a),a=(-2 ) and ( e^+a&e^-a + e^-a&e^+a ) = [ 2(2a)]^1/2(_x ) . then we can write = [ 2(2a)]^n/2with [ va ] = _ = 1^n ( _ ) , and [ vb ] = _
= 1^n(b__+1 ) , where we have identified @xmath45 . the transfer matrix ( [ t ] )
may then be expressed as = [ 2(2a)]^n/2 . due to the cyclic property of the trace , we may rewrite the partition function ( [ psum ] ) as z(a , b)=[2(2a)]^mn/2^m , where @xmath10 is defined by the hermitian matrix [ v ] = _ a/2_a/2 with [ vahalf ] _
a/2=_=1^n(_/2 ) , so that @xmath46 . if @xmath47 are the @xmath48 eigenvalues of @xmath10 , we have z(a , b)=[2(2a)]^mn/2_k=1 ^ 2^n_k^m . our task is therefore to find the eigenvalues of @xmath10 .
define the @xmath49 matrices ( @xmath50 ) _ 2 - 1 & = & _ 1_-1 _ , + _ 2 & = & _ 1_-1 _ , which obey [ gaga ] \{_,_}=2_. define further the matrix [ ux ] _ = _
1_n = i^n_1_2_2n , _
^2= , which anticommutes with every @xmath51 , \{_,_}=0 .
now we can write for the matrices appearing in the exponents of ( [ vb ] ) and ( [ vahalf ] ) [ x ] _ & = & -[_2,_2 - 1 ] , = 1, ,n , + [ zz1 ] _ _ + 1&=&-[_2 + 1,_2 ] , = 1, ,n-1 , + [ zz2 ] _
n_1&=&_. so far our treatment has been rather similar to huang s write up @xcite of kaufman s approach @xcite .
our subsequent treatment rests on the observation that the formulation ( [ v ] ) of @xmath10 with ( [ vb ] ) and ( [ vahalf ] ) involves only the product @xmath52 of all @xmath11 and bilinears @xmath53 , see ( [ ux ] ) , ( [ x])-([zz2 ] ) .
this will allow us to find two commuting algebras of projected bilinears of the @xmath11 .
define the projectors ^ ( _ ) , for which hold @xmath54 @xmath55 and = 0 . with their help ,
define the matrices [ jgaga ] _
= - [ _ , _ ] , _ ^=^ _ ,
so that [ jprops ] _ = _
^++_^- , _ _ ^=_^. since _
^=-_^ , there are @xmath56 such independent matrices of each kind @xmath12 and @xmath13 .
it is straightforward to show that their algebra decomposes into two commuting parts , = 0 , which obey identical algebras [ jalgebra ] [ _ ^,_^ ] = i(__^+__^-__^-__^ ) .
next note that with ( [ x])-([zz2 ] ) , ( [ jgaga ] ) and ( [ jprops ] ) we can write _ & = & 2(_2,2 - 1^++_2,2 - 1 ^ - ) , = 1, ,n , + _ _ + 1&=&2(_2 + 1,2^++_2 + 1,2 ^ - ) , = 1, ,n-1 , + _
n_1&=&-2_(_1,2n^++_1,2n^- ) = -2(_1,2n^+-_1,2n^- ) .
this allows us to express @xmath57 from ( [ vahalf ] ) and @xmath37 from ( [ vb ] ) in terms of the @xmath14 , _
a/2=_=1^n=_a/2^+_a/2 ^ - with [ vapm ] _
a/2^=_=1^n(_2,2 - 1^ ) , and [ vb3 ] & = & + & & _
= 1^n-1 + & = & ^+^- with [ vbpm ] ^=(2b_1,2n^ ) _ = 1^n-1(2b_2 + 1,2^ ) .
the rescaled transfer matrix @xmath10 defined in ( [ v ] ) reads then = ^+^- with [ tvv ] ^=_a/2^^_a/2^ , [ ^+,^-]=0 .
in this section we formulate matrices @xmath58 whose definitions are similar to the expressions of @xmath59 through ( [ vapm ] ) , ( [ vbpm ] ) and ( [ tvv ] ) . however , the @xmath17 have convenient periodicity properties which allow for explicit diagonalization .
our treatment closely parallels that of kaufman @xcite as written up by huang @xcite .
however , we will be more explicit and also carefully analyze some subtleties of the analysis concerning the representation of the relevant similarity transformation matrices @xmath60 and the sign of @xmath5 , see below
. define @xmath61 matrices @xmath16 by their elements ( j_)_ij = -i(_i_j-_i_j ) , where greek and latin indices run from 1 to @xmath62 . since @xmath63 , there are @xmath64 such independent matrices .
as can be easily checked , they obey the algebra [ sonalgebra ] [ j_,j_]=i(_j_+_j _ -_j_-_j _ ) , which is , if we set @xmath65 , identical to the algebras ( [ jalgebra ] ) of @xmath12 and @xmath13 .
now consider the matrices s=(ic_j _ ) , for which holds s^t & = & ( ic_j_)^t=(ic_j_^t ) + & = & ( -ic_j_)=s^-1 . since ( s)^2=ss^t = ss^-1=1 , and because @xmath66 is smoothly connected to the unit matrix , we have @xmath67 .
for real parameters @xmath68 , the @xmath66 are also real and form the group so@xmath69 of @xmath61 orthogonal matrices with unit determinant . the algebra ( [ sonalgebra ] ) is therefore called the lie algebra of so@xmath69 . here
, we let the @xmath68 be arbitrary complex numbers , so the matrices @xmath66 form the group so@xmath70 of complex @xmath61 matrices with s^t = s^-1 , s=1 .
define the so@xmath1 matrices [ vpm ] v^=v_a/2v_b^v_a/2 with v_a/2=_=1^n(j_2,2 - 1 ) and
v_b^=(2b j_1,2n ) _ = 1^n-1(2b j_2 + 1,2 ) , in analogy with @xmath59 , @xmath71 and @xmath72 in ( [ tvv ] ) , ( [ vapm ] ) and ( [ vbpm ] ) . since @xmath43 and @xmath25 are real , the matrices @xmath73 , @xmath74 and @xmath17 are not only orthogonal , but also hermitian , so the @xmath17 have only real eigenvalues and in each case a complete set of orthonormal eigenvectors .
since the algebras of the three sets of matrices @xmath12 , @xmath13 and @xmath16 are identical and because @xmath17 will turn out to be explicitly diagonalizable , we will be able to find all eigenvalues of @xmath10 .
the structure of @xmath73 and @xmath74 is given by v_a/2= ( & & & + & & & + & & & + & & & ) with = ( & i + -i & ) and v_b^= ( ( 2b)&&&&i2b + & & & & + & & & & + & & & & + i2b&&&&(2b ) ) with = ( 2b&i2b + -i2b&2b ) .
carrying out the matrix multiplication in ( [ vpm ] ) is straightforward and gives v^= ( a&b&0&&0&b^ + b^&a&b&&0&0 + 0&b^&a&&0&0 + & & & & & + 0&0&0&&a&b + b&0&0&&b^&a ) with a=2b(2&i2 + -i2&2 ) and b=2b(-2&-i^2 + i^2&-2 ) .
let us make the following ansatz for the normalized eigenvectors of @xmath17 , [ psi ] = ( zu + z^2u + + z^nu ) , where @xmath75 is a complex number and @xmath76 is a two - component vector that we assume to be normalized , @xmath77 .
the condition v^=leads to the @xmath3 equations ( za+z^2bz^nb^)u&=&zu , + ( z^2a+z^3b+zb^)u&=&z^2u , + ( z^3a+z^4b+z^2b^)u&=&z^3u , + & & + ( z^n-1a+z^nb+z^n-2b^)u&=&z^n-1u , + ( z^nazb+z^n-1b^)u&=&z^nu . the equations with @xmath78 through @xmath79 on the right hand side are identical , which leaves three independent equations [ abbeq1 ] ( a+zbz^n-1b^)u&=&u , + ( a+zb+z^-1b^)u&=&u , + [ abbeq3 ] ( az^1-nb+z^-1b^)u&=&u .
these equations can simultaneously be solved when [ z ] z^n=1 .
since then they become identical , [ laeq ] ( a+zb+z^-1b^)u = u . the sign @xmath80 in ( [ z ] ) is associated with @xmath17 .
altogether , there are @xmath49 values of @xmath75 that solve ( [ z ] ) , [ zk ] z_k = e^ik / n , k=0, ,2n-1 .
even @xmath81 lead to @xmath82 while odd @xmath81
lead to @xmath83 , i.e.k=1,3,5, ,2n-1 & & , + k=0,2,4, ,2n-2 & & . for each @xmath84
, we still have to find the associated two eigenvalues @xmath85 and @xmath86 and their corresponding eigenvectors @xmath87 and @xmath88 in ( [ laeq ] ) , i.e. in [ evecs ] m_ku_k^=_k^u_k^with the hermitian matrix m_k = a+e^ik / nb+e^-ik / nb^=(d_k&o_k + o_k^*&d_k ) with [ dk1 ] d_k=22b-22b and [ ok1 ] + & & + i(22b -22b ) .
explicit evaluation gives [ detm ] m_k = d_k^2-o_ko_k^*=1 , implying [ up to an overall sign , fixed by @xmath89 , see ( [ dk1 ] ) above ] eigenvalues of the form [ lak ] _
k^=e^+_k , _ k^=e^-_k with real @xmath90 . the value of @xmath90 can be found from m_k = e^+_k+e^-_k = 2_k .
so that [ gak ] _
k=22b-22b , and from ( [ detm ] ) follows then that we may write [ ok2 ] o_k = ie^i_k_k with some phase @xmath91 to be considered further below .
if @xmath90 is a solution , then also @xmath92 is one , but this has already been taken into account in ( [ lak ] ) .
let us fix the sign of @xmath90 by defining @xmath93 for @xmath94 and then analytically continuing to other values of @xmath25 .
for @xmath95 , this means @xmath96 . on the other hand , for @xmath5
this means [ ga0 ] _ 0=2(-b ) .
our sign convention for the @xmath90 and in particular for @xmath5 will allow us to treat all @xmath90 on an equal footing for both @xmath97 and @xmath98 , i.e. irrespective of the temperature .
@xmath90 for @xmath99 and @xmath100 as a function of @xmath24 in the model with @xmath101 .
the solid lines from the top are @xmath102 , @xmath103 , , @xmath5 , while the dashed lines from the bottom are @xmath104 , @xmath105 , @xmath106 , @xmath107 .
apart from @xmath108 and @xmath109 , all @xmath110 are two - fold degenerate . for @xmath111 ,
the lines merge into two continua above and below @xmath112.,title="fig:",width=302 ] @xmath90 for @xmath99 and @xmath100 as a function of @xmath24 in the model with @xmath101 .
the solid lines from the top are @xmath102 , @xmath103 , , @xmath5 , while the dashed lines from the bottom are @xmath104 , @xmath105 , @xmath106 , @xmath107 .
apart from @xmath108 and @xmath109 , all @xmath110 are two - fold degenerate . for @xmath111 ,
the lines merge into two continua above and below @xmath112.,title="fig:",width=302 ] it is obvious from ( [ dk1 ] ) that for @xmath113 [ gaksym ] _
k=_2n - k , i.e. the eigenvalues @xmath114 with @xmath113 occur in pairs .
since for @xmath115 [ extending for the moment ( [ gak ] ) to non - integer values of @xmath81 ] , = > 0 , we have 0<|_0|<_1<b , + 0<-_0<_1<0 , which leaves only s_=1 .
this means that the matrices @xmath60 are elements of so@xmath1 and can be represented as s_=(ic_^j _ ) with certain unknown complex parameters @xmath117 .
let us summarize the crucial result of this section : applying similarity transformations @xmath118 with certain matrices @xmath119 to @xmath120 with @xmath121 and @xmath122 , we obtain @xmath123 with @xmath90 as defined in ( [ gak ] ) and the subsequent sign convention .
this is what we will use in the next section to determine all eigenvalues of @xmath10 and therefore the partition function @xmath124 .
now use the same parameters @xmath117 from the last section to define the @xmath4-dimensional transformation matrix = _ + _ - , _
= ( ic_^_^ ) , and write _
= ^-1=_+^+_+^-1_-^-_-^-1 _ s^+_s^- . the factors _
s^=(ic_^_^)_a/2^^_a/2^(-ic_^_^ ) have the same structure as v_s^=(ic_^j_)v_a/2v_b^v_a/2 ( -ic_^j _ ) from the last section .
now imagine using the baker - campbell - hausdorff formula @xcite + & & + \{[a,[a , b]]+[b,[b , a ] ] } + ) to work out all products of exponentials in @xmath125 . since the @xmath12 , @xmath13 and @xmath16 obey identical algebras , the result is _
^=(_=1^n _ _ 2,2 - 1^ ) , so that _ & = & + & = & with the same @xmath90 as defined in ( [ gak ] ) and the subsequent sign convention . to diagonalize @xmath126 , define another similarity transformation _
= _ _ _ ^-1 with @xmath127 and its inverse given by _ ^1=2
^ -n/2_=1^n(i _ ) .
since _ _ _
^-1= _ , this transformation takes @xmath126 into _ & = & _ _ _ ^-1 = with _ = _ 1_n .
the matrix @xmath128 is diagonal , but we still have to determine its elements .
@xmath129 is a diagonal matrix with elements @xmath130 and @xmath131 occurring in equal numbers . for each element
holds : if an even ( odd ) number of @xmath9 provides a factor @xmath131 , the matrix element of @xmath129 is @xmath130 ( @xmath131 ) .
this means : ( i ) a matrix element of @xmath132 is @xmath28 ( @xmath112 ) if an even ( odd ) number of @xmath9 provides a factor @xmath131 ; ( ii ) a matrix element of @xmath133 is @xmath28 ( @xmath112 ) if an odd ( even ) number of @xmath9 provides a factor @xmath131 .
it follows that the @xmath48 eigenvalues of @xmath10 split into @xmath134 eigenvalues of the form [ oddevs ] ( _ = 1^n()_2 - 1 ) , and @xmath134 eigenvalues of the form [ evenevs ] ( _ = 1^n()_2 - 2 ) , where in the first ( second ) case all sign combinations with an even ( odd ) number of minus signs occur .
this is reflected by the indices `` e '' and `` o '' in our result for the partition function , + & = & [ 2(2a)]^mn/2 + & & \ {
_ k=1^n+_k=1^n+_k=1^n-_k=1^n}. + the last term within the braces has a sign differing from that in @xcite .
this is due to our different sign convention for @xmath5 .
the eigenvalues of @xmath6 are of course obtained by multiplying ( [ oddevs ] ) and ( [ evenevs ] ) with the trivial factor @xmath135^{n/2}$ ] .
the results for the eigenvalues of @xmath6 and the partition function are the starting point for the analysis of the thermodynamic properties of the two - dimensional ising model , the most interesting case being the thermodynamic limit @xmath136 .
such analyses can now proceed as usual ( see e.g.@xcite ) and will not be repeated here .
let us briefly remark on the difficulties encountered when trying to solve the two - dimensional model with magnetic field or the three - dimensional model in our approach .
in the first case , we need besides @xmath137 and @xmath138 another matrix = _ = 1^n(c
_ ) , and in the second case another matrix = _ = 1^n(d__+n ) , if the three - dimensional model is obtained from the two - dimensional one by letting sites @xmath40 and @xmath139 for @xmath50 interact with each other ( assuming we have chosen @xmath3 such it can be divided by @xmath140 ; the lattice is then @xmath141 and has in the @xmath140 and @xmath142 directions the character of a screw on a torus , while in the @xmath2 direction it remains periodic ) .
for the three - dimensional model with magnetic field , we need both @xmath143 and @xmath144 . both the @xmath9 and the @xmath145 ( with @xmath146 ) are not part of the algebra of @xmath12 and @xmath13 .
in fact , it turns out that amending the lie algebra with either of these classes of matrices makes the number of elements in the new algebra grow proportional to @xmath48 instead of @xmath147 .
this eliminates the possibility of using an extended version of the algebra of matrices @xmath16 with side length growing proportional to @xmath3 .
but this was essential for the convenient periodicity properties of the matrices @xmath17 .
l. onsager , phys .
65 * , 117 ( 1944 ) .
kramers and g.h .
wannier , phys .
* 60 * , 252 ( 1941 ) ; * 60 * , 263 ( 1941 ) .
w. lenz , phys .
z. * 21 * , 613 ( 1920 ) ; e. ising , z. phys . *
31 * , 253 ( 1925 ) .
b. kaufman , phys .
rev . * 76 * , 1232 ( 1949 ) .
baxter , _ exactly solved models in statistical mechanics _
( academic press , london , 1982 ) .
schultz , d.c .
mattis , and e.h .
lieb , rev .
phys . * 36 * , 856 ( 1964 ) ; c.j .
thompson , j. math .
* 6 * , 1392 ( 1965 ) ; r.j .
baxter , ann .
( n.y . ) * 70 * , 193 ( 1972 ) ; m.j . stephen and l. mittag , j. math .
* 13 * , 1944 ( 1972 ) .
b. kastening , phys .
e * 64 * , 066106 ( 2001 ) , cond - mat/0111380 .
k. huang , _ statistical mechanics _ , 2nd ed .
( wiley , new york , 1987 ) . j.e .
campbell , proc .
london math .
* 28 * , 381 ( 1897 ) ; * 29 * , 14 ( 1898 ) ; h.f .
baker , _ ibid .
_ , * 34 * , 347 ( 1902 ) ; * 3 * , 24 ( 1905 ) ; f. hausdorff , ber .
leipzig , math .- naturwiss . kl . * 58 * , 19 ( 1906 ) .
|
the partition function of the two - dimensional ising model with zero magnetic field on a square lattice with @xmath0 sites wrapped on a torus is computed within the transfer matrix formalism in an explicit step - by - step approach inspired by kaufman s work . however , working with two commuting representations of the complex rotation group so@xmath1 helps us avoid a number of unnecessary complications .
we find all eigenvalues of the transfer matrix and therefore the partition function in a straightforward way .
|
the authors wish to thank tom bullock for comments on an earlier version of this paper . t.h . and j.s .
acknowledge financial support from the academy of finland ( grant no .
|
we show that quantum theory contains observables that are as incompatible as any probabilistic physical theory can have .
we define the joint measurability region of a probabilistic theory which describes the amount of added noise needed to make any pair of observables in the theory jointly measurable .
we then show that the joint measurability region of quantum theory is contained in the joint measurability region of any other probabilistic theory .
quantum theory has a number of important features not known in classical physics , ranging from the superposition and indeterminacy principles formulated by the pioneers to the more recently discovered no - cloning and no - broadcasting theorems .
it is an old problem to identify operationally significant properties of quantum theory that distinguish it from other probabilistic theories . in recent years
many features have been under intensive investigation from this perspective , including information processing @xcite , optimal state discrimination @xcite , entropy @xcite , purification @xcite and discord @xcite .
it has been found that some properties are quite generally valid in any non - classical ( no - signaling ) probabilistic theories while others are specifically quantum .
the existence of pairs of incompatible observables marks one of the most striking distinctions between quantum theory and classical physical theories .
there are many manifestations of incompatibility , perhaps the most famous being the heisenberg uncertainty principle @xcite . in this letter
we demonstrate that quantum theory contains observables that are as incompatible ( in a sense to be defined ) as observables in any probabilistic theory can be .
our aim is thus to compare quantum theory to other possible probabilistic physical theories .
we first need to set some minimal constraints .
a probabilistic theory is a framework that provides a description of physical systems in terms of states and observables with the following general properties : + + ( i ) the states of a system are represented by the elements of a convex subset of a real vector space . + ( ii ) an observable is represented as an affine mapping from the set of states into the set of probability distributions on some outcome space . for simplicity , we restrict ourselves here to observables with a finite or countable number of outcomes . + ( iii ) any affine mapping from the set of states into the set of probability distributions is a valid observable . + we consider a particular probabilistic theory ( pt ) as given by a family of convex sets of states with associated sets of observables that share some properties specific to that pt
. one may think of each pair consisting of a set of states with associated set of observables as an _ instance _ of a pt representing a particular type of physical system .
given a pt , we denote by @xmath0 the probability of obtaining a measurement outcome @xmath1 when an observable @xmath2 is measured in a state @xmath3 .
hence , @xmath4 and @xmath5 .
we will typically label the measurement outcomes by integers .
in quantum theory the states are described by density operators and observables correspond to povms @xcite .
their duality is given by the trace formula ( with @xmath3 a density operator and @xmath6 a povm ) @xmath7 } \ , .\ ] ] another example of a probabilistic theory is a classical theory , where the states are probability measures on a phase space @xmath8 and observables are traditionally represented as functions @xmath9 ; the associated affine maps from states @xmath3 to probability distributions are then given by the formula @xmath10 continuing our discussion on general probabilistic theories , we note that it follows from the required properties ( i)-(iii ) that the set of observables is a convex set ; a mixture of two observables is an observable .
physically mixing corresponds to an experiment where we switch between two measurement apparatuses with a random probability .
we can directly write a mixture of two observables with the same set of measurement outcomes . if the sets of measurement outcomes differ , we can still write a mixture by first adding enough outcomes and then embedding both sets into @xmath11 . another consequence of the basic requirements is that every constant mapping @xmath12 , where @xmath13 is a fixed probability distribution , is an observable and we call it a _ trivial observable_. a trivial observable @xmath14 corresponds to a dice rolling experiment , where we randomly pick the measurement outcome according to a given fixed probability distribution , without manipulating the state at all . in quantum theory , trivial observables are described by povms @xmath15 such that each operator @xmath16 is a multiple of the identity operator , i.e. , @xmath17 for some @xmath18 with @xmath19 .
the concept of _ joint measurement _ can be defined in any probabilistic theory .
two observables @xmath20 and @xmath21 are _ jointly measurable _ if there exists an observable @xmath2 such that @xmath22 in this case @xmath2 is called a _ joint observable _ of @xmath20 and @xmath21 . if @xmath20 and @xmath21 are not jointly measurable , then we say that they are _
incompatible_.
any probabilistic theory contains jointly measurable pairs of observables .
namely , a trivial observable @xmath12 is jointly measurable with any other observable ; we can write a joint observable @xmath23 for the trivial observable and any other observable @xmath20 . this simply corresponds to an experiment where we measure @xmath20 and simultaneously roll a dice .
it is a well known fact that , in quantum theory , an observable which is jointly measurable with all other observables is necessarily a trivial observable .
indeed , any povm element of such an observable commutes with all projections and must therefore be a scalar multiple of the identity ( e.g. ( * ? ? ?
* theorem iv.1.3.1 ) ) .
the following simple observation is a key ingredient for our discussion .
[ prop : sum=1 ] let @xmath20 and @xmath21 be two observables and @xmath24 .
then @xmath25 and @xmath26 are jointly measurable for any choice of trivial observables @xmath27 and @xmath28
. this proposition can be proved with the following construction .
first , let @xmath29 and @xmath30 be the probability distributions related to @xmath27 and @xmath28 .
we define an observable @xmath2 by formula @xmath31 for a fixed @xmath3 , the right hand side is clearly a probability distribution .
moreover , the right hand side is an affine mapping on @xmath3 ; therefore @xmath2 is an observable .
the marginal observables are @xmath32 this proves prop .
[ prop : sum=1 ] . the physical idea behind this construction is the following .
in each measurement run we flip a coin and , depending on the result , we measure either @xmath20 or @xmath21 in the input state @xmath3 . in this way we get a measurement outcome for either @xmath20 or @xmath21 .
in addition to this , we roll a dice and pretend that this is a measurement outcome for the other observable . in this way we get an outcome for both observables simultaneously .
the overall observable is the one given in formula . for two observables @xmath20 and @xmath21 ,
we denote by @xmath33 the set of all points @xmath34\times[0,1]$ ] for which there exist trivial observables @xmath35 such that @xmath25 and @xmath36 are jointly measurable , and we call @xmath33 the _ joint measurability region _ of @xmath20 and @xmath21 .
the joint measurability region thus characterizes how much noise ( in terms of trivial observables ) we need to add to be able to have a joint realization of @xmath20 and @xmath21
. clearly , @xmath20 and @xmath21 are jointly measurable if and only if @xmath37 .
the joint measurability region @xmath33 is a convex region which can be plotted in the plane .
this follows as a simple application of prop . 2 of @xcite ; with the help of this lemma it also follows from prop . [
prop : sum=1 ] that @xmath38\times[0,1 ] : \lambda+\mu \leq 1 \ } \subseteq j({\mathfrak{m}}_1,{\mathfrak{m}}_2 ) \ , .\ ] ] for two orthogonal spin-@xmath39 measurements is a quadrant of the unit disk .
the region @xmath40 ( light ) is a subset of @xmath33 for any pair @xmath41 , while the surplus region ( dark ) depends on the specific pair under consideration.,width=94 ] as an example , suppose that we are within quantum theory and @xmath20 and @xmath21 correspond to spin-@xmath39 measurements in two orthogonal directions , say @xmath42 and @xmath43 -axes .
we then describe them with two povms @xmath44 and @xmath45 , where @xmath46 and @xmath47 are the usual pauli matrices in @xmath48 .
it has been shown in @xcite that for the uniformly distributed trivial observable @xmath49 ( hence describing an unbiased coin ) , the two observables @xmath50 and @xmath51 are jointly measurable if and only if @xmath52 . by prop .
3 of @xcite this inequality is a necessary condition for the joint measurability of any pair @xmath53 and @xmath54 , where @xmath55 are arbitrary trivial observables .
therefore , we conclude that @xmath56\times[0,1 ] : \lambda^2+\mu^2\leq 1\ } \ , .\ ] ] this region is depicted in fig .
[ fig : qubit ] .
we are now ready to define the main concept of this letter .
a global joint measurability feature of a probabilistic theory pt is characterized by the intersection of all the sets @xmath33 across all instances of pt , and we denote @xmath57\times[0,1 ] : ( \lambda,\mu)\in j({\mathfrak{m}}_1,{\mathfrak{m}}_2 ) \\ & \textrm{for all pairs of observables $ { \mathfrak{m}}_1 $ and $ { \mathfrak{m}}_2$}\\ & \textrm{in all instances of pt } \}.\end{aligned}\ ] ] we call @xmath58 the _ joint measurability region _ of @xmath59 .
we always have @xmath60 , but @xmath58 can be larger than @xmath40 .
the larger the surplus region is , the more jointly measurable the theory is globally ; see fig . [
fig : regiong ] .
if @xmath61 , this means that there is a pair of observables @xmath20 and @xmath21 such that the mixtures @xmath25 and @xmath36 are incompatible with any choice of trivial observables @xmath27 and @xmath28 . since @xmath58 can be defined in any probabilistic theory , we can compare the joint measurability regions of different probabilistic theories .
we obviously have @xmath62\times [ 0,1]$ ] in any probabilistic theory where all measurements are jointly measurable , such as the classical probability theory .
in the most incompatible case we would have @xmath63 .
we will next show that quantum theory is , globally , as incompatible as a probabilistic theory can be .
[ th : main ] in quantum theory @xmath64 . in particular , @xmath65 for any probabilistic theory pt . in quantum theory
every observable @xmath2 corresponds to a unique povm @xmath6 by equation .
we will prove that for any pair @xmath66 , there are quantum observables @xmath67 and @xmath68 such that the mixtures @xmath69 and @xmath70 are incompatible with any choice of trivial observables @xmath55 . our proof is based on a recent result @xcite on the joint measurability region for two complementary observables , which is a generalization of the result illustrated in fig .
[ fig : qubit ] .
we have earlier seen that @xmath60 , so we need to show that @xmath71 .
let @xmath66 , i.e. , @xmath72 .
fix @xmath73 such that @xmath74 .
we then choose @xmath75 to be a positive integer satisfying @xmath76 ( this can be done since the left hand side @xmath77 when @xmath78 . )
we will consider a quantum system that is described by a @xmath75-dimensional hilbert space @xmath79 .
let @xmath80 be an orthonormal basis for @xmath79 .
we define another orthonormal basis @xmath81 for @xmath79 by @xmath82 the orthonormal bases @xmath80 and @xmath81 are mutually unbiased , i.e. , @xmath83 .
we define two povms @xmath67 and @xmath68 by @xmath84 we thus obtain a pair of @xmath75-outcome observables on @xmath79 . since @xmath67 and @xmath68 consist of projections and @xmath85 , it follows that they are incompatible . as proved in @xcite , the observables @xmath86 and @xmath87 are incompatible for any choice of trivial observables @xmath55 whenever @xmath88 since @xmath89 we conclude that @xmath90 . in the proof of theorem [ th : main ]
we have used quantum observables with arbitrarily many ( but a finite number of ) outcomes .
it is natural to ask if this is a necessary trick in order to reach the conclusion @xmath64 . to present a partial answer to this question ,
let us investigate the joint measurability region in the case of pairs of binary quantum observables .
our aim is to show that @xmath91\times[0,1 ] : \lambda^2+\mu^2\leq 1\ } \subseteq j ( { \mathsf{m}}_1 , { \mathsf{m}}_2)\ ] ] for any binary observables @xmath67 and @xmath68 , regardless of the dimension of the hilbert space .
in other words , we will show that two orthogonal spin observables are as incompatible as any binary observables can be . to this end , let us note that two binary quantum observables are incompatible if and only if they enable a violation of the bell - chsh inequality @xcite .
we must therefore look at the bell expression @xmath92 let us denote @xmath93 and @xmath94 . using , e.g. , ( * ? ? ? * theorem 1 ) one verifies that by varying the observables @xmath95 and @xmath96 and the bipartite state the numbers @xmath97 and @xmath98 obtain all of the values from the disc @xmath99 .
if we now mix the observables @xmath95 with the trivial observable @xmath100 with some weights @xmath101 and @xmath102 we see that the pair @xmath103 turns into @xmath104 , thus changing the bell expression from @xmath105 to @xmath106 .
we must therefore determine those @xmath107 for which @xmath108 for all @xmath103 satisfying @xmath99 ( see fig .
[ fig : bell ] ) .
but the boundary curve for this region is obtained when the equations @xmath109 and @xmath110 have at most one common solution . by inserting @xmath111 into the first equation the problem reduces to determining when the discriminant is negative or zero , and one readily verifies that this is the case exactly when @xmath112 . in conclusion ,
given any pair of binary observables @xmath67 and @xmath68 , and weights @xmath101 and @xmath102 with @xmath113 , the mixtures @xmath114 and @xmath115 can not be used to violate the bell - cshs inequality and must therefore be jointly measurable .
we note that in the case @xmath116 the same result using a different technique has been obtained by banik _
et al .
_ @xcite .
we note that in the proof of theorem [ th : main ] we have used quantum observables acting on an arbitrarily high ( but finite ) dimensional ( but finite ) hilbert space
. we can again ask if this is a necessary trick in order to reach the conclusion @xmath64 . since
two mutually unbiased bases are expected to be among the most incompatible observables in a fixed dimension @xmath75 , our construction suggests that the conclusion @xmath117 can be reached only if one considers arbitrarily high dimensions .
a proof of this fact is , however , lacking .
we show finally that the conclusion @xmath117 can be reached by using a _ single pair _ of incompatible observables if we consider an infinite dimensional system and observables with infinite number of outcomes .
let @xmath118 be an infinite dimensional hilbert space and write it as a direct sum of finite @xmath75-dimensional hilbert spaces @xmath79 , @xmath119 . in each @xmath79
consider a pair of mutually unbiased orthonormal bases @xmath120 and @xmath121 , where the latter is obtained from the first one by the formula .
we define two povms @xmath122 and @xmath123 via @xmath124 these observables act in the infinite dimensional hilbert space @xmath118 and @xmath75 in is an index labeling the different outcomes .
the outcome space of @xmath125 and @xmath126 is @xmath127 .
[ th : totalmub ] the observables @xmath125 and @xmath126 defined in satisfy @xmath128 .
let @xmath29 and @xmath30 be two probability distributions defined on @xmath129 .
assume that @xmath130 and define two observables @xmath131 via @xmath132 we need to show that @xmath133 and @xmath134 are incompatible . to prove this
, we make a counter assumption that @xmath131 are jointly measurable .
this implies that for any projection @xmath135 on @xmath118 , the projected observables @xmath136 and @xmath137 acting on a subspace @xmath138 are jointly measurable .
( if @xmath139 is a joint observable of two observables @xmath140 , then @xmath141 is a joint observable of @xmath142 in @xmath138 . )
especially , the projections of @xmath133 and @xmath134 to any subspace @xmath79 should be jointly measurable .
but from the result cited in the proof of theorem [ th : main ] we know that for @xmath75 large enough , the projections to @xmath79 are incompatible .
hence , @xmath133 and @xmath134 are incompatible .
we note that the observables @xmath125 and @xmath126 defined in are not the only pair that satisfy @xmath128 .
namely , we can modify @xmath125 and @xmath126 in any chosen subspace @xmath79 but the conclusion @xmath128 is still true since it depends on the fact that @xmath125 and @xmath126 contain mutually unbiased bases in arbitrarily high dimension .
an interesting question within quantum theory is to find a characterization of all pairs of quantum observables @xmath140 that satisfy @xmath143 .
do they exist in finite dimensions ?
can they have finite numbers of outcomes ?
|
the fluorescence of the nitrogen gas in the atmosphere in the visible ( vis ) and ultraviolet ( uv ) regions was independently proposed in the sixties of last century by greisen @xcite , delvaille et al .
@xcite , suga @xcite , and chudakov @xcite as a way to detect high energy showers created by cosmic rays impinging on the earth atmosphere .
nowadays the air fluorescence method is a well established technique used by the experiments hires @xcite , which is the continuation of the pioneering fly s eye experiment @xcite , auger @xcite , and by the telescope array project @xcite .
the main emission of the fluorescence light is in the uv region between @xmath2 and @xmath3 .
the transparency of the atmosphere at those wavelength is limited by the presence of molecular oxygen and ozone , which absorb the uv photons . the rayleigh scattering , whose cross section scales as @xmath4 , and the mie scattering further reduce the number of photons which arrive at the detector .
we introduce an extinction length @xmath5 , function of the wavelength @xmath6 , which enters in the transmission of the light through the relationship @xmath7 , where @xmath8 is the intensity of a collimated light beam of wavelength @xmath6 after having travelled for the distance @xmath9 in the atmosphere , and @xmath10 is the initial intensity ( at @xmath11 ) . for the uv , @xmath12 .
this fact poses two main drawbacks .
first , the detector can not be placed at sea level , but at altitude .
for example , auger is at about @xmath13 above sea level , hires at @xmath14 .
second , the detection range of the instruments extends to a distance of the order of @xmath15 .
auger , for example , is capable to detect the highest energy air showers at a distance of the order of @xmath16 @xcite .
this is correlated to the total observation rate , which goes roughly as @xmath17 , which is therefore limited also by the atmospheric transmission .
the atmosphere is very transparent in some regions of the near - infrared ( nir ) region , below @xmath18 .
the molecules which present absorption bands are water and , in a minor extent , co@xmath19 .
several high transparency windows are present : from 0.8 to @xmath20 , from 0.95 to @xmath21 , from 1.15 to @xmath22 , from 1.5 to @xmath23 , from 2.0 to @xmath24 ( see , for example , ref.@xcite ) . moreover , the rayleigh scattering is negligible in the nir , because of the @xmath4 dependence . if the air fluoresces at those wavelengths , then there is the possibility to detect high energy cosmic rays impinging on the atmosphere at a very long distance from the detector , thus increasing the observation rate .
the spectroscopy of the molecule of nitrogen is rather complex and has been extensively studied in the last century .
an exhaustive review can be found in ref.@xcite . despite of this , to our knowledge there are no investigations on the infrared fluorescence emission from n@xmath19 or air excited by ionizing radiation .
the only research which extends somewhere in the nir region was performed in the sixties by davidson and oneil @xcite , who used photomultipliers to reach the maximum wavelength of @xmath25 . in this paper
we present our first measurements of the fluorescence light in the nir region of air , nitrogen , and oxygen .
the light emitted by the gas , excited by a @xmath0 electron beam , was detected by an ingaas photodiodes , sensitive down to about @xmath26 .
spectra have been recorded at atmospheric pressure and room temperature by mean of a fourier transform infrared spectrometer .
we also report on the light yield , which has been obtained by comparing the nir signal with the known fluorescence output in the uv region , measured by means of a silicon photodiode .
a schematic of the experimental setup is shown in fig.[fg : setup ] .
an electrostatic gun @xcite accelerated electrons , emitted by a hot filament , up to @xmath27 kinetic energy .
electrons exited the gun through a cvd diamond window @xcite , @xmath28 thick , whose external side had been metalized to screen the fluorescence light emitted by the diamond itself .
the gun was operated at @xmath0 in pulsed mode , with pulses @xmath29 long and rate of @xmath30 . in those conditions ,
the electron current behind the diamond windows was a few hundreds @xmath31 .
the chamber containing the gas was a stainless - steel cylinder with the axis along the beam path and volume about @xmath32 .
the chamber size was a compromise between the range of the electrons in the gas and the need of an adequate solid angle for the detection of the light .
the ionization cloud was almost entirely contained in the chamber , and the electrons that hit the materials of the chambers did not affected neither disturbed the measurements , as proved in section 3 .
perpendicular to the axis , one opposed to the other , two quartz windows allowed to view inside the chamber .
they were @xmath33 thick in order to stop the x - rays generated in the chamber , that could alter the signals of the photodiodes . by construction
, they looked exactly at the same gas region , a small fraction ( about @xmath34 ) of the chamber volume .
the light to the spectrometer exited through a third quartz window on the closing flange of the vessel . inside the chamber ,
an electrode collected the ionization charge for the beam monitoring and normalization , which was fed into a charge amplifier .
the output signal @xmath35 was displayed on a digital oscilloscope and also used as a feedback to stabilize the electron gun current against long term drifts .
@xmath35 was compared to a reference voltage level , and the difference signal was used to opportunely bias the grid in front of the filament , which regulated the electron beam current intensity before acceleration .
gas from certified bottles was flown in the chamber at room temperature and atmospheric pressure with a flux of @xmath36 .
we used n@xmath19 ( impurity concentration @xmath37 ) , o@xmath19 ( impurity concentration @xmath38 ) , and dry air , which is a mixture of @xmath39 nitrogen and @xmath40 oxygen ( impurity concentration @xmath41 ) .
the fluorescence light was detected by two solid state photodiodes ( pds ) , which could be placed on the two lateral windows of the chamber or inside the spectrometer .
the uv light was detected by a squared si pd @xcite , @xmath42 area , with a quantum efficiency ( qe ) of about @xmath43 from @xmath2 to @xmath44 .
the infrared radiations was detected by a round ingaas pd @xcite , @xmath45 diameter , sensitive from about @xmath46 to @xmath26 and cooled down to about @xmath47 by a built - in peltier cooler .
its qe is above @xmath39 from @xmath48 to @xmath49 .
the qe curves of both pds are shown in fig.[fg : pd_qe ] .
each photodiode was readout by a dedicated charge preamplifier , based on the ua1 hybrid @xcite .
the gain was measured by injecting a known calibration charge .
the output signals were displayed and measured by means of a digital oscilloscope .
since the si pd is also sensitive to the nir radiation , we used some colored glass filters to select only the uv photons .
the combinations of two filters , the filter fgb25 ( by thorlabs ) and the filter ha30 ( by hoya corp . ) , produced the transmission curve shown in fig.[fg : pd_qe ] .
only two bands are allowed , from @xmath2 to @xmath46 , which is the region of the uv fluorescence , and from @xmath50 to @xmath51 , where the fluorescence light output of air and nitrogen is negligible @xcite .
spectra were recorded and analyzed by a fourier transform infrared ( ftir ) spectrometer ( mod .
equinox55 by bruker optics ) which is essentially a two - arm michelson interferometer with a @xmath52 he - ne laser as reference light source .
for the equinox55 the minimum measurable wavelength was @xmath52 .
for details of this well known spectroscopic technique see for example ref.@xcite .
the spectrometer was controlled and the data were elaborated by the software opus 5.5 ( by bruker optics ) .
the laser light covered the same path of the radiation under measurement , and reached the detector , producing a very large noise . to filter the he - ne light
, we placed a colored glass filter in front of the photodiode , which transmitted only photons with wavelength longer than @xmath53 ( filter fgl780 by thorlabs ) .
the ftir spectrometer recorded the interferogram of the light source as a function of the wavenumber @xmath54 . in the calculation of the fourier transform of the interferogram to obtain the spectrum , we applied the apodization function blackman - harris ( 3-terms ) , and the phase correction with the mertz method ( no peak search ) @xcite .
the spectrometer resolution on @xmath55 ranged between 20 and @xmath56 but the final spectrum resolution was worse because of the low spatial coherence of the light source and of the mathematical operations involved in the interferogram - to - spectrum conversion .
note that a constant resolution @xmath57 on @xmath55 implies a non constant wavelength resolution @xmath58 , given by @xmath59 . to reduce fluctuations and noise in the spectrum ,
the final result was obtained by averaging together many different spectra ( between 10 and 20 ) .
as preliminary operation , we verified that no light emission occurred from the cvd diamond window or from the quartz windows or other materials of the chamber . to accomplish that , we evacuated the chamber with a rotative pump and run the electron gun at the highest beam current and energy .
we could not observe any light signal .
figures [ fg : spettron2 ] , [ fg : spettroo2 ] , and [ fg : spettrodryair ] show the measured spectra of nitrogen , oxygen and dry air , respectively . the spectra were recorded at atmospheric pressure and room temperature ( @xmath60 ) , and measured with a resolution @xmath61 , apart from oxygen .
they are corrected for the transmission curve of the filter fgl780 and for the qe curve of the ingaas pd , and truncated to @xmath62 . to identify the atomic transitions
, we used the nist atomic spectra database @xcite .
molecular nitrogen has a very rich spectrum , with numerous band systems corresponding to electronic transitions of n@xmath19 and n@xmath63 .
the near infrared region is dominated by the first positive system ( @xmath64 ) ( fps ) .
a less important system is the @xmath65 meinel system ( ms ) of the ionized nitrogen .
the herman infrared system ( his ) also appears , which extends from @xmath50 to @xmath66 @xcite . the spectrum in fig.[fg : spettron2 ] is dominated by the unresolved band around @xmath67 , which has been already observed by @xcite .
the band is the overlap of two nearby atomic transitions at @xmath68 , characteristic of the forbidden doublet transition @xmath69_{32}$ ] , and of the fps ( 0 - 0 ) band ( fig.[fg : spettron2zoom ] ) .
the second more intense band is the ( 0 - 1 ) band of the fps centered at @xmath70 .
other minor bumps has been identified as belonging to the hir , to the ms ( at about @xmath71 ) and to the fps ( the ( 0 - 2 ) band centered at @xmath72 ) .
two very weak atomic lines of the neutral nitrogen atom are distinguishable at the side of the main band , at @xmath73 ( transition @xmath74 ) and at @xmath75 ( transition @xmath76 ) .
because of the weak intensity and the poor resolution , bands from about 1250 to @xmath77 can not be unambiguously identified .
since the oxygen fluorescence light output is sensibly lower than the nitrogen or dry air one , we run the spectrometer with a resolution @xmath78 .
molecular oxygen is a weak light emitter since transitions from excited states to the ground state are strongly forbidden , for most of the excited states . as a fact
, the only line we could clearly identify to belong to a molecular transition is the faint line at @xmath79 , which is the well - known @xmath80 magnetic dipole intercombination transition @xcite .
all other lines have been identified as atomic transitions from the excited neutral oxygen ( o i ) , created by electron - impact dissociation of the o@xmath19 molecule , or from the single ionized atom ( o ii ) .
the list of the lines and relative transitions is reported in table [ tab : o2lines ] .
the line at @xmath81 is the sum of two very close lines , which were not resolved by the spectrometer .
the line intensities , after the continuum subtraction , are normalized to the highest intensity line . [ cols="^,^,^,^",options="header " , ] we calculated the weighted average , which yields @xmath82 .
therefore , our estimate of the nir light yield @xmath83 is @xmath84
the light yield of dry air in the nir is about @xmath85 of the uv yield .
the qe of the ingaas semiconductor is about @xmath86 times higher than the uv photocathode of a photomultiplier , so the total number of photoelectrons is approximately the same in the two cases .
however , from a practical point of view , when compared to photomultipliers , operation with ingaas photodiodes poses several drawbacks , such as : - small dimension ( area @xmath87 ) ; - no multiplication ; - need of a low noise electronics .
therefore , detection of cosmic showers with ingaas pds is nowadays not compelling
. the situation could be different if we adopt si semiconductor , whose qe can extend down to the neighborhood of @xmath88 , being limited by the @xmath89 band - gap of the ordinary bulk silicon .
moreover , photodiodes can be produced with an internal gain , obtaining avalanche photodiodes ( apds ) , with gain in the range @xmath90 and surface area larger than @xmath91 @xcite .
however , limiting the wavelength range would decrease the light output .
the amount of nir light with @xmath92 is about @xmath93 of the total light yield in the @xmath94 range .
therefore , si apd in the nir is about a factor of 10 less sensitive with respect to the uv light , at parity of detecting surface area .
so far we have not mentioned the main argument in favor of the nir fluorescence , namely , the transparency of the atmosphere . the uv light produced by a shower @xmath95 far away from the detector is attenuated by the atmosphere by about a factor of 10 .
then , since atmosphere is transparent in the nir , detection of the nir fluorescence with a si apd is , in this case , equivalent to the detection of the uv light with photomultipliers . for the general case
, it is likely that the detection of uhecrs through the nir fluorescence with si apds is not much disadvantaged with respect to the uv fluorescence with photomultipliers .
the principal pro for the physics of uhecrs would be the increase of the event rate , which scales as @xmath17 . on the opposite side , however , it must be demonstrated that the large background noise of the atmosphere in the nir , which is higher than in the uv @xcite , can be handled and does not introduce other limitations .
the technological research on si apds towards the enhancement of the response in the nir is progressing and has already reached remarkable results , not yet transferred on commercial devices .
for example , several groups have obtained apds with qe above @xmath96 for @xmath97 by laser treatment of the si surface @xcite .
others used inclusions of sige absorbing layers to enhance the qe well beyond the si cutoff wavelength @xcite .
it is therefore thinkable that in the near future the si apds are much more performing than the actual ones .
for sake of completeness , we have also to mention the possibility to use nir photomultipliers , whose sensitivity now extends down to @xmath1 @xcite .
they have a small active area ( smaller than the ingaas pd ) and a qe @xmath98 .
therefore operation with such photomultipliers seems not competitive with respect to large area si apds .
we have measured for the first time the fluorescence light of the air in the near infrared region .
the nir photon yield is 1/5 of the uv yield .
we propose such fluorescence light as a new way to detect uhecrs .
since the atmosphere is more transparent in the nir than in the uv region , there is reasonable expectation that the new method could lead to an increase of the uhecr observable rate .
this paper is the first step of a research program which aims to this goal .
the authors are grateful to g.carugno and a.f.borghesani for their support and for useful discussions .
yu.ralchenko , a.e.kramida , j.reader , and nist asd team ( 2008 ) .
nist atomic spectra database ( version 3.1.5 ) , [ online ] .
available : http://physics.nist.gov/asd3 . national institute of standards and technology , gaithersburg , md .
|
we have investigated the fluorescence emission in the near infrared from the air and its main components , nitrogen and oxygen .
the gas was excited by a @xmath0 electron beam and the fluorescence light detected by an ingaas photodiode , sensitive down to about @xmath1 .
we have recorded the emission spectra by means of a fourier transform infrared spectrometer .
the light yield was also measured by comparing the near infrared signal with the known ultraviolet fluorescence , detected by a si photodiode .
the possibility of using the near infrared fluorescence of the atmosphere to detect ultra - high - energy cosmic rays is discussed , showing the pros and the cons of this novel method .
air fluorescence , nitrogen fluorescence , oxygen fluorescence , near - infrared , cosmic - ray detection , ultra - high - energy cosmic rays
|
the fate of planetary systems around sun - like stars can be studied by finding planets around their white dwarf ( wd ) remnants .
wds are physically small ( @xmath1 ) , which sharply limits their luminosity in the infrared , where the planet spectrum peaks @xcite .
compared to main - sequence ( ms ) stars , the significant gain in contrast makes wds excellent targets for photometric searches for planetary companions .
however , to this day , no planet detection around wds has been reported . _
spitzer _ irac imaging surveys by @xcite , @xcite , @xcite , and @xcite , and near - infrared imaging surveys by @xcite , @xcite , and @xcite were unsuccessful in finding planetary mass companions to wds with limits typically between 5 and @xmath2 .
here we focus on using low - mass wds to provide additional leverage for detecting planetary mass companions . timing measurements of pulsating wds is an alternative , powerful method to search for planets @xcite , but this method can be used only for a limited sample of stars
. there is a promising signature of a @xmath3 planet in the pulsation timing measurements of the wd gd 66 @xcite .
however , a full orbit has not been observed , and this detection remains provisional ( f. mullally 2009 , priv .
comm . ) .
the discovery of circumstellar debris disks around more than a dozen wds imply that 1% to 3% of wds with cooling ages @xmath4 500 myr have disks ( see * ? ? ?
* and references therein ) .
these disks most likely formed by tidally disrupted asteroids @xcite .
hence , there is indirect evidence that at least a few percent of wds have remnant planetary systems . of course
, the observed debris disk fraction may represent only the planetary systems that become unstable during post - ms evolution , and the true fraction of wds with planets may be much higher .
the reason for the lack of known planetary companions to wds may be a target selection effect @xcite .
@xcite find that less than 3% of the stars with @xmath50.5 @xmath4 [ fe / h ] @xmath4 0.0 have doppler - detected planets , whereas 25% of the stars with [ fe / h ] @xmath6 + 0.3 have gas giant planets .
based on the age - metallicity relation derived by @xcite , @xcite find that only 3% of the stars in the solar neighborhood have had [ fe / h ] @xmath6 + 0.3 in the past 10 gyr .
therefore , a random selection of wds would be dominated by wds that come from low metallicity progenitors .
the best way to find planetary companions to wds may be to search for companions to wds that form from metal - rich stars .
low - mass helium - core wds ( @xmath7 ) can be produced from interacting binary systems , and traditionally all of them have been attributed to this channel .
however , a low - mass wd can also result from a single star that experiences severe mass loss on the first ascent giant branch .
@xcite identified several low - mass wds in the upper part of the wd cooling sequence of the metal - rich cluster ngc 6791 , and they argued that there is a mechanism in clusters to produce low - mass wds without requiring binary star interactions .
however , this argument is now challenged by @xcite , who argue that the bright peak observed in the wd luminosity fuction of this cluster is better explained as binary wds .
@xcite also argue that the low - mass wds found by @xcite can be explained as the product of binary star evolution .
unlike the relatively faint low - mass wds in ngc 6791 , it is possible to obtain radial velocity measurements of low - mass wds in the field and test the binary formation scenario .
around half the low - mass wds observed by @xcite lack detectable radial velocity companions . the eso snia progenitor survey ( spy ) searched for radial velocity variations in more than a thousand wds using the very large telescope @xcite .
they found that about 58% of the low - mass wds in their sample do not show radial velocity variations @xcite .
their detailed analysis of the detection probability of ms , wd , and brown dwarf companions show that their observations of low - mass wds are inconsistent with the idea that all of them reside in close binary systems .
the spy observations are less sensitive to brown dwarf companions , because of their smaller mass , but even in this case 100% binary fraction rate can be ruled out ( r. napiwotzki 2007 , private communication ) .
@xcite argue that these single low - mass wds might come from old metal - rich stars that truncate their evolution prior to the helium flash from severe mass loss .
the fraction of single low - mass wds among the field wd sample is also compatible with the fraction of metal - rich dwarfs among the field dwarf population .
a strong implication of the @xcite scenario is that single low - mass wds are good targets for planet searches because they may arise from metal - rich progenitors .
this leads to an inversion of prior strategies for seeking planetary companions to wds , which avoid low - mass wds on the grounds that they are the products of binary interactions . instead
, some of these objects may be the descendants of metal - rich stars that are known to have a high frequency of planet detection .
if the progenitors of these single low - mass wds have planets , the planets are more likely to survive the post - ms evolution , as these he - core wds do not go through the asymptotic giant branch phase .
an alternative formation scenario for single low - mass wds is proposed by @xcite . in this scenario ,
low - mass wds form from sun - like stars with close planetary companions .
a common envelope phase with a massive planet can expel the stellar envelope and form a low - mass wd .
the planet may ( not ) survive the common envelope phase . however , if there are additional planets in wide orbits , they are likely to survive @xcite .
both scenarios proposed by @xcite and @xcite suggest that low - mass wds may have planets . here
we present a _
irac search for substellar companions to 14 low - mass wds .
our sample selection and observations are discussed in section 2 , whereas the detectability of planets around our targets and the results of our search are discussed in section 3 and 4 .
@xcite presented a radial velocity survey of 71 wds including 13 low - mass wds with no detectable radial velocity variations . @xcite and @xcite obtained irac observations of 3 of these stars ( wd 0316 + 345 , wd 1353 + 409 , and wd 1614 + 136 ) .
we target the remaining 10 single low - mass wds for follow - up observations .
in addition , we target 4 low - mass wds from the palomar - green survey @xcite . @xcite
presented near - infrared photometry for two of these stars which did not reveal any companions .
the physical parameters of our targets , including masses , distances , and wd cooling ages , are presented in table 1 .
all of our targets are hydrogen - rich da wds .
lccrrr pg 0132 + 254 & 19960 & 0.41 & 180 & 60 & 1 + pg 0237 + 242 & 69160 & 0.40 & 881 & 20 & 1 + wd 0339 + 523 & 12640 & 0.33 & 103 & 180 & 2 + wd 0437 + 152 & 10410 & 0.38 & 153 & 340 & 3 + wd 0453 + 418 & 13660 & 0.44 & 44 & 150 & 2 + wd 0509@xmath5007 & 32200 & 0.40 & 117 & 10 & 4 + pg 0808 + 595 & 27330 & 0.42 & 238 & 10 & 1 + pg 0943 + 441 & 12820 & 0.41 & 30 & 340 & 1 + wd 0950@xmath5572 & 12400 & 0.44 & 58 & 230 & 2 + pg 1229@xmath5013 & 19430 & 0.41 & 59 & 70 & 1 + hs 1653 + 7753 & 29400 & 0.32 & 180 & 10 & 5 + pg 1654 + 637 & 15070 & 0.44 & 98 & 200 & 1 + pg 2226 + 061 & 15280 & 0.44 & 65 & 190 & 1
+ pg 2257 + 162 & 24580 & 0.43 & 212 & 20 & 1 + observations reported here were obtained as part of the cycle 5 spitzer fellowship program 474 .
we obtained 3.6 , 4.5 , 5.8 , and 8 @xmath8 m images with integration times of 30 or 100 seconds per dither , with five or nine dithers per object .
we use the iraf phot and idl astrolib packages to perform aperture photometry on the individual bcd frames from the latest available irac reduction pipeline ( 17.2 or 18.5 for our targets ) .
since our targets are relatively faint , we use the smallest aperture ( 2 pixels ) for which there are published aperture corrections . following the irac calibration procedure , corrections for the location of the source in the array
were taken into account before averaging the fluxes of each of the dithered frames at each wavelength .
channel 1 ( 3.6@xmath8 m ) photometry is also corrected for the pixel - phase - dependence .
the results from iraf and idl reductions are consistent within the errors .
the photometric error bars are estimated from the observed scatter in the 5 ( or 9 ) images corresponding to the dither positions .
we also add the 3% absolute calibration error in quadrature . finally , we divide the estimated fluxes by the color corrections for a rayleigh - jeans spectrum @xcite .
these corrections are 1.0111 , 1.0121 , 1.0155 , and 1.0337 for the 3.6 , 4.5 , 5.8 , and 8 @xmath8 m bands , respectively .
we present the irac photometry of our targets in table 2 .
most of our targets are detected in 2mass , at least in the @xmath9 band .
several targets also have sloan digital sky survey ( sdss ) photometry available .
we obtained @xmath10 and @xmath11 band photometry of hs 1653 + 7753 using the mdm 2.4 m telescope equipped with the echelle ccd on ut 2009 april 5 .
we use observations of the standard star field pg 1633 + 099 @xcite to calibrate the photometry .
the photometric reductions were performed by j. thorstensen , and kindly made available to us .
hs 1653 + 7753 has @xmath12 15.08 mag and @xmath13 15.33 mag .
we estimate an internal accuracy of 0.01 mag for the photometry , but to be conservative , we adopt errors of 0.03 mag .
h. c. harris kindly provided us unpublished @xmath14 photometry of wd 0339 + 523 .
we use the optical photometry from @xcite for the remaining targets .
lrrrr pg 0132 + 254 & 54.7 @xmath15 3.2 & 32.9 @xmath15 1.4 & 21.1 @xmath15 11.0 & 25.6 @xmath15 11.2 + pg 0237 + 242 & 54.2 @xmath15 3.4 & 34.4 @xmath15 1.8 & 24.2 @xmath15 10.8 & 22.5 @xmath15 8.4 + wd 0339 + 523 & 103.9 @xmath15 5.0 & 66.1 @xmath15 3.8 & 33.7 @xmath15 19.2 & 23.7 @xmath15 20.6 + wd 0437 + 152 & 79.0 @xmath15 4.1 & 49.2 @xmath15 2.2 & 31.3 @xmath15 17.7 & 22.0 @xmath15 19.3 + wd 0453 + 418 & 449.7 @xmath15 14.4 & 287.3 @xmath15 10.4 & 176.7 @xmath15 25.9 & 102.9 @xmath15 46.8 + wd 0509@xmath5007 & 333.3 @xmath15 11.1 & 213.4 @xmath15 11.5 & 141.8 @xmath15 18.9 & 73.5 @xmath15 18.3 + pg 0808 + 595 & 48.3 @xmath15 5.3 & 27.8 @xmath15 2.4 & 13.5 @xmath15 13.2 & 13.4 @xmath15 4.2 + pg 0943 + 441 & 881.2 @xmath15 32.4 & 549.1 @xmath15 19.4 & 367.1 @xmath15 16.7 & 212.7 @xmath15 19.7 + wd 0950@xmath5572 & 269.6 @xmath15 14.4 & 173.7 @xmath15 7.7 & 121.9 @xmath15 19.3 & 121.9 @xmath15 40.1 +
pg 1229@xmath5013 & 237.6 @xmath15 9.2 & 150.3 @xmath15 6.8 & 80.4 @xmath15 9.4 & 73.8 @xmath15 32.1 + hs 1653 + 7753 & 161.3 @xmath15 5.1 & 106.5 @xmath15 3.9 & 66.9 @xmath15 5.7 & 39.5 @xmath15 7.9 + pg 1654 + 637 & 89.6 @xmath15 4.7 & 53.9 @xmath15 2.3 & 37.1 @xmath15 14.6 & 15.2 @xmath15 11.1 + pg 2226 + 061 & 198.6 @xmath15 14.6 & 122.1 @xmath15 7.1 & 74.5 @xmath15 24.6 & 41.2 @xmath15 21.0 + pg 2257 + 162 & 410.7 @xmath15 17.9 & 282.4 @xmath15 14.1 & 190.1 @xmath15 30.4 & 113.2 @xmath15 36.6 + in addition , we obtained @xmath16 imaging observations of hs 1653 + 7753 using the mmt and magellan infrared spectrograph ( mmirs ; * ? ? ?
* ) on the mmt on ut 2009 june 13 .
the observations were obtained under cloudy conditions , and they included 10 exposures of 30 s in @xmath9 , 10 exposures of 20 s in @xmath17 , and 25 exposures of 45 s in @xmath18 .
figure 1 presents the spectral energy distributions of twelve of our fourteen low - mass wds .
two wds , with more complicated seds , are presented below .
d. koester kindly provided pure - hydrogen atmosphere wd model spectra for our targets .
we perform synthetic photometry on these models using the appropriate transmission curves for the @xmath19 , sdss , 2mass , and irac filters .
the resulting fluxes are then compared with the observations to find the normalization factor for the wd models .
we weight fluxes by their associated error bars .
the solid lines in figure 1 present the appropriate wd model for each star normalized to match the observations .
the use of optical and near - infrared photometry helps us constrain the predicted mid - infrared photospheric fluxes for wds .
the mid - infrared photometry of all but one of the twelve wds presented in this figure is consistent with the predicted photospheric flux from wds ; we do not detect excess mid - infrared flux from planets or substellar companions . the thermal emission from planetary companions
would be detected most strongly at 4.5 @xmath8 m @xcite .
one of the objects in figure 1 , wd 0950@xmath5572 shows a mid - infrared flux excess .
figure 2 presents a contour map of the immediate field around this wd ; it is near a bright star and even the smallest aperture that we use ( 2 pixels ) is contaminated by flux from this nearby bright star .
therefore , the observed excess flux is most likely not intrinsic to the wd environment . the usual method for estimating the upper mass limit for substellar companions in photometric surveys is to calculate the total age of the white dwarf , i.e. the wd cooling age plus the ms lifetime , and then use evolutionary models for substellar objects and giant planets to predict their brightness at the age and distance of the wd ( see for example
* ; * ? ? ?
but there is a problem for low mass wds : if they form through common envelope evolution , it is not possible to know their progenitor mass and their ms lifetimes .
hence only a lower limit to the age of the system , i.e. the wd cooling age , is known .
even if the low - mass wds form directly from single metal - rich progenitors , currently we do not have initial - final mass relations that would give us progenitor masses and lifetimes for such systems .
therefore , all we can do is to calculate companion upper mass limits for a range of possible ages starting from the wd cooling ages to several gyr .
all of the objects in our sample are hotter than 10,000 k ( table 1 ) , therefore their wd cooling ages are less than 350 myr @xcite . a relatively massive planet or brown dwarf can stay hot and easily detectable by @xmath20 for billions of years .
some brown dwarfs or planets may also be reheated during the red giant phase of the progenitor s evolution , either by direct radiation or by impacts from planets and debris belts destabilized during stellar mass loss @xcite .
collisions between two planets may create mergers that would essentially reset the cooling clock of the planet and make the remnant planet overluminous by a factor of a hundred .
however , @xcite predict that only about 2% of young wds may have overluminous planetary companions that form through collisions .
@xcite report the discovery of a unique first - ascent giant star with a substantial dust disk .
they suggest that the circumstellar material around this star may originate from a common - envelope phase with a low - mass stellar or substellar companion .
the gaseous and dusty material is similar to the primordial disks observed around t tauri stars .
they speculate that these disks may form a second generation of planets around first - ascent giant stars .
single low - mass wds may descend from such stars @xcite .
the discovery of a close - in substellar companion to the low - mass wd 0137@xmath5049 implies that this substellar companion has gone through a common - envelope phase that resulted in the ejection of the envelope of the progenitor first - ascent giant star @xcite .
if the first - ascent giant stars with gaseous / dusty disks are common and if these stars form low - mass single wds , then these wds may have second generation planets as well .
these planets would be as young as the cooling age of the wd , and therefore warmer and more luminous .
for the 11 stars presented in figure 1 ( excluding wd 0950@xmath5572 ) , we use @xcite and @xcite models to put upper limits on companions that would escape detection .
i. baraffe kindly provided us substellar cooling models updated to include fluxes in the irac bandpasses for 0.1 - 5 gyr old objects with @xmath21 1 - 100 @xmath22 .
figure 3 shows an example of predicted excess flux from planetary and brown dwarf companions around one of our targets , pg 2226 + 061 . by demanding a 3@xmath23 excess from any possible companion ,
we exclude 100 myr old , @xmath24 planets around pg 2226 + 061 . depending on the unknown ages of the possible planetary companions ,
this limit goes up to more than 30 @xmath22 for 5 gyr or older planets . in all cases ,
we are able to rule out brown dwarf companions with @xmath25 k around this wd .
we achieve similar limits on substellar companions for the rest of our sample .
these limits and our search radii are presented in table 3 . since the total ages
( wd cooling age + ms lifetime ) are unknown , we present companion mass limits for a range of ages .
if the planets are as old as the wd cooling ages ( 20 - 340 myr ) , we can rule out planets more massive than @xmath26 to @xmath27 around 10 of our targets ; second - generation massive planets are not common around low - mass wds . however ,
if the planets are as old as 5 gyr , only massive brown dwarf companions are ruled out .
lrllllr pg 0132 + 254 & 430 & @xmath28 & @xmath29 & @xmath30 & @xmath31 & 800 k + pg 0237 + 242 & 2110 & @xmath32 & @xmath33 & @xmath34 & @xmath34 & + wd 0339 + 523 & 250 & @xmath28 & @xmath29 & @xmath30 & @xmath31 & 800 k + wd 0437 + 152 & 370 & @xmath35 & @xmath29 & @xmath36 & @xmath31 & 800 k + wd 0453 + 418 & 110 & @xmath37 & @xmath38 & @xmath39 & @xmath40 & 700 k + wd 0509@xmath5007 & 280 & @xmath29 & @xmath31 & @xmath32 & @xmath41 & 1600 k + pg 0808 + 595 & 570 & @xmath38 & @xmath36 & @xmath40 & @xmath42 & 1100 k + pg 0943 + 441 & 70 & @xmath43 & @xmath44 & @xmath29 & @xmath36 & 700 k + pg 1229@xmath5013 & 140 & @xmath43 & @xmath35 & @xmath38 & @xmath36 & 700 k + pg 1654 + 637 & 240 & @xmath43 & @xmath28 & @xmath44 & @xmath30 & 700 k + pg 2226 + 061 & 160 & @xmath37 & @xmath29 & @xmath30 & @xmath40 & 700 k + one of our targets , hs 1653 + 7753 , shows excess flux in the mid - infrared .
the contour maps of the irac images show that there is a nearby faint source in the irac beam .
the 2mass images do not reveal any nearby , faint sources .
we obtained mmirs @xmath16 imaging of hs 1653 + 7753 to resolve this nearby source .
figure 4 displays @xmath16 and irac 4.5@xmath8 m images of the immediate field around hs 1653 + 7753 .
our mmirs data shows that there is a faint source to the southeast of the wd at @xmath45 separation .
this source is relatively bright in the mid - infrared , and contaminates the irac photometry even in a 2 pixel aperture .
we use a 3 pixel ( @xmath46 ) aperture to measure the total flux from the two sources in our irac data .
the nearby source is clearly resolved in the @xmath9-band data , but it is not detected well in the @xmath17 and @xmath18 band images .
this prevents us from using point spread function fitting photometry .
the fwhm of the near - infrared images is @xmath47 ( 6 pixels ) .
we use a @xmath47 aperture to measure the flux from the wd only , and a @xmath46 aperture to measure the total flux from the two sources .
the @xmath48 field of view of mmirs enables us to use about 20 nearby 2mass stars to calibrate the photometry .
we checked the photometry for several 2mass stars using @xmath47 and @xmath46 apertures , and found them to be consistent within the errors .
we measure @xmath49 , @xmath50 , and @xmath51 mag for the wd ( using a @xmath47 aperture ) , and @xmath52 , @xmath53 , and @xmath54 mag for the total flux in a @xmath46 aperture .
the latter are consistent with 2mass photometry for hs
1653 + 7753 within the errors .
( lower points ) and 3.6@xmath55 ( upper points ) apertures is shown .
the solid lines show the expected photospheric flux from the wd .
the dashed lines show the combined flux from the wd and 1600 k - 2000 k brown dwarf companions ( top panel ) , and a background object with a blackbody temperature of 2200 k ( bottom panel ) . ]
figure 5 presents the spectral energy distribution of hs 1653 + 7753 .
we show the near - infrared photometry for both @xmath47 and @xmath46 apertures , and the mid - infrared photometry for a @xmath46 aperture .
the combined flux from the wd and possible brown dwarf companions @xcite are shown in the top panel .
if the two objects are bound , the observed mid - infrared flux excess implies a relatively warm brown dwarf .
however , the excess flux ( at 4.5 @xmath8 m ) is not explained well by these models .
the bottom panel in figure 5 presents the combined flux from the wd and a background object with a blackbody temperature of 2200 k. the mid - infrared flux excess is explained better by a 2200 k background object .
therefore , the resolved object near the wd is most likely a background source .
hs 1653 + 7753 has @xmath56 and @xmath57 mas yr@xmath58 in the usno - b catalog @xcite . using 2mass and mmirs @xmath9-band images
, we measure @xmath59 and @xmath60 mas yr@xmath58 .
if hs 1653 + 7753 and the nearby source are not bound , they will be moving further apart .
follow up @xmath9-band observations in a few years will show if this system is bound .
pg 2257 + 162 is classified as a tentative wd + low - mass ms binary by @xcite based on its 2mass near - infrared colors .
it shows significant flux excesses in the near- and mid - infrared , and it is not resolved in the irac images .
the spectral energy distribution of this wd is presented in figure 6 .
it is clear from this figure that the excess comes from an object with a blackbody temperature of 3300 k. if the unresolved companion is physically associated with the wd , it is most likely a low - mass stellar companion .
we are currently obtaining radial velocity observations of this wd using the 1.5 m tillinghast telescope to see if this system is a physical binary or not .
our irac survey of 14 low - mass wds did not reveal any substellar or planetary mass companions .
one of our targets , hs 1653 + 7753 , has a nearby source that is likely to be a background object . another target ,
pg 2257 + 162 , possibly has a low - mass stellar companion .
in addition , the irac photometry for one of our targets , wd 0950@xmath5572 is affected by a nearby star . the remaining 11 low - mass wds do not show any mid - infrared flux excesses .
adding the three low - mass wds observed by @xcite and @xcite to our sample brings the sample size to 14 wds with no detectable mid - infrared flux excess .
eleven of these targets do not show radial velocity variations , ruling out companions more massive than @xmath61 @xcite .
therefore , these 11 low - mass wds appear to have no stellar ( including another wd ) or substellar companions to the limits given in table 3 .
the non - detection of stellar , substellar , or planetary companions to our targets is not inconsistent with the idea that single low - mass wds descend from single metal - rich stars through extreme mass loss .
@xcite predict that if a substellar object with @xmath62 goes through a common - envelope phase with a star , it will either evaporate or spiral - in during the process .
more massive objects are expected to survive this evolution , as evidenced by wd 0137@xmath5049b
( @xmath63 , * ? ? ? * ) , although the theoretical 32 @xmath22 limit needs to be tested by observations .
@xcite use wd 1614 + 136 to demonstrate that their common envelope phase scenario is feasible .
they assume that wd 1614 + 136 is a remnant of a 1 @xmath64 star with a core - mass of 0.33@xmath64 .
they find that the minimum companion mass required to expel the envelope of the progenitor is 21 @xmath22 .
no companion is detected in irac observations of this star , and @xcite rule out @xmath65 companions for ages @xmath66 gyr .
this limit goes up to about 40 @xmath22 for 5 gyr old companions .
a well tuned scenario involving a common envelope phase between a @xmath675 gyr old 1 @xmath64 star and a 21 - 40 @xmath22 brown dwarf at a certain orbital separation can explain the single , low - mass wd 1614 + 136 .
our survey is sensitive to massive brown dwarf companions around the majority of our targets even at 5 gyr ( see table 3 ) .
we rule out @xmath68 companions around 8 of our targets at 5 gyr . like wd 1614 + 136 ,
only a well tuned common - envelope phase scenario involving 20 - 40 @xmath22 companions would explain the low mass wds in our sample .
@xcite find that the driest part ( the minimum number ) of the brown dwarf desert occurs at 31 @xmath22 .
thus , it seems unlikely that the scenario proposed by @xcite will explain all single , low - mass wds .
however , given the caveats in understanding the common - envelope phase evolution and our age dependent limits on possible companions , we can not rule out this scenario . an alternative scenario for the formation of single low - mass wds involves binary mergers of even lower mass wds @xcite .
@xcite find that such mergers result in rotational velocities on the order of 1000 km s@xmath58 , unless the wds are efficient in loosing angular momentum through mass loss during the merging process .
@xcite measure rotational velocities of 50 km s@xmath58 for two single low - mass wds and they suggest that efficient angular momentum loss should take place if these wds formed through binary mergers .
@xcite find this formation scenario unlikely .
currently , there are no known binary low - mass wd progenitor systems with combined mass @xmath69 and with merger time shorter than a hubble time ( see table 1 in * ? ? ?
. radial velocity studies of extremely low - mass ( @xmath70 ) wds are therefore important to search for such systems @xcite and to see if this scenario is likely to explain at least a fraction of single , low - mass wds .
support for this work was provided by nasa through the spitzer space telescope fellowship program , under an award from caltech .
we thank the mmirs commissioning team for obtaining the observations , and t. von hippel for a careful reading of an earlier version of this manuscript .
we also thank an anonymous referee for useful suggestions .
baraffe , i. , chabrier , g. , barman , t. s. , allard , f. , & hauschildt , p. h. 2003 , , 402 , 701 bedin , l. r. , salaris , m. , piotto , g. , cassisi , s. , milone , a. p. , anderson , j. , & king , i. r. 2008 , , 679 , l29 bergeron , p. , wesemael , f. , & beauchamp , a. 1995 , , 107 , 1047 burleigh , m. r. , clarke , f. j. , & hodgkin , s. t. 2002 , , 331 , l41 burleigh , m. r. , et al .
2008 , , 386 , l5 burrows , a. , sudarsky , d. , & lunine , j. i. 2003 , , 596 , 587 burrows , a. , sudarsky , d. , & hubeny , i. 2006 , , 640 , 1063 debes , j. h. , & sigurdsson , s. 2002 , , 572 , 556 debes , j. h. , sigurdsson , s. , & hansen , b. 2007 , , 134 , 1662 duncan , m. j. , & lissauer , j. j. 1998 , icarus , 134 , 303 farihi , j. , becklin , e. e. , & zuckerman , b. 2005 , , 161 , 394 farihi , j. , becklin , e. e. , & zuckerman , b. 2008 , , 681 , 1470 farihi , j. , jura , m. , & zuckerman , b. 2009 , , 694 , 805 fischer , d. a. & valenti , j. 2005 , , 622 , 1102 friedrich , s. , zinnecker , h. , correia , s. , brandner , w. , burleigh , m. , & mccaughrean , m. 2006 , asp conference series , 999 grether , d. , & lineweaver , c. h. 2006 , , 640 , 1051 hansen , b. m. s. , kulkarni , s. , & wiktorowicz , s. 2006 , , 131 , 1106 hogan , e. , burleigh , m. r. , & clarke , f. j. 2009 , , in press iben , i. j. , tutukov , a. v. , & yungelson , l. r. 1997 , , 475 , 291 ignace , r. 2001 , , 113 , 1227 jura , m. 2003 , , 584 , l91 kalirai , j. s. , bergeron , p. , hansen , b. m. s. , kelson , d. d. , reitzel , d. b. , rich , r. m. , & richer , h. b. 2007 , , 671 , 748 kilic , m. , brown , w. r. , allende prieto , c. , pinsonneault , m. h. , & kenyon , s. j. 2007a , , 664 , 1088 kilic , m. , stanek , k. z. , & pinsonneault , m. h. 2007b , , 671 , 761 kilic , m. , brown , w. r. , allende prieto , c. , swift , b. , kenyon , s. j. , liebert , j. , & ageros , m. a. 2009a , , 695 , l92 kilic , m. , gould , a. , & koester , d. 2009b , , 705 , 1219 koester , d. , rollenhagen , k. , napiwotzki , r. , voss , b. , christlieb , n. , homeier , d. , & reimers , d. 2005 , , 432 , 1025 landolt , a. u. 1992 , , 104 , 340 lpine , s. , & shara , m. m. 2005 , , 129 , 1483 liebert , j. , bergeron , p. , & holberg , j. b. 2005 , , 156 , 47 livio , m. , & soker , n. 1983 , , 125 , l12 maxted , p. f. l. , & marsh , t. r. 1998 , , 296 , l34 maxted , p. f. l. , marsh , t. r. , & moran , c. k. j. 2000 , , 319 , 305 maxted , p. f. l. , napiwotzki , r. , dobbie , p. d. , & burleigh , m. r. 2006 , , 442 , 543 mccook , g. p. , & sion , e. m. 2006 , vizier online data catalog , 3235 , 0 mcleod , b. a. , fabricant , d. , geary , j. , martini , p. , nystrom , g. , elston , r. , eikenberry , s. s. , & epps , h. 2004 , , 5492 , 1306 melis , c. , zuckerman , b. , song , i. , rhee , j. h. , & metchev , s. 2009 , , 696 , 1964 monet , d. g. , et al . 2003 , , 125 , 984 mullally , f. , kilic , m. , reach , w. t. , kuchner , m. j. , von hippel , t. , burrows , a. , & winget , d. e. 2007 , , 171 , 206 mullally , f. , reach , w. t. , degennaro , s. , & burrows , a. 2009 , , 694 , 327 napiwotzki , r. , et al .
2001 , astronomische nachrichten , 322 , 411 napiwotzki , r. , et al . 2007 ,
15th european workshop on white dwarfs , 372 , 387 nelemans , g. , & tauris , t. m. 1998 , , 335 , l85 nelemans , g. , & tout , c. a. 2005 , , 356 , 753 parriott , j. , & alcock , c. 1998 , , 501 , 357 reach , w. t. , et al .
2005 , , 117 , 978 reid , i. n. , turner , e. l. , turnbull , m. c. , mountain , m. , & valenti , j. a. 2007 , , 665 , 767 wachter , s. , hoard , d. w. , hansen , k. h. , wilcox , r. e. , taylor , h. m. , & finkelstein , s. l. 2003 , , 586 , 1356 winget , d. e. , et al . 2003 , scientific frontiers in research on extrasolar planets , 294 , 59
|
the formation scenarios for _ single _ low - mass ( @xmath0 ) white dwarfs include enhanced mass loss from a metal - rich progenitor star or a common envelope phase of a solar - like star with a close - in massive planet or a brown dwarf .
both scenarios suggest that low - mass white dwarfs may have planets . here
, we present a _ spitzer _ irac search for substellar and planetary mass companions to 14 low - mass white dwarfs .
one of our targets , hs 1653 + 7753 , displays near- and mid - infrared flux excess .
however , follow - up mmt observations show that this excess is due to a nearby resolved source , which is mostly likely a background object .
another target , pg 2257 + 162 , shows flux excess compatible with a late - type stellar companion .
we do not detect substellar companions to any of the remaining targets .
in addition , eight of these stars do not show any radial velocity variations , ruling out stellar mass companions including other white dwarfs .
we conclude that a significant fraction of the low - mass white dwarfs in our sample do not have stellar or massive brown dwarf companions .
|
despite considerable experimental and theoretical efforts over the last few decades , unconventional superconductivity remains one of the most interesting open puzzles in solid - state physics .
while it has been proposed that in some unconventional superconductors several interactions may be responsible for superconductivity , spin fluctuations are argued by a large fraction of the solid - state community to be the dominant mechanism driving cooper pairing in heavy - fermion systems , cuprates , two - dimensional organic charge - transfer salts and the iron - based superconductors @xcite .
such a statement is based on a few properties that have been found to be characteristic for these material classes like the fact that the superconducting phase is located in close vicinity to an antiferromagnetic ordered state and the phase transition can be easily tuned by application of pressure or charge doping @xcite .
assuming spin , charge , and orbital fluctuations provide the dominant mechanism driving superconductivity , various theoretical approaches have been developed in the last few decades in order to predict superconducting gap functions , critical temperatures , and functional characteristics of thermodynamic quantities .
here we shall focus on the random phase approximation ( rpa ) for the hubbard model . in the single - orbital case on a three - dimensional cubic lattice this approach leads to a strong enhancement of the singlet coupling in the proximity of a spin - density - wave instability yielding to a @xmath0 symmetry of the gap function @xcite as observed , for instance , in the high - t@xmath1 cuprates @xcite .
also for superconducting organic charge - transfer salts @xcite this approximation seems to correctly predict the symmetry of the superconducting gap@xcite .
recently , such an approximation has been successfully used for the determination of the superconducting gap structure in iron - based superconductors and heavy - fermion systems @xcite . while the properties of high - t@xmath1 cuprates or organic charge - transfer salts are well captured by single - orbital models @xcite , it is crucial to consider multiorbital models@xcite when aiming for a proper description of heavy - fermion systems and iron - based superconductors @xcite . here
the fermi surfaces consist of several sheets , which emerge from various orbitals at the fermi level that participate in the interactions driving superconductivity .
the matrix - rpa formulation for the single- and multiorbital hubbard model , like other rpa approaches , is built around the idea that terms involving particle - hole susceptibilities associated with a given wave vector @xmath2 and frequency @xmath3 add up coherently , while other terms have a _ random phase _ which suppresses their contribution .
neglection of these _ random phase _
terms in the calculation of the two - particle pairing vertex in the single - orbital case results in a summation of infinite orders of diagrams of pure bubble and ladder topology , yielding an easy - to - evaluate scalar closed - form expression for the pairing interaction . in the multiorbital case , the susceptibilities acquire a complex orbital dependence , and the interaction strength has to be replaced by full interaction matrices operating on the different orbitals . in this work
we examine the relation between ( i ) the diagrammatic representation of the random phase approximation of the two - particle vertex function in the context of the multiorbital hubbard model considering inter- and intraorbital couplings and ( ii ) the corresponding matrix representation of the fluctuation exchange approximation @xcite .
the latter has been very commonly used to analyze superconductivity in multiband systems , with the assumption that it was equivalent to the usual bubble and ladder diagrams that occur in the one - band rpa .
in contrast to the derivation based on the separation into the different fluctuation channels @xcite , we follow the original diagrammatic approach @xcite and write down all low - order contributions to the two - particle vertex that can be rewritten as products of the interactions and the bare particle - hole susceptibilities . applying this prescription to the single - orbital case
, we recover the well - known result for the interaction vertex where only diagrams of pure bubble or ladder topology are incorporated .
in contrast , in the multiorbital case we find that , in addition to the diagrams that have the same topology as in the single - orbital case , there are additional diagrams that have the structure of vertex corrections .
we analyze the contributions of such diagrams to certain elements of the pairing interaction and discuss their physical importance for the superconducting state .
understanding the relationship between the usual diagrammatic perturbation - resummation approach and the rpa matrix formulation provides insight into the physical processes that are captured in the latter .
we are interested in relating the matrix - rpa form of the pairing vertex illustrated in fig .
[ fig : rpa](a ) to its diagrammatic representation . to illustrate our basic point , it is sufficient to consider a two - orbital problem , where the noninteracting part of the hamiltonian @xmath4 is diagonal in spin @xmath5 but allows for hopping between the two orbitals ( @xmath6 or @xmath7 ) .
our discussion can easily be generalized to more than two orbitals .
the noninteracting hamiltonian is given by @xmath8 for the interacting part of our model hamiltonian @xmath9 we first restrict ourselves to the intra- and interorbital coulomb interactions @xmath10 and @xmath11 , as the effects we wish to discuss are included : @xmath12 we will discuss later what happens when hund s rule exchange @xmath13 and pair hopping @xmath14 are included . for this model , the matrix - rpa pairing interaction in the singlet channel derived in the fluctuation exchange ( flex ) approximation @xcite in the vicinity of the critical temperature , where the anomalous green s function vanishes , has the following form : @xmath15_{l_1l_2l_3l_4}. \end{split}\end{aligned}\ ] ] the spin and charge ( also commonly called _ orbital _ ) interaction matrices @xmath16 and @xmath17 , respectively , are @xmath18 matrices with indices @xmath19 . their explicit form is given by @xmath20 [ eq : interactionmatrices ] the rpa spin ( @xmath21 ) and charge ( @xmath22 ) susceptibilities for total spin zero
have the form @xmath23^{-1 } \chi^0(q)\\ \chi_0^{\textnormal{rpa}}(q)=&\left[1+u^c \chi^0(q)\right]^{-1 } \chi^0(q)\end{aligned}\ ] ] [ eq : rpasusceptibilities ] where the matrix elements of the bare susceptibility @xmath24 are given by @xcite ( see fig . [
fig : rpa](b ) ) : @xmath25 here @xmath26 denotes the bare green s function @xmath27 for an electron with momentum @xmath28 and matsubara energy propagating between the @xmath29 and @xmath30 orbitals .
the weights @xmath31 are the @xmath29 elements of the eigenvectors with eigenvalue @xmath32 , which are determined by diagonalization of the noninteracting tight - binding hamiltonian ( eq . [ eq : noninthamil ] ) . , ( b ) @xmath33 , and ( c ) @xmath34 . ]
we begin our study of the relationship between the matrix formulation of the rpa pairing interaction @xmath35 ( eq . [ eq : singletvertex ] ) and the usual diagrammatic perturbation theory for @xmath36 by examining two second - order contributions to @xmath37 , namely , the index combinations @xmath38 , @xmath39 and @xmath40 , @xmath41 using the standard diagrammatic rules for @xmath42 , the last three diagrams shown in fig . [ fig : combined](a ) correspond to the last three terms in eq .
( [ eq : gamma1111 ] ) for @xmath43 . the first diagram in fig .
[ fig : combined](a ) gives @xmath44 . for a singlet and even frequency particle pair
the pauli principle implies a certain symmetry of the vertex function : @xmath45 as a consequence , @xmath42 is invariant under @xmath46 , and the contribution from the first diagram in fig .
[ fig : combined](a ) corresponds indeed to the first contribution to the rpa expression in eq .
( [ eq : gamma1111 ] ) .
vertex corrections which contribute to @xmath47 and @xmath48 . ]
we investigate now the contributions to @xmath49 ( eq . [ eq : gamma1121 ] ) and @xmath50 , which are related by symmetry in the singlet channel such that @xmath51 ( see eq .
[ eq : paulisymmetry ] ) . in the diagrammatic expansion of @xmath33 we have in second order a term of the form @xmath52 as shown in fig .
[ fig : combined](b ) . by symmetry
this diagram corresponds to the first term in eq .
[ eq : gamma1121 ] : @xmath53 the remaining contributions to @xmath49 ( eq . [ eq : gamma1121 ] ) are proportional to @xmath54 and @xmath55 respectively .
as the interorbital coulomb repulsion preserves spin and orbital indices at each primitive vertex , a scheme based on only bubble and ladder diagrams leads to no contributions to @xmath49 .
however , both terms in eq .
( [ eq : gamma1121 ] ) can be rewritten as shown in fig .
[ fig : combined](c ) ( left and right diagrams ) and have the meaning of vertex corrections in the standard diagrammatic expansion .
we note that the same type of vertex correction diagrams appear in functional renormalization - group calculations for the multiorbital hubbard model @xcite . for higher orders , the feynman diagrams that contribute to @xmath56 consist of the familiar bubble and particle - hole ladder diagrams as well as diagrams that can be rewritten as vertex corrections arising from @xmath11 .
examples of higher - order vertex contributions are illustrated in fig .
[ fig : vertcorr ] .
note that there is a simple way to recognize the feynman diagrams that are included in the matrix - rpa : if one can assign the internal propagators into pairs , such that the first propagator is created at the same interaction vertex at which the second propagator is annihilated and vice versa , then each pair corresponds to the bare susceptibility as given in eq .
( [ eq : baresuscep ] ) and the contribution of the considered feynman diagram to the pairing interaction can be entirely written in terms of the interactions and the bare susceptibilities . if a hund s rule exchange @xmath13 and a _ pair hopping _
term @xmath14 are included in the multi - orbital interaction ( see , e.g. , refs . ) , the interaction part of the hamiltonian in real space is given by @xmath57 and additional vertex terms appear in the diagrammatic expansion of the pairing interaction .
we analyze the contributions to the interorbital pair - scattering process following ref .
: we retain only the diagonal terms of the bare susceptibilities @xmath58 with @xmath59 , assuming that other terms are suppressed by matrix element effects ( see also eq .
[ eq : baregreensfunc ] ) .
we find that the dominant second - order contribution ( @xmath60 ) to the pair - scattering term @xmath61 originates in equal parts from ladder and vertex diagrams @xcite .
summing higher - order terms as shown in fig .
[ fig : jp ] we obtain the dominant term linear in @xmath14 : @xmath62 which contributes to @xmath63 ( compare also to the results of the three - orbital matrix - rpa as discussed in ref . ) . for inversion - symmetric systems the two contributions in eq .
( [ eq : leadingtermjprime ] ) are equal , as the bare susceptibility ( eq . [ eq : baresuscep ] ) obeys @xmath64 .
a numerical analysis of the importance of these interorbital pair - scattering terms @xmath61 , which we have shown to contain contributions arising from vertex correction diagrams , for a realistic model of the iron pnictides has recently been conducted by kemper _
et al . _
@xcite . there
, the authors found that in the presence of a nonvanishing pair hopping these terms critically influence the nodal structure of the superconducting gap ( see also fig .
9 of ref . )
moreover , another recent study by kontani and onari @xcite has shown that exactly the same terms may play a role in enhancing the electron - phonon coupling .
diagrams involving one factor of the pair hopping @xmath14 are summed they yield the leading @xmath14 result . ]
based on a simple two - orbital hubbard model , we have analyzed the relationship between the matrix rpa and the feynman diagram technique for the calculation of the superconducting pairing interaction . in the single - orbital case ,
the diagrams corresponding to the matrix - rpa have been found to have a simple well - known structure containing only bubble or ladder subunits . in the multiorbital case ,
the interorbital coulomb interaction @xmath11 and an exchange pair - hopping interaction @xmath14 generate additional diagrams that , as we have shown , have the structure of vertex corrections .
the matrix rpa in the context of the multiorbital hubbard model is therefore a significant generalization of the commonly used random phase approximation : it comprises infinite - order diagrams that can be written as products of the interactions and the bare susceptibilities .
these include both the usual particle - hole bubble or ladder contributions , and the vertex corrections discussed here , all of which may be summed to infinite order in the usual rpa manner .
we have therefore shown that the commonly used matrix rpa expression for the pairing vertex in the multiorbital hubbard model provides a hitherto unappreciated method to keep additional contributions to infinite order beyond rpa .
these additional diagrams are important in view of recent studies , as they significantly influence the node structure of the superconducting gap function @xcite and may also enhance phonon - mediated interactions @xcite .
we would like to thank k. zantout , a. rmer , y. wang , p. lange , p. kopietz and c. gros for useful discussions .
, d.g . and r.v .
thank the german research foundation ( deutsche forschungsgemeinschaft ) for support through grants sfb / tr49 and spp1458 .
p.j.h . acknowledges support through department of energy de - fg02 - 05er46236 .
t.a.m . and d.j.s .
acknowledge support through the center for nanophase materials science at ornl , which is sponsored by the division of scientific user facilities , u.s .
m.a . and r.v .
further acknowledge partial support by the kavli institute for theoretical physics at the university of california , santa barbara under nsf grant no .
phy11 - 25915 .
hirschfeld , _ using gap symmetry and structure to reveal the pairing mechanism in fe - based superconductors _ , comptes rendus physique * 17 * , 197 ( 2016 ) .
r. h. mckenzie , _ similarities between organic and cuprate superconductors _ , science * 278 * , 5339 ( 1997 ) .
d. guterding , s. diehl , m. altmeyer , t. methfessel , u. tutsch , h. schubert , m. lang , j. mller , m. huth , h. o. jeschke , r. valent , m. jourdan , h .- j .
elmers , _ evidence for eight node mixed - symmetry superconductivity in a correlated organic metal _ , phys .
116 * , 237001 ( 2016 ) .
a. f. kemper , t. a. maier , s. graser , h .-
cheng , p. j. hirschfeld , and d. j. scalapino , _ sensitivity of the superconducting state and magnetic susceptibility to key aspects of electronic structure in ferropnictides _ , new j. phys .
* 12 * , 073030 ( 2010 ) .
y. wang , a. kreisel , v. b. zabolotnyy , s. v. borisenko , b. bchner , t. a. maier , p. j. hirschfeld , and d. j. scalapino , _ superconducting gap in lifeas from three - dimensional spin - fluctuation pairing calculations _ , phys .
b * 88 * , 174516 ( 2013 ) .
k. suzuki , h. usui , s , iimura , y. sato , s. matsuishi , h. hosono , and k. kuroki , _ model of the electronic structure of electron - doped iron - based superconductors : evidence for enhanced spin fluctuations by diagonal electron hopping _ ,
* 113 * , 027002 ( 2014 ) .
d. guterding , h. o. jeschke , p. j. hirschfeld , and r. valent , _ unified picture of the doping dependence of superconducting transition temperatures in alkali metal / ammonia intercalated fese _ ,
b * 91 * , 041112(r ) ( 2015 ) .
h. arai , h. usui , k. suzuki , y. fuseya , and k. kuroki , _ theoretical study of correlation between spin fluctuations and @xmath70 in isovalent - doped 1111 iron - based superconductors _
b * 91 * , 134511 ( 2015 ) .
k. kuroki , h. usui , s. onari , r. arita , and h. aoki , _ pnictogen height as a possible switch between high - t@xmath1 nodeless and low - t@xmath1 nodal pairings in the iron - based superconductors _ ,
b * 79 * , 224511 ( 2009 ) .
t. maehira , t. hotta , k. ueda , and a. hasegawa , _ relativistic band - structure calculations for cetin@xmath73 ( t = ir and co ) and analysis of the energy bands by using tight - binding method _ , j. phys .
. jpn . * 72 * , 854 ( 2003 ) .
f. wang , h. zhai , y. ran , a. vishwanath , and d .- h .
lee , _ functional renormalization - group study of the pairing symmetry and pairing mechanism of the feas - based high - temperature superconductor _ , phys .
102 * , 047005 ( 2009 ) .
|
in the framework of a multiorbital hubbard model description of superconductivity , a matrix formulation of the superconducting pairing interaction that has been widely used is designed to treat spin , charge , and orbital fluctuations within a random phase approximation ( rpa ) . in terms of feynman diagrams , this takes into account particle - hole ladder and bubble contributions as expected .
it turns out , however , that this matrix formulation also generates additional terms which have the diagrammatic structure of vertex corrections .
here we examine these terms and discuss the relationship between the matrix - rpa superconducting pairing interaction and the feynman diagrams that it sums .
|
understanding topological order in extended systems is one of the major challenges in modern physics .
such an issue has immediate spinoff in condensed matter physics@xcite , but encompasses important aspects of quantum information as well@xcite .
topologically ordered systems are ( generically ) gapped systems characterized by a specific degeneracy of the ground states@xcite .
the main attraction for quantum computation applications relays on the intrinsic robustness of such an order to external perturbations .
indeed such a property is nothing but a rephrasing that no local order parameter can be defined to characterize topologically ordered states ( elitzur theorem ) .
this turns the main virtue of topological phases into a bottleneck , because any search for such kind of order in actual physical systems is problematic . for the same reason ,
also more generic spin liquids , that may be not topologically ordered , but nevertheless possess a gap and no local order parameter , are long - sought states in nowadays condensed matter physics .
@xcite in the last few years , topological order in many - body system has been studied resorting to new approaches .
it is believed that topological order can be characterized by the entanglement encoded in the states of the system@xcite .
more precisely , it is understood that such an order is related to a long range entanglement , meaning that topological states can not be adiabatically connected to non - topological ones resorting to quantum circuits made of local ( on the scale of the range of interaction in the underlying hamiltonian ) unitary gates@xcite .
accordingly , any observable that would be able , in principle , to detect the topological order , is intrinsically non local@xcite .
here we push forward the idea that progress can be made on this aspect , by analyzing how topologically ordered ground states change by varying the control parameter , within the quantum phase ( instead of looking at a single copy of ground state corresponding to a fixed value of the control parameter)@xcite .
in particular , we explore whether the ground state ` evolution ' by changing the hamiltonian s control parameter @xmath3 can be achieved by locc , after having bi - partitioned system .
the tool we exploit is provided by the differential local convertibility@xcite : a given physical system is partitioned into two parties , @xmath4 and @xmath5 , limited to local operations on their subsystems and classical inter - party communication ( locc ) .
we comment that the notion of locality we refer does depend on the partition that has been employed ; therefore the locc can indeed involve a portion of the system that can be very non - local on the scale fixed by the interactions in the hamiltonian .
this should be contrasted with the protocols defining the topological order in terms of local unitary transformations mentioned above@xcite .
assuming that @xmath4 and @xmath5 can share an entangled state ( ancilla ) , the differential local convertibility protocol can be feasibly expressed through a specific behavior of the rnyi entropies @xmath6 .
the differential convertibility holds if and only if @xmath7 , @xmath8 , or @xmath9 , @xmath8@xcite . in the present paper
we will be using the latter characterization .
such an approach was first discussed in the realm of quantum critical phenomena in the ref.@xcite . in particular , we note that paramagnets and phases with non vanishing local order parameters are indeed locally convertible@xcite . in this paper , we focus on spin systems in one spatial dimension , where topological order is protected by the symmetry of the system@xcite .
we shall see that such symmetry - protected - topological - phases are not convertible .
we shall see that such property holds on spatial scales that are smaller than the correlation length of the system . to address the question
, we refer to specific spin systems providing paradigmatic examples in this context .
first we will consider the cluster - ising model@xcite ( see also@xcite ) ) .
the physical platform for that is provided by cold atoms in a triangular optical lattice @xcite .
the model is particularly interesting in quantum information since it describes how the @xmath0-cluster states are quenched by a qubit - qubit exchange interaction@xcite .
our second example is the @xmath10 model , that is a well known model for studying the haldane order in @xmath0 quantum magnets@xcite .
quantum computation protocols based on haldane - type states have been provided in@xcite . the main numerical tool for the analysis is the dmrg technique with matrix product states ( mps ) variational ansatz@xcite .
for both the cluster - ising and the @xmath11-d models , we analyze rnyi entropies , entanglement spectrum and differential local convertibility . in sections
[ convert_clusterising ] and [ convert_lambda - d ] the differential local convertibility of the cluster - ising and @xmath10 will be analysed .
the scenario emerging from our study will be discussed in section [ discussion ] . in the appendix [ app_clusterising ] , [ app_lambda - d ] , we discuss edge states , correlation lengths and string order paramaters of the models we deal with .
the differential local convertibility for partition sizes larger than the correlation length of the system will be discussed in appendix [ other_partitions ] . in the appendix [ large_alpha ]
we discuss the rnyi entropy in the large @xmath12 limit .
the hamiltonian we consider is @xmath13 where @xmath14 , @xmath15 , are the pauli matrices and , except otherwise stated , we take open boundary conditions @xmath16 .
the phase diagram of ( [ eq : ham_cluster ] ) has been investigated in @xcite . for large @xmath17
the system is an antiferromagnet with local order parameter .
for @xmath18 the ground state is a a cluster state .
it results that the correlation pattern characterizing the cluster state is robust up to a critical value of the control parameter , meaningfully defining a cluster phase " with vanishing order parameter and string order@xcite . without symmetry , the cluster phase is a ( non topological ) quantum spin liquid , since there is a gap and no symmetry is spontaneously broken .
protected by a @xmath19 symmetry , the cluster phase is characterized by a topological fourfold ground state degeneracy , reflecting the existence of the edge states ( see appendix [ app_clusterising])@xcite . in the dmrg , we resolve the ground state degeneracy , by adding a small perturbation @xmath20 to the hamiltonian .
we find that the symmetric partition @xmath21 displays local convertibility , fig.[convert_cluster ] : @xmath22 .
this is indeed a fine tuned phenomenon since the cluster phase results non - locally convertible , for a generic block of spins , both of the type @xmath23 and the @xmath24 , fig.[convert_cluster ] .
we remark that such a property holds even for size region @xmath4 smaller than the correlation length ( see appendix [ other_partitions ] for other partitions ) .
indeed , the entanglement spectrum is doubly degenerate in all the cluster phase , as far as the size of the blocks @xmath4 and @xmath5 are larger than the correlation length . in contrast , the antiferromagnet is locally convertible , with non degenerate entanglement spectrum . ) .
we characterize the differential local convertibility in terms of the the slopes of the rnyi entropies .
panel ( a ) is for bipartition @xmath21 , @xmath25 .
there is differential local convertibility throughout the two different phases because , for fixed @xmath17 , @xmath26 does not change sign with @xmath12 . panel ( b ) is for bipartition @xmath23 , @xmath27 , @xmath28 .
panel ( c ) is for @xmath29 , being one blocks @xmath30 with @xmath31 , @xmath32 , and @xmath33 . in all of these cases
, @xmath26 changes sign .
@xmath34 , @xmath35 , and @xmath36 the larger and the smaller eigenvalues of reduced density matrix @xmath37 , respectively ; @xmath38 . in convertible phases , we observe that the change in the largest eigenvalues is faster than the rate at which the smallest eigenvalues are populated .
in contrast the non differential local convertibility arises because the sharpening of the first part of the spectrum is over - compensated by the increasing of the smallest @xmath37 . ]
in this section , we study the local convertibility of the @xmath10 model hamiltonian describing an interacting spin-@xmath39 chain with a single ion - anisotropy @xmath40.\end{aligned}\ ] ] where @xmath41 , @xmath42 are spin-@xmath39 operators : @xmath43 and @xmath44 .
the hamiltonian above enjoys several symmetries : the well known @xmath45 , and the link inversion symmetry @xmath46 ( see appendix [ app_lambda - d ] ) .
the phase diagram has been investigated by many authors@xcite ( see appendix [ app_lambda - d ] ) .
here we consider @xmath47 . for small / large @xmath48 and fixed @xmath11 ,
the system is in a polarized state along @xmath49 or @xmath50 , respectively . for large @xmath11 and
fixed @xmath48 , the state displays antifferomagnetic order . at intermediate @xmath48 and @xmath11 ,
the state is a diluted anti - ferromagnet with strong quantum fluctuations , defining the haldane phase , lacking of local order parameters and string order . for open boundary conditions ( we apply in the present paper ) , the haldane ground state displays a fourfold degeneracy , that can not be lifted without breaking the afore mentioned symmetry of the hamiltonian
this is the core mechanism defining the haldane phase as a symmetry - protected - topological ordered phase@xcite . without symmetry ,
the ground state is gapped and no symmetry is spontaneously broken , making the haldane phase a quantum spin liquid .
[ sweeps ] and change @xmath48 ; the haldane phase is approximately located in the range @xmath51 . _ 2 ) _ fix @xmath52 , varying on @xmath11 ; the haldane phase is located in the range @xmath53 .
, title="fig:",height=207 ] we sweep through the phase diagram in the two following two ways : _ 1 ) _ fix @xmath54 and change @xmath48 ; the haldane phase is approximately located in the range @xmath51 . _ 2 ) _ fix @xmath52 , varying on @xmath11 ; the haldane phase is located in the range @xmath53 ( see fig.[sweeps ] ) .
we analyzed all the four states separately adding the perturbation to the hamiltonian @xmath55 with a small coupling constant .
we find that the nel , ferromagnetic , and the large @xmath48 phases are locally convertible ( seefig.[conver_lambdad_symm ] : @xmath56 , @xmath57 ) .
consistently with @xcite , all of the haldane ground states are characterized by doubly degenerate entanglement spectrum for the symmetric @xmath23 partitions with @xmath58 , for both sweep ways ( fig.[conver_lambdad_symm ] : @xmath59 and @xmath60 ) ( see @xcite for a recent progress on the understanding of double degenerate entanglement spectrum ) .
such a property is not recovered both in the cases of asymmetric @xmath23 and @xmath61 partitions the entanglement spectrum is not found doubly degenerate , because the state of the system breaks the link inversion symmetry@xcite ( fig.[conver_lambdad_symm ] : @xmath62 , @xmath63 ) .
see @xcite for the analysis of the entanglement spectrum close to the quantum phase transitions .
we find that the haldane phase is not locally convertible ( see fig.[conver_lambdad_symm ] : @xmath64 , @xmath65 , [ fig_sweep1 ] , [ fig_sweep2 ] ) .
we remark that for both ways to partition the system the non - local - convertibility phenomenon is found even in the case of sizes of @xmath5 smaller than the correlation length @xmath66(see [ app_lambda - d ] for the behavior of @xmath66 ) . as for the model eq.[eq : ham_cluster ] , we find that the symmetric bipartition @xmath58 displays local convertibility as a fine - tuned effect , that is broken for generic partitions ( see appendix [ other_partitions ] for other partitions ) . .
the sweep ( 1 ) through the @xmath10 phase diagram is considered ( see also appendix [ app_lambda - d ] for the schematic phase diagram).the upper panel displays the results for the symmetric case @xmath21 .
the bottom panel refers to the antisymmetric case @xmath67 , @xmath68 .
the reny entropies are presented in @xmath56 , @xmath64 .
the signs distributions of the derivatives of the rnyi entropies are shown in @xmath57 , @xmath65 . and
eigenvalues of reduced density matrix @xmath37 and the entanglement spectrum are shown in @xmath59,@xmath60 , @xmath62 , @xmath63 as in fig.[convert_cluster ] .
the features of differential local convertibility are characterized by the slopes of the reny entropies and correspond to specific features of the entanglement spectrum as explained in fig.[convert_cluster ] . ]
sweep ( 1 ) through the @xmath10 model : @xmath54 , @xmath69 ( the schematic phase diagram is reported in appendix [ app_lamda - d ] ) .
the sign distribution of the derivative of the rnyi entropies @xmath70 for partitions @xmath61 , @xmath31 and @xmath68 ( @xmath71 ) presented in @xmath57 .
the features of differential local convertibility are characterized by the slopes of the reny entropies and correspond to specific features of the entanglement spectrum as explained in fig.[convert_cluster ] .
the @xmath72 are presented in @xmath56 for @xmath73 decreasing from top to low .
all such quantitates are calculated for the ground state in @xmath74 sector . ]
sweep ( 2 ) through the @xmath10 model : @xmath52 , @xmath75 ( see fig.[sweeps ] for a schematic phase diagram ) .
the sign distribution of the derivativative of the rnyi entropies @xmath76 for partitions @xmath61 , @xmath31 and @xmath68 ( @xmath71 ) presented in @xmath57 .
the features of differential local convertibility are characterized by the slopes of the reny entropies and correspond to specific features of the entanglement spectrum as explained in fig.[convert_cluster ] .
the @xmath72 are presented in @xmath56 for @xmath77 increasing from low to top .
all such quantitates are calculated for the ground state in @xmath74 sector . ]
we explored quantitatively the notion of locc in @xmath0-topologically ordered systems . in particular , we analyzed to what extent different ground states of the hamiltonian within the topological phase can be ` connected ' by ( entanglement assisted)-locc between two parts @xmath4 , @xmath5 in which the system has been divided .
this issue is analyzed through the notion of differential local convertibility that can be expressed in terms of the properties of the renyi entropies . with this tool
, we claim that progress can be made to detect quantum phases without local order parameter ( contributing to the long - sought hunt for quantum spin liquids topologically ordered or not)@xcite . in the present paper we specifically analyzed the question whether this method is useful to study topologically ordered phases assisted by symmetries . to this end
, we considered two very different models that are , at the same time , paradigmatic for the analysis of the notion of symmetry - protected - topological order : the spin-@xmath1 cluster - ising and the spin-@xmath39 @xmath10 chains .
we find the symmetry - protected - topological phases are characterized by non local convertibility , meaning that the hamiltonian is more effective to drive the system through different topological states than locc between @xmath4 and @xmath5 .
the phases with local order parameter turn out locally convertible .
the convertibility property is encoded in the specific response of the entanglement spectrum to the perturbation : the distribution of eigenvalues of the reduced density matrix get sharpened in the most represented eigenvalues ; this implies that high @xmath78rnyi entropies ca decrease ; at the same time , the spectrum acquires a tail made of small eigenvalues , because more states are involved by increasing the correlation length ( see fig.([convert_cluster ] ) @xmath79 , @xmath80 and fig.([conver_lambdad_symm ] ) @xmath81 , @xmath82 ) ) .
the local convertibility is achieved , when the sharpening is compensated by the tail of the distribution ( see fig.[convert_cluster ] , fig.[conver_lambdad_symm])@xcite .
we remark the counter intuitive phenomenon for which some quantum correlations encoded in @xmath72 , for certain @xmath12 , decrease , in - spite of the increase in correlation length ( see appendix [ app_lambda - d ] ) .
indeed , similar phenomenon was discovered in the @xmath83 toric code @xcite , corroborating the scenario described in the present paper . in particular , it was noticed that , despite non topologically ordered , also certain spin liquids result non - locally convertible for specific perturbations .
then , an interesting question that arises is concerning the stability of the local convertibility by changing the perturbation .
we observe that , on the other hand , that spin liquids with some symmetry protection , are indeed stable , spanning a well defined quantum phase , distinct from a paramagnet .
the results of the present letter indicate that the same symmetry protection is also able to protect their non local convertibility .
the non local convertibility occurs in the topological phases even for subsystems sizes _ smaller than the correlation length _ of the system ( see appendix [ app_lambda - d ] , [ app_clusterising ] for the behavior of @xmath66 in the models we analyzed ) ; the degeneracy in the entanglement spectrum , in contrast , is exhibited when the afore mentioned size is much larger than @xmath66 .
clearly , this paves the way to experimental tests through local measures on spatial region of sizes made of few spins , with the assistance of the protocols to address the rnyi entropies provided recently@xcite .
incidentally , we note that for the symmetric partition @xmath23 there is differential local convertibility , but this is a fine tuned effect that disappears if the two blocks have different size , see fig.[convert_cluster ] and fig.[conver_lambdad_symm ] . ultimately , the size and type of subsystems @xmath4 and @xmath5 on which the differential local convertibility is displayed depends on the recombination of the edge states that form at edges of the bipartition ( see @xcite for an extensive discussion of such edge state recombination phenomenon ) . with our results we can claim that the local convertibility can characterize the phase , independently of the way the system is partitioned .
we believe that such a scenario provides a valuable assistance to standard routes in experimental solid state physics to disclose topological order in the system .
our work opens several questions that will be subject of future investigation .
in particular , it is important to establish the precise relation between the ground state adiabatic evolution and the differential local convertibility@xcite .
another interesting question is the role of differential convertibility in 2d symmetry protected topologically ordered systems like topological insulators .
we thank m.c .
banuls , l. cincio , f. franchini , t. shi , h.h .
tu , s. santra and p. zanardi for discussions .
this work was supported in part by the national basic research program of china grant 2011cba00300 , 2011cba00301 the national natural science foundation of china grant 61073174 , 61033001 , 61061130540 , 11175248 , and the 973 program ( 2010cb922904 ) .
research at perimeter institute for theoretical physics is supported in part by the government of canada through nserc and by the province of ontario through mri .
the cluster - ising hamiltonian is @xmath13 without symmetry , the cluster phase in cluster - ising model is a ( non topological ) quantum spin liquid , since there is a gap and no symmetry is spontaneously broken .
protected by a @xmath19 symmetry , the cluster phase is characterized by a topological fourfold ground state degeneracy in open boundary conditions , reflecting the existence of the edge states .
such a degeneracy fans out from @xmath18 where @xmath84 majorana fermions are left free at the free ends of the chain .
the cluster phase can be characterized via a string order .
the two phases are separated by a continuous quantum phase transition with central charge @xmath85 .
indeed the hamiltonian ( [ eq : ham_cluster ] ) is equivalent to three decoupled ising chains @xcite .. after the jordan - wigner transformation @xmath86 , @xmath87 , @xmath88 , the hamiltonian of cluster - ising model can be written as @xmath89.\end{aligned}\ ] ] where @xmath90 and @xmath91 are two different majorana fermion operators .
although local order parameters do not exit to characterize the topological phase , the topological order in cluster - ising model(see fig.[stat_cluster ] ) can be detected by the edge states ( a1 ) and string order parameters ( a3 ) .
shows there is edge state in the cluster phase whereas there is no edge state in ising antiferromagnetic phase .
@xmath57 shows the correlation length of @xmath92 displaying a critical behavior .
@xmath59 is the string order parameter @xmath93 . ]
the @xmath10 hamiltonian is @xmath40.\end{aligned}\ ] ] the hamiltonian above enjoys several symmetries , including the time reversal @xmath94 , parity @xmath95 , @xmath96 generating @xmath45 , and the link inversion symmetry @xmath46 . for small / large @xmath48 and fixed @xmath11 ,
the system is in a polarized state along @xmath49 or @xmath50 , respectively . for large @xmath11 and
fixed @xmath48 , the state displays antifferomagnetic order . at intermediate @xmath48 and @xmath11 ,
the state is a diluted anti - ferromagnet with strong quantum fluctuations , defining the haldane phase .
there is no local order parameters to characterize the haldane phase in the @xmath10 model , too . with symmetry protection
, the topological order in haldane phase can be detected by the edge states and string order parameters defined in fig .
[ stat_lambdad ] ( see ref.[sorensen ] ) . without symmetry ,
the ground state is gapped and no symmetry is spontaneously broken , making the haldane phase a quantum spin liquid . in fig.[sweeps ] , we display the schematic phase diagram of the @xmath10 . model
. the sweep ( 1 ) through the @xmath10 phase diagram is considered ( see text ) . in @xmath56
we shows the haldane phase edge states ; we do not find edge states in the other phases . in @xmath57 the string order parameters @xmath97 . in @xmath59 the correlation length of @xmath98 . ]
in the main text we have shown that differential local convertibility method works well with subsystem size smaller than the correlation length . in this section
we present the results of subsystem size larger or equivalent to the correlation length ( figs.[convert_large_cluster ] , [ convert_large_lambda - d_1 ] , [ convert_large_lambda - d_2 ] ) . .
we characterize the differential local convertibility in terms of the the slopes of the rnyi entropies . @xmath26
changes sign in the cluster phase . @xmath59 and @xmath60 display the largest and the smaller eigenvalues of reduced density matrix @xmath37 , respectively ; @xmath38 . in convertible phases , we observe that the change in the largest eigenvalues is faster than the rate at which the smallest eigenvalues are populated .
in contrast the non differential local convertibility arises because the sharpening of the first part of the spectrum is over - compensated by the increasing of the smallest @xmath37 . ]
sweep ( 1 ) through the @xmath10 model : @xmath54 , @xmath69 .
the sign distribution of the derivative oconvertibilityf the rnyi entropies @xmath70 for partitions @xmath61 , @xmath99 and @xmath100 ( @xmath71 ) presented in @xmath57 .
the features of differential local convertibility are characterized by the slopes of the reny entropies and correspond to specific features of the entanglement spectrum as explained in fig.[convert_cluster ] .
the @xmath72 are presented in @xmath56 for @xmath101 decreasing from top to low .
all such quantitates are calculated for the ground state in @xmath74 sector . ]
sweep ( 2 ) through the @xmath10 model : @xmath52 , @xmath75 .
the sign distribution of the derivativative of the rnyi entropies @xmath76 for partitions @xmath61 , @xmath99 and @xmath100 ( @xmath71 ) presented in @xmath57 .
the features of differential local convertibility are characterized by the slopes of the reny entropies and correspond to specific features of the entanglement spectrum as explained in fig.[convert_cluster ] .
the @xmath72 are presented in @xmath56 for @xmath102 increasing from low to top .
all such quantitates are calculated for the ground state in @xmath74 sector . ]
in the main text , by calculating the rnyi entropies with different parameters , we have generally shown that with fixed bipartition of the spin chain , states with symmetry protected topological order can not convert to each other via locc ( assisted by entanglement ) , which is different from the stats with local order .
indeed , to arrive at such a conclusion rigorously , we have to calculate infinite rnyi entropies with @xmath12 from @xmath103 to @xmath104 .
from the definition of rnyi entropy @xmath105 we can see that if we directly calculate the rnyi entropy with very large @xmath12 numerically , the numerator and denominator are both infinitely large such that computer can not give correct results .
therefore in the @xmath106 limit , we can apply the lhospital s rule to obtain @xmath107 , where @xmath108 is the largest eigenvalue .
notice that the @xmath109 is a smooth and monotonic function of @xmath12 , therefore we can arrive at the rigorous conclusion numerically by going to the numeric limit of @xmath12 assisted with the verification by @xmath108 , see fig .
[ lam_max ] bipartition .
the left panel is the sign distribution of rnyi entropy derivative which characterizes the region with non local convertibility .
the right panel is the largest eigenvalue , whose slope has the opposite sign with the rnyi entropy derivative in the large @xmath12 limit . comparing the two panels we can see that our conclusion is still correct even we go to infinite @xmath12 . ]
99 h. l. stormer , d. c. tsui , and a. c. gossard , rev .
mod . phys . * 71 * , s298 ( 1999 ) .
m. z. hasan and c. l. kane , rev .
phys . * 82 * , 3045 ( 2010 ) .
s. yan , d. a. huse , and s. r. white , science , * 332 * , 1173 ( 2011 ) .
wen , _ quantum field theory of many - body systems
_ ( oxford university press , 2004 ) .
m. h. freedman , a. kitaev , and z. wang , commun .
phys . * 227 * , 587 ( 2002 ) ; c. nayak , s. h. simon , ady stern , m. freedman , s. das sarma , rev . mod . phys . *
80 * , 1083 ( 2008 ) .
a. y. kitaev , ann .
( n. y. ) * 303 * , 2 ( 2003 ) .
l. amico , r. fazio , a. osterloh , and v. vedral , rev .
phys . * 80 * , 517 ( 2008 ) .
j. eisert , m. cramer , and m. b. plenio , rev .
phys . * 82 * , 277 ( 2010 ) .
gu and x.g .
wen , phys .
b * 80 * , 155131 ( 2009 ) .
a. hamma , r. ionicioiu , and p. zanardi , phys .
a * 337 * , 22 ( 2005 ) ; a. hamma , r. ionicioiu , and p. zanardi , phys .
a * 71 * , 022315 ( 2005 ) ; a. kitaev and j. preskill , phys .
lett . * 96 * , 110404 ( 2006 ) ; m. levin and x .- g .
wen , phys .
96 * , 110405 ( 2006 ) .
s. t. flammia , a. hamma , t. l. hughes , and x .-
wen , phys .
lett . * 103 * , 261601 ( 2009 ) .
s. v. isakov , m. b. hastings , and r. g. melko , nature phys . * 7 * , 772 ( 2011 ) .
a. hamma , l. cincio , s. santra , p. zanardi , and l. amico , phys . rev
. lett . * 110 * , 210602 ( 2013 ) .
s. turgut , j. phys . a : math
. theor . * 40 * , 12185 ( 2007 ) ; m. klimesh , arxiv:0709.3680 ( 2007 ) .
j. cui , m. gu , l .- c .
kwek , m. f. santos , h. fan , v. vedral , nat commun * 3 * , 812 ( 2012 ) .
we remark that both the paramagnet and the ordered phases turn out locally convertible if symmetry breaking is taken into account f. franchini j. cui , l. amico , h. fan , m. gu , a. hamma , v.e .
korepin , l .- c .
kwek , v. vedral , arxiv:1306.6685 .
x. chen , z - c .
gu , x - g .
wen , phys .
b * 82 * , 155138 ( 2010 ) ; _ ibid _ phys . rev .
b * 83 * , 035107 ( 2011 ) ; _ ibid _ arxiv:1103.3323 ( 2011 ) .
w. son , l. amico , r. fazio , a. hamma , s. pascazio and v. vedral , europhys .
, * 95 * , 50001 ( 2011 ) .
p. smacchia , l. amico , p. facchi , r. fazio , g. florio , s. pascazio , and v. vedral phys . rev . a * 84 * , 022304 ( 2011 ) .
f. d. m. haldane , phys .
lett . * 61 * , 1029 - 1032 ( 1988 ) .
mikeska and a.k .
kolezhuk , _ quantum magnetism _ , u. schollwck , j. richter , d.j.j .
farnell , r.f .
bishop eds ( springer , berlin , 2004 ) . c. becker , p. soltan - panahi , j. kronjger , s. drscher , k. bongs , and k. sengstock , new j. phys .
* 12 * , 065025 ( 2010 ) .
j. k. pachos and m. b. plenio , * 93 * , 056402 ( 2004 ) . h. j. briegel and r. raussendorf , * 86 * , 910 ( 2001 ) .
j. m. renes , a. miyake , g. k. brennen , and s. d. bartlett , new j. phys . * 15 * , 025020 ( 2013 ) .
f. verstraete , v. murg , and j. i. cirac , adv . phys . * 2 * , 143 ( 2008 ) .
a. c. doherty and s. d. bartlett , * 103 * , 020506 ( 2009 ) .
s. o. skrvseth and s. d. bartlett , * 80 * , 022316 ( 2009 ) .
w. chen , k. hida , and b. c. sanctuary , phys .
b. * 67 * , 104401 ( 2003 ) . c. degli esposti boschi , e. ercolessi , g. morandi , in _ symmetries in science xi _
, 145 - 173 , ( kluwer 2004 ) ; arxiv : cond - mat/0309658
. s. hu , b. normand , x. wang , and l yu , phys .
b * 84 * , 220402(r ) ( 2011 ) .
w. son , l. amico , and v. vedral , quant .
11 * , 1961 ( 2012 ) .
f. pollmann , e. berg , a. m. turner , m. oshikawa , phys .
b * 85 * , 075125 ( 2012 ) .
kou and x - g .
wen , phys .
b * 80 * , 224406 ( 2009 ) .
f. pollmann , e. berg , a. m. turner , m. oshikawa phys .
b * 81 * , 064439 ( 2010 ) .
w. li , a. weichselbaum , and j. von delft , arxiv : 1306.5671 .
l. lepori , g. de chiara , a. sanpera , arxiv:1302.5285 ; g. de chiara , l. lepori , m. lewenstein , a. sanpera , arxiv:1104.1331 .
e. polizzi , f. mila , and e. s. sorensen , phys .
b * 58 * , 2407 ( 1998 ) .
a. j. daley , h. picher , j. schachenmayer , and p. zoller , phys .
* 109 * , 020505 ( 2012 ) ; d. abanin , and e. demler , phys .
lett . * 109 * , 020504 ( 2012 ) .
we note that at very large @xmath12 the small eigenvalues produce a log divergence in @xmath72 .
the correct behavior can be achieved by analyzing @xmath108 ( see appendix [ large_alpha ] ) .
|
we consider @xmath0 hamiltonian systems whose ground states display symmetry protected topological order .
we show that ground states within the topological phase can not be connected with each other through locc between a bipartition of the system .
our claim is demonstrated by analyzing the entanglement spectrum and rnyi entropies of different physical systems providing examples for symmetry protected topological phases .
specifically , we consider spin-@xmath1 cluster - ising model and a class of spin@xmath2 models undergoing quantum phase transitions to the haldane phase .
our results provide a probe for symmetry - protected topological order . since the picture holds even at the system s _ local scale _ , our analysis can serve as as local experimental test for topological order .
|
sbs 0335 - 052 ( izotov et al . 1990 ) is a blue compact dwarf ( bcd ) galaxy that , at an abundance of @xmath5 ( melnick et al .
1992 , izotov et al .
1997 ) , is the second lowest metallicity galaxy known after i zw 18 .
sbs 0335 - 052 hosts an exceptionally powerful episode of star formation , and thuan et al .
( 1997 ) proposed , on the basis of hst observations , that the present burst is possibly the first star - forming episode in the history of the galaxy .
the same authors deduce that the star formation occurs in six super - star clusters ( sscs ) not older than 25 myr and located within a region smaller than 2 , or 520 pc at a distance of 54.3 mpc that we will assume in this paper ( @xmath6km / s , see izotov et al .
though the presence of dust mixed with the star forming clusters is evident from the hst images , the iso observation by thuan et al .
( 1999 ) of a very strong continuum at @xmath7 m was a surprise .
this was successfully modeled with a modified black body and a screen extinction of @xmath8 , and interpreted as due to thermal emission by silicate dust ( m@xmath9 ) warmed by embedded clusters of star formation completely invisible in the optical . because of the lack of chemical enrichment , low - metallicity bcds such as sbs 0335 - 052 are also hypothetical candidates for young or primeval galaxies ( pgs ) undergoing their first burst of star formation .
while the majority of bcds show an underlying extended low - surface - brightness component with colors indicative of an intermediate - age or old stellar population ( loose & thuan 1985 ; kunth et al .
1988 ; papaderos et al . 1996 ; telles & terlevich 1997 ) , there is still debate whether such a component has been detected in the two most metal - deficient bcds presently known , izw18 ( z@xmath10/50 , searle & sargent 1972 ) and sbs 0335052 .
high - resolution hst optical images of izw18 ( hunter & thronson 1995 ) show colors of the underlying diffuse component consistent with a sea of unresolved b or early a stars , with no evidence for stars older than a few tens of myr , although aloisi et al .
( 1999 ) and stlin ( 2000 ) have claimed to detect a 1 - 4 gyr stellar population in that bcd .
similar observations of sbs0335 - 052 ( thuan et al . 1997 ) show colors much bluer than those expected for such an intermediate - age stellar population .
the optical evidence is only suggestive , however , since a large fraction of the extended emission in sbs0335 - 052 is of gaseous origin ( thuan et al .
1997 ) .
near - infrared ( nir ) colors can help resolve the age question , since they are extremely effective indicators of stellar population age , _ given the metallicity_. while nir colors of _ evolved _ stellar populations vary little ( griersmith et al . 1982 ; frogel 1985 ; giovanardi & hunt 1988 , 1996 ) , recent evolutionary synthesis models ( leitherer et al .
sb99 ; krger et al .
1995 ) show that the nir colors of _ young _ stellar populations differ significantly from those of older ones .
both models find that longward of 1@xmath11 m , the nebular continuum dominates the emission during the early phases ( a few myr ) of the starburst .
therefore nir colors provide a unique and effective diagnostic , and when combined with optical colors , are good indicators of stellar population age .
nir spectroscopy has a three - fold importance for our analysis .
first , it enables the estimate of the nebular contribution to the emission in order to decontaminate the broadband colors .
second , it is an almost extinction - free probe of the starburst parameters in what are typically dusty environments .
finally , it offers unique constraints on the star - formation history of the galaxy .
in particular the equivalent width of @xmath12 is possibly the best age indicator for young starbursts since it measures the ratio between young blue massive stars and the evolved red stellar population .
the nir co absorption features are the strongest signature of cold evolved stars , as are the he ricombination lines for hot massive stars , and [ feii ] for supernova remnants .
finally the infrared is the only spectral region where the molecular gas can be detected in its warm phase through @xmath13 emission lines .
it is therefore of crucial importance to further investigate sbs 0335 - 052 in the near - infrared to study the stellar populations in this extremely low - metallicity system , as well as investigate the putative hidden star formation suggested by the large extinction derived from the mid - infrared spectral energy distribution . with this aim
we have obtained nir spectra in the @xmath1 and @xmath2 bands , and broadband images in @xmath0 , @xmath1 , @xmath2 , and @xmath3 .
we present here the results of our new observations .
the sbs0335 - 052 system consists of two widely separated ( 22kpc in the east - west direction ) star - forming components within a large 64@xmath1421kpc hi cloud ( pustilnik et al .
we denote the eastern component by sbs0335 - 052 , and the western one ( lipovetsky et al .
1999 ) by sbs0335 - 052w .
the present paper is organized as follows . in sect.2
we present our data along with a description of the observations and data reduction .
sect.3 is a description of the nir spectrum and of the properties of the galaxy as can be derived from the analysis of the spectrum .
the same is done in sect.4 for the near - infrared images .
the characteristics of the galaxy derived from the observations are combined with data from the literature , and then used to constrain a starburst model in sect.5 . in sect.6
we summarize the results of our work .
our data were acquired at eso , la silla , with the ntt and sofi , and at ukirt , mauna kea , with ircam3 .
the former camera@xmath15spectrometer is based on a 1024@xmath141024 hawaii hgcdte array , and the latter camera on a 256@xmath14256 nicmos3 .
low - resolution ( r=600 ) spectra of sbs 0335 - 052 covering the @xmath1 and @xmath2 infrared bands were obtained with sofi at the ntt in two different occasions : first in august 1998 with 45 m of integration , and later in december 1998 with similar conditions and the same amount of integration time . in both cases we used a slit 1 wide ( 5 arcmin length ) oriented along the `` major axis '' of the galaxy @xmath16 .
since the source is much smaller than the length of the slit , the nod - on - slit technique was used to acquire the sky spectrum .
the data reduction followed the standard process of flat - fielding and sky - subtraction .
a lamp spectrum was used to correct the 2d spectra for distortions along the slit , and to have an accurate wavelength calibration .
1d spectra were extracted with an aperture of 15 centered on the brightest peak .
particular attention was given to achieving a good estimate of the continuum slope .
the 1d spectra were divided by the spectrum of a solar - type star observed under the same conditions to correct for the telluric absorption features .
then the solar spectrum was used to reestablish the original slope as described by maiolino et al .
the two spectra have been averaged with equal weights .
the spectrum was flux calibrated extracting the photometry from the corresponding apertures in our h and k images .
this also lead to the adjustment of the continuum that was otherwise too blue .
comparison of the spectrum with the photometric values was always done by convolving the spectrum with the filter bandpass measured in the laboratory .
the 1d combined spectrum flux and wavelength calibrated is shown in fig .
[ lowres ] . a medium - resolution
( r=1400 ) spectrum in the @xmath3 band was observed with sofi in october 1999 under exceptional seeing conditions ( 05 average of the night ) ; the total integration time was in this case 1h .
we used the same slit and same pa as in the observations described above , and the same procedure was used for the data reduction . at this spatial resolution ,
two peaks of emission separated by about 15 were clearly resolved in the 2d spectrum .
two 1d spectra centered on the two peaks were extracted with apertures of 08 ( south - east ) and 06 ( north - west ) .
we will refer to these spectra as a and b respectively in the rest of the paper . since the observed band is well transmitted by the atmosphere no correction or readjustment of the slope was applied .
the spectra were flux calibrated extracting the photometry from the corresponding apertures in the 0073@xmath17 image ( see next section ) . the flux- and wavelength - calibrated spectra are shown in fig.[higres ] .
all the spectra have been smoothed to lower resolution to reduce the noise . including the errors on the photometry , centering and size of the apertures the flux calibration of the spectra
can be estimated to be accurate within 10% .
we acquired @xmath0 ( 1.2@xmath11 m ) , @xmath1 ( 1.6@xmath11 m ) , and @xmath2 ( 2.2 @xmath11 m ) images of the nominal central ( e ) and western ( w ) component of sbs0335 - 052 with the 3.8-m united kingdom infrared telescope ( ukirt ) equipped with ircam3 .
the plate scale is 028 , with a total field - of - view of 72 ; the e and w components are separated by roughly 84 , too large to be accommodated by the ircam3 field - of - view , so we performed separate sets of observations for the two components .
the source was placed in several different positions on the array , taking care not to compromise the flat field by having overlap in the source positions . at the beginning of each observing sequence
, dark exposures were acquired with the same parameters as the subsequent science frames .
total on - source integration times were 55min and 48min in @xmath2 for the e and w components , respectively ; 30 and 34min in @xmath1 ; and 19 and 27min in @xmath0 .
individual frames were dark - subtracted and flat - fielded with a subset of the remaining frames in the sequence , after editing them for stars ( to avoid `` holes '' in the reduced frames ) .
the reduced frames were then registered and averaged ; observations obtained on different nights were combined with a similar algorithm , but with the flux scale given by the relative photometric calibration , and weighted by the total integration time .
photometric calibration was performed by observing standard stars , every few hours , from the ukirt faint standard list . as for the source observations ,
each standard was measured in several different positions on the array .
we corrected for atmospheric extinction by fitting simultaneously , for each band , an extinction coefficient common to all nights and a `` zero point '' that varied from night to night .
the resulting extinction coefficients are 0.049 , 0 , and 0.046 mag / airmass in @xmath0 , @xmath1 , and @xmath2 , respectively .
formal photometric accuracy , as judged by the nightly dispersion of the standard stars is 0.025mag , or better , in all three bands . we have corrected all photometry for galactic extinction ( @xmath18 mag ) using the extinction law of cardelli et al . (
1989 ) .
the final ukirt deep @xmath2-band image is shown in fig.[kimg_uk ] , with the contours of the hst @xmath19-band image from thuan et al .
( 1997 ) superimposed . as measured from the combined frames , seeing was 11 in @xmath0 , 081 in @xmath1 , and 075 in @xmath2 .
the combined images have a 1@xmath20 sensitivity ( per pixel ) of 22.4 mag / arcsec@xmath21 in @xmath0 , 21.4 mag / arcsec@xmath21 in @xmath1 , and 21.4 mag / arcsec@xmath21 in @xmath2 .
three images were also acquired in october 1999 with sofi at the ntt . a @xmath0 image with a scale of 0.146 arcsec / pix and a total integration of 300 sec , a @xmath3 image with the same scale and integration time and a @xmath3 image with a scale of 0073 /pixel and integration time of 400 sec .
the eso @xmath22 filter has a central wavelength @xmath23 of 2.162@xmath11 m , and a fwhm @xmath24 of 0.275@xmath11 m ( c.f .
, ukirt @xmath2 has @xmath25 m , and fwhm@xmath26 m ) .
the jitter - on - source technique was always used .
all the images were reduced following the standard procedures of flat - field , sky - subtraction and realignment .
photometric calibration was achieved through large - aperture photometry of the ukirt images .
the seeing measured on the combined images is 049 ( @xmath0 ) , 041 ( @xmath3 ) and 033 ( 0073 @xmath3 ) ! hereafter , these sofi images will be called high - resolution ( @xmath3 only , 0073/pix ) , medium - resolution ( @xmath0 and @xmath3 , 0.146 arcsec / pix ) , and the ukirt images low - resolution ( @xmath0 , @xmath1 , @xmath2 0.28 arcsec / pix ) .
fig.[kimg_hr ] shows the high - resolution @xmath3 sofi image , together with the ssc nomenclature of thuan et al .
the figure shows , at a significantly higher resolution and better seeing than fig .
[ kimg_uk ] , the same central structure as seen in the visible and in the deeper ukirt image .
the lack of extended structure in the high - resolution image is due to its lower sensitivity .
.recombination lines in sbs 0335 - 052 from the low - resolution spectrum , @xmath27 in @xmath11 m and fluxes measured in 10@xmath28erg / s / cm@xmath21 [ cols="^,^,^",options="header " , ] the @xmath30 and @xmath31 colors of the region extended to the northwest beyond sscs 4 + 5 in the eastern source ( see fig .
[ vmijvk ] ) seem to imply a young age similar to that estimated for the sscs , but with an excess in @xmath2 .
although such an excess could be imputed to our relatively low sensitivity in @xmath2 to the blue colors observed in sbs 0335 - 052 , we find a @xmath321.3 , even in relatively high surface brightness regions where our deep ukirt @xmath2 image is sufficiently sensitive .
the @xmath4 colors of the extended region ( nw ) , shown in fig .
[ jmhhmk ] are very red , and @xmath33 is blue ; the red @xmath4 is an indication that there may be hot dust also in the extended region .
we can estimate what fraction of an evolved stellar population could be present in sbs 0335 - 052 , given the observed nir colors .
the nir is the best spectral region to perform this estimate , since it is more sensitive than the optical to evolved stars .
first , we find absolutely no color evidence for an evolved stellar population ( age greater than 10myr ) in any of the regions examined , including the diffuse low - surface - brightness component .
hence , the fraction of hidden " evolved stars that could be present is essentially dictated by the uncertainties on our nir colors . in the sscs and the global colors , these are nominally around 0.07 mag , since each magnitude has an uncertainty of roughly 0.05 mag or better , including calibration errors .
therefore , by adding the colors of an evolved ( @xmath34 1gyr ) to the 3 myr sb99 color , we can estimate what fraction of an evolved stellar population would be responsible for a color change in @xmath33 and @xmath4 that is equivalent to our measurement uncertainty .
it turns out that @xmath33 changes by around 0.07 mag for a 1618% contribution of an evolved population to the 3 myr sb99 color ; the concomitant @xmath4 color change is smaller ( 0.03 mag ) . because we are observing the _ total _ colors , the fraction becomes @xmath35 or 15% ; that is to say we can not exclude a 15% or smaller contribution to the nir colors ( of the sscs or globally ) that derives from an evolved stellar population .
this upper limit is unchanged if we add evolved nir colors to a 3 myr solar - abundance stellar population ( sb99 ) , rather than to a metal - poor one .
we are also interested in the fraction of evolved stars that could be present in the diffuse region that extends to the northwest . as shown in fig .
[ jmhhmk ] , the uncertainties there are larger than those in the sscs .
if we take an error of 0.25 mag in @xmath33 , we derive a ratio of 0.9 between evolved and young stellar population colors .
this implies that 47% of the extended emission could derive from an evolved population , given the @xmath33 uncertainties in the extended region .
however , the @xmath33 colors are not the whole story .
we find an @xmath4 color of @xmath36 in the outer regions , whether we use the integrated profiles or the photometry of the color images to measure it .
such a red color is not reproducible by adding an evolved stellar population to the 3 myr nir colors ; indeed , the @xmath37 color of an evolved stellar population is _ bluer _ than that of stars@xmath15gas at 3 myr .
it is difficult to imagine a scenario involving a significant evolved stellar population in which the colors are conveniently uncertain in @xmath33 , but provide such a red @xmath4 . to investigate the hot dust that we have tentatively invoked to explain the red @xmath4 colors , we have combined the mid - infrared iso spectrum of sbs 0335 - 052 ( thuan et al . 1999 ) with the @xmath2-band emission after accounting for the observed gas fraction .
after subtracting the gas contribution , we have assumed that half of the remaining @xmath2 band flux is due to hot dust .
the combined sed has been fitted with two modified blackbodies at different temperatures and a foreground dust screen .
dust temperatures and amplitudes , dust emissivity index , and visual extinction @xmath38 were left as free parameters , and we have adopted the extinction curve given in draine ( 1989 ) . the resulting fit is shown in fig . [ dust ] as a dotted line .
the warm and hot components have temperatures of 226 k and 668 k , the cooler temperature being similar to 260 k found by thuan et al .
the fitted emissivity is around @xmath39 ( instead of the @xmath40 used by thuan et al .
1999 ) , and the amplitude of the hotter component is 9% that of the cooler one at 10@xmath11 m . at shorter wavelengths , the ratio of the hotter to the cooler component increases , being 31% at 7.7@xmath11 m and 68% at 6.7@xmath11 m .
the fitted extinction @xmath38 is around 12 mag , but more uncertain ( @xmath411 mag ) than the temperatures ( uncertainties @xmath4110 k ) . in any case , without a significant extinction , the sed is very poorly fit as it is too narrow around its peak at 10 @xmath11 m . with two dust components ,
we find a slightly lower temperature ( for the cooler dust ) and lower extinction than the @xmath38 of 21 mag found by thuan et al .
nevertheless , both models show that the mid - infrared extinction is high , in variance with what we infer from the optical / nir recombination lines .
the most likely explanation is the clumpy nature of the interstellar medium , since we do not detect peaks of emission at 2@xmath11 m corresponding to hidden sscs .
there may be a relatively high quantity of dust globally , but not locally in the vicinity of the sscs we detect .
the fit of the mid - infrared spectrum around 10 @xmath11 m is poor in detail because the draine extinction curve overpredicts silicate absorption there ( lutz et al .
1996 ) , and indeed fits of the sed in sbs 0335 - 052 with the galactic center curve give better results ( thuan et al .
however , it is clear that a hot - dust component can fit the shorter - wavelength points very well , and consistently with the @xmath2 emission after gas subtraction .
indeed , such dust seems to be necessary to explain the red @xmath4 colors of sbs 0335 - 052 .
we can summarize the results of our work as follows : 1 .
the near - infrared spectrum of sbs 0335 - 052 is representative of an extreme and young starburst galaxy .
the @xmath12 equivalent width indicates a very young age , and is one of the largest ever observed in an extragalactic object .
the he i recombination lines support the presence of young massive stars .
recombination and @xmath13 lines are strong , [ feii ] and stellar absorption bands are absent , which also implies a strong uv field and a young stellar population ; there is no spectroscopic evidence for stars older than 56 myr .
modeling the @xmath12 and @xmath2 continuum leads to a burst age@xmath425 myr and a star formation rate that can be as high as 10 @xmath43/yr .
the spatial profile of the @xmath2 continuum is shifted by about 60 pc northwest of the emission line peak and it is much closer to it than the optical peak at a distance of 200 pc .
this is explained with the dominant nebular contribution to @xmath2 .
all optical and nir emission lines coincide and have consistent spatial profiles .
the near - infrared photometry of sbs 0335 - 052 is highly contaminated by nebular emission .
the optical and nir colors are unusual , and after correction for the gas contribution can be only understood as due to a stellar population not older than 4 myr , together with hot dust at 670 k. 4 . on the basis of the nir colors , the stellar population in sbs 0335 - 052w appears to be older than that in sbs 0335 - 052 ; the colors are redder in @xmath33 and bluer in @xmath4 and correspond to an approximate age of 1020 myr according to sb99 . however , optical recombination line equivalent widths suggest that the stellar population can not be older than 10 myr , and our models suggest that the age is more likely around 57 myr . 5 .
the gas - corrected spectrum from 2 to 17@xmath11 m is well fit by two thermal components ( hot dust at 670 k , and cooler dust at 225 k ) , together with @xmath44mag of visual extinction in a foreground screen .
we do not find evidence at 2@xmath11 m for optically hidden star formation .
judging from the uncertainties on the nir photometry , the possible contribution from an evolved stellar population in sbs 0335 - 052 can not exceed 15% . sbs 0335 - 052 appears to be an unusual object , as there is no evidence , even in the near - infrared , for an evolved stellar population deriving from an earlier burst .
this is in contrast to results for the lowest - metallicity galaxy known , i zw 18 , in which aloisi et al .
( 1999 ) and stlin ( 2000 ) claimed to detect a 14 gyr underlying stellar population .
if the star - formation rates derived from our modeling are representative , then stars are forming in sbs 0335 - 052 extremely rapidly , but how such a rate is connected to the formation of sscs is an open question .
it could be that under certain conditions , star formation , galaxy gas consumption , and evolution proceed very quickly , while in others , evolution is slower .
future work will be aimed at better quantifying what these conditions might be , and how they relate to metallicity .
we thank gian - paolo tozzi who generously shared with us part of his observing time at the ntt during the run of august 1998 .
txt thanks the partial financial support of nsf grant ast-9616863 and the hospitality of the observatoire de paris - meudon and the institut dastrophysique de paris during his sabbatical year .
we are also grateful to george rieke who made available to us the results of his starburst model , to daniel schaerer who gave us his low - metallicity stellar tracks and to leonardo testi for useful comments on the manuscript .
|
we present near - infrared ( nir ) spectra , deep @xmath0 , @xmath1 and @xmath2 low - resolution images , and high - resolution images in @xmath0 and @xmath3 of the low - metallicity blue dwarf galaxy sbs 0335 - 052 .
these new nir data , together with optical and mid - infrared data from the literature , are used to constrain the properties and the star formation history of the galaxy .
the nir spectral characteristics and broadband colors of this galaxy are those of an extremely young starburst .
we find that the nir emission is dominated by star formation occuring in clusters younger than 5 myr ; no optically hidden star formation is revealed in our data .
we also find that a warm - dust component is necessary to explain the @xmath4 color , and is consistent with the mid - infrared spectral energy distribution .
finally , we quantify the possible contribution from an evolved stellar population .
|
electron scattering off of spin fluctuations in magnetic metals results in an `` anomalous '' contribution to electric resistivity.@xcite the analysis of this spin - disorder resistivity ( sdr ) is of interest because it can provide material - specific information on the character of spin fluctuations which is not easily accessible by other means .
scattering on spin disorder is also an important factor degrading the performance of magnetoresistive nanostructures in spintronic devices .
the magnitude of the spin - disorder contribution to resistivity is comparable to the phonon contribution near and above the curie temperature @xmath2.@xcite ( magnetic scattering amplitudes have no small parameter unless the local moments are small . ) it is usually assumed that sdr is constant well above @xmath2 . in this region matthiessen s rule is valid , and the phonon contribution can be fitted to the bloch - grneisen formula . the excess resistivity in the whole temperature range may be attributed to spin disorder,@xcite although one may expect deviations from matthiessen s rule at low temperatures where transport is carried by weakly interacting spin channels.@xcite in addition , it was argued that in some cases ( such as ni ) spin disorder may change the character of states on the fermi level and thereby appreciably change the phonon contribution itself.@xcite many authors have studied sdr theoretically using the @xmath3-@xmath4 model hamiltonian.@xcite in this approach the @xmath5 shells in transition metals ( or @xmath6 shells in rare earth metallic magnets ) are assumed to be localized at atomic sites and partially filled , forming magnetic moments @xmath7 that are coupled to the current - carrying conduction electrons by exchange interaction @xmath8 , where @xmath9 is the local @xmath3-@xmath4 exchange coupling constant and @xmath10 the spin - density operator of the conduction electrons at site @xmath11 .
thermal fluctuations of the @xmath4-electron spins generate an inhomogeneous exchange potential ; in the born approximation the sdr is then determined by the conduction electron band structure and the spin - spin correlation functions of @xmath4-electron spins.@xcite if the scattering is approximated as being elastic , only equal - time spin correlators have to be considered .
further , if the mean - field approximation ( mfa ) is used for @xmath5 spin statistics , the sdr behaves as @xmath12 $ ] , where @xmath13 is the magnetization at temperature @xmath14 and @xmath15 .
@xcite note that above @xmath2 sdr is constant and equal to @xmath16 .
the shape of the fermi surface of conduction electrons is immaterial to this prediction as long as the scattering is elastic .
@xcite the effects of magnetic short - range order ( msro ) on sdr have also been investigated within the @xmath3-@xmath4 model .
@xcite this problem has attracted considerable attention in connection with a `` bump '' in the resistivity that is observed near @xmath2 in some magnetic metals ( although it is usually quite small).@xcite the analysis of critical msro effects showed that a cusp may appear near @xmath2 due to long - wave critical fluctuations,@xcite although it should usually be strongly suppressed by finite mean - free path and cancelations due to fermi surface integration.@xcite it was also found that the effect of msro and even its sign are sensitive to such details of the model as the conduction band occupation and the form of the scattering ( pseudo)potential.@xcite although the @xmath3-@xmath4 model provided physical insight into the mechanism of sdr , it suffers from serious limitations .
first , the distinction between localized and conduction electrons is not justified in transition metals where @xmath5 electrons are itinerant and form the fermi surface .
even if the current is dominated by light @xmath3-like bands that can be distinguished from heavy @xmath4-like bands , the relaxation rate is dominated by interband ( @xmath3-@xmath4 ) scattering .
@xcite second , at elevated temperatures the scattering potential generated by spin disorder is of the order of the exchange splitting , which is not small compared to the bandwidth .
this invalidates the born approximation which is usually made in model calculations .
third , the @xmath3-@xmath4 model does not properly take into account the change of electronic structure due to disorder .
the first - principles approach to sdr is free from all these limitations and can be used for quantitative calculations of sdr .
this is of particular interest for the theory of itinerant magnets , because , as mentioned above , sdr depends on spin - spin correlation functions . different theories of itinerant magnetism make conflicting predictions for such properties as the degree of msro , the mean - squared magnetic moment , and their temperature dependence;@xcite these quantities are quite hard to measure directly . by calculating sdr for a particular model of spin fluctuations and comparing the results with experiment , one can attempt to validate or rule out different spin - fluctuation models .
earlier we have calculated the temperature dependence of sdr in fe and ni using supercell calculations within the tight - binding linear muffin - tin orbital ( tb - lmto ) method using the mean - field distribution for spin orientation statistics .
@xcite good agreement with experiment was obtained for fe , but for paramagnetic ni the sdr was found to be significantly overestimated . in this paper
we analyze the temperature dependence of sdr for fe and ni in greater detail and consider the effects of magnetic ordering , msro , and local moment reduction .
we also study the importance of the basis set size and self - consistency of the atomic potentials .
our approach is based on noncollinear density functional theory ( dft ) .
all the valence electrons are treated on the same footing , while the scattering potentials are determined by the self - consistent electron charge and spin densities .
we use the tb - lmto method @xcite which represents the electronic density of the crystal as a superposition of overlapping atomic spheres ; the electronic density inside each sphere is spherically symmetric .
this method is known to work very well in close - packed materials , and it allows us to introduce spin disorder in various ways . in this work we used the rigid spin approximation which assumes that the spin density in each atomic sphere remains collinear , while the spin densities of different atomic spheres become noncollinear at finite temperatures . in the simplest model the electron charge and spin densities in all atomic spheres
are taken from the ground state and frozen , while the directions of the spin moments in different spheres are randomized with the angular distribution function taken from mfa at the given temperature .
this model is expected to work reasonably well for fe which has a fairly stable local moment.@xcite in section [ mfa ] we show that this is indeed the case ; however , for ni the paramagnetic sdr calculated in this way is about twice too large . in order to explain this discrepancy , the dependence of sdr on the degree of msro and on the magnitude of the local moment is studied in sections [ msro ] and [ reduc ] .
we use the supercell approach and calculate the areal conductance of a layer of spin - disordered metal fm(d ) sandwiched between fully ordered semi - infinite leads fm(o ) made of the same metal ( see fig . [ fig1 ] ) .
the resistivity is then proportional to the slope of the inverse conductance as a function of the disordered layer thickness , once the ohmic limit is reached . for the given thickness of the fm(d ) layer ,
the conductance of the system was averaged over several disorder configurations ( typically 15 ) .
the planar system is represented by a laterally periodic prism with an axis along the [ 001 ] crystallographic direction , and care is taken to make sure that the conductance scales as the cross - section of the prism . to calculate the conductance we use the principal - layer green s function technique @xcite and the landauer - bttiker formalism @xcite in the implementation allowing for noncollinearity in the active region.@xcite this technique was employed before to study the effects of substitutional disorder on transport in magnetic multilayers ; @xcite it is similar to the supercell kubo - greenwood method used to calculate the residual resistivity of binary alloys.@xcite if the atomic potentials in the supercell are not converged to self - consistency with the given spin disorder configuration , care needs to be taken to ensure local charge neutrality .
indeed , fm(d ) and fm(o ) materials have different fermi levels that must normally be matched by the contact voltage . in order to enforce charge neutrality in the fm(d ) region ,
a constant potential shift was introduced in this region so that the charge in the central part of fm(d ) averaged over disorder realizations was zero .
this potential shift plays the role of the contact voltage .
note that no matter how the fm(o)/fm(d ) interfaces are treated ( self - consistently or not ) , they add contact resistances to the circuit
. however , since the _ resistivity _ of the fm(d ) material is extracted from the thickness dependence of the resistance in the ohmic limit , the simplified treatment of interfaces has no effect on the results .
in this section we analyze the temperature dependence of sdr for iron and nickel using mfa for thermal spin disorder ; the spin - spin correlator is purely local in this approximation .
first we consider the paramagnetic state where the angular distribution function is isotropic , and the resulting sdr is temperature - independent .
we need to make a physically reasonable choice of atomic potentials for the conductance calculations .
it is known that the local moments in fe are quite stable;@xcite in particular , the dlm method , which employs the coherent potential approximation for spin - disordered states , shows only a small reduction of the local moment in paramagnetic fe compared to its ground - state value.@xcite as seen below , direct averaging of self - consistent local moments in the paramagnetic states gives a similar result .
therefore , for fe it is reasonable to use frozen atomic potentials taken from the zero - temperature ground state in all calculations .
we have also checked the effect of self - consistency on sdr in fe and found it to be small ( see below ) .
the situation is entirely different for ni , where the local moment depends on the magnetic state ; in particular , it vanishes altogether in the paramagnetic dlm approximation .
@xcite since longitudinal spin fluctuations ( that are absent in our approach ) can at least partially restore the local moments , @xcite it is not _ a priori _ clear how the atomic potentials should be modified for ni . in this section
we use frozen atomic potentials ; the necessary corrections are discussed later . fig .
[ fig2 ] shows the inverse areal conductance for paramagnetic fe and ni as a function of the disordered fm(d ) region thickness .
here we used the frozen ground - state atomic potentials and the lmto basis including @xmath3 , @xmath17 , and @xmath4 orbitals ( @xmath18 ) .
the supercell cross - sections contained @xmath19 ( for fe ) and @xmath20 ( for ni ) cubic unit cells with edges oriented along the [ 100 ] directions .
almost perfect ohmic behavior is apparent for both fe and ni , which establishes the validity of our approach . of the fm(o)/fm(d)/fm(o ) systems as a function of the fm(d ) layer thickness for bcc fe
( black filled symbols ) and fcc ni ( red or gray empty symbols ) obtained with @xmath18 .
circles and squares correspond , respectively , to the paramagnetic state and to the lowest temperature for which the calculations were made .
@xmath19 and @xmath20 supercells were used for fe and ni , respectively , with edges along the [ 100 ] directions .
straight lines show the linear fitting ; error bars are smaller than the size of the symbols.,scaledwidth=45.0% ] table [ tab1 ] lists the values of sdr found for paramagnetic fe and ni using different supercell cross - sections , lmto bases truncated at @xmath18 and @xmath21 ( the latter includes @xmath6 orbitals ) , as well as the value found using self - consistent ( rather than frozen ) atomic potentials for fe .
it is seen that the results are well converged with respect to the supercell cross - section , and even @xmath22 supercells provide sufficient accuracy .
this is reasonable because the mean - free path in the paramagnetic state is small .
.spin - disorder resistivity in @xmath23 cm for paramagnetic bcc fe and fcc ni .
the calculated values are given for basis sets with @xmath18 and 3 , as well as for different lateral cell sizes with edges along the [ 100 ] directions .
sc denotes calculations with self - consistent potentials .
standard deviations of sdr due to limited disorder sampling are included .
[ cols="<,^,^,^,^,^",options="header " , ]
numerous previous studies @xcite based on the @xmath3-@xmath4 model concluded that sdr in the paramagnetic state is essentially proportional to @xmath24 where @xmath25 is the spin of the partially filled @xmath5 shell .
this dependence is easy to understand based on the fermi golden rule with averaging over the initial states of the @xmath5 spin . in our treatment based on noncollinear dft
, the exchange - correlation field with randomized directions on different sites plays the role of the @xmath3-@xmath4 hamiltonian .
however , contrary to the @xmath3-@xmath4 model , the @xmath5 spin is treated classically , i. e. @xmath26 is just a classical vector and not an operator .
the fermi golden rule in our case would give a paramagnetic sdr proportional to @xmath27 .
thus , if the @xmath28 factor were correct , noncollinear dft calculations would underestimate the paramagnetic sdr by a factor @xmath1 .
this factor is close to 2 for fe and more than 3 for ni . in reality , the calculated sdr agrees well with experiment for fe and is _ _
over__estimated for ni ( if the local moment reduction is not included ) .
we believe that these results provide clear evidence against the @xmath28 factor which appears if the local moments are treated as local atomic spins . instead , the classical description of the local magnetic fluctuations in the spirit of the dlm approach is supported by our results .
we suggest that the itinerancy of the @xmath5 electrons is crucial for this behavior .
qualitatively , one can argue that the low - energy fluctuations in fe or ni on the scale of @xmath29 ( which the resistivity is most sensitive to ) are similar to classical rotations of the local moments rather than quantum fluctuations of localized spins
. it would be interesting to investigate this issue for magnets with a varying degree of localization , including rare - earth systems .
some poorly controlled assumptions are involved in the extraction of @xmath30 from the experimental data.@xcite first , it is assumed that @xmath30 is constant in the wide temperature range above @xmath2 where the total resistivity is linear in @xmath14 .
this assumption implies that the local moments ( or at least their mean - squared average ) are constant in this range .
spin fluctuation theories for itinerant metals show that the local moments may change with temperature above @xmath2 .
@xcite such change will contribute to the slope of @xmath31 above @xmath2 , and hence the separation of @xmath30 from the phonon contribution would be inaccurate .
on the other hand , it has been argued that the phonon contribution to the resistivity may be sensitive to spin disorder , because the latter may change the character of states at the fermi level.@xcite in particular , in ni the filled majority - spin @xmath4 states may be lifted up to the fermi level by spin disorder , thereby facilitating interband @xmath3-@xmath4 scattering by phonons .
this effect may therefore introduce an unusual temperature dependence of the phonon contribution , which makes spin disorder and phonon effects non - additive , even if the scattering rates themselves obey matthiessen s rule . since we have not studied this effect here
, our comparison of sdr with experiment for ni is incomplete .
however , the phonon contribution can be expected to follow the bloch - grneisen form above @xmath2 with the electron - phonon scattering renormalized by spin disorder ; therefore , the influence of spin disorder on the phonon contribution should not invalidate the procedure used for subtracting this contribution above @xmath2 . in conclusion
, we have calculated the spin - disorder resistivity of fe and ni in the whole temperature range up to @xmath2 using both the mean - field approximation and the nearest - neighbor heisenberg model to represent the canonical ensemble of classical spin configurations .
we found that sdr is insensitive to the magnetic short - range order ( msro ) in fe and ni .
the sdr in fe depends linearly on @xmath32 which implies that the main effect of spin disorder is to introduce scattering , which is proportional to the variance of the random potential . for ni
the calculated temperature dependence is more complicated ; at elevated temperatures close to @xmath2 the sdr grows faster than expected .
this faster increase of sdr may be explained by the reduction of the exchange splitting which lifts the heavy bands up to the fermi level , thereby increasing the scattering rate .
the results for fe are in very good agreement with experiment if the atomic potentials are taken from zero temperature and frozen , but for ni the sdr calculated in this way is strongly overestimated .
this disagreement is attributed to the reduction of the local magnetic moment in ni .
comparison with experimental sdr leads to a value of @xmath0 above @xmath2 , which may be compared with experiment .
we are grateful to v. p. antropov and e. y. tsymbal for useful discussions .
this work was supported by the nebraska research initiative , nsf epscor , and nsf mrsec .
k. d. b. is a cottrell scholar of research corporation .
the work was completed utilizing the research computing facility of the university of nebraska - lincoln .
a portion of this research at oak ridge national laboratory s center for nanophase materials sciences was sponsored by the scientific user facilities division , office of basic energy sciences , u.s .
department of energy .
9 b. r. coles , adv .
* 7 * , 40 ( 1958 ) .
n. f. mott , adv .
* 13 * , 325 ( 1964 ) .
s. v. vonsovskii , _ magnetism _ ( halsted press , new york 1974 ) .
r. j. weiss and a. s. marotta , j. phys .
chem . solids * 9 * , 302 ( 1959 ) .
a. fert and i. a. campbell , phys .
* 21 * , 1190 ( 1968 ) ; j. phys .
* 6 * , 849 ( 1976 ) .
t. kasuya , progr .
* 16 * , 58 ( 1956 ) .
g. de gennes and j. friedel , j. phys .
solids * 4 * , 71 ( 1958 ) .
i. mannari , progr .
* 26 * , 51 ( 1961 ) .
s. v. vonsovskii and yu .
izyumov , sov .
uspekhi * 5 * , 547 ( 1963 ) .
m. e. fisher and j. s. langer , phys .
20 * , 667 ( 1968 ) .
j. b. gibson , j. phys .
solids * 1 * , 27 ( 1956 ) .
p. l. rossiter , p. wells , j. phys .
c : solid st .
* 4 * , 354 ( 1971 ) .
s. alexander , j. s. helman , and i. balberg , phys .
b * 13 * , 304 ( 1976 ) .
m. kataoka , phys .
b * 63 * , 134435 ( 2001 ) .
k. akabli and h. t. diep , j. appl . phys . * 103 * , 07f307 ( 2008 ) .
d. a. goodings , phys . rev . *
132 * , 542 ( 1963 ) .
t. moriya , _ spin fluctuations in itinerant electron magnetism _
( springer , berlin , 1985 ) .
v. p. antropov , phys .
b * 72 * , 140406 ( 2005 ) . c. s. wang , r. e. prange , and v. korenman , phys . rev .
b * 25 * , 5766 ( 1982 ) .
a. v. ruban , s. khmelevskyi , p. mohn , and b. johansson , phys .
b * 75 * , 054402 ( 2007 )
. a. l. wysocki , j. k. glasbrenner , and k. d. belashchenko , phys .
b * 78 * , 184419 ( 2008 )
. a. l. wysocki , k. d. belashchenko , j. p. velev and m. van schilfgaarde , j. appl .
* 101 * , 09g506 ( 2007 ) .
o. k. andersen , phys .
b * 12 * , 3060 ( 1975 )
. b. l. gyorffy , a. j. pindor , j. staunton , g. m. stocks , and h. winter , j. phys .
phys . * 15 * , 1337 ( 1985 ) .
i. turek , v. drchal , j. kudrnovsk , m. ob , and p. weinberger , _ electronic structure of disordered alloys , surfaces and interfaces _
( kluwer , boston , 1997 ) . j. kudrnovsk , v. drchal , c. blaas , p. weinberger , i. turek , and p. bruno , phys . rev .
b * 62 * , 15084 ( 2000 ) .
s. datta , _ electronic transport in mezoscopic systems _
( cambridge university press , cambridge , 1995 ) . m. van schilfgaarde ( unpublished )
. v. drchal , j. kudrnovsk , p. bruno , p. h. dederichs , i. turek , and p. weinberger , phys .
b * 65 * , 214414 ( 2002 ) .
r. h. brown , p. b. allen , d. m. nicholson , and w. h. butler , phys .
rev . lett . * 62 * , 661 ( 1989 ) .
i. turek , j. kudrnovsk , v. drchal , l. szunyogh , and p. weinberger , phys .
b * 65 * , 125101 ( 2002 ) .
j. banhart , h. ebert , p. weinberger , and j. voitlnder , phys .
b * 50 * , 2104 ( 1994 ) .
buruzs , l. szunyogh , and p. weinberger , phil . mag . * 88 * , 2615 ( 2008 ) . j. b. staunton and b. l. gyorffy , phys .
* 69 * , 371 ( 1992 ) .
j. staunton , b. l. gyorffy , a. j. pindor , g. m. stocks , and h. winter , j. phys .
phys . * 15 * , 1387 ( 1985 ) .
h. a. mook , j. w. lynch , and r. m. nicklow , phys .
lett . * 30 * , 556 ( 1973 ) .
j. crangle and g. m. goodman , proc .
london , ser .
a * 321 * , 477 ( 1971 ) .
the mfa criterion of validity is @xmath33 where @xmath34 is the effective number of neighbors inside the interaction range . v.
p. antropov , b. n. harmon , and a. n. smirnov , j. magn .
mater . * 200 * , 148 ( 1999 ) .
r. l. mcgreevy , j. phys . : condens .
matter * 13 * , r877 ( 2001 ) . v.
l. moruzzi , _ et .
b * 34 * , 1784 ( 1986 ) .
|
spin - disorder resistivity of fe and ni and its temperature dependence are analyzed using noncollinear density functional calculations within the supercell method .
different models of thermal spin disorder are considered , including the mean - field approximation and the nearest - neighbor heisenberg model .
spin - disorder resistivity is found to depend weakly on magnetic short - range order . if the local moments are kept frozen at their zero - temperature values , very good agreement with experiment is obtained for fe , but for ni the resistivity at elevated temperatures is significantly overestimated .
agreement with experiment for fe is improved if the local moments are iterated to self - consistency .
the overestimation of the resistivity for paramagnetic ni is attributed to the reduction of the local moments down to @xmath0 .
overall , the results suggest that low - energy spin fluctuations in fe and ni are better viewed as classical rotations of local moments rather than quantized spin fluctuations that would require an @xmath1 correction .
|
the aggregation of many individual preferences into a single , `` social '' preference is a long - studied problem in mathematical social sciences which has more recently also been considered from a statistical mechanics viewpoint . here by `` preference ''
we mean simply a ranking of say @xmath0 objects , i.e. a complete ordering e.g. from the favorite downwards .
a minimal consistency requirement for preferences is that they should be transitive , i.e. if @xmath3 is preferred to @xmath4 and @xmath4 to @xmath5 , then @xmath3 should be preferred to @xmath5 .
it is however well known that as soon as @xmath0 exceeds 2 the construction of the aggregate preference runs into the condorcet problem @xcite : starting from transitive individual preferences , the social preference may turn out to be intransitive , i.e. contain a cycle of the type @xmath6 ( meaning a preferred to b etc . ) .
the probability with which an intransitive ranking emerges from random individual preferences ( ` impartial culture assumption ' @xcite ) has been studied in the mathematical economics literature in the past @xcite . remarkably , this bottleneck is present to different degrees for all aggregation methods one considers and can be removed only at the cost of loosening some of the requirements that a social choice should satisfy . among the different aggregation methods ,
the pairwise - majority rule ( pmr ) has been the most studied as it appears to be more robust to the above mentioned requirements . according to pmr , the social preference corresponds to that derived by comparing objects pairwise via a simple majority rule .
recently it has been possible to quantify the extent to which pmr is effective as a social choice rule by computing the probability that the aggregate of @xmath7 random preferences is transitive within a statistical mechanics framework @xcite .
it is similarly important to understand how the situation changes if voters interact before casting their ballot . in @xcite ,
a random - field type of interaction with conformism has been considered , and it has been shown , among other results , that an interacting population may reach consensus on one of @xmath0 issues more easily the larger is @xmath0 .
a scenario that has received much attention recently is the so - called deliberative approach @xcite , where voters discuss the alternatives before the vote and eventually change their preferences .
the idea is that discussion should lead to preference harmonization and structuring and thus drastically reduce the probability of an intransitive social choice . in this paper
we study the deliberative scenario in a dynamical model of @xmath8 interacting agents or voters , whose preferences are initialized to transitive random ones to simulate the original diversity of opinions .
the key ingredient of the model is that an agent interacts only with agents whose preferences are sufficiently close to his . in such `` neighborhoods '' , by conformity he aligns to the pmr - aggregate preference ( if transitive ) .
this generalizes a classical social interaction mechanism investigated previously in e.g. @xcite . as a measure of similarity between orderings we use the kemeny distance @xcite , widely used in the social sciences ( it is related to their hamming distance )
this choice is arbitrary and it is likely that distances more sensitive to the position in the orderings may give different results .
a social preference is then formed via pmr after every agent has updated his preference .
we are interested in studying the behavior of a specific macroscopic observable , namely the probability of a transitive aggregate ranking as a function of time ( number of system - wide updates ) and of the range of interaction ( the maximum distance within which agents interact ) .
our main result is that there is an optimal interaction range ( or more properly a window of ranges ) for which the condorcet problem is much less likely to occur than in the non - interacting case , and that the frustration decreases as a function of both time and @xmath0 .
the model will be fully defined in the following section .
its complexity has prevented analytical approaches on our side .
our results are thus obtained by means of numerical simulations .
consider @xmath8 agents ( labeled by @xmath9 ) each of whom ranks @xmath0 alternatives @xmath10 .
a ranking is a complete transitive ordering such as @xmath11 . at time zero
, every agent possesses a transitive preference ranking selected randomly with uniform probability among the @xmath12 possible ones .
every ranking can be split uniquely into pair - wise comparisons of @xmath13 pairs of distinct alternatives .
we label pairs as @xmath14 and denote by @xmath15 agent @xmath16 s preferences on pair @xmath17 . specifically , @xmath18 if @xmath19 and @xmath20 otherwise ( ties are excluded ) .
the dissimilarity between the rankings of agents @xmath16 and @xmath21 is given by their kemeny distance @xmath22 where the sum runs over all @xmath13 pairs .
@xmath23 measures simply the normalized number of pairs that are ranked differently by agents @xmath16 and @xmath21 , irrespective of the position in the alignment , or in other words it is the number of adjacent pairwise switches needed to convert one preference order into the other .
we introduce an interaction range @xmath24 and define the neighborhood of @xmath16 as the set of agents whose rankings have a kemeny distance of at most @xmath1 from his : @xmath25 at every time step , an agent is selected randomly and interaction takes place .
specifically , the agent changes his ranking to that derived from a pmr among agents in his neighborhood if the latter is transitive , otherwise he keeps his preference unchanged . in subsequent interactions
the agents enters with his new ranking . after a system - wide update ( sweep )
is performed a global vote by pmr takes place , the social ranking is computed and agents move into the next round .
we describe details of the pmr procedure below ( see also @xcite ) .
we notice that as @xmath1 decreases from 1 to 0 the interaction becomes more and more of a local nature .
however it is to be expected that the number of surviving rankings decreases in time as the local interaction mimics conformist behavior on the side of agents .
furthermore , neighborhoods evolve in time ( also inside a single system - wide update ) .
the basic function which is numerically evaluated in this study is the probability that pmr yields a social transitive ranking , denoted by @xmath26 .
it is evaluated by counting the number of times a collective transitive order is obtained through pmr , out of a large number of samples .
we monitor the evolution of @xmath26 in time specifically varying the number @xmath27 of sweeps . in absence of interaction
, @xmath26 decays as @xmath0 increases , though less fast than the nave guess @xmath28 corresponding to the ratio of the number of transitive rankings to the total number of binary vectors encoding different rankings of @xmath0 objects @xcite . coming to the details of pmr voting , for each pair @xmath29 ,
let @xmath30 where @xmath31 is heaviside function .
clearly , @xmath32 if the majority of agents prefers @xmath33 over @xmath34 , whereas @xmath35 if the majority ranks @xmath34 over @xmath33 . the aggregate order ( either local or global ) emerges from separate majority votes over all pairs , that is from a computation of the @xmath36 , for all @xmath29
. there is a simple method to check whether a ranking defined through the different @xmath36 s is transitive .
indeed each @xmath36 can be seen as the element of a @xmath37 matrix since @xmath38 .
let us then write explicitly @xmath36 as @xmath39 and let @xmath40 one easily understands ( e.g. by induction starting from small @xmath8 and @xmath0 ) that if the pmr ranking is transitive , then the @xmath0-vector @xmath41 will contain once and only once each of the integers @xmath42 as elements .
in fig . [ figure3 ] we display the time evolution ( in units of sweeps ) of @xmath43 as a function of @xmath1 for a system of size @xmath44 ( the choice of @xmath45 alternatives is here only a matter of convenience ; qualitatively identical results occur for larger values of @xmath0 ) .
@xmath26 as a function of the interaction range @xmath1 and of the number of sweeps @xmath27 .
parameters : @xmath44 , @xmath46 ; average over @xmath47 samples.,width=377 ] it is evident that after just a cycles the dependence of @xmath43 on the radius of interaction becomes stable .
it is clear , also from this graph , the onset of a peculiar regime in a window of interaction radii comprised between @xmath48 and @xmath49 , where the probability of getting a transitive ranking is considerably enhanced with respect to the non interacting case .
it is also worth noting that , until @xmath1 is sufficiently small , @xmath43 is almost constant and coincident with the corresponding probability in the non interacting case ; then it abruptly increases and reaches the value of @xmath50 at @xmath51 . for larger ranges ,
@xmath43 decreases again to the non - interacting case . in order to characterize the scaling of the distribution with increasing system size , in fig .
[ figure2 ] we show the dependence of @xmath43 on the number of agents @xmath8 and on the radius of interaction @xmath1 ( the asymptotic states is reached in all cases shown ) .
the reference value for a non - interacting system is @xmath52 .
@xmath43 as a function of the radius of interaction @xmath1 and the number of agents @xmath8 after @xmath53 sweeps .
average over @xmath47 samples .
for @xmath54 and @xmath55 one recovers the value observed for the non interacting case.,width=377 ] remarkably , the existence of an optimal interaction range is reinforced by increasing the number of alternatives , as shown in fig .
[ figure4 ] .
@xmath26 normalized by the corresponding @xmath56 for non interacting agents as a function of @xmath1 .
parameters are the same as in the previous figure , except for @xmath57 , for which @xmath58 and the average is performed over 500 samples . here , @xmath59.,width=377 ] note that for @xmath60 the probability of getting a transitive majority ranking is still that of the non - interacting case . for @xmath61
this probability increases , with a gradual gain which abruptly falls down for larger @xmath1 . to shed light on these observations , we report in fig .
[ figure5 ] , for different @xmath0 , the @xmath1-dependence of the average fraction of neighbors , i.e. the average number of agents with preference orders whose kemeny distance does not exceed @xmath1 .
average fraction of neighbors as a function of @xmath1 and @xmath0 .
parameters are the same as in the previous figure.,width=377 ] this parameter is computed , after @xmath27 sweeps , by recording the number of neighbors of each agent in a sample and averaging over samples . by increasing @xmath0 there is a marked tendency of the curve to acquire a sigmoidal shape .
the average fraction of neighbors increases gradually , with a maximum rate at @xmath62 , and saturates to @xmath63 for large enough @xmath1 .
this result indicates that there is a cooperative transition between a local regime and a global , effectively long range , regime . in the former ,
each agent interacts and confronts its preference order only with a small mass of the other agents ; in the latter , each agent s opinion is influenced by the opinions of most of the others .
note that in the cases @xmath64 and @xmath60 the @xmath1 dependence is monotonously increasing , but no sign of cooperativity is present yet , in accordance with the result of the previous figure . to investigate further this transition
, we consider in fig .
[ figure6 ] the @xmath1 dependence of the mean fraction of different transitive preference orders present in a population of agents after a long update cycle .
mean fraction of different transitive preference orders surviving after @xmath59 sweeps as a function of @xmath1 .
parameters are the same as in the previous figure.,width=377 ] remember that initially , before the update cycle starts , transitive preference orders are randomly distributed among the agents .
below @xmath65 the interaction preserves all the possible diversity of preference orders , while above the cooperative transition has taken place and there is a reduction of diversity . for @xmath61 at @xmath51 only about ten percent of the possible different transitive rankings are still present among the agents .
note that for sufficiently large @xmath1 one observes an increase of different transitive preference order , as is to be expected in view of the fact that the interaction mechanism in systems with very large and very small @xmath1 is qualitatively similar .
this phenomenon should be confronted with the result in fig .
[ figure7 ] , where the @xmath1-dependence of the fraction of samples with only one surviving preference order .
fraction of samples in which only one ranking survives after @xmath59 sweeps as a function of @xmath1 .
parameters are the same as in the previous figure.,width=377 ] note , that at @xmath55 this fraction coincides , for different @xmath0 , with the corresponding @xmath26 values of the non interacting case .
figures [ figure6 ] and [ figure7 ] show that update interaction cycles induce , if the cooperative transition is established , a reduction of the different transitive preference orders among which to choose in the pmr . moreover , for large enough interaction radii , when practically each agents interacts with all the others , there is a marked tendency towards the emergence of a dominant transitive preference order . as is to be expected , this tendency is contrasted by the increase in @xmath0 , the number of alternatives . in fig .
[ figure8 ] and [ figure9 ] we show the pair distribution function of the kemeny distances among individual rankings , as a function of @xmath1 .
we consider in these figures the discriminating cases : @xmath60 and @xmath46 .
pair distribution function @xmath66 of the kemeny distances between the preference orders of pairs of different agents , after the interaction process .
@xmath60 @xmath59 ; @xmath44 and @xmath47 samples.,width=10 ] pair distribution function @xmath66 of the kemeny distances @xmath67 between the preference orders of pairs of different agents , after the interaction process .
@xmath46 @xmath59 ; @xmath44 and @xmath47 samples.,width=10 ] in both cases , if @xmath1 is less than @xmath68 , the pair distribution has a maximum for @xmath69 ( corresponding to @xmath70 adjacent pairwise switches needed to convert one ranking into the other ) .
this is essentially the pair distribution function corresponding to the initial random assignment of preference orders . as an effect of the interaction
it is seen that , when the cooperative transition is completed , i.e. for large enough @xmath1 , a peak at @xmath71 emerges , corresponding to the formation of a big clique of agents with the same transitive preference order .
this points to an interesting herding effect , whose details deserve further investigation .
we have studied the emergence of the condorcet problem in a deliberative multi - agent scenario . we have studied in particular if and how the frustration expressed by condorcet s paradox can be mitigated by interaction among the agents , before the global pmr voting takes place .
we have introduced an interaction scheme based on local pmr voting among `` neighboring '' agents , whose preference orders are close , to mimic conformist behavior .
our results point to the existence of two regimes controlled by the interaction range @xmath1 , with a crossover from one to the other at intermediate @xmath1 . for low @xmath1 ,
the probability of getting a transitive outcome is unaffected ( with respect to the non - interacting case ) , whereas for sufficiently large @xmath1 a marked enhancement in @xmath26 is observed , which increases with @xmath0 , the number of alternatives to be ranked .
a herding phenomenon is furthermore observed which reduces the repertoire of different surviving rankings .
so , if the radius of interaction is too large it is difficult to have a transitive pmr outcome , but in the case it is reached that happens because all the agents practically vote in the same way .
these results support the claims that deliberative systems reduce the chance for the formation of cycles in the social choice , provided the interaction range lies in the optimal window
. it would be important to get a deeper insight on the microscopic details of this model .
in particular the computation of two point correlations , like the probability that after an interaction cycle two agents taken at random belong to the same neighborhood , or the probability that after the interaction two initially separated neighborhoods become overlapping .
the further investigation of the clustering dynamics ( including coalescence of neighborhoods ) could provide important insight also in directions different than statistical mechanics @xcite .
reference can be made to the original essay : m.j .
a. n. de caritat , marquis de condorcet , _ essai sur lapplication de lanalyse la probabilit des decision rendues a la pluralit des voix._. l imprimrie royale , paris , 1785 . j. moulen , l. diffo lambo , _ thorie du vote _ , hermes , paris , 2001 .
p. dasgupta and e. maskin .
290 * 92 ( 2004 ) g. weisbuch , g. deffuant , j.p .
interacting agents and continuous opinions dynamics_. working papers 01 - 11 - 072 , santa fe institute , 2001 j.g .
kemeny and l.j .
snell , _ preference ranking : an axiomatic approach _
models in the social sciences , mit press , cambridge , ma , 1978 .
d.g . saari and v.r .
soc . choice and welfare
* 17 * 403 ( 2000 )
|
we present numerical results on a complex dynamical model for the aggregation of many individual rankings of @xmath0 alternatives by the pairwise majority rule under a deliberative scenario .
agents are assumed to interact when the kemeny distance between their rankings is smaller than a range @xmath1 .
the main object of interest is the probability that the aggregate ( social ) ranking is transitive as a function of the interaction range .
this quantity is known to decay fast as @xmath0 increases in the non - interacting case . here
we find that when @xmath2 such a probability attains a sharp maximum when the interaction range is sufficiently large , in which case it significantly exceeds the corresponding value for a non - interacting system .
furthermore , the situation improves upon increasing @xmath0 .
a possible microscopic mechanism leading to this counterintuitive result is proposed and investigated . social choice , condorcet paradox , pairwise majority rule 89.65.-s , 05.10.-a , 02.50.ey
|
time evolution of an isolated quantum system can be described by unitary operators .
quantum dynamics corresponds then to an evolution in the space of quantum pure states , since a given initial state @xmath0 is mapped into another pure state @xmath1 , where @xmath2 . here
@xmath3 represents a hermitian hamiltonian of the system and the time @xmath4 is set to unity .
if the underlying classical dynamics is chaotic the hamiltonian @xmath3 or the evolution operator @xmath5 can be mimicked by ensembles of random unitary matrices @xcite . in particular , spectral properties of an evolution operator of a deterministic quantum chaotic system coincide with predictions obtained for the dyson ensembles of random unitary matrices @xcite .
the symmetry properties of the system determine which ensemble of matrices is applicable .
for instance , if the physical system in question does not possess any time - reversal symmetry , one uses random unitary matrices of the circular unitary ensemble ( cue ) . if such a symmetry exists and the dimension of the hilbert space is odd one uses symmetric unitary matrices of the circular orthogonal ensemble ( coe ) @xcite . if the quantum system @xmath6 is not isolated , but it is coupled with an environment @xmath7 , its time evolution is not unitary .
one needs then to characterize the quantum state by a density operator @xmath8 , which is hermitian , @xmath9 , positive , @xmath10 and normalized , tr@xmath11 .
time evolution of such an open system can be described in terms of master equations @xcite , which imply that the dynamics takes place inside the set of quantum mixed states .
the coupling of the system @xmath6 with an environment @xmath7 can heuristically be described by adding to the hamiltonian an anti - hermitian component , @xmath12 , where @xmath13 is an operator representing the interaction between both systems @xcite .
the corresponding ensembles of non - hermitian random matrices with spectrum supported on the lower half of the complex plane were studied in @xcite . for any positive value of the coupling strength parameter @xmath14 the dynamics of the system is not unitary and eigenvalues of the evolution operator move from the unit circle inside the unit disk
see fig .
[ fig_schv]b. a similar situation occurs if one takes into account dissipation in the system .
such a dynamics of eigenvalues of a non - unitary evolution operator in the complex plane was analyzed by grobe et al .
@xcite and later reviewed by haake @xcite .
time evolution of an open quantum system can also be described in terms of a global unitary dynamics @xmath15 , which couples together a system @xmath16 with an other subsystem @xmath17 , followed by averaging over the degrees of freedom describing the auxiliary subsystem .
technically , the image of an initial state @xmath8 of the system is obtained by a partial trace over the subsystem @xmath17 , @xmath18 $ ] , where @xmath19 denotes the initial state of the environment .
the map @xmath20 defined in this way is completely positive and preserves the trace , so it is often called a _ quantum operation _ @xcite . note that in this approach both _ interacting _ subsystems @xmath16 and @xmath17 are set on an equal footing .
the second system , usually referred to as an environment , is in fact treated symmetrically , and one may also consider a dual operation , in which the partial trace is taken over the principal subsystem @xmath16 - compare fig .
[ fig_schv]c . characterized by a hamiltonian @xmath3 and a unitary evolution operator @xmath2 ; b ) open quantum system @xmath6 .
the influence of an environment @xmath7 can be described by an anti - hermitian part of the hamiltonian , @xmath21 ; and c ) interacting systems @xmath6 and @xmath7 , in which the global evolution is unitary , and the non - unitarity of the evolution of @xmath16 is due to the partial trace over the subsystem @xmath17 .
panels a ) , b ) and c ) show exemplary spectra of the corresponding evolution operators , which belong to the unit disk on the complex plane @xmath22.,scaledwidth=50.0% ] a quantum map can be described by a _
@xmath23 , which acts on the space of density operators .
if @xmath24 denotes the size of a density matrix @xmath8 , the superoperator is represented by a matrix @xmath23 of size @xmath25 .
in general such a matrix is not unitary , but it obeys a quantum analogue of the frobenius perron theorem , so its spectrum is confined to the unit disk @xcite .
spectral properties of superoperators representing some exemplary interacting quantum systems were analyzed in @xcite .
it is worth to add that spectra of quantum superoperators are already experimentally accessible : weinstein et al .
@xcite study spectra of superoperators corresponding to an nmr realization of exemplary quantum gates . for a quantum operation @xmath23 there exists an invariant state @xmath26 . in a generic case of a typical ( random ) operation
such an invariant state is unique @xcite .
if the action of the map is repeated @xmath27 times any initial quantum state @xmath8 converges to @xmath28 exponentially with the discrete time @xmath27 .
the rate of this convergence is governed by spectral properties of the superoperator , which can be characterized by the _ spectral gap _ , defined as the difference between moduli of the two largest eigenvalues .
the main aim of this work is to analyze spectra of evolution operators representing interacting quantum systems .
we demonstrate that under the condition of strong classical chaos and strong decoherence these spectral properties are universal and correspond to an ensemble of random operations @xcite . in other words
, we explore the link between quantum chaotic dynamics and ensembles of random matrices . the analysis performed earlier for unitary quantum dynamics @xcite ( see fig . [
fig_schv]a ) is extended for a more general case of non - unitary time evolution of interacting quantum systems .
this problem can be described by an approach closely related to the one used earlier to characterize quantum dissipative dynamics . to describe spectra of such non - unitary evolution operators grobe et al .
@xcite applied random matrices of the complex ginibre ensemble @xcite .
the key idea of this work can be visualized in fig .
[ fig_relations ] , which shows a bridge established between interacting quantum systems , appropriate ensembles of random operations and ensembles of ginibre matrices .
since a superoperator describing one - step evolution operator can be represented as a real matrix @xcite , we are going to apply random matrices of _ the real ginibre ensemble _ @xcite .
in particular , we investigate time evolution of initially random pure states in a deterministic model of quantum baker map periodically subjected to quantum measurements , study the speed of their convergence to the invariant state and compare the results with those obtained for an appropriate ensemble of random operations .
this paper is organized as follows . in section
ii we recall necessary definitions and introduce several versions of a deterministic quantum system model : the quantum baker maps subjected to measurement process . in section iii
we introduce various ensembles of random maps . in section iv the spectra and spectral gaps of superoperators corresponding to baker map are analyzed and compared with the spectra of random operations . in section v we continue to analyze spectra of random quantum operations and give a proof for the real ginibre conjecture put forward in @xcite , and we study the fraction of the real eigenvalues . for large dimension
@xmath24 this ratio is found in agreement with the predictions obtained for the real ginibre ensemble .
we shall start reviewing the necessary notions and definitions .
let us define the set of quantum states @xmath29 which contains all hermitian , positive operators @xmath8 of size @xmath30 with trace set to unity . a quantum linear map @xmath23 acting on @xmath29 is called _ completely positive _ ( cp ) if positivity of the extended map , @xmath31 holds for an arbitrary dimension @xmath32 of the extension , and _ trace preserving _ ( tp ) if @xmath33 any cp tp map is called quantum operation or _
stochastic map_. if a quantum operation @xmath23 preserves the identity , @xmath34 , the map is called bistochastic . according to the dilation theorem of stinespring @xcite any cp map may be represented by a finite sum of @xmath32 kraus operators , @xmath35 if the kraus operators @xmath36 satisfy the identity resolution , @xmath37 , the map @xmath23 is trace preserving .
the corresponding superoperator can be expressed as a sum of the tensor products @xcite , @xmath38 where the @xmath39 denotes the complex conjugation .
let @xmath40 with @xmath41 denote the spectrum of @xmath23 ordered with respect to the moduli , @xmath42 .
a quantum stochastic map @xmath23 sends the compact , convex set @xmath29 of mixed density matrices into itself .
hence such a map has a fixed point , the invariant quantum state @xmath26 .
thus the spectrum of any superoperator @xmath23 representing a quantum operation contains an eigenvalue @xmath43 equal to unity , while all other eigenvalues belong to the unit disk . in the case of unitary dynamics
the leading eigenvalue is degenerated , but for a random stochastic map the invariant state is generically unique @xcite . in this case
any pure state , @xmath44 , converges to the equilibrium state @xmath28 if transformed several times by the map @xmath23 . to analyze the rate of convergence to @xmath28 we analyze an average trace distance to the invariant states , @xmath45 where @xmath4 denotes the discrete time ( i.e. the number of consecutive actions of the map @xmath23 ) , while the average is taken over the ensemble of initially random pure states , @xmath46 . in the case of a generic map an exponential convergence to equilibrium ,
@xmath47 was reported @xcite .
the convergence rate depends on spectral properties of the superoperator @xmath23 .
the spectrum can be characterized by the _ spectral gap _ , @xmath48 , which generically determines the convergence rate , @xmath49 .
the formalism of discrete quantum maps is applicable to describe a deterministic quantum system , periodically interacting with an environment . in this work we concentrate on quantum dynamical systems , the classical analogues of which
are known to be chaotic . following the model of balazs and voros
@xcite we consider the unitary operator describing the one step evolution model of quantum baker map , @xmath50 .\ ] ] here @xmath51 denotes the fourier matrix of size @xmath24 , @xmath52_{jk}=\exp({\sf i } j k / 2\pi n)/\sqrt{n}$ ] and it is assumed that the dimension @xmath24 of the hilbert space @xmath53 is even .
the standard quantum baker map @xmath54 may be generalized to represent an asymmetric classical map , @xmath55\ ] ] where @xmath56 is an integer asymmetry parameter chosen in such a way that the ratio @xmath57 is integer .
the standard model , obtained in the case @xmath58 corresponds to the classically chaotic dynamics characterized by the dynamical entropy @xmath3 equal to @xmath59 @xcite .
this system can be considered as a @xmath60dimensional lift of an @xmath61dimensional non - symmetric shift map , @xmath62\\&\\~kx - k+1 & : x\in((k-1)/k,1]\end{array}\right . .\ ] ] chaos in such a system can be characterized by its dynamical entropy , equal to the mean lyapunov exponent , averaged with respect to the invariant measure of the classical system . since the uniform measure is invariant with respect to the map @xmath63 , the dynamical entropy @xmath64 is equal to the mean logarithm of the slope @xmath65 , @xmath66 the entropy is maximal in the case @xmath67 , while in the limit @xmath68 the entropy tends to zero .
hence the larger value of the parameter @xmath69 is , the weaker chaos in the classical system becomes .
in the case of the quantum system acting on the @xmath24dimensional hilbert space the largest possible value of the asymmetry parameter reads @xmath70 .
thus the limiting case of the classically regular system can not be obtained for any finite @xmath71 the limit of vanishing dynamical entropy , @xmath72 can be approached only in the classical limit @xmath73 of the quantum system . a generalized variant of a non - unitary baker map introduced by saraceno and vallejos described a dissipative quantum system @xcite .
in this work we study another model of non - invertible quantum baker map analyzed in @xcite , which is deterministic , conserves the probability , and is capable to describe projective measurements or a coupling with an external subsystem .
in general there exist @xmath32 different outcomes of the measurement process and thus the map is described by a collection of @xmath32 kraus operators .
the simplest nontrivial case of @xmath74 corresponds to dividing of the phase space into two parts , which we can choose to be the lower and the upper part .
such a measurement scheme allows one to write down the quantum operation corresponding to the sloppy baker map , in which both pieces of the classical phase space are not placed precisely one by another , but in each step an overlap of a positive width takes place . in the classical model the upper piece of the phase space
is shifted down by @xmath75 - see fig .
[ fig_sloppy]c - so the invariant measure lives in the rectangle of the width @xmath76 .
baker map @xmath77 , c ) irreversible sloppy baker map @xmath78 in which in each step the upper part of the phase space is shifted down by @xmath75 , d ) double sloppy baker map @xmath79 in which both parts of the phase space are shifted vertically by @xmath80.,scaledwidth=40.0% ] to represent the shift in the quantum analogue of the map one uses a unitary translation operator @xmath15 such that @xmath81 and any momentum eigenstate @xmath82 is shifted by one , @xmath83 .
hence the shift down by @xmath75 is realized by the unitary operator , @xmath84 .
thus the stochastic map describing the quantum sloppy map @xcite
@xmath85 consists of two kraus operators , which act on the unitarily rotated state @xmath86 , @xmath87f_n,\nonumber\\ d_t&=v^{-n\delta/2 } f_{n}^{\dagger}\left[\begin{array}{cc}0&0\\0&\mathbbm{1}_{n/2}\end{array}\right]f_n . \label{krau2}\end{aligned}\ ] ] the operator @xmath88 describes the projection on the lower part of the phase space , while the definition of the operator @xmath89 includes also the operator representing the shift of the upper domain down by @xmath75 . observe that the parameter @xmath90 may take any real value from the unit interval @xmath91 $ ] .
however , the case @xmath92 corresponds to the baker map without the shift but with a measurement , so it does not reduce to the standard unitary baker map @xmath77 .
one can also consider another classical model of double sloppy map , in which both domains are simultaneously shifted by @xmath80 towards the center of the phase space @xcite fig .
[ fig_sloppy]d . to write down the corresponding quantum model
@xmath79 one needs thus to modify both kraus operators , @xmath93 and @xmath94 .
both variants of the model can be further generalized by allowing for a larger number @xmath32 of measurement operators , represented by projectors on mutually orthogonal subspaces . for simplicity
we assume here that the dimensionalities of all these subspaces are equal and read @xmath95 .
varying the parameter @xmath32 one may thus control the degree of the interaction of the baker system with the environment and study the relation between the decoherence in the interacting quantum system and the spectrum of the corresponding superoperator . increasing the asymmetry parameter @xmath69 one can decrease the degree of classical chaos . to increase the degree of chaos
one may just apply the quantum baker map twice , since the classical dynamical entropy of such a composite map is equal to @xmath96 .
in general one can allow for an arbitrary number of @xmath97 of unitary evolutions , and replace unitary @xmath54 by @xmath98 .
alternatively one can say that the non - unitary measurement operation is performed only once every @xmath97 periods of the unitary evolution .
choosing the parameter @xmath97 to be of order of @xmath24 one can assure that the quantum dynamics is as chaotic as allowed by the quantum theory , what can be quantitatively characterized by the quantum dynamical entropy @xcite .
therefore the generalized model of quantum sloppy baker map we are going to analyze here depends on the classical asymmetry parameter @xmath69 , the width of the classical shift @xmath90 , the number of free evolutions @xmath97 , and the quantum parameter @xmath32 denoting the number of measurement operators , @xmath99 additionally , for each set of parameters of the model one can choose the appropriate set of projection operators @xmath100 which correspond to the shift applied on one or on two parts of the classical phase space .
note that the measurement process can also be interpreted as an interaction with a measurement apparatus , described by an auxiliary hilbert space of @xmath32 dimensions .
thus the model ( [ mapgen ] ) represents an interacting quantum system and belongs to the general class of quantum operations defined by ( [ kraus1 ] ) .
a rich structure of the model and the possibility to tune independently several parameters of the quantum system allows us to treat this model as a valuable playground to investigate spectral properties of superoperators , which represent non - unitary dynamics of interacting quantum systems .
parameter @xmath74 , and the shift width @xmath101 are kept fixed .
a generic spectrum for @xmath102 is shown on panel b ) .
the subplots a ) and c ) are obtained for the cases of a weak classical chaos for @xmath103 and @xmath104 respectively , while the last case d ) shows the spectrum for the double sloppy map @xmath79 for @xmath105 and @xmath106 . , scaledwidth=49.0% ] we constructed quantum operations representing the generalized sloppy baker map ( [ mapgen ] ) for several sets of the parameters of the model . in each case the superoperator @xmath23 was obtained according to the expression ( [ phimap ] ) and diagonalized to yield the complex spectrum belonging to the unit disk .
in the case of several measurement operators , @xmath107 , the quantum baker map represents a non - unitary dynamics . under the condition of classical chaos
the leading eigenvalue @xmath108 is not degenerated and all remaining eigenvalues are located inside the disk of the radius equal to the modulus of the subleading eigenvalue @xmath109 .
the spectra of the superoperator of the generalized sloppy baker map ( [ mapgen ] ) were found to depend weakly on the shift parameter @xmath110 however , other parameters of the model ( namely @xmath111 and @xmath32 ) influence properties of the spectrum considerably - see fig .
[ qbm64s_klm ] .
as the asymmetry parameter @xmath69 increases the difference between the sizes of two domains which form the classical phase space becomes larger .
in the extreme limit of @xmath112 the classical system becomes only marginally chaotic , the eigenvalues are attracted to the unit circle and the spectral gap @xmath48 disappears . on the other hand ,
if we increase the degree of the classical chaos by increasing the number @xmath97 of unperturbed unitary evolutions , the size of the spectral gap does not change , but the spectrum fills the complex disc of radius @xmath113 almost uniformly .
eventually , an increase of the number @xmath32 of the measurements results in a faster decoherence in the system .
this is reflected by an increase of the spectral gap @xmath114 .
in fact the radius @xmath115 of the disk supporting the spectrum decreases with @xmath32 as @xmath116 .
this observation - demonstrated in fig .
[ p2nde ] - will be explained in section [ sec : ginib ] .
in this section we propose three different ensembles of random stochastic maps acting on the space @xmath29 of mixed states of size @xmath24 with different physical interpretation .
we assume that all unitary matrices @xmath5 used below are drawn according to the haar measure on the unitary group of corresponding dimension unless stated otherwise .
_ 1 . environmental representation _ of a random stochastic map @xcite . choose a random unitary matrix @xmath5 of composite dimension @xmath117 and construct a random map as @xmath118 .
\label{phie}\ ] ] it is assumed here that the environment , initially in an arbitrary pure state @xmath119 is coupled with the system @xmath8 by a random global unitary evolution @xmath5 . the stochastic map is obtained by performing the partial trace over the @xmath32-dimensional environment .
_ 2 . random external fields _ defined as a convex combination of @xmath32 unitary evolutions @xcite @xmath120 where @xmath121 are positive components of an arbitrary probabilistic vector of size @xmath32 , @xmath122 .
all unitaries @xmath123 are independent random haar matrices .
random external fields form an example of bistochastic maps .
they represent the physically relevant case in which the quantum system is subjected randomly with one of @xmath32 given unitary operations and can also be interpreted as quantum iterated function systems @xcite .
_ 3 . projected unitary matrices _ acting on states of a composite dimension , @xmath124 .
all @xmath32 kraus operators are formed by unitarily rotated projection operators , @xmath125 for @xmath126 which leads to the map @xmath127 where @xmath5 is a fixed random unitary matrix . here
@xmath128 denote projective operators on @xmath69 dimensional mutually orthogonal subspaces , which satisfy the identity resolution , @xmath129 this ensemble of bistochastic maps corresponds to a model of deterministic quantum systems , in which unitary dynamics is followed by a projective measurement . , and
b ) ensemble @xmath130 ( for definitions see sec .
ii ) . all maps act on quantum states of size @xmath131 , while the parameter of the model reads @xmath132 note that in all cases the spectrum is contained in the disk of radius @xmath133 apart of the leading eigenvalue marked by @xmath134.,scaledwidth=50.0% ] in the ensembles of random maps defined above the integer number @xmath135 determines the number of kraus operators and serves as the only parameter of each ensemble of random maps .
observe that in the special case @xmath136 the dynamics reduces to the unitary evolution , so both variants of the model are used to describe quantum systems with or without a generalized antiunitary symmetry @xcite . as shown in @xcite the flat measure in the set of stochastic matrices
is obtained for the coupling of the system with an environment of dimension @xmath137 , so that the choi matrix , @xmath138 of size @xmath25 has full rank . here
@xmath139 represents the maximally entangled state on the bipartite hilbert space , @xmath140 . due to the theorem of choi the condition of complete positivity of the map is equivalent to positivity of the choi matrix
, @xmath141 in general the discrete parameter @xmath32 characterizes the strength of the non - unitary interaction and we shall vary it from unity ( unitary dynamics ) to @xmath25 , which describes a generic random stochastic map .
we have generated several realizations of random maps from the ensembles @xmath142 and @xmath130 introduced above .
exemplary spectra of superoperators for maps pertaining ensembles obtained for @xmath74 are shown in fig .
[ rqo_exspectra ] . in the latter case we superimposed the spectra from two realizations of the map @xmath130 since by construction @xmath143 eigenvalues of the superoperator are equal to zero . in general ,
the spectra of random maps could be used to describe the spectra of deterministic system ( [ mapgen ] ) under the condition of classical chaos and large decoherence .
numerical results performed for various models of quantum maps reveal an exponential decay of the mean trace distance ( [ avdist ] ) to the invariant state .
a comparison of such a time dependence of the mean trace distance @xmath144 for quantum baker maps and random quantum maps is shown in fig .
[ pure_decay ] .
interestingly , for a fixed system of size @xmath24 the convergence rate @xmath145 increases with the number of the measurements as @xmath146 , but it seems not to depend on the assumption , whether the global random evolution matrix @xmath5 is taken from the orthogonal or unitary circular ensemble .
0.2 cm of random pure states to the invariant state for a ) exemplary random quantum maps @xmath142 and b ) quantum baker map with parameters : @xmath147 figure drawn in a log - scale shows the exponential decay for dimension @xmath148 and parameters @xmath149 and @xmath150 ( @xmath151 ) .
the average is taken over the set of @xmath152 initial projectors .
, scaledwidth=50.0% ] further numerical investigations confirm an expected relation between the size rate of the convergence of a typical quantum state towards the fixed point of the map and the spectral gap . as shown in fig .
[ p2nde ] obtained for random operations as well as the generalized quantum baker map the radius @xmath115 of the disk in the complex plane , which contains all but the leading eigenvalue , decreases with the number of measurements as @xmath153 in the next section we present an explanation of this relation based on the theory of random matrices . as a function of the dimension @xmath32 of the environment for random quantum operation @xmath142 for @xmath154 and for the sloppy baker map for @xmath155 ( @xmath156 ) .
solid line shows the fit according to eq .
looking at the spectrum of a random stochastic map @xmath142 shown in fig .
[ rqo_exspectra ] one can divide the entire spectrum into three parts : i ) a single eigenvalue @xmath108 ; ii ) @xmath157 real eigenvalues distributed along the real line with a density @xmath158 , iii ) remaining @xmath159 complex eigenvalues @xmath40 , the distribution of which can be described by a density @xmath160 on a complex plane .
any density operator @xmath8 of size @xmath24 can be represented using the generalized bloch vector representation @xmath161 here @xmath162 denotes the generators of @xmath163 such that @xmath164 and @xmath165 since @xmath166 @xmath167 for @xmath168 the real vector @xmath169 $ ] is called the generalized bloch vector . thus @xmath170 the bloch vector can also be used to represent an arbitrary operation . using kraus operators @xmath36 one represents the element @xmath171 of the superoperator @xmath23 in a form @xmath172 where @xmath173 .
this square matrix of order @xmath174 will be called @xmath175 . in a similar way we introduce the vector @xmath176 and find that the remaining elements of the matrix @xmath23 do vanish ,
@xmath177 hence the superoperator @xmath23 can be represented as a real asymmetric matrix @xmath178 , \label{super2}\ ] ] where the @xmath179 dimensional vector @xmath176 represents a translation vector while the @xmath180 real matrix @xmath175 is a real contraction @xcite .
thus the spectrum of @xmath23 consists of the leading eigenvalue , equal to unity , and the spectrum of @xmath175 .
note that the complex eigenvalues of the real matrix @xmath175 appear in conjugated pairs , @xmath181 and @xmath182 , which is a consequence of the fact , that the map @xmath23 sends the set of hermitian operators into itself , so as seen above the superoperator can be represented by a real matrix @xcite . since a map acting on states of size @xmath24 is represented by a superoperator of dimension @xmath25 the following normalization relation holds , @xmath183 . in the case that @xmath175 has only real eigenvalues one can bring @xmath175 by an orthogonal transformation @xmath184 to lower triangular form @xmath185 where @xmath186 while @xmath187 has elements only below the diagonal . thus @xmath188{\cal o}^{-1}.\end{aligned}\ ] ] hence the measure @xmath189 is given by @xmath190 where the vandermonde determinant is the jacobian of the transformation from @xmath191 to @xmath192 thus the measure @xmath193 has the form @xmath194 where in the last factor the positivity conditions of the corresponding choi matrix is averaged by integrated over the measure @xmath195 this factor is expected to be a smooth function of the eigenvalues @xmath196 in the case that @xmath175 has a certain number of complex conjugate eigenvalues @xmath197 is of similar form , but the product of differentials @xmath198 has to be interpreted as exterior product @xcite .
it turns out that for large dimension @xmath199 the measure @xmath193 is given by the real ginibre ensemble with the bulk of eigenvalues inside a certain disk in the complex plane . to prove this
let us go back to the matrix representation of @xmath23 in terms of @xmath32 kraus operators @xmath200 then @xmath201 where @xmath202 are the matrix elements of kraus operators and
@xmath39 denotes the complex conjugation ; @xmath203 the kraus operators obey @xmath204 thus it is natural to assume that @xmath202 represent @xmath24 columns of a matrix @xmath5 drawn from a circular unitary ensemble of dimension @xmath117 i.e. @xmath205 using formulas by mello @xcite for the first four moments of @xmath206 we are able to find exactly the first two moments of matrix elements @xmath207 .
for example for @xmath208 here @xmath209 means the averge over the unitary group .
this implies here : @xmath210 in this way we can derive the exact second moments : @xmath211 we see that in the limit of large @xmath24 the first two cumulants are identical to those of the gaussian distribution ( with variance @xmath212 ) @xmath213 the factor @xmath214 is due to the symmetry property @xmath215 .
we will argue below that for large @xmath32 in addition the higher cumulants can be neglected .
hence the superoperator @xmath23 associated with a random map can be described ( up to the one eigenvalue @xmath61 ) by the real ginibre ensemble with eigenvalues inside a disk of radius @xmath216 where @xmath32 is the number of random kraus operators defining @xmath217 this can also be seen by going back to the real matrix representation - .
let us now argue that for large @xmath32 we can neglect higher cumulants .
first of all for large @xmath24 the elements @xmath218 forming a minor of @xmath219 are essentially independent gaussian variables with zero mean and variance @xmath220 .
thus as a consequence of the central limit theorem for large @xmath32 @xmath207 as sum of @xmath32 essentially independent identically distributed variables is again gaussian with variance @xmath221 .
also in the bulk the different matrix elements of @xmath23 are independent .
the average of @xmath207 is given by @xmath222 .
to investigate the density of complex eigenvalues @xmath22 of the superoperator @xmath23 we analyzed their radial probability distribution @xmath223 , where @xmath224 .
the real eigenvalues are taken into account for this statistics .
[ ep_joint ] shows a comparison of numerical data obtained for several realizations of quantum baker map , projective random operations and real random matrices pertaining to the ginibre ensemble .
the data are represented in the rescaled variable @xmath225 so that the radius of the disk of eigenvalues is set to unity . in all three cases displayed in the figure the radial density grows linearly , which corresponds to the flat distribution of eigenvalues inside the complex disk , in agreement with the predictions of the ginibre ensemble .
these results obtained for @xmath226 show a smooth transition of the density in the vicinity of the boundary of the disk at @xmath227 , which becomes more abrupt for larger @xmath24 . in the asymptotic case @xmath228 , the density of rescaled eigenvalues is described by the _
circular law _ of girko , @xmath229 derived for complex ginibre matrices . , projective random operations @xmath230 and the spectra of real random matrices of the ginibre ensemble ( @xmath156 ) .
the size of each matrix is @xmath231 , the number of kraus operators is @xmath232 so the density is shown as a function of the rescaled radius @xmath233 .
the tail of the distribution beyond the point @xmath227 ( equivalent to @xmath234 ) does not violate therefore the quantum analogue of the frobenius - perron theorem.,scaledwidth=46.0% ] the spectra of real random ginibre matrices display a more subtle structure .
a finite fraction of all eigenvalues are real , in analogy to the mean number of real roots of a real polynomial @xcite .
real eigenvalues of a real ginibre matrix cover the real axis with a constant density . furthermore , for large dimension the density of complex eigenvalues is known to be asymptotically constant in the complex disk except for a small region near the real axis @xcite . to check for what random operations these effects can be observed in the spectrum of the superoperator we analyzed the average number @xmath235 of real eigenvalues of the superoperator @xmath23 . for any realization of @xmath23
we have @xmath236 where @xmath237 is the number of real eigenvalues of the real matrix of size @xmath238 .
these data are compared with predictions for the real ginibre ensemble , hereafter denoted by @xmath239 . the following expression for the mean number of real eigenvalues of a real ginibre matrix of size @xmath174
was derived in @xcite @xmath240 these analytical results suggest to introduce a rescaled ratio @xmath241 to make easier a comparison of data obtained for various systems of size @xmath24 .
numerical results presented in fig .
[ real_per ] show that the superoperators associated with random maps are characterized by a non - zero fraction of real eigenvalues . in the case of strong interaction with the ancilla of the size
@xmath137 the dynamical matrix @xmath242 has full rank and the rescaled fraction of real eigenvalues of @xmath23 coincides with the prediction for the real ginibre ensemble . of the superoperator for random operations with @xmath137 ( @xmath243 ) , @xmath244 @xmath230 and real ginibre matrices ( @xmath151 ) as a function of the matrix size @xmath24 .
solid horizontal line at @xmath245 represents the asymptotic value of the normalized ratio implied by ( [ ed_eq]).,scaledwidth=50.0% ] to demonstrate further spectral features characteristic of the real ginibre ensemble we analyzed spectra of superoperators and investigated the cross - section of the probability distribution @xmath246 along the imaginary axis .
numerical data of this distribution denoted as @xmath247 obtained for an ensemble of random maps @xmath142 acting on the states of size @xmath248 are shown in fig .
[ ginoe_cusps ] . in order to compare these data with predictions of the real ginibre ensemble
we need to assure a suitable normalization .
let us rescale the imaginary axis as @xmath249 , so that the rescaled formula ( 25 ) of @xcite takes the form @xmath250 making use of the standard estimations @xmath251 ( see formula ( 7.1.13 ) at page 298 of abramowitz and stegun @xcite ) , and the definition of the complementary error function @xmath252 , one obtains from eq .
( [ eq_erfc ] ) an explicit form for lower and upper bounds for the rescaled distribution in the vicinity of the real axis @xmath253 as shown in fig . [ ginoe_cusps ]
these bounds are rather precise and describe well the numerically observed density @xmath247 of complex eigenvalues of the superoperators along the imaginary axis . of complex eigenvalues along the rescaled imaginary axis obtained for an ensemble of random operations @xmath142 of dimension @xmath254 and with parameter @xmath255 ( @xmath156 )
are compared with lower and upper bounds ( [ bounds ] ) obtained for small @xmath256 from the real ginibre ensemble and represented by thick lines.,scaledwidth=50.0% ]
in this work we analyzed spectra of non - unitary evolution operators describing exemplary quantum chaotic systems and the time evolution of initially pure states . we have chosen to work with a generalized model of quantum baker map subjected to measurements @xcite , which allows one to control the degree of classical chaos and the strength of the interaction with the environment .
the size of the quantum effects , proportional to the ratio of the planck constant to the typical action in the system , is controlled by the size @xmath24 of the hilbert space used to describe the quantum system .
the classical limit of the quantum model corresponds to the limit @xmath228 .
due to a quantum analogue of the frobenius
perron theorem the evolution operator has at least one eigenvalue equal to unity , while all other eigenvalues are contained in the unit disk in the complex plane . in a generic case
the leading eigenvalue is not degenerated and the corresponding eigenstate represents the unique quantum state invariant with respect to the evolution operator .
investigating the time evolution of initially random pure states we found out that in a generic case they converge exponentially fast to the invariant state .
the rate of this relaxation to the equilibrium depends on the size of the spectral gap , equal to the difference between the moduli of the first and the second eigenvalues of the evolution operator . in particular ,
the relaxation rate @xmath145 was found to depend on the number of measurement operators @xmath32 as @xmath257 .
spectral properties of evolution operators of deterministic quantum systems interacting with the environment were compared with spectra of suitably defined ensembles of random matrices .
note that an idea to apply random matrices to model evolution operators of open deterministic quantum systems was put forward by pepowski and haake @xcite , but random maps used therein are not necessarily completely positive .
this property is by construction fulfilled by the ensemble of random stochastic maps introduced in @xcite and by two other ensembles of random bistochastic maps used in this work . in general
, the spectra of non - unitary operators corresponding to quantum deterministic systems display a wide variety of structures , which depend on classical parameters as the degree of chaos of the corresponding classical system characterized quantitatively e.g. by its dynamical entropy .
the spectra depend also on quantum parameters as the dimension of the hilbert space and the character of the interaction with the environment , which governs the strength of the decoherence effects . however , under an assumption of strong classical chaos and a uniform coupling of the system analyzed with all the states of the @xmath32dimensional environment the spectral properties the corresponding evolution become _ universal _ : the spectrum consists of a leading eigenvalue equal to unity , while all other eigenvalues cover the complex disk of radius @xmath258 . in the asymptotic limit @xmath228
the density of complex eigenvalues becomes uniform in the disk , besides the region close to the real axis .
as the size of the environment @xmath32 is equal to @xmath25 , which implies strong decoherence , the dynamical matrix @xmath242 describing the quantum map @xmath23 has full rank , so it can be considered as generic . in this very case
the spectral statistics of this region of the complex spectrum of the superoperator and the fraction of its real eigenvalues coincides with predictions of the _ real ginibre ensemble _ , a proof of which is given in this work .
financial support by the transregio-12 project der deutschen forschungsgemeinschaft and the special grant number dfg - sfb/38/2007 of polish ministry of science and higher education is gratefully acknowledged .
f. haake , _ quantum signatures of chaos _ , ii ed .
springer , berlin ( 2006 ) .
stckman , _ quantum chaos _ , cambridge univ .
press ( 1999 ) .
f. j. dyson , _
phys . _ * 3 * , no . 1 , 140 , 157 , 166 ( 1962 ) . m. l. mehta , _ random matrices _ , iii ed .
academic , new york ( 2004 ) .
r. alicki , k. lendi , _ quantum dynamical semigroups and their applications _ , springer , berlin ( 1987 ) . c. h. lewenkopf , h. a. weidenmller , _ ann . of phys .
_ , * 212 * , issue 1 , 53 - 83 ( 1991 ) .
f. haake , f. izrailev , n. lehmann , d. saher , h .- j .
sommers , _ z. phys .
, 88 , 359 - 370 , ( 1992 ) .
y. v. fyodorov , h .- j .
sommers , _
j. phys . a : math .
* 36 * , 3303 ( 2003 ) .
r. grobe , f. haake , h .- j .
sommers , _ phys .
lett . _ * 61 * , 1899 ( 1986 ) .
m. a. nielsen , i. l. chuang , _ quantum computation and quantum information _ , cambridge univ . press ( 2000 ) .
i. bengtsson , k. yczkowski , _ geometry of quantum states _ , cambridge univ .
press ( 2006 ) .
w. bruzda , v. cappellini , h .- j .
sommers , k. yczkowski , _ phys .
lett a _ * 373 * , 320 - 324 ( 2009 ) .
a. oziski , p. pakoski , k. yczkowski , _ phys . rev .
e _ * 66 * 065210 ( 2002 ) .
i. garca - mata , m. saraceno , m. e. spina , _ phys .
lett . _ * 91 * , 064101 ( 2003 ) .
i. garca - mata , m. saraceno , _ modern phys .
_ * b 19 * , 341 , ( 2005 ) .
j. m. pedrosa , g. g. carlo , d. a. wisniacki , l. ermann , _ phys .
e _ * 79 * , 016215 ( 2009 ) .
y. s. weinstein , t. f. havel , j. emerson , n. boulant , m. saraceno , s. lloyd , d. g. cory , _ j. chem .
* 121 * , 6117 ( 2004 ) .
j. ginibre , _ j. math .
phys . _ * 6 * , 440 ( 1965 ) .
b. m. terhal , d. divincenzo , _ phys .
_ * a61 * , 22301 ( 2000 ) .
n. lehmann , h .- j .
sommers , _ phys .
lett . _ * 67 * , 941 ( 1991 ) .
sommers , w. wieczorek , j. phys .
a * 41 * 405003 ( 2008 ) .
w. f. stinespring , _ proc . of the ams _ , 211 - 216 ( 1955 ) .
n. l. balazs , a. voros , _ ann . phys . _ ( n. y. ) * 190 * , 1 ( 1989 ) .
e. ott , _ chaos in dynamical systems _ , ii ed .
cambridge university press ( 2002 ) .
m. saraceno , r. o. vallejos , _ chaos _ * 6 * , 1054 ( 1996 ) .
r. alicki , a. oziski , p. pakoski , k. yczkowski , _ j. phys . _ * a 37 * 5157 ( 2004 ) .
m. smaczyski , _ master thesis _
, jagiellonian university , cracow 2009 , see : + ` http://chaos.if.uj.edu.pl/\simkarol/prace/s09.pdf ` .
w. somczyski , k. yczkowski , _ phys .
lett . _ * 80 * , 1880 ( 1998 ) .
f. benatti , v. cappellini , m. de cock , m. fannes , d. vanpeteghem , _ rev .
phys . _ * 15 * , 847 - 875 ( 2003 ) . v. cappellini , _ acta phys .
polon . _ * a 112 * , 589 - 605 ( 2007 ) .
a. oziski , k. yczkowski , w. somczyski , _ phys . rev _ * e 68 * , 046110 ( 2003 ) .
p. a. mello , _
j. phys . a : math .
* 23 * 4061 - 4080 ( 1990 ) .
e. bogomolny , o. bohigas , p. leboeuf , _ phys .
lett . _ * 68 * , 2726 - 2729 ( 1992 ) .
g. akemann , e. kanzieper , _ j. stat .
phys . _ * 129 * , 1159 - 1231 ( 2007 ) .
a. edelman , e. kostlan , _ bull .
soc . _ * 32 * , 1 - 37 ( 1995 ) .
a. edelman , _ journal of multivariative analysis _ , * 60 * , 203 - 232 ( 1997 ) .
sommers , _
a _ , * 40 * , 671 - 676 ( 2007 ) .
m. abramowitz , i. a. stegun , _ handbook of mathematical functions _ , nbs , washington , dc ( 1972 ) .
p. pepowski , f. haake , _
_ * a 26 * , 2473 ( 1993 ) .
|
we analyze a model quantum dynamical system subjected to periodic interaction with an environment , which can describe quantum measurements . under the condition of strong classical chaos and strong decoherence due to large coupling with the measurement device , the spectra of the evolution operator exhibit an universal behavior .
a generic spectrum consists of a single eigenvalue equal to unity , which corresponds to the invariant state of the system , while all other eigenvalues are contained in a disk in the complex plane .
its radius depends on the number of the kraus measurement operators , and determines the speed with which an arbitrary initial state converges to the unique invariant state .
these spectral properties are characteristic of an ensemble of random quantum maps , which in turn can be described by an ensemble of real random ginibre matrices .
this will be proven in the limit of large dimension .
|
several considerations ( e.g. , uv - completions for the standard model ) and observations ( e.g. , dark matter ) lead us to assume the existence of particles beyond the standard model . in order to have evaded detection so far , such new particles can be either very heavy , or rather light ( e.g. , at sub - ev scale ) if they have extremely small coupling to known particles @xcite .
the most renowned example of such proposed weakly interacting slim particles , `` wisps '' for short , is arguably the axion @xcite , which is a consequence of the peccei - quinn solution to the strong cp - problem .
in addition , strong interest has emerged recently for so - called axion-_like _ particles ( alps ) , whose mass - coupling relation is relaxed compared to the qcd axion : alps can , e.g. , appear in intermediate string scale scenarios @xcite , and constitute dark matter @xcite .
moreover they could explain yet puzzling observations in some astrophysical processes such as anomalous white dwarf cooling @xcite and the transmissibility of the universe to high - energetic photons @xcite .
further wisps can be particles of hidden sectors in string- and field - theoretic extensions of the standard model , see , e.g. , @xcite : particularly hidden photons ( hps ) , i.e. , gauge bosons of an extra u(1 ) gauge group as well hidden sector matter . the latter can acquire an electromagnetic fractional charge and thus can constitute so - called minicharged particles ( mcps ) , see @xcite for an overview .
hidden photons are also a viable dark matter candidate @xcite and could be responsible for the phenomenon of dark radiation @xcite .
in addition , wisps can appear as scalar modes in theories of massive gravity @xcite . if such wisps exist , it is expedient to search for them also by their interactions with photons . amongst others ,
this is advantageous because photons can be easily produced at high rates and do not have tree - level self - interactions within the standard model .
thus , beyond - standard - model physics becomes readily accessible .
experiments particularly apt to look for wisps with photons are of the ` light - shining - through - a - wall'-type ( lsw ) @xcite : laser photons can be converted into a wisp in front of a light - blocking barrier ( generation region ) and reconverted into photons behind that barrier ( regeneration region ) .
depending on the particle type , these conversion processes are induced by magnetic fields or are manifest as oscillations .
the most sensitive lsw laboratory setup thus far is the first stage of the any light particle search ( alps - i ) @xcite at desy . with major upgrades in magnetic length , laser power and the detection system
, the proposed alps - ii experiment aims at improving the sensitivity by a few orders of magnitude for the different wisps . following last year s workshop contribution @xcite we shortly present an updated status of alps - ii .
the reason for proposing to realize an upgraded version of alps is its sensitivity to particularly interesting parameter regions for various wisps , as indicated in the previous section .
three key ingredients are responsible for this sensitivity boost @xcite , cf . tab .
[ tab : param ] : foremost , the magnetic length of alps - ii is expected to be 468 tm . this can be achieved by a string of 10 + 10 hera dipoles which can be taken from a reserve of 24 spare magnets manufactured for hera at desy .
note that the setup of this string requires an aperture increase by straightening the beam pipe to avoid clipping losses for the laser .
the viability of this undertaking has been proven with the straightened alps - i magnet achieving quench currents above values measured for the unbent magnet @xcite .
secondly , the effective photon flux in the setup is planned @xcite to be increased through a higher power buildup in the production cavity and by virtue of ` resonant regeneration ' @xcite , i.e. , an optical resonator in the regeneration region , locked to the resonator in the generation region . to assure long - term stability of the cavity mirrors
, alps - ii will employ infrared light at 1064 nm wavelength ( instead of green 532 nm for alps - i ) .
thirdly , alps - ii will feature a nearly background - free transition edge sensor allowing for high detection efficiency even to infrared light . whilst this sensor is under development
, the alps - i ccd can be used as a fall - back option .
.parameters of the alps - i experiment in comparison to the alps - ii proposal .
the second column shows the dependence of the reachable alp - photon coupling on the experimental parameters .
the last column lists the approximate sensitivity gain for alp searches compared to alps - i . for hidden photons
, there is no gain from the magnetic field .
thus the sensitivity gain follows as above except for the factor coming from @xmath0 and amounts to 147 .
[ cols="<,^,^,^,^,^ " , ] alps - ii is set out to be realized in three stages : alps - iia is already well under way and will search for hidden photons in a 10m+10 m lsw - setup without magnets .
it is meant to demonstrate the viability of the optics setup , particularly the locking scheme of the regeneration cavity . as physics benefit
, it will be sensitive to the dark radiation hint @xcite as well as to parameter regions favored in intermediate string scale scenarios and a small region compatible with hp dark matter , see fig .
[ fig : hps ] . as already successfully used at alps - i @xcite ,
also gas will be inserted into the vacuum tube at each stage to close the gaps arising from at minima of the conversion probability .
alps - iib , still without magnets is planned to test a smooth operation at 100m+100 m length in the hera tunnel .
only the final stage , alps - iic , is planned to operate with magnets and thus explore the parameter space of axion - like particles , cf .
[ fig : alps ] .
a further field of search for physics beyond the standard model in which lasers - based setups could yield insight , is the search for a quantum theory of gravity .
as there is no tree - level background within the standard model , it is tempting to explore if measuring the cross - section of photon - photon scattering at high energies can teach us about quantum gravity ( qg ) , since the small qg signal is easier accessible without a tree - level background . to achieve high photon energies in a collider mode ,
different future options are advanced : compton - backscattering is possible within linacs or from wake - field - accelerated charged particles using pulsed high - intensity lasers @xcite . the tiny cross section for photon - photon scattering through graviton exchange @xcite
may be drastically enhanced in scenarios with extra dimensions @xcite .
this is not a new idea , but strongly deserves a revisit @xcite in the context of uv - complete theories , that do not rely on an energy cutoff scale , such as asymptotically safe gravity , see @xcite for reviews .
it is worth exploring the possibility further if laser - based searches could eventually even shed light onto qg .
laser - based laboratory searches can be a strong tool to address various questions of beyond the standard model physics . in this contribution
we have discussed the status and prospect of alps - ii , which is designed to explore particularly motivated parameter regions of different wisps .
further , we have briefly pointed at the possibility of even testing quantum gravity scenarios with purely laser - based setups .
j. jaeckel _ et al .
nucl . part .
sci . * 60 * , 405 ( 2010 ) ; a. ringwald , arxiv:1210.5081 [ hep - ph ] .
f. wilczek , phys .
lett . *
40 * , 279 ( 1978 ) ; s. weinberg , phys .
lett . * 40 * , 223 ( 1978 ) .
m. cicoli _ et al .
_ , jhep * 1210 * , 146 ( 2012 ) [ arxiv:1206.0819 [ hep - th ] ]
. j. redondo , theese proceedings ; p. arias _ et al .
_ , jcap * 1206 * , 013 ( 2012 ) .
j. isern , these proceedings ; j. isern _ et al .
_ , astrophys . j. * 682 * ( 2008 ) l109 [ arxiv:0806.2807 [ astro - ph ] ] .
m. meyer _ et al .
_ , these proceedings arxiv:1211.6405 ; d. horns and m. meyer , jcap * 1202 * , 033 ( 2012 ) .
j. redondo _ et al .
. phys . * 52 * , 211 ( 2011 ) ; p. arias _ et al . _ phys .
d * 82 * , 115018 ( 2010 ) .
f. karbstein , these proceedings ; b. dobrich _ et al .
_ , phys .
* 109 * , 131802 ( 2012 ) & arxiv:1203.4986 g. mueller , p. sikivie , d. b. tanner and k. van bibber , aip conf .
proc . * 1274 * , 150 ( 2010 ) .
m. kando _ et al .
_ , aip conf .
* 1024 * , 197 ( 2008 ) ; t. tajima _ et al .
_ , prog . theor
. phys . * 125 * , 617 ( 2011 ) .
|
high - intensity photon beams can provide for a viable probe for many particles of standard model extensions .
this workshop contribution briefly reviews the status of the second stage of the any light particle search ( alps - ii ) at desy , an experiment of the light - shining - through - a - wall type , as well as an idea to test asymptotically safe quantum gravity in a photon - scattering experiment .
|
there are several definitions of orthogonal designs ( ods ) in the literature @xcite the well known being as given in @xcite : [ cod ] a complex orthogonal design ( cod ) @xmath11 ( in short @xmath12 ) for @xmath0 transmit antennas is defined as a @xmath13 matrix such that ( i ) the nonzero entries of @xmath12 are the complex variables @xmath14 and their conjugates and ( ii ) @xmath15 where @xmath16 stands for the complex conjugate transpose and @xmath17 is the @xmath18 identity matrix .
the matrix @xmath12 is also said to be a @xmath19 $ ] cod and its rate in complex symbols per channel use is @xmath20 when @xmath21 are real variables , the designs are called real orthogonal design ( rod ) . space - time block codes ( stbcs ) from cods have been widely studied for square designs , i.e. , @xmath22 since they correspond to minimum decoding delay codes for co - located multiple - antenna coherent communication systems .
however , non - square designs naturally appear and important in the following situations . 1 . in coherent co - located mimo systems , for a specified number of transmit antennas , non - square designs can give much higher rate than the square designs @xcite .
2 . in non - coherent mimo systems with non - differential detection , non - square designs with @xmath23 lead to low decoding complexity stbcs @xcite .
space - time - frequency codes can be viewed as non - square designs @xcite .
4 . in distributed space - time coding for relay channels , rectangular designs naturally appear @xcite .
rate @xmath3 non - square cods have been proposed for use in analog transmission with application to channel feedback @xcite .
the rate of the square cods falls exponentially with increase in the number of transmit antennas .
the following theorem relates the rate of a square od , real / complex , with the number of transmit antennas . for an integer @xmath24
where @xmath25 and @xmath26 and @xmath27 being integers with @xmath28 the hurwitz - radon number @xmath29 is defined as @xmath30 [ thm1 ] the maximal rate of a square rod for @xmath0 transmit antennas is given by @xmath31 where @xmath32 is the hurwitz - radon number of @xmath0 , while that of a square cod is given by @xmath33 where @xmath34 is the exponent of @xmath35 in the prime factorization of @xmath4 several authors have constructed square cods achieving maximal rate @xcite . in @xcite ,
the following induction method is used to construct square cods for @xmath36 antennas , @xmath37 , starting from @xmath38,~ g_a= \left[\begin{array}{rr } g_{a-1 } & -x_{a}^*i_{2^{a-1 } } \\ x_{a}i_{2^{a-1 } } & g_{a-1}^\mathcal{h } \end{array}\right],\ ] ] where @xmath39 is a @xmath40 complex matrix .
note that @xmath39 is a square cod in @xmath41 complex variables @xmath42 .
it is clear from theorem [ thm1 ] as well as the construction given by that the square ods , real and complex are not bandwidth efficient and naturally one is led to study non - square ods in order to obtain codes with higher rate .
it is known that @xcite there always exists a rate-1 rod for any number of transmit antennas .
in fact , the existence of rate-1 @xmath43 $ ] rod is equivalent to that of a @xmath44 $ ] square rod . for a rate-1 rod , the minimum value of decoding delay @xmath45 as a function of @xmath0 , denoted by @xmath46 , is given by @xmath47 it is known @xcite that the maximal rate of a non - square cod is equal to @xmath48 when the number of transmit antennas is @xmath49 or @xmath50 .
liang in @xcite has given an explicit construction of non - square cods achieving this maximal rate for any number of antennas .
there is also another construction of these codes given by lu et al @xcite .
the former construction is algorithmic in nature while the latter one is based on _ patch - up _ of several matrices .
the minimum decoding delay for the maximal rate non - square cods is , in general , not known .
the following theorem states what is known about the minimum delay of these code . for details ,
see @xcite .
[ delaycod ] a tight lower bound on the decoding delay of maximum rate non - square cods for @xmath0 antennas is @xmath51 for @xmath52 or @xmath53 .
moreover , if @xmath0 is congruent to @xmath54 or @xmath8 modulo @xmath6 , then this lower bound is achievable .
if @xmath0 is congruent to @xmath35 modulo @xmath6 , the minimum decoding delay is upper bounded by @xmath55 . as the rate of these maximal rate codes is close to @xmath10 for large number of antennas , it is sufficient to focus on rate @xmath10 codes when large number of antennas is under consideration . then , a natural problem to study is construction of rate @xmath10 non - square cods with the decoding delay as small as possible .
tarokh et al @xcite have given a class of rate @xmath10 codes obtained from rate-1 rods , which has lower delays than those of maximal - rate codes constructed in @xcite for number of transmit antennas more than 5 .
for example , for four transmit antennas , the rate @xmath10 code is @xmath56.\ ] ] notice that , contrary to the definition in @xcite of a cod , in this code , the variables appear in a column more than once and each entry of all the columns of the design matrix is scaled by @xmath57 in order to satisfy the condition @xmath58 of a cod .
we call such designs _ scaled _ cod which is not a cod in the conventional sense as in @xcite ( definition [ cod ] ) .
we define the class of scaled cods as follows : [ scod ] a _ @xmath59-scaled complex orthogonal design _ , for a positive integer @xmath60 ( @xmath59-scaled cod ) @xmath12 is a @xmath61 matrix in @xmath62 complex variables @xmath63 such that a non - zero entry of the matrix is a variable or its complex conjugate , or the negative of these and all the entries of any subset of columns of the matrix is scaled by @xmath64 satisfying the condition : @xmath65 the matrix @xmath12 is also said to be a @xmath19 $ ] @xmath66scaled cod .
notice that a @xmath59-scaled cod with with no column scaled by @xmath64 is a cod . in columns with scaling by @xmath64
all the variables appear exactly @xmath59 times . in this paper
we consider only the case @xmath67 and call these codes simply scaled - cod . _
contributions of this paper : _ the contributions of this paper may be summarized as follows : * for the rate @xmath10 scaled cods of @xcite , all the columns are scaled by the factor @xmath68 which led to the reduced delay compared to the codes of liang and a main result of this paper is that by having only a subset of the columns scaled by @xmath57 further reduction in delay by 50% is possible .
we use following notations to refer to the rate 1/2 cods given by tarokh et al @xcite , the maximal rate codes given in lu et al @xcite and the codes of this paper . *
* @xmath69 is the maximal rate cod for @xmath0 transmit antennas with the decoding delay as specified in the theorem [ delaycod ] .
* * @xmath70 is the rate @xmath10 scaled cods for @xmath0 transmit antennas constructed by tarokh et al @xcite .
* * @xmath71 is the rate @xmath10 scaled cods for @xmath0 transmit antennas constructed in this paper .
+ note that as @xmath0 increases , the maximal rate of @xmath69 approaches @xmath10 , thus two codes @xmath69 and @xmath70 can be compared for large value of @xmath72 based on their delays .
it is not difficult to see that the decoding delay of @xmath70 is less than that of @xmath69 for large @xmath0 .
we provide an explicit construction of rate @xmath10 scaled cods for any number of transmit antennas such that the decoding delay of these codes is @xmath46 when @xmath0 is the number of transmit antennas , whereas the delay for the codes @xmath70 is @xmath73 the table [ recparameters ] at the top of the next page shows that for large values of @xmath72 but for a marginal decrease in the rate with respect to @xmath74 the codes of this paper are the best codes . * for the case of 9 transmit antennas our rate-@xmath3 code is shown to be of minimal - delay .
+ c||rrrrrrrrrrrrrrrr @xmath0 & 5 & 6 & 7 & 8 & 9 & 10 & 11 & 12 & 13 & 14 & 15 & 16 + decoding delay of @xmath71 & 8 & 8 & 8 & 8 & 16 & 32 & 64 & 64 & 128 & 128 & 128 & 128 + decoding delay of @xmath70 & 16 & 16 & 16 & 16 & 32 & 64 & 128 & 128 & 256 & 256 & 256 & 256 + @xmath69 & 15 & 30 & 56 & 56 & 210 & 420 & 792 & 792 & 3003 & 6006 & 11440 & 11440 + @xmath75 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 + @xmath76 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 & 1/2 + @xmath77 & 2/3 & 2/3 & 5/8 & 5/8 & 3/5 & 3/5 & 7/12 & 7/12 & 4/7 & 4/7 & 9/16 & 9/16 + * as a byproduct of the above mentioned construction , a general construction of square real orthogonal designs ( rods ) is presented which includes as special cases well known constructions of adams , lax and phillips @xcite and geramita and pullman @xcite .
* even though scaling only a subset of columns allowed us to decrease the delay , it is shown that such a scaling limits the rate of the design strictly to @xmath78 in other words , the maximal rate of the scaled - cods is @xmath3 when scaling is present in atleast one column .
* zero entries in a design increase the peak - to - average power ratio ( papr ) in the transmitted signal and it is preferred not to have any zero entries in the design .
this problem has been addressed for square designs in @xcite
. all the known maximal rate non - square designs have zero entries .
our initial construction of rate-@xmath3 scaled cods have zero entries in the design matrix which will lead to higher peak - to - average power ratio ( papr ) in contrast to the designs @xmath79 however , we show that by post - multiplication of appropriate matrices , our construction leads to designs with no zero entries without change in the parameters of the design .
the remaining part of the paper is organized as follows : in section [ sec2 ] , we present the main result of the paper given by theorem [ rate12cod ] . before this , construction of a new set of maximal - rate square rods is given in subsection [ subsec2 - 1 ] .
in subsection [ subsec2 - 2 ] , construction of two new sets of rate-1 rods from the maximal - rate square rods of subsection [ subsec2 - 1 ] is presented and in subsection [ subsec2 - 3 ] , construction of the low - delay rate-1/2 scaled - cods is achieved using rate-1 rods of the previous subsection . in subsection
[ subsec2 - 4 ] , it is shown that the maximal rate for scaled - cods is @xmath78 for the special case of 9 transmit antennas , in section [ sec3 ] , it is shown that our construction is of minimal delay . in section [ sec4 ] , we show that the codes discussed so far can be made to have no zero entries in by appropriate preprocessing without affecting the parameters of the design . concluding remarks constitute section [ sec5 ] . in appendix
[ appendixii ] , it is shown that the well known constructions of square rods by adams - lax - phillips and geramita - pullman are special cases of our construction .
in this section , we construct a rate-1/2 scaled cods for any number of transmit antennas with @xmath5 reduction in decoding delay compared to the rate-1/2 codes constructed by tarokh et al @xcite .
the construction of these codes is done in the following three steps : + step 1 : construction of a new set of square rods ( subsection [ subsec2 - 1 ] ) .
+ step 2 : construction of two new sets of rate-1 rods from the square rods of step 1 ( subsection [ subsec2 - 2 ] ) .
+ step 3 : construction of low - delay rate-1/2 scaled cods using rate-1 rods ( subsection [ subsec2 - 3 ] ) . before explaining these steps , we first build up some preliminary results needed to describe these steps .
let @xmath80 be the finite field with two elements with addition and multiplication denoted by @xmath81 and @xmath82 for @xmath83 we consider logical operations also on the elements of this field : @xmath84 and @xmath85 represent respectively the logical disjunction ( or ) of @xmath86 and @xmath87 and complement or negation of @xmath86 , i.e. , @xmath88 for any finite subset @xmath89 of the set of natural numbers @xmath90 , let @xmath91 be the least integer such that @xmath92 for any @xmath93 often we identify each element of @xmath89 with an element of @xmath94 using the following correspondence : @xmath95 such that @xmath96 .
the all zero vector and all one vector in @xmath94 are denoted by @xmath97 and @xmath98 respectively . for @xmath99 , @xmath100 and @xmath101 represent respectively the @xmath102 complement of @xmath103 in @xmath94 , @xmath104 complement of @xmath103 in @xmath94 and hamming weight of @xmath103 . in other words
, @xmath105 and @xmath106 let @xmath107 and @xmath108 .
then , @xmath109 denote the component - wise modulo-2 addition and component - wise multiplication ( and operation ) of @xmath103 and @xmath110 respectively i.e. , @xmath111 and @xmath112 .
let @xmath113 .
for a set @xmath114 , we define @xmath115 , @xmath116 , @xmath117 for any @xmath118 and @xmath119 denotes the number of elements in the set @xmath120 . for any two sets @xmath121 with @xmath122 , the set @xmath123 , consists of those elements of @xmath124 , which are not in @xmath89 . for two integers @xmath125
, we use the notation @xmath126 , to indicate that @xmath127 for any matrix of size @xmath128 the rows and columns of the matrix are labeled by the elements of @xmath129 and @xmath130 respectively . if @xmath131 is a @xmath61 matrix in @xmath62 real variables @xmath132 , such that each non - zero entry of the matrix is @xmath133 or @xmath134 for some @xmath135 , obviously , it is not necessary that @xmath131 is a rod .
for example , @xmath136 is not a rod . in the following ( lemma [ proper ] )
, we derive a necessary and sufficient condition for the matrix @xmath131 to be a rod . + a submatrix @xmath137 of size @xmath138 constructed by choosing any two rows and any two columns of @xmath131
is called _ proper _ if * none of the entries of @xmath137 is zero and * it contains exactly two distinct variables .
consider the following matrix in three real variables @xmath139 and @xmath140 @xmath141 the sub - matrix @xmath142 is _ proper _ while @xmath143 is not . if @xmath144 , then we write @xmath145 whenever @xmath146 or @xmath147 ( in case of cods , @xmath145 if @xmath148 ) . if @xmath131 is a rod and if @xmath137 is a @xmath149 _ proper _ sub - matrix of @xmath131 , containing two variables , say @xmath150 and @xmath151 , @xmath152 positive integers , then @xmath153 . in other words , @xmath137 is a rod by itself in two variables .
the following lemma gives a characterization of rods in term of proper @xmath149 matrices .
[ proper ] let @xmath131 be a @xmath61 matrix in @xmath62 real variables @xmath132 , such that each non - zero entry of the matrix is @xmath133 or @xmath134 for some @xmath135
. then the following two statements are equivalent : + 1 ) @xmath131 is a rod .
+ 2 ) ( i ) each variable appears exactly once along each column of @xmath131 and atmost once along each row of @xmath131 , + ( ii ) if for some @xmath154 , @xmath144 and @xmath155 , then there exists @xmath156 such that @xmath157 and @xmath158 , + ( iii ) any proper @xmath149 sub - matrix of @xmath131 is a rod .
square rods have been constructed by several authors , for example , adams et al @xcite and geramita et al @xcite .
all these designs are constructed recursively and the basic blocks of these designs are the rods of order @xmath159 and @xmath160 .
it is known that these designs are obtained by left ( or right ) regular representations of the field of real numbers , the field of complex numbers , the quaternion algebra and the octonion algebra respectively . in this subsection ,
we take a different approach towards the construction of rods and that lead to a new class of rods constructable recursively of which the constructions of @xcite and @xcite are special cases .
if @xmath161 is a square real design of size @xmath162 $ ] in @xmath62 real variables @xmath21 , then whenever @xmath163 , we write @xmath164 @xmath161 is uniquely determined by @xmath165 and @xmath166 . +
the approach we take is identifying a pair of functions @xmath165 and @xmath166 that defines a square rod . towards that end , for any @xmath167 identifying @xmath168 with @xmath169 we have @xmath170 identified with a subset of @xmath171 .
we define two maps which are used for the construction of our square rods as follows .
+ let @xmath172 be an injective map defined on @xmath173 with the image denoted by @xmath174 and @xmath175 be another injective map defined on @xmath176 in the following theorem , we define the two maps @xmath165 and @xmath166 given in which define the matrix @xmath177 in terms of the two maps and and identify the conditions for the resulting @xmath161 to be a square rod .
[ ratesquarerod ] let @xmath178 and @xmath161 be a real design with @xmath179 be non - zero if and only if @xmath180 .
when @xmath163 , let @xmath181 if @xmath182 is odd , for all @xmath183 then , @xmath161 is a square rod of size @xmath184 $ ] .
we use the lemma [ proper ] to prove that @xmath161 is a rod .
first , for a fixed @xmath185 define @xmath186 it is clear that @xmath187 .
moreover , as @xmath188 is injective , we have @xmath189 whenever @xmath190 .
therefore , each column of the matrix @xmath161 contains all the variables @xmath191 and these variables appear exactly once .
similarly , it follows that the variables appear atmost once in any row of @xmath161 .
secondly , assume that @xmath163 and @xmath192 , then we show that there exists @xmath156 such that @xmath193 let @xmath194 .
we have @xmath195 therefore , @xmath196 .
similarly , @xmath197 .
thirdly , we show that any proper @xmath149 sub - matrix of @xmath161 is a rod .
it is enough to prove that @xmath198 , or equivalently , @xmath199 is an odd number . but
@xmath194 and @xmath200 where @xmath201 if @xmath202 is a multiple of @xmath35 .
we can write as @xmath203 which is an odd number as both @xmath204 and @xmath205 are the elements of @xmath206 .
thus , it is enough to construct @xmath188 and @xmath207 satisfying the property stated in theorem [ ratesquarerod ] , in order to construct a square rod .
this we achieve by making use of another set of two maps @xmath208 and @xmath209 as follows . on @xmath210
we define a map @xmath211 given by @xmath212 for @xmath213 , we have @xmath214 .
define @xmath215 where the @xmath35 s complement of an element @xmath103 of @xmath94 is performed in @xmath216 note that the map @xmath217 is well defined even though it is defined in terms of @xmath208 which is defined on @xmath218 [ propertyv8 ] let @xmath219 .
identify @xmath168 with @xmath94 .
then @xmath220 is an odd integer .
we prove it only for @xmath221 .
for all other values of @xmath34 , one can prove it similarly .
write @xmath222 .
let the radix-2 representation of @xmath223 be @xmath224 and that of @xmath225 be @xmath226 where each @xmath133 or @xmath227 takes value from the set @xmath80 .
the following table describes the map @xmath228 for @xmath229 . @xmath230 using this table , and we express @xmath227 in term of @xmath231 and @xmath232 for @xmath233 and @xmath35 as follows .
@xmath234 hence @xmath235 let @xmath236 .
now @xmath237 is odd if and only if @xmath238 but @xmath239 .
+ now @xmath240 is equal to @xmath241 if and only if + @xmath242 i.e. , @xmath243 for all @xmath244 , which implies that @xmath245 . + as @xmath246 , we have @xmath247 the second map @xmath209 is defined on the set @xmath248 given by @xmath249 as injective map @xmath250 given by @xmath251 [ propertyphi ] let @xmath248 be the set given in and @xmath252 .
then + ( i ) @xmath253 is odd for all @xmath254 .
+ ( ii ) @xmath255 is odd for all @xmath256 .
+ there are only finitely many possibilities for @xmath103 and @xmath110 and it can be checked that both the statements ( i ) and ( ii ) hold for all possible cases .
now , we define the maps @xmath188 and @xmath207 in terms of the maps @xmath208 and @xmath209 as follows : the map @xmath188 defined over @xmath173 is given by @xmath257 where @xmath258 for an element @xmath259 , either @xmath223 or @xmath260 for some @xmath261 and @xmath262 .
let @xmath263 be the map defined on the set @xmath206 given by @xmath264 the map @xmath207 is defined by @xmath265 the following theorem shows that the maps @xmath188 and @xmath207 defined by and satisfy the conditions of theorem [ ratesquarerod ] and hence define a rod .
[ orthog ] identify @xmath206 with a subset of @xmath171 , @xmath178 .
then @xmath266 is odd for all @xmath183 and hence from theorem [ ratesquarerod ] the matrix @xmath161 defined by @xmath188 and @xmath207 by and is a square rod . for @xmath267 and @xmath160 ,
the statement holds by lemma [ propertyv8 ] .
hence we assume that @xmath268 .
as @xmath269 , it is enough to prove that + ( i ) @xmath270 is odd for all @xmath271 .
+ ( ii)@xmath272 is odd for all @xmath273 . + to prove ( i ) , let @xmath274 . if @xmath275 , we have @xmath276 which is an odd number by lemma [ propertyv8 ] .
+ on the other hand , if @xmath277 , @xmath278 , then @xmath279 where the @xmath102 complement of an element is performed in @xmath94 .
we have @xmath280 where the @xmath102 complement of @xmath281 is performed in @xmath282 .
hence @xmath283 is odd by lemma [ propertyphi ] . in order to prove the part ( ii ) , we have following three cases : + ( i ) @xmath284 , + ( ii ) @xmath285 for some @xmath286 , @xmath287 , + ( iii ) @xmath288 for some @xmath289 .
+ in all the three cases , we have @xmath246 . by lemma [ propertyv8 ] , ( i )
+ for the second case , let @xmath290 .
we have @xmath291 . as @xmath292 ( the all zero vector in @xmath171 ) for @xmath287 , we have @xmath293 . but
@xmath294 is odd for all @xmath286 , hence @xmath283 is an odd number . for ( iii ) , let @xmath290 .
we have @xmath295 if @xmath296 , we have @xmath297 and @xmath298 .
thus @xmath299 is an odd number by lemma [ propertyphi ] .
if @xmath300 , it follows that @xmath301 which is an odd number by lemma [ propertyphi ] . from theorem [ ratesquarerod ]
it follows that the matrix @xmath161 defined by @xmath188 and @xmath207 by and is a square rod .
the square rods of theorem [ orthog ] will be denoted by @xmath302 throughout .
the rod @xmath303 of size @xmath304 $ ] is given by at the top of the next page .
as another example the rod @xmath305 of size @xmath306 $ ] is given by in the following page . in appendix
[ appendixi ] , it is shown that the rods @xmath302 can be constructed recursively .
@xmath307\end{aligned}\ ] ] ' '' '' ' '' '' one can define the functions @xmath188 and @xmath207 different from the one given above and can have a square rod different from @xmath302 . in appendix [ appendixii ] , we provide three different pairs of such functions and these are shown to give the well known adams - lax - phillips construction from octonions and quaternions and geramita and pullman s construction of square rods .
transition from a square rod to rate-1 rod is illustrated in @xcite using column vector representation of a rod . in a similar way
, we construct a rate-1 rod @xmath308 of size @xmath309 $ ] for @xmath0 transmit antennas from a rod of size @xmath310 $ ] where @xmath0 is any non - zero positive integer , not necessarily power of 2 .
any square rod of order @xmath46 obtained via a suitable pair of mapping @xmath311 and @xmath312 satisfying the conditions of theorem [ ratesquarerod ] ( for instance , @xmath313 obtained in the previous subsection or @xmath314 and @xmath315 obtained in appendix [ appendixii ] ) can be used for this purpose .
we refer to any such design by @xmath316 consisting of @xmath0 real variables @xmath317 .
let @xmath318 be @xmath46 real variables which constitute the matrix @xmath308 .
the matrix @xmath308 is obtained as follows : make @xmath319 if the @xmath244-th row of @xmath316 does not contain @xmath320 .
otherwise , @xmath321 or @xmath322 if @xmath323 or @xmath324 respectively .
the construction of the matrix @xmath308 ensures that it is a rate-1 rod . using , and theorem [ orthog ] ,
we have @xmath325 where @xmath326 for @xmath327 . + let @xmath328 be a matrix which is similar to the matrix @xmath308 , given by @xmath329 where @xmath330 @xmath328 is also a rate-1 rod .
both @xmath308 and @xmath328 are the new sets of rate-1 rods that are used in the following subsection to construct our codes . as examples , the rods @xmath331 and @xmath332 of size
@xmath333 $ ] obtained using @xmath303 are given by and respectively and the rods @xmath334 and @xmath335 of size @xmath336 $ ] obtained using @xmath305 are given by and respectively .
@xmath337\ ] ] @xmath338\ ] ] @xmath339\ ] ] in this subsection , we construct a rate-1/2 scaled cod with the help of rate-1 rods @xmath308 and @xmath328 constructed in the previous subsection .
let @xmath340 be complex variables .
the @xmath341 rate-@xmath3 cods @xmath342 and @xmath343 and the @xmath344 column vector @xmath345 in four variables , shown below , are the basic ingredients for our construction of rate-@xmath3 scaled cods .
@xmath346\ ] ] @xmath347\ ] ] @xmath348^\mathcal{t}.\ ] ] let @xmath244 be any non - negative integer . define the following matrices : @xmath349 one can easily verify that the matrices given by @xmath350,~~ \left[\begin{array}{cc } \overline{a}(1 ) & -\overline{a}(0)\\ \overline{a}(0 ) & \overline{a}(1 ) \end{array}\right],~~\ ] ] are scaled cods as the columns of the matrices are orthogonal to each other and the norm of each column is equal to square root of the sum of the norms of the variables of the design .
+ by relabeling the variables in the matrices @xmath351 and @xmath352 , it follows that @xmath353\ ] ] is a scaled cod whenever @xmath354 is odd and @xmath355,~~\ ] ] is a scaled cod for all values of @xmath244 and @xmath356 , @xmath357 .
let @xmath0 be an integer such that @xmath358 .
we construct a matrix @xmath71 of size @xmath359 as follows : let @xmath360 and @xmath361 and @xmath362 are the two rate-1 rods of size @xmath363 $ ] in @xmath364 real variables @xmath365 constructed in the previous subsection .
let @xmath366 and @xmath367 be the matrices formed by substituting @xmath368 with @xmath369 in the matrix @xmath361 and @xmath370 in the matrix @xmath362 respectively for @xmath371 to @xmath372 .
note that the size of both @xmath366 and @xmath367 is @xmath373 .
let @xmath374 .
let @xmath375 and @xmath376 , each of size @xmath377 , be defined as follows : @xmath378~~ o_8= \left [ \begin{array}{c } a(1 ) \\
a(3 ) \\ . \\ . \\ . \\
a(u-1 ) \end{array } \right].\end{aligned}\ ] ] define the matrix @xmath71 as @xmath379.\ ] ] note that the number of rows and columns of the matrix @xmath71 are @xmath380 and @xmath381 respectively .
the following theorem is the main result of this paper .
[ rate12cod ] let @xmath0 be a positive integer and @xmath71 be the matrix as defined in .
then @xmath71 is a rate-1/2 scaled cod of size @xmath382 $ ] . for @xmath383
, one can construct rate-1/2 cod of size @xmath382 $ ] from a cod of size @xmath384 $ ] given in .
we assume that @xmath358 .
let @xmath385 .
we have @xmath386.\ ] ] from the construction of @xmath375 and @xmath376 given in , we have + @xmath387 . from equation , we have @xmath388 thus it is enough to show that @xmath389 where @xmath390 is a matrix of size @xmath391 containing zero only .
let the @xmath356-th column of @xmath366 and @xmath367 be @xmath392 and @xmath393 respectively .
then we show that @xmath394 for all @xmath395 .
+ let @xmath396 . for convenience ,
we write @xmath397 for @xmath398 .
we have @xmath399,\\ \vspace*{.40 cm } o_8^h= \left [ \begin{array}{cccc } a^h(1 ) & a^h(3)&\cdots & a^h(u-1 ) \end{array } \right],\\ \vspace*{.40 cm } h_t(j)= \left [ \begin{array}{c } s(0,j)\overline{a}({2(0\oplus \gamma(j))+1 } ) \\ s(1,j)\overline{a}({2(1\oplus \gamma(j))+1 } ) \\ . \\ .
\\ s(i , j)\overline{a}({2(i\oplus \gamma(j))+1 } ) \\ . \\ .
\\ s(\frac{u}{2}-1,j)\overline{a}({2\bigl((\frac{u}{2}-1)\oplus \gamma(j)\bigr)+1 } ) \end{array } \right ] , \\ \vspace*{.40 cm } \hat{h}_t(j)= \left [ \begin{array}{c } \hat{s}(0,j)\overline{a}({2(0\oplus \gamma(j ) ) } ) \\
\hat{s}(1,j)\overline{a}({2(1\oplus \gamma(j ) ) } ) \\ . \\ .
\\ \hat{s}(i , j)\overline{a}({2(i\oplus \gamma(j ) ) } ) \\ . \\ . \\ \hat{s}(\frac{u}{2}-1,j)\overline{a}({2((\frac{u}{2}-1)\oplus \gamma(j ) ) } ) \end{array } \right ] , \end{array}\end{aligned}\ ] ] where @xmath400 and @xmath401 are defined in and respectively .
we have @xmath402 now @xmath403 hence @xmath404 we illustrate our main result in the following example . for @xmath405 transmit antennas ,
we have rate-@xmath10 scaled cod of size @xmath406 $ ] given by @xmath407.\ ] ] while the known rate @xmath10 scaled cod for @xmath405 transmit antenna is given by @xcite @xmath408.\ ] ] where the decoding delay is @xmath409 . for @xmath410 transmit antennas ,
the rate-1/2 scaled cod given by tarokh et al @xcite of size @xmath411 $ ] is given in appendix [ appendixvi ] , while the new rate-@xmath10 code of size @xmath412 $ ] is given by .
@xmath413.\ ] ] it has been shown by liang @xcite that the maximal rate of a cod for @xmath0 transmit antennas is @xmath414 when @xmath415 or @xmath416 .
the following result says when there is scaling of atleast one column then the maximal rate is @xmath78 the maximal rate of a scaled cod , with scaling of at least one column , for @xmath0 transmit antennas is @xmath78 let @xmath71 be a scaled cod for @xmath0 transmit antennas and there exists at - least one column of the matrix such that whenever a variable appears in that column , it is scaled by @xmath417 since all the variables appearing in a column is scaled by @xmath57 , each variable must appear twice in that column .
let the number of distinct complex variables in @xmath71 is @xmath62 .
then @xmath418 , i.e. , @xmath419 .
thus if one allows to incorporate a factor of @xmath57 for all the entries of any column , there wo nt be any improvement in the rate , on the other hand , as we will have observed , we can construct some rate 1/2 codes with lesser decoding delay .
in this section , it is shown that the low - delay rate @xmath10 cod developed in the previous section is of minimal delay for 9 transmit antennas . to prove this , we need some preliminary facts regarding the interrelationship between real and complex ods and certain bilinear maps .
the rod and bilinear maps are intimately related in the sense that a rod of size @xmath19 $ ] exists if and only if there exists a normed bilinear map of the same size .
the normed bilinear maps have been studied extensively and one find a good introduction to this topic in the book by shapiro @xcite .
as the results from the theory of normed bilinear maps is used to prove our claim , these maps have been defined below and some facts are stated regarding these maps . a bilinear map @xmath420 ( over a field @xmath421 ) is a map @xmath422 such that it is linear in both @xmath103 and @xmath110 , i.e. , @xmath423 and @xmath424 for all @xmath425 and @xmath426 the space @xmath427 is called the target space of @xmath420 . if the vector spaces under consideration are inner product spaces , for example , when the field is real numbers or complex numbers , the euclidean norm of a vector @xmath103 is denoted by @xmath428 . if a bilinear map preserves the norm , then it is called a normed bilinear map .
more precisely , a _ normed real bilinear map _ ( nrbm ) of size @xmath19 $ ] is a map @xmath429 such that @xmath420 is bilinear and normed i.e. , @xmath430 .
if @xmath431 implies @xmath432 or @xmath433 , then such a map is called a nonsingular map .
the following theorem gives a lower bound on @xmath45 for fixed values of @xmath0 and @xmath62 .
if there exists a nonsingular bilinear map of size @xmath19 $ ] over @xmath434 , then @xmath435/(x^n , y^k)$ ] .
let @xmath436 be positive integers .
then the three quantities @xmath437 @xmath438 and @xmath439 are defined by * @xmath440/(x^n , y^k)\}$ ] , * @xmath441 there is a nonsingular bilinear map @xmath442 $ ] over @xmath443 } , * @xmath444 there is a normed bilinear map @xmath442 $ ] over @xmath445 } , the following basic facts about these quantities are well known @xcite .
+ @xmath446 and @xmath447 if and only if @xmath448 where @xmath449 is the hurwitz - radon function .
it follows from the definition of @xmath450 that [ ncirck ] @xmath450 is a commutative binary operation .
+ @xmath451 if @xmath452 then @xmath453 + @xmath454 @xmath455 if and only if @xmath456 and @xmath457 .
+ @xmath458 if @xmath459 then @xmath460 let us compute @xmath461 .
we observe that @xmath462 , but @xmath463 .
so , @xmath464 . the relation between rods and nrbms has been observed by wang and xia in @xcite .
for the sake of completeness and since the proof of this fact gives the explicit relation between the nrbm and the row - vector representation matrices of the rod which is in correspondence with it , we give here the proof of the following lemma .
[ rod_nrbl ] a rod of size @xmath19 $ ] exists if and only if there exists a normed real bilinear map of size @xmath19 $ ] .
let @xmath465 @xmath466 and @xmath467 be real column vectors .
let @xmath124 be a rod of size @xmath19 $ ] given by @xmath468 where the @xmath61 real matrices are the dispersion matrices or weight matrices defining the design @xmath124 @xcite .
let @xmath469 we show that @xmath420 is a normed real bilinear map of size @xmath19 $ ] .
let @xmath470 .
then @xmath471 where the @xmath472 real matrices @xmath473 are the row vector representation @xcite of the design @xmath474 this representation shows that @xmath420 is bilinear .
moreover , @xmath420 is normed , since @xmath475= @xmath476 . to show that the converse holds , let @xmath420 be the normed bilinear map given by @xmath477 as @xmath420 is linear in both @xmath478 and @xmath479 , we have @xmath480 where @xmath124 is an @xmath61 matrix where each entry of the matrix is a real linear combination of the variables @xmath481 .
as @xmath420 is normed , we have @xmath482 . but @xmath483 .
then , @xmath484 .
so , we have @xmath485 as @xmath479 consists of variables , the above equation is equivalent to @xmath486 .
we now prove the main result of this section .
the minimum decoding delay of the rate-@xmath10 cod admitting linear combination of complex variables for @xmath405 transmit antennas is @xmath487 . in this proof , we assume that the rod and cod admit linear combination of two or more variables as its entries .
suppose , there exists a cod of size @xmath488 $ ] where @xmath103 is an integer less than @xmath160 .
this implies existence of a rod of size @xmath489 , x < 8 $ ] which is obtained by replacing each complex entry by its @xmath149 real matrix representation . in the remaining part of the proof ,
we show that such a rod does not exist thus proving the theorem .
+ if a rod of size @xmath489 $ ] exists , then there also exists a normed bilinear map of the same size by lemma [ rod_nrbl ] which implies that @xmath490 . as @xmath491 , we have @xmath492 i.e. , @xmath493 for @xmath103 being an integer .
thus we have three possible choices for @xmath103 , namely @xmath494 and @xmath495 .
+ by proposition [ ncirck ] , we have @xmath496 for @xmath497 and @xmath495 respectively . in all the cases ,
@xmath498 which contradicts the fact that @xmath490 .
in this section , we study the papr properties of scaled cods constructed in this paper .
note that in the construction of @xmath70 given in @xcite , even though the delay is more , there are no zero entries in the design matrix . on the contrary , in our construction of low - day codes
@xmath71 there are zero entries . to be specific ,
observe that the first eight columns of rate-@xmath10 code @xmath499 given in contains as many zero as the number of non - zero entries in it , while there is no zero in the remaining columns of the matrix .
when the number of transmit antennas @xmath0 is more than @xmath495 , the total number of zeros in the codeword matrix is equal to @xmath500 .
hence the fraction of zeros in the codeword matrix is equal to @xmath501 for @xmath502 .
now in the remaining part of this section , we show that one can further reduce the number of zeros in @xmath71 by suitably choosing a post - multiplication matrix to it without increasing signaling complexity @xcite of the code .
as seen easily , only the first eight column contain zeros while the others do not .
moreover , the zeros in @xmath241-th column and the @xmath503th column occupy complementary locations , so is also for the pairs of columns given by @xmath504 and @xmath505 .
what it essentially suggests is that we can perform some elementary column operations which will result in a code in which all the entries are non - zero . in other words ,
if the rate-1/2 cod is of size @xmath61 , then we post - multiply it with a matrix @xmath506 of size @xmath18 given by @xmath507\end{aligned}\ ] ] where @xmath124 is a matrix of size @xmath508 given by @xmath509\end{aligned}\ ] ] and @xmath510 is the @xmath511 identity matrix , then all the entries of the scaled cod given by @xmath512 are non - zero .
we formally present this fact as : the matrix @xmath506 when post - multiplied with a rate-1/2 scaled cod @xmath71 given by @xmath513.\ ] ] always gives a cod with no zeros .
moreover , the matrix @xmath506 does not depend on any particular construction procedure ( namely the maps @xmath188 and @xmath207 ) used to obtain the constituent rate-1 rods .
it is clear that the first @xmath160 columns of the matrix has @xmath5 zeros in it and in the remaining @xmath514 columns formed by @xmath366 and @xmath367 , there are no zeros as both these matrices are constructed from rate-1 rod by substituting all the variables in it with appropriate @xmath160-tuple column vectors .
here neither rate-1 rod nor the @xmath160-tuple column vector has any any zero in it .
therefore , the matrix @xmath506 gives a rate @xmath10 scales cod without any zeros irrespective of how the rate-1 rods are obtained for the construction of @xmath71 .
the rate-1/2 code with no zero entry for @xmath405 transmit antennas is given by @xmath515\ ] ]
this paper gives rate-@xmath10 cods for @xmath0 transmit antennas with decoding delay equal to @xmath46 .
the decoding delay of these codes is half of that of the rate-1/2 cods given in tarokh et al @xcite .
as the maximal rate of a scaled cod is close to @xmath10 for large number of transmit antennas , the codes constructed in this paper are better than the codes constructed by liang @xcite or lu et al @xcite when considered for large number of transmit antennas .
another advantage with the designs reported in this paper is that they do not contain zero entries leading to low papr .
all the four constructions namely adams , lax and phillips construction from quaternions , octonion , geramita - pullman construction and the construction given in this paper will give the same square rod if number of transmit antennas is less than or equal to @xmath160 .
therefore , these four construction will generate the same rate @xmath10 scaled cod if the number of transmit antennas ( of the scaled cod ) is less than or equal to @xmath487 . for more than @xmath487 antennas , rate-1/2 scaled cods will vary with the methods chosen for the construction of rate-1 rods .
due to the largeness of the matrices involved , it is not possible to display two distinct rate-1/2 scaled cods for @xmath516 transmit antennas , obtained by two different construction procedures for rate-1 rods .
it is not known whether the low - delay for rate @xmath10 scaled cods we have achieved is minimal delay except for the case of @xmath405 transmit antennas .
we conjecture that @xmath46 is the minimum value of the decoding delay of rate-1/2 scaled cods for any @xmath0 transmit antennas
. it will be interesting to see whether this is indeed true .
an interesting direction for further research would be to investigate whether the necessary conditions given in theorem [ ratesquarerod ] on the maps @xmath188 and @xmath207 are indeed sufficient also . 160 x. b. liang `` orthogonal designs with maximal rates , '' _
ieee trans.inform .
vol.49 , pp . no . 10 ,
2468 - 2503 , oct .
2003 vahid tarokh and il - min kim , `` existence and construction of noncoherent unitary space - time codes , '' _ ieee trans . on inform .
3112 - 3117 , dec . 2002
. r. v. j. r. doddi , v. shashidhar , md .
zafar ali khan and b. sundar rajan , `` low - complexity , full - diversity , space - time frequency block codes for mimo - ofdm , '' _ proc .
ieee globecom 2004 _ , dallas , texas , nov.29-dec.3 , pp .
204 - 208 , 2004 .
yindi jing and b. hassibi , `` distributed space - time coding in wireless relay networks , '' _ ieee trans . on wireless communications _ , vol .
3524 - 3536 , dec . 2006 .
jinhui chen and dirk t. m. slock , `` orthogonal space - time block codes for analog channel feedback , '' proceedings of ieee international symposium on information theory , ( isit 2008 ) , toronto , canada , july 6 - 11 , 2008 , pp.1473 - 1477 .
o. tirkkonen and a. hottinen , `` square matrix embeddable stbc for complex signal constellations space - time block codes from orthogonal design , '' _ ieee trans .
vol 48,no .
384 - 395 , feb . 2002 .
v. tarokh , h. jafarkhani , and a. r. calderbank , `` space - time block codes from orthogonal designs , '' _ ieee trans .
theory , _ vol .
1456 - 1467 , july 1999 .
l. c. tran , t. a. wysocki , a. mertins and j. seberry , complex orthogonal space - time processing in wireless communications , springer - verlag , 2006 .
j. f. adams , p. d. lax , and r. s. phillips , `` on matrices whose real linear combinations are nonsingular , '' _ proc .
, _ vol . 16 , 1965 , pp .
318 - 322 .
kejie lu , shengli fu and xiang - g xia , `` closed - form designs of complex orthogonal space - time block codes of rates @xmath517 for @xmath518 or @xmath519 transmit antennas , '' _ ieee trans .
theory , _ vol .
51 , no.5 , pp .
4340 - 4347 , dec 2005 .
sarah spence adams , nathaniel karst and jonathan pollak , `` the minimum decoding delay of maximal rate complex orthogonal space - time block codes , '' _ ieee trans .
theory , _ vol .
53 , no.8 , pp . 2677 - 2684 , aug 2007 .
a. v. geramita and n. j. pullman , `` a theorem of hurwitz and radon and orthogonal design , '' _ proc .
soc . , _ vol .
42 , no . 1 , 1974 , pp .
51 - 56 .
smarajit das and b. sundar rajan , `` square complex orthogonal designs with low papr , '' proceedings of ieee international symposium on information theory , ( isit 2007 ) , nice , france , june 24 - 29 , 2007 , pp .
2626 - 2630 .
smarajit das and b. sundar rajan , `` square complex orthogonal designs with low papr and signaling complexity , '' to appear in ieee transactions on wireless communications .
smarajit das and b. sundar rajan , `` low - delay , high - rate , non - square stbcs from scaled complex orthogonal designs , '' proceedings of ieee international symposium on information theory , ( isit 2008 ) , toronto , canada , july 6 - 11 , 2008 , pp .
1483 - 1487 . c. yuen , y. l. guan and t. t. tjhung , `` power - balanced orthogonal space - time block code , '' accepted for publication in ieee trans .
vehicular technology .
haiquan wang and xiang - gen xia , `` upper bounds of rates of complex orthogonal space - time block codes , '' ieee trans .
theory , 49 ( 2003 ) , pp .
2788 - 2796 .
d. b. shapiro , compositions of quadratic forms , berlin , germany : walter de gruyter , 2000 .
in this appendix we show that the rods @xmath302 can be constructed recursively .
let @xmath520 for @xmath267 and @xmath160 .
the four square ods @xmath521 are shown below . @xmath522
it follows that @xmath523 for @xmath267 or @xmath160 .
the expression for @xmath302 of order @xmath524 as given in theorem [ orthog ] gives rise to the following recursive construction of @xmath302 .
@xmath525 , \quad r_{4n}= \left[\begin{array}{cc } r_{2n } & x_{\rho(n)+1}i_{2n}\\ -x_{\rho(n)+1}i_{2n } & r_{2n}^\mathcal{t } \end{array}\right],\ ] ] @xmath526 , \quad r_{16n}= \left[\begin{array}{cc } r_{8n } & t_8(y_2,y_3,y_4,y_5)\otimes i_n \\ t_8(-y_2,y_3,y_4,y_5)\otimes
i_n & r_{8n}^\mathcal{t } \end{array}\right]\\\ ] ] ' '' '' given two matrices @xmath527 of size @xmath528 and @xmath529 of size @xmath530 , we define the _ kronecker product or tensor product _ of @xmath531 and @xmath529 as the following @xmath532 matrix : @xmath533 let @xmath17 be an identity matrix of size @xmath0 .
define @xmath534 let @xmath535 be real variables .
define @xmath536 we have four rods of order @xmath537 with @xmath538 as given in which are respectively @xmath539 and @xmath540 . + assuming that a square rod of order @xmath541 @xmath542 which has @xmath32 real variables , is given , then we construct @xmath543 of order @xmath544 , @xmath545 , @xmath546 and @xmath547 respectively given by , as shown at the top of the next page where @xmath548 and @xmath549
in this appendix we show that the well known constructions of square rods by adams - lax - phillips using octonions and quaternions as well as the construction by geramita and pullman are nothing but our construction corresponding to specific choices of the functions @xmath188 and @xmath207 defined by and .
it turns out to be convenient to use the map @xmath550 instead of the map @xmath551 note that both @xmath188 and @xmath552 act on the set @xmath173 and are injective .
now given @xmath188 and @xmath552 , we have @xmath553 . with this new definition
, we can reformulate the criteria given in theorem [ orthog ] as follows .
@xmath554 in the following lemma , we define @xmath188 and @xmath552 in three different ways and these maps are shown to satisfy the relation given in .
although both @xmath188 and @xmath552 are different for all the three cases for arbitrary values of @xmath524 , @xmath188 is the identity map when @xmath267 or @xmath160 .
hence @xmath555 if @xmath556 and is given by .
[ threedesign ] let @xmath178 , @xmath25 , @xmath557 .
let @xmath188 and @xmath552 be two maps defined over @xmath173 in three different ways as given below . identify @xmath558 and @xmath559 as subsets of @xmath94 .
then @xmath560 is odd for all @xmath561 , @xmath562 .
+ for @xmath563 \(i ) @xmath564 \(ii ) @xmath565 where @xmath566 ( iii ) @xmath567 we give proof only for the case ( i ) . the cases ( ii ) and ( iii ) can be proved similarly .
it is enough to prove that + ( b1 ) @xmath568 is odd for all @xmath254 , @xmath569 and + ( b2 ) @xmath570 is odd for all @xmath561 , @xmath571 .
let @xmath572 such that @xmath573 and @xmath574 .
+ similarly , let @xmath575 such that @xmath576 let @xmath577 and @xmath578 . from the definition of @xmath579
, it follows that + @xmath580 first we prove ( b1 ) .
let @xmath581 with @xmath582 .
we have @xmath583 but @xmath584 is an odd number by lemma [ propertyv8 ] .
+ if @xmath585 , we have @xmath586 by ( a4 ) .
+ to prove ( b2 ) , let @xmath587 and @xmath588 .
write @xmath589 , @xmath590 with @xmath591 .
we have two cases : + ( c1 ) : @xmath592 , ( c2 ) : @xmath593 , @xmath594 . + * case ( c1 ) : * we have @xmath595 but @xmath596 by ( a3 ) + and @xmath597 by ( a1 ) , + thus @xmath598 .
+ now @xmath599 by ( a5 ) . + but @xmath600 by ( a1 ) and @xmath601 by ( a2 ) .
+ hence @xmath602 is an odd number .
* case ( c2 ) : * we consider two following cases : + ( i ) @xmath603 and ( ii ) @xmath604 .
note that @xmath605 is always non - zero .
+ let @xmath606 .
+ * case ( i ) : * we have @xmath607 which is an odd number by lemma [ propertyv8 ] . + * case ( ii ) : * since @xmath604 , therefore @xmath608 .
we have @xmath609 @xmath610 ' '' '' by lemma [ threedesign ] and theorem [ ratesquarerod ] , the matrix @xmath161 defined by the two functions @xmath188 and @xmath552 is a square rod in all the three cases .
we refer to these three different rods by @xmath611 and @xmath612 corresponding to the pair of functions defined in ( i ) , ( ii ) and ( iii ) respectively .
observe that our construction @xmath302 is different from any of @xmath611 and @xmath612 for general values of @xmath613 now , we proceed to show that the designs @xmath611 and @xmath612 are essentially the adams - lax - phillips construction using octonions and quaternions and the geramita - pullman construction respectively with change in sign of some rows or columns .
the adams - lax - phillips construction from octonions is given by induction from order @xmath537 to @xmath547 as follows @xcite : denoting the square rod of order @xmath537 resulting from the adams - lax - phillips construction using octonions by @xmath614 which has @xmath32 real variables , the square rod of order @xmath547 with @xmath615 real variables @xmath616 @xmath617 @xmath618 is given by @xmath619 \end{aligned}\ ] ] with @xmath620 .
+ with re - arrangement of variables and change in signs , we rewrite the design @xmath621 as @xmath622 \end{aligned}\ ] ] with @xmath623 and @xmath624 the reason why we consider this rearranged version is that we show in lemma [ ato2 ] that @xmath625 is same as @xmath626 with @xmath627 + [ ato2 ] let @xmath628 and a power of @xmath629 also , let @xmath625 be the square rod of order @xmath524 as given in lemma [ threedesign ] ( i ) , and @xmath630 be the square rod given in which is of order @xmath547 . then @xmath631 for @xmath632 .
we prove it by induction on @xmath524 .
for @xmath267 and @xmath160 , @xmath633 and the cod @xmath634 of order @xmath524 is also given by @xmath635 .
hence the lemma holds for @xmath267 and @xmath160 . assuming that the lemma holds for @xmath636 , i.e. , @xmath637 of order @xmath0 , we have to prove that the lemma also holds for @xmath638 i.e.
, @xmath639 let @xmath640\ ] ] where @xmath641 , @xmath642 are square matrices of size @xmath643 .
it is easy to check that the location of non - zero variables in the matrix @xmath644 coincide with that of @xmath630 .
therefore it is enough to show the signs ( positive / negative polarity ) of the corresponding entries in the two designs are same i.e. , 1 .
@xmath645 for @xmath646 , + 2 .
@xmath647 for @xmath648 , + 3 .
@xmath649 + if @xmath650 , + 4 .
@xmath651 for @xmath652 , + 5 .
@xmath653 if @xmath654 or @xmath655 , + 6 .
@xmath656 + if @xmath657 .
+ note that + 1 ) & 2 ) together imply @xmath658 , + 3 ) & 4 ) together imply @xmath659 and + 5 ) & 6 ) together imply @xmath660 .
+ let @xmath661 .
+ then @xmath662 and @xmath663
. + to prove 1 ) , we have to show that @xmath664 for @xmath646 .
+ we have @xmath665 and @xmath666 .
so @xmath667 and @xmath668 .
i.e. , @xmath669 .
+ thus it is enough to prove that @xmath670 now @xmath671 , @xmath160 divides @xmath672 and @xmath673 , hence the statement holds .
the statement 2 ) is true as @xmath674 for @xmath648 .
+ in order to prove 3 ) , we must have @xmath675 i.e. , @xmath676 . as @xmath677
, we have @xmath665 with @xmath678 .
so @xmath160 divides @xmath679 as @xmath160 divides both @xmath680 and @xmath681 .
so @xmath682 i.e. , @xmath683 .
thus it is enough to prove that @xmath684 .
it is indeed true as @xmath685 is a multiple of @xmath160 . to prove 4 ) , we have to show that @xmath686 we have @xmath687 for some @xmath688 with @xmath678 and @xmath689 .
let @xmath690 and @xmath25 .
+ if @xmath691 , we have @xmath692 and @xmath693 .
one can easily see that the above statement holds .
+ on the other hand , if @xmath694 , we have @xmath695 and @xmath696 . in this case too , the statement holds . to prove 5 ) , we have to show that @xmath697 i.e. , @xmath698 . now for @xmath654 or greater than @xmath546 , @xmath699 . to prove 6 )
, we have to show that @xmath700 i.e. , @xmath701 . but @xmath702 for all @xmath703 .
adams - lax - phillips has also provided another construction of square rods using quaternions which is explicitly shown in @xcite . assuming that a square rod of order @xmath537 @xmath704 which has @xmath32 real variables , is given , then a square rod of order @xmath547 with @xmath705 real variables @xmath133 for @xmath706 @xmath707 is given by , as shown at the top of this page where the two matrices @xmath708 and @xmath709 are given by @xmath710 @xmath711 respectively with @xmath623 .
+ the geramita - pullman construction of square rod is also given by induction explicitly in @xcite .
consider a recursive construction of square rod of order @xmath537 to @xmath547 as follows : @xmath712 which has @xmath32 real variables is given , then a square rod @xmath713 of order @xmath547 with @xmath705 real variables @xmath133 for @xmath706 is given by @xmath714 \end{aligned}\ ] ] with @xmath623 .
+ it can be checked that both adams - lax - phillips construction from quaternions and geramita - pullman s construction given in @xcite differ from the constructions of @xmath715 and @xmath716 defined above only in rearrangement of variables and in signs of some of the rows or columns of the design matrix .
[ atqgp ] let @xmath628 and @xmath717 and @xmath612 be the square rod of order @xmath524 as given in lemma [ threedesign ] ( ii ) and ( iii ) , and also let @xmath715 and @xmath716 be the square rods given in and in which are of order @xmath547
. then @xmath718 and @xmath719 for @xmath627 similar to that of lemma [ ato2 ] and hence omitted .
[ adpgpexamples ] square rod @xmath720 of size @xmath304 $ ] by adams - lax - phillips construction from octonion is given by .
@xmath721\end{aligned}\ ] ] ' '' '' square rod @xmath722 of size @xmath304 $ ] by adams - lax - phillips construction from quaternion is given by .
@xmath723\end{aligned}\ ] ] ' '' '' square rod of @xmath724 size @xmath304 $ ] by geramita - pullman construction is the same as @xmath725 square rod of @xmath726 size @xmath306 $ ] by geramita - pullman construction is given by . ' '' ''
@xmath727.\ ] ]
|
the maximal rate for non - square complex orthogonal designs ( cods ) with @xmath0 transmit antennas is @xmath1 if @xmath0 is even and @xmath2 if @xmath0 is odd , which are close to @xmath3 for large values of @xmath4 a class of maximal rate non - square cods have been constructed by liang ( ieee trans .
inform .
theory , 2003 ) and lu et .
al .
( ieee trans .
inform .
theory , 2005 ) have shown that the decoding delay of the codes given by liang , can be reduced by @xmath5 when number of transmit antennas is a multiple of @xmath6 .
adams et .
al .
( ieee trans .
inform .
theory , 2007 ) have shown that the designs of liang are of minimal - delay for @xmath0 equal to @xmath7 and @xmath8 modulo 4 and that of lu et.al .
are of minimal delay when @xmath0 is a multiple of @xmath9 however , these minimal delays are large compared to the delays of the rate @xmath10 non - square cods constructed by tarokh et al ( ieee trans .
inform .
theory , 1999 ) from rate-@xmath7 real orthogonal designs ( rods ) . in this paper
, we construct a class of rate-@xmath10 non - square cods for any @xmath0 with the decoding delay equal to @xmath5 of that of the delay of the rate-1/2 codes given by tarokh et al .
this is achieved by giving first a general construction of rate-1 square real orthogonal designs ( rods ) which includes as special cases the well known constructions of adams , lax and phillips and geramita and pullman , and then making use of it to obtain the desired rate-@xmath3 non - square cod . for the case of 9 transmit antennas ,
our rate-@xmath3 cod is shown to be of minimal - delay .
the proposed construction results in designs with zero entries which may have high peak - to - average power ratio ( papr ) and it is shown that by appropriate postmultiplication , a design with no zero entries can be obtained with no change in the code parameters .
|
the detection of planets orbiting other stars has revealed a great diversity of both planetary mass and location , including several populations which have no analogue in our own solar system .
of particular interest is the discovery of substantial numbers of sub - jovian planets with orbital periods shorter than that of mercury , using both the radial velocity and transit techniques ( howard et al .
2010 ; mayor et al . 2011 ; borucki et al .
2011 ; batalha et al .
2013 ) .
this population proved to be a surprise for models of planetary systems whose short period populations are generated by migration inwards from larger radii ( e.g. ida & lin 2008 ) .
however , the properties of the observed systems do match the expectations of simple in situ assembly models ( hansen & murray 2012 , 2013 ; chiang & laughlin 2013 ) , although the amount of mass in rocky material required for such models is sometimes larger than what one might anticipate from a simple minimum - mass solar nebula model .
hansen & murray ( 2012 ) suggested that such conditions could be realised if solid material is concentrated radially in the inner parts of the gas disc prior to the late - stage assembly into solid planets .
this is not an outrageous expectation as it is well known that small bodies in gas discs are potentially subject to dynamically important aerodynamic drag forces ( whipple 1972 ; weidenschilling 1977a ; takeuchi & lin 2002 ; bai & stone 2010 ) .
however , the particular details of how such a model might set the stage for planet formation are still unclear . in this paper
we attempt to outline a simple model that provides such a framework .
several authors have discussed the possibility that planet formation may be seeded by special locations in the disc which allow radially migrating material to collect and assemble ( haghighipour & boss 2003 ; masset et al .
2006 ; hasegawa & pudritz 2011 ; kretke & lin 2012 ) .
recently , two groups of authors ( boley & ford 2013 ; chatterjee & tan 2013 ) have proposed models that use this idea as the basis for a scenario for inside - out planet formation , in which solid material migrates inward due to drag forces until it it stalls due to a pressure maximum , either at the inner edge of the dead zone or near an evaporation or sublimation front .
the resulting accumulation of material causes planetesimals to form and to extend the truncation of the disc outwards , resulting in what these authors call ` systems with tightly - packed inner planets ' , or ` stips ' .
these proposals may indeed provide compact planetary systems , but they may , in fact , produce systems that are too radially concentrated . the distribution of known extrasolar planets ( figure [ am ] ) shows two distinct length scales for giant planets ( an overdensity near 0.05 au and an increase in frequency @xmath3 au ) but no corresponding special location for planets with mass @xmath4 , as might be expected if the pile - up is generated by a feature in the underlying disc structure , such the inner edge of a dead zone .
indeed , attempts to characterise these planetary systems consistently find relatively smooth distributions .
chiang & laughlin ( 2013 ) construct a ` minimum mass kepler nebula ' by considering the ensemble properties of the kepler planet candidate sample , and derive an empirical distribution consistent with an power - law distribution of surface density , @xmath5 .
hansen & murray ( 2012 , 2013 ) recover a very similar underlying model in matching their assembly simulations for these planets to the observations , and found that initial density profiles that were too steep produced planetary systems that were too compact to match the observations .
the inside - out packing of planets in the stips models produce a spacing of planets that is dictated by the considerations of dynamical stability of interacting neighbouring planets ( chatterjee & tan 2013 ) .
monte carlo models of the in situ assembly of the observed planetary systems ( hansen & murray 2013 ) are able to match the observed distributions of periods and period ratios of the observed kepler systems . however , the resulting distribution in planetary spacings peaks at factors of 23 times larger than the threshold for dynamical stability , suggesting that the final configuration is not as tightly packed as it would be if it were assembled sequentially from the inside out . .
[ am],width=317 ] in this paper , we consider an alternative notion namely that the solid material is not trapped , but rather migrates rapidly to the inner edge of the protoplanetary disc , from whence it is returned to the outer disc , to begin the inspiral anew .
we then examine the consequences of this hypothesis and the manner in which the resulting steady - state distribution of solids can form the initial conditions for the later stages of planetary formation .
such a model finds support within our own solar system via the evidence for widespread processing at high temperatures in chondrites ( e.g. scott & krot 2005 ) and in the observation that the chemical constituents of the comet wild-2 show evidence of substantial chemical processing in the hotter inner regions of the protosolar nebula ( brownlee et al . 2006 ; mckeegan et al . 2006 ; brownlee , joswiak & matrajt 2012 ) .
indeed , such considerations have long spurred suggestions that material was recirculated in the protosolar nebula ( liffman & brown 1995 ; shu , shang & lee 1996 ; shu et al .
2001 ; hu 2010 ; salmeron & ireland 2012 ) .
these models propose that solid material in the inner parts of the protoplanetary nebula are levitated and carried outwards by either the magnetic funnel of the stellar wind or by a magnetically driven outflow from the disk , ultimately to be returned to the protoplanetary nebula on larger scales .
protostellar outflows are a common feature of young stars , and the entrainment of solid material into such ` winds ' offers a mechanism to redistribute material from the locations near the star to more distant locations .
the most prominent of these models is that of shu et al . , which proposes that the chemical processing observed in the constituents of asteroids ( at @xmath6 au from the sun ) may have , in large part , occurred at distances of a few stellar radii , with the material subsequently swept up into the outflow until the material decoupled from the wind and fell back to the disc .
one of the principal motivations of this model was to describe the heating and chemical alteration of metereotic components such as chondrules by processes near the magnetic x - point where the disc and the stellar magnetosphere meet .
recent cosmochemical and age - dating evidence ( connelly et al .
2008 ; desch et al .
2012 ) favor more traditional explanations for these effects , but this does not necessarily invalidate the dynamical component of the x - wind model . an alternative launching mechanism can be found in models where the inner parts of protoplanetary discs are dominated by strong magnetic fields which launch magnetocentrifugal winds ( knigl & pudritz 2000 ; bai & stone 2013 ; armitage , simon & martin 2013 ) from a wider range of radii , which may also entrain solid materials and return them to the outer disc .
our goal is to examine the effect that such a model has on the global inventory of solid material in the protoplanetary disc , and how it might set the stage for subsequent evolution .
thus , in [ model ] , we describe a simple model of an evolving gaseous protoplanetary disc , within which we describe the dynamics of small ( dust ) particles of various size , which migrate radially inwards until they reach the inner edge of the disc , at which point they are redistributed outwards to begin the cycle again . in [ results ]
we describe the resulting evolution of gas and dust components and how they set a context for planet formation through the onset of gravitational instability and the formation of planetesimals .
in order to describe the dynamics of the solid particles , we need to adopt a model for the evolution of the gaseous protoplanetary disc , which acts as the background for the evolution of the solids .
we follow the description of alexander & armitage ( 2007)hereafter aa07 which provides a simple model of a protoplanetary disc that includes the viscous evolution as well as the evolution through the transition disc phase and eventual dissipation of the disc .
the gas surface density @xmath7 evolves according to the traditional one - dimensional diffusion equation @xmath8 - \dot{\sigma}_{w},\end{aligned}\ ] ] where @xmath9 is the cylindrical radius , @xmath10 is the kinematic viscosity and @xmath11 is the mass loss rate due to a wind from the disc , which is a function of @xmath9 and time @xmath12 . the viscosity is based on the @xmath13 prescription @xmath14 , where @xmath15 is the orbital frequency and @xmath16 is the disc scaleheight . following aa07 , we adopt @xmath17 , which yields a viscosity that grows linearly with @xmath9 .
we will assume @xmath18 as our default viscosity parameterisation .
the aa07 model also includes a disc wind , driven by photoionization from the central star . the inclusion of the evaporation yields an evolution that begins with the standard accretion and viscous spreading , but eventually results in a gap opened in the gas disc at @xmath19au . at this point
the gas interior to the gap accretes rapidly onto the star , leaving an outer disc and an inner hole .
the outer disc is then slowly removed by evaporation . for simplicity , we restrict our model to the diffuse radiation field used in aa07 , as the additional direct radiation component applies to the later dissipation of the outer disc once the inner disc has been evacuated , a phase of evolution of less interest to us .
we use the wind profile as outlined in the appendix a of aa07 . an alternative mode of disc clearing
is presented by armitage , simon & martin ( 2013 ) asm hereafter motivated by studies which suggest that the conservation of magnetic flux in a disc leads to an increase in the strength of the magnetic disc wind as the gas surface density decreases .
this in turn drives an evolution of the viscosity parameter @xmath13 , and an increase in the rate of viscous evolution with time .
this can result in a qualitatively different clearing , by removing the outer disc first and then finally the inner disc .
we incorporate this as an alternative disc model by neglecting the photoevaporation but including an evolution of @xmath13 as parameterised by asm .
this evolution is based on the global conservation of magnetic flux in the disc , which amounts to the conservation of the integral of @xmath20 in our model .
our default initial disc is @xmath21 which is of similar form to that of aa07 but more compact , as befits our interest in the inner regions of the disc . our model for the initial deposition of solid particles
is drawn from hughes & armitage ( 2012 ) , setting the condensible mass fraction to be @xmath22 interior to the water ice line ( which we take to be at 3 au ) , 0.0106 between the water ice line and the methane ice line ( taken to be at 15 au ) , and 0.0149 exterior to that .
we initially neglect grain growth and so we assume all the condensible material is deposited at the beginning and we thereafter simply follow the dynamical evolution .
we will discuss growth effects in [ sizes ] .
the solid particles that condense out of the gas phase must eventually grow into larger bodies to form planets .
however , there are several significant hurdles along the path to planet formation .
the first is that , as small particles grow larger , their coupling to the gas weakens and the resulting difference in orbital velocity means that they are subject to aerodynamic drag forces .
this causes them to spiral inwards towards the star ( whipple 1972 ; weidenschilling 1977a ) , where they may be lost .
we incorporate this into our model using the formalism of aa07 , based largely on takeuchi & lin ( 2002 ) , in which the radial velocity of each particle is given by @xmath23 the quantity @xmath24 represents the radial velocity of the gas due to the viscous evolution , and @xmath25 is the keplerian orbital velocity at distance @xmath9 .
thus , as the parameter @xmath26 ( defined below ) becomes non - negligible , the solids will begin to evolve radially with respect to the gas .
the non - dimensional stopping time @xmath27 is given by @xmath28 over most of the disc , where @xmath29 and @xmath30 are the density and radius of a given grain .
this quantity determines whether the particles are well coupled to the gas ( @xmath31 ) or not .
the quantity @xmath26 reflects the differential motion between a particle in a true keplerian orbit and the gas , which is partially supported by a radial pressure gradient , @xmath32 the nature of the drag force changes depending on the size of the particle relative to the mean free path in the gas . over the bulk of our disc ,
the drag force is of the epstein form ( equation [ epstau ] ) .
however , in the outer part of the disc , at late times , it can become of stokes form @xmath33 where @xmath34 is proton mass , @xmath35 is the collision cross - section of molecular hydrogen
see discussion in chiang & youdin ( 2010 ) ] , and @xmath15 is the orbital frequency . we must also account for diffusive effects due to gas turbulence , which acts against particle density gradients established by differential drag .
thus , we include a diffusive velocity term @xmath36\ ] ] where @xmath37 is the local dust density , and @xmath38 is the turbulent diffusivity . the default assumption is to equate this with the viscosity @xmath10 ( defined as the schmidt number @xmath39 ) .
deviations from unity are possible , but our focus is on the smaller particles , while proposed corrections are most important for larger particles in the marginally coupled limit , as discussed by youdin & lithwick ( 2007 ) .
we assume the diffusion coefficient is the same for all particle sizes , but that the diffusion velocity of each particle size is determined by the surface density of that species only .
our model thus far represents a standard treatment of the evolution of both gas and solids in an evolving protoplanetary disc . to this
, we now add the hypothesis that solid material is entrained by a wind or jet in the inner disc ( liffman & brown 1995 ; shu et al .
1996 ; 2001 ; salmeron & ireland 2012 ; bai & stone 2013 ) and transferred back to the outer disc .
the original shu et al .
model focussed primarily on redistrbution only as far as the asteroid belt , but hu ( 2010 ) , motivated by the stardust results , reviewed the aerodynamic requirements for redistribution in an x - wind model and found that , with uncertainties due to outflow and disc geometries , particles of size in the range of microns to millimeters could be transferred as far as the kuiper belt regions of discs .
larger particles are less well coupled to the outflow and likely to return to the disc on smaller scales , while smaller particles are sufficiently well entrained that they are easily lost from the system entirely .
thus , it is possible to invoke such models to explain the observed composition of the comet wild-2 as well .
we incorporate these effects into our simple model by assuming that , when solid particles reach 0.05 au , they are instantly transferred back to the outer disc , with an equal probability per unit area of landing between 35 and 90 au .
this keeps the entire solid inventory circulating between the inner and outer disc , although the gas disc is slowly being accreted onto the star or evaporated .
we will examine the effects of varying the properties of the redistribution in
[ loss ] .
our hypothesis of wind entrainment and circulation of solid material provides a potential solution to the first hurdle to planet formation , in that it prevents the accretion of the solid material by the star .
there is a second well - known hurdle to planet formation , which is the difficulty of converting small , dust - size particles into larger bodies with the densities and tensile properties of planetesimals . despite decades of work
, there is little convincing evidence that this transformation can be achieved by incremental processes , and thus the best hope for forming planetesimals is the gravitational instability of a solid dust layer that sediments to the midplane ( goldreich & ward 1973 ) .
however , this model has also proven difficult to implement because turbulent gas motions can prevent a sufficiently dense dust layer from forming ( weidenschilling 1980 ) . in order for the solid surface densities to be high enough to overcome turbulent stirring , youdin & shu ( 2002 ) demonstrated that the local solid densities had to be enhanced by factors of at least several over traditional protoplanetary metallicities .
there have been a variety of proposals ( see chiang & youdin 2010 for a review ) for achieving such enhancements .
one class of models rely on the backreaction on the gas from radial particle streaming ( youdin & chiang 2004 ; youdin & goodman 2005 ; johansen et al .
2007 ) to promote gravitational instability .
however , this only works if the particles are big enough to decouple from the gas on timescales similar to the orbital time ( @xmath40 ) , which requires growth to sizes well in excess of the micron centimeter sizes considered here .
another set of proposals includes growth of secular modes ( ward 2000 ; youdin 2011 ) to produce particle clumping .
however , these models also require either large particles or low turbulent viscosities ( @xmath41 ) to produce realistic growth . in our simple model ,
the global dust inventory remains in circulation , even while the disc slowly loses gas by accretion and wind loss . as a result , the traditional scenario for direct gravitational instability remains potentially viable , because the required overdensities will eventually come to pass , as the global solid / gas ratio is monotonically increasing , although the locations where the instability occurs will depend on the evolutionary details .
these locations are what is of interest to us , as they will then dictate the initial conditions for the later stages of planetesimal accumulation .
therefore , we adopt the youdin & shu ( 2002 ) ys02 model of planetesimal formation . interpreting the ys02 critical dust surface density in terms of our disc model yields an expression for the critical dust surface density @xmath42 where @xmath43 - \sqrt{1 + 2 \psi}$ ] , and @xmath44 . as an example , for a gas surface density of @xmath45 at @xmath46 , this requires @xmath47 for gravitational instability .
this represents an enhancement factor of 37 over the original , in situ , ratio , although the precise value of the necessary enhancement may be uncertain ( e.g. gomez & ostriker 2005 ; shi & chiang 2013 ) .
whenever the local surface density of dust exceeds the critical value , we assume the excess is transformed into planetesimals by gravitational instability and no longer evolves with drift or with the gas .
this assumption is a good one until the planetesimals assemble into larger bodies , at which point they may migrate again by exchanging angular momentum with the disc gas .
we shall assume that this latter process takes @xmath48myr the point at which our calculation ends .
therefore , the final result of our calculation is a deposition profile of planetesimals that result from the global evolution of the dust and gas populations , which may then serve as the initial conditions for the later assembly of larger bodies .
we begin with a reference model , in which the gas disc given by equation ( [ init ] ) is evolved , along with our model for dust drift , diffusion and formation of planetesimals included .
however , in this reference model , there is no redistribution of material from the inner disc to the outer disc .
thus , the reference model represents the expectations for planetesimal formation if we adopt the standard picture of viscous gas evolution and dust drift .
this model also requires a choice of the size of solid particles , as they will determine the overall drift rates .
figure [ ref1 ] shows the evolution of a dust population of size 1 mm , where particles are simply removed when they reach the inner edge .
the dust population is substantially depleted within 0.3 myr and completely removed within 1 myr .
other sizes behave in a similar fashion , albeit with different timescales larger dust
is removed more rapidly and smaller dust is retained longer . , width=317 ] the gas disc shows the expected evolution , with the bulk of the mass moving inwards and accreting onto the star , with a small fraction moving outwards to conserve angular momentum .
we see that the effect of the wind starts to have an influence after @xmath49 myr , and that this rapidly causes a gap to form at @xmath50 , and the gas interior to this is quickly accreted onto the star .
the outer gas disc is slowly eaten away from the inside out by the wind mass loss .
the radial drift of the dust will enhance solid surface densities and potentially generate sufficient overdensities for planetesimal formation ( youdin & chiang 2004 ) , but studies in evolving gas discs have not found significant mass deposition ( hughes & armitage 2010 ) . in our model too , we find little planetesimal formation .
our models do satisfy the analytic criteria for particle pile - up identified by youdin & chiang , but the enhancements are not sufficient to generate gravitational instability without recycling ( however , see [ wind ] ) .
the essential problem is that the radial migration of dust is too rapid in a protoplanetary disc for easy planetesimal formation .
figure [ x1 ] now demonstrates the effect of including our model for redistribution of the solid material back to the outer disc .
this model uses the same parameters as the model shown in figure [ ref1 ] , with dust of size 1 mm .
after several radial travel times , the dust establishes a steady - state surface density profile in the disc , which is steeper than the original dust profile .
as the gas disc evolves and the gas surface density decreases , the dust profile will eventually exceed the critical density at particular locations and thus planetesimals will form through gravitational instability .
the deposition of planetesimals begins at @xmath51 au but eventually extends inwards to @xmath52 au .
the steep dust density profile can be understood by the dynamics of dust in the @xmath53 limit , which leads to a drift velocity which scales @xmath54 in the case where the gas surface density @xmath55 . given this density profile , the dust population achieves steady state flux when @xmath56 . .
the dotted line indicates the surface density of planetesimals at 0.2 myr ( early deposition ) and the dashed line the same but at 4 myr ( final deposition ) .
[ x1],width=317 ] the speed with which material moves radially through the disc is the result of particle size , and so the normalisation of the steady state surface density profile varies with the size of the particle ( the overall slope is the same as long as @xmath53 ) .
figure [ solid2 ] shows the final planetesimal deposition profiles of models in which all the particles have radii of @xmath57 , @xmath58 and @xmath59 respectively .
also shown are estimates of the required rocky inventories for the planetary systems seen by radial velocity ( rv ) studies and kepler , as well as for our own terrestrial and giant planets .
for the inner disc we use the estimate of chiang & laughlin ( 2013 ) , while we use a modification of the weidenschilling ( 1977b ) estimate of the minimum mass solar nebula ( in that we restrict ourselves to the rocky component here ) .
we see that circulating populations of small particles in the @xmath53 regime are able to provide an inventory of planetesimals that matches the requirements imposed by both terrestrial and extrasolar observations .
note in particular that this model provides a rationale for the smooth , quasi - power - law distributions usually adopted in planet formation models .
cores for the outer solar system planets ) .
the dot - dashed line is the solid surface density estimate by pollack et al .
( 1996 ) for the formation of giant planet cores .
the green curve represents the planetesimal deposition in our model for a dust component composed of 1 cm ( large ) particles , the blue curve is for a model in which particles have size 1 mm , and the red curve shows the equivalent result when the particles are composed of 10@xmath60 m ( small ) particles , which move through the disc more slowly .
[ solid2],width=317 ] a proper model for the evolution of the small bodies would also include a model for particle growth over time , which would couple the chemistry and the dynamics of small bodies in the disc .
such a calculation is well beyond our simple model , but we can anticipate some of the consequences of grain growth by considering a composite model of sizes designed to examine the differential motion of both small and large particle components in [ singlesize ] .
our single size distribution shown in figure [ x1 ] was chosen on the basis that the evidence from solar system meteorites suggests that the size distribution of chondrules peaks between 0.11 mm ( scott & krot 2005 ) and thus a dust population of such particles offers a reasonable simple model of the underlying constituents .
however , if this is the result of aerodynamic sorting in the formation of the chondrites ( e.g. cuzzi , hogan & shariff 2008 ) , then the original size distribution of the nebular solids may be different .
thus , we consider a model in which the solids are distributed in proportions of 10% large ( 1 cm ) particles , 20% medium ( 0.5 mm ) particles and 70% small ( 10 @xmath60 m ) particles .
these proportions are applied at all radii in the initial construction of the dust profiles .
this is intended to model a more primitive dust distribution , to see how the resulting planetesimal population is affected .
the resulting distribution is shown in figure [ x2 ] .
, width=317 ] in figure [ x2 ] , we see that the first planetesimals to form are those at @xmath61 au , after only 0.3 myr .
the composition of these planetesimals is dominated by the large particles , whose profile is most concentrated in the inner disc and whose radial mobility is responsible for achieving the overdensity necessary for gravitational collapse . however , by 1 myr we see that the planetesimal formation is starting to extend outwards to the terrestrial planet domain , and eventually the final distribution extends out to @xmath62 au .
figure [ size ] shows the distribution of sizes that are removed from the dust disc by planetesimal formation over the full 5 myr history of the disc .
we see that larger particles dominate the inventory at small scales , with progressively smaller particles being incorporated further out and at later times . particles .
we see that solids that form on scales @xmath63au are dominated by the smaller particles , while the material at @xmath64au is dominated by the larger particles .
[ size],width=317 ] in summary then , the choice of size distribution does not significantly change the amount of mass distributed in planetesimals inside 1 au , but can influence the location of the first planetesimals to form ( larger particles move the location inwards ) and how much mass is formed in planetesimals are scales of several au ( smaller particles enhance the mass on larger scales ) .
figure [ mass ] shows the enclosed mass in planetesimals as a function of @xmath9 in the model of figure [ x2 ] .
we see that @xmath65 is deposited within 1 myr inside 1 au . after 2 myr ,
this is supplemented by an additional @xmath66 that is deposited out as far as 10 au . for comparison
, we also show the final mass deposition profile from the single size model in figure [ x1 ] . in this case the bulk of the mass is deposited inside 2 au .
the mass inventory in planetesimals laid down in this fashion may provide the starting conditions necessary for a variety of planetary formation scenarios .
the assembly simulations discussed in hansen & murray ( 2012 , 2013 ) require anywhere from 2550 @xmath1 of rocky material interior to 1 au , depending on the mass inventory of the observed system .
assembly of giant planet cores in the solar system requires 1020 @xmath1 of material on scales @xmath610 au , which is also realised if there is sufficient mass in particles smaller than 1 mm . , after 1 myr of the discs evolution .
the dotted line represents the same quantity after 2 myr , and the solid line represents the final deposition profile , after 5 myr , when the gas disc is largely depleted .
the long dashed line indicates the final planetesimal deposition in the case when the particles are all of size 1 mm .
[ mass],width=317 ] we have also investigated an alternative model due to armitage , simon & martin ( asm ) , in which the disc is cleared not by photoevaporation , but rather by the amplification of the magnetic field in the disc and a resulting acceleration of the viscous evolution at late times .
our calculations suggest that the specific mechanism of disc clearing does not qualitatively affect the mass deposition interior to 1 au , but may influence the amount of planetesimals formed on larger scales .
figure [ xbz ] shows the equivalent of figure [ x2 ] , but using the asm model rather than the aa07 model , with an initial parameterisation equivalent to the @xmath67 model of asm ( chosen to clear the disc on a similar timescale to the photoevaporative model ) .
we include the same distribution of particle sizes , and find that the deposition of planetesimals once again starts on scales of @xmath52au after 0.3 myr and spreads slowly outwards on timescales of myr .
the only real difference between the final planetesimal mass profiles is a slightly more extended disc at the outer edge , driven by the fact that the late - time decrease in the gas surface density is more rapid in this model .
, width=317 ] the role of disc winds could have a larger effect on small scales if they are capable of removing angular momentum without driving mhd turbulence .
bai & stone ( 2013 ) find that turbulence can be suppressed if there is sufficient vertical field threading the disk , with the disk wind able to carry away the angular momentum necessary to facilitate accretion .
therefore , we have run the same model as in figure [ x1 ] but with the diffusive contribution to the particle velocity set to zero interior to 10 au .
the resulting evolution is shown in figure [ xdiff ] .
two differences in the planetesimal deposition are notable .
the first is that the deposition occurs much more rapidly ( starting at 0.1 myr and substantially finished by 1 myr ) .
the second is that the profile extends inwards farther , indeed as far as the wind launching radius . ,
width=317 ] both features are the result of the fact that the turbulent diffusion acts to smooth out a particle pileup acting against the inspiral exterior to a local maximum in the dust density , but accelerating the inspiral interior to it . without this contribution , our dust evolution looks a lot more like the model presented in youdin & chiang ( 2004 ) , wherein the required overdensity for gravititional instability results purely from the concentration of the solids towards the interior of the disk .
however , the depletion of the dust is again quite rapid and possibly at odds with observations of young star forming regions .
our simple model assumes that the gas is slowly lost from the system , but that solids circulate indefinitely .
the result is that gravitational instability and planetesimal formation is inevitable .
how robust is this consequence in the face of the inevitably imperfect retention of solids in a more realistic model ?
some solid material may remain entrained with the gas that is accreted onto the star , and some material may be carried to infinity by the outflow .
figure [ x3 ] shows the evolution of the same model as figure [ x1 ] ( dust contained in 1 mm particles ) , but with a model which incorporates mass loss . in this case
mass is lost from the inner edge of the disc at the same rate as it is added back at the outer edge i.e. a 50% retention fraction with each cycle . not surprisingly , the dust is depleted more rapidly , and with much less planetesimal formation .
however , the dashed line does show that some planetesimals do form , with a concentration between 0.51 au .
the final amount of mass in solids is @xmath68 , which actually makes an excellent initial condition for assembly of systems like that of the terrestrial planets of our own solar system ( hansen 2009 ) , although it lacks the solid inventory on smaller scales that other systems require .
since the loss rate is determined by the rate at which material passes through the inner edge of the disc , this means that larger particles are lost more rapidly and smaller particles are retained longer . but would still be quite appropriate to the formation of the solar system terrestrial planets .
[ x3],width=317 ] figure [ losses ] shows the impact of dust loss on our simple model of a composite population .
the solid curve indicates the planetesimal deposition result from the closed box model of figure [ x2 ] , at ages of 1 myr and 2 myr .
the dotted line is the equivalent for a model in which particles of all sizes are subject to the same 50% loss on each passage used in the model in figure [ x3 ] .
we see that the biggest change is the reduction in the amount of mass deposited into planetesimals at small radii .
this is because these are composed preferentially of the larger particles , which have shorter circulation times and therefore make more passages through the inner disc per unit time . as such , they are preferentially depleted in this simple model .
perhaps a more realistic model is to deplete only the smallest particles , as hu ( 2010 ) demonstrates that these are the most likely to be strongly coupled to the outflow and lost from the system .
the dashed line in figure [ losses ] indicates the consequence for this model , in which the overall deposition is only slightly reduced from the closed box model .
this suggests that the broad features of the planetesimal deposition are robust to a moderate amount of loss in the circulation and that the component of particles with sizes @xmath0 cm or larger are the most likely to be depleted in the final accumulation .
, width=317 ] the recirculation of solids requires that gas drag from the outflow accelerate particles until the outflow gas density drops low enough for the particle to decouple from the drag force and fall back to the disk .
each time the gas lifts a particle out of the potential well it saps the energy from the outflow .
although the total mass fraction of the disk in solids is small , if dust particles make many passages through the disk and are lifted each time by the outflow , then the energy required to lift the dust to large radii many times may start to affect the energetics of the outflow . for a single size particle distribution , with a characteristic disc traversal time @xmath27
, the particles will make @xmath69 passages through the disk during a global lifetime t ( which may be taken as the time to start forming planetesimals , which remove the dust from circulation ) .
this means that gas drag from the wind must supply sufficient binding energy to lift a mass @xmath70 to the outer part of the disk .
as long as this is substantially less than the gas mass removed in the outflow , it will not have a significant effect on the outflow dynamics .
for our default model with mm - size particles , @xmath71 years and @xmath72 years , so that dust makes @xmath62 passages through the disk before being absorbed into planetesimals . for global metallicities
@xmath73 , this means that the outflow dynamics are not strongly affected as long as the outflow carries more than 10% of the gas mass that flows inwards from the disk to the star . for dust populations with a size range ,
larger particles make more passages through the disk in a finite time , and smaller particles fewer .
thus , for distributions considered here ( e.g. in figure [ x2 ] ) , with more of the mass in smaller particles , the energy requirements for recirculation are well within the energy supplied by the outflow . on timescales
@xmath74myr , the mass flux in the outflow must necessarily drop as the gas disk mass drops , but the circulating dust mass also drops as the mass turns into planetesimals on this timescale , so that the ability of the outflow to lift solid material does not appreciably limit the end result of the circulation .
the results of these calculations suggest that a protoplanetary disc with a circulating dust component will naturally produce conditions that allow for the formation of planetesimals via gravitational instability , and with a substantial mass deposition interior to 1 au .
this offers a promising route to the formation of planetary systems by in situ assembly on these scales ( hansen & murray 2012 ; 2013 ; chiang & laughlin 2013 ) .
of course , there are still several uncertainties that need to be addressed in matching the properties of the initial planetesimal population to that of the final planetary system .
the formation of planetesimals is a necessary step on the road to forming a planet , but the conditions under which the subsequent assembly takes place will determine what kind of planet ultimately results .
the first planetesimals to form in this model do so quite early , within 0.5 myr of the original condensation from the nebula .
this occurs when there is still substantial gas present in the disc and thus the future assembly into protoplanets will likely be affected by interactions with the gas , which can affect both the mass inventory and potentially the radial location .
figure [ inout ] shows the relative mass in both gas and planetesimals in both the composite size model and the one in which all dust is @xmath51 mm , for both the inner disc ( defined here as @xmath75au ) and the outer disc ( @xmath76 ) . in the case of the inner disc ,
the amount of mass in solids becomes comparable to the gas within 0.5 myr . in models in which the disc clears via photoevaporation , such as those of aa07
, the gas interior to 1 au is not evaporated from the system , but accreted onto the star .
thus , the eventual disposition of the remnant gaseous disc will require it to either be accreted or cross the orbits of the final planets to get to the central object . as a consequence , this remnant gas disc may have dynamical consequences for the final planetary configuration , but it does not have sufficient mass to overwhelm the rocky inventory available for forming planets . indeed , much of this surviving gas may ultimately be accreted by the planets rather than the star , since observations suggest that many of the newly discovered planets are predominantly rocky but possess sufficiently massive gaseous envelopes that they must have accreted some gas from the nebula ( e.g. lissauer et al . 2013 ) . in the case of the model shown in figure [ x1 ] ,
the gas / planetesimals mass ratio is @xmath77 interior to 1 au at the point when photoevaporation decouples the outer gas disk from the inner gas disk .
lopez & fortney ( 2013 ) suggest a characteristic value of the hydrogen to rock ratio is more like @xmath78 , which would require capture of the envelope on slightly shorter timescales @xmath6 myr ( figure [ inout ] ) . in heavy elements .
we see that formation of gas giant planets requires formation within @xmath0 1.8 myr even with very efficient accumulation of mass from the disc onto the planet .
the gas mass in the inner disc is exceeded by the planetesimal mass within 0.7 myr , and reaches 1% at 2.53 myr .
[ inout],width=317 ] exterior to 1 au , there is a delay of 12 myr before a population of planetesimals begin to accumulate .
figure [ inout ] demonstrates that the gas / rock ratio is more variable on these scales , depending on the exact size distribution of the underlying solids . allowing for more massive discs can allow for the formation of rocky cores more rapidly , but
all models show the same requirement that the assembly of giant planet cores needs to occur within a narrow window of 12 myr , in order to capture sufficient gas from the nebula .
this echoes the well - known constraint on the formation of giant planets ( e.g. pollack et al .
1996 ) . the process can be helped if we allow for a higher metallicity in the gas .
figure [ xz ] shows the same model as in figure [ x1 ] , but we have increased the original dust sedimentation by a factor of two to mimic an enhanced metallicity .
we see that the planetesimal formation in this model meets the surface density requirements of pollack et al .
( 1996 ) within 0.2 myr , suggesting that there is sufficient mass to form giant planet cores and to accrete the gaseous envelopes .
the fact that this process is enhanced in high metallicity models may help to explain the correlation between host star metallicity and giant planet frequency .
, width=317 ] the last planetesimals to form are those on scales @xmath62 au , which result from the onset of gravitational instability during the last stages of the gas disc clearing .
the gas - poor origins of this outermost rocky component provides a natural environment for the origin of kuiper belt analogues , and for the assembly of the gas - poor ice giants of the outer solar system ( e.g. goldreich , lithwick & sari 2004 ) .
these effects may be further enhanced in cases where the evaporation is enhanced by an external evaporation field ( throop & bally 2005 ) .
the composition of the material that assembles into planetesimals in a composite model also varies radially .
interior to @xmath79au , the planetesimals are dominated by larger particles ( @xmath80 cm sizes ) , while the planetesimals are dominated by medium sized particles between 0.51 au , and smaller particles exterior to that .
this natural aerodynamic sorting by particle size results from the different normalisations of the steady state surface density distributions , as determined by the radial drift velocity .
similar sorting is observed in the chondritic material of the solar system ( e.g. scott & krot 2005 ) , in which chondrite families show an internal size consistency although there is some variation between families , and has spurred models of sorting by turbulent processes ( e.g. cuzzi , hogan & shariff 2008 ) .
as we have shown , a dust population of this size can produce a planetesimal disc well suited to the in situ assembly of compact planetary systems .
however , composite size distributions can also match this observation since mm - scale particles dominate the planetesimals that form on scales @xmath0 1 au , with larger(smaller ) particles favouring planetesimals on smaller(larger ) scales .
furthermore , isotopic analysis of solar system materials also suggests that the process of planetesimal assembly lasted for at least a few myr , in that chondrules are inferred to be identifiably older ( e.g. connelly et al . 2008 ) than the oldest calcium - aluminum rich inclusions ( cai ) .
this is also nicely matched by the timescales in our model , with planetesimals at scales @xmath80au taking @xmath63myr to form , although the innermost materials assemble within 0.5 myr .
, while the dotted line is the mass retained in the dust component .
we see that the dust mass remains substantial for several myr and only starts to decrease significantly once the outer parts of the disc start to form planetesimals .
[ dust],width=317 ] our baseline gas disc evolution is drawn from aa07 , whose goal was to study the evolution of dust that is observed in young protostellar systems through the reprocessing of the stellar emission .
this work was motivated by the observed persistence of infrared emission from dust grains throughout the tt phase ( andrews & williams 2005 ) , contrary to the expectation that the mm - sized dust grains should be removed by the radial drag forces discussed above ( takeuchi & lin 2005 ) .
aa07 discuss replenishment by evolution and growth in the grain population .
while such processes are quite likely , our model also alleviates much of the observational discrepancy , since dust is not lost from the system but continuously recycled back to large radii .
the formation of planetesimals will also remove dust ( until such time as it is replenished by planetesimal collisions ) , so that our model also predicts a systematic loss of observable dust over time , shown in figure [ dust ] .
we see that the dust decreases by two orders of magnitude on a timescale @xmath6myr , which is more palatable than the timescale for loss via radial drift , which can be one to two orders of magnitude smaller .
another consequence of such a model is that the material accreted onto the star should be depleted in heavy elements .
our nominal model contains an initial mass of 0.03@xmath81 of gas , most of which is ultimately accreted onto the star . for an original global metallicity fraction @xmath82 , this implies a total mass of @xmath83 in potential condensibles .
however , using the condensation model of hughes & armitage ( 2012 ) , we only condense 114 @xmath1 into dust , with the difference being volatile material that remains in the gas phase in the inner disc .
if we assume that only this remnant metal inventory is accreted with the gas , then we find that the material accreted onto the star should slightly depress the observed overall metallicity and enhance the volatile to refractories ratio slightly . to estimate the amount of the change , we assume that the surface convection zone during this protostellar stage is @xmath84 .
if we add the above gas to this convection zone , we decrease the net metallicity by @xmath85 , or by 0.07 dex .
some authors have claimed evidence for such metallicity trends ( ramirez et al 2009 , ramirez et al .
2011 , gonzalez hernandez et al .
2013 , ramirez et al .
2013 ) in nearby stars , and have concluded that the sun is depleted in refractories due to the formation of the terrestrial planets .
however , it is not clear how easy it is to define a suitable control set of stars without planets , given that observations suggest at least 20%50% of sun - like stars ( howard et al .
2010 ; mayor et al .
2011 ) host low mass planetary systems , whose detectability is not assured .
nevertheless , our results suggest that such enrichments are observable , and do not require the assumption of late time accretion to avoid dilution of the signal , because the material removed from the accreted gas is more than an order of magnitude larger than that estimated using our , relatively low mass , terrestrial planet system ( e.g. chambers 2010 ) .
we have presented a simple model for the evolution of a protoplanetary disc in which dust particles undergo radial drift inwards , but are then recycled to the outer parts of the nebula through the action of a stellar or disc wind .
although this model is quite simplistic , it provides a natural framework for the deposition of tens of earth masses of material into planetesimals on scales of 0.11 au .
this matches the required mass inventory to assemble the observed planets in situ .
there are also a variety of subsidiary issues that suggest further study is warranted .
the retention of solids while gas is lost produces a natural evolution of the solid / gas ratio towards the limit where gravitational instability and planetesimal formation is likely to set in , obviating the need to invoke other physical mechanisms that require the existence of large particles or anomalously low viscosities .
much of the planetesimal reservoir is deposited within 1 myr , which allows for the capture of residual gas from the nebula to explain the observed low mass hydrogen envelopes , and matches the timescales inferred from the cosmochemical age dating of solar system meteoritic components . furthermore .
the gas mass on these scales is less than the mass in planetesimals , so that the resulting planets are likely to be as observed with substantial hydrogen envelopes that are nevertheless a minority constituent by overall mass .
furthermore , we find that increasing the metallicity of the disk has a larger effect on the mass of planetesimals formed on scales of several au , and thus provides a rationale for why the giant planet frequency correlates with metallicity ( gonzalez 1997 ; santos et al .
2004 ; fischer & valenti 2005 ; johnson et al .
2010 ) more strongly than the frequency of lower mass planets ( sousa et al . 2008 ; bouchy et al . 2009
; mayor et al . 2011 ; buchave et al . 2012 ) .
if the solid retention is not perfect , and loss rate is size dependant , it provides an aerodynamic sorting mechanism that may explain the characteristic sizes of chondrules in the solar system .
large particles ( dimensions of cm or larger ) make more passages through the disc and are thus likely to be more depleted via loss at the inner edge .
similarly , entrainment in the outflows is more likely to remove small particles ( hu 2010 ) , which suggests that particles in the size range 0.01 - 1 mm may have the greatest chances of retention and survival .
the circulation of solid material also helps to explain the apparent chemical homogeneity of the solar system solid inventory ( e.g. villeneuve et al .
2009 ) and the ubiquity of material processed at high temperatures ( e.g. brownlee et al .
2012 ) .
there several ways in which this calculation could be improved .
the size evolution of the dust component has been ignored , although this is likely to provide an important feedback loop that may help to regulate the radial profile of the eventual formed planetesimals .
we have also not extended the model forward to consider the formation of larger protoplanets and planets from our initial conditions . nevertheless , we consider the above results encouraging in the sense that they manage to generate conditions that may plausibly be used to match to observed systems , at the reasonable price of invoking an assumption that has already proven useful in other contexts and may be required anyway to explain the well - mixed compositions of solar system bodies .
the author thanks phil armitage , andrew youdin and the referee for helpful comments .
this research has made use of the nasa exoplanet archive , which is operated by the california institute of technology , under contract with the national aeronautics and space administration under the exoplanet exploration program .
99 alexander r. d. , armitage p. j. , 2007 , mnras , 375 , 500 andrews s. m. , williams j. p. , 2005 ,
apj , 631 , 1134 armitage , p. j. , simon , j. b. & martin , r. g. , 2013 , apj , 778 , l14 bai x .-
n . , stone j. m. , 2010 , apj , 722 , 1437 bai x .-
n . , stone , j. m. 2013 , apj , 769 , 76 batalha n. et al .
, 2013 , apjs , 204 , 24 borucki w. j. et al . , 2011 ,
apj , 736 , 19 boley a. c. , ford e. b. , 2013 , arxiv:1306.0566 bouchy f. et al . , 2009 , a&a , 496 , 527 brownlee d. et al .
, 2006 , science , 314 , 1711 brownlee d. , joswiak d. & matrajt g. , 2012 , m & ps , 47 , 453 buchave l. a. et al . , 2012 , nature , 486 , 375 chambers j. e. , 2010 , apj , 724 , 92 chatterjee s. , tan j. c. , 2014 , apj , 780 , 53 chiang e. , laughlin g. , 2013 , mnras , 431 , 3444 chiang e. , youdin a. , 2010 , areps , 38 , 493 connelly j. n. , amelin y. , krot a. n. & bizzarro m. , 2008 , apj , 675 , l121 cuzzi j. n. , hogan r. c. & shariff k. , 2008 , apj , 687 , l432 desch s. j. , morris m. a. , connolly h. c. & boss a. p. , 2012
, met . & planet .
sci . , 47 , 1139 fischer d. a. & valenti j. , 2005 , apj , 622 , 1102 goldreich p. , lithwick y. & sari r. , 2004 , ara&a , 42 , 549 goldreich p. , ward w. r. , 1973 , apj , 183 , 1051 gomez g. o. , ostriker e. c. , 2005 , apj , 630 , 1093 gonzalez g. , 1997 , mnras , 285 , 403 gonzalez hernandez j. i. , et al . , 2013 , a&a , 552 , a6 haghighipour n. , boss a. p. , 2003 , apj , 583 , 996 hansen b. , 2009 , apj , 703 , 1131 hansen b. , murray n. , 2012 , apj , 751 , 158 hansen b. , murray n. , 2013 , apj , 775 , 53 hasegawa y. , pudritz r. e. , 2011 , mnras , 417 , 1234 howard a. , et al . , 2010 , science , 330 , 653 hu r. , 2010 , apj , 725 , 1421 hughes a. l. h. , armitage p. j. , 2010 , apj , 719 .
1633 hughes a. l. h. , armitage p. j. , 2012 , mnras , 423 , 389 ida s. , lin d. n. c. , 2008 , apj , 685 , 584 johansen a. et al . , 2007 , nature , 448 , 1022 johnson j. a. , aller k. m. , howard a. w. & crepp j. r. , 2010 , pasp , 122 , 905 konigl , a. , & pudritz , r. e. , 2000 , in protostars and planets iv , ed . mannings , v. , boss , a. p. & russell , s. s. , university of arizona press , p 759 kretke k. a. , lin d. n. c. , 2012 , apj , 755 , 74 liffman , k. & brown , m . , 1995 ,
icarus , 116 , 275 lissauer j. j. , et al . , 2013 , apj , 770 , 131 lopez , e. d. & fortney , j. j. , 2013 , arxiv:1311.0329 masset f. s. , morbidelli a. , crida a. & ferreira j. , 2006 , apj , 642 , 478 mayor m , et al .
2011 , arxiv:1109.2497 mckeegan k. d. et al . , 2006 , science , 314 , 1724 pollack j. , hubickyj o. , bodenheimer p. , lissauer j. j. , podolack m. & greensweig y. , 1996 , icarus , 124 , 62 ramirez i. , melendez j. & asplund m. , 2009 , a&a , 508 , l17 ramirez i. , melendez j. , cornejo d. , roederer i. u. & fish j. r. , 2011 , apj , 740 , 76 ramirez i. , melendez j. & asplund m. , 2014 , a&a , 561 , 7 salmeron , r. & ireland , t. r. , 2012 , earth .
lett . , 327 , 61 santos n. c. , israelian g. & mayor m. , 2004 , a&a , 415 , 1153 scott e. r. d. , krot a.n . , 2005 , in chondrites and the protoplanetary disk , asp conf .
vol 341 , p 15 shi j .-
, chiang e. , 2013 , apj , 764 , 20 shu f. h. , shang h. & lee t. , 1996 , science , 271 , 1545 shu f. h. , shang h. , gounelle m. , glassgold a. & lee t. , 2001 , apj , 548 , 1029 sousa s. g. et al .
, 2008 , a&a , 487 , 373 takeuchi t. , lin d. n. c. , 2002 , apj , 581 , 1344 takeuchi t. , lin , d. n. c. , 2005 , apj , 623 , 482 throop h. b. & bally j. , 2005 , apj , 623 , l149 villeneuve j. , chaussidon m. & libourel g. , 2009 , science , 325 , 985 ward w. r. , 2000 , in origin of the earth & moon , ed .
r. canup & k. righter , univ .
arizona press , p75 weidenschilling s. j. , 1977a , mnras , 180 , 57 weidenschilling s. j. , 1977b , ap&ss , 51 , 153 weidenschilling s. j. , 1980 , icarus , 44 , 172 whipple f. l. , 1972 , in from plasma to planets , ed .
a. evlius , wiley ( new york ) , p 211 youdin a. n. , 2011 , apj , 731 , 99 youdin a. n. , chiang e. i. , 2004 , apj , 601 , 1109 youdin a. n. , goodman j. , 2005 , apj , 620 , 459 youdin a. n. , lithwick y. , 2007 , icarus , 192 , 588 youdin a. n. , shu f. h. , 2002 , apj , 580 , 494
|
we examine the consequences of a model for the circulation of solids in a protoplanetary nebula in which aerodynamic drag is counterbalanced by the recycling of material to the outer disc by a protostellar outflow or a disc wind .
this population of circulating dust eventually becomes unstable to the formation of planetesimals by gravitational instability , and results in the ultimate deposition of @xmath0 3050 @xmath1 in planetesimals on scales @xmath2 .
such a model may provide an appropriate justification for the approximately power law initial conditions needed to reproduce observed planetary systems by in situ assembly .
[ firstpage ] accretion discs planets and satellites : formation protoplanetary discs
|
kinetic equations describe the time evolution of a certain chosen set of physical variables .
the choice of physical variables in principle is arbitrary , but often in practice is governed by the measurement of interest .
typical examples are the order parameter of a complex system or the coordinate of a brownian particle in a heat reservoir .
the kinetic approach is usually implemented through a proper separation of the microscopic equations of motion of the chosen physical variables into regular and random parts .
an averaging over the random part then generates the effective partition function for the regular part .
this averaging is often referred to as a coarse - graining .
one typical application of the kinetic approach is when the physical variables of interest possess energy in relative excess or deficiency to the rest of a large system .
kinetic theory then describes the approach to equilibrium of the chosen physical variables , as for example in the kinetics of phase transitions or in brownian motion . in the former case , the system is able to release energy to the environment due to some change in its internal state .
provided the environment is disproportionately large relative to the system , the process is irreversible . for a continuous transition , the focus of the present work
, this process of equilibration can be described by the monotonic change of an appropriate order parameter , which is the chosen physical variable .
many systems are known to relax in this manner .
phenomenologically , they are successfully described by the time - dependent ginzburg - landau theory @xcite .
here we are interested in examining under which circumstances physical variables whose microscopic dynamics is second order in time , as for example the higgs order parameter of spontaneous symmetry breaking , may have a dynamics which is effectively first order in time as in ginzburg - landau phenomenological models . qualitatively it is not difficult to argue the plausibility of this standard picture for the higgs symmetry breaking scenario . a single variable , the higgs order parameter , is modeled to control the release of energy to all the modes that couple to it . by basic notions of equipartition ,
one anticipates that some portion of the order parameter s energy will flow irreversibly to any given mode . provided the higgs order parameter couples to a sufficient number of modes
, the motion of the order parameter will be overdamped . in particle physics models ,
higgs symmetry breaking is accompanied by mass generation .
thus the natural couplings for the higgs field @xmath0 to bosonic fields @xmath1 is @xmath2 , gauge fields @xmath3 is @xmath4 and fermionic fields @xmath5 is @xmath6 . for a microscopic realization of time dependent ginzburg - landau theory for the higgs scalar order parameter in a particle physics setting ,
these are the most obvious types of couplings to investigate . in this paper
we will examine the case of purely bosonic couplings in the `` symmetry restored '' regime .
that is , we will study the relaxation of an order parameter which is initially away from the only minimum of the free energy density describing the system .
much of the formalism required for this has already been done in @xcite but we will extend that calculation to the overdamped regime . in an upcoming paper ,
we plan to study the symmetry broken case . to our knowledge
, this paper is the first study of overdamping in quantum field theory with realistic couplings between system and environment , as inspired by particle physics .
overdamping has been studied in quantum mechanical reaction rate theory for a particle escaping from a metastable state ( for a review please see @xcite ) .
this is sometimes referred to as the kramer s problem , with the overdamped limit also called the smoluchowski limit .
quantum mechanical models describing this problem are commonly of the system - heat bath type .
microscopic quantum mechanical models have been constructed along these lines , in which the particle ( system ) is coupled to a set of otherwise free harmonic oscillators ( heat bath ) .
such microscopic system - heat bath models are often referred to as caldeira - leggett ( cl ) models . in many cases they
have been exactly solved @xcite .
the overdamped limit has been derived in these models for the case where the coupling is linear with respect to the oscillator variables but arbitrary with respect to the particle variable @xcite .
a caldeira - leggett type model has also been formulated for the case where the system is a self interacting scalar quantum field coupled linearly to a set of otherwise free fields and the overdamped limit has been obtained @xcite .
this model does provide a microscopic quantum mechanical realization of time - dependent ginzburg - landau dynamics in scalar quantum field theory .
however , since the couplings between system and environment variables are linear , it should be considered as a first step toward more realistic treatments .
more importantly , the calculational method used in @xcite can not be extended to the case when the system variable couples quadratically to other fields .
although the analysis of overdamping in this paper has general applicability , it was motivated by the warm inflation scenario of the early universe @xcite . in @xcite
it was realized that the standard higgs symmetry breaking scenario , when put into a cosmological setting , provides suitable conditions for the universe to enter a de sitter expansion phase and then smoothly exit into a radiation dominated phase .
the overdamped motion of the order parameter in this scenario may sustain the vacuum energy sufficiently long for de sitter expansion to solve the horizon and flatness problems .
simultaneously , the relaxational kinetics of the order parameter can maintain the temperature of the universe and permit a smooth exit from the de sitter phase into the radiation dominated phase .
finally , the thermal fluctuations of the order parameter provide the initial seeds of density perturbations , which in addition could be scale free under specified conditions @xcite .
an elementary analysis of this scenario , based on friedmann cosmology for general realizations of order parameter kinematics , indicated that if the universe s temperature does not fall too much during de sitter expansion , then the cosmological expansion factor from the de sitter phase should be of order the lower bound set by observation @xcite .
although this is not a tight constraint of this scenario , it is a natural one .
an analysis of cobe data motivated by this expectation did indicate a slight preference for a small super - hubble suppression scale , which could be interpreted as arising from a de sitter expansion with duration near its lower bound @xcite .
furthermore , the overdamped limit required by warm inflation , when expressed in different terms , was noted @xcite to be an adiabatic limit , for which known methods from dissipative quantum field theory @xcite are presumed valid .
these facts provide further motivation to seek a microscopic model of the scenario , which is the goal of the present work .
the calculational methods used here , based on schwinger s close - time path formalism , were developed in @xcite .
there are several other works in the literature that apply this formalism to a variety of different situations .
[ see , for example , the works of refs . @xcite . ]
the new feature of the present paper is to shift focus to a kinematic regime dominated by strong dissipation , in order to establish under which conditions this regime leads to overdamped motion .
this approach will allow us to have a unique understanding of the microphysical origin of such dynamical behavior , which is in general invoked phenomenologically in applications ranging from condensed matter physics to inflationary cosmology .
the paper is organized as follows . in sec .
ii our model of interacting bosons is presented and the effective action is computed perturbatively for a homogeneous time dependent background field configuration @xmath7 . in sec .
iii the effective langevin - like equation of motion is obtained for @xmath0 in the symmetry - restored phase . in sec .
iv the overdamped limit of this equation of motion is derived and regions of validity are given . in sec .
v the results of the previous sections , which are for minkowski space , are extrapolated into a cosmological setting and a preliminary examination is made of the warm inflation scenario . in sec .
vi concluding remarks are given .
two appendixes are included to clarify a few technical details , like the evaluation of the imaginary part of the self - energies and to stress the importance of taking fully - dressed field propagators to properly describe dissipation in the adiabatic approximation for the field configuration .
let us consider the following model of a scalar field @xmath0 in interaction with @xmath8 scalar fields @xmath9 : @xmath10 = \frac{1}{2 } ( \partial_\mu \phi)^2 - v[\phi ] + \sum_{j=1}^n \left\ { \frac{1}{2 } ( \partial_\mu \chi_j)^2 - v[\chi_j ] \right\ } - v_{\rm int } [ \phi,\chi_j]\ : , \label{nfields}\ ] ] where @xmath11 = \frac{m^2}{2}\phi^2 + \frac{\lambda}{4 ! } \phi^4 \ ; , \label{vphi}\ ] ] @xmath12 = \frac{\mu_j^2}{2}\chi_j^2 + \frac{f_j}{4 ! } \chi_j^4 \ : , \label{vchi}\ ] ] and @xmath13 = \sum_{j=1}^n \frac{g_j^2}{2 } \phi^2 \chi_j^2 \ : .
\label{vint}\ ] ] for the most part , we will consider all coupling constants positive : @xmath14 , @xmath15 and @xmath16 @xmath17 .
writing @xmath18 in ( [ nfields ] ) , where @xmath19 is a background field configuration and @xmath20 are small fluctuations around @xmath19 , we obtain the expression for the 1-loop effective action @xmath21 $ ] , valid to second order in the fluctuations , by performing the functional ( gaussian ) integrations in @xmath20 and @xmath9 : @xmath22 = s[\varphi ] + \frac{1}{2 } i { \rm tr } \ln \left [ \box + v''(\varphi ) \right ] + \frac{1}{2 } i \sum_{j=1}^n { \rm tr } \ln \left [ \box + \mu_j^2 + g_j^2 \varphi^2 \right ] \ : , \label{action1}\ ] ] where @xmath23 = \int d^4 x { \cal l}[\varphi,0]$ ] , @xmath24 \big|_{_{\phi = \varphi}}$ ] , and @xmath25 + \frac{1}{2 } i \sum_{j=1}^n { \rm tr } \ln \left [ \box + \mu_j^2 + g_j^2 \varphi^2 \right]= } \nonumber \\ & & = - i \ln \int d \eta \prod_j d \chi_j \exp \left\ { - \frac{i}{2 } \eta \left [ \box + v''(\varphi ) \right ] \eta - \frac{i}{2 } \chi_j \left [ \box + \mu_j^2 + g_j^2 \varphi^2 \right ] \chi_j \right\ } \ : . \label{trln}\end{aligned}\ ] ] neglecting contributions to ( [ action1 ] ) which are independent of @xmath19 , we can expand the logarithms in ( [ action1 ] ) in powers of @xmath19 , obtaining , in the graphic representation : @xmath26_{_{\rm 1-loop}}= $ } \thicklines \put(0,0){\line(-1,-1){10 } } \put(10,0){\circle{20 } } \put(0,0){\line(-1,1){10 } } \put(7,15){$\phi$ } \put(25,0){+ } \put(50,0){\line(-1,-1){10 } } \put(60,0){\circle{20 } } \put(50,0){\line(-1,1){10 } } \put(70,0){\line(1,1){10 } } \put(70,0){\line(1,-1){10 } } \put(57,15){$\phi$ } \put(57,-20){$\phi$ } \put(85,0){+ } \put(100,0){$\sum_{j=1}^n \big[$ } \put(150,0){\line(-1,-1){10 } } \put(150,0){\line(-1,1){10 } } \put(160,0){\circle{20 } } \put(157,15){$\chi_j$ } \put(175,0){+ } \put(200,0){\line(-1,-1){10 } } \put(210,0){\circle{20 } } \put(200,0){\line(-1,1){10 } } \put(220,0){\line(1,1){10 } } \put(220,0){\line(1,-1){10 } } \put(207,15){$\chi_j$ } \put(207,-20){$\chi_j$ } \put(240,0){$\big]$ } \put(245,0){$+\ : { \cal o}(\lambda^3 ) + { \cal o}(g_j^6)\ : , $ } \end{picture } \label{graphs1l}\ ] ] where we have identified the propagators in the internal lines .
external lines are @xmath27 for the @xmath0-graphs and @xmath28 for the @xmath29-graphs . before presenting our derivation of the effective nonequilibrium equation of motion for @xmath19 , we contrast our approach with earlier works in the literature .
we closely follow the method of ref .
@xcite in the derivation of the evolution equation for @xmath19 .
in particular , it was shown in @xcite that for slowly changing fields , dissipative terms vanish if they are computed perturbatively with bare propagators .
there are several issues related to this result .
@xcite argue , in the context of a toy model , that dissipative effects can not be studied within perturbation theory : perturbation theory breaks down before dissipative effects can be observed .
this shows that dissipation is a nonperturbative effect in quantum field theory . in @xcite
it was shown that dissipative terms can be derived once a consistent `` dressing '' of propagators is used .
this is an explicit way of considering the effect of quasi - particles ( or single - particle states ) in the evolution of the system , described by @xmath19 , in interaction with a thermal bath which represents fluctuations of @xmath0 and of others fields to which it may be coupled .
it seems reasonable to expect that dissipation effects are closely related to the effect of collisions which dress the field propagators .
take for example the case of a bare propagator expressed in terms of the spectral density @xmath30 , where there is a one to one correspondence between the energy and the momentum of a given state .
this completely neglects the spreading of possible energy states due to interactions . in a full `` dressing '' of propagators ,
this is accounted for through the introduction of a lifetime ( decay width ) for single - particle states , such that the full ( dressed ) spectral density @xmath31 is smeared out .
in particular , particle lifetimes are crucial in the study of relaxation time - scales in quantum many - body theory @xcite .
also , the reason why we can get dissipation within our approach can be traced back to the very way that transport coefficients are derived in quantum field theory . as we will show later ,
the assumption of a slowly moving field is consistent with overdamping in a strong dissipative environment , justifying the adiabatic approximation we adopted . in this regime
, there is a close relation between the dissipation we compute and the shear viscosity computed from the kubo formula @xcite .
as explained in @xcite , diagrams contributing to the shear viscosity have near on - shell singularities for free bare propagators .
full resummed propagators regulate these singularities through an explicit thermal lifetime of single particle excitations .
analogous singularities are exhibited by our expressions for dissipation terms if bare propagators are used .
additional issues concerning the relation of our dissipation terms with the shear viscosity will be discussed in the following two sections . from the above discussion
, we rewrite the lagrangian density in ( [ nfields ] ) as @xmath32 where @xmath33 and @xmath34 are the self - energies for the @xmath0 and @xmath9 fields , respectively .
this way we can work with full ( dressed ) propagators for the @xmath0 and @xmath9 fields ( note the implicit resummation of diagrams involved in this operation ) and at the same time keep consistency by considering @xmath35 and also @xmath36 as interaction terms .
this method has already been adopted before in many different contexts [ see , for example @xcite ] . in terms of the self - energies
the field propagators are written as @xmath37 for both @xmath0 and @xmath9 fields , a finite lifetime of single particle excitations , given in terms of the imaginary part of the self- energies , first appear at the two - loop order .
we thus restrict , for simplicity , the evaluation of @xmath33 and @xmath34 up to the two - loop level .
diagrammatically , the self - energies are given by @xmath38 @xmath39 and @xmath40 @xmath41 the setting sun ( non - local ) diagrams in ( [ selfphi ] ) and ( [ selfchi ] ) ( the two last terms in ( [ selfphi ] ) and ( [ selfchi ] ) ) contribute imaginary terms to the self - energies , from which we can write the decay widths @xmath42 , @xmath43 , for the @xmath0 and @xmath9 fields , respectively , in terms of the on - shell expressions @xcite ( @xmath44 ) : @xmath45 and @xmath46 where @xmath47 is given by the solution of @xmath48 .
explicit expressions for @xmath49 in the @xmath50 model have been obtained in @xcite and @xcite .
we follow @xcite to compute @xmath42 and @xmath43
. a straightforward extension of the computation can be applied to our model of interacting @xmath51 fields .
some of the details are shown in appendix a , where we evaluate the imaginary contribution coming from the mixed setting sun diagrams in @xmath33 and @xmath34 [ the last diagrams in ( [ selfphi ] ) and ( [ selfchi ] ) ] .
even though in general there are no simple way of expressing the results , if we adopt the zero space momentum ( @xmath52 ) approximation for the imaginary part of the self - energies , we can find simple approximated expressions for both ( [ widthphi ] ) and ( [ widthchi ] ) , respectively , given at finite temperature ( @xmath53 ) by ( for @xmath54 ) @xmath55 \label{gammaphi}\end{aligned}\ ] ] and @xmath56 \ : .
\label{gammachi}\end{aligned}\ ] ] in the above expressions , @xmath57 and @xmath58 are the thermal masses for @xmath0 and @xmath9 , respectively .
@xmath59 for @xmath60 and @xmath61 otherwise .
@xmath62 is the dilogarithm function .
some useful approximations for @xmath62 are @xmath63 z + { \cal o } ( z^2)$ ] , for @xmath64 , and @xmath65 , for @xmath66 . ] .
this approximation for the decay widths , in terms of the zero space - momentum expression for the imaginary part of the self - energies , is common to computations of transport coefficients and contrast densities in field theory @xcite .
however , wang , heinz and zhang @xcite showed that this approximation may lead to errors in the calculation of the contrast density in the @xmath50 model .
in fact , the expressions for @xmath67 can be fast changing for some momentum range and values of the masses .
for example , in fig .
1 we plot the value of ( the on - shell ) @xmath68 , obtained numerically , as a function of the momentum , normalized by its @xmath69 expression ( for @xmath70 ) . even though @xmath68 can depart considerable from its @xmath71 value , we will show later that , for a range of small thermal masses , this approximation results in a small error ( @xmath72 ) in the expression for the dissipation coefficient , when compared with the computation using the complete @xmath73 expressions for @xmath68 . in the analysis presented in the next sections
, it will also be sufficient to use the leading order high temperature expressions for the finite temperature effective ( renormalized ) masses , @xmath57 and @xmath74 , appearing in ( [ widthphi ] ) - ( [ gammachi ] ) , ( obtained from the 1-loop diagrams in ( [ selfphi ] ) and ( [ selfchi ] ) , respectively ) , given by ) and ( [ selfchi ] ) , as in the effective action , can be dealt with by the usual introduction of the appropriate renormalization counterterms in the initial lagrangian , for the masses , coupling constants and the wave - function .
in particular , we note that the imaginary terms in the self - energies expressions , coming from the setting - sun diagrams , are finite .
@xmath75 , @xmath76 , @xmath77 , @xmath78 and @xmath14 in ( [ mt ] ) and ( [ mut ] ) and in our later results are to be interpreted as the corrected and not as bare quantities . ]
@xmath79 and @xmath80 in order to obtain the evolution equation for the field configuration @xmath19 , we use the real - time schwinger s closed - time path ( ctp ) formalism @xcite . in the ctp formalism
the time integration is along a contour @xmath81 from @xmath82 to @xmath83 and then back to @xmath84 . for reviews
please see , for example , refs .
@xcite . in the ctp formalism
the field propagators are given by @xcite ( with analogous expressions for @xmath85 ) : @xmath86 where @xmath87 and @xmath88 indicate chronological and anti - chronological ordering , respectively .
@xmath89 is the usual physical ( causal ) propagator .
the other three propagators come as a consequence of the time contour and are considered as auxiliary ( unphysical ) propagators .
the expressions for @xmath90 in terms of its momentum - space fourier transforms are given by @xmath91 where @xmath92 in terms of the decay width @xmath42 , the expression for the full dressed propagators at finite temperature where obtained in @xcite , from which we have @xmath93 e^{-i ( \omega_\phi - i \gamma_\phi ) ( t - t ' ) } + n(\omega_\phi + i \gamma_\phi ) e^{i ( \omega_\phi + i \gamma_\phi ) ( t - t ' ) } \right\ } \nonumber \\ & g_\phi^{<}({\bf q } , t - t ' ) & = g_\phi^{>}({\bf q } , t'-t ) \ : , \label{g><}\end{aligned}\ ] ] where @xmath94 is the bose distribution and @xmath95 is the particle s energy , or dispersion relation , @xmath96 . for @xmath85 , @xmath97 .
with fields in the forward and backward segments of the ctp time contour identified as @xmath98 and @xmath99 , respectively , the classical action can be written as @xmath100 = \int d^4 x \left\ { { \cal l}[\phi_{+},\chi_+ ] - { \cal l}[\phi_{-},\chi_- ] \right\ } \ : , \label{s+-}\ ] ] the evaluation of the effective action at real time can be done exactly as in @xcite .
there are also a number of other works using schwinger s closed - time path formalism to obtain the real - time effective action for field configurations .
e.g. , refs .
] here we will concentrate on the evaluation of the effective equation of motion in the strong dissipative regime . in the evaluation of the effective action there
appear several imaginary terms , once the @xmath29 fields and the fluctuations around the @xmath19 background are integrated out .
these imaginary terms can be interpreted as coming from functional integrations over gaussian stochastic fields , as can be visualized by introducing the new field variables : @xmath101 in terms of these new variables the equation of motion is obtained by @xcite @xmath102 } { \delta \varphi_{\delta}}|_{\varphi_{\delta } = 0 } = 0 ~~ , \label{motion}\ ] ] where @xmath103 are stochastic fields , related to each distinct dissipative kernel appearing in ( [ motion ] ) . at 1-loop order ,
the leading contributions to the dissipative terms in the equation of motion come from the diagrams : @xmath104 $ } \end{picture } \label{leading}\ ] ] the explicit expression corresponding to these terms appearing in the effective equation of motion , eq .
( [ motion ] ) , is ( as obtained in @xcite for a similar case ) @xmath105_{x , x'}^2 + \sum_{j=1}^n 2 g_j^4 { \rm i m } \left[g_{\chi_j}^{++}\right]_{x , x'}^2 \right\ } \theta(t - t ' ) } \nonumber \\ & & \sim \varphi_c^2 ( t ) \dot{\varphi}_c ( t ) \ ; \left\ { \frac{\lambda^2}{8 } \beta \int \frac{d^3q}{(2 \pi)^3 } \frac{n_\phi(1 + n_\phi)}{\omega_\phi^2 ( { \bf q } ) \gamma_\phi ( { \bf q } ) } + \sum_{j=1}^n \frac{g_j^4}{2 } \beta \int \frac{d^3 q}{(2 \pi)^3 } \frac{n_{\chi_j } ( 1 + n_{\chi_j})}{\omega_{\chi_j}^2 ( { \bf q } ) \gamma_{\chi_j } ( { \bf q } ) } \right\ } \nonumber\\ & & + { \cal o}\left(\lambda^2 \frac{\gamma_\phi}{\omega_\phi}\right ) + { \cal o } \left(g_j^4 \frac{\gamma_{\chi_j}}{\omega_{\chi_j } } \right ) \nonumber \\ & & + \varphi_c^3 ( t ) \int_{-\infty}^t d t ' \int \frac{d^3 q } { ( 2 \pi)^3 }
\left\ { \frac{\lambda^2}{2 } { \rm i m } \left[g_\phi^{++}({\bf q},t - t')\right]^2 + 2 \sum_{j=1}^n g_j^4 { \rm i m } \left[g_{\chi_j}^{++}({\bf q},t - t')\right]^2 \right\}\ : , \label{dissterm}\end{aligned}\ ] ] where in the lhs of the above equation , we used the compact notation @xmath106_{x , x'}^2 = \int \frac{d^3 k}{(2 \pi)^3 } \exp\left [ i { \bf k } .
( { \bf x } - { \bf x}')\right ] \int \frac{d^3 q}{(2 \pi)^3 } g_{\phi,\chi_j}^{++}({\bf q } , t - t')g_{\phi,\chi_j}^{++ } ( { \bf q } - { \bf k},t - t ' ) \ ; , \label{g2}\ ] ] with @xmath107 obtained from ( [ g of k ] ) and ( [ g > < ] ) . in the rhs of ( [ dissterm ] ) , we have taken the limit of homogeneous fields , for details see the appendix b. we have also made use of the approximation for slowly moving fields : @xmath108 . in the next section we show that this approximation is consistent with strong dissipation .
after performing the time integration and retaining the leading terms in the coupling constants , we obtain the result given in ( [ dissterm ] ) . the last term , proportional to @xmath109 , will correspond to the finite temperature correction to the quartic @xmath0 self - interaction ( see appendix b ) .
the final equation of motion , at leading order in the coupling constants , at high temperatures ( @xmath110 ) and in the adiabatic limit , can then be written as @xmath111 where @xmath57 is given by ( [ mt ] ) , @xmath112 is the temperature - dependent effective ( renormalized ) quartic coupling constant ) . the apparent instability from these terms for high
@xmath113 is only an artifact of the loop expansion .
as shown in @xcite for the @xmath114 model , once higher order corrections are accounted for , @xmath112 is always positive even in the @xmath115 limit . using full dressed propagators we are automatically taking into account these higher order corrections , through the appearance of thermal masses in ( [ lambdat ] ) .
however , in the multi - field case there is the possibility of vacuum instability due to the @xmath0 couplings to the @xmath9 fields .
this appears as a constraint in our estimates below . ] : @xmath116 + { \cal o}\left(\frac{m}{t}\right ) \right\ } - \nonumber \\ & - & 6 \sum_{j=1}^n g_j^4 \left\ { \frac{t}{8 \pi \mu_j(t ) } + \frac{1}{8 \pi^2 } \left[\ln\left(\frac{\mu_j(t)}{4 \pi t } \right ) + \gamma \right ] + { \cal o}\left(\frac{\mu_j(t)}{t}\right ) \right\}+ \nonumber \\ & + & { \cal o}(\lambda^3 , g^4 f ,
\lambda^2 g^2 , g^6 ) \ : , \label{lambdat}\end{aligned}\ ] ] in ( [ eq motion ] ) , @xmath117 is a stochastic field associated with the imaginary terms in the effective action coming from the real - time evaluation of the diagrams ( [ leading ] ) .
its two - point correlation function is given by @xcite @xmath118_{x , x'}^2 + 2 \sum_{j=1}^n g_j^4 { \rm re}\left[g_{\chi_j}^{++}\right]_{x , x'}^2\ : .
\label{corr noise}\ ] ] note that since we are considering homogeneous field configurations , @xmath117 is a space uncorrelated stochastic field , but it is colored ( time dependent ) and gaussian distributed , with probability distribution given by ( @xmath119 is a normalization constant ) @xmath120 = n_1^{-1 } \exp\left\ { - \frac{1}{2 } \int d^4 x d^4 x ' \xi_1 ( x ) \left [ \frac{\lambda^2}{2 } { \rm re } \left[g_\phi^{++}\right]_{x , x'}^2 + 2 \sum _ j g_j^4 { \rm re } \left[g_{\chi_j}^{++}\right]_{x , x'}^2 \right]^{-1 } \xi_1 ( x ' ) \right\ } \ : .
\label{p1}\ ] ] as shown in @xcite , the dissipative coefficient in ( [ eq motion ] ) , written explicitly in eq .
( [ disscoef ] ) below , and the noise correlation function eq .
( [ corr noise ] ) ( in the homogeneous limit ) , are related by a fluctuation - dissipation expression valid within our approximations ( 1-loop order at @xmath121 and for @xmath122 ) : @xmath123 in @xcite it was also shown that as @xmath124 , @xmath125 , and the integrand in ( [ corr noise ] ) becomes sharply peaked at @xmath126 . in this limit , we can obtain an approximate markovian limit for ( [ corr noise ] ) .
we can read the dissipation coefficient @xmath127 , which appears in eq .
( [ eq motion ] ) , from ( [ dissterm ] ) decay rate coming from the @xmath128 interaction was left out . here
we give the correct expressions for @xmath42 , @xmath129 and for the dissipation .
] , @xmath130 for the model we are interested in , with lagrangian density given by ( [ nfields ] ) , with a large number of @xmath29 fields coupled to @xmath0 , and for @xmath70 and @xmath131 , we can use the obtained expressions for @xmath42 and @xmath43 , to show that @xmath132 . since the dissipation coefficient , eq .
( [ disscoef ] ) , goes as @xmath133 , @xmath43 will give the dominant contribution to @xmath127 .
an explicit expression for @xmath127 , can be obtained by using the @xmath69 approximation for @xmath134 and @xmath135 , or , equivalently , eqs .
( [ gammaphi ] ) and ( [ gammachi ] ) for @xmath42 and @xmath43 , respectively , in eq .
( [ disscoef ] ) . at the high temperature limit , @xmath136 and for @xmath54
, we then obtain the following approximate expression for @xmath127 ( using @xmath137 , for @xmath64 ) @xmath138 } \ln \left ( \frac{2t}{m_t}\right ) \right . \nonumber \\
& + & \left .
\sum_{j=1}^n \frac{4 g_j^4}{f_j^2 + 8 g_j^4 \left[1- \frac{6}{\pi^2 } { \rm li}_2 \left(\frac{\mu_j ( t)}{m_t } \right ) \right ] } \ln \left ( \frac{2t}{\mu_j ( t)}\right ) \right\ } \ : .
\label{aproxdisscoef}\end{aligned}\ ] ] in order to test the validity of the above approximate expression for @xmath127 , we have computed it numerically .
the two expressions are shown in fig .
2 , for @xmath70 , @xmath131 , and @xmath139 , where , for simplicity , we have also considered @xmath140 and @xmath141 for all @xmath9 fields ( @xmath142 ) .
we see that the above approximation for @xmath127 fits reasonably well the full expression for the dissipation coefficient in the high temperature region , having a @xmath143 discrepancy for @xmath144 .
it is interesting to note the close relation of the above expression for the dissipation coefficient with that obtained for the shear viscosity evaluated , _
e.g. _ , from a kubo formula @xcite : @xmath145 \rangle \ ; , \label{shear1}\ ] ] where @xmath146 , with @xmath147 the space components of the energy - momentum tensor . in our case , with lagrangian given by ( [ nfields ] ) , @xmath148 in order to compute the shear viscosity in ( [ shear1 ] ) to lowest order , we must evaluate the diagrams ( [ leading ] ) , which , as shown in @xcite , have near on - shell singularities coming from the product of ( bare ) propagators .
these singularities are softened once explicit lifetimes for excitations are included through dressed propagators . taking this into account
, we obtain the following expression for the shear viscosity @xmath149 ( in analogy with the evaluation of @xmath149 in the @xmath50 single field case ) , @xmath150 \;. \label{shear2}\ ] ] compare the above expression with ( [ disscoef ] ) .
the evaluation of ( [ shear2 ] ) leads to the standard result for the shear viscosity being proportional to @xmath151 and inversely proportional to the coupling constants .
however , eq . ( [ shear2 ] ) , as shown by jeon in @xcite , does not represent the unique contribution to @xmath152 at this order of coupling constants . due to the near on - shell singularities and the way they are regulated by the thermal width
, there is an entire class of diagrams , called ladder diagrams ( diagrams with insertions of loops between the two propagators in ( [ leading ] ) ) , contributing to @xmath149 at the same order . by using a formal resummation of vertices ,
jeon was able to perform the summation of the whole set of ladder diagrams in the simple @xmath50 theory , showing that the true value of the shear viscosity is about four times larger than the one loop result in the high temperature limit . since our expression for the dissipation coefficient exhibits the same properties of @xmath149 , we expect that these higher loop ladder diagrams will also give a significant contribution to the value of @xmath127 in ( [ disscoef ] ) .
however , as we are dealing with the more complicated situation of several interacting fields , we will not attempt here to evaluate these contributions .
from the example of the shear viscosity calculation in the single field case , these ladder contributions will only add to the one - loop result for the dissipation coefficient , not changing qualitatively our results .
thus , eq . ( [ disscoef ] ) represents , at least , a _ lower bound _ for the dissipation , applicable in the strong dissipation regime , as we will show next .
we now investigate the validity and limits of applicability of our main approximations , in particular the adiabatic approximation . in order to arrive at the expression for the dissipation , eq .
( [ disscoef ] ) , and to write the equation of motion for @xmath153 as in eq .
( [ eq motion ] ) , we assumed that the field @xmath153 changes adiabatically [ see ( [ dissterm ] ) ] : @xmath154 this approximation for the field configuration has recently been the focus of some attention in the literature @xcite .
the authors in @xcite , working with soft field modes set by a coarse graining scale @xmath155 , showed that the adiabatic approximation breaks down once the field configurations ( soft modes ) oscillate with the same time scale as the dissipative kernels ( with time scale given by @xmath156 ) .
however , here we work in a very different context .
we are mainly concerned with the overdamped motion of the homogeneous field configuration @xmath153 , _ i.e. _ , when its oscillatory motion is suppressed .
therefore , the dynamic time - scale for @xmath153 must be much larger than the typical collision time - scale ( @xmath157 ) .
note that this is a much stronger condition than the simple requirement that the field should change slowly in time , with time scale set by the frequency @xmath158 .
thus , we must examine when the condition @xmath159 is satisfied .
we choose @xmath160 as the smallest of the two thermal decay widths @xmath161 , as it will set the largest time - scale for collisions for the system in interaction with the thermal bath .
note that in the evaluation of the dissipation coefficient in ( [ dissterm ] ) , the leading contribution to the first time derivative of @xmath153 is of order @xmath162 .
as discussed earlier in connection with the shear viscosity coefficient , the dependence of the dissipation coefficient on the decay width @xmath160 comes from using it as the regulator of on - shell singularities present in ( [ leading ] ) at first order in the time derivative . in appendix
b we present an argument justifying the need of regulating with the decay width and also compute the next order contribution in the adiabatic approximation , showing the consistency of the results . since the stronger the dissipation the more efficient the adiabatic approximation , the parameter range where ( [ kinetic ] ) is valid leads naturally to the regime where @xmath153 undergoes overdamped motion ( in the sense of eq .
( [ over ] ) below ) .
if we consider the ensemble average of the equation of motion ( [ eq motion ] ) : @xmath163 } { \delta \varphi_{\delta}}|_{\varphi_{\delta } = 0 } \right\rangle = 0 \ ; , \label{aver}\ ] ] where @xmath164 means average over the stochastic fields , then we define the overdamped regime when the ( averaged ) background configuration @xmath153 satisfies @xmath165 we also restrict our study to the high - temperature and ultra - relativistic region : @xmath166 . we take the couplings @xmath167 , @xmath15 such that @xmath168 . also , for simplicity , as before , we take all @xmath169 . at high temperatures we can then write for ( [ mt ] ) and ( [ mut ] ) ( @xmath170 ) the expressions @xmath171 where @xmath8 is the number of fields coupled to @xmath0 , and @xmath172 we will examine the condition for strong dissipation with overdamped motion for three particular choices of parameters , showing that there is a region of parameter space consistent with this regime . using ( [ over ] ) , we can write the equivalent expression for ( [ kinetic ] ) : @xmath173 in the estimates below , we evaluated both @xmath127 [ from eq .
( [ disscoef ] ) ] , and @xmath174 ( computed at @xmath175 ) numerically .
the three cases analyzed are : * case 1 : * @xmath176 : in this case we obtain that @xmath177 @xmath178 note that the last condition is written as a constraint for the positivity of @xmath112 . with these values and for the case @xmath139
, we obtain the results shown in fig .
3a , where we have plotted both sides of eq .
( [ kinetic2 ] ) .
the region of parameters satisfying eq .
( [ kinetic2 ] ) is given by the intersection of the region below the solid lines ( the function @xmath129 ) with the region above the dashed line ( @xmath179 computed for different values of @xmath153 ) . * case 2 : * @xmath180 : as above , this is shown in fig .
3b . the region satisfying eq .
( [ kinetic2 ] ) is given again by the intersecting region below the solid line and above the dashed lines .
in both figs . 3a and 3b ,
the results are shown up to the value of @xmath57 satisfying the constraint for the positivity of @xmath112 .
* case 3 : * @xmath181 : this case follows a slightly different philosophy , of fixing the corrected coupling as opposed to the bare coupling .
we have , @xmath182 @xmath183 with the additional constraint , @xmath184 the results for this case are shown in fig .
3c , with the same interpretation as in cases 1 and 2 : the region satisfying eq .
( [ kinetic2 ] ) is given by the intersecting region below the solid line and above the dashed lines .
the results are shown up to the value for @xmath57 satisfying the condition ( [ lambcond ] ) .
we note that the case @xmath185 is the one with the broadest range of validity in parameter space , as seen in fig .
3c , followed by the case @xmath186 , shown in fig .
3b . for @xmath187 ,
the condition for adiabaticity is only possible for fairly large field amplitudes , which may be beyond the validity of a perturbative evaluation of the effective action .
we will come back to this issue in the next section . in any case
, we stress that there are several regimes where the adiabatic approximation is valid . in all cases , the smaller @xmath8 the smaller the region of parameters that satisfies ( [ kinetic ] ) .
in particular , for @xmath188 , we find no parameter range satisfying ( [ kinetic ] ) and therefore , the adiabatic approximation .
this is consistent with the intuition that dissipation is caused by the decay of the @xmath0 field into @xmath29 fields and is more efficient the larger the number of decay channels available .
we also obtain that @xmath153 is always somewhat large ( @xmath189 ) for the range of physical parameters satisfying ( [ kinetic ] ) , for both cases analyzed , being even higher for case 1 .
if in ( [ kinetic2 ] ) we use @xmath42 instead of @xmath43 , the region of parameters improves considerably ; since @xmath132 for large @xmath8 , it allows much smaller values of @xmath190 .
it should be recalled that @xmath42 determines the relaxation time scale for the @xmath0 field . finally ,
as discussed earlier , the expression we quoted for @xmath127 gives only a lower bound for the dissipation coefficient .
as in the case studied by jeon in @xcite , higher loop ladder diagrams can lead to a considerably higher value for @xmath127 . for several interacting fields , simple estimates
show that these ladder diagrams scale at most as @xmath8 .
therefore , they may well be of the same order as the leading order 1-loop contribution to the dissipation coefficient , given by the @xmath29-sector .
we leave a more detailed analysis of the contributions coming from ladder diagrams to a future work .
additional contributions to the dissipation coefficient in ( [ disscoef ] ) only improve our estimates , enlarging the region of parameter space satisfying the adiabatic approximation ; the ratio @xmath190 decreases , broadening the conditions under which the field undergoes overdamped motion ( strong dissipative regime ) .
it is worth mentioning that the @xmath128 coupling constant in eq .
( [ nfields ] ) can be negative and this also leads to interesting results . as an illustrative example , consider an even number of @xmath29-fields with the sign of the @xmath128 coupling distributed so that @xmath191 and @xmath192 . in order that the potential be strictly positive
, it requires @xmath193 which for large @xmath8 implies @xmath194 . in the alternating sign regime ( asr )
the thermal masses are @xmath195 and @xmath196 following an analysis similar to above and for case 3 ( @xmath197 ) , we find a solution regime within the perturbative amplitude expansion , @xmath198 , @xmath199 , for @xmath200 and @xmath201 .
for example , these conditions are satisfied for @xmath202 , @xmath203 . in this example
@xmath204 , but this can be modified in several ways .
in general , when the @xmath128 couplings are distributed between positive and negative strengths , it controls the growth of @xmath57 due to the cancellation of thermal mass contributions from the @xmath29-fields . restricting the magnitude of @xmath57 , in turn , increases the parameter regime and duration of overdamped motion .
this example demonstrates another regime of overdamped motion in our model for small field amplitudes @xmath205 .
the fact that overdamping in ( [ eq motion ] ) for much of the parameter space demands large field amplitudes , at least within the approximation scheme used here , is a direct consequence of having a field dependent dissipation @xmath206 . since in ( [ graphs1l ] ) we are considering a perturbative expansion for the 1-loop effective action in the field amplitudes ( that is , in powers of @xmath207 and @xmath208 ) , the need for large field amplitudes may place doubts on the validity of our calculations for a considerable portion of the parameter space .
below we address this issue in two different ways ; first by comparing our results with an improved one - loop approximation and then by using the subcritical bubbles method @xcite to test the validity of the effective potential for large - amplitude fluctuations .
we start by computing the analog of ( [ over ] ) in the context of the whole one - loop approximation , _
i.e. _ , when @xmath207 and @xmath208 in ( [ trln ] ) are taken as part of field - dependent masses . for this , let us give an alternative computation of the evolution equation for @xmath153 in terms of the tadpole method of weinberg @xcite : in the shifting of the scalar field , @xmath209 , the requirement @xmath210 leads , at the one - loop order , to the evolution equation for @xmath153 ( for homogeneous fields ) @xmath211 where @xmath212 and @xmath213 are given in terms of the coincidence limit of the ( causal ) two - point green s functions @xmath214 and @xmath215 , respectively , which satisfy , in the fully dressed propagator form ( see , e.g. , ringwald in @xcite ) @xmath216 g^{++}_\phi ( x , x ' ) + \int d^4 z \sigma_\phi ( x , z ) g^{++}_\phi ( z , x ' ) = i \delta ( x , x ' ) \label{gphi}\ ] ] and @xmath217 g_{\chi_j}^{++ } ( x , x ' ) + \int d^4 z \sigma_{\chi_j } ( x , z ) g^{++}_{\chi_j } ( z , x ' ) = i \delta ( x , x')\ ; , \label{gchi}\ ] ] where , in ( [ gphi ] ) and ( [ gchi ] ) , @xmath218 and @xmath219 are the ( causal ) self - energies for the @xmath0 and @xmath9 fields , respectively . by expressing @xmath220 and @xmath221 in terms of mode functions
, we can then evaluate the averages in ( [ eqnew ] ) .
an explicit expression can be obtained in the approximation ( equivalent to the adiabatic approximation ) @xmath222 and @xmath223 , for which there is a wkb solution for the mode functions of the fields . in this paper , however , we will not carry out this calculation .
a detailed study of this , in the context of an expanding background and along the proposals made in the next section , will be presented in a forthcoming paper . for
now , we can present the result of this calculation , by using the simplest formulation proposed in @xcite , based on a relaxation - time approximation of the kinetic equation , for the calculation of the averages in ( [ eqnew ] ) .
we can then show that the ( ensemble averaged ) evolution equation for @xmath224 can be expressed , in the quasi - adiabatic approximation ( hydrodynamical regime of @xcite ) , by @xmath225 where @xmath226 , is the field derivative of the 1-loop effective potential , @xmath227 where @xmath228 and @xmath229 are the field dependent frequencies , with masses given in terms of the thermal ones , eqs .
( [ mt ] ) and ( [ mut ] ) . also , @xmath127 in ( [ eqveff ] ) is the same as in ( [ disscoef ] ) , but now with the masses replaced by the field dependent ones . in terms of ( [ eqveff ] ) , in the overdamping approximation
, the condition ( [ kinetic ] ) becomes @xmath230 using eq .
( [ kin2 ] ) , in the high temperature approximation for the fields , @xmath231 , we can show that the results obtained earlier , in terms of the amplitude expansion for the effective action , for instance , the results expressed in fig . 3 ( with @xmath57 replaced with the field dependent mass @xmath232 ) , remain approximately the same , for the cases where @xmath233 .
thus , at least for these values of the field amplitude , higher order corrections do not add to the effective potential .
in other words , at leading order in the high - temperature expansion , the field derivative of @xmath234 can be just expressed as in ( [ over ] ) , @xmath235 .
we can also address the issue of high - amplitude fluctuations by adopting a method suggested in ref .
@xcite , where it was applied to test the validity of the 1-loop approximation to the electroweak effective potential .
we note that the results from this approach are entirely consistent with nonperturbative computations based on lattice gauge theories performed by k. kajantie _
_ @xcite .
the interactions of the field @xmath19 with a thermal environment will promote fluctuations around the perturbative vacuum .
the subcritical bubbles method models these fluctuations as unstable spherically - symmetric configurations with a distribution of sizes and amplitudes .
for details see refs .
@xcite . using a distribution function for these configurations , it is possible to compute the rms amplitude of the fluctuations @xcite , @xmath236 where @xmath237 is the thermal average defined in ref . @xcite .
for the effective potential of eq .
( [ over ] ) , we obtain , @xmath238 since the perturbative approach for the computation of the effective potential relies on a saddle - point approximation to the partition function , it will only be valid for small - amplitude fluctuations about the perturbative vacuum . for potentials which exhibit spontaneous symmetry breaking ,
it is customary to choose the maximum amplitude to be at the inflection point , @xmath239 . here
, since we have a potential with positive - definite curvature , we will conservatively assume that the perturbative expansion is valid for fluctuations dominated by the quadratic term of the effective potential , that is , for @xmath240 the condition for the validity of the 1-loop approximation for the effective potential is then written as @xmath241 it is straightforward to apply this condition to the 3 cases analysed above . since case 3 is the one with a larger range of parameters satisfying the adiabatic condition , we use it as an illustration . from eqs .
( [ 3lam ] ) and ( [ 3mass ] ) , we can write eq .
( [ subcond ] ) , after dividing by @xmath242 , as the inequality @xmath243^{1/2 } > { { g^4}\over { 144\sqrt{6}}}~,\ ] ] where , @xmath244 .
this condition is easily satisfied for a large range of parameters . in particular , for @xmath245 , which are values inside the region of parameters allowed for overdamping shown in fig .
3c for @xmath246 , we obtain , @xmath247 and @xmath248 , well within the range of validity of the small - amplitude approximation .
we thus conclude that it is possible to attain the adiabatic limit of strong dissipation within the 1-loop approximation scheme adopted here .
the calculation in secs .
ii - iv presented a microscopic quantum field theory model of strong dissipation in minkowski spacetime .
this section addresses the application of this calculation to the cosmological warm inflation scenario .
although we will not present a detailed extension of our previous results to an expanding spacetime , we will argue that most of the modifications are quite straight - forward up to the requirements for the warm inflation scenario .
consider the standard friedmann cosmology with robertson - walker metric @xmath249.\ ] ] we restrict our analysis to flat space , @xmath250 , and quasistatic de sitter expansion , @xmath251 for notational convenience , the origin of cosmic time is defined as the beginning of our treatment . for this metric , the minimally coupled lagrangian for the model in eq .
( [ nfields ] ) is @xmath252 \right .
\nonumber \\ & & - \left .
\sum_i \frac{g_i^2}{2 } \chi^2_i({\bf x},t ) \phi^2({\bf x},t ) \right\}. \label{langran}\end{aligned}\ ] ] there exists an alternative derivation of the ensemble average of eq .
( [ eq motion ] ) , which was presented for a single scalar field , as an intuitive argument in @xcite . in the context of a single scalar field
, the method is to work directly with the operator equation of motion for @xmath253 .
the operator @xmath0 is reexpressed as the sum of a c - number @xmath254 , representing the classical displacement , plus a shifted operator @xmath255 , @xmath256 with @xmath257 .
a thermal average is taken of this equation of motion , in which thermal expectation values involving @xmath255 are computed such that @xmath254 is treated as an adiabatic parameter . to the order of perturbation theory considered in the previous sections , for the single scalar field
the intuitive derivation in @xcite gives the same effective equation of motion as the ensemble average of eq .
( [ eq motion ] ) as shown in @xcite .
no new considerations are needed to apply this intuitive derivation to the model in eq .
( [ nfields ] ) .
the @xmath258 fields are treated as quantum fluctuations similar to @xmath255 . from the treatment in @xcite
, it follows that the expressions for @xmath259 and @xmath127 in eq .
( [ eq motion ] ) will arise from the thermal averages , @xmath260 and @xmath261 , taken with respect to the instantaneous background @xmath254 .
although the approach in @xcite immediately isolates the dissipative term and the finite temperature renormalizations at the level of the equation of motion , it is not systematic to all orders . furthermore , it can not treat noise and it is valid only in the adiabatic approximation .
these limitations can be accounted for in the closed - time - path formalism used in this paper .
a recent work @xcite has discussed some of the difficulties associated with extending this formalism to an expanding background in order to treat noise and dissipation . our goal at present is more modest . as an easier first step ,
the intuitive derivation of @xcite is extended to an expanding background .
the exact operator equations of motion from the lagrangian eq .
( [ langran ] ) are @xmath262 and @xmath263 the objective is to displace the operator @xmath253 by a x - independent c - number at time @xmath264 , @xmath265 , and then determine the evolution of the expectation value @xmath266 by solving eqs .
( [ phieq ] ) and ( [ chieq ] ) perturbatively .
thus @xmath253 is reexpressed as eq .
( [ phiceta ] ) . with this definition of @xmath254 , for flat ,
@xmath250 , nonexpanding , @xmath267 , spacetime , the resulting equation of motion is the same as the ensemble average of the equation of motion , eq .
( [ eq motion ] ) . for the case of expanding spacetime , @xmath268 , in order to obtain the equation of motion for @xmath254 , thermal expectation values
must be taken of eqs .
( [ phieq ] ) and ( [ chieq ] ) .
provided the temperature , @xmath269 , of the thermal bath is time independent , ( _ i.e. _ , rapid equilibration time scales ) , thermal expectation values of terms linear in @xmath253 can be replaced by @xmath254 , just as for the nonexpanding case . in evaluating @xmath270 and @xmath271 ,
if the characteristic time scale for the quantum fluctuations is much faster than the expansion time scale , @xmath272 , the calculation is no different from the minkowski space situation .
this criteria is self - consistently satisfied provided @xmath273 where the left hand side is given in eqs .
( [ gammaphi ] ) and ( [ gammachi ] ) for our model .
these arguments suggest that at leading nontrivial order , the effective equation of motion for @xmath254 in an expanding de sitter spacetime , under the same conditions required for eq .
( [ eq motion ] ) plus the additional condition eq .
( [ ggth ] ) is @xmath274{\dot \varphi}_c(t ) + m_t^2 \varphi_c(t)+ \frac{\lambda_t}{6 } \varphi^3_c(t)=0 .
\label{phiheq}\ ] ] further justification that eq .
( [ phiheq ] ) is the appropriate replacement of eq .
( [ eq motion ] ) , for expanding de sitter space , can be obtained from @xcite , where an effective equation of motion similar to eq .
( [ phiheq ] ) was obtained for a model like eq .
( [ langran ] ) . however , the coupling between fields was linear , @xmath275 , which is analytically much more tractable than the present case of quadratic coupling , @xmath276 .
the entire discussion above assumes that the temperature has a well defined meaning in an expanding background .
furthermore , eq . ( [ phiheq ] ) has been motivated under the restriction eq .
( [ ggth ] ) .
as will be discussed next , condition ( [ ggth ] ) is a specific example of a general microscopic property argued in @xcite to be a necessary condition for warm inflation . as such
, when ( [ phiheq ] ) is applied to the warm inflation scenario , condition ( [ ggth ] ) imposes no additional restriction .
the warm inflation picture requires that an order parameter , in a strongly dissipative regime , slowly rolls down a potential , liberating vacuum energy into radiation energy , @xmath277 . the nonisentropic expansion which underlies warm inflation imposes that the rate of radiation production is sufficient to compensate for red - shift losses due to cosmological expansion , @xmath278 to give meaning to temperature
, the newly liberated radiation must thermalize at a scale @xmath279 which is faster than the expansion scale , @xmath280 minimally this requires an energy transfer rate from vacuum to radiation that is faster than the expansion rate , which in our model implies the condition ( [ ggth ] ) .
thus , eq . ( [ rholth ] ) is necessary to justify a temperature parameter @xmath113 , which , combined with condition ( [ ggth ] ) , are sufficient to justify the arguments leading to eq .
( [ phiheq ] ) . to completely justify a temperature parameter for an expanding background spacetime
, it is required studying the thermalization of the radiation , once it is liberated .
general arguments , as well as specific calculations @xcite at high temperature , indicate that this rate is set by the temperature @xmath281 , for some appropriate , model dependent , coefficient @xmath282 .
this minimally requires @xmath283 .
however , @xmath282 may be very small , as for example in eqs .
( [ gammaphi ] ) and ( [ gammachi ] ) .
thus the correct constraint is @xmath284 this problem will not be considered further here .
( [ ggth ] ) will be our only criteria for thermalization .
this is equivalent to assuming that the thermalization rate is at least as fast as the energy transfer rate .
( [ phiheq ] ) is accepted as the macroscopic equation governing the evolution of the order parameter @xmath254 , it can be used as a given input to construct warm inflation scenarios as in @xcite .
the microscopic origin of the equation can be forgotten up to restrictions on parameters and the self consistency condition eq .
( [ ggth ] ) . for a general equation like eq .
( [ phiheq ] ) , the warm inflation scenario requires the strong dissipative regime @xcite : @xmath285 { \dot \varphi}_c \gg { \ddot \varphi}_c \ ; , \label{ddpledp}\ ] ] with @xmath286 for our model . for the derivation in secs .
ii - iv , where @xmath267 , this condition is sufficient to satisfy the adiabatic condition , eq . ( [ kinetic ] ) , which is required for the consistency of the microscopic calculation . as such , this model provides an example of a general point conveyed in @xcite , that warm inflation defines a good regime for application of finite temperature dissipative quantum field theory methods .
the study of warm inflation in @xcite also found that to satisfy observational constraints on the expansion factor , it requires @xmath287 thus , warm inflation is an extreme example of dissipative dynamic during de sitter expansion . as demonstrated in @xcite , dissipation is generally prevalent during inflation . the microscopic model in this paper could be used to examine the general case , but then the condition ( [ eggh ] ) can be relaxed . here
, only the warm inflation regime will be further examined .
thus in the limit given by eq .
( [ eggh ] ) and based on the remaining discussion in this section , the equation of motion for the order parameter @xmath254 in our model for the warm inflation scenario turns out to be eq .
( [ over ] ) but with the additional constraint eq .
( [ ggth ] ) .
the other input for constructing warm inflation scenarios is the free energy in the expanding environment for the model ( [ langran ] ) .
it already has been argued above that temperature is a good parameter for describing the state of the radiation in the warm inflation regime .
it also follows from the above that the change in temperature can be treated adiabatically in the thermodynamic functions , since this requires @xmath288 , which is automatically satisfied due to eq .
( [ rholth ] ) .
therefore the free energy density should be well represented by the minkowski space expression , with temperature treated as an adiabatic parameter .
for the model in secs .
ii - iv , the free energy density is @xmath289 where the factor @xmath290 , in the last term , comes from the functional integration over the @xmath29 fields and the @xmath0-field s fluctuations .
having established this to be the free energy density for the warm inflation scenario , the other thermodynamic functions such as pressure , energy density and entropy density can be easily obtained . with the free energy ( [ freeeng ] ) and the order parameter equation of motion , eq .
( [ phiheq ] ) , determined , the time evolution of the three unknowns : temperature @xmath291 , scale factor @xmath292 and order parameter @xmath254 , can be obtained from eq .
( [ phiheq ] ) plus any two independent equations from friedmann cosmology along with a self consistency check for adiabaticity , eq .
( [ ddpledp ] ) . at this point
the procedure in @xcite can be followed .
however due to the microscopic origin of this model , additional self consistency checks are necessary for adiabaticity , given by eq .
( [ kinetic ] ) and thermalization , eq .
( [ ggth ] ) .
observationally interesting expansion factors will require @xmath293 , in which case the condition ( [ ggth ] ) immediately implies the microscopic adiabatic condition ( [ kinetic ] ) .
up to this point , the formulation of warm inflation in conjunction with a microscopic dynamics has been general . in the remainder of this section
, some demonstrative calculations of this cosmology will be presented based on our microscopic model .
an exhaustive analysis of the parameter space will not be performed . in this first examination ,
the emphasis is to understand the interplay between the microscopic and macroscopic physics of warm inflation for generic potentials , which in particular , have curvature scale of order the temperature scale . for such potentials , thermal fluctuations that displace @xmath294 substantially from the origin
are exponentially suppressed .
however , it is such fluctuations that allow enough time , during the roll down back to the origin , for the universe to inflate sufficiently . as such , this elementary fact , in any case , quells significant interest in comparing the cases we will examine to observation .
it should be noted that the order parameter in this symmetry restored warm inflation regime is configured similar to those in the chaotic inflation scenario @xcite .
however , in the chaotic inflation scenario the potentials are ultra - flat .
such potentials permit large fluctuations of the order parameter and in fact prefer them . the dissipative model in this paper
could be studied for the case of ultra - flat potentials , perhaps motivated by supersymmetric model building .
this would extend the pure quantum mechanical , new - inflation type dynamics of chaotic inflation into the intermediate regime discussed in @xcite .
this will not be examined here . proceeding with our demonstrative examination of warm inflation ,
let the origin of time be the beginning of the inflationlike regime ( bi ) and also the beginning of our treatment .
the basic picture of the particular warm inflation scenario studied here is as follows . at @xmath264
the initial conditions are arranged so that the field is displaced from the origin @xmath295 , the temperature of the universe is @xmath296 and since the universe is at the onset of the inflation - like regime , by definition this means the vacuum energy density equals the radiation energy density , @xmath297 . for @xmath298 the field will relax back to the origin within a strongly dissipative regime and in the process liberate vacuum energy into radiation energy .
simultaneously , the scale factor will undergo inflation - like expansion . during the roll - down period ,
the vacuum energy first dominates until at some point it is superseded by the radiation energy . at this point
the universe smoothly exits the inflation - like regime into the radiation dominated regime . from our model in the previous sections
, we will consider the case of @xmath299 @xmath29-bosons ( @xmath300 ) with @xmath301 and @xmath302 @xmath29-bosons ( @xmath29 ) with @xmath303 .
for this model , the dissipative dynamics of @xmath153 , expressed through @xmath304 and @xmath127 , is controlled by the @xmath299 @xmath29-bosons .
the other @xmath302 @xmath29-bosons only serve as additional fields in the radiation bath . for this purpose , from eq .
( [ gammachi ] ) , for @xmath305 the @xmath29 and @xmath300 bosons will be equally effective in thermalizing the radiation energy .
in this paper we will examine this scenario in the regime @xmath306 and in the high temperature limit @xmath307 . also , for ease of presentation , we will write the expression for @xmath308 at @xmath309 . although with these simplifications the results will not be cosmologically interesting , it is a good example to demonstrate the general procedure . in this regime , the effective equation of motion for @xmath153 , from eq .
( [ over ] ) , is : @xmath310 where @xmath311 formally the friedmann cosmology for the warm inflation scenario associated with the above equation was called the quadratic limit in @xcite .
the macroscopic and microscopic requirements of warm inflation will imply various parametric constraints which are as follows .
( [ p4ggp2 ] ) will be satisfied by requiring : @xmath312 where the parameter @xmath313 has been introduced . as shown in @xcite , throughout the inflation - like period
until just before it ends , the temperature drops slightly faster than @xmath0 . as such , the thermal mass term , @xmath314 , will continue to satisfy eq .
( [ p4ggp2 ] ) given that initially it does .
the number of e - folds , @xmath315 obtained during the roll - down is , from @xcite , @xmath316 where @xmath317 with @xmath318 the planck mass .
the microscopic condition , eq . ( [ ggth ] ) , requires @xmath319 the threshold condition for inflation , @xmath320 , implies @xmath321 finally , the validity of the perturbative derivation in the previous sections will require @xmath322 where @xmath323 and @xmath324 .
( [ micro ] ) can be turned into an equality , in which case , along with eqs .
( [ p4erp2 ] ) , ( [ rne ] ) , and ( [ rverr ] ) , they determine the boundary of the allowed parameter space .
thus there are six constraining equations for the eleven quantities @xmath325 and @xmath296 .
we will let @xmath296 set the overall scale and will fix @xmath326 .
then , based on the constraint equations , this determines the remaining parameters .
in particular , we have for @xmath112 : @xmath327 this expression is suggestive of the case 3 analyzed in sec .
iv . for such , taking then @xmath328 , we get the additional parameters : @xmath329 @xmath330 and @xmath331 based on these equations , it is not difficult to find parametric regimes in which the warm inflation scenario is realized , but it is only for @xmath332 . as such , this simple case has no observational relevance .
there are a few improvements that could be made to our analysis that would increase @xmath315 .
firstly our estimates above ignore the effects of the thermal mass term , @xmath333 , on the dynamics .
its contribution to the energy density and pressure are @xmath334 and @xmath335 respectively .
thus it helps the @xmath336 term to drive inflation .
secondly , recall that the dissipative coefficient @xmath337 . in the above analysis
, we fixed @xmath338 .
however , during the roll - down , temperature does fall by a factor of order ten , which in turn would increase @xmath127 .
finally the parameter regime could be extended to include both positive and negative @xmath128 couplings , such as the asr case described in subsection iv.a . as noted there , in this regime
the duration of overdamped motion can be increased significantly within the perturbative amplitude expansion .
this directly corresponds to increasing the e - folds @xmath315 .
a more elementary modification is to extend the region of validity to larger displacements of @xmath153 .
the extension to this larger regime can be treated by a summation of the complete one - loop series as outlined in subsection iv.b .
overdamped motion for much larger displacements of @xmath153 can also be attained by a modification to the @xmath128 interaction in eq .
( [ nfields ] ) as @xmath339 in this distributed mass model ( dmm ) , a given @xmath9-field is thermally excited when its effective mass @xmath340 .
the contribution from the thermally excited @xmath29-fields to the effective dynamics of @xmath153 is similar to our calculations in sections ii - iv . as such , an effective equation of motion similar to eq .
( 4.4 ) can be obtained for this modified model .
given an appropriate distribution of mass coefficients @xmath341 along the path of @xmath153 , from an arbitrarily large initial displacement @xmath342 , @xmath254 could undergo overdamped motion along its entire path .
details will be presented elsewhere on the warm inflation scenario which considers these various cases . for
this `` symmetry restored '' case , initial fluctuations of @xmath254 , @xmath342 , are strongly suppressed with probability @xmath343 $ ] , where the volume @xmath344 .
the most optimistic initial conditions have probability @xmath345 $ ] . thus ,
unless a viable mechanism is found to justify a large enough initial value of @xmath153 , the regime investigated here may not be very relevant for practical applications of warm inflation . in any case , the microscopic dynamics of the symmetry restored regime investigated here is similar to more realistic scenarios in the symmetry broken regime , where the field has an average initial value close or identical to zero .
an important difference being that the initial state in the latter case has no boltzmann suppression .
this section has made an initial examination of treating strong dissipation from first principles during a de sitter expansion regime .
further results will be presented elsewhere as well as a calculation similar to this one , for the symmetry broken case .
in this paper a microscopic quantum field theory model has been presented , describing overdamped motion of a scalar field .
commonly , such behavior is treated phenomenologically by ginzburg - landau order parameter kinetics .
our model provides a first principle explanation of how kinetics equivalent to the ginzburg - landau type , which is first order in time , arise for inherently second order dynamical systems .
the microscopic treatment of this problem , in principle , should be well controlled , due to its fundamental reliance on the adiabatic limit , and our model exemplifies this expectation .
the calculational method for treating dissipation in this paper has one distinct difference from several other related works . in our calculation
, we consider the effect of particle lifetimes in the effective equation of motion . to our knowledge , this effect has been discussed in only a few works in the past @xcite .
ii - iv presented a general , flat - space treatment , which offers a microscopic justification to the often used limit of diffusive ginzburg - landau scalar field dynamics .
we have shown how it is possible to obtain an effective evolution for the scalar field which is first - order in time , due to its own thermal dissipation effects , interpreted microscopically as its decay into many quanta . in a sense , the field acts as its own brakes ,
the slowing of its dynamics being attributed to the highly viscous medium where it propagates , a densely populated sea of its own decay products .
the application that we considered in sec .
v was in expanding spacetime , for the cosmological warm inflation scenario .
although we did not formally derive the extension of our flat - space model of secs .
ii - iv to an expanding background , we did present heuristic arguments that validate this extension for the special needs of warm inflation .
the results of the simple analysis in sec .
v are strongly dependent on initial conditions and may be difficult to implement for models of observational interest .
nevertheless , these results will provide useful guidance both for modifications of this model and for our next study of the symmetry broken regime .
the direct significance of the present study to inflationary cosmology would be to the initial state problem @xcite in scenarios during symmetry breaking .
the initial conditions required for warm inflation in the symmetry broken case are similar to new inflation .
the requirement is a thermalized inflaton field , which at the onset of the warm inflation regime is homogeneous with expectation value @xmath346 .
although we have made no detailed application of our results to this problem , some general features are evident from the analysis in sec .
v. in particular , both the suppression of large fluctuations and thermalization are mutually consistent with strong dissipative dynamics .
many of the difficulties that have been discussed @xcite in association with the initial state problem , are eliminated in the strong dissipative regime .
in addition , the damping of fluctuations should simplify the formal problem of coupling this model to classical gravity .
thus , the strong dissipative regime appears to have the correct features both to carry the universe into an inflation - like phase and then to smoothly exit into a hot big - bang regime .
we thank r. holman for helpful discussions .
ab was supported by a department of energy grant .
mg was partially supported at dartmouth college by the national science foundation through a presidential faculty fellows award no .
phy-9453431 and by the national aeronautics and space administration grant no .
mg thanks both fermilab and the osservatorio di roma for their kind hospitality during the completion of this work .
ror was partially supported by fundao de amparo pesquisa no estado do rio de janeiro - faperj and by conselho nacional de desenvolvimento cientfico e tecnolgico - cnpq .
we now give a brief overview of the calculation of the imaginary part of the two - loop , setting sun , self - energy terms in ( [ selfphi ] ) and ( [ selfchi ] ) .
let us express generically those diagrams in terms of field propagators with masses @xmath347 and @xmath348 and external lines of type @xmath349 .
for an interaction between @xmath349 and @xmath350 fields of the form @xmath351 ( for @xmath352 , @xmath353 and for @xmath354 , @xmath355 ) the imaginary part of the two - loop sunset diagram for the @xmath349 field can be expressed by ( see for example , in ref .
@xcite , appendix g and also wang and heinz in @xcite ) : @xmath357 } \nonumber \\ & & \times \left [ 1 + n(\sigma_{l } e_{l}^t)\right ] \left [ 1 + n(\sigma_{k+l } e_{k+l}^t)\right ] \delta(e_q^s - \sigma_l e_l^t - \sigma_{k+q } e_{k+q}^s - \sigma_{k+l } e_{k+l}^t ) \ : , \hspace{2.5 cm } \label{b1}\end{aligned}\ ] ] where @xmath358 , @xmath359 is the bose - distribution function and @xmath360 is a symmetry factor : for @xmath361 , @xmath362 and for @xmath363 , @xmath364 . by expanding in the sum in @xmath365 and retaining only the on - shell , energy conserving processes ( corresponding to the scattering processes @xmath366 , @xmath367 ) , we obtain the result : with @xmath369 for @xmath370 and @xmath371 for @xmath372 .
@xmath373 for @xmath361 has been obtained in details in refs . @xcite and @xcite . in particular , wand and heinz in @xcite
have discussed and obtained in detail the kinematic limits , for the @xmath50 model , of the integration on the momenta in ( [ b2 ] ) , implicit in the dirac delta - function . here
we obtain the results for the case of @xmath372 . in ( [ b2 ] ) , by defining in the three - dimensional momentum integrations the angular differentials @xmath374 and @xmath375 ( @xmath376 , we are then able to perform the angular integrals in ( [ b2 ] ) . from the constraint in the integration limits for @xmath377 and @xmath378 , which comes from the delta - function ,
we obtain the following result for @xmath379 , @xmath380 \left [ 1 + n(e_{l}^t)\right ] \right .
\nonumber \\ &
\times & \left . { \rm ln } \left ( \frac{e^{\beta e_{k+l}^t } - 1}{e^{\beta e_{k+l}^t}- e^{\beta ( e_{l}^t- e_q^s ) } } \right ) + \left ( \int_{-q}^\infty dk \int_{u(k+q , q)}^\infty \frac{l dl}{e_l^t } - \int_{q}^\infty dk \int_{u(-k+q , q)}^\infty \frac{l dl}{e_l^t } \right ) \right . \nonumber \\ & \times & \left .
\left [ 1 + n(e_q^s - e_l^t)\right ] \left [ 1 + n(e_{l}^t)\right ] { \rm ln } \left ( \frac{e^{\beta e_{k}^s } - 1}{e^{\beta e_{k}^s}- e^{\beta ( e_{q}^s- e_l^t ) } } \right ) \right\ } \ ; , \label{b3}\end{aligned}\ ] ] @xmath382 \left(k^2 m_s^2 + 2 m_s^2 m_t^2 + 2 q^2 m_t^2 \right .
\nonumber \\ & - & \left .
2 kq m_t^2 + 2 m_t^2 \sqrt{(k - q)^2 + m_s^2 } \sqrt{q^2+m_s^2 } \right)^{\frac{1}{2 } } \right\}\;. \label{b4}\end{aligned}\ ] ] in the first term of ( [ b3 ] ) , we can make the change of integration variables @xmath383 , @xmath384 , to obtain @xmath385
. doing the same for the remaining integrals , @xmath386 , relabeling @xmath387 and @xmath388 back to @xmath378 , @xmath377 , and performing a new change of integration variables , @xmath389 , @xmath390 , we are able to compute the @xmath391 integrations . from this point on , the integrations are equivalent to the ones in @xcite , once we change the integration limits in the @xmath377 and @xmath378 integrals , and take into account the function @xmath381 , eq .
( [ b4 ] ) .
the results shown in sec .
ii for @xmath42 , @xmath43 , eqs .
( [ gammaphi ] ) and ( [ gammachi ] ) , are obtained once the limit of vanishing @xmath392 ( @xmath52 ) , is taken in the above equations .
let us now compute the expression on the lhs of ( [ dissterm ] ) in the adiabatic approximation , stressing the need for regularizing the propagators with quasi - particle lifetimes .
let us write the lhs of ( [ dissterm ] ) in terms of the adiabatic series : @xmath393_{x , x'}^2 \theta ( t - t ' ) } \nonumber \\ & & = \sum_n \int d^3 x ' \int \frac{d^3 k}{(2 \pi)^3 } e^{i { \bf k}.({\bf x } - { \bf x } ' ) } \frac{1}{n ! } \left .
\frac{\partial^n \left(\varphi_c^2 \right)}{\partial t^{\prime \ ,
n } } \right|_{t'=t } \int \frac{d^3 q}{(2 \pi)^3 } \int_{-\infty}^t d t ' ( t'-t)^n \nonumber \\ & & \times { \rm i m } \left [ g^{++ } ( { \bf q } , |t- t'| ) g^{++ } ( { \bf q } - { \bf k } , |t - t'| ) \right ] \ ; , \label{a1}\end{aligned}\ ] ] where @xmath394 can be read from ( [ g of k ] ) and ( [ g > < ] ) and it refers generically to the @xmath0 or @xmath9 field propagators .
the approximation of considering a homogeneous field , @xmath395 , is equivalent to taking the limit @xmath396 for the external momentum in @xmath397 . in this case
the @xmath398 and @xmath377 integrations in ( [ a1 ] ) can be done trivially .
however , it is known that taking the limit of zero external momentum ( @xmath399 ) @xcite in self - energy expressions , which are given in terms of products of non - local propagators , can be problematic .
this is related to the non - analyticity of these expressions around the origin .
in particular , gross , pisarski , and yaffe in @xcite argue that a correct way of taking the limit @xmath399 , at finite temperatures , is to first take @xmath400 and then @xmath396 .
they also argue that the non - analyticity problem , which comes from a failure to do a self- consistent calculation , would be eliminated once fully - dressed propagators are taken consistently into account .
we note , in particular , that for fully - dressed propagators the decay width @xmath160 works as a regulator . in ( [ a1 ] ) ,
the limit @xmath400 is implicit in the adiabatic approximation , where the fields are required to change slowly in time , while the limit @xmath396 is implemented by approximating the fields to be homogeneous . in evaluating ( [ a1 ] ) , we will first compute the time integral , expand in terms of the `` regulator '' @xmath160 , and then finally take the limit @xmath396 .
this is the opposite of what was done in obtaining eq .
( [ dissterm ] ) . from this lesson
we will see both the importance of considering fully - dressed propagators and how consistent results can be obtained once the limits are taken in the correct order
. this will be crucial when evaluating the dissipation contribution ( @xmath401 , in ( [ a1 ] ) ) .
consider the time integral on the rhs of eq .
( [ a1 ] ) and expand it to first order in the `` regulator '' , @xmath160 ( which is of order @xmath402 , @xmath403 ) .
higher order terms in @xmath160 need to be considered in conjunction with higher order loop terms , for consistency .
for @xmath404 , the zeroth - order in the adiabatic approximation , we obtain @xmath405 } \nonumber \\ & & \simeq \frac{1}{4 \omega_{{\bf q } } \omega_{{\bf q } - { \bf k } } } \left [ - \frac{1}{\omega_{{\bf q } } + \omega_{{\bf q } - { \bf k } } } + 2 \frac{\omega_{{\bf q } - { \bf k } } n(\omega_{{\bf q } } ) - \omega_{{\bf q } } n(\omega_{{\bf q } - { \bf k}})}{\omega_{{\bf q}}^2 - \omega_{{\bf q } - { \bf k}}^2 } \right ] + { \cal o } \left [ \left ( \frac{\gamma_{{\bf q } } + \gamma_{{\bf q } - { \bf k}}}{\omega_{{\bf q } } + \omega_{{\bf q } - { \bf k } } } \right)^2 \right ]
\;. \label{a2}\end{aligned}\ ] ] @xmath406 \to - \frac{\left[1 + 2 n(\omega_{{\bf q } } ) \right]}{8 \omega_{{\bf q}}^3 } - \beta \frac{n(\omega_{{\bf q } } ) \left[1 + n ( \omega_{{\bf q } } ) \right]}{4 \omega_{{\bf q}}^2 } + { \cal o } \left ( \frac{\gamma^2}{\omega^2 } \right ) \;. \label{a3}\ ] ] @xmath407 } \nonumber \\ & & \simeq - \left [ 1 + 2 n(\omega_{{\bf q}})\right ] \frac { 3 \omega_{{\bf q}}^2 + \omega_{{\bf q}-{\bf k}}^2}{\omega_{{\bf q } } \left(\omega_{{\bf q}}^2 - \omega_{{\bf q}-{\bf k}}^2 \right)^2 \left(\omega_{{\bf q } } - \omega_{{\bf q}-{\bf k } } \right ) } \frac{\gamma_{{\bf q } } + \gamma_{{\bf q } - { \bf k}}}{\omega_{{\bf q } } + \omega_{{\bf q } - { \bf k } } } + ( \omega_{{\bf q } } \rightleftharpoons \omega_{{\bf q}-{\bf k } } ) \nonumber \\ & & - \beta \gamma_{{\bf q } } n(\omega_{{\bf q } } ) \left[1 + n(\omega_{{\bf q}})\right ] \frac{1}{\left(\omega_{{\bf q}}^2 - \omega_{{\bf q}-{\bf k}}^2 \right)^2 } + ( \omega_{{\bf q } } \rightleftharpoons \omega_{{\bf q}-{\bf k } } ) + { \cal o } \left [ \left ( \frac{\gamma_{{\bf q } } + \gamma_{{\bf q } - { \bf k}}}{\omega_{{\bf q } } + \omega_{{\bf q } - { \bf k } } } \right)^3 \right ] \;. \label{a4}\end{aligned}\ ] ] we can see , in contrast with ( [ a2 ] ) , that the limit @xmath396 is divergent .
this divergence is reminiscent of the on - shell singularity which is present in the integral in ( [ a4 ] ) when bare propagators are used , thus showing the importance of @xmath160 as a regulator . by first taking the homogeneous limit @xmath396 and then expanding in @xmath160
, we obtain the result given in the text , which is the first term on the rhs in ( [ dissterm ] ) .
this result is consistent with a recent calculation in ref .
@xcite , which addresses the time derivative expansion of the effective action , for a given scalar field model .
the authors of @xcite also discuss their work in the context of the non - analyticity problem in finite temperature qft .
in fact , we note that the second order term in the adiabatic approximation can be associated with the first order term in the time derivative expansion of the effective action , @xmath410 ( for a time - dependent , space - homogeneous field configuration ) . j. schwinger , jour . math . phys .
( n.y . ) * 2 * , 407 ( 1961 ) ; p. m. bakshi and k. t. mahanthappa , jour . math .
* 4 * , 1 ( 1963 ) ; * 4 * , 12 ( 1963 ) ; l. v. keldysh , zh . eksp .
. fiz . * 47 * , 1515 ( 1964 ) ; a. niemi and g. semenoff , ann .
( ny ) * 152 * , 105 ( 1984 ) ; nucl .
b230 * , 181 ( 1984 ) .
m. gleiser , e. w. kolb , and r. watkins , nucl .
b364 * , 411 ( 1991 ) ; g. gelmini and m. gleiser , nucl . phys . *
b419 * , 129 ( 1994 ) ; m. gleiser and e. w. kolb , phys .
* 69 * , 1304 ( 1992 ) ; n. tetradis , z. phys .
* c57 * , 331 ( 1993 ) .
d. j. gross , r. d. pisarski and l. g. yaffe , rev .
* 53 * , 43 ( 1981 ) ; h. a. weldon , phys . rev . *
d28 * , 2007 ( 1993 ) ; ibid . *
d47 * , 594 ( 1993 ) ; t. s. evans , can . j. phys . * 71 * , 241 ( 1993 ) ; a. das and m. hott , phys . rev . *
d50 * , 6655 ( 1994 ) ; * figure 3 * : results for the adiabatic condition , eq .
( [ kinetic2 ] ) .
the dashed lines correspond to @xmath412 , for different values for @xmath153 .
the solid line corresponds to @xmath413 , evaluated at @xmath175 .
all cases shown are for @xmath139 .
the region satisfying the adiabatic condition is the intersection of the region above the dashed lines with the region below the solid line .
3a is for @xmath414 , fig .
3b is for @xmath415 and fig .
3c is for @xmath416 .
|
we study the conditions under which an overdamped regime can be attained in the dynamic evolution of a quantum field configuration . using a real - time formulation of finite temperature field theory ,
we compute the effective evolution equation of a scalar field configuration , quadratically interacting with a given set of other scalar fields .
we then show that , in the overdamped regime , the dissipative kernel in the field equation of motion is closely related to the shear viscosity coefficient , as computed in scalar field theory at finite temperature .
the effective dynamics is equivalent to a time - dependent ginzburg - landau description of the approach to equilibrium in phenomenological theories of phase transitions .
applications of our results , including a recently proposed inflationary scenario called `` warm inflation '' , are discussed .
pacs number(s ) : 98.80 cq , 05.70.ln , 11.10.wx in press physical review d , 1998
|
atmospheric muons are among the final products of cosmic ray induced air - showers .
the absolute muon flux and its momentum dependence are mainly determined by the flux of nucleons entering the atmosphere and the inclusive meson production cross sections in high - energy hadronic interactions .
the ratio of the fluxes of positive to negative muons , denoted as charge ratio in the following , reflects the proton to neutron ratio at the top of the atmosphere , folded with the production and decay spectra of charged pions and kaons . while the knowledge of the primary cosmic ray spectrum below a few 100 has improved considerably in the recent past @xcite
, large uncertainties still exist in the primary energy range between 0.1 and 500 responsible for the production of secondaries with momenta in the range under study here .
moreover , the details of high energy hadronic interactions still lack theoretical understanding and there is little experimental data in the relevant energy and phase space regions @xcite .
therefore the ground - level muon flux and charge ratio are widely used to tune or verify the parameters of atmospheric cascade calculations @xcite .
currently these calculations are of great interest , as they predict the absolute atmospheric neutrino fluxes @xcite which are needed to interpret the observed muon neutrino flux deficit @xcite and to evaluate the backgrounds for neutrino astronomy .
the muon flux and charge ratio have been extensively studied with different experimental techniques @xcite .
however , results show discrepancies of about 10@xmath420% with respect to each other , which exceed the uncertainties assigned to the individual measurements and thus indicate the presence of systematic effects not accounted for . here a new measurement of the atmospheric muon flux is presented using the precise muon spectrometer of the l3 detector located at the lep collider at cern , near geneva , switzerland .
special attention is given to the precise determination of all relevant detector and environmental parameters needed to convert the raw - data distributions into an absolute surface level flux .
the large statistics available permits extensive studies of the residual systematic uncertainties .
the momentum distribution of atmospheric muons is measured with the upgraded l3 setup of the l3 detector @xcite known as l3+c @xcite .
the parts of the detector used in this analysis are sketched in figure [ fig : l3c ] . after passing through the stratified rock overburden , called `` molasse ''
, the arrival time @xmath5 of a muon is measured with a resolution of 1.7 ns by a 202 m@xmath2 scintillator array placed on top of the l3 detector .
the array is composed of 34 modules , each read out by two photomultipliers in coincidence to reduce noise . inside a volume of about 1000 m@xmath6 with a magnetic field of 0.5 t ,
the coordinates and slopes of a muon track are measured in up to six drift chambers in the bending plane and up to eight times in the non - bending plane .
these chambers are arranged concentrically around the lep beam line in two groups of eight octants , each containing three layers of drift cells . by subtracting the @xmath5 time from the arrival times of the drift electrons at the sense wires , a track position in each chamber
can be reconstructed with a precision of about 60 @xmath7 m in the bending plane and 1 mm in the non - bending plane .
only three points are needed to determine the radius of the track in the magnetic field , therefore the momentum of a muon traversing two octants can be measured twice . this redundancy is used to evaluate the detector efficiencies and the resolution of the apparatus .
the best resolution is obtained when fitting the six points together over the full track length of 11 meters .
the multiple scattering and energy loss inside the l3 inner detectors , as well as the effect of the inhomogeneous magnetic field are taken into account using the procedure proposed in reference @xcite .
equipped with a trigger and data acquisition system independent of the normal l3 data - taking , l3+c recorded @xmath8 atmospheric muon triggers during its operation in the years 1999 and 2000 .
the l3+c experiment was located 450 m above sea level at a longitude of 6.02@xmath9 e and a latitude of 46.25@xmath9 n. for vertically incident muons , the mean energy loss in the molasse ( @xmath10= 6854 g@xmath11 ) and the magnet ( @xmath10=1227 g@xmath11 ) is 19 at low momenta and reaches 57 at 1 .
the geometrical acceptance of the l3+c detector and the stochastic energy loss in the molasse overburden are evaluated using the following simulation procedure : monte carlo events are generated on the surface using a parameterization of the zenith angle and momentum dependence of the muon spectrum as obtained with the ` corsika ` @xcite program .
these simulated muons are then tracked through a ` geant ` @xcite model of the l3+c environment which includes the molasse , access shafts and the concrete structures around the cavern which hosts the apparatus . finally , the detector response is simulated with a detailed ` geant ` description of the l3 detector .
the generated detector signals are reconstructed with the same program used for the data . in total @xmath12
reconstructed monte carlo events are used in this analysis .
the data analysis is restricted to events with three position measurements in at least one octant , a scintillator hit and good running conditions during data - taking .
a total of @xmath13 reconstructed muon tracks are retained .
the shielding of the 30 m of molasse overburden absorbs most of the charged air - shower particles other than muons .
the number of muons produced in @xmath14 collisions by lep is negligible compared to the flux of atmospheric muons .
therefore no background rejection is needed .
the data selection focuses on two topics .
firstly , fiducial volume cuts are defined to assure a good description of the data by the simulation .
secondly , selection cuts are imposed on the track quality to enhance the momentum and angular resolution .
these selection criteria are : * the muon track positions must be measured in six layers in the bending plane . *
the momentum resolution , calculated from the quality of the track position measurements , should not exceed its nominal value by more than 50% .
* at least four position measurements ( two in each octant ) should be present in the non - bending plane .
* the @xmath15 of a fit of the tracks to a circle within an octant must satisfy @xmath16 .
* the difference between the two photomultiplier time measurements from the same scintillator module must be below 8 ns .
after these cuts , @xmath17 data events remain for the muon spectrum analysis .
+ the selection efficiencies depend on the charge , momentum and direction of the muon . for muons with momenta above 80 , the average efficiency is 7.6% for the fiducial volume cut and 33.3% for the quality selection .
+ 4 % of the raw events are multi - muon events .
each individual muon is counted as an input to the spectrum data .
the single - octant resolution is inferred directly from the data by comparing the two independent curvature measurements of muons traversing two octants .
an example of the curvature difference is shown in figure [ fig : resoexa ] .
these measurements are used to tune the detector simulation , from which the resolution of the full fit is determined . the relative momentum resolution @xmath18 as a function of momentum at the detector - level is shown in figure [ fig : reso](a ) . and @xmath19 .
the fraction of events with width @xmath20 is denoted by @xmath21 . ] the maximum detectable momentum of the spectrometer , defined as the momentum at which @xmath18 reaches unity , is 0.78 for muons measured in only one octant and about 5 for muons measured in two octants .
the efficiency of each subdetector is studied by exploiting redundancies in the measurement process . for about 50% of the tracks , the muon arrival time is also deduced from the muon chambers .
these tracks are used to determine the scintillator efficiencies as a function of time and position on the array .
a mean efficiency of 95.6% is found at the start of data - taking decreasing continuously to 94.5% towards the end of 2000 .
the possibility of reconstructing a muon within a single octant is used to scan the drift - layer performance of the facing octant . on average , a fraction of 10.5% of the drift cells are found to have an efficiency lower than 80% .
these regions are excluded in both the data and monte carlo reconstruction . under these conditions the trigger efficiency is determined from redundant trigger classes to be 99.85% on average . during data - taking , the total effective running time was continuously measured with a 10 mhz live - time counter , which is disabled whenever the trigger system is not ready to accept new data . in addition
, each second an external trigger signal was sent to the l3+c trigger system .
the number of these external triggers on tape compared to the total number of running seconds gives another estimate of the effective running time and agrees with the value from the live - time counter within 0.02% . using the possibility to measure a muon independently in two detector parts ,
the selection efficiencies are determined in the following way : the detector is subdivided into two hemispheres , @xmath22 and @xmath23 , and the conditional hemisphere selection probabilities @xmath24 are measured for data and monte carlo separately as a function of the muon charge @xmath25 , momentum @xmath26 and zenith angle @xmath27 . in the absence of correlated inefficiencies , the total selection efficiency for accepting a track in the two hemispheres is given by the product @xmath28 . the ratio @xmath29 is used to correct the differences between data and monte carlo .
depending on the zenith angle range and the data - taking year , @xmath21 varies from 0.84 to 0.90 .
a large fraction of this correction factor originates from a defect in the tdcs used to read out the muon chambers , giving rise to an 8% inefficiency for the full track selection . the relation between the momentum distribution measured in l3+c and the muon surface spectrum
is given by @xmath30 here @xmath31 is the vector of events @xmath32 with measured momenta between @xmath33 $ ] .
the effective live - time is given by @xmath34 and @xmath35 denotes the migration matrix , _
i.e. _ the conditional probability of measuring a momentum @xmath36 given a surface momentum @xmath37 .
@xmath38 is the diagonal matrix of geometrical acceptances as a function of the surface momentum and @xmath39 is the diagonal matrix of detector efficiencies as a function of momentum at the detector - level .
the vector @xmath40 contains the true surface spectrum integrated over a surface momentum bin .
the complete detector matrix , @xmath41 , is evaluated from the measured detector efficiencies and the detector simulation as follows : @xmath42 where @xmath43 is the surface area used in the monte carlo generator , @xmath44 the solid angle of the zenith bin under study , @xmath24 includes the scintillator and trigger efficiencies and @xmath45 is the selection efficiency correction discussed above .
@xmath46 denotes the number of selected monte carlo events found within a detector - level momentum bin @xmath22 , which were generated within the momentum bin @xmath23 at the surface , and @xmath47 is the total number of monte carlo events generated within this surface momentum bin . the effective acceptance of this analysis is calculated by summing over the columns of the detector matrix , which yields the geometrical factor for a muon being registered in the detector and fulfilling the selection cuts .
the product of the effective acceptance and the live - time gives the total exposure , shown in figure [ fig : reso](b ) for positive and negative muons as a function of surface momentum .
it rapidly decreases at low energies due to the momentum cut - off caused by the molasse overburden .
below 200 , positive and negative muons have different acceptances , because the magnetic field bends their tracks in opposite directions and correspondingly into different detector regions . at large momenta
the acceptance decreases with the performance of the full detector fit and a more difficult reconstruction caused by the increasing production of delta rays .
the measurement equation ( [ eq : spec ] ) is solved using the least squares method , by minimizing @xmath48 where @xmath49 contains the statistical errors of the data and the detector matrix : @xmath50\ , m_j^2}\;. \label{eq : newerr}\ ] ] in the first step , the statistical monte carlo variances @xmath51 $ ] are set to zero , such that equation ( [ eq : lsqchi2 ] ) becomes linear with respect to the surface spectrum @xmath52 and its solution is @xmath53 with covariance matrix @xmath54=\left(\mathbf{d}^t \mathbf{w } \mathbf{d}\right)^{-1}\ , .
\label{eq : vm}\ ] ] here @xmath55 denotes the diagonal weight matrix containing the statistical errors of the data and the monte carlo , @xmath56 .
the minimization of equation ( [ eq : lsqchi2 ] ) is then repeated using the solution @xmath57 of the previous iteration for the calculation of the errors in equation ( [ eq : newerr ] ) .
this process is repeated until the maximum relative difference to the result of the previous iteration is below @xmath58 .
typically four iterations are needed .
uncertainties on the live - time and the trigger and scintillator efficiencies give rise to a normalization uncertainty of 0.7% .
the uncertainty of the detector acceptance is assessed in three studies : first , the results obtained for statistically independent data subsamples , as for instance the data collected in 1999 and 2000 or in different detector parts , are compared .
second , the muon flux and charge ratio are measured as a function of the azimuthal angle . at large momenta , geomagnetic effects and the variation of the molasse overburden are not important , and
therefore a flat distribution is expected .
finally , the stability of the measured flux and charge ratio with respect to a variation of the selection criteria is investigated . from these studies ,
additional normalization uncertainties in the absolute muon flux are derived . these range from 1.7% to 3.7% depending on the zenith angle . for the charge ratio normalization uncertainties between 1.0% and 2.3%
are estimated .
figure [ fit : rvst ] shows an example of the stability of the measured muon flux within the two data - taking years for zenith angles between @xmath59 and @xmath60 degrees .
since the muon production is known to change with atmospheric conditions @xcite , the observed variation of the muon flux is compared to an air - shower simulation with the ` target ` @xcite program using atmospheric density profiles measured in balloon flights close to the experiment @xcite .
good overall agreement between data and monte carlo is observed .
however , the full comparison to the rates in 26 weeks and 14 momentum bins yields a @xmath61 of 526/364 .
the assumption that this large value is caused by detector inefficiencies not accounted for , leads to an additional normalization uncertainty of 0.3% , which is well within the above estimated uncertainties .
+ comparison of data and monte carlo is also shown . ]
due to the steepness of the muon spectrum , even small uncertainties in the absolute momentum scale can introduce a considerable bias in the muon flux measurement . the uncertainty on the l3 magnetic field strength introduces a momentum scale bias of less than 0.4% @xcite .
furthermore , the momentum measurement is subject to uncertainties of the detector alignment .
a systematic shift of the chamber positions may introduce a constant offset @xmath62 .
the measurement of the curvature , @xmath63 , and the alignment related momentum scale uncertainty , @xmath64 is given by @xmath65 and depends on the muon charge .
within one octant , the alignment is measured by an optical alignment system @xcite with a precision corresponding to 0.19 @xmath66 @xcite .
the relative alignment of the muon chamber octants , relevant for this analysis , is determined from the data itself with a precision between 0.075 and 0.152 @xmath66 @xcite , depending on the zenith angle .
the uncertainty due the molasse overburden affects the conversion of the measured flux at the detector to the surface .
the results of two survey drillings at different locations close to l3+c provide an absolute measurement of the l3+c matter overburden .
the influence of molasse inhomogeneities and of surface installations not included in the l3+c simulation is estimated by studying the variance of the muon flux as a function of the azimuthal angle near the momentum threshold .
this leads to an uncertainty of the average rock density of 2% , which is equivalent to an energy loss uncertainty of 0.4 in the vertical direction .
good agreement between the muon energy - loss calculation used here @xcite and other approaches @xcite is found .
the residual differences correspond to a momentum scale uncertainty below or less than 0.3% in the vertical direction . the relative momentum scale uncertainties for vertically incident muons are displayed in figure [ fig : scsys](a ) . at low energies
the molasse uncertainty contributes the most , whereas above 100 the alignment uncertainties dominate .
[ fig : crats ] the limited monte carlo statistics affects the precision of the detector matrix @xmath67 .
below 200 , it dominates the total statistical uncertainty in the denominator of equation ( [ eq : lsqchi2 ] ) , contributing about 0.5% to the total uncertainty per zenith angle bin . in order to estimate the influence of the uncertainty of the momentum resolution on the measured muon flux
, the minimization of equation ( [ eq : lsqchi2 ] ) is repeated with different detector matrices , for which the momentum resolution is altered by @xmath38% .
this corresponds to the estimated uncertainty of its monte carlo prediction .
as expected , no differences are found at low momenta . above 200 ,
the observed relative flux change @xmath68 is well described by @xmath69 with @xmath70 .
the observed difference between the high - energy muon flux measured in different detector regions leads to a somewhat larger value of @xmath71 .
the total uncertainties of the muon flux and charge ratio are obtained by adding the individual contributions in quadrature .
the different sources of the vertical uncertainties are shown in figures [ fig : scsys](b ) and ( c ) .
the muon flux uncertainty is dominated by the uncertainty of the molasse overburden at low momenta and by the alignment and resolution uncertainty at high momenta .
the minimal uncertainty is 2.3% at 150 in the vertical direction .
the vertical charge ratio uncertainty is below 2% up to momenta of 100 . above this momentum ,
it rises rapidly with the alignment uncertainties .
these uncertainties are fully correlated between different momenta for a given zenith angle bin . as
approximately the same detector parts are used to measure the muons in neighboring zenith angles , the systematic uncertainties are also correlated with respect to the zenith angle .
the estimated correlation coefficients are listed in table [ tab : zsyscorr ] .
the understanding of the detector is validated by analyzing the muons produced at lep via the process @xmath72 recorded during the lep calibration runs at a mean centre - of - mass energy of 91.27 .
the selection criteria include the requirement of a muon track close to the collision point and an event - time in coincidence with the lep beam crossing time .
the number of selected muons with a momentum above 60% of the beam energy is converted to an absolute cross section resulting in @xmath73 here the quoted systematic uncertainty includes only sources which are not relevant to the muon spectrum measurement , such as the luminosity .
+ using the lep precision measurements @xcite , the standard model expectation of @xmath74 is calculated @xcite to be @xmath75 which is in excellent agreement to the value measured here .
+ thus this study verifies the l3+c acceptance calculation and a normalization uncertainty of @xmath765.2% at 68% c.l . can be stated .
although this number is larger than the estimated systematic uncertainty of the muon flux normalization , it provides an absolute systematic cross - check qualitatively different from the relative studies described above .
+ the momentum distribution of the selected events , displayed in figure [ fig : z0e ] , shows good agreement between the data and the simulation . from the peak position of the data , an absolute momentum scale uncertainty of @xmath76370 and a single octant alignment uncertainty of @xmath760.1 @xmath66
is derived . + as can be seen in figure figure [ fig : reso](a ) , the momentum resolution derived from the @xmath77 muon sample agrees well with the one measured with atmospheric muons . ) .
the arrow indicates the low momentum cut . ]
the muon fluxes , @xmath78 , conventionally multiplied by the third power of the momentum , and the charge ratios , @xmath79 , are listed for each zenith angle bin in tables [ tab : flux10]@xmath4[tab : flux03 ] with their statistical and systematic uncertainties .
the statistical correlation coefficients , @xmath80 , between neighboring momentum bins , as derived from equation ( [ eq : vm ] ) , are also given .
due to the limited detector resolution these correlations are inevitable .
however , the momentum binning is chosen such that only neighboring bins have a significant correlation .
the average momenta , @xmath81 , within a momentum range @xmath82 $ ] are calculated @xcite by fitting the phenomenological muon flux function from reference @xcite to our data and solving @xmath83 it should be noted that the fluxes are neither corrected for the altitude of l3+c nor for the atmospheric profile ( bottom ) to 58@xmath84 ( top ) .
the inner bars denote the statistical uncertainty , the full bars show the total uncertainty . for better visibility , an offset of 0.05 @xmath2@xmath11s@xmath66sr@xmath66
was added consecutively and lines are shown to guide the eye . ] to avoid additional theoretical uncertainties .
instead , we quote the average atmospheric mass overburden @xmath10 above l3+c , which was continuously measured with balloon flights from close to the experiment to altitudes of over 30 km @xcite .
the parameterization of reference @xcite is used to describe the mass profile @xmath10 in g@xmath11 as a function of the altitude @xmath85 in km above sea level : @xmath86 a fit to the live - time weighted balloon data yields @xmath87 , @xmath88 , @xmath89 , @xmath90 and @xmath91 .
the measured muon fluxes at the l3+c altitude are shown in figure [ fig : spectrum ] for each zenith angle bin . as no previous continuous zenith angle measurements exist in the large energy range examined here ,
only the vertical flux can be compared to other experiments , as shown in figure [ fig : vertflux ] .
only measurements providing an absolute normalization @xcite are taken into account .
the data are extrapolated to sea level using the muon flux predictions of the ` target ` @xcite program . the comparison to low energy experiments @xcite gives a good overall agreement with this analysis above about 40 .
at lower momenta , a systematic slope difference seems to be present , which corresponds to about three standard deviations of the systematic molasse uncertainty estimated above . only three previous experiments measured a normalized spectrum at high energies .
the shape of the kiel measurements @xcite agree with this result over the full momentum range , but a lower flux normalization is determined by l3+c .
+ the data obtained with the mars apparatus @xcite significantly disagree with this result , both in shape and normalization . + above momenta of 50 , the recent muon flux measurement from bess - tev @xcite is in very good agreement with this result .
the measured charge ratios at the l3+c altitude are shown in figure [ fig : crat ] for each zenith angle bin up to momenta of 500 . in the considered momentum range ,
the charge ratio is independent of the momentum within the experimental uncertainties . the mean value in the vertical direction is found to be 1.285 @xmath3 0.003 ( stat . )
@xmath3 0.019 ( sys . ) with a @xmath929.5/11 .
this is in good agreement with the average of all previous measurements , 1.270 @xmath3 0.003 ( stat . ) @xmath3 0.015 ( sys . )
it is worth noting , that the precision of the data of a single l3+c zenith angle bin is comparable to the combined uncertainty of all data collected in the past .
( bottom ) to 58@xmath84 ( top ) .
the inner bars denote the statistical uncertainty , the full bars show the total uncertainty . for better visibility , an offset of 0.5
was added consecutively and lines are shown to guide the eye . ]
we wish to acknowledge the contribution of all the engineers and technicians who have participated in the construction and the maintenance of the l3 and l3+c experiments . those of us who are not from member states thank cern for its hospitality and help .
furthermore , we would like to thank the bartol group for providing the ` target ` program .
+ t. k. gaisser , m. honda , p. lipari and t. stanev , in proceedings of the 27th icrc ( hamburg , 7 - 15 aug , 2001 ) 1643 , and references therein .
r. engel , t. k. gaisser and t. stanev , phys .
b * 472 * ( 2000 ) 113 .
m. honda , t. kajita , k. kasahara and s. midorikawa , phys .
d * 52 * ( 1995 ) 4985 .
v. agrawal , t. k. gaisser , p. lipari and t. stanev , phys .
d * 53 * ( 1996 ) 1314 .
e. v. bugaev _ et al .
_ , phys .
d * 58 * ( 1998 ) 054001 .
g. fiorentini , v. a. naumov and f. l. villante , phys .
b * 510 * ( 2001 ) 173 .
t. k. gaisser and m. honda , ann .
nucl . part .
* 52 * ( 2002 ) 153 , and references therein .
super - kamiokande collaboration , y. fukuda _ et al .
_ , phys .
* 81 * ( 1998 ) 1562 .
soudan 2 collaboration , w. w. m. allison _ et al .
_ , phys .
b * 391 * ( 1997 ) 491 .
macro collaboration , m. ambrosio _ et al .
_ , phys .
b * 434 * ( 1998 ) 451 .
t. hebbeker and c. timmermans , astropart .
phys . * 18 * ( 2002 ) 107 .
l3 collaboration , b. adeva _ et al .
_ , nucl .
instrum .
a * 289 * ( 1990 ) 35 .
l3+c collaboration , o. adriani _ et al .
_ , nucl .
instrum .
a * 488 * ( 2002 ) 209 .
v. innocente and e. nagy , nucl .
instrum .
a * 324 * ( 1993 ) 297 .
d. heck _ et al .
_ , `` corsika : a monte carlo code to simulate extensive air showers , '' forschungszentrum karlsruhe , fzka 6019 ( 1998 ) .
r. brun _ et al .
_ , `` geant 3 , '' cern report dd / ee/84 - 1 ( 1987 ) .
s. bottai and l. perrone , nucl .
instrum .
meth . a * 459 * ( 2001 ) 319 .
p. h. barrett _ et al .
_ , rev .
* 24 * ( 1952 ) 133 .
service de climatologie , geneva . c. brouwer _
_ , nucl .
instrum .
a * 313 * ( 1992 ) 50 .
p. duinker _ et al .
_ , nucl .
instrum .
meth . a * 273 * ( 1988 ) 814
. m. fabre , `` the dimuon mass resolution of the l3 experiment at lep , '' ph.d .
thesis , eth
zrich , 1992 .
m. unger , `` measurement of the momentum spectrum of atmospheric muons with the l3 detector , '' ph.d .
thesis , humboldt universitt berlin , 2004 , desy - thesis-2004 - 008 .
w. lohmann , r. kopp and r. voss , `` energy loss of muons in the energy range 110000 gev , '' cern report 85 - 03 ( 1985 ) .
d. e. groom , n. v. mokhov and s. i. striganov , atom .
data nucl .
data tabl .
* 78 * ( 2001 ) 183 .
d. chirkin and w. rhode , `` muon monte carlo : a new high precision tool for muon propagation through matter , '' desy - proc-2002 - 01 ( 2002 ) .
particle data group collaboration , k. hagiwara _ et al .
_ , phys .
d * 66 * , ( 2002 ) 010001 , and references therein .
d. y. bardin _ et al .
_ , comput . phys
. commun .
* 133 * ( 2001 ) 229 .
s. jadach , b. f. l. ward and z. was , comput .
* 79 * ( 1994 ) 503 .
t. sjstrand , comput .
* 82 * ( 1994 ) 74 .
k. maeda , fortsch .
* 21 * ( 1973 ) 113 .
o. c. allkofer , k. carstensen and d. w. dau , phys .
b * 36 * ( 1971 ) 425 . c. a. ayre
_ et al . _ , j. phys .
g * 1 * ( 1975 ) 584 .
b. j. bateman _ et al .
_ , phys .
b * 36 * ( 1971 ) 144 .
j. green _ et al . _ ,
d * 20 * ( 1979 ) 1598 .
de pascale _ et al .
_ , j. geophys .
* 98 * ( 1993 ) 3501 .
caprice collaboration , j. kremer _ et al .
_ , phys .
* 83 * ( 1999 ) 4241 .
caprice collaboration , m. boezio _ et al .
_ , phys .
d * 67 * ( 2003 ) 072003 .
bess collaboration , t. sanuki _ et al .
_ , phys .
b * 541 * ( 2002 ) 234 [ erratum - ibid .
b * 581 * ( 2004 ) 272 ] .
bess collaboration , s. haino _ et al .
_ , `` measurements of primary and atmospheric cosmic - ray spectra with the bess - tev spectrometer '' [ arxiv : astro - ph/0403704 ] .
p.achard o.adriani m.aguilar-benitezm.van den akker j.alcaraz g.alemannij.allabya.aloisio m.g.alviggih.anderhub v.p.andreevf.anselmoa.arefiev t.azemoon t.aziz p.bagnaiaa.bajo g.baksayl.baksayj.bhrs.v.baldew s.banerjee sw.banerjee a.barczyk r.barillre p.bartalini m.basilen.batalovar.battistona.bay f.becattiniu.beckerf.behnerl.bellucci r.berbeco j.berdugo p.berges b.bertuccib.l.betevm.biasinim.bigliettia.biland j.j.blaising s.c.blyth g.j.bobbink a.bhml.boldizsarb.borgia s.bottaid.bourilkovm.bourquins.braccinij.g.bransonf.brochu j.d.burgerw.j.burgerx.d.cai m.capellg.cara romeog.carlinoa.cartacci j.casausf.cavallarin.cavallo c.cecchi m.cerradam.chamizot.chiarusiy.h.chang m.chemarina.chen g.chen g.m.chen h.f.chen h.s.cheng.chiefari l.cifarellif.cindoloi.clarer.clare g.coignet n.colino s.costantini b.de la cruzs.cucciarellij.a.van dalen r.de asmundisp.dglon j.debreczenia.degr k.dehmeltk.deiters d.della volpe e.delmeire p.denes f.denotaristefania.de salvo m.diemoz m.dierckxsens l.k.dingc.dionisi m.dittmara.doriam.t.dovad.duchesneau m.dudai.duranb.echenarda.elinea.el hageh.el mamounia.engler f.j.eppling p.extermanng.faber m.a.falagans.falcianoa.favaraj.fay o.fedinm.felcinit.ferguson h.fesefeldt e.fiandrinij.h.field f.filthautw.fisheri.fiskg.forconi k.freudenreichc.furettayu.galaktionovs.n.ganguli p.garcia-abiam.gataullins.gentiles.giaguz.f.gongh.j.graboschg.grenier o.grimmh.groenstege m.w.gruenewald m.guida y.n.guos.guptav.k.gupta a.gurtul.j.gutayd.haasch.hallerd.hatzifotiadouy.hayashiz.x.het.hebbekera.herv j.hirschfelderh.hoferh.hofer,jun .
m.hohlmanng.holzner s.r.houa.x.huoy.hun.ito b.n.jin c.l.jing l.w.jonesp.de
jongi.josa-mutuberrav.kantserov m.kaurs.kawakamim.n.kienzle-focaccij.k.kimj.kirkbyw.kittela.klimentov a.c.knige.koka.kornm.kopalv.koutsenko m.krberh.h.kuang r.w.kraemera.krgerj.kuijpers a.kunin p.ladron de guevarai.laktinehg.landim.lebeaua.lebedevp.lebrunp.lecomte p.lecoq p.le coultre j.m.le goffy.leih.leichr.leiste m.levtchenkop.levtchenkoc.li l.liz.c.lis.likhoded c.h.linw.t.linf.l.lindel.listaz.a.liuw.lohmanne.longo y.s.lu c.luci l.luminariw.lustermannw.g.ma x.h.may.q.mal.malgeria.malinin c.maaj.mans j.p.martin f.marzano k.mazumdarr.r.mcneil s.melex.w.mengl.merola m.meschini w.j.metzgera.mihula.van milh.milcentg.mirabelli j.mnichg.b.mohanty b.monteleonig.s.muanzaa.j.m.muijsb.musicar m.musy s.nagyr.nahnhauerv.a.naumov s.natalem.napolitanof.nessi-tedaldih.newman a.nisatit.novakh.nowak r.ofierzynski g.organtinii.palc.palomaresp.paoluccir.paramatti j .-
t.paulm.pauluzzic.pausf.paussm.pedaces.pensottid.perret-gallix b.petersend.piccolo f.pierella m.pierim.pioppip.a.pirou e.pistolesiv.plyaskin m.pohl v.pojidaevj.pothierd.prokofiev j.quartieric.r.qingg.rahal-callotm.a.rahaman p.raics n.rajar.ramelli p.g.rancoitar.ranieri a.raspereza k.c.ravindranp.razisd.ren m.rescignos.reucroftp.rewiersmas.riemannk.rilesb.p.roea.rojkovl.romero a.rosca c.rosemannc.rosenblecks.rosier-leess.rothj.a.rubio g.ruggiero h.rykaczewski r.saidia.sakharovs.saremi s.sarkarj.salicio e.sanchezc.schferv.schegelskyv.schmittb.schoeneichh.schopperd.j.schotanusc.sciaccal.servolic.q.shens.shevchenkon.shivarovv.shoutko e.shumilov a.shvorobd.sonc.sougap.spillantini m.steuerd.p.stickland b.stoyanova.straessnerk.sudhakarh.sulankeg.sultanovl.z.suns.sushkovh.suter j.d.swainz.szillasix.w.tangp.tarjanl.tauscherl.taylorb.tellili d.teyssier c.timmermanssamuel c.c.ting s.m.ting s.c.tonwar j.tth g.trowitzschc.tullyk.l.tungj.ulbricht m.ungere.valente h.verkooijenr.t.van de waller.vasquezv.veszpremig.vesztergombii.vetlitsky d.vicinanza g.viertel s.villam.vivargent s.vlachosi.vodopianov h.vogelh.vogt i.vorobiev a.a.vorobyov m.wadhwar.g.wangq.wangx.l.wangx.w.wang z.m.wangm.weberr.van wijkt.a.m.wijnenh.wilkenss.wynhoff l.xia y.p.xuj.s.xuz.z.xu j.yamamoto b.z.yang c.g.yang h.j.yangm.yangx.f.yangz.g.yaos.c.yeh z.q.yu an.zaliteyu.zalitec.zhangf.zhangj.zhangs.zhangz.p.zhang j.zhaos.j.zhoug.y.zhur.y.zhuh.l.zhuangq.q.zhua.zichichib.zimmermann m.zllera.n.m.zwart .
20.0&27.0 & 23.18 & 0.217 & 0.4&[1.5ex ] -0.24 & 4.4 & 1.274 & 0.7&[1.5ex ] -0.24 & 1.5 + 27.0&34.5 & 30.47 & 0.231 & 0.3&[1.5ex ] -0.22 & 3.7 & 1.284 & 0.6&[1.5ex ] -0.22 & 1.5 + 34.5&42.0 & 38.02 & 0.244 & 0.4&[1.5ex ] -0.24 & 3.3 & 1.295 & 0.8&[1.5ex ] -0.24 & 1.5 + 42.0&50.0 & 45.78 & 0.252 & 0.5&[1.5ex ] -0.26 & 3.0 & 1.269 & 0.9&[1.5ex ] -0.26 & 1.5 + 50.0&58.5 & 54.04 & 0.257 & 0.5&[1.5ex ] -0.31 & 2.8 & 1.286 & 1.0&[1.5ex ] -0.31 & 1.5 + 58.5&68.5 & 63.25 & 0.261 & 0.5&[1.5ex ] -0.31 & 2.6 & 1.298 & 1.0&[1.5ex ] -0.31 & 1.5 + 68.5&81.5 & 74.63 & 0.266 & 0.5&[1.5ex ] -0.30 & 2.5 & 1.273 & 1.1&[1.5ex ] -0.29 & 1.5 + 81.5&100 & 90.13 & 0.262 & 0.6&[1.5ex ] -0.26 & 2.4 & 1.293 & 1.2&[1.5ex ] -0.26 & 1.5 + 100&132 & 114.5 & 0.264 & 0.5&[1.5ex ] -0.22 & 2.3 & 1.286 & 1.0&[1.5ex ] -0.21 & 1.6 + 132&200 & 161.3 & 0.250 & 0.5&[1.5ex ] -0.22 & 2.2 & 1.287 & 1.1&[1.5ex ] -0.22 & 2.1 + 200&300 & 243.0 & 0.233 & 0.8&[1.5ex ] -0.27 & 2.2 & 1.327 & 1.7&[1.5ex ] -0.27 & 3.7 + 300&500 & 381.9 & 0.203 & 1.2&[1.5ex ] -0.28 & 2.4 & 1.276 & 2.6 & & 7.2 + 500&1000 & 687.2 & 0.151 & 2.3&[1.5ex ] -0.30 & 3.5 + 1000&3000 & 1599 & 0.087 & 6.0 & & 8.4 + 27.0&34.5 & 30.47 & 0.237 & 0.3&[1.5ex ] -0.23 & 4.5 & 1.265 & 0.6&[1.5ex ] -0.22 & 2.3 + 34.5&42.0 & 38.02 & 0.254 & 0.4&[1.5ex ] -0.24 & 4.1 & 1.261 & 0.8&[1.5ex ] -0.24 & 2.3 + 42.0&50.0 & 45.78 & 0.259 & 0.5&[1.5ex ] -0.26 & 3.8 & 1.265 & 1.0&[1.5ex ] -0.26 & 2.3 + 50.0&58.5 & 54.04 & 0.264 & 0.5&[1.5ex ] -0.30 & 3.7 & 1.297 & 1.1&[1.5ex ] -0.30 & 2.3 + 58.5&68.5 & 63.25 & 0.266 & 0.6&[1.5ex ] -0.31 & 3.5 & 1.250 & 1.1&[1.5ex ] -0.31 & 2.3 + 68.5&81.5 & 74.63 & 0.270 & 0.6&[1.5ex ] -0.29 & 3.4 & 1.319 & 1.2&[1.5ex ] -0.29 & 2.3 + 81.5&100 & 90.13 & 0.271 & 0.6&[1.5ex ] -0.25 & 3.3 & 1.259 & 1.3&[1.5ex ] -0.25 & 2.3 + 100&132 & 114.5 & 0.270 & 0.5&[1.5ex ] -0.21 & 3.2 & 1.296 & 1.1&[1.5ex ] -0.21 & 2.5 + 132&200 & 161.3 & 0.262 & 0.6&[1.5ex ] -0.21 & 3.1 & 1.281 & 1.2&[1.5ex ] -0.21 & 3.0 + 200&300 & 243.0 & 0.241 & 0.9&[1.5ex ] -0.27 & 3.1 & 1.273 & 1.8&[1.5ex ] -0.27 & 4.8 + 300&500 & 382.1 & 0.219 & 1.3&[1.5ex ] -0.27 & 3.2 & 1.389 & 2.7 & & 8.3 + 500&1000 & 687.8 & 0.165 & 2.4&[1.5ex ] -0.30 & 4.2 + 1000&3000 & 1604 & 0.093 & 6.5 & & 9.2 + 27.0&34.5 & 30.47 & 0.245 & 0.4&[1.5ex ] -0.24 & 4.3 & 1.265 & 0.7&[1.5ex ] -0.24 & 2.3 + 34.5&42.0 & 38.02 & 0.257 & 0.4&[1.5ex ] -0.26 & 3.9 & 1.296 & 0.9&[1.5ex ] -0.26 & 2.3 + 42.0&50.0 & 45.78 & 0.265 & 0.5&[1.5ex ] -0.28 & 3.6 & 1.302 & 1.1&[1.5ex ] -0.28 & 2.3 + 50.0&58.5 & 54.04 & 0.273 & 0.5&[1.5ex ] -0.31 & 3.4 & 1.267 & 1.1&[1.5ex ] -0.31 & 2.3 + 58.5&68.5 & 63.25 & 0.279 & 0.5&[1.5ex ] -0.31 & 3.2 & 1.287 & 1.1&[1.5ex ] -0.30 & 2.3 + 68.5&81.5 & 74.64 & 0.282 & 0.5&[1.5ex ] -0.27 & 3.1 & 1.272 & 1.1&[1.5ex ] -0.27 & 2.3 + 81.5&100 & 90.14 & 0.283 & 0.5&[1.5ex ]
-0.23 & 3.0 & 1.288 & 1.1&[1.5ex ] -0.23 & 2.4 + 100&132 & 114.5 & 0.284 & 0.5&[1.5ex ] -0.19 & 2.9 & 1.303 & 0.9&[1.5ex ] -0.19 & 2.6 + 132&200 & 161.3 & 0.274 & 0.5&[1.5ex ] -0.19 & 2.8 & 1.325 & 1.0&[1.5ex ] -0.19 & 3.8 + 200&300 & 243.0 & 0.255 & 0.7&[1.5ex ] -0.23 & 2.8 & 1.300 & 1.4&[1.5ex ] -0.23 & 6.8 + 300&500 & 382.0 & 0.220 & 1.1&[1.5ex ] -0.25 & 2.9 & 1.437 & 2.2 & & 12.0 + 500&1000 & 687.3 & 0.172 & 1.9&[1.5ex ] -0.29 & 4.2 + 1000&3000 & 1599 & 0.091 & 5.7 & & 11.3 + 34.5&42.0 & 38.02 & 0.263 & 0.4&[1.5ex ] -0.28 & 3.9 & 1.248 & 0.8&[1.5ex ] -0.28 & 1.3 + 42.0&50.0 & 45.78 & 0.269 & 0.5&[1.5ex ] -0.29 & 3.5 & 1.288 & 1.1&[1.5ex ] -0.29 & 1.3 + 50.0&58.5 & 54.04 & 0.274 & 0.5&[1.5ex ] -0.32 & 3.3 & 1.280 & 1.1&[1.5ex ] -0.32 & 1.3 + 58.5&68.5 & 63.25 & 0.284 & 0.6&[1.5ex ] -0.31 & 3.1 & 1.285 & 1.2&[1.5ex ] -0.31 & 1.3 + 68.5&81.5 & 74.64 & 0.280 & 0.6&[1.5ex ] -0.28 & 2.9 & 1.274 & 1.2&[1.5ex ] -0.28 & 1.3 + 81.5&100 & 90.14 & 0.288 & 0.6&[1.5ex ] -0.24 & 2.8 & 1.302 & 1.3&[1.5ex ] -0.24 & 1.3 + 100&132 & 114.5 & 0.290 & 0.5&[1.5ex ] -0.20 & 2.6 & 1.280 & 1.1&[1.5ex ] -0.20 & 1.4 + 132&200 & 161.3 & 0.279 & 0.6&[1.5ex ] -0.20 & 2.6 & 1.313 & 1.1&[1.5ex ] -0.20 & 2.0 + 200&300 & 243.1 & 0.264 & 0.8&[1.5ex ] -0.27 & 2.5 & 1.301 & 1.7&[1.5ex ] -0.27 & 4.7 + 300&500 & 382.1 & 0.232 & 1.3&[1.5ex ] -0.30 & 2.6 & 1.318 & 2.6 & & 11.3 + 500&1000 & 686.8 & 0.167 & 2.5&[1.5ex ] -0.37 & 4.2 + 1000&3000 & 1587 & 0.087 & 7.9 & & 10.4 + 34.5&42.0 & 38.02 & 0.255 & 0.5&[1.5ex ] -0.32 & 4.6 & 1.279 & 1.0&[1.5ex ] -0.32 & 1.0 + 42.0&50.0 & 45.78 & 0.265 & 0.5&[1.5ex ] -0.32 & 4.2 & 1.284 & 1.1&[1.5ex ] -0.32 & 1.0 + 50.0&58.5 & 54.04 & 0.275 & 0.6&[1.5ex ] -0.35 & 3.9 & 1.263 & 1.1&[1.5ex ] -0.35 & 1.0 + 58.5&68.5 & 63.25 & 0.283 & 0.6&[1.5ex ] -0.33 & 3.7 & 1.261 & 1.2&[1.5ex ] -0.33 & 1.0 + 68.5&81.5 & 74.64 & 0.283 & 0.6&[1.5ex ] -0.29 & 3.6 & 1.271 & 1.2&[1.5ex ] -0.29 & 1.0 + 81.5&100 & 90.14 & 0.290 & 0.6&[1.5ex ] -0.26 & 3.4 & 1.281 & 1.2&[1.5ex ] -0.26 & 1.0 + 100&132 & 114.6 & 0.290 & 0.5&[1.5ex ] -0.20 & 3.3 & 1.256 & 1.0&[1.5ex ] -0.20 & 1.1 + 132&200 & 161.3 & 0.292 & 0.5&[1.5ex ] -0.22 & 3.2 & 1.280 & 1.1&[1.5ex ] -0.22 & 1.4 + 200&300 & 243.1 & 0.271 & 0.8&[1.5ex ] -0.28 & 3.2 & 1.297 & 1.6&[1.5ex ] -0.28 & 3.1 + 300&500 & 382.2 & 0.244 & 1.2&[1.5ex ] -0.31 & 3.3 & 1.428 & 2.5 & & 7.4 + 500&1000 & 687.5 & 0.191 & 2.2&[1.5ex ] -0.36 & 4.0 + 1000&3000 & 1594 & 0.087 & 7.6 & & 8.4 + 34.5&42.0 & 38.02 & 0.253 & 0.6&[1.5ex ] -0.38 & 4.6 & 1.273 & 1.2&[1.5ex ] -0.38 & 1.8 + 42.0&50.0 & 45.78 & 0.265 & 0.6&[1.5ex ] -0.35 & 4.1 & 1.292 & 1.1&[1.5ex ] -0.35 & 1.8 + 50.0&58.5 & 54.04 & 0.275 & 0.5&[1.5ex ] -0.38 & 3.7 & 1.263 & 1.1&[1.5ex ] -0.38 & 1.8 + 58.5&68.5 & 63.25 & 0.281 & 0.6&[1.5ex ] -0.34 & 3.4 & 1.273 & 1.2&[1.5ex ] -0.34 & 1.8 + 68.5&81.5 & 74.64 & 0.291 & 0.6&[1.5ex ] -0.31 & 3.1 & 1.263 & 1.2&[1.5ex ] -0.31 & 1.8 + 81.5&100 & 90.15 & 0.293 & 0.6&[1.5ex ] -0.27 & 2.9 & 1.260 & 1.2&[1.5ex ] -0.27 & 1.8 + 100&132 & 114.6 & 0.298 & 0.5&[1.5ex ] -0.22 & 2.7 & 1.266 & 1.0&[1.5ex ] -0.22 & 1.8 + 132&200 & 161.4 & 0.298 & 0.5&[1.5ex ] -0.23 & 2.5 & 1.281 & 1.0&[1.5ex ] -0.23 & 1.9 + 200&300 & 243.2 & 0.285 & 0.7&[1.5ex ] -0.29 & 2.4 & 1.267 & 1.6&[1.5ex ] -0.29 & 2.8 + 300&500 & 382.3 & 0.256 & 1.1&[1.5ex ] -0.31 & 2.6 & 1.394 & 2.5 & & 5.7 + 500&1000 & 688.2 & 0.203 & 2.0&[1.5ex ] -0.35 & 3.5 + 1000&3000 & 1601 & 0.120 & 5.3 & & 7.8 + 42.0&50.0 & 45.79 & 0.256 & 0.6&[1.5ex ] -0.41 & 4.7 & 1.262 & 1.3&[1.5ex ] -0.41 & 1.4 + 50.0&58.5 & 54.05 & 0.269 & 0.6&[1.5ex ] -0.42 & 4.3 & 1.284 & 1.2&[1.5ex ] -0.42 & 1.4 + 58.5&68.5 & 63.26 & 0.282 & 0.6&[1.5ex ] -0.36 & 4.0 & 1.261 & 1.3&[1.5ex ] -0.36 & 1.4 + 68.5&81.5 & 74.65 & 0.290 & 0.6&[1.5ex ] -0.31 & 3.7 & 1.277 & 1.3&[1.5ex ] -0.31 & 1.4 + 81.5&100 & 90.16 & 0.299 & 0.6&[1.5ex ] -0.26 & 3.5 & 1.267 & 1.2&[1.5ex ] -0.26 & 1.4 + 100&132 & 114.6 & 0.307 & 0.5&[1.5ex ] -0.22 & 3.3 & 1.287 & 1.0&[1.5ex ] -0.22 & 1.4 + 132&200 & 161.4 & 0.312 & 0.5&[1.5ex ] -0.24 & 3.1 & 1.274 & 1.0&[1.5ex ] -0.23 & 1.4 + 200&300 & 243.2 & 0.300 & 0.8&[1.5ex ] -0.32 & 3.0 & 1.278 & 1.6&[1.5ex ] -0.31 & 2.1 + 300&500 & 382.5 & 0.277 & 1.2&[1.5ex ] -0.35 & 3.1 & 1.317 & 2.4 & & 4.7 + 500&1000 & 688.5 & 0.231 & 2.0&[1.5ex ] -0.38 & 3.8 + 1000&3000 & 1602 & 0.117 & 6.3 & & 8.7 + 42.0&50.0 & 45.79 & 0.256 & 1.4&[1.5ex ] -0.49 & 5.6 & 1.281 & 2.8&[1.5ex ] -0.50 & 1.4 + 50.0&58.5 & 54.05 & 0.255 & 0.9&[1.5ex ] -0.45 & 5.3 & 1.269 & 1.7&[1.5ex ] -0.45 & 1.4 + 58.5&68.5 & 63.26 & 0.273 & 0.8&[1.5ex ] -0.38 & 5.1 & 1.290 & 1.6&[1.5ex ] -0.37 & 1.4 + 68.5&81.5 & 74.65 & 0.284 & 0.8&[1.5ex ] -0.32 & 4.9 & 1.280 & 1.5&[1.5ex ] -0.32 & 1.4 + 81.5&100 & 90.16 & 0.294 & 0.7&[1.5ex ] -0.25 & 4.7 & 1.278 & 1.4&[1.5ex ] -0.25 & 1.4 + 100&132 & 114.6 & 0.305 & 0.5&[1.5ex ] -0.22 & 4.6 & 1.258 & 1.1&[1.5ex ] -0.21 & 1.4 + 132&200 & 161.5 & 0.316 & 0.6&[1.5ex ] -0.22 & 4.5 & 1.254 & 1.2&[1.5ex ] -0.22 & 1.6 + 200&300 & 243.4 & 0.315 & 0.8&[1.5ex ] -0.30 & 4.4 & 1.285 & 1.7&[1.5ex ] -0.30 & 2.6 + 300&500 & 382.7 & 0.297 & 1.3&[1.5ex ] -0.33 & 4.4 & 1.332 & 2.6 & & 5.6 + 500&1000 & 689.3 & 0.243 & 2.3&[1.5ex ] -0.36&4.8 + 1000&3000 & 1604 & 0.153 & 5.9 & & 9.4 +
|
the absolute muon flux between 20 and 3000 is measured with the l3 magnetic muon spectrometer for zenith angles ranging from @xmath0 to @xmath1 . due to the large exposure of about 150 m@xmath2srd , and the excellent momentum resolution of the l3 muon chambers , a precision of 2.3% at 150 in the vertical direction is achieved .
the ratio of positive to negative muons is studied between 20 and 500 , and the average vertical muon charge ratio is found to be 1.285 @xmath3 0.003 ( stat . ) @xmath3 0.019 ( syst . ) . _
the l3+c group dedicates this publication to the late bianca monteleoni . _
|
the experimental evidence for neutrino mass , acquired in recent decades , provides concrete evidence for physics beyond the standard model ( sm ) ( see e.g. @xcite ) .
although the requisite new degrees of freedom can not yet be determined it is clear that additional particles are likely to exist , in order to generate the masses .
similarly , there is by now a large amount of evidence for an additional galactic constituent , an unknown substance referred to as dark matter ( see e.g. @xcite ) .
this may or may not require new degree s of freedom , but the hypothesis that the dark matter is comprised of a stable ( or long lived ) new particle species provides a simple explanation for this observed feature of the universe . given that these two indicators for beyond - sm physics can be explained by extending the particle spectrum of the sm , it is natural to ask if the requisite new particles can be related .
could the mechanism of neutrino mass be related to the existence of a stable dark - matter candidate ?
a particularly simple model realizing this idea was proposed by ma in 2006 @xcite .
this model extends the sm to include an additional sm - like scalar doublet and gauge - singlet fermions , all of which are odd under a discrete @xmath4 symmetry .
the extended field content allows for radiative neutrino mass , generated at the one - loop level , while the lightest beyond - sm field is absolutely stable .
one thus arrives at a simple synergetic model of radiative neutrino mass and dark matter . in this work
we generalize ma s approach .
we present a class of related models , all of which generate neutrino mass via a loop diagram with the same topography as ma s , whilst simultaneously admitting stable dark - matter candidates .
the loop - diagram employed by ma contains a mass insertion on the internal fermion line ( see figure [ fig : yukawa_loop ] ) , and our generalizations fall naturally into two categories ; those which break lepton - number symmetry via a majorana mass insertion , and those with a dirac mass insertion , such that lepton - number symmetry is broken at a vertex .
although the basic mechanism is very similar in both cases , this difference modifies one s expectations for the beyond - sm field content and the associated phenomenology .
it turns out that , in both cases , this approach is very general and many realizations are possible .
however , restricting attention to models in which the beyond - sm multiplets are no larger than the adjoint representation significantly reduces the possibilities .
as we shall see , there are only two such ( minimal ) models with a majorana mass - insertion , both of which are known @xcite .
we find seven additional models with a mass insertion of the dirac type , all of which achieve radiative neutrino mass and dark - matter candidates .
we detail these models , finding a subset that are compatible with direct - detection experiments .
there are cases with an inert singlet , an inert doublet , and an inert triplet ; the models therefore provide a natural setting for inert @xmath0-tuplet theories of dark matter , such that radiative neutrino mass is also achieved .
interestingly , three of the new models admit a simple extension that can explain the origin of the ( formerly imposed ) discrete symmetry . by upgrading the discrete symmetry to a gauged @xmath5 symmetry , and extending the field content by a single sm - singlet scalar ,
the discrete symmetry can arise as an accidental symmetry of the low - energy lagrangian , after @xmath5 symmetry breaking takes place .
this provides a simple explanation for the discrete symmetry .
though we focus on models with representations no larger than the adjoint , we also briefly discuss cases where the beyond - sm fields can be quadruplet and/or quintuplet representations of @xmath6 .
we present the candidate models in these cases , and mention some key issues , based on the lessons learned from our preceding studies . despite the use of larger multiplets , these models can still be of interest ; in addition to allowing for radiative neutrino mass , the exotics with larger electric - charges in these multiplets can enhance the @xmath7 and/or @xmath8 signal from dark - matter annihilation when they appear inside loops @xcite .
this can provide a simple connection between the mechanism of neutrino mass and the astrophysical gamma - ray signal @xcite . before proceeding we note that the connection between radiative neutrino mass and dark matter has been explored in a number of different models , including ref .
@xcite , which precedes ma s work ; for other examples see @xcite .
previous works on inert - singlet models @xcite , inert - doublet models @xcite , and inert - triplet models @xcite are also well known .
additional relevant works dealing with inert - multiplet dark matter and/or radiative neutrino mass are cited in the text .
note also that refs .
@xcite have detailed the one - loop realizations of the @xmath9 operator for neutrino mass .
inert scalar dark matter can also help cure the little hierarchy problem found in low - scale seesaws @xcite .
the present work follows on from the generalized tree - level seesaws presented in ref .
@xcite .
the plan of this paper is as follows . in section [ sec : rad_nu_plus_dm ]
we discuss ma s model and present the generalizations that similarly achieve radiative neutrino mass and dark - matter candidates .
section [ sec : fermion_dm ] considers the case of fermionic dark - matter , while section [ sec : scalar_dm ] discusses scalar dark - matter .
one of the generalized models is presented in more detail in section [ sec : real_and_complex_scalar ] . in section [ sec : gauge_symmetry ] we show that some of the models allow a simple extension , such that the discrete symmetry appears as an accidental symmetry in the low - energy theory . models with exotics forming larger @xmath10 representations are discussed in section [ sec : non - minimal_models ] ( and explicitly displayed in the appendix ) .
we conclude in section [ sec : conc ] .
we are interested in the class of models that generate neutrino mass radiatively by the diagram in figure [ fig : yukawa_loop ] . here
@xmath11 is a beyond - sm fermion and @xmath12 are new scalars ( which can be identical in some cases ) .
a basic feature of this diagram is that the three vertices can all involve two beyond - sm fields .
consequently one can always consider a discrete @xmath4 symmetry whose action on the beyond - sm multiplets is \ { , s_1 , s_2 } & & - \ { , s_1 , s_2 } , while all sm fields transform trivially .
the lightest field within the multiplets @xmath11 and @xmath12 will thus be stable , and provided this field is electrically neutral and colorless , one arrives at a dark - matter candidate .
these comments are generic for all models of this type ; the connection between loop masses and dark matter is simple to realize in this class of models . figure [ fig : yukawa_loop ] produces majorana neutrino masses , so the loop diagram must contain a source of lepton number violation . choosing the convention for lepton - number symmetry such that the new fermion @xmath11 has the same value as the sm leptons , there are two ways to explicitly break lepton - number symmetry ; it can be broken at either the mass insertion or at one of the vertices .
the simplest models , in terms of the requisite number of beyond - sm multiplets , are those with a lepton number violating ( majorana ) mass insertion . in this case one
has @xmath13 and minimal cases occur for @xmath14 .
thus , only two beyond - sm multiplets are required . the general loop - diagram for this subset of models is given in figure [ fig : mass_loop ] .
we consider this case first .
we seek models that achieve neutrino mass via figure [ fig : mass_loop ] and give rise to dark - matter candidates .
clearly the fermion must form a real representation of the sm gauge symmetry , @xmath15 , in order to allow a bare majorana mass .
as we are considering dark - matter candidates we do not consider colored fields .
note also that @xmath16 must be odd - valued to ensure there is no fractionally charged particles ( the lightest of which would be stable and thus cosmologically excluded ) .
odd - valued @xmath16 also ensures that @xmath17 contains an electrically neutral component , so no additional constraint is imposed by this demand . with this information
one can obtain the viable combinations of @xmath17 and @xmath18 that generate figure [ fig : mass_loop ] . the basic lagrangian terms are & & i|_r^d__r - _ r + |d^s|^2 -m_s^2 |s|^2 + |l _
r+_(s^h)^2 + , [ eq : ma_lagrangian ] where @xmath19 ( @xmath20 ) is the sm lepton ( scalar ) doublet and @xmath21 denotes the charge - conjugate of @xmath18 .
it turns out that the possible combinations for @xmath11 and @xmath18 are not restricted by quantum numbers ; one can consider increasingly large multiplets , presumably up to some unitarity limits @xcite , and realize a model with figure [ fig : mass_loop ] and a dark - matter candidate .
however , if we restrict our attention to models with @xmath22 , such that no new multiplet is larger than the adjoint representation , there are only two possibilities .
the first case is ma s original proposal , which employs an additional ( inert ) scalar doublet @xmath23 , and a gauge - singlet fermion @xmath24 @xcite .
this model is the prototype for the class we are considering .
the second model also employs the scalar doublet @xmath23 , but instead utilizes the triplet fermion @xmath25 @xcite , familiar from the type - iii seesaw @xcite .
thus , both of the models are known in the literature , and there are no additional possibilities unless one considers larger multiplets . in each of these models
the dark - matter candidate can be a neutral component of @xmath18 or a majorana fermion .
there is , however , an important difference between the singlet case , @xmath26 , and the other model ; the singlet does not participate in weak interactions and is therefore brought into thermal contact with the sm sector via the yukawa coupling . for fermionic dark - matter , this can produce conflict between the need to keep the yukawa coupling large to ensure thermal dark - matter , and the need to suppress the yukawa coupling to limit the size of flavor changing effects .
this issue does not arise in the triplet fermion model , as the fermions can maintain equilibrium with the sm sector via weak interactions in these cases , even if the yukawa couplings are suppressed .
for an analysis of ma s model , incorporating recent lhc data on the higgs , see e.g. ref .
also note that loop effects can induce observable interactions between dark matter and experimental detectors in ma s model @xcite .
having exhausted the minimal models with a lepton - number violating mass insertion , we now consider models with a dirac mass insertion ; i.e. the beyond - sm fermion has nonzero hypercharge . in this case
the general mass - diagram has the form shown in figure [ fig : yukawa_loop ] .
the fields @xmath17 and @xmath27 are no longer related by charge conjugation , so the mass insertion is of the dirac type and @xmath11 is a vector - like fermion .
in addition one requires @xmath28 .
we again consider a @xmath4 symmetry under which the sm fields transform trivially but the new fields are odd .
the lagrangian contains the following pertinent terms : & & i|^d _ - m _ + _
i=1,2\{|d^s_i|^2 - m_i^2 |s_i|^2 } + & & + _ 1|l _ rs_1 + _ 2 |l^c_l_2 + _
s_1_2 h^2 + , [ eq : l_ykawa_lagrange ] where , in our convention , lepton number symmetry is broken by a yukawa coupling .
with the @xmath4 symmetry present , there are no terms in the scalar potential that are linear in _ just one _ of the new scalars @xmath12
. therefore the beyond - sm scalars do not acquire an induced vev and parameter space exists for which @xmath29 , so the @xmath4 symmetry remains exact .
in selecting viable multiplets one must ensure that no new multiplet contains a fractionally charged field , to avoid a ( cosmologically excluded ) stable charged field . to ensure that the lightest @xmath4-odd field is a neutral dark - matter candidate one must demand that the new multiplets contain at least one neutral field .
note that the neutral field does not have to appear as an explicit propagating degree of freedom inside the loop diagram ; it is sufficient merely that the loop - diagram exists and that the particle content includes a neutral field that can play the role of dark matter . with these conditions in mind we search for viable combinations of the beyond - sm multiplets that realize figure [ fig : yukawa_loop ] .
we find that the size of the beyond - sm multiplets is not restricted by our demands ; one can consider increasingly large multiplets and realize the loop diagram .
however , if attention is restricted to models in which none of the beyond - sm multiplets are larger than the adjoint representation , only seven distinct models are found . these are listed in table [ l_yukawa ] . of the seven models ,
one employs the exotic lepton triplet @xmath30 , studied in refs .
@xcite , three contain an exotic vector - like ( and sm - like ) lepton doublet @xmath31 @xcite , and two contain a charged lepton doublet @xmath32 @xcite .
there is also a model with a sm - like charged singlet fermion , @xmath33 , which already appeared in ref .
@xcite .
neutrino masses take a standard calculable form in these models . for example , in models @xmath34 and @xmath35 only singly - charged exotics propagate in the loop , and the sm neutrino mass matrix is given by , and should be multiplied by an extra factor of 2 for model @xmath34 . ]
( _ ) _ & & + & & here @xmath36 is the mass for the charged component of the exotic fermion @xmath37 , and summation is implied for the repeated index @xmath38 ( which labels the exotic - fermion generations ) .
the masses @xmath39 refer to the charged scalar mass - eigenstates , which are linear combinations of the charged scalars @xmath40 and @xmath41 .
the mixing results from the @xmath42-term in eq . , which takes the explicit form @xmath43 , for models @xmath34 and @xmath35 .
if all the exotics are at the tev scale one requires dimensionless couplings of @xmath44 to obtain @xmath45 ev .
the scenario with all exotics at the tev scale is most interesting from a phenomenological perspective .
however , strictly speaking one only requires the lightest exotic to have a mass of @xmath46 in order to realize a dark - matter candidate .
the other exotics can be much heavier , allowing larger dimensionless couplings . before moving on to discuss dark matter in detail
, we note that , of the models in table [ l_yukawa ] , only models @xmath34 , @xmath35 and @xmath47 are expected to produce ( dominant ) radiative neutrino masses in the absence of the @xmath4 symmetry , as we briefly discuss in appendix [ app : mass_no_z2 ] .
m. c. gonzalez - garcia , m. maltoni , j. salvado and t. schwetz , jhep * 1212 * , 123 ( 2012 ) [ arxiv:1209.3023 [ hep - ph ] ] .
a. h. g. peter , arxiv:1201.3942 [ astro-ph.co ] .
e. ma , phys .
d * 73 * , 077301 ( 2006 ) [ hep - ph/0601225 ] .
e. ma and d. suematsu , mod .
a * 24 * , 583 ( 2009 ) [ arxiv:0809.0942 [ hep - ph ] ] .
j. kopp , e. t. neil , r. primulando and j. zupan , phys .
dark . univ .
* 2 * ( 2013 ) 22 [ arxiv:1301.1683 [ hep - ph ] ] . c. weniger , jcap * 1208 * , 007 ( 2012 ) [ arxiv:1204.2797 [ hep - ph ] ] .
l. m. krauss , s. nasri and m. trodden , phys .
d * 67 * , 085002 ( 2003 ) [ hep - ph/0210389 ] . c. boehm , y. farzan , t. hambye , s. palomares - ruiz and s. pascoli , phys .
d * 77 * , 043516 ( 2008 ) [ hep - ph/0612228 ] ; m. lindner , d. schmidt and t. schwetz , phys .
b * 705 * , 324 ( 2011 ) [ arxiv:1105.4626 [ hep - ph ] ] ; f. -x .
josse - michaux and e. molinaro , phys .
d * 84 * , 125021 ( 2011 ) [ arxiv:1108.0482 [ hep - ph ] ] ; s. kanemura , t. nabeshima and h. sugiyama , phys .
d * 85 * , 033004 ( 2012 ) [ arxiv:1111.0599 [ hep - ph ] ] ; m. gustafsson , j. m. no and m. a. rivera , arxiv:1212.4806 [ hep - ph ] ; m. aoki , j. kubo and h. takano , arxiv:1302.3936 [ hep - ph ] ; y. kajiyama , h. okada and t. toma , arxiv:1303.7356 [ hep - ph ] . v. silveira and a. zee , phys .
b * 161 * , 136 ( 1985 ) ; j. mcdonald , phys . rev .
d * 50 * ( 1994 ) 3637 [ hep - ph/0702143 [ hep - ph ] ] ; c. p. burgess , m. pospelov and t. ter veldhuis , nucl .
b * 619 * ( 2001 ) 709 [ hep - ph/0011335 ] .
r. barbieri , l. j. hall and v. s. rychkov , phys .
d * 74 * ( 2006 ) 015007 [ hep - ph/0603188 ] ; l. lopez honorez , e. nezri , j. f. oliver and m. h. g. tytgat , jcap * 0702 * ( 2007 ) 028 [ hep - ph/0612275 ] . q. -h .
cao , e. ma and g. rajasekaran , phys .
d * 76 * ( 2007 ) 095011 [ arxiv:0708.2939 [ hep - ph ] ] ; s. andreas , m. h. g. tytgat and q. swillens , jcap * 0904 * ( 2009 ) 004 [ arxiv:0901.1750 [ hep - ph ] ] ; l. lopez honorez and c. e. yaguna , jcap * 1101 * ( 2011 ) 002 [ arxiv:1011.1411 [ hep - ph ] ] ; m. gustafsson , s. rydbeck , l. lopez - honorez and e. lundstrom , phys . rev .
d * 86 * ( 2012 ) 075019 [ arxiv:1206.6316 [ hep - ph ] ] ; s. kashiwase and d. suematsu , phys . rev .
d * 86 * ( 2012 ) 053001 [ arxiv:1207.2594 [ hep - ph ] ] .
m. cirelli , n. fornengo and a. strumia , nucl .
b * 753 * , 178 ( 2006 ) [ hep - ph/0512090 ] ; m. cirelli , a. strumia and m. tamburini , nucl .
b * 787 * , 152 ( 2007 ) [ arxiv:0706.4071 [ hep - ph ] ] .
e. ma , phys .
* 81 * ( 1998 ) 1171 [ hep - ph/9805219 ] .
f. bonnet , m. hirsch , t. ota and w. winter , jhep * 1207 * , 153 ( 2012 ) [ arxiv:1204.5862 [ hep - ph ] ] .
m. fabbrichesi and s. petcov , arxiv:1304.4001 [ hep - ph ] .
k. l. mcdonald , jhep
* 1307 * , 020 ( 2013 ) [ arxiv:1303.4573 [ hep - ph ] ] . k. hally , h. e. logan and t. pilkington , phys .
d * 85 * , 095017 ( 2012 ) [ arxiv:1202.5073 [ hep - ph ] ] .
r. foot , h. lew , x. g. he and g. c. joshi , z. phys .
c * 44 * , 441 ( 1989 ) .
s. -y . ho and j. tandean , arxiv:1303.5700 [ hep - ph ] .
d. schmidt , t. schwetz and t. toma , phys .
d * 85 * , 073009 ( 2012 ) [ arxiv:1201.0906 [ hep - ph ] ] . c. -k .
chua and s. s. c. law , phys .
d * 83 * , 055010 ( 2011 ) [ arxiv:1011.4730 [ hep - ph ] ] ; a. delgado , c. garcia cely ,
t. han and z. wang , phys .
d * 84 * , 073007 ( 2011 ) [ arxiv:1105.5417 [ hep - ph ] ] ; s. s. c. law and k. l. mcdonald , phys .
b * 713 * , 490 ( 2012 ) [ arxiv:1204.2529 [ hep - ph ] ] ; i. baldes , n. f. bell , k. petraki and r. r. volkas , arxiv:1304.6162 [ hep - ph ] ; g. bambhaniya , j. chakrabortty , s. goswami and p. konar , arxiv:1305.2795 [ hep - ph ] .
e. del nobile , r. franceschini , d. pappadopulo and a. strumia , nucl .
b * 826 * , 217 ( 2010 ) [ arxiv:0908.1567 [ hep - ph ] ] .
a. joglekar , p. schwaller and c. e. m. wagner , jhep * 1212 * ( 2012 ) 064 [ arxiv:1207.4235 [ hep - ph ] ] . c. arina , r. n. mohapatra and n. sahu , phys .
b * 720 * ( 2013 ) 130 [ arxiv:1211.0435 [ hep - ph ] ] .
e. aprile _ et al .
_ [ xenon100 collaboration ] , phys . rev .
lett . * 107 * ( 2011 ) 131302 [ arxiv:1104.2549 [ astro-ph.co ] ] ; phys .
lett . * 109 * , 181301 ( 2012 ) [ arxiv:1207.5988 [ astro-ph.co ] ] .
e. m. dolle and s. su , phys .
d * 80 * ( 2009 ) 055012 [ arxiv:0906.1609 [ hep - ph ] ] .
t. hambye , f. -s .
ling , l. lopez honorez and j. rocher , jhep * 0907 * ( 2009 ) 090 [ erratum - ibid .
* 1005 * ( 2010 ) 066 ] [ arxiv:0903.4010 [ hep - ph ] ] .
t. araki , c. q. geng and k. i. nagao , phys .
d * 83 * ( 2011 ) 075014 [ arxiv:1102.4906 [ hep - ph ] ] ; t. araki , c. q. geng and k. i. nagao , int .
j. mod .
d * 20 * ( 2011 ) 1433 [ arxiv:1108.2753 [ hep - ph ] ] .
r. foot , a. kobakhidze , k. l. mcdonald and r. r. volkas , phys .
d * 76 * ( 2007 ) 075014 [ arxiv:0706.1829 [ hep - ph ] ] ; r. foot , a. kobakhidze , k. l. mcdonald and r. r. volkas , phys .
d * 77 * , 035006 ( 2008 ) [ arxiv:0709.2750 [ hep - ph ] ] .
y. kajiyama , h. okada and k. yagyu , arxiv:1303.3463 [ hep - ph ] . s. kanemura and h. sugiyama , phys .
d * 86 * ( 2012 ) 073006 [ arxiv:1202.5231 [ hep - ph ] ] .
a. djouadi , o. lebedev , y. mambrini and j. quevillon , phys .
b * 709 * , 65 ( 2012 ) [ arxiv:1112.3299 [ hep - ph ] ] .
k. kumericki , i. picek and b. radovcic , phys .
d * 86 * , 013006 ( 2012 ) [ arxiv:1204.6599 [ hep - ph ] ] ; i. picek and b. radovcic , phys .
b * 719 * , 404 ( 2013 ) [ arxiv:1210.6449 [ hep - ph ] ] .
the same model is mentioned in y. liao , jhep * 1106 * , 098 ( 2011 ) [ arxiv:1011.3633 [ hep - ph ] ] .
m. cirelli and a. strumia , new j. phys .
* 11 * ( 2009 ) 105005 [ arxiv:0903.3381 [ hep - ph ] ] . w. konetschny and w. kummer , phys .
b * 70 * , 433 ( 1977 ) ; t. p. cheng and l. f. li , phys .
d * 22 * , 2860 ( 1980 ) ; m. magg and c. wetterich , phys .
b * 94 * , 61 ( 1980 ) ; j. schechter and j. w. f. valle , phys . rev .
d * 22 * , 2227 ( 1980 ) ; g. lazarides , q. shafi and c. wetterich , nucl .
b * 181 * , 287 ( 1981 ) ; c. wetterich , nucl .
b * 187 * , 343 ( 1981 ) ; r. n. mohapatra and g. senjanovic , phys .
d * 23 * , 165 ( 1981 ) . s. s. c. law and k. l. mcdonald , arxiv:1303.6384 [ hep - ph ] .
x. -g .
he , eur .
j. c * 34 * , 371 ( 2004 ) [ hep - ph/0307172 ] .
k. s. babu , s. nandi and z. tavartkiladze , phys .
d * 80 * , 071702 ( 2009 ) [ arxiv:0905.2710 [ hep - ph ] ] .
i. picek and b. radovcic , phys .
b * 687 * , 338 ( 2010 ) [ arxiv:0911.1374 [ hep - ph ] ] ; k. kumericki , i. picek and b. radovcic , phys .
d * 84 * , 093002 ( 2011 ) [ arxiv:1106.1069 [ hep - ph ] ] .
|
we present a class of models with radiative neutrino mass and stable dark - matter candidates .
neutrino mass is generated by a one - loop diagram with the same topography as ma s 2006 proposal ( which used an inert scalar - doublet and singlet fermion ) .
we generalize this approach and determine all variants with new fields no larger than the adjoint representation .
when the neutrino mass diagram contains a majorana mass insertion there are two possibilities , both of which are known .
if the mass insertion is of the dirac type there are seven additional models , two of which are excluded by direct - detection experiments .
the other five models are also constrained , such that only scalar dark - matter is viable .
there are cases with an inert singlet , an inert doublet , and an inert triplet , providing a natural setting for inert @xmath0-tuplet models of dark matter , with the additional feature of achieving radiative neutrino mass .
we show that some of the models admit a simple explanation for the ( requisite ) discrete symmetry , and briefly discuss cases with representations larger than the adjoint , which can admit a connection to the astrophysical gamma - ray signal . a class of inert n - tuplet models with radiative neutrino mass and dark matter + sandy s. c. law@xmath1 and kristian l. mcdonald@xmath2 + * department of physics , national cheng - kung university , + tainan 701 , taiwan + @xmath3 arc centre of excellence for particle physics at the terascale , + school of physics , the university of sydney , nsw 2006 , australia +
|
the concept of @xmath2-ordered hamiltonian graphs has been recently introduced by ng and schultz @xcite .
a simple graph @xmath0 is @xmath2-_ordered _ ( resp .
@xmath2-_ordered hamiltonian _ ) if for any sequence @xmath3 , @xmath4 ,
, @xmath5 of @xmath2 vertices of @xmath0 there is a cycle ( resp . a hamiltonian cycle ) in @xmath0 containing these vertices in the given order .
note that being @xmath12-ordered hamiltonian is equivalent to being hamiltonian .
a natural direction of research related to this new hamiltonian property is to generalize existing results implying graph hamiltonicity and obtain results implying @xmath2-ordered hamiltonicity . in @xcite ng and schultz
generalize classical theorems of dirac and ore and give minimum vertex degree conditions that guarantee a graph is @xmath2-ordered hamiltonian .
these conditions were improved by faudree @xcite .
another series of results appearing in @xcite describes various forbidden subgraphs that force a graph to be @xmath2-ordered or @xmath2-ordered hamiltonian .
there are many open questions about whether these properties can be ensured by sufficient connectivity in a graph ( see @xcite ) . in this paper
we extend a well - known result on hamiltonicity of the third power of a simple graph , which is defined below . [ g_power ]
let @xmath0 be a simple graph with vertex set @xmath13 and edge set @xmath14 .
the @xmath15-_th power _ of @xmath0 , denoted @xmath16 , is the simple graph with the same vertex set @xmath13 and with the edge set @xmath17 our goal is to explore the @xmath2-ordered hamiltonicity of graph powers . in section [ ppaths ]
we give bounds on the smallest power of the path @xmath18 that is @xmath2-ordered hamiltonian . in section
[ main ] we prove the main theorem , which states that @xmath19 is @xmath2-ordered hamiltonian for a connected graph @xmath0 on at least @xmath2 vertices . in section
[ four ] we prove the 4-ordered hamiltonicity of the fourth power of a connected graph . in section
[ cycles ] we discuss the @xmath2-ordered hamiltonicity of powers of the cycle @xmath20 . in section [ conclusion ]
we pose several open questions related to our results .
all graphs considered in this paper are finite simple graphs .
the distance between vertices @xmath21 , @xmath22 of a graph @xmath0 is denoted by @xmath23 .
the number of vertices of @xmath0 is denoted by @xmath24 . if @xmath25 is a path with endpoints @xmath26 and @xmath27 , then @xmath28 denotes the path @xmath25 starting at @xmath26 and ending at @xmath27 . a graph @xmath0 is _ hamilton - connected _ if for any pair @xmath29 of vertices of @xmath0 there is a path in @xmath0 between @xmath21 and @xmath22 containing all vertices of @xmath0 .
such path is called a _
hamiltonian path_. the following theorem is often included as an exercise in graph theory textbooks .
[ g3c ] if @xmath0 is a connected graph on 2 or more vertices , then @xmath1 is hamilton - connected .
let @xmath30 be a spanning tree of @xmath0 .
notice that it suffices to show that @xmath31 is hamilton - connected since @xmath31 is a subgraph of @xmath1 with the same vertex set .
we will show by induction that if @xmath30 is a tree , then @xmath31 is hamilton - connected .
this is obvious if @xmath30 has only 2 vertices .
suppose that the assertion is true for trees with fewer than @xmath32 vertices .
let @xmath3 and @xmath4 be distinct vertices of @xmath30 .
choose an edge @xmath33 of @xmath30 such that @xmath34 consists of two connected components @xmath35 and @xmath36 satisfying @xmath37 and @xmath38 . for @xmath39 let @xmath40 if
@xmath41 or if @xmath42 , otherwise let @xmath43 be a vertex of @xmath44 such that @xmath45 .
let @xmath46 be a hamiltonian path in @xmath47 between @xmath48 and @xmath43 ( if @xmath49 , then @xmath46 is the `` empty '' path starting and ending at @xmath48 ) .
notice that @xmath50 , so @xmath51 is a hamiltonian path in @xmath31 between @xmath3 and @xmath4 .
since hamilton - connectedness implies hamiltonicity , we have the following immediate corollary : [ g3 ] if @xmath0 is a connected graph on 3 or more vertices , then @xmath1 is hamiltonian .
the proof of theorem [ g3c ] has been included for two reasons .
first , it uses the fact that it suffices to prove the statement only for trees instead of arbitrary graphs .
we will use the same idea in the proof of theorem [ g4 ] .
second , it is based on a simple induction argument , which can not be applied if we need to keep track of the order in which certain vertices are contained in the cycle .
the following theorem about 2-connected graphs was proved by fleischner @xcite in 1974 .
a simpler proof can be found in @xcite .
[ g2 ] if @xmath0 is a 2-connected graph on 3 or more vertices , then @xmath52 is hamiltonian .
another result relevant to the discussion of this paper was proved by lou , xu , and yao @xcite .
[ 2edges ] let @xmath0 be a connected graph on 3 or more vertices .
for any two edges @xmath53 and @xmath54 of @xmath0 , there is a hamiltonian cycle in @xmath7 containing @xmath53 and @xmath54 . for @xmath55 and a graph @xmath0 ,
let @xmath56 be the smallest integer @xmath57 such that @xmath58 is @xmath2-ordered hamiltonian . in this paper
we give bounds on @xmath56 for an arbitrary connected graph @xmath0 and in the special cases of @xmath0 being a path or a cycle .
let @xmath18 be the path on @xmath15 vertices . in this section we present bounds on @xmath59 .
[ pathlowerbound ] for @xmath55 and @xmath60 , the inequality @xmath61 holds . first , consider the case of even @xmath2 , and let @xmath62 .
we show that @xmath63 is not @xmath64-ordered .
let @xmath3 , @xmath65 ,
, @xmath66 , @xmath67 , @xmath68 , , @xmath69 , @xmath4 , @xmath70 , , @xmath71 be consecutive vertices of the path @xmath18 .
suppose that @xmath72 is a cycle in @xmath63 containing the vertices @xmath3 , @xmath4 ,
, @xmath71 in order . for @xmath73 ,
let @xmath74 denote the part of @xmath72 between , and including , @xmath48 and @xmath75 ( indices taken modulo @xmath64 ) .
put @xmath76 . since @xmath77 , there is an index @xmath78 such that @xmath74 does not contain a vertex of @xmath79 . in @xmath18 , the set @xmath79 is located between @xmath48 and @xmath75 ,
so @xmath74 must contain an edge @xmath80 that `` skips over '' @xmath79 .
the edge @xmath80 connects one of @xmath3 , @xmath65 , ,
@xmath66 and one of @xmath4 , @xmath70 , , @xmath71 because @xmath81 .
also , the edge @xmath80 connects two non - consecutive vertices of @xmath3 , @xmath4 ,
, @xmath71 since the distance in @xmath18 between two consecutive vertices of this sequence is at least @xmath82 .
thus , the cycle @xmath72 can not contain @xmath80 , which contradicts @xmath74 containing @xmath80 .
we treat the case of odd @xmath2 in a similar way .
let @xmath83 .
we show that @xmath84 is not @xmath85-ordered .
let @xmath3 , @xmath65 ,
, @xmath86 , @xmath67 , @xmath68 , , @xmath69 , @xmath4 , @xmath70 , , @xmath71 be consecutive vertices of the path @xmath18 .
suppose that @xmath72 is a cycle in @xmath84 containing the vertices @xmath3 , @xmath4 ,
, @xmath86 in order . as in the previous case
, we find an edge @xmath80 of @xmath72 connecting one of @xmath3 , @xmath65 , , @xmath86 and one of @xmath4 , @xmath70 ,
, @xmath71 .
again , this edge connects two non - consecutive vertices of @xmath3 , @xmath4 , , @xmath86 since @xmath87 for @xmath88 .
we obtain a similar contradiction .
we conclude that @xmath89 and @xmath90 .
the theorem follows .
next , we present a construction which shows that @xmath91 is @xmath2-ordered hamiltonian .
[ pathupperbound ] for @xmath92 , the inequality @xmath93 holds .
let @xmath94 .
we show that @xmath95 is @xmath2-ordered hamiltonian .
label consecutive vertices of @xmath18 with consecutive integers between @xmath96 and @xmath15 .
let @xmath3 , @xmath4 ,
, @xmath5 be a sequence of @xmath2 vertices of @xmath18 .
we view the @xmath48 s as both numbers and vertices .
first , we construct a cycle @xmath72 in @xmath95 containing the vertices @xmath3 , @xmath4 , ,
@xmath5 in order .
we mark some of the @xmath48 s using the following procedure : 1 .
[ step1 ] initially , put @xmath97 , where @xmath98 and @xmath99 are the smallest and the largest among the @xmath48 s , respectively .
mark each @xmath100 for which no other element of @xmath13 is congruent to @xmath48 modulo @xmath101 .
2 . [ step2 ]
now , add @xmath98 to @xmath13 , and mark @xmath98 if no other element of @xmath13 is congruent to @xmath98 modulo @xmath101 .
[ step3 ] repeat step [ step2 ] with @xmath99 .
( note that if @xmath102 and @xmath98 is marked at step [ step2 ] , then @xmath99 is not marked . )
let @xmath103 denote the set of residues modulo @xmath101 present among elements of @xmath13 .
also , let @xmath104 be the number of elements of @xmath13 that are not marked .
[ justalemma ] after the above procedure , the inequality @xmath105 holds .
consider the situation immediately after step [ step1 ] , when @xmath106 .
let @xmath107 be the number of residues modulo @xmath101 that appear exactly once among the elements of @xmath13 , and let @xmath108 be the number of residues that appear more than once .
then @xmath109 and @xmath110 since exactly @xmath107 elements are marked at step [ step1 ] .
thus @xmath111 each of the @xmath108 `` repeated '' residues appears at least twice , so @xmath112 . substituting this relation into the above equation , we get @xmath113 , either @xmath98 adds a new residue to @xmath13 , or @xmath98 is not marked , but not both .
thus , the sum @xmath114 is increased by @xmath96 .
similarly , at step [ step3 ] the sum @xmath114 is increased by @xmath96 again . hence after the entire procedure , we have @xmath115 where @xmath13 is now the set of all the @xmath48 s .
let @xmath116 , , @xmath117 be the vertices of @xmath13 that are not marked .
select @xmath104 distinct residues @xmath118 , , @xmath119 modulo @xmath101 not contained in @xmath103 .
note that we can choose these residues because @xmath120 .
set @xmath121 note that the residues @xmath122 , , @xmath123 are all distinct since at most one instance of each residue modulo @xmath101 in @xmath13 is marked . for @xmath124 ,
let @xmath74 be the path in @xmath95 from @xmath48 to @xmath75 whose interior vertices are all the vertices of @xmath18 between @xmath48 and @xmath75 congruent to @xmath125 modulo @xmath101 ( indices taken modulo @xmath2 ) .
let @xmath72 be the cycle obtained by joining the paths @xmath126 , @xmath127 ,
, @xmath128 .
clearly , @xmath72 contains the vertices @xmath3 , , @xmath5 in order .
we need to check that the paths @xmath74 are interior vertex - disjoint and that no path @xmath74 contains a vertex of @xmath13 in its interior .
the former property of the @xmath74 s follows from the fact that the values of @xmath125 are all distinct .
we check the latter property .
suppose that @xmath74 contains a vertex @xmath129 in its interior .
then @xmath130 if @xmath48 is marked , then @xmath131 , so , by the above congruence , @xmath48 can not be marked .
if @xmath48 is not marked , then @xmath132 , which contradicts the above congruence . finally , we extend @xmath72 to a hamiltonian cycle in @xmath95 .
if @xmath72 does not contain all vertices of @xmath18 , choose a vertex @xmath133 adjacent in @xmath18 to a vertex @xmath134 .
let @xmath22 and @xmath135 be the two vertices adjacent to @xmath136 in @xmath72 .
it is easy to check that either @xmath137 or @xmath138 is at most @xmath101 . without loss of generality ,
let @xmath139 .
we can add @xmath140 to @xmath72 by replacing the edge @xmath141 with the path @xmath142 to produce a new cycle @xmath143 in @xmath95 .
clearly , @xmath143 still contains the vertices @xmath3 , , @xmath5 in order . in this way we can add all vertices of @xmath18 to the cycle
we now prove our main result which gives an upper bound on @xmath56 for a connected graph @xmath0 on at least @xmath2 vertices .
we begin by proving three technical lemmas .
[ tl0 ] let @xmath144 be a tree on at least @xmath145 vertices , and let @xmath22 be a vertex of @xmath144 .
then @xmath146 has a hamiltonian path whose endpoints @xmath147 and @xmath148 satisfy @xmath149 and @xmath150 .
let @xmath151 , , @xmath152 be the connected components of @xmath153 .
let @xmath154 be the vertex of @xmath155 adjacent to @xmath22 in @xmath144 , and let @xmath156 be a vertex of @xmath155 adjacent to @xmath154 , or equal to @xmath154 if @xmath157 .
let @xmath158 be a hamiltonian path in @xmath159 starting at @xmath154 and ending at @xmath156 ; such a path exists by theorem [ g3c ] .
note that @xmath160 for @xmath161 , so the concatenation @xmath162 yields the desired path with @xmath163 and @xmath164 .
[ tl1 ] let @xmath2 and @xmath165 be positive integers , and let @xmath0 be a connected graph on at least @xmath2 vertices .
suppose that , for every sequence @xmath166 of @xmath2 vertices of @xmath0 , there exists a subtree @xmath167 and a cycle @xmath72 in @xmath168 satisfying the following conditions : 1 .
@xmath72 contains @xmath3 , ,
@xmath5 in order ; 2 .
@xmath72 contains every leaf of @xmath79 ; 3 .
if @xmath136 is a leaf of @xmath79 , then @xmath136 is adjacent in @xmath72 to a vertex @xmath135 such that either @xmath169 , or @xmath170 and @xmath135 is not a leaf of @xmath79 .
then @xmath58 is @xmath2-ordered hamiltonian .
we begin by extending @xmath72 to a hamiltonian cycle in @xmath168 .
let @xmath135 be a vertex of @xmath79 such that @xmath171 . by condition ( ii ) ,
@xmath135 is not a leaf .
therefore , @xmath172 has at least two connected components , and some part of @xmath72 between a leaf in one component and a leaf in another component does not contain @xmath135 .
this part contains an edge @xmath173 such that @xmath135 lies on the unique path in @xmath79 between @xmath136 and @xmath140 .
let @xmath143 be the cycle obtained from @xmath72 by replacing the edge @xmath173 with the path @xmath174 .
clearly , @xmath143 satisfies conditions ( i ) and ( ii ) . since @xmath175
, it follows that @xmath176 and @xmath177 , hence @xmath143 satisfies condition ( iii ) . in this way
, we can add all remaining vertices of @xmath79 to obtain a hamiltonian cycle @xmath178 of @xmath168 satisfying conditions ( i)(iii ) with @xmath72 replaced by @xmath179 .
let @xmath30 be a spanning tree of @xmath0 containing @xmath79 .
consider the graph @xmath180 obtained by removing the edges of @xmath79 from @xmath30 .
let @xmath181 , , @xmath182 be the connected components of @xmath180 such that @xmath183 .
let @xmath43 denote the unique vertex of @xmath184 .
note that the vertices @xmath67 , , @xmath185 are distinct .
put @xmath186 and @xmath187 for @xmath188 , so that @xmath189 .
we construct a sequence @xmath190 , @xmath191 , , @xmath192 , where @xmath193 is a hamiltonian cycle of @xmath194 satisfying the following conditions : 1 .
@xmath193 contains the vertices @xmath3 , , @xmath5 in order ; 2 .
if @xmath136 is a common leaf of @xmath79 and @xmath44 , then @xmath136 is adjacent in @xmath193 to a vertex @xmath135 such that either @xmath195 , or @xmath196 and @xmath135 is not a leaf of @xmath79 . the cycle @xmath197 satisfies the above conditions .
we construct @xmath198 by inserting the vertices of @xmath199 between two consecutive vertices of @xmath193 . by lemma [ tl0 ]
, there is a hamiltonian path @xmath200 in @xmath201 whose endpoints @xmath147 and @xmath148 satisfy @xmath202 and @xmath203 .
we consider two cases .
case 1 : @xmath204 is not a leaf of @xmath44 .
then @xmath205 has at least two connected components , and one of the parts of @xmath193 between a vertex in one component and a vertex in another other component does not contain @xmath204 .
this part contains an edge @xmath173 such that @xmath204 belongs to the unique path between @xmath136 and @xmath140 in @xmath44 . without loss of generality ,
suppose that @xmath206 .
let @xmath198 be the cycle obtained by replacing the edge @xmath173 with the path @xmath207 in @xmath193 .
then @xmath198 is a hamiltonian cycle in @xmath208 because @xmath209 and @xmath210 .
since @xmath198 contains the vertices of @xmath193 in the same order as @xmath193 , it follows that @xmath198 satisfies condition ( i ) . to show that @xmath198 satisfies condition ( ii )
, we need to consider the case when @xmath136 or @xmath140 is a common leaf of @xmath79 and @xmath211 .
if @xmath136 is a common leaf of @xmath79 and @xmath211 , then @xmath136 is also a leaf of @xmath44 , so either @xmath212 , or the vertex @xmath213 adjacent to @xmath136 in @xmath193 has the property of condition ( ii ) . in the former case
, we have @xmath214 , and @xmath147 is not a leaf of @xmath79 . in the latter case , the vertex @xmath135 , which is adjacent to @xmath136 in @xmath198 , has the required property .
the case when @xmath140 is a common leaf of @xmath79 and @xmath211 is treated similarly .
case 2 : @xmath204 is a leaf of @xmath44 . then , by condition ( ii ) , @xmath204 is adjacent in @xmath193 to a vertex @xmath135 such that either @xmath215 , or @xmath216 and @xmath135 is not a leaf of @xmath79 .
let @xmath198 be the cycle obtained by replacing the edge @xmath217 with the path @xmath218 in @xmath193 .
then @xmath198 is a hamiltonian cycle in @xmath208 because @xmath219 and @xmath220 . as in the previous case
, it is easy to see that @xmath198 satisfies condition ( i ) . to show that @xmath198 satisfies condition ( ii )
, we need to consider the case when @xmath135 is a common leaf of @xmath79 and @xmath211 . in this case
, @xmath135 is also a leaf of @xmath44 , so either @xmath221 , or the vertex @xmath222 adjacent to @xmath135 in @xmath193 has the property of condition ( ii ) . in the former case
, we have @xmath223 , and @xmath147 is not a leaf of @xmath79 , and in the latter case the vertex @xmath224 , which is adjacent to @xmath135 in @xmath198 , has the required property . since @xmath189
, it follows that @xmath192 is a hamiltonian cycle in @xmath225 , and hence in @xmath58 , containing @xmath3,
,@xmath5 in order . the lemma follows .
[ tl2 ] for @xmath226 , let @xmath0 be a connected graph on at least @xmath101 vertices .
then there exists a map @xmath227 such that , for every two vertices @xmath136 and @xmath140 and every integer @xmath228 , there exists a sequence @xmath229 , @xmath230 , , @xmath231 , @xmath232 of distinct vertices such that @xmath233 for @xmath234 , @xmath235 for @xmath236 , and @xmath237 ( resp .
, @xmath238 ) if @xmath239 ( resp . , @xmath240 ) .
let @xmath30 be a spanning tree of @xmath0 .
since @xmath241 for all @xmath136 and @xmath242 , it suffices to prove the lemma with @xmath0 replaced by @xmath30 .
we define a map @xmath243 such that , for every vertex @xmath136 , the following conditions are satisfied : 1 .
there is a sequence @xmath244 , , @xmath245 of vertices such that @xmath246 for @xmath247 , and the sequence @xmath248,
,@xmath249 contains all elements of @xmath250 ; 2 . if there is a vertex @xmath251 such that @xmath252 , then there is a vertex @xmath253 such that @xmath254 and @xmath255 .
let @xmath256 be a subtree of @xmath30 with @xmath257 .
construct a sequence @xmath258 of subtrees of @xmath30 as follows : for @xmath259 , choose a vertex @xmath260 adjacent to a vertex @xmath261 , and let @xmath262 be the tree obtained by adjoining @xmath263 to @xmath264 by means of the edge @xmath265 .
first , we choose a different value of @xmath266 for each @xmath267 .
clearly , this assignment satisfies condition ( a ) above : we can take @xmath268 to be the @xmath78-th closest vertex of @xmath256 to @xmath136 . for @xmath259 ,
define @xmath269 as follows .
let @xmath270 , , @xmath271 be the sequence described in condition ( a ) .
set @xmath272 and @xmath273 for @xmath247 .
set @xmath274 .
it is easy to verify that the sequence @xmath275 , , @xmath276 satisfies conditions ( a ) and ( b ) .
also , note that if , before @xmath269 was defined , @xmath277 was the only vertex mapped by @xmath278 to @xmath266 , then we can set @xmath279 to satisfy condition ( b ) for the vertex @xmath136 .
we now prove by induction that for every @xmath280 and every two vertices @xmath136 and @xmath140 of @xmath264 , there exists a sequence @xmath281 of vertices of @xmath264 satisfying the conditions of the lemma .
this assertion is true for @xmath282 because we can set @xmath283 and @xmath284 since @xmath285 .
suppose the assertion is true for some @xmath286 .
let @xmath136 and @xmath140 be two vertices of @xmath262 , and let @xmath287 be a element of @xmath288
. without loss of generality , assume that @xmath289 .
if @xmath290 , then we can find the desired sequence in @xmath264 . now
, suppose that @xmath291 .
let @xmath242 be the element of the sequence @xmath244 , @xmath292 , @xmath293 such that @xmath294 .
note that in our construction , @xmath293 is defined since @xmath295 , so @xmath242 is well defined . by the inductive hypothesis , there is a sequence @xmath296 in @xmath264 such that @xmath297 for @xmath298 , @xmath235 for @xmath299 , and @xmath300 if @xmath240 .
then the sequence @xmath301 satisfies the conditions of the lemma because @xmath302 , and @xmath303 if @xmath304 .
therefore , the assertion is true for @xmath305 , and setting @xmath306 yields the lemma .
we now prove the main theorem .
[ maintheorem ] for @xmath55 , let @xmath0 be a connected graph on at least @xmath2 vertices . then @xmath307 is @xmath2-ordered , and @xmath308 is @xmath2-ordered hamiltonian .
put @xmath309 .
let @xmath310 be a sequence of @xmath2 vertices of @xmath0 .
let @xmath30 be a spanning tree of @xmath0 , and let @xmath79 be the smallest subtree of @xmath30 containing all the @xmath48 s .
then all leaves of @xmath79 are among the @xmath48 s .
let @xmath311 be a map satisfying the conditions of lemma [ tl2 ] for @xmath312 .
first , we construct a cycle @xmath72 in @xmath313 containing the vertices @xmath3 , ,
@xmath5 in order .
put @xmath314 , and mark each @xmath100 for which no other element @xmath315 satisfies @xmath316 .
let @xmath317 , and let @xmath104 be the number of unmarked vertices . [ justalemma2 ] @xmath318 .
the inequality is derived in the same way as the inequality ( [ eq13 ] ) in the proof of lemma [ justalemma ] .
values of @xmath278 play the role of residues modulo @xmath101 . let @xmath116 , , @xmath117 be the unmarked elements of @xmath13 .
choose a sequence @xmath118 ,
, @xmath119 of distinct elements of @xmath319 .
such a sequence exists because @xmath320 note that @xmath122,
,@xmath123 are all distinct since the values of @xmath321 are distinct for all marked vertices @xmath48 . for @xmath124 ,
let @xmath74 be the path in @xmath313 between @xmath48 and @xmath75 traversing the sequence @xmath322 obtained by applying lemma [ tl2 ] to @xmath323 , @xmath324 , and @xmath325 ( indices taken modulo @xmath2 ) . for @xmath326 , we have @xmath327 , so the values of @xmath278 are different for interior vertices of @xmath74 and interior vertices of @xmath328 , hence @xmath74 and @xmath328 are interior vertex disjoint .
also , no path @xmath74 contains a vertex @xmath329 in its interior . indeed ,
if @xmath48 is marked , then @xmath330 for every interior vertex @xmath135 of @xmath74 , so @xmath331 . if @xmath48 is not marked , then @xmath332 for every interior vertex @xmath135 of @xmath74 , so @xmath331 .
it follows that joining the paths @xmath126 , , @xmath128 yields a cycle @xmath72 in @xmath313 containing the vertices @xmath3 , ,
@xmath5 in order .
since @xmath72 is a cycle in @xmath333 , it follows that @xmath333 is @xmath2-ordered .
note that @xmath72 contains the leaves of @xmath79 because all of them are among the @xmath48 s .
also , note that for a leaf @xmath48 of @xmath79 , we have either @xmath334 or @xmath335 , so the vertex adjacent to @xmath48 in one of @xmath336 and @xmath74 is at most distance @xmath337 away from @xmath48 .
applying lemma [ tl1 ] with @xmath338 , we conclude that @xmath339 is @xmath2-ordered hamiltonian .
[ maincor ] for a connected graph @xmath0 on at least @xmath2 vertices , the inequality @xmath340 holds .
the obtained upper bound on @xmath56 for an arbitrary connected graph @xmath0 is not tight for small @xmath2 . indeed ,
corollary [ maincor ] states that @xmath341 and @xmath342 , whereas corollary [ g3 ] implies @xmath343 , and our next result yields @xmath344 .
[ g4 ] if @xmath0 is a connected graph on 4 or more vertices , then @xmath7 is 4-ordered hamiltonian . as in the proof of theorem [ g3c ]
, it suffices to show that @xmath345 is 4-ordered hamiltonian , where @xmath30 is a spanning tree of @xmath0 .
the argument will be based on the following technical lemma which sets up an application of lemma [ tl1 ] .
[ t - bar ] let @xmath346 be a tree with at most 4 leaves , and let @xmath3 , @xmath4 , @xmath65 , @xmath70 be a sequence of 4 vertices of @xmath346 such that all leaves of @xmath346 are among @xmath3 , , @xmath70 .
then there is a hamiltonian cycle @xmath72 in @xmath347 containing @xmath3 , , @xmath70 in the given order and having the following property : for each vertex @xmath48 that is a leaf of @xmath346 , there is a vertex @xmath348 adjacent to @xmath48 in @xmath72 such that @xmath349 or else @xmath350 and @xmath348 is not a leaf of @xmath346 .
figure [ types ] shows the three possible shapes of @xmath346 and locations of @xmath3 , ,
. we will consider each case separately .
case 1 : @xmath346 has 2 leaves , i.e. @xmath346 is a path .
let @xmath147 , @xmath148 , @xmath351 , @xmath352 be the vertices @xmath3 , @xmath4 , @xmath65 , @xmath70 in the order in which they appear in @xmath346 , so that @xmath147 and @xmath352 are the two leaves .
we will construct 6 internally disjoint paths @xmath353 in @xmath347 , @xmath354 , where @xmath353 is a path between @xmath348 and @xmath355 and @xmath356 by convention .
we start by choosing integers @xmath26 and @xmath27 such that @xmath357 and @xmath358 .
then we choose an integer @xmath287 such that @xmath359 and @xmath360 . finally , we choose an integer @xmath361 such that @xmath362 , @xmath363 , and @xmath364 .
let @xmath365 be the path @xmath366 in @xmath347 , where @xmath367 let @xmath368 be the path @xmath369 in @xmath347 , where @xmath370 let @xmath371 be the path @xmath372 in @xmath347 , where @xmath373 figure [ paths ] shows the paths @xmath365 , @xmath368 , and @xmath371 .
the conditions on @xmath26 , @xmath27 , @xmath287 , and @xmath361 ensure that these paths are internally disjoint .
notice also that we allow @xmath2 , @xmath374 , @xmath108 , @xmath57 , @xmath375 , @xmath376 , and @xmath107 to be zero ( for example , if @xmath377 , then @xmath378 ) .
let @xmath379 , @xmath380 , and @xmath381 be the paths in @xmath347 consisting of all vertices of @xmath346 between @xmath147 and @xmath148 , @xmath148 and @xmath351 , and @xmath351 and @xmath352 , respectively , that are not contained in @xmath365 , @xmath368 , and @xmath371 ( the resulting paths are valid since each of @xmath365 , @xmath368 , and @xmath371 uses every fourth vertex on each of the three intervals )
. now denote @xmath382 and form a cycle @xmath72 in @xmath347 containing @xmath3 , , @xmath70 in the given order by linking the paths @xmath383 , @xmath384 , @xmath385 , and @xmath386 .
this cycle can be easily extended to a hamiltonian cycle since every vertex @xmath387 lies between vertices @xmath101 and @xmath388 that are adjacent in @xmath72 , so @xmath21 can be inserted between @xmath101 and @xmath388 .
the conditions of the lemma are satisfied since every leaf of @xmath346 is adjacent in @xmath72 to a non - leaf that is at most distance 3 away .
case 2 : @xmath346 has 3 leaves . without loss of generality , assume that @xmath3 , @xmath4 , and @xmath65 are the leaves of @xmath346 .
let @xmath389 be the only vertex of @xmath346 of degree 3 .
we consider two subcases .
case 2.1 : @xmath70 lies between @xmath3 and @xmath389 ( the case when @xmath70 lies between @xmath65 and @xmath389 is analogous ) or @xmath390 .
let @xmath391 be the path in @xmath346 between @xmath3 and @xmath4 , and let @xmath392 be the hamiltonian cycle in @xmath393 . the part of @xmath392 between @xmath4 and @xmath70 that does not contain @xmath3 contains either the vertex @xmath389 or two vertices @xmath388 and @xmath22 that are adjacent to @xmath389 in @xmath391 ( suppose that @xmath388 is closer to @xmath4 than @xmath22 ) .
let @xmath394 be the path in @xmath346 between @xmath389 and @xmath65 , and let @xmath395 be the hamiltonian path in @xmath396 between vertices @xmath136 and @xmath135 , where @xmath397 and @xmath398 ( if @xmath65 is the only vertex of @xmath399 , then let @xmath395 be the `` empty '' path starting and ending at @xmath65 ) . form the hamiltonian cycle @xmath72 in @xmath347 as follows : start by going from @xmath3 to @xmath4 along the part of @xmath392 that does not contain @xmath70 , then continue along @xmath392 until either @xmath388 or the vertex preceding @xmath389 in @xmath392 is encountered , then proceed to @xmath136 and go to @xmath135 along @xmath395 , and finally , proceed to @xmath22 or @xmath389 and finish the cycle by going to @xmath3 along the remaining part of @xmath392 ( see figure [ cyc21 ] ) .
clearly , the conditions of the lemma are satisfied since every leaf of @xmath346 is adjacent in @xmath72 to a vertex of @xmath346 that is at most distance 2 away .
case 2.2 : @xmath70 lies between @xmath4 and @xmath389
. for @xmath400 let @xmath401 denote the path in @xmath346 between @xmath48 and @xmath329 .
let @xmath402 , @xmath43 , and @xmath348 denote vertices of @xmath403 , @xmath404 , and @xmath405 , respectively , such that @xmath406 .
let @xmath126 be the hamiltonian path in @xmath407 between @xmath389 and @xmath408 .
let @xmath127 be the hamiltonian path in @xmath409 between @xmath410 and @xmath411 , where @xmath412 and @xmath413 ( if @xmath4 is the only vertex of @xmath414 , then let @xmath127 be the `` empty '' path starting and ending at @xmath4 ) .
let @xmath415 be the hamiltonian path in @xmath416 between @xmath147 and @xmath148 ( again , if @xmath65 is the only vertex of @xmath417 , then let @xmath415 be the `` empty '' path starting and ending at @xmath65 ) .
below is the procedure to construct a hamiltonian cycle @xmath72 in @xmath347 with the desired properties . 1 .
start at @xmath3 .
if @xmath418 is divisible by 4 , then go along @xmath126 to @xmath408 , otherwise go along @xmath126 to @xmath389 .
2 . if @xmath419 , then use the path @xmath420 or the path @xmath421 to get within distance 3 of @xmath70 .
3 . proceed to @xmath410 and go along @xmath127 to @xmath411 ( or stay at @xmath410 if @xmath422 ) .
if @xmath423 , then proceed to either @xmath424 or @xmath425 , where @xmath426 and @xmath427 ( one of @xmath424 and @xmath425 has not been used yet ) and use the path @xmath428 or @xmath429 to get to one of @xmath67 , @xmath68 , or @xmath430 ( notice that it is impossible to encounter @xmath389 at this step since @xmath389 was encountered in step 1 unless @xmath418 is divisible by 4 , in which case @xmath68 or @xmath430 will be encountered ) .
proceed to @xmath147 and go along @xmath415 to @xmath148 ( or stay at @xmath147 if @xmath431 ) .
6 . if @xmath419 , then proceed to @xmath389 , @xmath67 , or @xmath68 ( one of these has not been used yet ) , and follow the path @xmath432 , @xmath433 , or @xmath434 to get within distance 4 of @xmath70 .
proceed to @xmath70 .
if @xmath435 , then go to step 9 .
one of the vertices @xmath424 , @xmath425 , @xmath436 , @xmath437 has not been used yet ; let @xmath438 be this vertex ( @xmath439 ) .
use the path @xmath440 to get to one of @xmath389 , @xmath67 , @xmath68 , @xmath430 .
if @xmath389 was encountered in step 1 , then proceed to @xmath408 and finish the cycle by going along the remaining part of @xmath126 to @xmath3 .
otherwise , proceed to @xmath389 and go along the remaining part of @xmath126 to @xmath3 .
let us verify that the conditions of the lemma are satisfied .
the leaf @xmath3 is adjacent in @xmath72 to two vertices of the path @xmath391 , at least one of which is a non - leaf that is at most distance 3 away .
the same is true of @xmath4 unless @xmath441 and @xmath442 , in which case @xmath4 is adjacent in @xmath72 to @xmath389 or @xmath410 , one of which is at most distance 2 away from @xmath4 .
finally , @xmath65 is adjacent in @xmath72 to two vertices of the path @xmath443 , one of which is a non - leaf that is at most distance 3 away .
case 3 : @xmath346 has 4 leaves @xmath3 , @xmath4 , @xmath65 , and @xmath70 . there is either one vertex of degree 4 , which we denote by @xmath444 , or two vertices of degree 3 , which we denote by @xmath444 and @xmath424 , where @xmath445 .
let @xmath446 be the path in @xmath346 between @xmath444 and @xmath424 ( if the degree of @xmath444 is 4 , then let @xmath25 be the `` empty '' path consisting of @xmath444 alone )
. without loss of generality , assume that @xmath444 is closer to @xmath3 than @xmath424 . for @xmath447 , let @xmath46 be the path in @xmath346 between @xmath48 and the closest endpoint of @xmath25 .
let @xmath448 , @xmath449 , @xmath450 , and @xmath451 denote the vertex of @xmath452 , @xmath453 , @xmath454 , and @xmath455 , respectively , the distance from which to the closest endpoint of @xmath25 is @xmath78 .
case 3.1 : @xmath3 and @xmath4 are on one side of @xmath25 , and @xmath65 and @xmath70 are on the other side of @xmath25 ( the case when @xmath3 and @xmath70 are on one side of @xmath25 is analogous ) . for @xmath456 ,
let @xmath74 be the hamiltonian path in @xmath457 between @xmath458 and @xmath444 , and for @xmath459 let @xmath74 be the hamiltonian path in @xmath460 between @xmath458 and @xmath461 ( if @xmath462 , then let @xmath74 be the `` empty '' path consisting of @xmath48 alone ) . below is the procedure to construct a hamiltonian cycle @xmath72 in @xmath347 with the desired properties
start at @xmath3 and go along @xmath126 to @xmath463 .
2 . proceed to @xmath464 and go along @xmath127 to @xmath444 .
3 . if @xmath465 , then use the path @xmath466 to get within distance 1 of @xmath424 .
4 . proceed to @xmath467 and go along @xmath415 to @xmath468 ( or stay at @xmath467 if @xmath469 ) .
proceed to @xmath470 and go along @xmath471 to @xmath472 ( or stay at @xmath470 if @xmath473 ) .
6 . if @xmath474 , then proceed to @xmath424 or @xmath425 ( one of these vertices has not been used yet ) and use the path @xmath475 or the path @xmath476 to get within distance 1 of @xmath444 .
proceed to @xmath463 and go along the remaining part of @xmath126 to @xmath3 .
let us verify that the conditions of the lemma are satisfied . if the distance between @xmath48 and the closest endpoint of @xmath25 is greater than 1 , then @xmath48 is adjacent in @xmath74 , and hence in @xmath72 , to a vertex that is at most distance 2 away . if @xmath477 , then @xmath3 is adjacent in @xmath72 to @xmath464 , which is distance 2 away . if @xmath478 , then @xmath4 is adjacent in @xmath72 to @xmath444 . if @xmath479 , then @xmath65 is adjacent in @xmath72 to @xmath424 or @xmath425 , which are distance 1 and 2 away , respectively . if @xmath426 and @xmath480 , then @xmath70 is adjacent in @xmath72 to either @xmath424 or @xmath425 , which are distance 1 and 2 away , respectively . finally , if @xmath426 and @xmath481 , then @xmath70 is adjacent in @xmath72 to @xmath463 , which is distance 2 away .
case 3.2 : @xmath3 and @xmath65 are on one side of @xmath25 , and @xmath4 and @xmath70 are on the other side of @xmath25 .
let @xmath126 be the hamiltonian path in @xmath482 between @xmath463 and @xmath444 .
for @xmath483 , let @xmath74 be the hamiltonian path in @xmath484 between @xmath458 and @xmath461 , where @xmath485 is the endpoint of @xmath46 different from @xmath48 ( if @xmath486 , then let @xmath74 be the `` empty '' path consisting of @xmath48 alone ) .
below is the procedure to construct a hamiltonian cycle @xmath72 in @xmath347 with the desired properties . 1
. start at @xmath3 and go along @xmath126 to @xmath444 .
2 . if @xmath487 , then use the path @xmath488 to get within distance 3 of @xmath424 .
3 . proceed to @xmath464 and go along @xmath127 to @xmath489 ( or stay at @xmath464 if @xmath490 ) .
4 . if @xmath465 , then proceed to @xmath424 or @xmath425 ( one of them has not been used yet ) and use the path @xmath491 or the path @xmath429 to get within distance 3 of @xmath444 . 5 . proceed to @xmath467 and go along @xmath415 to @xmath468 ( or stay at @xmath467 if @xmath469 ) .
6 . if @xmath465 , then proceed to @xmath67 or @xmath68 ( one of them has not been used yet ) and use the path @xmath492 or the path @xmath493 to get within distance 3 of @xmath424
otherwise , if @xmath494 , then proceed to @xmath67 ( in this case we proceeded from @xmath489 or @xmath464 directly to @xmath467 , so @xmath67 has not been used ) . 7 . if @xmath495 and @xmath437 is encountered at the end of step 6 or
if @xmath481 , then proceed to @xmath470 and go along @xmath471 to @xmath472 ( or stay at @xmath470 if @xmath496 ) , otherwise proceed to @xmath472 and go along @xmath471 to @xmath470 ( or proceed to @xmath470 and stay there if @xmath473 ) .
if @xmath495 , then one of the vertices @xmath424 , @xmath425 , @xmath436 , @xmath437 has not been encountered yet ; denote this vertex by @xmath438 ( if the last encountered vertex is @xmath472 , then @xmath437 has been already encountered , so @xmath497 ) .
proceed to @xmath438 and use the path @xmath498 to get within distance 3 of @xmath444 .
proceed to @xmath463 .
( if @xmath499 , then the last encountered vertex is @xmath470 , so @xmath463 is at most distance 4 away , and if @xmath481 , then @xmath500 . ) 10 .
finish the cycle by going along the remaining part of @xmath126 to @xmath3 .
let us verify that the obtained cycle @xmath72 satisfies the conditions of the lemma .
let @xmath401 be the path in @xmath346 between @xmath48 and @xmath329 .
if @xmath474 , then @xmath3 and @xmath70 are each adjacent in @xmath72 to two vertices of @xmath501 , one of which is a non - leaf that is at most distance 3 away , and @xmath4 and @xmath65 are each adjacent to two vertices of @xmath502 , one of which is a non - leaf that is at most distance 3 away .
now suppose that @xmath503 . if @xmath504 , then @xmath48 is adjacent in @xmath74 , and hence in @xmath72 , to a vertex of @xmath346 that is at most distance 2 away . if @xmath477 , then @xmath3 is adjacent in @xmath72 to @xmath444 . if @xmath478 , then @xmath4 is adjacent in @xmath72 to @xmath467 , which is distance 2 away . if @xmath505 , then @xmath65 is adjacent in @xmath72 to @xmath470 , which is distance 2 away . finally , if @xmath506 , then @xmath70 is adjacent in @xmath72 to @xmath463 , which is distance 2 away .
theorem [ g4 ] now follows from lemma [ t - bar ] and lemma [ tl1 ] .
in this section , we compute @xmath507 and give a lower bound on @xmath508 . [ cyc3 ]
let @xmath20 denote the cycle on @xmath15 vertices . if @xmath509 , then @xmath510 is 5-ordered hamiltonian .
let @xmath3 , @xmath4 , , @xmath511 be a sequence of 5 vertices of @xmath20 , and let @xmath147 , @xmath148 , , @xmath512 be the same 5 vertices in the order in which they appear in the cycle @xmath20 . for @xmath513
, let @xmath46 denote the portion of @xmath20 between @xmath348 and @xmath514 containing no other vertices of the sequence @xmath147 , , @xmath512 ( indices taken modulo 5 ) .
we will construct 10 internally - disjoint paths @xmath353 in @xmath510 , where @xmath353 is a path between @xmath348 and @xmath355 for @xmath515 .
we adopt the convention that @xmath516 .
first , we construct @xmath517 for @xmath513 .
let @xmath518 , @xmath519 ,
, @xmath520 be integers such that @xmath521 and @xmath522 for all @xmath78 ( note that @xmath523 is the length of @xmath46 ) . for @xmath513 ,
let @xmath517 be the path @xmath524 , where @xmath525 , @xmath526 , and @xmath527 we allow @xmath528 if @xmath529 and @xmath530 if @xmath531 .
figure [ paths_c ] shows the paths @xmath517 and @xmath532 .
since latexmath:[$a_{i+1 } + 1\not\equiv for @xmath513 , let @xmath534 be the path in @xmath510 between @xmath348 and @xmath514 containing all vertices of @xmath46 that are not contained in @xmath535 and @xmath536 .
define @xmath537 by @xmath382 and form a cycle @xmath72 in @xmath510 containing @xmath3 ,
, @xmath511 in order by linking the paths @xmath383 , @xmath384 , ,
finally , we extend @xmath72 to a hamiltonian cycle using the following procedure . 1 .
if @xmath72 is not hamiltonian , choose a vertex @xmath133 of @xmath20 adjacent in @xmath20 to a vertex @xmath539 .
2 . at most one vertex
adjacent to @xmath388 in @xmath510 is more than distance 3 away from @xmath140 .
since @xmath388 is adjacent to two vertices in @xmath72 , we can choose a vertex @xmath101 adjacent to @xmath388 in @xmath72 such that @xmath540 .
3 . replace the edge @xmath541 of @xmath72 with the path @xmath542 .
4 . if @xmath72 is not hamiltonian , return to step 1 .
since during the procedure we insert the remaining vertices into @xmath72 without changing the order of the vertices already in @xmath72 , the order in which @xmath3 , , @xmath511 are contained in @xmath72 is preserved , so @xmath510 is 5-ordered hamiltonian .
this completes the proof of the theorem .
[ ham3 ] if @xmath0 is a hamiltonian graph on 5 or more vertices , then @xmath1 is 5-ordered hamiltonian .
let @xmath72 be a hamiltonian cycle in @xmath0 .
then @xmath543 is 5-ordered hamiltonian .
since @xmath543 is a subgraph of @xmath1 with the same vertex set , it follows that @xmath1 is 5-ordered hamiltonian , too .
the next proposition yields a lower bound on @xmath508 .
[ sharp2 ] if @xmath495 and @xmath15 is sufficiently large , then @xmath544 is not @xmath64-ordered .
let @xmath3 , @xmath65 , @xmath511 , , @xmath66 , @xmath147 , @xmath148 , , @xmath545 , @xmath4 , @xmath70 , , @xmath546 , @xmath71 , @xmath547 , @xmath67 , @xmath68 , , @xmath425 be the vertices of @xmath20 ( in this order ; see figure [ cm ] ) .
suppose @xmath544 is @xmath64-ordered ; then there is a cycle @xmath72 in @xmath544 that contains @xmath3 , @xmath4 , , @xmath71 in order . for @xmath73 ,
let @xmath74 be the portion of @xmath72 between @xmath48 and @xmath75 that does not contain any other vertices of the sequence @xmath3 ,
, @xmath71 ( indices taken modulo @xmath64 ) . denote @xmath548 and @xmath549 .
if @xmath15 is sufficiently large , then we have @xmath550 for all @xmath78 .
therefore , each @xmath74 must contain a vertex of @xmath79 or a vertex of @xmath144 .
since @xmath551 , there are at least @xmath552 paths @xmath74 that do not contain vertices of @xmath79 and hence consist only of vertices in @xmath553 .
each such @xmath74 must contain at least @xmath554 vertices of @xmath144 , otherwise there would be @xmath108 consecutive vertices of @xmath144 not contained in @xmath74 , and hence @xmath74 would have two adjacent vertices that are more than distance @xmath108 apart in @xmath20 .
we conclude that at least @xmath552 of the paths @xmath74 have at least @xmath554 interior vertices , thus @xmath555 if @xmath15 is sufficiently large a contradiction .
it follows that @xmath544 is not @xmath64-ordered .
@xmath556 for sufficiently large @xmath15 .
we conclude the paper by giving a brief summary of the results and posing questions for further research . we have shown that @xmath557 for any connected graph @xmath0 on at least @xmath2 vertices .
this bound is almost tight because @xmath558 for sufficiently large @xmath15 . what is the best possible upper bound on @xmath56 for an arbitrary connected graph ? for @xmath559 ,
the optimal bound @xmath344 follows from theorem [ g4 ] , and tightness follows from theorem [ pathlowerbound ] . in the case of the path @xmath18 , the value of @xmath59 obeys the inequalities @xmath560 for sufficiently large @xmath15 .
thus @xmath59 can take one of two possible values .
note that @xmath561 and @xmath562 , so both inequalities can be tight for small @xmath2 . what is the exact value of @xmath59 ? another interesting problem is to advance our results on the value of @xmath508 .
theorem [ cyc3 ] and corollary [ sharp2 ] imply @xmath563 .
determine the exact value or give bounds on @xmath508 .
finally , a more difficult task is to generalize theorem [ g2 ] to imply @xmath2-ordered hamiltonicity of powers of @xmath145-connected graphs . give bounds on @xmath56 for a @xmath145-connected graph @xmath0 on at least @xmath2 vertices
this research was begun the 2002 research experience for undergraduates at the university of minnesota duluth .
i would like to thank the program director joseph gallian for suggesting this problem and encouraging me to work on it .
i would also like to thank advisors geir helleloid and phil matchett for many useful comments on this paper .
finally , i want to thank all other student participants and visitors for a wonderful summer .
|
it is known that if @xmath0 is a connected simple graph , then @xmath1 is hamiltonian ( in fact , hamilton - connected ) . a simple graph is @xmath2-ordered hamiltonian if for any sequence @xmath3 , @xmath4 , , @xmath5 of @xmath2 vertices there is a hamiltonian cycle containing these vertices in the given order . in this paper , we prove that @xmath6 is @xmath2-ordered hamiltonian for a connected graph @xmath0 on at least @xmath2 vertices .
we further show that if @xmath0 is connected , then @xmath7 is 4-ordered hamiltonian and that if @xmath0 is hamiltonian , then @xmath1 is 5-ordered hamiltonian .
we also give bounds on the smallest power @xmath8 such that @xmath9 is @xmath2-ordered hamiltonian for @xmath10 and @xmath11 .
_ dedicated to pavlo pylyavskyy on the occasion of his 21st birthday . _
|
black holes and related classical solutions are a topic of long - standing interest in string theory @xcite .
their study has shed light on old questions @xcite in black hole physics ( see e.g. @xcite ) as well as dualities @xcite and other stringy issues @xcite .
indeed , it was an investigation of black holes that first lead to maldacena s conjecture @xcite ( based on earlier work , e.g. @xcite ) relating string theory in asymptotically anti - de sitter space to a conformal field theory on the boundary at spatial infinity . for evidence supporting this conjecture , see @xcite .
it is therefore natural to investigate asymptotically anti - de sitter ( ads ) black holes in light of maldacena s conjecture .
previous work @xcite has analyzed the ( 2 + 1)-dimensional btz black holes @xcite in this way , using the fact that the classical black hole solutions are certain quotients of ads@xmath0 to identify associated states in the conformal field theory ( cft ) .
recall , however , that the nonextremal btz black holes have two asymptotically anti - de sitter regions . as a result ,
the conformal boundary of such spacetimes is not the usual @xmath1 of the universal cover of ads@xmath0 , but two copies of this cylinder .
this means that , strictly speaking , such black holes are not described by quite the same conformal field theory as ads@xmath0 and the corresponding states do not lie in the same hilbert space .
we note that the @xmath2 black hole also has only a single asymptotic region and so again can not lie in the same hilbert space as the btz black holes .
in contrast , there are other asymptotically ads black holes which have only a single asymptotic region .
some examples were constructed in @xcite as quotients of ads@xmath0 . for such black holes ,
we expect the state in the boundary cft to be approximately described by the result of a quotient - like operation on the original vacuum @xmath3 of the conformal field theory .
we do not consider here any effects which may result from additional winding modes in the quotient spacetime .
now , the boundary state corresponding to the btz black hole has been characterized as thermal @xcite .
this is a result of the boundary cft having two disconnected components and the fact that such black holes correspond to states of the boundary cft in which the two boundary components are entangled .
thus , the boundary states are not pure states on either boundary component separately . on the other hand , as discussed in @xcite , there is no reason to expect single - exterior black holes to be mixed states in any corresponding sense .
in particular , as the boundary theory is now exactly the same as that of either ads@xmath0 or the @xmath2 black hole , one expects to be able to interpret single - exterior black holes as ( pure state ) excitations of these ground states . in this paper
we investigate the boundary states for certain single - exterior , asymptotically ads@xmath0 black holes by using a simplified model of the boundary conformal field theory .
our main focus is on a class of spacetimes referred to as @xmath4 geons , which are analogous to the asymptotically flat @xmath5 geon @xcite . in section [ classic ]
we discuss the structure and construction of the @xmath4 geons as quotient spaces of ads@xmath0 . in section [ quantum ]
we first motivate and define our model cft and then verify that the @xmath4 geon corresponds to a pure state of our model theory .
we also verify that the expectation value of the cft hamiltonian in this state coincides with the mass of the black hole in the limit where the black hole horizon circumference is much greater than the length scale associated with the ads space .
this corresponds to the limit where @xmath6 or @xmath7 is much greater than @xmath8 in terms of the left and right momentum , onebrane , and fivebrane quantum numbers of the associated @xcite six - dimensional black string .
note that taking such a limit is also important to remove quantum corrections to the entropy of such black strings . after developing our technology in the context of the @xmath4 geons
, we then briefly address the single - exterior black holes of refs .
@xcite in section [ swedish ] .
we close with some comments on the extrapolation of our results to the full cft of maldacena s conjecture , the encoding of the geon topology in the boundary state , and other issues in section [ disc ] .
our attention will be focused on the special case of maldacena s conjecture for the spacetime ads@xmath9 .
we use units in which @xmath10 .
the ( 2 + 1)-dimensional newton s constant is denoted by @xmath11 .
in this section we describe the quotient constructions of the spinless nonextremal btz hole and the @xmath4 geon from three - dimensional space , and the extension of this quotient construction to the conformal boundaries of the spacetimes .
the material for the btz hole is familiar @xcite , but a review is needed in order to establish the relationship between the btz hole and the geon .
we also mention generalizations of the geon construction to spacetimes with additional internal dimensions , in particular the internal factor @xmath13 that arises in string theory @xcite .
recall that the three - dimensional space @xmath12 can be defined as the surface @xmath15 in @xmath16 with the global coordinates @xmath17 and the metric @xmath18 the positive parameter @xmath19 is the inverse of the gaussian curvature .
@xmath12 is a smooth three - dimensional spacetime with signature @xmath20 .
it is maximally symmetric , and ( the connected component of ) the isometry group is ( the connected component of ) @xmath21 . from now on we set @xmath22 .
it is useful to introduce on @xmath12 the coordinates @xmath23 by @xcite [ sausage - coords ] @xmath24 with @xmath25 and the identifications @xmath26 these coordinates can be understood as global on @xmath12 , apart from the elementary coordinate singularity at @xmath27 .
the metric reads @xmath28 \ \ .
\label{sausage - metric}\ ] ] we define the time orientation on @xmath12 so that the killing vector @xmath29 points to the future , and a spatial orientation so that , for @xmath30 , the pair @xmath31 is right - handed . dropping from ( [ sausage - metric ] ) the conformal factor @xmath32 yields a spacetime that can be regularly extended to @xmath33 .
the timelike hypersurface @xmath33 in this conformal spacetime is by definition the conformal boundary of @xmath12 : we denote this conformal boundary by @xmath34 .
@xmath34 is a timelike two - torus , coordinatized by @xmath35 with the identifications @xmath36 and the metric on @xmath34 is flat , @xmath37 @xmath34 inherits from @xmath12 a time orientation in which the vector @xmath29 points to the future , and a spatial orientation in which @xmath38 points to the right .
the above definition of a metric on @xmath34 relies on a particular coordinate system .
the isometries of @xmath12 act on the metric ( [ boundary - metric ] ) as an @xmath21 group of conformal isometries , and the metric on @xmath34 is thus invariantly defined only up to such transformations .
we therefore understand the metric ( [ boundary - metric ] ) as a representative of its @xmath21 equivalence class .
the above constructions adapt in an obvious way to the universal covering space of @xmath12 , which we denote by @xmath14 , and to its conformal boundary , which we denote by @xmath39 . when the last identifications in ( [ sausage - idents ] ) and ( [ boundary - idents ] ) are dropped , the coordinates @xmath23 can be regarded as global on @xmath14 , and the coordinates @xmath35 can be regarded as global on @xmath39 .
@xmath39 has topology @xmath40 , and the metric ( [ boundary - metric ] ) on @xmath39 is globally hyperbolic .
@xmath39 is time - oriented , with @xmath29 pointing to the future , and space - oriented , with @xmath38 pointing to the right .
the isometries of @xmath14 clearly induce conformal isometries of @xmath39 , and the metric on @xmath39 is invariantly defined only up to these conformal isometries .
we now describe the spinless nonextremal btz black hole and its conformal boundary .
we denote by @xmath41 and @xmath42 the killing vectors on @xmath14 that are respectively induced by the killing vectors @xmath43 of @xmath16 . the conformal killing vectors that @xmath41 and @xmath42 induce on @xmath39 are respectively @xmath44 @xmath41 and @xmath42 are clearly mutually orthogonal , and @xmath45 and @xmath46 are similarly mutually orthogonal .
we denote by @xmath47 the largest connected region of @xmath14 that contains the hypersurface @xmath48 and in which @xmath41 is spacelike . as @xmath49 , we see from ( [ sausage - coords ] ) that @xmath47 is isometric to the subset @xmath50 of the surface ( [ embedding - surface ] ) in @xmath16 .
@xmath47 intersects every constant @xmath51 hypersurface for @xmath52 , but the only one of these hypersurfaces that is entirely contained in @xmath47 is @xmath48 .
the conformal extension of @xmath47 to @xmath39 intersects @xmath39 in the two disconnected diamonds [ drldefs ] @xmath53 by construction , @xmath45 is spacelike with respect to the metric ( [ boundary - metric ] ) in @xmath54 and @xmath55 . in the orientations on @xmath54 and @xmath55 induced by that on @xmath39 , @xmath45 points to the right in @xmath54 and to the left in @xmath55 .
@xmath46 is future timelike in @xmath54 , and past timelike in @xmath55
. now , let @xmath56 be a prescribed positive parameter , and let @xmath57 be the group of isometries of @xmath14 generated by @xmath58 .
@xmath59 preserves @xmath47 , and its action on @xmath47 is free and properly discontinuous .
the quotient space @xmath60 is the spinless , nonextremal btz black hole .
the horizon - generating killing vector , induced by @xmath42 , is respectively future and past timelike in the two exterior regions , and spacelike in the black and white hole interiors .
the horizon circumference is @xmath56 , and the mass is @xmath61 , where @xmath11 is the ( 2 + 1)-dimensional newton s constant .
a conformal diagram is shown in figure 1 .
. a conformal diagram of the btz hole .
each point in the diagram represents a suppressed @xmath62 .
the involution @xmath63 introduced in section [ subsec : geon ] consists of a left - right reflection about the dashed vertical line , followed by a rotation by @xmath64 on the suppressed @xmath62 . in each of the two exterior regions of the hole ,
the geometry is asymptotic to the asymptotic region of @xmath14 .
one can therefore attach to each of the two exterior regions a conformal boundary that is isometric to @xmath39 .
what is important for us is that these conformal boundaries can be identified as the quotient spaces @xmath65 and @xmath66 , where @xmath67 and @xmath68 are the restrictions to respectively @xmath54 and @xmath55 of the conformal isometry group of @xmath39 generated by @xmath69 @xcite . to see this explicitly ,
consider @xmath54 , and cover @xmath54 by the coordinates @xmath70 defined by [ alphabeta - r ] @xmath71 \ \ , \\
\beta & = & \ln \tan\left[(\theta+t)/2\right ] \ \ .\end{aligned}\ ] ] both @xmath72 and @xmath73 take all real values , and the metric ( [ boundary - metric ] ) on @xmath54 reads @xmath74 @xmath75 and @xmath76 are future - pointing null vectors , and @xmath77 the generator @xmath69 of @xmath67 acts in these coordinates as @xmath78 , and the metric ( [ dr - metric1 ] ) is not invariant under @xmath67 , but the conformally equivalent metric @xmath79 is .
the quotient space @xmath65 , with the metric induced by ( [ dr - metric1-conf ] ) , is thus isometric to @xmath39 with the metric ( [ boundary - metric ] ) .
note that the vector on @xmath65 induced by @xmath46 is a future timelike killing vector in the metric induced from ( [ dr - metric1-conf ] ) , and the isometry with ( [ boundary - metric ] ) takes this vector to the vector @xmath80 on @xmath39 .
similar observations apply to @xmath66 , the main difference being that the timelike killing vector induced by @xmath46 is now past - pointing , and mapped to @xmath81 under the isometry with @xmath39 .
we now turn to the @xmath4 geon .
consider on @xmath14 the isometry @xmath82 that is the composition of @xmath83 and the map @xmath84 .
the group generated by @xmath82 acts on @xmath47 freely and properly discontinuously .
we define the @xmath4 geon as the quotient space of @xmath47 under this group . as @xmath85
, @xmath82 induces on the btz hole an involutive isometry , which we denote by @xmath86 , and the @xmath4 geon is precisely the quotient space of the btz hole under the @xmath87 isometry group generated by @xmath88 .
the action of @xmath63 on the btz hole is easily understood in the conformal diagram , as shown in figure 1 and described in the caption .
the conformal diagram of the @xmath4 geon is shown in figure 2 .
it is clear that the @xmath4 geon is a black hole spacetime with a single exterior region that is isometric to one exterior region of the btz hole .
the geon is time orientable and admits a global foliation with spacelike hypersurfaces of topology @xmath89@xmath90point at infinity@xmath91 , whence its name ; it is , however , not space orientable .
it shares all the local isometries of the btz hole .
however , as @xmath82 inverts the sign of the killing vector @xmath42 on @xmath14 , @xmath42 does not induce a globally - defined killing vector on the geon , while it does induce a globally - defined killing vector on the btz hole .
the quotient construction from the btz hole to the @xmath4 geon is highly similar to the quotient construction from the kruskal manifold to the @xmath5 geon in four spacetime dimensions @xcite .
a conformal diagram of the @xmath4 geon .
the region not on the dashed line is identical to that in the diagram of figure 1 , each point representing a suppressed @xmath62 in the spacetime . on the dashed line , each point in the diagram represents again an @xmath62 in the spacetime , but with only half of the circumference of the @xmath62 s in the diagram of figure 1 .
the map @xmath63 can clearly be extended to the conformal boundary of the btz hole , where it defines an involution @xmath92 that interchanges the two boundary components .
the conformal boundary of the geon can thus be understood as the quotient of the conformal boundary of the btz hole under the @xmath87 generated by @xmath92 .
the conformal boundary of the geon is clearly isometric to a single component of the conformal boundary of the btz hole , and the quotient construction induces on it a time orientation .
although the boundary of the geon is space orientable , the quotient construction does not induce a choice for the space orientation : the reason is that @xmath92 interchanges the orientations of the two boundary components of the btz hole . in string theory
, there is interest in spacetimes that are metric products of @xmath14 and a compact `` internal '' space .
the quotient constructions of the btz hole and the @xmath4 geon can clearly be extended to such a spacetime and its boundary by taking the identification group @xmath93 to act trivially on the internal dimensions . however ,
if the internal space admits suitable isometries , other extensions of interest exist . specifically ,
if the internal space admits an involutive isometry @xmath94 , the composition @xmath95 of @xmath94 and @xmath82 is an isometry on the full spacetime , and taking @xmath95 to generate the identification group yields a generalization of the @xmath4 geon .
an equivalent construction is to consider first the product spacetime of the btz hole and the internal space , and quotient this by the @xmath87 generated by the map @xmath96 that is the composition of @xmath63 and @xmath94 .
the resulting geon has again a single exterior region . defining the conformal boundary in terms of a conformal rescaling of the ( 2 + 1)-dimensional part of the metric
, one finds that the conformal boundary is connected , and its geometry is that of one component of the conformal boundary of the btz hole with the constant internal space .
note that while the geometry of the geon depends on @xmath94 , the geometry of its conformal boundary does not .
in particular , if the internal space is orientable and @xmath94 reverses the orientation , the above construction yields a space - orientable geon .
the case that will concern us below is a ten - dimensional black hole spacetime ( considered in the context of string theory in ref .
@xcite ) in which the internal space is the metric product of @xmath97 and @xmath98 , with a round metric on the former and a flat metric on the latter . if the @xmath98 factor further factorizes into a metric product of @xmath62 and @xmath99 , the reflection on the @xmath62 , @xmath100 .
] of the @xmath62 is a nonorientable involutive isometry of the internal space .
other nonorientable involutive isometries can be found by composing this inversion with orientable involutive isometries acting on the other factors , such as the antipodal map on the @xmath97 . as smooth orientable quotients of ads@xmath9 ,
all of these spacetimes provide exact classical solutions of string theory .
we now turn to the question of what sort of quantum state in the boundary cft of maldacena s conjecture @xcite is induced by the quotient constructions of section [ classic ] .
we specialize to the internal space @xmath13 and , in order to arrive at an orientable spacetime in which we might discuss orientable string theory , we further assume the metric on the @xmath98 to factorize in such a way that a reflection of an @xmath62 provides an internal involutive isometry @xmath94 as discussed in section [ subsec : internal - dims ]
. we will not consider the full cft suggested by the conjecture but , instead , we consider a simplified linear field theory which we expect to capture the central features of interest .
we consider a set of free scalar fields on the boundary cylinder @xmath101 of @xmath14 , but with certain refinements .
the point is that , as discussed in section [ subsec : internal - dims ] , the internal isometry @xmath94 does not affect how the identifications of the full spacetime @xmath102 project to identifications of @xmath39 .
nevertheless , @xmath94 is expected to affect the full conformal field theory of ref .
@xcite on @xmath39 .
thus , our model must contain enough additional structure to faithfully represent the action of @xmath103 .
recall that the @xmath4 geon is a quotient of the btz black hole by the involution @xmath104 .
let us take a moment to consider first how this @xmath105 quotient would be reflected in a boundary cft . for a linear field
, it is natural to think of the field on both the btz hole and geon boundaries as being the _ same _ operator - valued distribution on the boundary of the btz hole , but merely smeared against different classes of test functions .
the detailed correspondence is given by lifting a test function from the geon boundary to the btz black hole boundary and dividing by @xmath106 to ensure canonical normalization of the field .
since the geon fields are embedded in this way in the algebra of btz fields , any state on the btz boundary directly induces a state on the geon boundary . as in @xcite ,
it is sufficient to think of a free scalar field @xmath107 on the geon boundary as being a symmetrization of the corresponding field @xmath108 residing on the boundary of the btz black hole : @xmath109 where @xmath110 is the covering map from the btz black hole boundary to the geon boundary .
now , in our construction of the orientable geon from ( btz hole)@xmath111 , the involutive spacetime isometry acts on the btz hole dimensions by @xmath104 and on the internal toroidal dimensions by reflecting one of the @xmath62 s of the @xmath98 .
thus , we must include in our model some feature that corresponds to this internal topology .
we recall that the topology of the internal torus is captured @xcite by the fact that the boundary cft should be a nonlinear sigma model whose target space is a symmetric product of copies of the @xmath98
. a given @xmath62 factor of the internal space is represented as a symmetric product of @xmath62 factors in the target space of the sigma model .
it is clear that , when acting on the boundary field theory , the involution that exchanges a point @xmath112 with its image @xmath113 should act nontrivially on this part of the sigma model , reflecting the appropriate @xmath62 factors in the target space .
this is in direct analogy with the constructions of @xcite , where the involutions acted only on the internal @xmath97 .
we will model this feature by replacing the part of the sigma model associated with the appropriate @xmath62 factors by a single scalar field @xmath114 . to tighten the analogy with the sigma model
, one might like to think of the field space of @xmath114 as compactified to a circle of the same size as the internal @xmath62 ( say , length @xmath115 ) .
however , this would present certain problems for our quotient construction .
setting aside the reflection of this @xmath62 for the moment , consider the analogue of ( [ sym ] ) for a field @xmath116 which is periodic in field space with this period .
the resulting geon field @xmath117 would then have period @xmath118 . on the other hand ,
the quotient by @xmath119 does not change the size of the internal @xmath62 factors . note that the @xmath120 normalization factor in ( [ sym ] ) is fixed by the commutation relations and the behavior of the green s functions . in order to have a faithful representation of the involution @xmath121
, we choose to ignore the compactness of the @xmath62 and to allow our field @xmath114 to take values on a real line . in analogy with the @xmath62 it replaces
, this line will be reflected through the origin by the action of the involution on our field theory .
for contrast , the rest of the sigma model will be replaced by a single free scalar field @xmath122 whose field space is not affected by the involution .
our simple model allows exact calculations to be done and captures many features of interest .
a notable exception , however , is that the central charge of our model is @xmath123 , while that of the cft in maldacena s conjecture is @xmath124 .
this we will correct by hand when considering the energies of our states in section [ energies ] . in finding the geon state in the cft we will proceed as indicated above , first calculating the boundary state of the btz hole in section [ qbtz ] , and then performing a final identification to yield the state for the geon in section [ geon ] .
though the state on the black hole boundary has been considered in @xcite , setting up the btz calculation in a different way will make our geon calculation particularly straightforward .
in addition , we will be able to see certain effects of the compact boundary that were neglected in @xcite .
sections [ qbtz ] and [ geon ] consider only the oscillator modes of our scalar fields .
the zero modes are more subtle and are treated separately in section [ zero ] .
section [ energies ] discusses the energy of our states and compares the result with the mass of the corresponding black hole or geon .
[ osc ] as discussed in section [ classic ] , the btz black hole is the quotient of the region @xmath125 under the discrete isometry group generated by @xmath126 .
similarly , the boundary of the btz hole may be thought of as the quotient of the region @xmath127 in the boundary @xmath39 of @xmath14 under the group generated by @xmath128 .
we would now like to consider the vacuum state @xmath129 that is defined on @xmath39 with respect to the timelike killing vector @xmath29 , and construct the state that @xmath129 induces on the black hole boundary . since the quotient of @xmath47 to the btz black hole does not act on the internal factors ,
the construction is identical for both @xmath122 and @xmath114 and the discussion below applies to either field .
recall that the null coordinates @xmath72 and @xmath73 ( [ alphabeta - r ] ) define a conformal mapping of @xmath54 onto minkowski space ( and similarly for @xmath55 ) . in terms of this minkowski space ,
the map @xmath128 is just a spatial translation , and when the overall scale of the metric is chosen as in ( [ dr - metric1-conf ] ) , the proper distance of the translation is @xmath130 .
thus , these identifications enact the usual compactification of minkowski space to @xmath1 .
the effect of the compactification on the scalar field theory is merely to remove all modes that are not appropriately periodic and to reinterpret the periodic modes , which are not normalizable on @xmath54 ( or @xmath55 ) , as normalizable modes on the cylinder .
the nontrivial part in this construction is that , as noted in section [ subsec : btz ] , the timelike killing vectors on the two boundary components of the btz black hole do not lift to the timelike killing vector @xmath29 on @xmath39 : the future timelike killing vector on the boundary component arising from @xmath54 lifts to @xmath131\eta$ ] , and that on the boundary component arising from @xmath55 lifts to @xmath132\eta$ ] .
thus , in order to interpret the state induced by @xmath129 on the btz hole boundary in terms of the btz particle modes , we must first write the state induced by @xmath129 on @xmath133 in terms of continuum - normalized particle states that are positive frequency with respect to @xmath46 on @xmath54 and with respect to @xmath134 on @xmath55 . this calculation is quite similar to expressing the minkowski vacuum in terms of rindler particle modes ( see e.g. @xcite ) . to begin , consider the mode functions [ rindler - modes ] @xmath135}^{i\epsilon\omega } \ \ , \\ & & u^{l}_{\omega , \epsilon } : = \frac{1}{\sqrt{4\pi\omega } } { \left [ \tan\biglb((-\theta + \epsilon t)/2 \bigrb ) \right]}^{-i\epsilon\omega } \ \ , \end{aligned}\ ] ] where @xmath136 , the index @xmath137 takes the values @xmath138 , and the modes with superscript @xmath139 ( @xmath140 ) have support in @xmath54 ( @xmath55 ) .
the @xmath139-modes are eigenfunctions of the vector field @xmath46 on @xmath54 with eigenvalue @xmath141 , and the @xmath140-modes are similarly eigenfunctions of the vector field @xmath134 on @xmath55 with eigenvalue @xmath141 .
the modes are continuum orthonormal on @xmath127 .
the modes with @xmath142 are right - moving and those with @xmath143 are left - moving , in the orientations on @xmath54 and @xmath55 induced by that on @xmath39 . ) , in this orientation @xmath45 is right - pointing in @xmath54 and left - pointing in @xmath55 .
] these properties for the @xmath139-modes become explicit by writing the modes in terms of the null coordinates @xmath70 ( [ alphabeta - r ] ) on @xmath54 as @xmath144 and @xmath145 .
analogous expressions hold for the @xmath140-modes on @xmath55 .
let now @xmath146 stand for the vacuum of the non - zero modes induced on @xmath147 by @xmath3 , and let @xmath148 stand for the vacuum of the @xmath149-modes ( [ rindler - modes ] ) .
we need to express @xmath150 in terms of @xmath148 and the excitations associated with the @xmath149-modes . to this end , we follow the method of unruh @xcite and build from the @xmath149-modes and their complex conjugates a complete set of linear combinations , called @xmath151-modes , that are bounded analytic functions in the lower half complex @xmath51 plane . by construction ,
the @xmath151-modes are purely positive frequency with respect to @xmath29 , and they thus share the vacuum @xmath146 . the relevant bogoliubov transformation can then be simply read from expressions of the @xmath151-modes . analytically continuing the @xmath149-modes ( [ rindler - modes ] ) between @xmath54 and @xmath55 in the lower half of the complex @xmath51 plane , we find that a complete set of @xmath151-modes is [ w - modes ] @xmath152 where @xmath153 and @xmath154 .
the creation and annihilation operators @xmath155 , @xmath156 for the @xmath151-modes are thus related to the creation and annihilation operators @xmath157 , @xmath158 for the @xmath149-modes by [ a - vs - b ] @xmath159 to relate the vacua , one notices that equations ( [ a - vs - b ] ) can be written as @xcite @xmath160 where @xmath161 is the ( formally ) hermitian operator @xmath162 and @xmath163 is defined by @xmath164 the vacuum @xmath146 , annihilated by the @xmath165 , is therefore related to the vacuum @xmath148 , annihilated by the @xmath158 , through @xmath166 in terms of the normalized @xmath167-particle states @xmath168 ( @xmath169 ) associated with the modes @xmath170 ( @xmath171 ) , this relation reads @xmath172 we can now pass from the field theory on @xmath127 to the field theory on the boundary of the btz hole .
let us refer to the non - zero modes of the field as oscillator modes .
for the oscillator modes , the effect of the periodic identifications is simply to replace the continuous index @xmath141 by the discrete values @xmath173 , where the index @xmath174 takes values in the positive integers , and to change the normalization factor in the @xmath149-modes ( [ rindler - modes ] ) from @xmath175 to @xmath176 .
for the oscillator modes , we then obtain from @xmath150 ( [ vacads - expanded ] ) the state @xmath177 @xmath178 clearly lies in the fock space @xmath179 , where @xmath180 and @xmath181 are the hilbert spaces of the oscillator modes of the scalar field on the respective @xmath1 conformal boundary components of the btz hole .
note that @xmath182 ( [ vacbtzhole - expanded ] ) is properly normalized and that it may be written as @xmath183 , where @xmath184 when both fields @xmath122 and @xmath114 are considered ( together with the zero mode states discussed below in section [ zero ] ) , @xmath185 gives the btz black hole quantum state in our model theory on @xmath186 .
we see from ( [ vacbtzhole - expanded ] ) that @xmath187 contains pairwise correlations between modes residing on the two boundary components , and when the modes on one component are traced over , the resulting state on the other component is thermal .
the expectation value of any operator associated with only one boundary component is thus identical to the expectation value in a thermal state .
this is in particular true of the stress - energy tensor . finally , to identify the temperature of the effective thermal state on a single boundary component , we recall that the discretized @xmath149-modes are eigenfunctions of @xmath188 with eigenvalue @xmath189 .
the expression ( [ vacbtzhole - expanded ] ) therefore implies that the thermal state has temperature @xmath190 with respect to @xmath188 , which translates into the temperature @xmath191 with respect to the killing vector @xmath29 in the form ( [ boundary - metric ] ) of the boundary metric .
this is the usual hawking temperature of ( the interior of ) the btz black hole with respect to a killing time coordinate that agrees with our @xmath51 on the boundary @xcite .
we shall discuss the energy expectation values further in section [ energies ] after having first addressed the zero modes .
we now construct the state of the oscillator modes in our model on the boundary of the @xmath4 geon .
the zero modes will be discussed below in section [ zero ] .
let @xmath192 and @xmath193 be the restrictions of the covering map @xmath110 of the btz hole boundary over the boundary of the geon to the left and right components of the btz hole boundary .
the geon boundary fields @xmath194 and @xmath195 are then related to the btz fields by [ geonfields ] @xmath196 \ \ , \\
\phi_{g}(x ) & : = & \frac{1}{\sqrt{2 } } \left [ \phi_{\rm btz}\biglb(\rho_r^{-1}(x)\bigrb ) + \phi_{\rm btz } \biglb(\rho_l^{-1}(x)\bigrb ) \right ] \ \ .
\label{geonfields - phi}\end{aligned}\ ] ] the argument @xmath112 of @xmath197 and @xmath198 takes values in a single copy of @xmath1 , but the field operators act in the hilbert space @xmath199 of the btz boundary theory . as the btz black hole state is symmetric with respect to the sign of @xmath114 , the geon state will not depend on which boundary component is called left or right .
what we wish to do is to calculate the restriction of the btz state to the algebra generated by the fields ( [ geonfields ] ) .
note that the restriction of a pure state to a subalgebra is not necessarily pure .
thus , a priori , the result could be either a pure state or a mixed state .
we proceed by introducing two more fields , @xmath200 and @xmath201 , through [ fakefields ] @xmath202 \ \ , \\
\tilde \phi_{g}(x ) & : = & \frac{1}{\sqrt{2 } } \left [ \phi_{\rm btz}\biglb(\rho_r^{-1}(x)\bigrb ) - \phi_{\rm btz } \biglb(\rho_l^{-1}(x)\bigrb ) \right ] \ \ .
\label{fakefields - phi}\end{aligned}\ ] ] these fields again live on @xmath1 and act in the hilbert space @xmath199 . the definitions
( [ geonfields ] ) and ( [ fakefields ] ) amount to writing the two fields @xmath203 on the two - component btz boundary as the four fields @xmath204 on a single copy of @xmath1 .
the hilbert space @xmath199 then factors as @xmath205 , where @xmath206 is the hilbert space of the geon fields ( [ geonfields ] ) while @xmath207 is the hilbert space of the fields ( [ fakefields ] ) .
the desired state in the geon boundary theory ( @xmath206 ) then follows by tracing over the hilbert space @xmath208 .
in fact , taking this trace will be trivial as we will see that the state @xmath187 is a tensor product state , containing no correlations between @xmath209 and @xmath208 . that this must be so follows from the observation that @xmath210 contains only two - particle correlations .
since this state vector is invariant under the operation of interchanging the right and left boundary components , it can only contain correlations between fields of the same parity under this operation .
note that the only difference between the fields @xmath122 and @xmath114 is in the signs in ( [ geonfields ] ) and ( [ fakefields ] ) , and that interchanging the tilded geon boundary fields for the untilded ones is equivalent to interchanging @xmath122 for @xmath114 on the btz boundary .
thus , the state of @xmath211 is identical to the state of @xmath201 ( on the btz boundary ) and the state of @xmath200 is identical to that of @xmath117 . as a result
, it will again be sufficient to treat only one of the fields @xmath122 and @xmath114 explicitly .
we choose the field @xmath122 , and then read off the state of @xmath114 from the results .
consider thus the field @xmath122 . a complete orthonormal basis of positive frequency oscillator modes on the geon boundary is given by the functions @xmath212 , which are the pushforward to the geon of the modes ( [ rindler - modes ] ) on the right btz boundary .
we denote the annihilation and creation operators associated with the field @xmath117 in this basis by @xmath213 and @xmath214 , and those associated with the field @xmath201 by @xmath215 and @xmath216 . as the properties of the involution @xmath217 ( whose quotient of the btz boundary yields the geon boundary ) imply that the pullback of these modes to the left btz boundary differs from ( [ rindler - modes ] ) by a @xmath64 rotation ( and a definition of left- and right - moving ) , we have @xmath218 , and we then find from ( [ geonfields - phi ] ) and ( [ fakefields - phi ] ) the relations [ d - vs - b ] @xmath219 \ \ , \\ d_{\tilde \phi , n,\epsilon } & = & { 1 \over \sqrt{2 } } \left [ b^r_{\phi , n , \epsilon } - ( -1)^n b^l _ { \phi , n , -\epsilon } \right ] \ \ . \end{aligned}\ ] ] using ( [ d - vs - b ] ) , the operator @xmath220 ( [ jbtz ] ) can be written in the form @xmath221 \ \ .\ ] ] the parts of ( [ diag ] ) referring to @xmath117 and @xmath222 each have the same form as the terms in ( [ jbtz ] ) , apart from some changes of signs . by the same methods as in section [ qbtz ] , we can therefore write @xmath182 in terms of the normalized @xmath167-particle states @xmath223 and @xmath224 associated with the operators @xmath214 and @xmath225 as @xmath226 } \exp(-\pi\omega_n q ) \right ) \ \ , \ ] ] where we have defined @xmath227 and @xmath228 .
this means in particular that @xmath229 is a direct product of a state in @xmath206 with a state in @xmath230 .
the restriction of ( [ diagstate ] ) to the field @xmath117 therefore yields the geon oscillator state for @xmath122 . for the field @xmath114 ,
the calculation is similar except in that the tilded and untilded fields are interchanged .
the geon oscillator state for @xmath114 can therefore be read off from the restriction of ( [ diagstate ] ) to @xmath201 .
thus , defining @xmath231 , the geon oscillator state including both fields is @xmath232 } \right ) \ \ , \label{vacbtzhole - proj}\ ] ] which is a normalized pure state in @xmath206 .
the correlations between the right - movers and the left - movers exhibited in ( [ vacbtzhole - proj ] ) are similar to the correlations found in scalar field theory on the ( interior of the ) @xmath5 geon spacetime and on an analogous rindler - type spacetime in ref .
@xcite . ) were encoded in the phase choices for the mode functions .
] we shall discuss this phenomenon further in section [ disc ] . in our calculations of the zero mode states below
, we replace the oscillator modes on the cads@xmath0 , btz hole , and geon boundaries by modes of finite frequency @xmath233 .
we then take the limit @xmath234 to give the state for the actual zero modes of our massless fields .
now , the reader may be concerned by the fact that modifying the zero modes on the boundary of the btz hole affects not only the zero mode on @xmath39 ; it will modify the oscillator modes as well .
however , the procedure below may be thought of as a condensed version of the more manifestly self - consistent procedure of giving our fields a finite mass @xmath235 ( which of course affects all of the modes together ) and then taking the @xmath236 limit . in this longer and more complicated calculation
, one would compute the state of the massive field on @xmath237 in terms of modes that are positive frequency with respect to @xmath46 on @xmath54 and @xmath134 on @xmath55 and then take the @xmath236 limit to yield the state of the massless field ( zero mode and all ) on @xmath238 .
the massless field state can then be compactified as before .
it will be clear that the calculation below gives identical results .
now , the fact that the zero mode energy eigenstates of a free field on @xmath1 are not normalizable will lead to some subtleties in our argument .
in particular , the ground state @xmath129 from which the btz and geon states are induced is non - normalizable .
thus , the btz and geon states are unlikely to be normalizable , and a limit taken in the hilbert space topology will not be useful
. we will proceed by considering the states as tempered distributions on the zero mode configuration space .
note that , in the topology of tempered distributions , a suitably rescaled version of the harmonic oscillator ground state does in fact converge to the free particle ground state @xmath239 , where @xmath240 is the free particle momentum .
the rescaling is necessary since @xmath241 is not a normalizable state in the hilbert space .
our limit will require a rescaling of the state as well , and , for this reason , we induce the btz and geon states from the state @xmath242 , which is @xmath243 times the normalizable ground state for the frequency @xmath233 zero modes .
more will be said about the precise form of this rescaling at the end of the calculation .
below , we first take the limit in the sense of ( smooth ) functions on the configuration space .
we then note that this convergence is sufficiently uniform to guarantee that the same limit is given by the topology of tempered distributions .
let us begin by replacing the zero mode on the boundary of @xmath14 by the @xmath244 oscillator mode @xmath245 , and similarly on the two components of the btz boundary and on the geon boundary . in this case , the btz zero modes are associated with the modes @xmath246^{i\omega/2 } \ \ , \\ u^l_{\omega } & = & { 1 \over { \sqrt{4 \pi \omega } } } \left [ { { \tan \biglb ( ( -\theta - t)/2 \bigrb ) } \over { \tan \biglb ( ( -\theta + t)/2 \bigrb ) } } \right]^{i\omega/2 } \ \ , \end{aligned}\ ] ] in the domains @xmath247 on the boundary of @xmath14 .
we refer to the creation and annihilation operators for such modes as @xmath248 and @xmath249 . as before , we need only explicitly calculate the btz state for one of our scalar fields .
the calculation proceeds much as in section [ qbtz ] , except that the zero mode does not have separate left- and right - moving parts .
thus , the operator @xmath250 which relates the zero mode part of the regulated btz zero mode state @xmath251 to that of the vacuum @xmath252 on the btz boundary takes the form @xmath253 with the corresponding form @xmath254 for the zero mode state in terms of normalized @xmath167-particle states . here ,
the fact that our state @xmath242 is @xmath255 times a normalized state can be seen explicitly .
we must now take the limit as the frequency @xmath233 is sent to zero .
to proceed , recall that the states @xmath256 may be thought of as the normalized occupation number states for a harmonic oscillator on the real line .
we therefore introduce the usual position states @xmath257 [ normalized to @xmath258 and momentum states @xmath259 [ normalized to @xmath260 for this particle , as well as the tensor products @xmath261 and @xmath262 . in the limit @xmath263
, the occupation number states must go over to energy eigenstates of the free particle .
moreover , since states with @xmath264 have positive parity , they must be proportional to the positive parity states @xmath265 for the appropriate momentum @xmath240 in the limit of small @xmath233 .
similarly , odd states with @xmath266 must become proportional to @xmath267 . to fix this remaining constant of proportionality ,
consider the even wave functions @xcite @xmath268^{-1/2 } h_{2k } ( x \sqrt{\omega } ) e^{-x^2 \omega/2}\ ] ] of the oscillator states . here
@xmath269 is the hermite polynomial of order @xmath174 . of course , since the states @xmath270 are normalized , the wave function at any point @xmath112 vanishes as @xmath263 .
in contrast , we have @xmath271 .
thus , if we fix a compact set @xmath272 , in the limit of small @xmath233 with @xmath273 held fixed we have @xmath274^{-1/2 } h_{2k } ( 0 ) \sqrt{\pi } \left ( { } ^{l , r}\langle x | p \rangle^{l , r}_+ \right)\ ] ] uniformly on @xmath275 . using the fact
that , in the @xmath276 limit at fixed @xmath112 , the coefficient in ( [ wfr ] ) is independent of @xmath277 and the creation operator goes over to @xmath278 times the momentum operator , one can show that the relative normalizations of the wave functions @xmath279 and @xmath280 are the same up to a factor of @xmath281 : @xmath282^{-1/2 } h_{2k } ( 0 ) \sqrt{\pi } \left ( { } ^{l , r}\langle x | p \rangle^{l , r}_- \right ) .\ ] ] let us now evaluate the part of the wave function @xmath283 that comes from harmonic oscillator states with energies @xmath284 in a small interval @xmath285 in the limit of small @xmath233 .
we include both even parity ( @xmath286 ) and odd parity ( @xmath287 ) states .
the associated momentum interval is @xmath288 where @xmath289 . for fixed @xmath290 the values of the wave functions @xmath291 for the allowed values of @xmath277 are nearly identical and are given by ( [ wfr ] ) .
now , the even hermite polynomials at zero are given @xcite by @xmath292 .
using stirling s approximation @xmath293 , and the fact that there are @xmath294 states of each parity in the allowed energy range , the contribution of these states is @xmath295 thus , summing over all such intervals @xmath296 and considering all @xmath297 gives the zero frequency state of the zero mode for either @xmath122 or @xmath114 : @xmath298 , the exponential cutoff @xmath299 and exponential falloff of the oscillator wave functions in the momentum representation can be used to show that that ( [ phiz ] ) is in fact the limit of @xmath300 in the sense of tempered distributions .
thus , expression ( [ phiz ] ) is the zero mode state on the btz boundary .
we see that the trace over either boundary component yields a thermal state at the same temperature as for the oscillators .
as expected , neither ( [ phiz ] ) nor the state traced over one component is normalizable .
we see that this is due to the precise correlation of the momenta on the right and left in ( [ phiz ] ) .
it turns out that something similar must happen whenever the zero mode spectrum is continuous .
this is because the state @xmath3 from which the btz state is induced is invariant under the action of the killing field @xmath46 ( [ eta ] ) . on the btz boundary this corresponds to the action of the difference @xmath301 between the right and
left boundary hamiltonians .
thus , @xmath302 will annihilate the btz state .
but , when the zero mode spectrum is continuous , @xmath302 has no normalizable eigenstates .
to arrive at the geon boundary state for , say , @xmath122 , we need only introduce the basis @xmath303 in terms of eigenvalues @xmath304 of the momenta conjugate to the zero modes of @xmath117 and @xmath201 . the btz state ( [ phiz ] ) for @xmath122 may be written latexmath:[\[\label{ppt }
dp d \tilde p \
\delta(p ) \exp(-\pi \tilde p^2/4 ) no correlations between @xmath201 and @xmath117 .
thus , we may read off from ( [ ppt ] ) the geon boundary states for both @xmath114 and @xmath122 . the ( normalized ) geon state for @xmath114
is given simply by the factor @xmath306 corresponding to @xmath201 in ( [ ppt ] ) , while the geon state for @xmath122 is given by the other factor @xmath307 .
note that the zero mode of @xmath122 is in its ground state .
one might expect this result to be maintained if the field space of @xmath122 could be compactified , in which case the ground state would of course be normalizable .
at this point , a comment is in order on the form of the factor @xmath308 by which we needed to rescale the normalized ground state .
the reader will note that the distribution @xmath309 is in fact the limit ( as a distribution over the configuration space ) of @xmath310 times the normalized harmonic oscillator ground state .
thus , the limit of @xmath242 as @xmath263 is not @xmath3 , but is larger by @xmath311 . that this extra rescaling is necessary results from the fact that the fluctuations of @xmath312 in our state are much smaller than the fluctuations of the momentum in the ground state of the harmonic oscillator .
the geon state for @xmath114 is normalizable but the corresponding state for @xmath122 is not .
we note that the @xmath114 zero mode state can in fact be calculated without dealing with distributions at all .
to do so , one first writes the operator @xmath250 ( [ kom ] ) in terms of creation and annihilation operators for the zero modes of @xmath211 and @xmath200 . as usual , this gives a sum of two operators , one involving @xmath114 and one involving @xmath313 . letting the exponential of ( @xmath281 times ) the @xmath114 part act on a normalized vacuum state gives a one - parameter family of normalized states that converges to the above result _ in the hilbert space norm _ as @xmath263 .
we consider this an important check on our use of distributions above .
note that , since @xmath314 must annihilate the full state @xmath315 , it then follows that @xmath316 ( and therefore @xmath317 is in the zero momentum state .
we now examine the expectation value of the energy ( expected energy ) in our quantum states . for the btz black hole
, we consider the hamiltonian associated with a single boundary component ( say , the one on the right ) .
note that it is the adm hamiltonian of a single asymptotic region that gives the classical mass of the black hole @xcite .
note also that the black hole mass is associated with the killing field @xmath29 of the boundary metric ( [ boundary - metric ] ) .
thus , we consider the notion of energy defined by this vector field .
the thermal behavior noted in section [ osc ] is therefore associated with a temperature @xmath318 .
because the total energy is a sum of the hamiltonians for the left- and right - moving modes separately , it is apparent from an examination of ( [ vacbtzhole - expanded ] ) and ( [ vacbtzhole - proj ] ) that the expected energy of the oscillator modes is _ exactly _ the same for the btz and geon states .
the extra minus signs in ( [ vacbtzhole - proj ] ) do not affect the expectation value and the hamiltonian of , say , the right - moving modes on our boundary component does not care whether a right - moving state there is correlated with a mode on another boundary component or with a left - moving mode on the same boundary component . in both cases ,
the result is just the expected energy in a thermal state of temperature @xmath319 .
the same is also true of the stress - energy tensor .
for a zero mode on a component of the btz boundary , one sees from the regulated expression ( [ bom ] ) that the state again acts like a thermal state at the same temperature .
while such a state is not normalizable in the @xmath263 limit , the expectation value of the energy associated with @xmath29 is finite and equal to @xmath320 . on the geon boundary ,
the expected energy of the @xmath114 zero mode is @xmath321 while that of the @xmath122 zero mode vanishes . for the oscillator modes of a free scalar field on a cylinder
, the energy expectation value in a thermal state is well known @xcite .
the circumference of our cylinder ( [ boundary - metric ] ) is @xmath130 , and the temperature is @xmath320 : with these parameters , one finds from the general formulas given in ref .
@xcite that the energy of our oscillator state relative to the ground state ( casimir ) energy is @xmath322 to connect these results within our model with the full cft , we recall that our model theory has central charge @xmath123 while the full cft of @xcite has central charge @xmath323 .
since the field space of @xmath114 represents one of the four @xmath62 factors of the internal torus ( and thus one fourth of the non - linear sigma model ) , we may expect that a central charge of @xmath324 is associated with the part of the sigma model that is similar to @xmath114 , while the remaining @xmath325 is associated with fields similar to @xmath122 .
we therefore model the energy of the full theory with @xmath326 copies of @xmath114 and @xmath325 copies of @xmath122 .
as noted above , the energy is the same for both @xmath122 and @xmath114 on the btz boundary .
thus , the total energy there is given by @xmath327 times the expression ( [ therm ] ) plus @xmath328 for the zero modes . for our @xmath4 geon , the zero mode of @xmath122 is in its ground state but the zero mode of @xmath114 has energy @xmath321 .
thus , the total energy is @xmath124 times expression ( [ therm ] ) plus @xmath329 . in the limit @xmath330 ( large black hole and large temperature of the thermal states )
the zero mode correction is negligible and ( using @xmath331 ) the expected energy reduces to @xmath332 . in our notation
( and for the spinless case we consider ) , @xmath333 and @xmath334 of @xcite are given by @xmath335 .
we also note that all energies in @xcite were computed with respect to a killing vector field that corresponds to @xmath336 , where @xmath139 is defined in @xcite . with this understanding
, our result in this limit agrees with @xcite ( which did not take into account the discrete nature of the field modes ) .
note that we have not set the three - dimensional newton s constant @xmath11 equal to one and , in fact , it is fixed @xcite by the relation between the anti - de sitter space and the central charge of the cft . since we have set the length scale @xmath19 of anti - de sitter space to one ,
@xmath337 is @xmath338 ( times the string scale ) .
it follows that the energy of our cft state also agrees with the classical black hole mass @xmath339 for @xmath330 .
this observation was made in @xcite ( in which this limit was taken implicitly ) in the context of btz holes .
note that taking @xmath330 gives the limit in which the black hole is much larger than the radius of curvature of the ads space , and it is in this regime that the energy of a thermal bath geon analogue of the scalar field vacuum on the @xmath5 geon constructed in @xcite . ] required to maintain equilibrium would be small compared to the mass of the black hole .
in this section we investigate a cft on the boundary of another geon - type , single - exterior , ( 2 + 1)-dimensional black hole spacetime : the spinless black hole with spatial topology @xmath340@xmath90point at infinity@xmath91 constructed in ref .
@xcite and analyzed in detail in ref .
we refer to this spacetime as the swedish geon . we will not be able to obtain the cft state as explicitly as for the @xmath4 geon , but we can reduce the problem of finding this state to a mathematical problem involving certain automorphic functions
. we will also be able to contrast the correlations present in the swedish geon state to those present in the @xmath4 geon state . as the swedish geon is space and time orientable
, we consider as a model theory a single conformal scalar field @xmath122 that lives on the boundary of the spacetime .
let us briefly recall the construction of the swedish geon and its conformal boundary @xcite .
let @xmath41 and @xmath341 be on @xmath14 the killing vectors respectively induced by the killing vectors @xmath342 of @xmath16 .
the conformal killing vectors induced on @xmath39 are respectively @xmath343 note that @xmath41 and @xmath45 are as in section [ classic ] .
the swedish geon is now defined as the quotient of a certain subset of @xmath14 under the infinite discrete group generated by @xmath344 and @xmath345 , where the parameter @xmath56 satisfies @xmath346 .
the geon is space and time orientable , it admits a global foliation with spacelike hypersurfaces of topology @xmath340@xmath90point at infinity@xmath91 , and it has a single exterior region isometric to that of a spinless nonextremal btz black hole with horizon circumference @xmath347 $ ] .
the singularities , and the exotic topology , are hidden behind the horizon .
it follows from the above that the conformal boundary of the swedish geon consists of just one copy of the conformal boundary of @xmath14 . as explained in detail in ref .
@xcite , this boundary emerges from the boundary @xmath39 of the original @xmath14 as the quotient of a set @xmath348 under the discrete group @xmath349 generated by @xmath350 and @xmath351 .
@xmath352 consists of a countable number of disconnected diamonds , each of them the domain of dependence of an open interval in the @xmath48 circle : the end points of the intervals are at the fixed points of @xmath349 on this circle . recall from section [ classic ] that it was possible to describe the @xmath4 geon boundary by considering just one of the diamonds of @xmath353 , and taking its quotient under the identification subgroup that maps this diamond to itself .
a similar description is possible for the swedish geon boundary @xcite . among the countably many diamonds constituting @xmath352 ,
let @xmath354 be the one that intersects @xmath48 in the interval @xmath355
< \theta < ( \pi/4 ) + \arccos[c/(\sqrt{2}s)]$ ] , where @xmath356 and @xmath357 .
it can be shown that the only elements of @xmath349 that leave @xmath354 invariant are powers of @xmath358 , and that the boundary of the swedish geon is the quotient of @xmath354 under the @xmath359 generated by @xmath360 .
now , @xmath360 can be written as @xmath361 , where @xmath362 \ \ .\ ] ] @xmath363 is a conformal killing vector on @xmath39 , and its fixed points at @xmath48 are precisely at the corners of @xmath354 , at @xmath364 $ ] .
the conformal killing vector @xmath363 is thus analogous to the conformal killing vector @xmath45 on @xmath54 in section [ classic ] . in particular
, @xmath354 admits a future timelike conformal killing vector orthogonal to @xmath363 , analogous to @xmath46 on @xmath54 , and this conformal killing vector defines the positive and negative frequencies on the swedish geon boundary .
consider our conformal scalar field @xmath122 on @xmath365 .
we introduce on @xmath354 the null coordinates @xmath366 : @xmath367 \ \ , \\ & & v : = t + [ \theta - ( \pi/4 ) ] \ \ , \end{aligned}\ ] ] which cover @xmath354 with @xmath368 $ ] and @xmath369 $ ] . in analogy with ( [ rindler - modes ] )
, one finds that a complete orthonormal basis for the oscillator modes of @xmath122 on @xmath365 , positive frequency with respect to the geon boundary time , is [ sgeon - modes ] @xmath370^{-2\pi i n/\gamma } \ \ , \label{sgeon - modes - u } \\ & & u_{n,- } : = \frac{1}{\sqrt{4\pi n } } \left [ \frac{v+\tan(v/2)}{v-\tan(v/2 ) } \right]^{-2\pi i n/\gamma } \ \ , \end{aligned}\ ] ] where @xmath174 takes values in the positive integers and @xmath371 the subscript @xmath372 ( @xmath373 ) yields the right - moving ( left - moving ) modes . when the metric on the geon boundary is written as in ( [ boundary - metric ] ) , the frequency with respect to @xmath29 is just @xmath174 .
the vacuum of the modes ( [ sgeon - modes ] ) is therefore the usual vacuum on the geon boundary for the oscillator modes of the field : we denote this vacuum by @xmath374 .
the usual vacuum for the zero modes is again nonnormalizable ; from now on we restrict the discussion to the oscillator modes .
we would now like to use unruh s analytic continuation method @xcite to find the oscillator mode state @xmath375 that is induced on the boundary of the geon by the usual oscillator mode vacuum @xmath150 on @xmath39 .
this means that we must form from the @xmath376-modes ( [ sgeon - modes ] ) and their complex conjugates linear combinations , the @xmath151-modes , that satisfy two requirements .
first , when analytically continued to the lower half of the complex planes in @xmath149 and @xmath377 , the @xmath151-modes must be bounded analytic functions : this guarantees that they are purely positive frequency linear combinations of the modes that define @xmath150 .
note that the @xmath151-modes may have singularities at certain real values of @xmath149 and @xmath377 , but apart from these singularities , the analytic continuation defines them as functions on all of @xmath39 and not just in the diamond @xmath378 .
second , the @xmath151-modes must accommodate the fact that the geon boundary field operator on @xmath352 is constructed by averaging @xmath122 over @xmath349 , as in ( [ geonfields - phi ] ) .
this means that the @xmath151-modes must be invariant over @xmath349 , while each of the @xmath376-modes ( [ sgeon - modes ] ) , when analytically continued from @xmath354 to @xmath352 , is individually invariant only under the subgroup of @xmath349 generated by @xmath360 .
it is easy to see that @xmath349 takes @xmath149-independent functions into @xmath149-independent functions and similarly @xmath377-independent functions into @xmath377-independent functions .
the @xmath151-modes can therefore be divided into right - movers , constructed from @xmath379 and their complex conjugates , and left - movers , constructed from @xmath380 and their complex conjugates . for concreteness , consider the right - movers . to put the problem into a mathematically familiar form , we replace @xmath149 by the coordinate @xmath381 . in terms of @xmath382
, @xmath39 corresponds to the compactification of the real line , and @xmath354 is covered by the interval @xmath383 , where @xmath384 .
analytic continuation of @xmath149 into the lower half - plane is equivalent to analytic continuation of @xmath382 into the upper half - plane .
the generators @xmath385 and @xmath34 act on @xmath382 as fractional linear transformations whose matrices [ which , as elements of @xmath386 , are defined only up to the overall sign ] are @xmath387 the @xmath151-modes are thus the bounded analytic functions in the upper @xmath382 half - plane that are invariant under the group generated by @xmath388 and @xmath389 .
and @xmath389 are boosts with magnitude @xmath56 , the fixed points of @xmath388 are at @xmath390 and @xmath391 , and the fixed points of @xmath392 are at @xmath393 .
this makes the @xmath151-modes automorphic functions @xcite on the noncompact riemann surface that is isomorphic to the time - symmetric hypersurface in the geon _
spacetime_. for a fundamental domain for this riemann surface , see ref .
@xcite . ] expressing the modes @xmath379 ( [ sgeon - modes - u ] ) in terms of @xmath382 , we see that finding the boboliubov transformation reduces to finding the coefficients @xmath394 in the expansions @xmath395 ^{2\pi n i/\gamma } \ \ , \label{sge - wexp}\ ] ] where the terms with positive @xmath174 come from the @xmath396 and the terms with negative @xmath174 come from the complex conjugates .
note that , by construction , each term on the right - hand - side of ( [ sge - wexp ] ) is invariant under the fractional linear transformation @xmath397 , whose fixed points are at @xmath398
. we shall not pursue the analysis further here , but we make one speculative comment . on the real axis , both @xmath388 and @xmath389 map the interval @xmath383 completely outside this interval .
if @xmath383 , both @xmath388 and @xmath389 thus take each term in the sum ( [ sge - wexp ] ) to a term whose magnitude differs by the factor @xmath399 .
this suggests ( but certainly does not prove ) that if the sum is to be invariant , the coefficients @xmath394 should be exponentially increasing in @xmath400 . if this is true , comparison with the relative weights of the terms in ( [ w - modes ] ) suggests ( but again certainly does not prove ) that @xmath375 might appear in some respects as a thermal state in a temperature proportional to @xmath401 .
we leave the examination of these speculations subject to future work .
we have seen that , in our model , an @xmath4 geon corresponds to a pure state of finite energy on the cads@xmath0 boundary . in particular , these states contain correlations between the right- and left - moving sectors such that , when one of these sectors is traced over , the other sector is left in a thermal state .
the expectation value of the hamiltonian in an @xmath4 geon state is exactly the same as the expectation value of the hamiltonian ( for either of the boundary components ) in the corresponding btz black hole state , and this value agrees with the the classical mass of the spacetimes in the limit where the black hole is much larger than the length scale of the ads space .
in our model , the zero mode parts of our states @xmath402 and @xmath403 were not normalizable .
this was due to the noncompact range of our fields and the resulting continuous spectrum of the zero mode hamiltonian .
for the same reason , the ground state @xmath404 of our model theory is again non - normalizable .
since our final states were in fact induced from the ground state , it is no surprise that our construction failed to generate normalizable states .
indeed , the surprise is that the field @xmath114 is in a normalizable state on the geon boundary . in all cases ,
we arrive at a generalized state that may be expressed in the usual way in terms of distributions .
while we were not able to complete a corresponding analysis of the swedish geon states , the fact that these black holes have only a single asymptotic region ( with topology @xmath1 ) leads one to once again expect pure states .
still , calculating one of these states would be of interest as it is far from clear what sort of correlations it would contain . in particular , in contrast with an @xmath4 geon , the identifications that lead to a swedish geon act separately on the right- and left - moving parts of the cft .
thus , there should be no correlations between right- and left - moving modes and the correlations must take a rather different form than for an @xmath4 geon . nonetheless , the hypothesis that a swedish geon is ` not too far ' from a thermal state is supported by the behavior of the ( as yet formal ) unruh modes of section [ swedish ] under analytic continuation . another motivation for studying
the swedish geon is that , since the identifications that yield an orientable swedish geon need not act on the internal factors , such geons may more readily allow a treatment of fields with compact target spaces .
one might also try to generalize our calculation to higher - dimensional single - exterior locally ads black holes constructed from the two - exterior locally ads black holes of refs .
@xcite via a suitable involution .
however , such single - exterior black holes are of somewhat less interest as their asymptotic topology is always different from the asymptotic topology of ads space .
thus , the black hole and the ads space will in any case not correspond to quite the same boundary field theory .
of course , the real interest is to extrapolate our results to the more complicated cft which forms the basis of maldacena s duality conjecture @xcite .
the main difficulty with our model was that we were unable to capture the compact nature of the moduli space of this theory .
it is unclear to us to what extent quotients of the type described here can actually be carried out in a nonlinear field theory . it is natural to assume , however , that the major qualitative difference between our model and the full theory is that the btz and geon states become normalizable , placing the fields @xmath117 and @xmath200 in their ground states
certainly , we would once again expect an @xmath4 geon to be associated with a pure state ( of finite energy ) in the usual hilbert space .
it should contain correlations of the sort found here , between identical right- and left - moving modes , again defining a thermal state with respect to either the right- or left - moving sector alone .
although we have not discussed fermions in our model , if they were included with antiperiodic boundary conditions the associated geon state would be interpreted as an excited state of the cft vacuum representing ads space .
on the other hand , fermions may also be included with periodic boundary conditions , in which case the ground state of the cft is associated @xcite with the @xmath2 btz black hole and our geon is a corresponding excited state .
these two cases would correspond to string theories twisted in different ways around the nontrivial topology of the black hole throat . as excited states , the @xmath4 geon states
are certainly not invariant under time translations .
in fact , an inspection of ( [ vacbtzhole - proj ] ) shows that they are not even stationary .
this is in accordance with the fact that the timelike killing field in the exterior region of the geon spacetime can not be extended to a globally - defined killing field on all of the spacetime .
now , in the theory considered here , all of the modes ( except the zero modes ) of our scalar fields are periodic in time with a common period .
thus , the oscillator part of our geon state is actually periodic in time .
this is not a feature of the classical geon spacetime and it is _ not _ a feature that one would expect to survive in the full boundary cft .
indeed , already in describing the original ads space we see that this periodicity must be broken ( by the anomalous dimensions of certain operators in the cft @xcite ) if the boundary theory is to describe aperiodic processes . in order to see what should be expected when this periodicity is broken
, we note that the construction of the geon boundary state is in direct parallel with the construction given in @xcite of vacua on the entire asymptotically flat @xmath5 geon spacetime and on an analogous rindler - type spacetime .
in those cases , it was found that the correlations between field modes became unobservable by localized detectors far from the preferred time @xmath48 and that the state behaved for many purposes as a thermal ( _ i.e. _ , mixed ) state .
clearly , we expect parallel results here .
it is true that , since any correspondence between the boundary cft and the bulk string theory will be nonlocal , the relevance of local detectors on the boundary is unclear .
however , one still expects that the geon state will , in an appropriate sense , approximate the btz state over a single boundary component at early and late times .
perhaps one of the most interesting aspects of our calculation is the way in which the reflection of an internal @xmath62 is represented in the boundary quantum state .
our results are similar to those of @xcite , in that the identifications on the internal dimensions are reflected in certain symmetries of the cft quantum states . in our model , this involved the state of the scalar field @xmath114 , which was associated with the @xmath62 factor on which our spacetime identifications act . by including more of the full nonlinear sigma model and thus capturing more of the internal dimensions
, we could arrive at similar states that correspond to other orientable @xmath4-like geons with different internal spaces , including orbifolds .
in contrast , the complicated topology of the swedish geons is associated only with the ads factor of the spacetime ( and not with the internal compact dimensions ) .
it would therefore be interesting to probe this issue further through a full calculation of a swedish geon boundary state .
finally , we note that if @xmath200 were the physical field on the geon , the oscillator state would differ only by removing the factors of @xmath405 . placing the zero mode in its ground state , one constructs in this way a state that one is tempted to associate with string theory on a non - orientable @xmath4 geon .
thus , one might speculate that it may be possible to describe states of non - orientable string theory in terms of the same cft hilbert space .
whether or not this happens in the full theory or is merely an artifact of our model must , of course , be left for future studies .
we would like to thank vijay balasubramanian , ingemar bengtsson , steve carlip , nico giulini , gary horowitz , albion lawrence , juan maldacena , and max niedermaier for discussions and correspondence . d.m .
was supported in part by nsf grant phy-9722362 and research funds provided by syracuse university . d.m . would like to thank the max - planck - institut fr gravitationsphysik for their hospitality during the final stages of this work .
i. klebanov , nucl .
* b496 * , 231 ( 1997 ) ( hep - th/9702076 ) ; s. gubser and i. klebanov , phys .
b * 413 * , 41 ( 1997 ) ( hep - th/9708005 ) ; m. douglas , j. polchinski , and a. strominger , j. high energy phys .
* 12 * , 3 ( 1997 ) ( hep - th/9703031 ) .
|
in the context of the conjectured ads - cft correspondence of string theory , we consider a class of asymptotically black holes whose conformal boundary consists of a _ single _ connected component , identical to the conformal boundary of space . in a simplified model of the boundary theory
, we find that the boundary state to which the black hole corresponds is pure , but this state involves correlations that produce thermal expectation values at the usual hawking temperature for suitably restricted classes of operators .
the energy of the state is finite and agrees in the semiclassical limit with the black hole mass .
we discuss the relationship between the black hole topology and the correlations in the boundary state , and speculate on generalizations of the results beyond the simplified model theory .
= 10000
|
the desire to create an optical directional coupler , a device composed by parallel optical waveguides close enough that leaked energy is transferred between them , led to the exploration of waveguide creation in semi - conductors via proton bombardment @xcite . at the time , power losses played an interesting role and the nascent mode coupling theory @xcite allowed the theoretical description of linear loses in such devices @xcite , @xmath2,\end{aligned}\ ] ] where the real numbers @xmath3 and @xmath4 are the effective linear loss , identical in all implanted waveguides , and the effective waveguide coupling strength , also indentical for the whole system , in that order .
this , to the best of our knowledge , was the first theoretical description of an experimental @xmath1-waveguide coupler including losses in the form of a schrdinger - like equation involving a non - hermitian hamiltonian .
almost twenty years later , the desire to create an intensity dependent switch working at low power levels took another team of researchers to explore twin core nonlinear couplers with gain and loss .
again , coupled mode theory allowed the description of such devices @xcite , @xmath5 where we have kept the notation used before and introduce the effective real part of the refractive index , @xmath6 , and the real function @xmath7 that describes an effective kerr nonlinearity induced change in the refractive index of the @xmath8th core , positive for self - focusing and negative for self - defocusing materials . at this point in history , there existed experimental and theoretical work describing an optical dimer where the waveguides present effective loss and gain and a nonlinearity but a little something was missing .
a couple of years later , a theory exploring a particular type of non - hermitian hamiltonians with real spectra was brought forward in quantum mechanics @xcite .
these hamiltonians were invariant under space - time reflection , received the name of @xmath0-symmetric , and opened new avenues of research in quantum mechanics as well as other areas of physics and matemathics , c.f . @xcite and references therein . in optics
, it took a few years more to propose two seminal ideas .
the first one regarded single elements and showed , in particular , that an optical planar slab waveguide composed of two media with linear gain and loss can be described by a schrdinger - like equation under dynamics dictated by a @xmath9-symmetric hamiltonian , where the optical refractive index played the role of quantum - like potential and propagation distance that of time @xcite .
the second one dealt with composite systems , where a mature mode coupling theory produced a theory of coupled optical pt - symmetric structures @xcite .
in the ten years following those first proposals for an optical realization of @xmath9-symmetry , work has been reported on slab waveguides @xcite , bragg scatterers @xcite , as well as linear @xcite and nonlinear @xcite coupled waveguides , to mention just a few .
research in this field is slowly getting to information technologies applications with recent proposals of all - optical @xmath0-symmetric logic gates @xcite and amplitude - to - phase converters @xcite . here
, we will provide a review of the optical @xmath0-symmetric dimer .
first , in section 2 , we will introduce a two - waveguide coupler where component waveguides show effective complex refractive indices with identical real part .
starting from this device , we will recover the effective mode coupling differential equation set for the linear @xmath0-symmetric dimer , which describes a nonunitary optical device showing symmetric effective loss and gain feasible of passive and active optical realizations .
then , we will recover the dispersion relation for the dimer that shows three regimes , one with real eigenvalues , the @xmath0-symmetric regime , another with fully degenerate eigenvalues equal to zero , the fully degenerate regime , and a third one with purely imaginary eigenvalues , the broken symmetry regime .
we will construct an analytic propagator that will show asymmetric amplifying oscillator , power amplification , and exponential amplification behaviors in each of these regimes .
we will also show that it is possible to uncouple the mode coupling differential equation set of the @xmath0-symmetric dimer .
the resulting second order differential equations and boundary conditions for the field amplitudes propagating through each waveguides take the form of nonlocal oscillators with positive potential , free particle traveling through a nonlocal medium , and nonlocal oscillators with inverted potential , in each of the regimes . in the final part of section 2 , we will bring forward the renormalized field approach that helps us cast the linear @xmath0-symmetric dimer as a nonlinear dimer with imaginary kerr nonlinearity , either in a self- or a cross - modulation scheme , and allows us to realize an asymptotic behavior that depends just on the gain to coupling strength ratio of the device . in section 3 , we will discuss the linear @xmath0-symmetric dimer when both waveguides show the same effective self - focusing kerr nonlinearity .
we will show the stable nonlinear modes of the device , discuss its dynamics in terms of the passive kerr two - waveguide coupler that allows for coherent and localized oscillations between the waveguide field modes , and show that the inclusion of symmetric gain and loss breaks these dynamics producing localization in the gain waveguide without showing an asymptotic behavior . in section 4
, we will extend the linear @xmath0-symmetric dimer to planar @xmath1-waveguide couplers using finite dimensional matrix representations of a complexified version of @xmath10 .
we have previously shown @xcite that the @xmath0-symmetric dimer and its extensions to planar @xmath1-waveguide couplers possess an @xmath11 symmetry realized in a finite dimensional non - unitary irreducible representation .
this representation is accomplished through complexification of @xmath10 , @xmath12 , and it allows us to provide the dispersion relation and a closed form analytic propagator , which have the same regimes and dynamics found for the dimer .
we will show that the renormalized field approach provides us with an asymptotic behavior that is independent of the initial field distribution and depends just on the waveguide number and the effective gain to coupling ratio .
then , in section 5 , we will introduce a modified version of the ehrenfest theorem suitable for non - hermitian hamiltonians and show how it can help us define the dynamics of an @xmath13-dimensional generalized stokes vector for the planar @xmath1-waveguide couplers discussed in section 4 . in section 6 , we will go back to the dimer but consider the propagation of quantum fields . in the quantum regime , spontaneous generation and absorption of electromagnetic radiation
should be considered when using media with linear gain or loss .
we will show the solution for the quantum linear @xmath0-symmetric dimer and discuss the generation of light from vacuum due to spontaneous processes in the absence of fields impinging at the device .
finally , we will produce a brief summary and discuss future avenues regarding non - hermitian optical systems .
a two - waveguide coupler can be described by the following differential equation system via mode coupling theory , @xmath14 where the complex numbers @xmath15 are the field amplitudes at each waveguide , the parameters @xmath16 and @xmath17 , which in the most general case can be complex , are the effective refractive indices and waveguide coupling , in that order , and the operator @xmath18 stands for the derivative with respect to the propagation distance @xmath19 .
note that using field amplitudes of the form @xmath20 with @xmath21 , reduces the system to an effective mode coupling matrix , @xmath22 the mode coupling matrix has eigenvalues @xmath23 the eigenvalues are real for the case of identical real part of the effective refractive indices , @xmath24 , and imaginary part less than the value of the effective coupling , @xmath25 .
the eigenvalues degenerate in the case @xmath24 and @xmath26 .
they are purely imaginary for @xmath27 and @xmath28 and , finally , complex elsewhere .
this general non - hermitian dimer has a rich structure that deserves further attention but , right now , we are interested in just the @xmath0-symmetric case . in quantum mechanics , pt - symmetry refers to space - time reflection symmetry @xcite . in discrete optical couplers we can consider waveguide permutation and propagation inversion as equivalent to space and time reflection , respectively .
then , in order to recover the standard linear @xmath9-symmetric dimer , we need to work with waveguides that have the same effective refractive indices , @xmath29 such that @xmath30 is purely imaginary and we can write a differential set @xcite , @xmath31 where permutation of the waveguides , @xmath32 , and propagation reversal , @xmath33 , leaves the system invariant . note that we have used the effective coupling parameter to scale the propagation distance , @xmath34 , such that we deal with a single parameter given by the effective refractive index to coupling ratio , @xmath35 with @xmath36 .
this mathematical model is equivalent to consider waveguides with effective pure linear loss and gain , @xmath37 and @xmath38 in that order . in optics ,
the linear @xmath0-symmetric dimer has been experimentally demonstrated in passive lossy waveguides @xcite , fig .
[ fig : figure1](a ) and fig .
[ fig : figure1](b ) , as well as active , pumped waveguides
@xcite , fig .
[ fig : figure1](c ) , and pumped whispering - gallery mode microcavities @xcite , fig .
[ fig : figure1](d ) , with linear gain .
the experimental demonstration of @xmath0-symmetric devices is not limited to optical resonators , they have also been realized with operational amplifiers in electronics @xcite .
-symmetric dimer .
( a ) passive laser engraved waveguides , ( b ) passive waveguides with metallic scatterers , ( c ) pumped active waveguides , ( d ) pumped active whispering - gallery mode microcavities . ] let us try to address propagation in the linear @xmath0-symmetric dimer .
first , note that we can cast its coupled differential equation system , eq .
( [ eq : ptdimer ] ) , in a vector form , @xmath39 similar to schrdinger equation .
now , the coupling matrix takes the role of a hamiltonian , @xmath40 where the operators @xmath41 with @xmath42 are pauli matrices . in any given case ,
it is straightforward to find the propagator for this @xmath43-independent schrdinger - like equation @xcite , @xmath44 such that the propagated fields through the device are given in terms of the initial field configuration , @xmath45 with the dispersion relation given by @xmath46 which can be real for @xmath47 , zero for @xmath48 , or purely imaginary for @xmath49 .
note that the eigenvalues of the coupling matrix are given by @xmath50 and they become fully degenerate at @xmath48 .
figure [ fig : figure2 ] shows the behavior of the coupling matrix eigenvalues as a function of the gain to coupling ratio as they go from purely real , fig .
[ fig : figure2](a ) , degenerate to zero , fig .
[ fig : figure2](b ) , and become purely imaginary , fig .
[ fig : figure2](c ) .
-symmetric regime , ( b ) fully degenerate regime , and ( c ) broken symmetry regime .
the black arrows show the direction of the eigenvalues as the gain to coupling ratio increases . ]
it is straightforward to show that in the @xmath0-symmetric regime , where the eigenvalues of the mode coupling matrix are real , the device behaves as an asymmetric oscillator with amplification , @xmath51 once the @xmath9-symmetry is broken , we have two distinct cases , the fully degenerate one where both eigenvalues are zero and the device shows amplification ruled by a power law , @xmath52 and the case of purely imaginary eigenvalues , where the amplification is exponential , @xmath53 figure [ fig : figure3 ] shows the absolute field amplitude propagating through a @xmath0-symmetric dimer when light impinges just at the first waveguide in a device with parameters in the regime with real eigenvalues , fig .
[ fig : figure3](a ) , fully degenerate eigenvalues , fig .
[ fig : figure3](b ) , and imaginary eigenvalues , fig .
[ fig : figure3](c ) .
now , while we find this algebraic approach short , elegant and elucidating , it is not the only available method to infer the properties of the @xmath0-symmetric dimer .
-symmetric regime , @xmath54 ,
( b ) fully degenerate regime , @xmath55 , and ( c ) broken symmetry regime , @xmath56 . ]
it is also possible to decouple the dimer differential equation set , eq .
( [ eq : ptdimer ] ) , into two second order differential equations that have the same form of the one - dimensional wave equation @xcite , @xmath57 the boundary conditions are given by the initial fields impinging the device , @xmath58 , and their first derivatives , @xmath59 obtained from the dimer mode coupling equation , eq.([eq : ptdimer ] ) .
note that the latter takes us afar from the standard one - dimensional wave equation because it refers to nonlocal media involving linear loss and gain .
nevertheless , we can follow the standard partial differential equation approach . in the @xmath0-symmetric case ,
the characteristic equation is positive , @xmath60 , and we can write the second order differential equation as that of a standard oscillator , @xmath61 thus , we can use its well known solution plus our particular boundary conditions to obtain the propagated fields , @xmath62 \sin\left ( \vert \omega \vert \zeta \right ) , \\ { { \mathcal{e } } _ { 2}}\left ( \zeta \right)&=&{{\mathcal{e}}_{2}}\left(0\right ) \cos\left ( \vert \omega \vert \zeta \right ) + \frac{1 } { \vert \omega \vert } \left [ \gamma { { \mathcal{e}}_{2}}\left(0\right ) + i { { \mathcal{e}}_{1}}\left(0\right ) \right ] \sin\left ( \vert \omega \vert \zeta \right).\end{aligned}\ ] ] these fields allow us to describe the dimer as an asymmetric periodic oscillator with amplification . it is not an harmonic oscillator due to the boundary condition on the first derivatives . in the fully degenerate case , where the characteristic equation is equal to zero , @xmath63 , @xmath64 we can think of light propagating through the dimer as a free particle through some nonlocal media with linear unitary gain and loss , that yields amplification following a power law , @xmath65 \zeta , \\ { { \mathcal{e}}_{2}}\left ( \zeta \right)&=&{{\mathcal{e}}_{2}}\left(0\right ) + \left [ { { \mathcal{e}}_{2}}\left(0\right ) + i { { \mathcal{e}}_{1}}\left(0\right ) \right ] \zeta.\end{aligned}\ ] ] finally , in the broken symmetry case , the characteristic equation is negative , @xmath66 , and we can write the second order differential equation as an inverted oscillator , @xmath67 that provides us with a device that amplifies initial fields following an exponential law , @xmath68 \sinh\left ( \vert \omega \vert \zeta \right ) , \\ { { \mathcal{e } } _ { 2}}\left ( \zeta \right)&=&{{\mathcal{e}}_{2}}\left(0\right ) \cosh\left ( \vert \omega \vert \zeta \right ) + \frac{1 } { \vert \omega \vert } \left [ \gamma { { \mathcal{e}}_{2}}\left(0\right ) + i { { \mathcal{e}}_{1}}\left(0\right ) \right ] \sinh\left ( \vert \omega \vert \zeta \right ) .\end{aligned}\ ] ] all these solutions are just the explicit form of the propagated field found earlier by a purely algebraic approach , eq . ( [ eq : propagatedfield ] ) eq . ( [ eq : propexpamp ] ) , and tap into the well known one - dimensional wave equation with the difference that an effective nonlocal active medium is provided by the first derivative boundary conditions . so far , we have managed to provide an algebraic propagator and describe the field behavior in the three possible regimes of the linear @xmath0-symmetric dimer .
now , we can bring forward a complementary view that can give us asymptotic information of the broken symmetry phases .
let us define instantaneous renormalized fields @xcite , @xmath69 such that the total renormalized field intensity at each propagation distance is always the unit , @xmath70 . in this picture , it is easier to realize that light intensity through the @xmath0-symmetric device behaves like a nonharmonic oscillator , fig .
[ fig : figure4](a ) .
furthermore , this allows us to conduct asymptotic analysis in the broken symmetry phases . in the fully degenerate phase , @xmath48 such that @xmath71 , it is possible to calculate the asymptotic behavior of the fields intensity as the scaled propagation distance goes to infinity and find out that the renormalized optical power is balanced in both waveguides , independent of the initial field distribution , @xmath72 as shown in fig .
[ fig : figure4](b ) .
in the broken symmetry regime , where the eigenvalues are purely imaginary , the asymptotic intensity distribution depends on the effective gain to coupling ratio and we can include the previous result , @xmath73 in other words , for a device long enough , in the broken symmetry region , the input field distribution has no effect on the output field intensity distribution .
the latter is governed only by the gain to coupling ratio of the device , @xmath74 , fig .
[ fig : figure4](c ) . solid blue line , and gain
, @xmath75 dashed red line , in the ( a ) @xmath0-symmetric regime , @xmath54 , ( b ) fully degenerate regime , @xmath55 , and ( c ) broken symmetry regime , @xmath56 , for an initial field impinging just at the first waveguide . ] furthermore , this approach also allows us to show that the linear @xmath0-symmetric dimmer is the irreducible form of two equivalent reducible nonlinear models , one with linear loss and gain , as well as an imaginary nonlinearity , @xmath76 \tilde{\mathcal{e}_{1}}(\zeta ) , \\ - i \partial_{\zeta } \tilde{\mathcal{e}_{2}}(z ) & = & \tilde{\mathcal{e}_{1}}(\zeta ) - 2 i \gamma \left [ 1- \vert \tilde{\mathcal{e}_{2}}(\zeta ) \vert^{2 } \right ]
\tilde{\mathcal{e}_{2}}(\zeta ) , \end{aligned}\ ] ] and the other corresponding to an imaginary cross - nonlinearity , @xmath77 we can follow a standard approach to deal with nonlinear @xmath0-symmetric dimers @xcite , and introduce a stockes vector , @xmath78 , with components given by , @xmath79 again , the matrices @xmath41 are pauli matrices , the notation @xmath80 is a column vector containing the renormalized field amplitudes equivalent to that defined in section 2 for the field amplitudes , and the new conjugate transpose notation @xmath81 is a row vector with the conjugate renormalized field amplitudes as components . in terms of the renormalized field amplitudes , @xmath82 ,
\\ \tilde{\mathcal{s}}_{z}(\zeta)&=&\vert \tilde{\mathcal{e}}_{1}(\zeta)\vert^2-\vert\tilde{\mathcal{e}}_{2}(\zeta)\vert^2.\end{aligned}\ ] ] in the case at hand , where the nonlinearity is reducible , the stokes vector norm is the unit and a constant of motion , @xmath83 the stokes vector components verify the following set of coupled differential equations : @xmath84 , \label{eq : rnsvy}\\ \partial_{\zeta}\tilde{\mathcal{s}}_{z}(\zeta)&= & -2 \left\ { \tilde{\mathcal{s}}_{y}(\zeta ) + \gamma \left[1 - \tilde{\mathcal{s}}_{z}^2(\zeta ) \right ] \right\}. \label{eq : rnsvz}\end{aligned}\ ] ] note that these equations have a set of stable points only outside the @xmath0-symmetric regime , @xmath85 , given by @xmath86 .
it is also possible to describe the asymptotic behavior of the renormalized stokes vector components , @xmath87 the stokes vector approach allows us to visualize the field propagation as a trajectory on a unit sphere .
any initial condition , in the fully degenerate and broken symmetry regimes , will converge asymptotically to the same stable point on the sphere of the differential equation , @xmath88 , in a dimer described by the gain to coupling ratio @xmath74 .
figure [ fig : figure5 ] shows the stokes vector propagation related to the examples given in fig .
[ fig : figure4 ] and an additional initial condition set to show that the asymptotic behavior is independent of the initial conditions .
this asymptotical behavior suggest the use of this device as an unidirectional variable amplitude coupler .
-symmetric regime , @xmath54 , ( b ) fully degenerate regime , @xmath55 , and ( c ) broken symmetry regime , @xmath56 , for the initial conditions @xmath89 and @xmath90 in black , and @xmath91 and @xmath92 in red . ]
now , we will discuss a variation of the @xmath0-symmetric dimer where the waveguides have an additional kerr nonlinearity , @xmath93 \mathcal{e}_{1}(\zeta ) + \mathcal{e}_{2}(\zeta ) , \\ - i \partial_{\zeta } \mathcal{e}_{2}(\zeta ) & = & \left [ -i \gamma + \kappa \vert \mathcal{e}_{2}(\zeta ) \vert^{2 } \right ] \mathcal{e}_{2}(\zeta ) + \mathcal{e}_{1}(\zeta).\end{aligned}\ ] ] for the sake of simplicity , we will consider the effective kerr nonlinearity to coupling strength ratio , @xmath94 , to be equal in both waveguides .
stability in this system has been formally discussed in the case of @xmath95 and it was found that the zero equilibrium state is neutrally stable in the @xmath0-symmetric regime , @xmath47 , and that the total light intensity at the waveguides is bounded from above by the initial intensity amplified by an exponential gain , @xmath96 @xcite . in order to deal with the dynamics , it is possible to introduce a stokes vector defined , now , in terms of the field amplitudes @xcite , @xmath97 \label{eq : svy } , \\
\mathcal{s}_{z}(\zeta)&= & \vert { \mathcal{e}}_{1}(\zeta)\vert^2-\vert { \mathcal{e}}_{2}(\zeta)\vert^2 .
\label{eq : svz}\end{aligned}\ ] ] thus , the norm of the redefined stokes vector , the total intensity at the waveguides , is no longer a constant of motion , @xmath98 and its dynamics , @xmath99 , \label{eq : svnly } \\
\partial_{\zeta } \mathcal{s}_{z}\left ( \zeta \right ) & = & - 2 \left[\mathcal{s}_{y } \left ( \zeta \right ) + \gamma \mathcal{s}\left ( \zeta \right ) \right ] , \label{eq : svnlz}\end{aligned}\ ] ] will not be restricted to the unit sphere .
these dynamics have been shown to be an optical simulation of a relativistic massless particle of negative charge in a pseudolectromagnetic field @xcite .
note that there is a set of stable points , @xmath100 , for the effective gain to coupling ratio @xmath47 , @xmath101 all these stable nonlinear modes are waveguide fields of the form @xmath102 with @xmath103 and phase difference constricted by the relation , @xmath104 there are no stable points outside the @xmath0-symmetric regime , @xmath85 . in order to create intuition ,
let us start with the passive self - focusing two - waveguide coupler , @xmath105 , that has two constants of motion in the form of the stokes vector norm , @xmath106 such that @xmath107 , and the hamiltonian - like quantity , @xmath108 note that the system is integrable and the conservation of the stokes vector norm allows for the parametrization @xmath109 and @xmath110 , such that we can write a hamilton - jacobi model , @xmath111 equivalent to that of a nonrigid pendulum or a bose - josephson junction @xcite .
this particular configuration allows for rabi oscillations , fig .
[ fig : figure6](a ) , below the critical effective kerr nonlinearity to coupling ratio , @xmath112 , fig .
[ fig : figure6](b ) , above this critical value the system can show both rabi and josephson oscillations , fig .
[ fig : figure6](c ) . in other words , the initial field amplitudes
either coherently oscillate between the waveguides or localize at the waveguide where they were originally prepared , depending on both the initial field distribution and the effective kerr nonlinearity of the device .
, with nonlinearities ( a ) below , @xmath113 , ( b ) at , @xmath114 , and ( c ) above , @xmath115 , the critical kerr nonlinearity to coupling strength ratio , @xmath114 .
the figure shows ( a ) stationary point , ( b ) trajectory infinitesimally near the stationary point , and ( c ) separatrix in black , ( a)-(b ) rabi and ( c ) josephson oscillations in red , and ( a)-(c ) rabi oscillations in blue . ]
if we now include the symmetric gain and loss , @xmath116 , the system is still integrable as we can write two constants of motion @xcite , @xmath117^{2 } + \left [ \kappa \mathcal{s}_{y}\left ( \zeta \right ) \right]^{2 } } , \\
\mathcal{s}_{j}\left ( \zeta \right ) & = & s\left ( \zeta \right ) + \frac{2 \gamma}{\kappa } \arctan \left [ \frac { \kappa \mathcal{s}_{x } \left ( \zeta \right ) - 2 } { \kappa \mathcal{s}_{y}\left ( \zeta\right ) } \right].\end{aligned}\ ] ] in the literature , it has been found that the system is stable in the interval @xmath118 for effective nonlinearity @xmath95 @xcite and numerical arguments have been given in the most general case @xcite .
figure [ fig : figure7 ] shows how the dynamics of the passive nonlinear dimer are affected by the addition of a small effective gain to coupling ratio to the system . below the critical nonlinearity for the passive system , fig .
[ fig : figure7](a ) , we can still find the coherent oscillation behavior of the linear @xmath0-symmetric dimer but the former stable point is no longer a fixed point of the system . as the nonlinearity increases
, we can see how device parameters and initial conditions start having an effect on the dynamics , fig .
[ fig : figure7](b ) , until a point where it is possible to have unstable light localization at the waveguide where light originally impinged , fig .
[ fig : figure7](c ) .
here , the constant of motion plays an important role as an accuracy test for the process of numerically solving the coupled nonlinear system . and variable effective nonlinearity to coupling ratio ( a ) @xmath113 , ( b ) @xmath114 , and ( c ) @xmath115 , for the same initial conditions than fig .
[ fig : figure6 ] . ] now , for a given effective kerr nonlinearity to coupling ratio , @xmath95 , we can approximate a critical effective gain to coupling ratio , @xmath119 , where a change of dynamics occur .
first , the reciprocity condition regarding the exchange of waveguides for a given initial field distribution breaks as we get closer to the approximate critical gain to coupling ratio , note the shift to the left in fig .
[ fig : figure8](a ) and [ fig : figure8](b ) , then , after we cross the critical value , the field intensity at the gain waveguide gets localized and experiences an exponential gain , while the field in the lossy waveguide diminishes independently of the initial field amplitude distribution , fig .
[ fig : figure8](c ) .
note , the dynamics above the critical gain to coupling ratio do not tend to the fixed point of the system as in the linear device ; the fields do not seem to show a constant behavior in the asymptotic limit .
this localization with amplification in the gain waveguide suggests the use of these devices as optical diodes @xcite . and variable effective gain to coupling ratio ( a ) @xmath120 , ( b ) @xmath121 , and ( c ) @xmath122 , for the same initial conditions than fig .
[ fig : figure7 ] . ]
as we said before , the linear dimer is the device with the smallest dimension showing @xmath0-symmetry . its algebraic structure , eq.([eq : hamiltonian ] ) , and
the fact that pauli matrices are the two by two scaled matrix representation of the @xmath10 group in dimension two , @xmath123 , suggest that a larger class of @xmath1-waveguide devices can be constructed with higher dimensional matrix representations of @xmath10 .
therefore , we can construct a mode coupling equation set describing a device of @xmath1 waveguides following eq .
( [ eq : ptdimer ] ) , @xmath124 now , we choose to introduce a different scaled propagation , @xmath125 , to recover a single parameter schrdinger - like equation with the form , @xmath126 where the effective refractive index to coupling ratio is the same that in the standard dimer , @xmath127 .
this mode coupling matrix with @xmath128 , describes planar @xmath1-waveguide couplers with identical real effective refractive indices and underlying @xmath10 symmetry that show harmonic oscillator behavior @xcite .
these devices have been used to produce perfect state transfer in both the classical and quantum regimes of optical circuits @xcite . in order to deal with a linear @xmath0-symmetric @xmath1-waveguide coupler
, we must consider pure imaginary effective refractive indices , @xmath129 , and , thus , the realization of higher finite - dimensional non - unitary representation of @xmath11 , implemented as a complexified version of @xmath10 , are involved . in matrix form ,
the group generators have the following elements , @xmath130_{m , n } & = & \frac{1}{2 } \left[\delta_{m-1,n } \sqrt { n \left(2j - n+1\right ) } ~ + \delta_{m+1,n } \sqrt { m \left(2j - m+1\right ) } \right],\\ \left [ \mathbbm{j}_{y}\right]_{m , n } & = & \frac{i}{2 } \left[\delta_{m-1,n } \sqrt { n \left(2j - n+1\right ) } ~ - \delta_{m+1,n } \sqrt { m \left(2j - m+1\right ) } \right ] , \\
\left [ \mathbbm{j}_{z}\right]_{m , n } & = & \delta_{m , n } \left ( j - m+1 \right ) , \qquad \qquad \qquad m , n = 1,2,\ldots , n,\end{aligned}\ ] ] with the kronecker delta given by @xmath131 and the bargmann parameter by @xmath132 .
these matrices fulfill the commutation relation @xmath133 = i \epsilon_{ijk } \mathbbm{j}_{z}$ ] , where @xmath134 is the levi - civitta symbol , and commute with the casimir operator @xmath135 , @xmath136=0 $ ] . in the standard differential form , this is equivalent to the coupled mode set , @xmath137 following the vector notation , we can construct a field vector as , @xmath138 where we can define the @xmath139th element of the standard basis as @xmath140_{n } = \delta_{j - m+1,n}\ , , \qquad m= -j , -j+1 , \ldots , j-1 , j,\end{aligned}\ ] ] such that we can define more helpful generators with their corresponding actions , @xmath141 where we have defined the ladder operators @xmath142 that fulfill @xmath143 = \pm \mathbbm{j}_{\pm}$ ] . following the gilmore - perelomov approach for @xmath10 @xcite , we can find the @xmath139th eigenvalue of the mode coupling matrix , @xmath144 and obtain the same structure found for the dimer .
all the eigenvalues will be real numbers for @xmath145 , fig .
[ fig : figure9](a ) , completely degenerate and equal to zero @xmath146 for @xmath147 , fig .
[ fig : figure9](b ) , and imaginary for @xmath148 , fig .
[ fig : figure9](c ) .
this so - called collapse of the eigenvalues is a direct consequence of the underlying symmetry .
-symmetric regime , ( b ) fully degenerate regime , and ( c ) broken symmetry regime .
the black arrows show the direction of the eigenvalues as the gain to coupling ratio increases .
theses cases show the results for a @xmath149 waveguide coupler that provides a bargmann parameter @xmath150 . ] furthermore , we can provide the propagation matrix elements , @xmath151_{m , n}&= & \omega^{-2j } \sqrt { \left ( \begin{array}{c } 2j \\ m-1 \end{array } \right ) \left ( \begin{array}{c } 2j \\ n-1 \end{array}\right ) } ~ \nonumber \\ & & \left ( \omega \cos \frac{\omega}{2 } \xi - \gamma \sin \frac{\omega}{2 } \xi \right)^{2(j+1)-m - n } \left ( i \sin \frac{\omega}{2 } \xi \right)^{m+n-2 } \nonumber \\ & & ~_{2}f_{1 } \left ( 1-m,1-n,-2j , \omega^{2 } \csc^{2 } \frac{\omega}{2 } \xi \right ) , \qquad \gamma < 1 , \label{eq : progsu2a}\\ \left [ \mathbb{u}(\xi ) \right]_{m , n}&= & ( -2)^{-2j } ~ \sqrt { \left ( \begin{array}{c } 2j \\ m-1 \end{array } \right ) \left ( \begin{array}{c } 2j \\ n-1 \end{array}\right ) } ~ \left ( \xi - 2 \right)^{2(j+1)-m - n } ~ \nonumber \\ & & \left ( i \xi \right ) ^{m+n-2 } ~_{2}f_{1 } \left ( 1-m,1-n,-2j , \frac{4}{\xi^{2 } } \right ) , \qquad \gamma = 1 , \label{eq : progsu2b}\\ \left [ \mathbb{u}(\xi ) \right]_{m , n}&= & \left(\vert \omega\vert \right)^{-2j } \sqrt { \left ( \begin{array}{c } 2j \\ m-1 \end{array } \right ) \left ( \begin{array}{c } 2j \\ n-1 \end{array}\right ) } ~ \nonumber \\ & & \left ( \vert\omega\vert \cosh \frac{\vert\omega\vert}{2 } \xi - \gamma \sinh \frac{\vert\omega\vert}{2 } \xi \right)^{2(j+1)-m - n } \left ( i \sinh \frac{\vert\omega\vert}{2 } \xi \right)^{m+n-2 } \nonumber \\ & & ~_{2}f_{1 } \left ( 1-m,1-n,-2j , \vert \omega \vert ^{2 } \mathrm{csch}^{2 } \frac{\vert\omega\vert}{2 } \xi \right ) , \qquad \gamma > 1 , \label{eq : progsu2c}\end{aligned}\ ] ] where the notation @xmath152 and @xmath153 stand for binomial coeffcient and gauss hypergeometric function , in that order .
again , we will have three distinct propagation behaviors as demonstrated for the dimer .
these behaviors are simpler to visualize if we define renormalized field amplitudes , @xmath154 now , we can see periodical amplified oscillations in the @xmath0-symmetric regime , @xmath47 , fig .
[ fig : figure10](a ) , amplification following a power law in the fully degenerate regime , @xmath48 , fig .
[ fig : figure10](b ) , asymmetric amplification following an exponential law in the broken symmetry regime , @xmath49 , fig .
[ fig : figure10](c ) .
waveguide coupler , bargmann parameter @xmath150 , in the ( a ) @xmath0-symmetric regime , @xmath54 , ( b ) fully degenerate regime , @xmath55 , and ( c ) broken symmetry regime , @xmath56 , for an initial field impinging just the first waveguide . ] also , we can derive the set of reducible coupled nonlinear equations for the renormalized field amplitudes , @xmath155 \tilde{\mathcal{e}}_{k}\left ( \xi \right ) + \nonumber \\ & & + \frac{1}{2 } \sqrt { ( k-1)(n - k-1)}~ \tilde{\mathcal{e}}_{k-1}\left(\xi\right ) + \nonumber \\ & & \frac{1}{2 } \sqrt { k ( n - k ) } ~ \tilde{\mathcal{e}}_{k+1}\left(\xi\right ) , \quad k=1,2 , \ldots , n.\end{aligned}\ ] ] it is cumbersome but possible to show that the asymptotic response of a @xmath1-waveguide coupler in the fully degenerate and broken symmetry regimes is independent of the input field amplitude distribution , @xmath156 which can be seen in fig .
[ fig : figure10](b ) for @xmath48 and fig .
[ fig : figure10](c ) for @xmath148 . as expected from the mathematical description , in the fully degenerate regime , fig .
[ fig : figure10](b ) , the extremal waveguides , those with major effective losses and gain , will transmit fields with smaller amplitudes than those in the central waveguides , because the asymptotic intensity distribution follows the binomial coefficient . in the broken symmetry regime ,
[ fig : figure10](c ) , the field intensity correlates with the strength of the gain or loss ; the most intense field will travel through the waveguide with the larger effective gain , and the less intense through the one with the larger effective loss . the equivalent three - dimensional formulation of the stokes vector for renormalized fields , @xmath157 yields components of the following form , @xmath158 , \\
\tilde{\mathcal{j}}_{y}\left(\xi\right ) & = & \frac{i}{2 } \sum_{k=1}^{2j+1 } \sqrt { k(2j - k+1 ) } \left [ \tilde{\mathcal{e}}_{k+1}^{\ast } \tilde{\mathcal{e}}_{k } - \tilde{\mathcal{e}}_{k}^{\ast } \tilde{\mathcal{e}}_{k+1 } \right ] , , \\
\tilde{\mathcal{j}}_{z}\left(\xi\right ) & = & \sum_{k=1}^{2j+1 } \left(j - k+1 \right ) \vert \tilde{\mathcal{e}}_{k}\left ( \xi \right ) \vert^{2}.\end{aligned}\ ] ] here , the conserved variable is the casimir operator , @xmath159 \vert \tilde{\mathcal{e}}\left ( \xi \right ) \rangle = j(j+1),\end{aligned}\ ] ] and it is important to emphasize that the norm of this three - dimensional stokes vector is no longer a constant of motion , @xmath160 the reason behind this is that a complex vector of dimension @xmath1 with unit norm can be represented as a point on the surface of a ball of unit radius in dimension @xmath161 .
for example , the renormalized fields through a two - waveguide coupler , two - dimensional complex vector of unit norm , can be represented on the surface of a three - dimensional ball ; in other words , a two - dimensional sphere , where the stokes vector norm is a constant of motion , @xmath162 = \frac{1}{2 } , \qquad j=\frac{1}{2}.\end{aligned}\ ] ] in general , we should use the surface of a @xmath163-ball of unit radius in order to describe properly the field amplitudes propagating through a @xmath1-waveguide coupler .
this does not make it simpler to visualize the dynamics , so , we favor a projection from @xmath164-dimensional to three - dimensional space with the price of loosing the unit norm for all cases but @xmath165 , where we can write @xmath166 with @xmath167 .
note that this projection also allows us to derive an asymptotic expression for the @xmath19-component of the stokes vector via eq.([eq : asympnplanar ] ) , and heuristically propose the rest , @xmath168 this was confirmed numerically over a random sample of initial states and gain to coupling ratios outside the @xmath0-symmetric regime , @xmath169 $ ] and @xmath170 $ ] .
figure [ fig : figure11 ] shows the propagation of the renormalized stokes vector in a six - waveguide coupler , @xmath150 , with parameters in the @xmath0-symmetric , @xmath171 , fig .
[ fig : figure11](a ) , fully degenerate , @xmath48 , fig .
[ fig : figure11](b ) , and broken symmetry , @xmath172 , fig .
[ fig : figure11](c ) , regimes for light impinging the first waveguide of the coupler in black . also , the renormalized stokes vector propation for an initial field amplitude distribution corresponding to the eigenstate of @xmath173 with eigenvalue @xmath174 is plotted in red to show the asymptotic behavior outside the @xmath0-symmetric regime .
, in the ( a ) @xmath0-symmetric regime , @xmath54 , ( b ) fully degenerate regime , @xmath55 , and ( c ) broken symmetry regime , @xmath56 , for the initial conditions @xmath175 in black , and the eigenstate of @xmath173 with eigenvalue @xmath174 , @xmath176 , in red . ]
so far , we have studied propagation of classical light through a class of @xmath0-symmetric devices with underlying @xmath11 symmetry that includes the linear and nonlinear @xmath0-symmetric dimer . we have seen that propagation on these linear and nonlinear devices is more involved than in the passive case , @xmath177 , but the linear algebra approach has helped us produce closed form propagators for the lineal case and the renormalized fields approach and its stokes vector representation has allowed us to find stationary states and visualize propagation dynamics . here , we shall try to understand the underlying reasons for this more complex propagation behavior .
as we mentioned earlier , the propagation of light through tight - binding @xmath1-waveguide couplers can be modeled by mode coupling theory in a form similar to the schrdinger equation , eq .
( [ eq : schrodingereq ] ) .
thus , finding a propagator , eq .
( [ eq : propagatedfield ] ) , provides us with the information of the complex field amplitudes as they propagate through each waveguide . in quantum mechanics
, we can also ask about the propagation of the expectation value , which in the optical picture translates to the following , @xmath178 for the operator @xmath179 , which can vary with propagation .
this is exactly what has been done when studying the propagation of the stokes vector defined as the expectation value of the @xmath10 generators , eqs .
( [ stockesvector ] ) and ( [ nstockesvector ] ) , with the peculiarity that we used renormalized field amplitudes instead of just the field amplitudes in the linear cases .
ehrenfest theorem relates the variation with propagation of the mean value with the dynamics of the model , @xmath180 in standard hermitian quantum mechanics , heisenberg equation provides the equation of motion for the operator but in our non - hermitian model we need to go beyond this . any non - hermitian operator @xmath179 can be decomposed , @xmath181 in an hermitian , @xmath182 , and skew - hermitian , @xmath183 , parts .
thus , we can define a commutator between non - hermitian operators , @xmath184_{nh } = \mathbbm{a}(z ) \mathbbm{b}(z ) - \mathbbm{b}^{\dagger}(z ) \mathbbm{a}^{\dagger}(z),\end{aligned}\ ] ] such that we can write a heisenberg - like equation of motion ruled by a non - hermitian hamiltonian , @xmath185_{nh } + \frac{\partial}{\partial z } \mathbbm{a}(z),\end{aligned}\ ] ] and recover a non - hermitian generalization of ehrenfest theorem for tight binding non - hermitian waveguide couplers , @xmath186_{nh } \right\rangle + \left\langle \frac{\partial}{\partial z } \mathbbm{a}(z ) \right\rangle.\end{aligned}\ ] ] let us consider as example the @xmath1-waveguide coupler of last section , eq .
( [ eq : hsu2 ] ) , where the effective mode - coupling matrix acting as hamiltonian has as hermitian part , @xmath187 , and a skew - hermitian part , @xmath188 . here
the ehrenfest theorem can be simplified to the expression , @xmath189 \right\rangle + i \left\langle \left\ { \mathbbm{a}(\xi ) , \mathbbm{h}_{s } \right\ } \right\rangle + \left\langle \partial_{\xi } \mathbbm{a}(\xi ) \right\rangle,\end{aligned}\ ] ] where the standard commutator , @xmath190 = \mathbbm{a } \mathbbm{b } - \mathbbm{b } \mathbbm{a}$ ] , and anti - commutator , @xmath191 , have been used .
we can see that the anti - commutator term in this expression will be proportional to the gain to coupling strength ratio , @xmath74 , and this is the culprit behind the more complex behavior of our general class of @xmath1-waveguide lattices with underlying complexified @xmath10 symmetry .
note that we can also use this result to derive the conserved quantities of the model by solving
@xmath192 \right\rangle + i \left\langle \left\ { \mathbbm{a}(\xi ) , \mathbbm{h}_{s } \right\ } \right\rangle + \left\langle \partial_{\xi } \mathbbm{a}(\xi ) \right\rangle = 0 $ ] . as a practical example , let us derive the equations of motion for the stockes vector for the linear @xmath0-symmetric dimer , @xmath193 with @xmath194 , where @xmath195 is the identity matrix , given by eqs.([eq : svx])([eq : sv0 ] ) . here , the hermitian and skew - hermitian parts of the mode coupling matrix are @xmath196 and @xmath197 .
now , any linear hermitian operator for this system can be written as the linear superposition of the matrices @xmath198 , @xmath199 and the propagation of its expectation values , according to the non - hermitian ehrenfest theorem , are given by the following expression , @xmath200 + \sum_{k=0,x , y , z } \mathcal{s}_{k}(\zeta ) \partial_\zeta a_{k}(\zeta ) + \nonumber \\ & & - 2 \gamma \left[a_{0}(\zeta ) \mathcal{s}_{z}(\zeta ) + a_{z}(\zeta ) \mathcal{s}_{0}(\zeta ) \right]\end{aligned}\ ] ] thus , noting that for the stokes vectors the coefficients are constant , @xmath201 , we can write the evolution for the components of the stokes vector without field renormalization , @xmath202 , \end{aligned}\ ] ] which are in complete agreement with what we obtain from the nonlinear @xmath0-symmetric dimer , eqs .
( [ eq : svnl0])([eq : svnlz ] ) , if we kill the effective nonlinearity to coupling strength ratio , @xmath203 .
note that the total intensity , @xmath204 , is not conserved as expected from non - hermitian dynamics .
note that , in this case , the total intensity , @xmath204 , coincides with the norm of the stokes vector , @xmath205 , and we can recover the renormalized stokes vector dynamics , eqs .
( [ eq : rnsv0])([eq : rnsvz ] ) , if we define a renormalized stokes vector , @xmath206 , and use the equations of motion found here . if we were to find a constant of motion , @xmath207 , then its components should satisfy , @xmath208.\end{aligned}\ ] ] a particular solution to this set of equation is @xmath209 in agreement with eq.(105 ) . for the general case of the planar @xmath1-waveguide
coupler , the situation is far more complex as we are dealing with square matrices of dimension @xmath1 . in order to construct any given hermitian operator of this dimension
, we need a basis with total of @xmath13 matrices , these are provided by the standard unitary group of degree @xmath1 , @xmath210 , plus the identity .
this way , we will work with a set of @xmath13 operators where the first four elements are the representation of @xmath10 in dimension @xmath1 plus the unity , @xmath211 with @xmath194 , that form @xmath212 , @xmath213 where we have just implicitly make the change @xmath214 , @xmath215 , @xmath216 .
thus , if we define a generalized stokes vector for the planar @xmath1-waveguide coupler , it will have dimension @xmath13 and the zeroth component will be the total intensity in the system , @xmath217 but in this case the zeroth component of the generalized stokes vector , @xmath218 , is still the total intensity but does not coincide with the norm of the generalized stokes vector , @xmath219 .
thus , a graphical representation on the sphere will just be a projection of the propagation dynamics occurring on a @xmath164-dimensional hypersphere as mentioned before .
note that for passive devices , @xmath177 , the propagation equations for the stokes vectors do not involve any other functions and the propagation dynamics is restricted to the subgroup @xmath10 of @xmath212 , recovering the results of ref .
let us turn our attention now to the propagation of nonclassical light . in the quantum regime , it is possible to describe two - waveguide couplers with the following effective hamiltonian @xcite , @xmath220 where we have kept the notation for the effective refractive indices and evanescent coupling strength , @xmath16 with @xmath221 and g , in that order . at this point
, we can use schwinger two - boson representation of @xmath10 @xcite , @xmath222 to write an effective hamiltonian with underlying @xmath10 symmetry , @xmath223 that answers to the effective schrdinger equation , @xmath224 in the single photon regime , @xmath225 we recover the differential equation set describing the standard two - waveguide coupler , @xmath226 this approach suffices for the analysis of ideal dimers without gain nor loses , @xmath227 , where the total photon number , @xmath228 of the initial state , @xmath229 , determines the dimension of the @xmath10 representation to be used , @xmath230 , with the eigenbasis of @xmath231 given by the following , @xmath232 such that we can use the results proposed for classical waveguides couplers with underlying @xmath10 symmetry @xcite to calculate relevant quantities like mean photon number at each waveguide . the inclusion of linear loses and gain is not a trivial matter and it is simpler to discuss in the heisenberg picture @xcite , @xmath233 + \partial_{z } \hat{o}(z),\end{aligned}\ ] ] where we have accounted for the change from time to distance propagation
let us go straight to the @xmath0-symmetric dimer , with identical real part of the refractive index and moving into a rotating frame , @xmath234 such that , again , we can define a scaled propagation , @xmath34 , and include spontaneous processes arising from the quantum description of materials with linear loss or gain processes @xcite , @xmath235 where the first term in the right hand side is related to propagation through the quantum two - waveguide coupler with linear loss and gain , @xmath129 , and the second term describe gaussian random processes of emission and absorption , a result arising from the linear materials in an equivalent treatment to that used in the quantum description of the laser @xcite , @xmath236 the formal solution for this differential equation yields the propagation of the annihilation operators , @xmath237 where we have obviated the common phase factor @xmath238 that does not play any important role .
note that we can use the propagator we already found for the classical dimer , @xmath239 in eq .
( [ eq : propagator ] ) , for the first term in the right hand side .
now , in order to realize the effect of processes induced by the linear materials , let us focus on spontaneous generation in the absence of fields in both waveguides .
in the classical case , there will be no light at all propagating through the waveguides but , in the quantum case , even with an initial vacuum state we can calculate the spontaneous generation at each waveguide @xcite , @xmath240_{12 } \right\vert^{2 } dt , \\
s_{2 } & = & 2 \gamma \int_{0}^{\zeta } \left\vert \left[e^{i \mathbbm{h } t}\right]_{22 } \right\vert^{2 } dt.\end{aligned}\ ] ] it shows in the symmetric regime , @xmath145 , @xmath241 , \\
s_{2 } & = & \frac{1}{4 \omega^{3 } } \left [ 2 \omega \zeta + \left(1- 2\gamma^{2 } \right ) \sin\left(2 \omega \zeta \right ) + 4 \gamma \omega \sin^{2 } \left(2 \omega \zeta \right ) \right],\end{aligned}\ ] ] a linear increase with a periodic modulation , fig .
[ fig : figure12](a ) . in the fully degenerate case ,
@xmath48 , the spontaneous generation , @xmath242 follows a cubic polynomial , fig .
[ fig : figure12](b ) , and in the broken symmetry regime , @xmath49 , @xmath243 , \\ s_{2 } & = & \frac{1}{4 \vert \omega \vert^{3 } } \left[- 2 \vert \omega \vert \zeta - \left(1- 2\gamma^{2 } \right ) \sinh \left(2 \vert \omega \vert \zeta \right ) + 4 \gamma \vert \omega \vert \sinh^{2 } \left(2 \vert \omega \vert \zeta \right ) \right],\end{aligned}\ ] ] it shows exponential amplification , fig .
[ fig : figure12](c ) . further discussion regarding the effect of spontaneous processes on the propagation of diverse nonclassical fields through a linear @xmath0-symmetric dimer can be found in ref .
@xcite . solid blue line , and gain
, @xmath244 dashed red line , in the ( a ) @xmath0-symmetric regime , @xmath54 , ( b ) fully degenerate regime , @xmath55 , and ( c ) broken symmetry regime , @xmath56 , for quantum vacuum fields in both waveguides . ]
we have presented a review of the @xmath0-symmetric dimer in its linear , nonlinear and quantum versions and show that it belongs to a symmetry class with underlying @xmath11 symmetry , realized as a complexification of the @xmath10 group , that allows the description of @xmath1-waveguide couplers .
we have aimed to present a coherent narrative of the different approaches to the optical @xmath0-symmetric dimer and relate them to the underlying symmetry of the model . in doing this
, we introduce the idea of using a non - hermitian version of ehrenfest theorem to approach the propagation dynamics of waveguide couplers described by non - hermitian mode coupling matrices .
the field is young and there still exist fundamental open questions on the subject such as the analytic determination of critical effective nonlinearity to coupling ratios for the kerr nonlinear @xmath0-symmetric dimer ; the need of a deeper understanding of the non - unitary finite dimensional representations of @xmath11 , realized without resorting to the complexified @xmath10 representations ; the generalization to propagation dependent photonic systems together with its possible applications , just to mention a few that we hope to address in future work .
dhm acknowledges financial support from conacyt @xmath245294921 phd grant .
sla acknowledges financial support from conacyt @xmath245243284 grant .
10 o. i. zaragoza gutirrez , l. f. salinas mendoza , and b. m. rodrguez - lara .
all - optical @xmath9-symmetric conversion of amplitude ( phase ) modulation to phase ( amplitude ) modulation .
, 24:3989 3998 , 2015 .
a. guo , g. j. salamo , d. duchesne , r. morandotti , m. volatier - ravat , v. aimez , g. a. siviloglou , and d. n. christodoulides .
observation of @xmath0-symmetric breaking in complex optical potentials .
, 103:093902 , 2009 .
|
optics has proved a fertile ground for the experimental simulation of quantum mechanics .
most recently , optical realizations of @xmath0-symmetric quantum mechanics have been shown , both theoretically and experimentally , opening the door to international efforts aiming at the design of practical optical devices exploiting this symmetry . here ,
we focus on the optical @xmath0-symmetric dimer , a two - waveguide coupler were the materials show symmetric effective gain and loss , and provide a review of the linear and nonlinear optical realizations from a symmetry based point of view .
we go beyond a simple review of the literature and show that the dimer is just the smallest of a class of planar @xmath1-waveguide couplers that are the optical realization of lorentz group in 2 + 1 dimensions .
furthermore , we provide a formulation to describe light propagation through waveguide couplers described by non - hermitian mode coupling matrices based on a non - hermitian generalization of ehrenfest theorem .
|
clusters of galaxies from low to high redshift provide a unique opportunity to trace back the evolution of galaxies in dense environments .
detailed studies of elliptical galaxies in nearby and intermediate redshift clusters showed that their photometric properties are homogeneous and evolve only mildly with redshift at @xmath8 .
this is consistent with the so - called ` passive evolution ' , namely the change of the photometric properties of elliptical galaxies by the aging of their stars only , with a formation epoch of at least @xmath9 gyr ago ( bower , lucey & ellis 1992 ; aragn - salamanca et al .
1993 ; ellis et al . 1997
; dickinson 1997b ; stanford , eisenhardt , & dickinson 1997 ) . on the other hand , some s0 and spiral galaxies seem to experience faster evolution ( e.g. , dressler et al .
1994 ; dressler and smail 1997 ) .
it is clear that pushing these studies to higher redshifts should provide stronger constraints on the epoch of cluster galaxy formation .
unfortunately , only a very limited number of clusters and cluster candidates are known beyond @xmath10 ( dickinson 1995 ; 1997b ) , and thus the identification of high - redshift clusters itself should be the next important step .
near - infrared ( nir ) observations are essential in identifying clusters of galaxies at z @xmath11 .
it is known that old passively - evolving galaxies observed at high redshifts have very red optical - nir colors ( aragn - salamanca et al . 1993 ; dickinson 1995 ; yamada and arimoto 1995 ) , since their optical flux rapidly decreases as the 4000@xmath12 break shifts toward longer wavelengths .
for example , a 3@xmath12gyr - old passively - evolving model galaxy observed at @xmath13 has @xmath14 which is redder than most galactic stars and most field galaxies .
therefore by selecting objects with red optical - nir colors , the contrast between cluster and field galaxies will be very much enhanced if clusters are dominated by old galaxies .
it has been known for some time that radio - loud agns ( both radio - galaxies and quasars ) are frequently located in cluster environments ( see , e.g. , yee & green 1987 ; ellingson , yee & green 1991 ; dickinson 1995 ) .
thus , targeted searches in the fields of high redshift radio - loud agn provide an efficient approach to the problem of finding the ( probably ) rare galaxy clusters at @xmath15 ( see dickinson 1997c for a review ) . in this _
letter _ we report the clustering of very red objects in the vicinity of a quasar at @xmath16 where some excess in the surface number density of galaxies had already been reported by hutchings et al .
( 1993 ) based on their optical data .
we obtained deeper @xmath17 and @xmath18 optical images and nir @xmath19-band data of the region and confirmed the clustering of galaxies near the quasar ; their optical - nir colors and magnitudes are consistent with bright cluster galaxies formed at higher redshift and observed at @xmath16 .
the cosmological parameters @xmath20km@xmath12s@xmath21mpc@xmath22 and @xmath23 have been assumed .
with the aim of identifying high redshift cluster galaxies with red optical - nir colors , we have targeted the field around the radio - loud mjy exists at ( @xmath24 ( gregory and condon 1991 ) , while the quasar locates at @xmath25 ( vron - cetty and vron 1996 ) .
we estimate that the logarithmic radio - optical luminosity ratio of the quasar is @xmath26 .
thus the 1335.8 + 2834 should be classified as radio - loud . in the recent vla first survey catalog ( white et al .
1997 ) , where the typical positional error is less than @xmath27arcsec , there is a radio source at @xmath28 . ]
quasar 1335.8@xmath02834 at @xmath29 .
hutchings , crampton & persam ( 1993 ) found that the surface density of galaxies within @xmath30@xmath31 separation ( @xmath32mpc at @xmath16 ) from the quasar is three times larger than the average density of their control fields . using narrow - band images centered on the [ oii]@xmath33 line redshifted to @xmath1 , they also found several emission - line galaxies near the quasar , suggesting the excess density is associated with the quasar redshift .
optical images were obtained with the isaac newton telescope at la palma on the 28th of february , 1995 .
a tek @xmath34 ccd with a projected pixel size of @xmath35arcsec was used .
the weather conditions were mostly photometric .
total exposures of @xmath36sec in the @xmath17 band and @xmath37sec in the @xmath18 band were gathered , subdivided in @xmath38 or @xmath39sec sub - exposures .
the fwhm sizes of the stars were typically 1.5 ( @xmath17 ) and 1.8 ( @xmath18 ) arcsec .
the frames were bias - subtracted , flat - fielded and median - stacked in the usual manner .
the @xmath19-band images were obtained with the university of hawaii 2.2 m telescope equipped with the quirc camera on the 29th of february , 1996 .
the detector was the @xmath34 hgcdte array with a projected pixel size of @xmath40arcsec .
since the field of view of this camera ( @xmath41arcmin@xmath42 ) is smaller than that of the optical ccd ( @xmath43arcmin@xmath42 ) , we covered only the field which contains the region where the surface density had been evaluated to be high from the optical data . in this letter , we concentrate mainly on the properties of the objects contained within the @xmath19-band field .
the net exposure time of the @xmath19-band image is @xmath44sec , subdivided in many @xmath45sec dis - registered exposures .
the fwhm size of the stellar images was @xmath46arcsec .
flat - field frames were built using a running median of @xmath47 images in a time sequence .
the flat - fielded images were then median - stacked and a constant sky value was subtracted .
the data reduction was done with iraf .
object detection was carried out using focas ( jarvis and tyson 1981 ) with a detection threshold of @xmath48 of the sky fluctuations per pixel .
a minimum of @xmath49 connecting pixels above the threshold was required for detection .
landolt ( 1992 ) standards in the @xmath17 and @xmath18 bands , and ukirt bright standards in @xmath19 were used for photometric calibration , yielding zero - point errors smaller than @xmath50mag .
we measured the colors of the objects within a fixed aperture with a diameter of @xmath51arcsec , twice the fwhm seeing in the @xmath18 band image ( the worse case ) , after correcting the seeing differences among the images .
the apphot task in the iraf was used for the photometry .
total magnitudes were also estimated using template growth curves determined from 10 isolated galaxies in the field .
the source detection is complete for objects brighter than 24.5 , 23.5 , and 20.0 magnitudes in @xmath17 , @xmath18 , and @xmath19 respectively .
these limits were estimated from the turn - off of the number counts in the ` field region ' ( western half of the @xmath19-band image ) , which avoids the region of the cluster candidate ( see ahead ) .
the counts are consistent , within @xmath52 , with published field number counts ( e.g. , smail et al .
1995 ; djorgovski et al .
the internal photometric errors in the aperture magnitudes were evaluated using the photon count statistics and also by changing the sky regions in the local sky subtraction procedure .
typical errors are @xmath53 mag in @xmath17 and @xmath54 mag in @xmath18 and @xmath19 at the completeness limits . the error in the @xmath19-band total magnitude ( used in the color - magnitude diagram below )
is estimated to be @xmath55 mag at @xmath56mag , where the uncertainties in our growth - curve - fitting procedure have been taken into account .
figure 1 shows the three - color image of the @xmath19-band field , and figure 2a shows the distribution of the objects detected in the @xmath19-band image . objects with different colors are shown as different symbols .
an excess in the number density of faint objects near the quasar ( denoted by ` qso ' ) is seen .
hutchings et al .
( 1993 ) divided the image into cells and found a @xmath57 surface density excess in the number of objects in a @xmath58arcsec@xmath42 region centered on the quasar .
we made a similar analysis on our @xmath17 and @xmath18 optical images and found the excess surface density in a similar region to be @xmath59 above the counts of the surrounding field .
moreover , red objects are seen much more frequently in the eastern ( left ) half of the image near the quasar .
the color of most objects in the western side of the field is @xmath60 ( see below ) .
it is clear that objects with red optical - nir colors are responsible for most of the excess in the number density . since stars earlier than m3v have colors @xmath60 ( johnson 1966 ; bessell 1990 ) , most of these red objects are probably galaxies ( see figure 4 ) .
an un - evolved elliptical galaxy ( the reddest case for a galaxy without extinction ) has @xmath61 at @xmath62 , thus these red objects are probably galaxies at higher redshifts . a fairly bright galaxy near the quasar ( g1 in figures 1 and 2 ) has @xmath63 , @xmath64 , and @xmath65 , which agree very well with the expectations for a brightest cluster galaxy at @xmath66 ( aragn - salamanca et al 1993 )
. a closer look at the spatial distribution of the overdensity indicates that the contrast is maximised when g1 is taken as the center , instead of the quasar .
figure 2b shows the surface - density - contrast profile , centered on g1 , of all the objects detected in the @xmath19 band as well as those with @xmath67 .
the excess surface density of @xmath19-band detected objects in a 100 arcsec box centered on g1 is @xmath68 above the counts of the surrounding field , which is significantly above that found in the optical images .
moreover , if we consider only the red objects , the contrast of the putative cluster is significantly enhanced to @xmath69 above the field counts , which clearly indicates that the use of optical - nir colors is a powerful technique to identify high - redshift clusters . -0.02 cm the size of the region with the excess density is @xmath70 arcmin in diameter , corresponding to @xmath71mpc at @xmath16 , which is consistent with that expected for a rich galaxy cluster .
if we count all the galaxies with @xmath72 as cluster members , nearly 30 galaxies remain after correcting for field contamination in this color range .
this roughly corresponds to the abell richness class 0 .
this could be a lower - limit for the cluster richness since we have only counted the red objects as cluster members .
a more accurate estimate of the richness is not possible without galaxy redshifts .
figure 3 shows the @xmath73 vs @xmath19 color - magnitude diagram for the objects in the @xmath19-band field .
we tentatively define the ` cluster region ' as the eastern half of the field , and show the objects in this area with filled circles .
the color distributions in the two regions are very different .
there are 37 galaxies with @xmath74 and @xmath75 in the eastern ` cluster region ' but only 6 such objects in the western ` field ' region . in the cluster region
, there is a group of objects with similar colors and magnitudes , @xmath76 and @xmath77@xmath78mag .
they are consistent with a homogeneous population of galaxies seen at a similar redshift , which suggests that the clustering is probably real .
these objects have @xmath79 .
the @xmath80 color is very redshift - sensitive at @xmath81 due to the fact that the @xmath82 break is located near or just inside the @xmath18 band at this redshift .
thus , the red @xmath80 colors strongly suggest that these objects are old galaxies at @xmath83 . for comparison , on figure 3 we also plot the model colors and magnitudes of passively - evolving elliptical galaxies calculated with the models introduced by kodama and arimoto ( 1997 ) .
the model parameters are chosen so that the color - magnitude relation at @xmath4 ( interpreted as a mass - metallicity sequence ) agrees with that of the coma cluster ( bower et al .
the color and luminosity evolution predicted by such models agrees very well with the observed one for cluster ellipticals ( see kodama and arimoto 1997 ) .
the solid lines indicate the expected colors of passively - evolving model galaxies with @xmath84 and @xmath85 mag at @xmath4 , which are formed at @xmath86 ( @xmath87gyr at @xmath4 ) . tilted straight lines indicate the expected color - magnitude relation at different redshifts ( see caption ) .
the colors of the reddest galaxies in the complete sample are consistent with passively - evolving old galaxies observed at @xmath5 .
they define a ` red envelope ' as expected if they are the oldest cluster galaxies at this redshift , although a clear color - magnitude sequence is not seen .
-0.02 cm the color - magnitude diagram also reveals a significant fraction of objects with @xmath88 and @xmath89 . since their color distribution seems somewhat separated from the redder galaxies , it is possible that these objects belong to a different system at lower redshift ( @xmath90 ) .
however , the sky distribution of these moderately red objects is similar to that of the reddest objects ( cf .
figure 2a ) , so it is possible that they are in the same system . without redshifts , these to possibilities can not be discriminated .
figure 4 shows the two - color diagram for the ` cluster ' ( west half ) and ` field ' ( east half ) galaxies .
the solid line shows the colors of elliptical galaxies observed at various redshifts derived from the same models as in figure 3 , representing the expected locus of the reddest galaxies observed at any redshift in the absence of reddening .
it is clear that the reddest galaxies in both @xmath73 and @xmath91 are compatible with being passively - evolving cluster ellipticals at @xmath5 . on the other hand
, there is a number of galaxies with moderately red @xmath73 color ( @xmath92 ) and bluer @xmath91 colors .
hutchings et al .
( 1993 ) had also noticed the bimodal distribution of @xmath91 colors in the excess - density region , which is not seen in their control field .
if they are cluster members , their bluer optical colors would suggest a certain amount of current or recent star forming activity .
we have found significant clustering of galaxies with very red optical - nir colors near the radio - loud quasar 1335.8@xmath02834 at @xmath5 .
the reddest objects ( @xmath93 ) are responsible for nearly half of the surface density excess , and their colors and magnitudes are consistent with those of passively - evolving old galaxies seen at @xmath94 . if they are at the quasar redshift , galaxy evolution models suggest that in the absence of reddening the reddest objects are already @xmath95@xmath96gyr old . the range of colors displayed by the galaxies can be readily explained by the presence of some field galaxies at a variety of redshifts and by cluster galaxies with a variety of star formation histories such as disk galaxies formed a few gyr before @xmath5 , or old galaxies with young bursts of star formation .
redshifts and morphologies are needed to study the evolutionary properties of the galaxies in detail .
other examples of clustering of red galaxies near a radio - loud object at comparable redshifts have been found by dickinson and his collaborators ( dickinson 1995 ; 1997a ) .
they observed the fields of powerful radio galaxies at intermediate and high redshifts with a similar optical and nir imaging strategy .
the cluster around 3c 324 is their most outstanding case .
redshifts for a few hundred objects confirmed a cluster at @xmath97 associated with 3c 324 , although many interlopers in another group or cluster at @xmath98 were also found .
a similar contamination by a group or cluster at a slightly different redshift can not be excluded in the case of 1335.8@xmath02834 without extensive redshift data .
hintzen , romanishin & valdes ( 1991 ) have also found significant clustering of galaxies around @xmath99@xmath100 radio - loud quasars .
the galaxies responsible for the excess they detected seemed to be more luminous in the observed @xmath17-band than present - day brightest cluster galaxies .
however , the galaxies responsible for the excess we have found in the field of 1335.8@xmath02834 do not seem to be over - luminous in the @xmath19-band . since the observed @xmath17-band samples ultraviolet light at these redshifts , evolutionary effects such as the presence of moderate amounts of young stars can boost up the observed flux considerably , while the nir should show significantly milder evolutionary effects .
excess in the galaxy number density has also been found for higher - redshift quasars or radio galaxies ( dressler et al .
1993 , 1994 ; hutchings 1995 ; pascarelle et al .
these authors claim that the excess galaxies are small compact objects which may be the ` building blocks ' of today s luminous galaxies .
in contrast , many of the galaxies found near 1335.8@xmath02834 are fairly luminous ( @xmath101 ) .
similarly , an excess of @xmath102 galaxies was found near radio - loud quasars at @xmath103 by aragn - salamanca , ellis , and obrien ( 1996 ) .
the situation is therefore not clear , and a systematic study of the fields of a large well - defined sample of known high redshift objects is clearly needed . in summary
, we have shown that the combination of deep optical and nir imaging is a very powerful tool in revealing high redshift clusters of galaxies and/or their progenitors .
such clusters can provide samples of high redshift galaxies for evolutionary studies at large look - back times , which can greatly contribute to our knowledge of galaxy formation and evolution at very early epochs . while this kind of study is very successful at finding candidate clusters , redshifts and morphologies will be ultimately needed .
this work was partially supported by a grant - in - aid for scientific research of the japanese ministry of education , science , sports and culture ( no.07041104 and no.08740181 ) .
part of this work was also supported by the foundation for the promotion of astronomy of japan .
ty was a special post - doctoral researcher of the institute of physical and chemical research ( riken ) and a part of this work was supported by this institute .
aas acknowledges generous financial support from the royal society and the particle physics and astronomy research council .
tk thanks to the japan society for the promotion of science postdoctoral fellowships for research abroad .
ko was a visiting astronomer of the institute for astronomy , university of hawaii and thanks their hospitality during his stay .
aragn - salamanca , a. , ellis , r. s. , couch , w. j. , & carter , d. 1993 , , 262 , 764 aragn - salamanca , a. , ellis , r.s . &
obrien , k.s .
1996 , mnras , 281 , 945 bessell , m. s. 1990 , , 102 , 1181 bower , r. g. , lucey , j. r. , & ellis , r. s. 1992 , , 254 , 601 dickinson , m. 1995 , in asp conf .
vol.86 fresh view of elliptical galaxies , eds .
a. buzzoni , a. renzini , & a. serrano , p.283 dickinson , m. 1997a , in hst and the high redshift universe , ed .
n. tanvir , a. aragn - salamanca , and j.v .
wall , ( singapore : world scientific ) , p.207 dickinson , m. 1997b , in galaxy scaling relations : origins , evolution and applications , ed
. l. da costa ( springer - verlag ) , in press , _ astro - ph/9703035 _ dickinson , m. 1997c , in the early universe with the vlt , ed .
j. bergeron , ( springer - verlag ) , p.274 djorgovski , s. g. , et al . 1995 , , 438 , l13 dressler , a. , & smail , i. 1997 , in hst and the high redshift universe , ed .
n. tanvir , a. aragn - salamanca , and j.v .
wall , ( singapore : world scientific ) , p.185 dressler , a. , oemler , a. , gunn , j. , & butcher , h. 1993 , , 404 , l45 dressler , a. , oemler , a. , butcher , h. , & gunn , j. 1994 , , 435 , l23 ellingson , e. , & yee , h. k. c. 1991 , , 371 , 49 ellingson , e. , yee , h.k.c .
& green , r.f .
1991 , , 371 , 36 ellis , r. s. , smail , i. , dressler , a. , couch , w. j. , oemler , a. , butcher , h. , & sharples , r. m. 1997 , apj , 483 , 582 gregory , p.c .
, & condon , j.j .
1991 , , 75 , 1011 hintzen p. , romanishin w. , & valdes , f. 1991 , , 366 , 7 hutchings , j. b. 1995 , , 110 , 994 hutchings , j. b. , crampton , d. , & persram , d. 1993 , , 106 , 1324 hutchings , j. b. , crampton , d. , & johnson , a. 1995 , , 109 , 73 jarvis , j. f. , & tyson , j. a. 1981 , , 86 , 476 johnson , h.l . 1966 , , 4 , 193 kodama , t. , & arimoto , n .
1997 , , 320 , 41 landolt , a. u. 1992 , , 104 , 340 pascarelle , s. m. , windhorst , r. a. , keel , w. c. , & odewahn , s. c. 1996 , nature , 383 , 45 smail , i. , hogg , d. w. , yan , l. , & cohen , j. g. 1995 , , 449 , l105 stanford , s. a. , eisenhardt , p. r. m. , & dickinson , m. 1995 , , 450 , 512 stanford , s.a . ,
eisenhardt , p.r.m . , and
dickinson , m. 1997 , apj , submitted vron - cetty , m .-
& vron , p. 1996
, , 115 , 97 white , r. l. , becker , r. h. , helfand , d. j. , gregg , m. d. 1997 , , 475 , 479 yamada t. , & arimoto , n. 1995 , in iau symp .
no.171 , new light on galaxy evolution , eds .
r. bender , & r. l. davies , p.472 yee , h. k. c. & green , r. f. 1987 , apj , 319 , 28
|
we have obtained new deep optical and near - infrared images of the field of the _ radio - loud _
quasar 1335.8@xmath02834 at @xmath1 where an excess in the surface number density of galaxies was reported by hutchings et al .
[ aj , 106 , 1324 ] from optical data .
we found a significant clustering of objects with very red optical - near infrared colors , @xmath2 and @xmath3 near the quasar .
the colors and magnitudes of the reddest objects are consistent with those of old ( 12 gyr old at @xmath4 ) passively - evolving elliptical galaxies seen at @xmath5 , clearly defining a ` red envelope ' like that found in galaxy clusters at similar or lower redshifts .
this evidence strongly suggests that the quasar resides in a moderately - rich cluster of galaxies ( richness - class @xmath6 ) .
there is also a relatively large fraction of objects with moderately red colors ( @xmath7 ) which have a distribution on the sky similar to that of the reddest objects .
they may be interpreted as cluster galaxies with some recent or on - going star formation .
|
the building blocks of mathematical morphogenesis were put several decades ago in the seminal works of thompson @xcite and turing @xcite . a particularly relevant problem in this context is the examination of the properties relating to the architecture of cell colonies , as already noted and investigated by eden @xcite . at some basic underlying level , one could say that turing and eden shared the goal of understanding how a macroscopic structure , in particular one breaking the initial homogeneity , could arise out of a multiplicity of simple interactions . in the introduction of @xcite
one can read turing thoughts on how a structure composed by randomly and isotropically proliferating cells could or could not break the initial spherical symmetry .
but then his approach moved to the use of reaction diffusion equations and to the search of a deterministic mechanism , the currently well known turing instability , able to give rise to pattern formation . on the other hand ,
eden focused his interest on the evolution of a pure growth process .
he concentrated on a probabilistic abstraction of a developing cell colony and studied stochastic symmetry breaking concomitant to growth .
in particular , eden studied the architecture of a lattice cell colony to which new cells were added following certain probabilistic rules .
the objective was determining the asymptotic colony profile .
if the growth rules are isotropic then the realizations of the growth process are either spherically symmetric or this spherical symmetry is broken by means of long range fluctuations .
these sorts of discrete models are usually studied using continuum equations for ease of analytical treatment @xcite .
while most of these developments have been carried out for systems in which there is an external input of mass , the universality of the surface fluctuations is expected to remain unchanged when the origin of mass appearance is internal .
in fact , the eden model in a strip geometry has been satisfactorily analyzed within the classical theoretical framework @xcite . and ,
as we already have mentioned , this model was introduced as a probabilistic abstraction of a proliferating cell colony .
so we expect our results to hold independently of the internal or external origin of mass input into the system
. this will be so at least in those cases in which the growing radial interface can be described as a monge patch in spherical coordinates .
growing clusters as the ones generated by diffusion limited aggregation processes yield more complex structures that can not be described in such simple terms , and thus fall beyond the scope of the present approach . indeed , the original eden problem can be greatly generalized by means of the use of stochastic partial differential equations .
they allow a systematic study of the properties of the colony periphery , particularly of the interface fluctuations . at the same time , they allow us to get rid of the undesirable lattice anisotropy . in this work
we will concentrate on the properties of the realizations of the stochastic growth process .
our goal is unveiling under which conditions the developing radial cluster asymptotically _ weakly _ converges to the concentrically propagating spherically symmetric profile .
let us emphasize that we are interested in determining the presence or absence of spherical symmetry in every realization of the stochastic growth process in the long time limit .
as we are considering isotropic growth which is free from deterministic instabilities , averaging the cluster profile over many realizations yields an immediate spherical symmetry .
the properties of the realizations can be ascertained , as we will see , calculating suitable correlations adapted from classical elements of stochastic growth theory @xcite , and keeping in mind that we should expect weak rather than strong convergence of the asymptotic profiles .
weak convergence is a mathematically totally precise concept which physically corresponds to the convergence resulting from the self - averaging of wildly oscillating quantities or structures @xcite .
apart from the overall shape , the microscopic fluctuations of growing radial clusters have been studied as well .
it turns out that these fluctuations sculpture a fractal surface which is statistically self - affine @xcite .
this microscopic roughness , in those cases in which the pointwise width @xcite is well defined ( which actually rules out all flat interfaces due to the microscopic properties of white noise ) is described by the familiar scaling of planar systems @xcite .
this is not so , however , for long range fluctuations @xcite .
long range radial correlations can not be in general deduced from planar scalings .
they still adopt a self - affine form for macroscopic length scales , but scale invariance is this time achieved for different values of the critical exponents .
the following sections will discuss the numerical values and physical meaning of these exponents .
we will see that they are able to characterize which regimes asymptotically weakly converge to the radially symmetric shape and which do not .
in planar situations of non - equilibrium growth the interface is parameterized by the height function @xmath0 , which indicates the displacement of the interface with respect to some hyperplane taken as origin , at some @xmath1dimensional spatial point @xmath2 and for some instant of time @xmath3 @xcite .
we illustrate this geometry in fig .
( [ rdfig ] ) , where we present the one - dimensional solution to the so - called random deposition equation @xmath4 where @xmath5 is a zero - mean gaussian noise whose correlation is @xmath6 .
this numerical approximation is found by means of spatially discretizing the system . ) at several times .
the values of the parameters are @xmath7 , @xmath8 , and the system has been discretized in @xmath9 spatial points .
we have used as initial condition @xmath10 . ] in these systems and for large spatiotemporal scales the height - height correlation function adopts the self - affine scaling form @xmath11 which is invariant to the transformation @xmath12 , @xmath13 and @xmath14 , for @xmath15 and @xmath16 , @xmath17 and @xmath18 , where the subscript @xmath19 denotes the classical value of an exponent for a given model .
@xmath20 is known as roughness exponent , @xmath21 as growth exponent and @xmath22 as dynamic exponent , @xmath23 and @xmath24 are scaling functions .
this scaling is valid for a large class of linear and nonlinear models .
if we consider the linear langevin equation @xmath25 where @xmath5 , as before , is a zero - mean gaussian noise whose correlation is @xmath6 and @xmath26 is a pseudo - differential operator to be understood in the fourier transform sense , the exponents read @xmath27 , @xmath28 and @xmath29 .
one would like to know if the radial counterpart of eq .
( [ langevin ] ) displays an analogous behavior to that dictated by scaling ( [ fv ] ) . in the radial extension of the theory one has a new degree of freedom which was trivial in the planar case : the rate of growth .
if we assume that the average radius ( measured as distance from the origin ) of the macroscopic radial form grows as a power law @xmath30 for @xmath31 and a growth index @xmath32 , then the equation for the radial surface fluctuation @xmath33 is @xmath34 according to the reparametrization invariance principle @xcite , and where @xmath35 denotes the set of angles parameterizing the @xmath1dimensional radial interface and @xmath36 is the jacobian determinant of the change of variables from @xmath37 to @xmath38 evaluated at @xmath39 @xcite .
the noise is simply @xmath40 . in order to visually compare planar and radial processes
we have plotted the field @xmath41 for @xmath42 in fig .
( [ rrdfig ] ) . neglecting the diffusion
constant makes this stochastic process the radial counterpart of the random deposition process plotted in fig .
( [ rdfig ] )
. we will therefore refer to it as the _ radial random deposition process_. , @xmath43 , @xmath44 , and the system has been discretized in @xmath45 spatial points .
we have used as initial condition @xmath46 ( note that we have used as initial time @xmath47 ) . ] as we have mentioned in the introduction , one of the motivations of this work is studying the large scale , long time properties of a randomly proliferating set of cells or entities in general .
such a process can be modelled in some detail using a master equation , which can be mapped into a stochastic partial differential equation provided with a multiplicative noise term @xcite . on the other hand ,
we focus in this work on stochastic partial differential equations provided with additive noise .
while this could seem as an apparent paradox , the underlying reason is that equations ( [ langevin ] ) and ( [ radeq ] ) are effective models .
so they are intended to describe just the long range properties of the growing interface , and in this limit fluctuations can be effectively considered additive . a nice example of this , a case in which a front driven by multiplicative noise is found to belong to a universality class described by an additive noise stochastic equation , is described in @xcite . the analysis of the exact solutions to eq .
( [ radeq ] ) for @xmath48 and long times shows the diffusion term is irrelevant in the large scale @xmath49 and the correlation becomes @xcite @xmath50 where @xmath51 . in terms of arc - length differences @xmath52
one finds @xcite @xmath53 this correlation is invariant to the transformation @xmath54 , @xmath55 and @xmath56 for @xmath57 and @xmath58 ( @xmath59 ) when @xmath60 ( @xmath61 ) .
we define @xmath62 ( @xmath63 ) when @xmath60 ( @xmath61 ) .
first , the functional form of this correlation is not reducible to scaling form ( [ fv ] ) , and second , the exponents are totally different ( this makes precise the analysis carried out in @xcite ) . in summary , long range
radial surface fluctuations are not describable in terms of scaling ( [ fv ] ) .
it is clarifying to consider a simplified problem : an abstraction in which eq .
( [ langevin ] ) is placed in a growing domain @xcite .
if this is realized by means of the transformation @xmath64 one finds @xmath65 where the mass source is explicitly included and we assume fast growth @xmath48 and the rough interface inequality @xmath66 .
in this case it can be shown that the correlation obeys scaling ( [ fv ] ) for short length scales @xmath67 and for long length scales @xmath68 it is @xcite @xmath69 where @xmath70 is the counterpart of the arc - length coordinate .
this correlation is scale invariant for the exponents @xmath71 but not for the exponents @xmath72 . this way of realizing domain growth agrees with reparametrization invariance @xcite , but is not unique .
it makes space and mass grow simultaneously @xcite , while it is possible to realize domain growth at constant mass by introducing a dilution term @xcite @xmath73 in this case the correlation is @xmath74 which is scale invariant for both @xmath72 and @xmath75 .
the first invariance can be used to argue that this correlation is a direct consequence of scaling ( [ fv ] ) , but the scaling form ( [ rd ] ) is definitely ambiguous in view of the second invariance .
both sort of dynamics , with and without dilution , are physically realizable ; while dilution dynamics is apparently better suited to describe some natural processes , no - dilution dynamics seems to agree better with the numerical simulation of some discrete models as discussed in @xcite .
together with the difference in the roughness exponent , very pronounced differences in the auto - correlation and persistence exponents separate the interface behavior in presence and absence of dilution @xcite .
this reveals that the non - uniqueness of the extension of a langevin equation to a growing domain implies in turn a universality class bifurcation .
one can apply what has been learned in the previous section about growing domains to the radial situation . including dilution in the radial equation for
interface motion yields a different dynamics from the one established by reparametrization invariance .
the correlation in this case becomes @xcite @xmath76 which is scale invariant for the exponents @xmath77 but not for @xmath72 .
the difference among @xmath78 and @xmath79 originates in the memory effects that affect the surface described by reparametrization invariance and disappear by virtue of dilution @xcite .
the difference between the sets @xmath71 and @xmath77 , and the classical set @xmath72 is of a different nature .
note that the langevin dynamics on a growing domain and influenced by dilution gave rise to the random deposition correlation ( [ rd ] ) : @xmath80 this correlation is invariant to the transformation @xmath81 for @xmath82 let us remark that scale invariance is achieved for any real number @xmath83 , thus ( [ scale1 ] ) and ( [ scale2 ] ) actually constitute a one - parameter family of scale transformations .
this is , there is one degree of freedom in the critical exponents , or in other words , the scaling form is insensitive to the dynamic exponent .
contrarily , in a radial geometry the correlation , eq . ( [ radcorr ] ) , is scale invariant only if @xmath84 : this is the constraint imposed by the geometrical factor @xmath85 . the jacobian determinant @xmath86 for a @xmath1dimensional interface , where @xmath87 is the set of polar angles , while it is of course independent of the azimuth angle @xmath88 .
so this constraint appears for @xmath89 , but it is not present for @xmath90 .
the reason is that the one - dimensional radial interface is topologically identical to the one - dimensional , topologically toroidal , interface which results from assuming periodic boundary conditions on the growing domain setting . in this case , the angle - like coordinate @xmath2 , to which the azimuth angle is akin , may be either dimensional or dimensionless ( this could be expressed by means of the dichotomy of lagrangian / eulerian coordinates similarly to the developments in @xcite ) . on the other hand , the polar angles , ranging only @xmath91 radians , can not be associated to periodic coordinates which could have dimensions .
this fixes the scaling @xmath92 .
one can use the ambiguity of eq .
( [ rd2 ] ) to claim that @xmath93 is a good dynamic exponent in this case , but that would be accepting that topology can modify the scaling .
also , the value of the auto - correlation exponent in this case suggests that @xmath92 is the good dynamic exponent @xcite . in any case
, it is now clear that there is no basis for claiming that correlation ( [ rd])-([rd2 ] ) is a consequence of scaling ( [ fv ] ) in the present framework .
this is because , although both forms are invariant to the transformation ( [ scale1 ] ) for @xmath93 ( for linear equations like ( [ langevin ] ) ) , self - similar form ( [ rd])-([rd2 ] ) preserves scale invariance independently of the value of @xmath22 ( as specified in eq .
( [ scale2 ] ) ) .
so this second form is much more general and the value @xmath93 plays no special role in it ( let us note again that the value @xmath92 plays a special role in the sense that it appears in the auto - correlation function @xcite ) . despite this ambiguity in the growing domain situation ,
it is clear that the radial correlation ( [ radcorr ] ) is just invariant for @xmath77 , and the classical exponents @xmath72 play no role in this case .
this can be seen by means of a direct application of transformation ( [ scale1 ] ) to this correlation and substituting @xmath20 and @xmath22 for the two different alternatives .
our results so far allow us to develop a consistent dimensional analysis of radial equations .
we start with eq .
( [ radeq ] ) performing the scale transformation @xmath94 and @xmath95 .
the angular variables are dimensionless and this , as specified by the factor @xmath85 , fixes the exponent @xmath92 .
the stochastic term becomes scale invariant for @xmath96 and the diffusion irrelevant for @xmath97 : this agrees with our previous analysis using the explicit solutions of the linear equations .
we can now use this technique to get some insight on nonlinear equations .
the kardar - parisi - zhang ( kpz ) equation @xcite for the radial surface fluctuation can be written as @xmath98 we note that scaling arguments were successfully employed in determining the critical behavior of kpz with a time dependent coefficient of the nonlinearity @xcite . if we select the same values of @xmath20 and @xmath22 as in the linear case in order to make the stochastic term scale invariant we find that diffusion and nonlinearity are irrelevant for @xmath99 and @xmath100 respectively .
this agrees with the naf dimensional analysis of kpz and thus we can not expect it to be correct .
in fact , the naf dynamical exponent @xmath101 found by means of a dimensional analysis of the kpz nonlinearity does not agree with the one measured in simulations and successfully predicted by the dynamic renormalization group analysis in one dimension @xcite .
it is plausible that the nonlinearity interacts with the noise to yield a correlation different from ( [ radcorr1 ] ) ( or from ( [ radcorr ] ) if dilution is included ) . in order to guess what form the correlation would have
we can try to get some illumination from scaling ( [ fv ] ) .
taking the short time limit in ( [ fv ] ) we find @xmath102 where the temporal prefactor expresses the time dependence of the variance of the interface center mass position @xcite .
using this form we modify , for instance , eq . ( [ radcorr ] ) to the new correlation @xmath103 which yields the exponents @xmath104 , @xmath92 and @xmath105 . in this case
we have the diffusion is irrelevant whenever @xmath99 and the nonlinearity is irrelevant if @xmath106 , where we have employed the relation @xmath107 in the derivation of the last inequality .
this last inequality could be considered as desirable but correlation ( [ radcorrkpz ] ) has a different drawback .
as we have said , the exponent @xmath108 describes the time power law dependence of the standard deviation of the interface center of mass position . for the eden model , which in planar format belongs to the kpz universality class @xcite , @xmath109 and so @xmath110 .
this exponent however does not agree with the one measured in simulations @xcite and this constitutes another disadvantage of the use of scaling ( [ fv ] ) in radial systems .
we note that considering the planar kpz equation on a growing domain does not simplify much things with respect to considering the full radial case @xcite .
together with the kpz equation , the villain - lai - das sarma ( vlds ) equation @xcite is one of most important nonlinear models for surface growth .
if one derives the vlds equation in the radial setting one finds irrelevance of the diffusion for @xmath111 and of the nonlinearity for @xmath112 . in this case
the well known hyperscaling relation for conserved growth models yields @xmath113 , and so equality of ( [ radcorrkpz ] ) with ( [ radcorr ] ) .
this suggests that the scaling properties derived herein for linear models are also valid for some nonlinear models as the vlds equation .
in this work we have examined the dynamic scaling of radial interfaces .
we have shown that long range fluctuations , which are describable by two - point correlation functions , become scale invariant for values of the critical exponents which may totally differ from the corresponding ones in a planar geometry . in rapidly growing radial systems
we have found @xmath114 all of which have a clear physical meaning .
the @xmath108 exponent , as we have discussed , describes the interface center of mass fluctuations .
the @xmath20 exponent associated to the radial surface fluctuation indicates how much the growing cluster deviates from a @xmath1dimensional growing hypersphere . for @xmath115
the growing cluster converges to a hypersphere concentrically growing at the deterministic rate @xmath116 , and in this sense the growth profile is flat .
this sort of convergence is of course _ weak _ : the profile converges to the hypersphere on ( any type of spatial ) average .
the microscopic roughness , which is given by the height - difference correlation , reduces to that of the planar profile @xcite and precludes strong or pointwise convergence .
this is so at least in situations in which the height - difference correlation is well defined ; otherwise a different scaling is possible even locally in space @xcite .
this makes precise and generalizes our previous results @xcite in the following way : a fast growth rate will make the interface weakly converge to a hypersphere .
we have further found the scaling relation @xmath117 , which quantitatively relates the center of mass fluctuations , growth rate and rate of ( weak ) convergence to / divergence from the hyperspherical profile .
these results , together with our previous ones @xcite , suggest the following generalization of the critical exponents for arbitrary values of @xmath118 @xmath119 which should be valid for linear models as well as some nonlinear equations as the vlds one , which present a conserved drift and its critical behavior is accessible to dimensional analysis . for the kpz equation , which nonlinearity introduces both mass and fluctuations into the system
, we expect a different behavior . as we have seen , the amount of mass entering the interface ( and of course the way it is introduced ) may change the critical behavior .
we have explicitly shown this by means of the linear models in absence and presence of dilution , which gave respectively rise to correlations ( [ radcorr1 ] ) and ( [ radcorr ] ) .
@xmath120 & @xmath121 & @xmath122 + classical exponents & roughness & super - roughness & weak convergence to the hyperspherical profile + we have summarized the different morphologies that can appear in table [ table ] . in the first column , for small @xmath118 , we have placed the regime in which the classical values of the exponents are recovered .
the other three columns represent situations in which this is not the case .
the `` roughness '' column describes the situation in which the radial fluctuations do not average each other out for long times , so in this regime the overall appearance of the cluster deviates from the hyperspherical profile .
the `` super - roughness '' column states a particular case of this last regime , in which the amplitude of the radial fluctuations grows faster than the arc - length sustaining them .
finally , the last column describes the regime in which the weak convergence to the hyperspherical profile is reached asymptotically in time .
note that the meaning of the exponent @xmath21 in ( [ gscaling ] ) changes for @xmath123 and for @xmath106 . for @xmath124
this exponent describes the short time dependence of the spatially averaged pointwise width with time , as in the classical situation . for @xmath106 ,
as we have mentioned , it describes the center of mass fluctuations , i. e. , the _ less _ dominant source of fluctuations is chosen in each regime . for @xmath123 and @xmath125 ( which implies the height difference correlation
is well defined and the pointwise width is finite ) there are always center of mass fluctuations .
for @xmath126 the situation is different as it is possible to find growth regimes with no center of mass fluctuations and which pointwise width increases as a power law of time .
this is precisely the weak converge to the hyperspherical profile : while microscopically the interface is rough , macroscopically it is a concentrically propagating hypersphere .
the change of meaning of the exponent @xmath21 coincides with a change of physics .
the long range shape of the interface is no longer affected by the diffusion mechanism , which becomes irrelevant , but for the system size evolution , which sculptures the resulting profile @xcite . and
so , the violation of scaling ( [ fv ] ) is associated with the advent of a new dominant physical mechanism in the interface macroscale : system size growth overtakes diffusion and determines the cluster macroscopic shape .
this work has been partially supported by the micinn ( spain ) through project no .
mtm2010 - 18128 .
weak convergence is formalized in mathematical analysis texts as convergence in the dual space .
intuitively , it corresponds to the spatial self - averaging of a wildly oscillating quantity .
a good example of this is given by the riemann - lebesgue lemma , which states the weak convergence of wildly oscillating trigonometric functions .
also , weak convergence is a relevant concept in physics : it might result , for instance , from the self - averaging of random fluctuations . as an example , some quantum mechanical quantities weakly converge to their classical counterparts .
|
the asymptotic shape of randomly growing radial clusters is studied .
we pose the problem in terms of the dynamics of stochastic partial differential equations .
we concentrate on the properties of the realizations of the stochastic growth process and in particular on the interface fluctuations .
our goal is unveiling under which conditions the developing radial cluster asymptotically _ weakly _ converges to the concentrically propagating spherically symmetric profile or either to a symmetry breaking shape .
we demonstrate that the long range correlations of the surface fluctuations obey a self - affine scaling and that scale invariance is achieved by means of the introduction of three critical exponents .
these are able to characterize the large scale dynamics and to describe those regimes dominated by system size evolution .
the connection of these results with mathematical morphogenetic problems is also outlined .
|
non - fermi liquid ( nfl ) or strange metal phase is often observed to precede many correlated phases , such as antiferro- or ferromagnetism , and unconventional superconductivity.@xcite it is characterized by deviation from the well - defined fermi - liquid ( fl ) predictions , such as the resistivity - temperature exponent ( @xmath1 ) being @xmath51 , while @xmath6 is expected from the fl theory .
systematic studies in various superconducting families have also revealed that superconducting transition temperature ( @xmath4 ) increases as the exponent @xmath1 decreases , i.e. , as the system deviates farther from the fl behavior.@xcite therefore , nfl state is considered an important problem towards the understanding of unconventional superconductivity . in the phase diagram , nfl state
is often seen to arise near the boundary between an order phase and the fl state , where superconductivity is optimized.@xcite yet , considerable counter - examples are also present where a nfl phase is present without a quantum phase transition@xcite or superconductivity.@xcite therefore , it has remained a ` chicken - and - egg ' type problem whether a quantum critical point drives the nfl state or vice versa , and how they conspire into superconductivity . to definitely answer this issue
, one can approach the nfl state from the fl regime .
the transition from the fl to nfl state is very adiabatic , i.e , the resistivity exponent @xmath1 _ continuously _ changes from 2 ( fl ) to 1 or even below 1 in the nfl state at a given temperature .
theoretical models , relying on either critical fluctuations,@xcite or marginal fluctuations,@xcite or other quantum fields,@xcite predict that when quantum fluctuations become gapless ( critical ) , the low - energy excitation spectrum becomes linear in energy in the entire brillouin zone ( bz ) , and the quasiparticle description fully breaks down . in such a case , the transport relaxation rate behaves as @xmath7 , giving a linear resistivity exponent ( @xmath1 = 1 ) . on the other hand ,
the weak correlation in the fl regime gives a quadratic energy dependence of the excitations , and the corresponding low - energy quasiparticles produce @xmath6 .
model descriptions thus treat these regimes separately , and often unable to reconcile the adiabatic crossover between the fl to nfl states within a unified , self - consistent framework.@xcite here we present a theory of the nfl phenomena arising from the _ momentum _ dependent self - energy due to canonical density - density fluctuations .
the many - body self - energy is calculated based on the momentum - resolved density - fluctuation ( mrdf ) model within the fluctuation - exchange approximation.@xcite the method is known to be appropriate for the intermediate coupling region where quantum fluctuations are of the order of the kinetic energy , stipulating a dynamically frustrated system .
we apply the result in various cuprate materials .
we find that the dominant source of the nfl state stems from large degeneracy in the quasiparticle density .
the nfl strength is incipiently controlled by the position of the dos degeneracy from the fermi level ( @xmath8 ) as well as its occupied momentum space volume ( i.e. the amount of degeneracy ) at a given condition
. one of our key results is that even in the extreme nfl state where @xmath91 ( near the optimal doping in cuprates ) , not all electronic states in the bz become incoherent .
conversely , in the fl state ( in overdoped region ) , not all quasiparticles are necessarily long - lived .
this can be understood from the behavior of the @xmath0-dependent self - energy @xmath10 , allowing a @xmath0-dependent exponent @xmath11 .
we find that @xmath11 strongly varies in the bz : the high - density region ( ` hot - spots ' ) governs @xmath12 , while the low - density region ( ` cold - spots ' ) gives @xmath13 .
this allows a coexistence and competition between the nfl and fl physics in the same system .
the spectral weight transfer between the ` hot- ' and ` cold - spots ' , caused by doping , temperature , and other parameters , manifests into an adiabatic transition between the fl and the nfl state .
the results are applicable to a wider class of correlated materials in which large density of states is caused by van - hove singularity ( vhs ) , or liftshitz points ( as in pnictides ) , spin - orbit couping ( in heavy - fermions and actinides ) and leads to strong anisotropic self - energy effects . we also make direct comparison with corresponding experiments . a recent angle - resolved photoemission spectroscopy ( arpes )
experiment also observed a similar @xmath0-dependence in the electronic life - time in la - based cuprate.@xcite it was found that the inverse of the quasiparticle lifetime changes from quadratic to linear as we move from nodal to the antinodal region , even in the overdoped sample where the transport data suggest a simple fl behavior .
again , angle - dependent magnetoresistance ( admr ) measurements on overdoped tl - based cuprate also exhibited the similar behavior , in that the scattering rate becomes quadratic to linear - in-@xmath14 as we moves from the nodal to the antinodal region of the sample.@xcite the rest of the paper is organized as follows . in sec .
[ sec : model ] , we discuss the mrdf model and the tight - binding dispersion .
momentum - dependent self - energy results and comparison with experiment is done in sec .
[ sec : selfenergy ] .
the overall fl / nfl behavior of a given system is characterized by the resistivity calculation and its doming dependence are discussed in sec .
[ sec : resistivity ] . in sec .
[ sec : materials ] , we study the materials dependence of the resistivity - temperature exponent and its dependence with superconducting transition temperature is presented . finally , we conclude in sec . [
sec : conclusions ] .
the robustness of the results against the value of the hubbard interaction @xmath15 is demonstrated in the appendix [ apx : u ] .
cuprate is a prototype of correlated superconducting family where the interplay between nfl , unconventional superconductivity , and various intertwined orders leads to a complex doping dependent phase diagram ( see fig . [ fig1]).@xcite yet , the band structure is rather straightforward with a single and strongly anisotropic band passing through the fermi level .
we consider a realistic band structure including upto fourth order tight - binding hoppings ( @xmath16 , @xmath17 , @xmath18 , and @xmath19 ) as @xmath20 . the second nearest neighbor hopping @xmath21 has a special importance in cuprates as it controls the flatness of the band near @xmath22 and its equivalent points .
this generates paramount degeneracy in the doss at the vhs position .
interestingly , an earlier density functional theory ( dft ) calculation demonstrated that the optimal @xmath4 in different cuprates scales almost linearly with the corresponding @xmath23 ratio.@xcite it is also known that the energy of the vhs in a given system moves closer to the fermi level with increasing doping , and crosses it in the vicinity of the optimal doping , see fig .
these results give an important clue that the density fluctuation driven nfl is concentrated at the vhs region , while the other parts of the bz continue to be fl - like .
their competition is controlled by chemical potential ( doping ) and @xmath21 ( materials , pressure ) .
the density fluctuations dominate in the magnetic channel , giving paramagnon excitations in cuprates .
these excitations are strongly dispersive , and present at all dopings.@xcite the coupling between density fluctuations and electrons gives rise to a complex self - energy , which can be calculated in a variety of numerical techniques.@xcite here we invoke a self - consistent momentum - resolved density fluctuation ( mrdf ) technique@xcite which is suitable for the intermediate coupling regime where kinetic energy and dynamical correlations are comparable .
the advantage of this technique is that it can efficiently calculate the full @xmath0-dependence of the self - energy .
the self - energy in the fluctuation - exchange approximation is obtained as ( see appendix c ) @xmath24,\end{aligned}\ ] ] where @xmath25 and @xmath26 are fermionic and bosonic distribution functions , respectively .
@xmath27 is the total number of lattice sites . @xmath28 and @xmath29^{-1}$ ]
are the self - energy dressed spectral weight and green s function , respectively .
the vertex correction is @xmath30 .
@xmath31 is the back - reaction potential of quasiparticle density fluctuation which is separated into the spin ( @xmath32 ) and charge ( @xmath33 ) density channels as @xmath34 , \label{vq}\end{aligned}\ ] ] where @xmath35 , and @xmath36 , and @xmath15 is the onsite hubbard interaction .
@xmath37 is the corresponding bare correlator ( same for both spin and charge channels ) , evaluated self - consistently , as @xmath38 the tight - binding parameters and the hubbard @xmath15 for different materials are deduced from dft calculations , and are listed in appendix a , and @xmath8 is calculated self - consistently at each doping .
the real part of the self - energy @xmath39 renormalizes the band as @xmath40 , where @xmath41 .
the corresponding spectral weight is determined by the combined effects of @xmath42 for @xmath43 , and @xmath44 at finite energy .
we fit @xmath45 ( where @xmath46 is a proportionality constant ) . @xmath47 and @xmath11 behave similarly ( since @xmath48 and @xmath49 are connected to each other via kramers - kronig relation ) : lower their values imply stronger the correlation strength at a given @xmath50 . in fig .
[ fig2 ] , we show the @xmath0- and doping dependence of @xmath49 and @xmath11 for a representative material la@xmath51sr@xmath52cuo@xmath53 ( lsco ) . the results are compared with the corresponding angle - resolved photoemission spectroscopy ( arpes ) data .
@xmath0-resolved values of the half - width half - maximum linewidth ( defined as @xmath54 ) of the arpes spectral weight show that the exponent @xmath11 gradually decreases as one moves from the nodal region ( @xmath13 ) to the antinodal one ( @xmath55).@xcite our theoretical result , shown in fig .
[ fig2](b ) , shows a similar behavior in that @xmath13 near the nodal region , which decreases to @xmath56 as one reaches the antinodal region .
the full @xmath0-space profile for @xmath11 is shown in figs .
[ fig2](c - e ) for three characteristic hole dopings for the same material ( black line gives the corresponding non - interacting fs ) .
the overall @xmath0-dependence of @xmath11 remains similar at all dopings : @xmath11 attains its minimum value around the antinodal region and maximum somewhere around the nodal region . in the underdoped region , where the vhs is well below @xmath8 , we find that the overall @xmath11 profile is less @xmath0-sensitive , and varies within the range of @xmath57 . near the optimal doping , where the vhs exactly crosses above @xmath8 , we find that the @xmath0-dependence of @xmath11 becomes strongest , and also @xmath11 obtains its minimum value of @xmath50.5 at the antinodal region .
this is also the minimum possible value of @xmath11 at all dopings and momenta for this material . in this region
, we find that the resistivity - temperature exponent also attains its minimum value of @xmath50.7 as shown in fig .
[ fig3](b ) . finally , as the vhs crosses above @xmath58 , again the value of @xmath11 increases , indicating that the strength of dynamical correlation is maximum around the optimal doping where the fermi surface degeneracy is highest.@xcite interestingly , in the overdoped region , where the resistivity data below shows an overall fl - behavior , the antinodal regions continue to show nfl self - energy behavior , in consistent with the arpes data on lsco at @xmath59.@xcite as mentioned above , the momentum profile of @xmath11 is similar to that of @xmath47 , and thus the spectral weight distribution on the fermi surface also becomes characteristically the same . in
what follows , the spectral weight at the antinodal region becomes minimum ( where @xmath47 is maximum ) and it gradually increases as one moves to the nodal region , giving a fermi arc shape of the fermi surface , a trademark feature of cuprate fermiology.@xcite
when the @xmath0-dependence of the self - energy is neglected , a direct relation between the microscopic single - particle spectral properties ( quantified by @xmath60 ) and the macroscopic transport behavior ( @xmath61 ) can be established . however , as the system acquires strong anisotropy in @xmath11 , it becomes less intuitive to deduce the overall correlation landscape of the system by the transport behavior .
we compute the temperature dependence of the resistivity by including the full momentum and energy dependence of the self - energy . within the linear response theory ,
the dc conductivity can be calculated from the current - current correlation function in the equivalent way as the density correlation is evaluated in eq . , and we obtain : @xmath62 where @xmath63 and @xmath64 have the usual meanings , and @xmath65 is the band velocity . neglecting any anomalous term , we obtain resistivity as @xmath66 . the results are presented in fig .
[ fig3](a ) for lsco at different dopings .
we find that the resistivity exponent becomes minimum near the optimal doping where the vhs crosses @xmath8 , see fig .
[ fig3](b ) . here , the system acquires dominant nfl - behavior with @xmath671 . at the same doping , the self - energy exponent @xmath11 in fig .
[ fig2](d ) not only obtains its minimum value ( min@xmath68\sim 0.5 $ ] ) , but also it occupies larger @xmath0-space area .
however , the other parts of the bz remains fl - like with @xmath11 as large as @xmath52 .
similarly , in both under- and overdopings , where @xmath69 , the antinodal region continues to have @xmath55 . finally , we repeat the calculation for the ybco material as a function of doping , using the corresponding realistic tight - binding parameter set,@xcite and the result is plotted in fig . [ fig3](b ) .
we consistently find that @xmath1 is minimum near its optimal doping as the corresponding vhs passes through @xmath8 .
admr technique has the ability to probe the angle - variation of the resistivity by tilting the magnetic field with the sample orientation .
this allows to effectively measure the scattering life - time @xmath70 as a function of fermi surface angle @xmath71 .
an earlier admr study on overdoped tl@xmath72ba@xmath72cuo@xmath73 found that @xmath74 varies as @xmath75 in the nodal region ( @xmath76 ) and it gradually changes to @xmath14 in the antinodal region ( @xmath77).@xcite this result is consistent with out finding of quasiparticle life - time variation shown in fig .
note that even the overall resistivity exponent is close to 2 in the overdoped region , however , its local variation reveals that both the single - particle life - time and scattering rate consistently remains incoherent in the antinodal region .
the celebrated paper by pavarini _ et al._@xcite pointed out an intriguing relationship between the @xmath78 ratio obtained in different materials with their @xmath4 .
@xmath78 triggers higher degeneracy in the dos ( see appendix b3 ) , and hence it is natural to expect that the strength of the nfl state would also increase .
we calculate the resistivity exponent @xmath1 for different values of @xmath23 by fixing the vhs at the @xmath79 , and the result is plotted in fig .
indeed , we find that with increasing @xmath23 , @xmath1 decreases , that means , the system becomes more nfl like .
with increasing @xmath78 , both the dos at vhs increases and the bandwidth decreases ( see appendix b3 ) , and thus the nfl phenomena also increases .
it is already known that the optimal @xmath4 increases with increasing @xmath23,@xcite and with decreasing @xmath1 .
interestingly , it is also observed that as @xmath1 decreases , the residual resistivity , @xmath80 as @xmath81 , also decreases , and @xmath4 increases .
this phenomena is consistently observed in various cuprates , pnictides and heavy - fermions.@xcite our results thus give the microscopic insight into this empirical observation .
finally , we discuss various other phenomena in cuprates and their impact on our conclusions .
the doping dependence of the dynamical correlation is different from that of the static correlation .
the dynamical correlation , arising from the density fluctuation is determined by the strength of degeneracy ( here vhs ) .
it becomes dominant near the optimal doping where the vhs lies at @xmath8 . on the other hand ,
static correlation is determined by the onsite coulomb interaction ( @xmath15 in eq . ) and it gives rise to a mott insulator state at half - filling . while the static correlation decreases with doping , kinetic frustration increases and becomes maximum near the optimal doping at the vhs point .
kinetic frustration also introduces other competing phases such as antiferromagnetism , charge - density wave order , and other orders , and can modify the results in the underdoped region . yet , without including any order parameter or quantum critical point , we get a good , quantitative agreement between the theory and experiment in the entire doping regime , suggesting that the dominant nfl physics stems from the df fluctuations in these systems .
we have also studied all the results presented here for different values of @xmath15 , and shown that the momentum profile of @xmath11 , doping dependence of @xmath1 etc .
remain the same for different values of @xmath15 , except that their overall strength decreases with decreasing @xmath15 , i.e. , the system moves towards fl - like as static correlation decreases .
the fl state is recovered in the weak - coupling region ( see appendix b ) .
the important message of our result is that for strongly anisotropic materials where the dynamical fluctuations have significant momentum dependence , the resistivity - temperature exponent is not a robust measure of the full correlation spectrum of the underlying quasiparticle states .
we found that even in the underdoped and overdoped regions , where resistivity exponent n@xmath822 , there are considerable amount of nfl - like incoherent states lying in the antinodal regions and vice versa .
similarly , in the extreme nfl region near the optimal doping regime ( determined by n@xmath51 ) , the nodal quasiparticles continue to behave fl - like ( with @xmath83 ) . both as a function of temperature and doping ( and other tunnings )
, the spectral weight is transfered between the nfl and fl regions and the system adiabatically transforms from a dominant nfl to a fl - like state , as seen in experiments .
our work suggests that the microscopic and macroscopic landscapes of the nfl behavior can be characteristically different and that a direct correspondence between @xmath0-resolved spectroscopy ( such as arpes , and quasiparticle interference ( qpi ) pattern ) and the transport , and thermodynamical properties are necessary to deduce the global and local nfl behavior of a given system
. 0.5 cm
@xmath16 & @xmath17 & @xmath18 & @xmath19 & u & reference + lsco & 0.4195 & -0.0375 & 0.018 & 0.034 & 1.6 & + ybco & 0.35 & -0.06 & 0.035 & -0.005 & 1.9 & + +
all results and conclusions presented above are obtained for material specific values of the hubbard @xmath15 . in the appendix , we investigate them for different values of @xmath15 and establish their robustness .
the following results also demonstrate the distinction between the doping dependence of the static correlation ( @xmath15 ) and the dynamical correlation ( @xmath84 in eq . 2 ) .
+ keeping all other parameters the same , we expect that the system would tend to transform from nfl to fl like as we decrease the values of @xmath15 .
this is precisely what we observe in fig .
[ figb1 ] where we plot the momentum profile of @xmath11 at a fixed doping of @xmath85 for lsco for three different values of @xmath15 . in all three cases ,
the momentum profile remains very much the same , as we expect , since the momentum dependence is governed by the band structure anisotropy and spectral density ` hot - spots ' .
we notice a characteristic change in the overall range of @xmath11 ( as highlighted by red circles in the adjacent colorbars ) .
we find that both the minimum and the maximum values of @xmath11 increases with decreasing @xmath15 .
in addition we also notice that the @xmath0-space area of the nfl region ( @xmath861 ) also decreases with decreasing @xmath15 , reflecting that the system moves towards fl as correlation weakens .
the result is confirmed by the resistivity exponent calculation as presented in fig .
reffigb2 .
+ we obtain the same conclusion in the resistivity - temperature exponent @xmath1 , calculated with the same parameter sets as in fig .
[ figb1 ] .
we find that the overall doping dependence of @xmath1 is similar for all three values of @xmath15 : it obtains the minimum value near the optimal doping where the vhs passes through the fermi level , irrespective of the values of @xmath15 . however , the overall value of @xmath1 increases with decreasing @xmath15 as the system moves towards the fl like with lowering its correlation strength .
+ finally we study the evolution of the @xmath78 vs. @xmath1 plot for different values of @xmath15 in fig .
[ figb3 ] .
we have learned in the main text that @xmath1 decreases as the @xmath78 ratio increases , keeping the corresponding vhs fixed at the fermi level for all cases .
this is because the dos at the vhs increases with increasing @xmath78 and the bandwidth simultaneously decreases .
therefore , the system becomes more nfl - like as @xmath78 increases .
this conclusion remains intact as we tune the values of @xmath15 . for different values of @xmath15 , the general trend in @xmath1 with @xmath78
remains the same , however the overall range of @xmath1 increases with decreasing @xmath15 as we also found in figs .
[ figb1 ] and [ figb2 ] before .
we start with the standard definition of spin / charge susceptibility @xcite which is given by @xmath87 where @xmath88 denotes the spin / charge density where indices i , j denote different components ( for example x , y , z components ) in case of spin susceptibility . charge and spin densities @xmath89
are given in the second quantized notation as @xmath90 where @xmath91s are the pauli spin matrices in 2d .
@xmath92 is the dressed quasi - particle creation operator ( sometimes called dyson orbitals ) at bloch momentum @xmath50 and spin @xmath93 .
since the ground state is spinless , both transverse and longitudinal spin - densities , as well as the charge density term yields the same bare susceptibility .
in general , we can write eq . as @xmath94 where the momentum conservation is imposed .
we can decompose the four - field terms into bi - linear terms within wick s theorem , and allow the spin - conservation condition for the ground state .
then we obtain , @xmath95 where we have identified the terms in the brackets as self - energy dressed green s functions such as @xmath96 the fourier transform of green s function is defined as @xmath97 and that of the susceptibility as @xmath98 here @xmath99 and @xmath100 are the fermionic and bosonic matsubara frequencies , respectively .
finally , we get @xmath101 from here onwards we drop the index @xmath93 and assume an implied sum over @xmath93 index .
it is not easy to perform the matsubara frequency summation using self - energy dressed green s function .
so we use its spectral representation as @xmath102 where @xmath103 is the corresponding spectral weight defined as @xmath28 , where @xmath104 is obtained by taking the analytical continuation to the real frequency @xmath105 , with @xmath106 being infinitesimal broadening .
so the susceptibility expression becomes @xmath107 the term in the bracket can be evaluate by matsubara technique @xcite and we arrive at the expression @xmath108 where @xmath25 denotes the fermi distribution function .
the computation of the susceptibility is done using analytical continuation to the real frequency as discussed before .
next we calculate the self energy starting from gw approximation @xcite which is given by @xmath109 where @xmath110 is the vertex correction discussed in the main text .
@xmath111 is the fluctuation - exchange potential which we obtain within the random - phase approximation as @xmath112 where @xmath113 and 1 for spin ( @xmath32 ) and charge ( @xmath33 ) density fluctuations .
again , to aid the matsubara frequency summation , we use the spectral representation of @xmath111 as @xmath114 we denote the fluctuation - exchange potential as @xmath115 . therefore , using eqs . and , we get @xmath116.\end{aligned}\ ] ] all other symbols were defined in the main text
. 1 g.r .
stewart , rev .
phys . * 73 * , 797 ( 2001 ) .
subir sachdev , rev .
phys . * 75 * , 913 ( 2003 ) .
t. das and c. panagopoulos , new j. phys .
* 18 * , 103033 ( 2016 ) .
l. taillefer , la physique au canada * 67 * , 109 ( 2011 ) .
t. shibauchi , a. carrington and y. matsuda , annu .
cond . mat .
phys . * 5 * , 113 - 135 ( 2014 ) .
n. doiron - leyraud , p. auban - senzier , s.r .
de cotret , a. sedeki , c. bourbonnais , d. jerome , k. bechgaard and l. taillefer , arxiv:0905.0964 ( 2009 ) .
p. coleman and a.j .
schofield , nature * 433 * , 226 - 229 ( 2005 ) m.b .
maple , r.e .
baumbach , n.p .
butch , j.j . hamlin and m. janoschek ,
j. low temp .
phys . * 161 * , 454 ( 2010 ) .
y. matsumoto _
et al . _ ,
science * 331 * , 316 - 319 ( 2011 ) .
sidorov , m. nicklas , p.g .
pagliuso , j.l .
sarrao , y. bang , a.v .
balatsky and j.d .
thompson , phys .
lett . * 89 * , 157004 ( 2002 ) .
et al . _ ,
b * 94 * , 094524 ( 2016 ) .
bhatt and p.a .
lee , j. appl . phys . * 52 * , 1703 ( 1981 ) ; e. miranda and v. dobrosavljevi , rep
. prog . phys . *
68 * , 2337 ( 2005 ) ; c. pfleiderer , p. bni , t. keller , u.k .
rler and a. rosch , science * 316 * , 1871 ( 2007 ) .
s. sachdev quantum phase transition ( cambridge univ .
press , cambridge , 1999 ) .
she , j. zaanen , a.r .
bishop and a. v. balatsky , phys .
b * 82 * , 165128 ( 2010 ) ; s .- x .
et al . _ ,
* 106 * , 047004 ( 2011 ) .
d. bergeron , d. chowdhury , m. punk , s. sachdev and a .- m.s .
tremblay , phys .
b * 86 * , 155123 ( 2012 ) p. gegenwart , q. si and f. steglich , nature physics * 4 * , 186 ( 2008 )
varma , p.b .
littlewood , s. schmitt - rink , e. abrahams and a.e .
ruckenstein , phys .
64 * , 497 ( 1990 ) . w. xu , k. haule and g. kotliar , phys . rev
. lett . * 111 * , 036401 ( 2013 ) ; x. deng , j. mravlje , r. zitko , m. ferrero , g. kotliar and a. georges , phys .
rev . lett . * 110 * , 086401 ( 2013 ) ; x. deng , a. sternbach , k. haule , d.n .
basov and g. kotliar , phys .
* 113 * , 246404 ( 2014 ) .
w. xu , g. kotliar and a.m. tsvelik , arxiv:1610.03720 .
phillips and c. chamon , phys .
* 95 * , 107002 ( 2005 ) .
d. dalidovich and s .- s .
lee , phys .
b * 88 * , 245106 ( 2013 ) ; s. sur and s .- s .
lee , phys .
b * 91 * , 125136 ( 2015 ) .
vidhyadhiraja , a. macridin , c. sen , m. jarrell and m. ma , phys .
lett . * 102 * , 206407 ( 2009 ) .
davison , k. schalm and j. zaanen , phys .
b * 89 * , 245116 ( 2014 ) ; j. zaanen , y .- w .
sun , y. liu , k. schalm , holographic dualtity for condensed matter physics ( cambridge univ .
press , 2015 ) .
maslov , a.v .
chubukov and r. saha , phys .
b * 74 * , 220402(r ) ( 2006 ) .
t. das , r.s .
markiewicz and a. bansil , advances in physics * 63 * , 151 - 266 ( 2014 ) .
t. das , j .- x .
zhu and m. j. graf , phys .
lett . * 108 * , 017001 ( 2012 ) .
t. das , and k. dolui , phys .
b * 91 * , 094510 ( 2015 ) , r.s .
dhaka , t. das , n.c .
plumb , z. ristic , w. kong , c.e .
matt , n. xu , k. dolui , e. razzoli , m. medarde , l. patthey , m. shi , m. radovic and j. mesot , phys .
b * 92 * , 035127 ( 2015 )
. j. chang _ et al .
_ , nature communications * 4 * , ( 2013 )
. m. abdel - jawad , m.p .
kennett , l. balicas , a. carrington , a.p .
mackenzie , r.h .
mckenzie , and n.e .
hussey , nature phys .
* 2 * , 821 ( 2006 ) .
m m j french , j g analytis , a carrington , l balicas , and n e hussey , new j. phys . * 11 * , 055057 ( 2009 ) .
b. keimer , s.a .
kivelson , m.r .
norman , s. uchida and j. zaanen , nature * 518 * , 179 ( 2015 ) .
e. fradkin , s.a .
kivelson , nature phys .
* 8 * , 865 - 866 ( 2012 ) .
e. pavarini , i. dasgupta , t. saha - dasgupta , o. jepsen and o.k .
andersen , phys .
lett . * 87 * , 047003 ( 2001 ) .
m.p.m . dean _
et al . _ , nature mat .
* 12 * , 1019 - 2023 ( 2013 ) .
the many - body self - energy can nowadays be calculated in variety methods such as fluctuation - exchange ( flex ) method , quantum monte - carlo@xcite , or dynamical mean - field theory ( dmft)@xcite , and others .
fermi surface degeneracy here does not refer to fermi surface nesting , rather indicate that the degeneracy of the band is highest on the largest parts of the fermi surface .
manuscript under preparation .
r. s. markiewicz , s. sahrakorpi , m. lindroos , hsin lin , and a. bansil , phys .
b * 72 * , 054519 ( 2005 ) .
t. das , r.s .
markiewicz and a. bansil , advances in physics * 63 * , 151 - 266 ( 2014 ) .
|
non - fermi liquid ( nfl ) state represents an ensemble of incoherent quantum fluids arising from the coupling between electrons and massless ( critical ) excitations , and is separated by phase boundary from the quasiparticle behavior in the fermi - liquid ( fl ) theory .
here we show that such sharp distinction breaks down in cuprates , and that both nfl and fl states coexists in different momentum ( @xmath0 ) regions at all dopings .
their coexistence originates from the strong anisotropy in the many - body self - energy , arising from dispersive density - density fluctuations . the self - energy attains maxima ( nfl - like ) in the region where density degeneracy is optimum ( antinodal region ) , while the nodal region remains fl - like at all dopings .
we attribute the global nfl / fl behavior via the calculation of the resistivity - temperature exponent ( @xmath1 ) .
surprisingly , we find that the entire brillouin zone becomes neither fully incoherent , nfl - like even at optimal doping with @xmath2 , nor fully fl - like even at overdoping ( @xmath3 ) .
as density degeneracy increases in different materials with increasing superconducting @xmath4 , @xmath1 decreases ; providing a microscopic explanation to this intriguing relationship .
all results , including coexistence of nfl- and fl - self - energies in the @xmath0-space , and their doping , materials dependencies are compared with available experimental data , followed by definite predictions for future studies .
|
during merging processes leading to the formation of clusters of galaxies , large amounts of gravitational energy are released on timescales of the order of @xmath3gyr .
most of this energy is contained in hot ( temperatures @xmath4kev ) x - ray emitting plasma which constitutes , along with the dark matter , the dominant fraction of the intracluster medium ( icm ; e.g. , @xcite ) .
in addition to the thermal gas , however , @xmath5 g magnetic fields and ultrarelativistic electrons are present within the icm as well , manifesting most clearly in extended diffuse radio structures such as giant and mini radio halos in the central parts of clusters , or radio relics at cluster peripheries ( see the reviews by @xcite ) .
these non - thermal constituents of the icm are believe to be related to the energy dissipation processes enabled by large - scale shocks formed at the outskirts of merging systems , and/or turbulence driven by various mechanisms at early post - merger stages of a cluster lifetime .
the presence of ( or rather the amount of ) hadronic cosmic rays in the diffuse cluster environment is still an open question , being currently probed with new - generation @xmath6-ray instruments ( see , e.g. , @xcite ) . besides
radio halos and relics , extended lobes and plumes of radio galaxies make up yet another class of diffuse non - thermal structures often found in the icm .
these are formed due to the jet activity of accreting supermassive black holes ( smbhs ) hosted by the cluster galaxies .
the large - scale morphologies of such lobes are rarely regular or symmetric , unlike in the case of ` classical doubles ' typical for poorer ( galaxy group ) environments .
instead , lobes of cluster radio galaxies are often bent ( ` tailed ' radio sources ) , irregular , or even amorphous , reflecting the dramatic impact of high - pressure ambient medium , high peculiar velocities of parent galaxies , and highly intermittent jet activity of cluster smbhs ( e.g. , @xcite ; see also the discussion in 5.3 below ) .
such structures may therefore be considered as useful probes of the dynamical state and structure of the icm , or even as effective tracers of galaxy clusters at moderate and high redshifts ( see , e.g. , @xcite ) . since they are best characterized at low radio frequencies , the operation of the next generation of low - frequency interferometers like the low frequency array @xcite and the murchison widefield array @xcite are in this context much anticipated .
recently , much attention has been given to radio sources located within the central parts of galaxy clusters , since the mechanical energy output of radio - loud active galactic nuclei ( agn ) which are hosted by central elliptical ( cd ) galaxies is widely believed to be responsible for suppressing the cluster cooling flows , quenching the star - formation in cd systems and the growth of their smbhs ( see , e.g. , @xcite ) . in order to investigate the details of such feedback processes ,
a precise spectral and morphological characterization of central radio structures and of their surroundings is however obligatory , and this requires deep high - resolution exposures with various instruments operating from radio to x - ray bands . without such extensive dataset ,
the physics of the lobes - icm interactions may remain elusive . in this paper
we analyze the peculiar radio structure observed in the direction of the central parts of the galaxy cluster abell585 ( @xmath0 ) .
the paper is organized as follows .
the galaxy cluster abell585 is briefly introduced in 2 .
multi - frequency data for the system are analyzed in 3 , and discussed further in 4 .
the interpretation of the obtained observational results is presented in 5 , and the summary of the studies is given in 6 .
in the paper we assume modern cosmology with @xmath7kms@xmath8mpc@xmath8 , @xmath9 , and @xmath10 @xcite .
all source positions are given in the j2000.0 coordinate system .
the cluster of galaxies abell585 ( another designation zwcl0727.1 + 4057 ; @xcite ) is centered at @xmath11@xmath12 and @xmath13@xmath14 ( the corresponding galactic longitude and latitude of 17788 and + 2444 , respectively ; @xcite ) . the cluster is classified as a ( b m ) type ii system .
the red magnitude of the cluster s tenth brightest member galaxy is 17.0mag , and there are in total 35 galaxies with apparent red magnitudes between mag@xmath15 and mag@xmath16 .
abell585 occupies an area of 26 in diameter and its richness class is 0 @xcite . using the photometric redshifts of galaxies from the sloan digital sky survey data release 6 ( sdss dr6 ; @xcite ) , @xcite identified a cluster named whlj073045.4@xmath17405038 , which overlaps with abell585 .
the photometric redshift of this sdss cluster is @xmath18 , and the spectroscopic redshift of the brightest cluster galaxy ( @xmath19mag ) is @xmath20 .
the total @xmath21-band cluster luminosity summed over all 27 member galaxies , which are spread over @xmath22mpc , is @xmath23 . in the catalogue of superclusters ( scl ; @xcite ) , abell585 together with abell580 and abell591 appear as a part of the scl074 , for which the central position is , approximately , [ @xmath11@xmath24 , @xmath13@xmath25 .
abell580 is centered at [ @xmath11@xmath26 , @xmath13@xmath27 at the redshift of @xmath28 , with the diameter @xmath29 , ( b m ) type ii
iii , and richness class 1 .
abell591 is centered at [ @xmath11@xmath30 , @xmath13@xmath31 at the redshift of @xmath32 , with the diameter of @xmath33 , morphology system of ( b m ) type iii , and richness class equal 1 .
the scl074 triplet may therefore represent a rather dynamic merging structure .
the analysis presented in this paper is based on the archival optical , x - ray , and multi - frequency radio data , as described in detail below .
a digital sky survey ( dss ) image of the central part of the galaxy cluster abell585 is shown in figure[fig : sdss ] , where the cluster member galaxies , as well as the background quasar b30727 + 409 ( see table1 ) , are marked with the cross symbols .
we use the sloan digital sky survey ( sdss ; @xcite ) database to locate galaxies and to establish their associations with the abell585 cluster based on spectroscopic redshifts from the sdss dr9 @xcite .
we defined all the member galaxies to have velocities in the range between @xmath34 and @xmath35 km / s ( i.e. about @xmath362000 km / s around the velocity corresponding to z=0.12 ) .
the redshifts of the member galaxies are provided in table1 along with the source positions , @xmath21-band magnitudes , and central black hole masses estimated by us from the stellar velocity dispersion using the @xcite relation except for the black hole mass for b30727 + 409 , estimated using the quasar emission lines ( see 5.1 ) . in total
, we found 21 galaxies within the 15 radius ( corresponding to a projected linear size of about 4mpc ) around the central cluster region , plus a luminous face - on spiral with bright optical core lacking any spectroscopic redshift determination , but with the sdss photometric redshift of @xmath37 and an optical apparent magnitude similar to the cluster member galaxies . [ cols="^,^,^,<,^,^ " , ] notes : @xmath38 the peak flux only available ; @xmath39 the original b3 flux density given in the ckl scale @xcite is multiplied by a factor of 1.129 relative to the common baars scale according to @xcite
. + references : ( 1 ) vlss : @xcite ; ( 2 ) 6c : @xcite ; ( 3 ) wenss : @xcite ; ( 4 ) b3 : @xcite ; ( 5 ) nvss : @xcite ; ( 6 ) first : @xcite ; ( 7 ) this paper ; ( 8) @xcite ; ( 9 ) gb2 : @xcite ; ( 10 ) @xcite ; ( 11 ) ocra - p : @xcite .
+ the area of interest has been observed a number of times with the vla in the past , due to the fact that the quasar b30727 + 409 is a vla phase calibrator often observed in a snapshot mode ( usually a few minutes of integration time ; for details see col.4 in table2 ) .
hence , for the purpose of our investigation , apart from the publicly available survey fits maps ( i.e. vlss , wenss , first , nvss and gb2 ) , we analyze available raw unpublished data from the vla archive ( these data are marked with number ` 7 ' in col .
10 of table2 ) , reducing them using standard procedures with the aips package . after preliminary cleaning of the maps with the routine imagr
, several iterations of self - calibration were performed to improve the maps quality .
finally , the maps were corrected for primary beam attenuation using the pbcor task .
all the resulting flux densities are given in the @xcite scale .
the flux densities obtained directly from the fits maps ( either the maps made for the purpose of this investigation , or the maps taken from the surveys ) with the astronomical image processing system ( aips ) task tvstat or jmfit ( in the case of a point source ) are assumed to have the absolute flux calibration errors of 5% ; the flux densities taken from the literature are assumed to have the errors as determined in the corresponding references . in the case of the flux density values which had not been measured directly from the fits maps ( e.g. those of the extended structure ) but estimated by other means , we propagate the errors of the directly measured flux densities .
the resulting flux densities ( along with the errors ) of the quasar and of the total structure ( i.e. , quasar @xmath17 extended radio tail ) are listed in table2 .
the quasar flux densities are peak and integrated values , in order to clarify any possible contamination from the diffuse emission .
the flux density errors of the quasar are calculated by adding the flux measurement errors and the jmfit errors in quadrature .
the multi - frequency radio data collected here can be divided into three categories : i ) those which resolve the quasar core as an isolated source and give proper values of its flux - density ( see table2 , col .
7 and 8) , ii ) those which give a proper overall image and flux - densities of the entire ` quasar + extended halo ' structure but have little ability to distinguish individual components as distinct sources ( see table2 , col .
9 ) , and iii ) those which have adequate angular resolution to separate and image properly both the quasar core and the extended features but may underestimate the flux densities of the extended components ( vla - b 1400mhz and vla - c 1425mhz observations ) . the former class of data ( i ) were used to determine the radio spectrum of the quasar and to examine whether the source is variable or not ( for detail see 4.1 ) .
the characterized shape of the quasar continuum allowed us next to subtract the contribution of the quasar core from the total ` quasar + extended halo ' fluxes evaluated using the second ( ii ) class of data . in this way we obtained the total spectrum of the extended structure .
the latter class of data ( iii ) allowed us finally to depict morphological details of the extended structure .
furthermore , the vlba data were used to investigate the small - scale structure of the quasar core ( for details see 5.1 ) .
figure[fig : perley ] presents our vla image at 1425mhz of the entire radio source in abell585 ( the largest angular extent of about 25 ) with an angular resolution of @xmath315 along with the @xmath35 angular resolution 1400mhz first map .
the diffuse feature seen in the nvss map is now clearly resolved into two cometary - like structures with bright compact ` heads ' in their northern parts ( hereafter the eastern one ` e ' and the western one ` w ' ) , and edge dimmed tails elongated to the south .
the high - frequency ( e.g. , 8439.9mhz ) maps allow us to measure precisely the positions of the heads flux maxima as [ @xmath11@xmath40 , @xmath13@xmath41 and [ @xmath11@xmath42 , @xmath13@xmath43 .
there are no optical counterparts at these positions in the dss and sdss maps . in order to measure the total flux density of the extended emission component ( the two cometary - like structures ) ,
we first measure the flux densities of the quasar core at different frequencies , and fit the spectrum with a single power - law function , which returns @xmath44 = - ( 0.077 \pm0.039 ) \times \log [ \nu/{\rm mhz } ] + ( 2.829 \pm 0.137)$ ] .
next we subtract the fitted quasar spectrum at a given frequency from the integrated flux of the entire structure , obtaining in this way the spectrum of the extended structure only , @xmath44 = - ( 0.408 \pm0.043 ) \times \log [ \nu/{\rm mhz } ] + ( 3.520 \pm 0.125)$ ] .
this estimate implies a total flux density of the extended structure at 1425mhz of about 171mjy , which is consistent with the total flux density measured in the vla c array map at 1425mhz for the western and eastern structures separately ( see figure[fig : perley ] , thick contours ) , i.e. @xmath45 and @xmath46mjy , respectively .
a small disagreement here may be due to the fact that , as indicated by the multi - epoch radio data , the b30727 + 409 core radio emission is variable , decreasing with time over the last decades , and this trend is particularly clear at 1.4 , 5 and 8ghz .
a spectral curvature for the extended component ( see the next section ) may also play a role .
+ the nvss and wenss images which have similar beam sizes and are sensitive to extended ( up to a few arcmin ) diffuse structures were used to create a spectral index map of abell585 .
first , using the aips task sad we subtracted the quasar b30727 + 409 from both maps .
subsequently , the higher resolution map was convolved to the resolution of the 327mhz map . since misalignment of the total - power map at two frequencies
could produce systematic errors in the spectral - index map , we co - registered the positions of several bright point - like field sources which surround the target source on both maps .
further , both maps were brought to a common scale using the aips task hgeom .
finally , the spectral index map was obtained using the aips task comb .
regions with flux density values below @xmath47rms were considered to be unreliable and blanked .
the final grey - scale map of the spectral index is shown in figure[fig : pol](a ) .
the map indicates that there is a monotonic steepening of the radio spectrum of the diffuse component from north to south , i.e. from compact ` heads ' of the two resolved structures along their tails .
the trend is not that clear in the case of the eastern tail because of an imperfect subtraction of a point source present around the tail s southern edge .
such a steepening is a classic spectral signature of an aging synchrotron - emitting plasma in the lobes and plums of radio galaxies , observed in almost all luminous classical doubles ( for which the flattest radio spectra are seen at the positions of terminal hotspots , where the injection of fresh particles takes place ; e.g. , @xcite ) and also low - power radio galaxies including tailed sources in galaxy clusters ( for which which the flattest radio spectra are seen around the jet base ; e.g. , @xcite ) .
the radio spectrum of the quasar core is inverted , as expected , with the 327 - 1400mhz spectral index of about @xmath48 ( defined here as @xmath49 for the flux spectral density @xmath50 ) .
the mean spectral index over the extended structure is characteristic for lobes and plumes of radio galaxies in general , @xmath51 , and implies a relatively young population of the radio - emitting electrons ( @xmath52myr ) .
we note in this context that rather steep radio spectra seen at the south - western outskirts of the structure ( @xmath53 ) may be due to a missing flux in the 1400mhz interferometric map only .
finally , using the vla 4710.1 and 8439.9mhz maps we evaluate the spectral index of the eastern ` head ' as @xmath54 .
maps of the linearly polarized intensity and fractional polarization were made by combining the nvss stokes @xmath55 and @xmath56 maps with the aips task comb .
this allows for the determination of the polarized flux density , the fractional polarization , and the @xmath57-vectors polarization angle .
the total intensity nvss map with the electric field @xmath57-vectors superimposed is shown in figure[fig : pol](b ) , while figure[fig : pol](c ) presents the linearly polarized intensity map with the vectors of fractional linear polarization superimposed . from these maps
it can be seen that the extended structure is mildly polarized .
the total integrated polarized flux intensity of the extended source is @xmath58mjy , which gives about 5% for the mean fractional polarization .
the integrated polarized flux density of the compact quasar core is @xmath59mjy , and the corresponding degree of linear polarization is therefore about 1% .
we found different orientations of the electric vectors in different regions of the structure , as well as changes in the levels of polarization across the source .
the polarization percentage increasing along the tails ( i.e. , away from spots w and e to the south ) , resembles typical polarization properties of tailed radio galaxies ( e.g. , @xcite ) .
the polarization @xmath57-vectors , however , are mostly perpendicular to the tails .
this indicates that the magnetic field vectors , in the case of negligible rotation measure ( rm ) values , would be transverse to the resolved structures , unlike in other tailed radio galaxies .
such an ` unusual ' property may on the other hand simply suggest a substantial rotation of the polarization plane inside the cluster abell585 . in order to determine the real distribution of the magnetic @xmath60-field directions ,
it is necessary to correct the radio data for the faraday rotation , and for that purpose follow - up radio polarization measurements at other wavelengths with high angular resolution would be needed .
the rm of the quasar is @xmath61radm@xmath62 @xcite .
@xcite claimed a comparable value of @xmath63radm@xmath62 for the polarized radio feature located north - west to the quasar ; this feature coincides with the diffuse tail extending from our spot / head e. the provided value has to be taken with caution , however , since it was estimated using only two nearby frequencies of 1365mhz and 1435mhz . ) . in several places of the figure
spectral index values are given explicitly .
* b ) * total intensity 1400mhz contours spaced by a factor of 2 starting at 1.35mjybeam@xmath8 , with the superimposed @xmath57-vectors , for which the length is proportional to the polarized intensity ( 10 corresponds to 5.5mjybeam@xmath8 ) .
* c ) * linearly polarized 1400mhz intensity contours spaced by a factor of 2 starting from 0.9mjybeam@xmath8 , with the vectors of the fractional linear polarization superimposed ( 10 corresponds to 28% of the fractional linear polarization ) .
the size of the beam is indicated by the circles in the left upper corners of the images . ]
b30727 + 409 is a luminous radio source with the 1400mhz power of @xmath64whz@xmath8 .
its sdss spectrum reduced through the standard procedure ( see @xcite ) allows us to estimate the mass of the central black hole @xmath65 when using the mgii line , and @xmath66 when using the civ line .
the resulting value is characteristic for radio - loud quasars at high redshifts ( e.g. , @xcite ) .
also the bolometric luminosity of the accreting matter in b30727 + 409 estimated from the mgii line is typical for quasars in general , namely @xmath67ergs@xmath8 , implying high accretion rate @xmath68 ( in the eddington units ) for the standard @xmath69 radiative efficiency of the accretion disk .
as mentioned previously , b30727 + 409 is a vla phase calibrator , but it was also a vlba phase calibrator .
the existing vlba maps show one - sided nuclear jet oriented toward the extended structure under investigation .
interestingly , the position angles of the nuclear jet and of the extended structure are similar : the former one is 127 , 125 , 123 at , respectively , 2268.3 , 4995.5 and 8338.3mhz , while the position angle of the extended structure as measured in the first map between the quasar core and the peak of the radio emission in the western ` head ' [ @xmath11@xmath70 , @xmath13@xmath71 is 134 .
this agreement may suggest that the extended radio structure observed around the central region of abell585 is not related to the cluster , but instead represents the large - scale jet / lobe of the background quasar b30727 + 409 , as in fact proposed in @xcite .
additionally in such a case , the x - ray emission seen in the rosat data likely extend along the radio structure , and could possibly be accounted for by the jet emission as well since large - scale quasar jets are established luminous x - ray emitters ( see @xcite for a review ) . the angular distance between the quasar core and the western radio ` head ' is 272 , which , if the radio structure is at the quasar redshift indeed , corresponds to the linear size of 1.3mpc .
the de - projected physical size of the large - scale radio structure of b30727 + 409 would be in such a case @xmath72mpc , for the anticipated quasar viewing angle @xmath73and the assumed source intrinsic symmetry .
in general , lobes with linear sizes exceeding mpc , called ` giant radio sources ' , are known to be associated with radio galaxies and quasars even at high redshifts ( see @xcite ) ; the largest object of this kind known to date is characterized by the projected size of about 4.7 mpc ( j1420@xmath740545 ; @xcite ) . on the other hand , morphological , spectral , and polarization properties of giant radio sources are very different from those of the extended radio structure revealed by the high - resolution radio maps discussed in this paper .
moreover , in the framework of the quasar scenario , the source energetics would be rather extreme , as discussed below .
the 1400mhz flux of the extended structure @xmath75mjy ( see 4.1 ) , if produced at the redshift of @xmath2 , gives the monochromatic radio luminosity @xmath76ergs@xmath8 .
since only a moderate beaming for quasar radio jets is expected on large scales , with the corresponding jet bulk lorentz factors @xmath77 @xcite , the total emitted radio power of the source would therefore read as @xmath78ergs@xmath8 for the bolometric correction factor @xmath79 ( following from the assumption that the jet radio continuum extends from @xmath80ghz down to mhz frequencies with the average spectral index of @xmath81 ) .
this emitted radio power would then require the total jet kinetic power @xmath82 in excess of @xmath83ergs@xmath8 ( assuming the standard @xmath84 jet radiative efficiency ) , which is close to the maximum rate of the smbh energy extraction in the system , @xmath85ergs@xmath8 @xcite .
this power would have to be continuously supplied by the matter accretion for at least @xmath86myr , in order to power the @xmath87mpc - scale flat - spectrum radio cavity , and this is assuming a rather high jet advance velocity of @xmath88 ( see the related discussion in @xcite ) .
the aforementioned x - ray emission is in this context even more problematic , since the observed x - ray flux @xmath89ergcm@xmath62s@xmath8 , if due to the quasar jet , implies a very high luminosity @xmath91ergs@xmath8 .
note that any significant beaming which could be invoked to reduce the total emitted power of a jet , would require at the same time very small jet viewing angle , and hence the de - projected linear size of the structure exceeding 10mpc .
for all these reasons , we consider the quasar scenario for the analyzed extended radio structure implausible . at the redshift of the galaxy cluster abell585
, the observed 1400mhz flux of the extended radio structure corresponds to the radio luminosity of @xmath92ergs@xmath8 , which is in the range of radio luminosities for typical cluster radio halos @xcite .
however , cometary - like morphologies of two distinct features with clearly defined ` hotspots ' or ` heads ' revealed on the high - resolution radio maps , their relatively small sizes @xmath93kpc ( if at @xmath94 ) , @xmath95 radio polarization , and relatively flat radio spectra ( @xmath81 ) , are all in a disagreement with such an association . as we mentioned in 2 , abell585 is a member of the scl074 supercluster which may represent a dynamic merging structure .
the energy dissipated in large - scale shocks during cluster - cluster interactions can be channeled into relativistic electrons and magnetic field ( e.g. @xcite ) , resulting in a formation of extended radio structures similar to those seen in e.g. abell85 ( * ? ? ?
* see 6 below ) . yet
the analyzed radio features can be hardly identified with any sort of radio relics , since those elongated or even filamentary radio structures with sizes @xmath96mpc are found mostly at the outskirts of galaxy clusters , and , even though polarized , are characterized by very steep radio spectra ( @xmath97 ; see , e.g. , @xcite ) .
radio - loud agn typically found in clusters of galaxies are low - power radio galaxies , whose jets and lobes are often distorted and bent , with the characteristic u- or c - shape large - scale morphologies .
such sources are denoted as narrow - angle tail ( nat ) and wide - angle tail ( wat ) radio galaxies , respectively .
the dramatic bending of the jets is due to a combination of a parent galaxy motion within the cluster , and local pressure gradients in the icm . for the early comprehensive and statistical study of nat sources
see , e.g. , @xcite and @xcite .
the two radio features resolved in our high - resolution radio maps , and especially the western one , do resemble nat sources taking into account their morphologies , location around the central parts of a dynamical cluster , radio powers ( @xmath98ergs@xmath8 each ) , and finally linear sizes of the order of hundreds of kpc .
the main problem with this interpretation is the lack of any obvious galaxies which could be identified as hosts of the two radio structures : as already mentioned above , no optical counterparts can be found in the sdss data at the positions of the two radio ` heads ' , which presumably would mark the positions of any putative radio cores . moreover , none of the galaxies located within the central parts of the cluster abell585 in close proximity to the radio features reveal any signatures of the agn activity ( see table1 ) .
we have verified this finding by a careful examination of the all - sky mid - infrared ( mir ) data provided by the nasa wide - field infrared survey explorer satellite ( wise ; @xcite ) gaussian ; nominal 5@xmath99 point source sensitivity from @xmath30.08 up to 6mjy ; see also ` http://wise2.ipac.caltech.edu/docs/release/allsky ` ] in the range @xmath100@xmath101 m .
indeed , all the mir emitters found in the region of interest correspond to the optically - detected galaxies and the quasar b30727 + 409 , and therefore we can exclude a possibility for the presence of heavily - obscured agn at the positions of the radio cores of the two analyzed radio features . but
can the two resolved radio features be associated with background radio galaxies located at redshifts high enough that their hosts are simply below the flux limit of the sdss ? using the @xmath21-band hubble diagram for 3c galaxies given in @xcite , and the detection limit in the @xmath21-band sdss images of @xmath102 @xcite , we estimate the minimum redshift of putative parent galaxies in this scenario as @xmath103 . at such high redshifts ,
the angular size of the radio tails ( @xmath104 ) would correspond to the linear size of @xmath105mpc .
keeping in mind that the largest head - tail radio galaxy known to date , peculiar 3c129 at @xmath106 , is only twice as long @xcite , we consider the possibility of two neighboring distant giant tailed radio sources to be very low .
the two cluster member galaxies closest to the cores of the analyzed radio structures with projected separations of about @xmath107kpc are the giant but otherwise standard elliptical with no prominent emission lines , and the bright face - on spiral with bright optical core ( lacking however any spectroscopic redshift determination ) , designated respectively as ` p ' and ` s ' in figure[fig : perley ] ( see also table1 ) .
let us therefore consider a scenario in which the radio structures in question are due to the jet activity of smbhs which experienced gravitational wave recoil kicks in galaxy merging process within the central parts of abell585 .
the gravitational recoil of merging black holes is widely considered in the literature , typically in the context of the cosmological evolution of smbhs and the galaxy formation ( e.g. , @xcite ) .
candidate agn related to this process have been identified based on peculiarities in their line emission @xcite , or the presence of double nuclei separated by projected distances of a few up to several kpc @xcite . if the gravitational recoil process is of any relevance for the systems analyzed in this paper , the recoil kicks have to be strong enough for a smbh to be able to escape the gravitational potential of a massive host galaxy ( see in this context @xcite ) .
recent numerical simulations indicate that such large recoil velocities exceeding 1000kms@xmath8 are possible under specific conditions of rapidly spinning black holes with comparable masses and a particular orientation of their spin vectors at small angles to the orbital plane ( e.g. , @xcite ) . and in fact dry mergers of massive ellipticals in central parts of galaxy clusters may provide the required conditions , since such systems are poor in cold gaseous content ( while this is the abundant cold gas which is believed to enable an efficient alignment of black hole spins perpendicular to the orbital plane during the merger ; see @xcite ) , but on the other hand to contain rapidly spinning smbhs on average ( see in this context , e.g. , @xcite ) .
one could speculate further that the recoiled smbhs traveling through the icm after leaving host galaxies accrete only the hot intracluster gas at low ( bondi ) rates , and as such form radiatively inefficient accretion disks ; this could potentially explain the apparent lack of any agn activity at mir - optical wavelengths at the position of the cores of the analyzed radio structures .
yet , on the other hand , the probability of having two major mergers in the central parts of a single cluster , within the period overlapping with radiative lifetimes of the two radio galaxies ( @xmath52myr ; see 4.2 ) , is very low .
there are also no obvious ` remnant ' pairs of elliptical hosts which should be expected around the positions of the e and w spots in the framework of the gravitational recoil scenario : as emphasized above , only one giant elliptical and one bright spiral are detected within the radius of @xmath108kpc from the radio cores .
note in this context that powering the lobes with the observed radio luminosities of @xmath98ergs@xmath8 requires jet kinetic luminosities of the order of , at least , @xmath109ergs@xmath8 , and therefore very high black hole masses for the anticipated low / moderate accretion rates .
hence any association of the discussed radio structures with some low - mass ( and as such undetected ) hosts seems implausible . in order to justify this statement
, we note that with the bondi accretion rate @xmath110 and the following maximum jet kinetic luminosity @xmath111 , one obtains the limiting black hole mass @xmath112 where @xmath113@xmath114 , @xmath115k , and @xmath116 are respectively the gas number density , temperature , and sound velocity expected for the central parts of a galaxy cluster .
the peculiar arcmin - scale radio structure observed in the direction of the central parts of the galaxy cluster abell585 ( @xmath0 ) is unassociated with the distant ( @xmath2 ) radio quasar b30727 + 409 , as we demonstrated here by means of a careful analysis of all the available radio data for the system .
this structure consists of two dominant features which resemble typical lobes of cluster radio galaxies of the nat type .
however , we are not able to identify host galaxies of the two features in the available optical ( sdss ) and infrared ( wise ) surveys .
we speculate if the analyzed systems are examples of extreme gravitational recoils of smbhs in the process of merging massive ellipticals within the central parts of a rich cluster environment , but we do not reach any robust conclusions regarding the origin of the detected radio features .
large - scale radio structures lacking obvious host galaxies may not be that unique for the local clusters .
for example , radio maps of b1753@xmath17580 found at the position of abell2289 @xcite show two edge - dimmed tails with bright ` heads ' and with a @xmath117 galaxy located between them @xcite , but without any detectable compact radio core .
@xcite presented a detailed study of the radio structure in cluster abell4038 resembling the evolved lobes of a luminous radio galaxy , but with the most likely host elliptical galaxy displaced by about 20kpc with respect to the central parts of the radio structure .
@xcite discussed three other possibly analogous systems in clusters abell13 , abell85 , and abell133 . inspecting first maps of several galaxy clusters from @xcite , we add the case of max bcg j130.34220 + 61.21246 ( @xmath118 ) to the literature examples , and
note the striking similarity between the radio structures discussed in this paper and those in the @xmath119 cluster maxbcg j250.34552 + 38.03597 that lack any optical identifications .
there may be more examples of such peculiar structures awaiting discovery with high - resolution multiwavelength radio observations of lesser - known clusters of galaxies in formation .
we thank the anonymous reviewer for comments which have significantly improved the paper . this research has made use of the nasa / ipac extragalactic database ( ned ) , which is operated by the jet propulsion laboratory , caltech , under contract with the national aeronautics and space administration .
we acknowledge use of the sloan digitized sky survey . .s .
, m.o . and m.j .
are supported by polish nsc grants dec-2012/04/a
/ st9/00083 and dec-2013/09/b / st9/00599 , respectively .
work by c.c.c . at nrl
is supported in part by nasa dpr s-15633-y .
|
in this paper we analyze the peculiar radio structure observed across the central region of the galaxy cluster abell585 ( @xmath0 ) . in the low - resolution radio maps ,
this structure appears uniform and diffuse on angular scales of @xmath1 , and is seemingly related to the distant ( @xmath2 ) radio quasar b30727 + 409 rather than to the cluster itself .
however , after a careful investigation of the unpublished archival radio data with better angular resolution , we resolve the structure into two distinct arcmin - scale features , which resemble typical lobes of cluster radio galaxies with no obvious connection to the background quasar .
we support this conclusion by examining the spectral and polarization properties of the features , demonstrating in addition that the analyzed structure can hardly be associated with any sort of a radio mini - halo or relics of the cluster . yet at the same time we are not able to identify host galaxies of the radio lobes in the available optical and infrared surveys .
we consider some speculative explanations for our findings , including gravitational wave recoil kicks of smbhs responsible for the lobes formation in the process of merging massive ellipticals within the central parts of a rich cluster environment , but we do not reach any robust conclusions regarding the origin of the detected radio features .
[ firstpage ] radiation mechanisms : non - thermal galaxies : active galaxies : clusters : individual : abell585 galaxies : jets quasars : individual : b30727 + 409 radio continuum : galaxies
|
# 1 ps . @headings oddfootevenfoot oddhead evenheadoddhead # # 1 ps .
@headings = 12 # 1@xmath4
= 11 # 1 ps . @headings oddfootevenfoot oddhead evenheadoddhead # # 1 ps .
@headings = 12
= 11 = 12 # 1#2#3_phys .
lett . _ * # 1*(19#2)#3 # 1#2#3_z .
phys . _ * # 1*(19#2)#3 # 1#2#3_phys .
lett . _ * # 1*(19#2)#3 # 1#2#3_rev .
phys . _ * # 1*(19#2)#3 # 1#2#3_phys .
rep . _ * # 1*(19#2)#3 # 1#2#3_phys .
rev . _ * # 1*(19#2)#3 # 1#2#3_nucl .
_ * # 1*(19#2)#3 # 1#2#3_sov
. j. nucl .
phys . _ * # 1*(19#2)#3 # 1#2#3_acta phys .
polon . _ * # 1*(19#2)#3 # 1#2#3_ann .
phys . _ * # 1*(19#2)#3 # 1#2#3#4#5#6 1.0 cm * differential distributions for * * heavy flavour production at hera * .6 cm * stefano frixione * .2 cm dip .
di fisica , universit di genova , genoa , italy + .5 cm * paolo nason * .2 cm cern th - division , ch-1211 geneva 23 , switzerland .5 cm * giovanni ridolfi * .2 cm infn , sezione di genova , genoa , italy .
cern - th/95 - 143/95 may 1995
experimental results on charm photoproduction at hera have recently become available [ [ zeuscharm ] ] .
it is likely that more detailed results on the differential distributions will appear in the near future . with respect to previous photoproduction experiments ( ref .
[ [ fixedtargetphotoproduction ] ] ) hera offers the new opportunity of a higher energy regime . a theoretical study of the total photoproduction cross section has already been given in ref .
[ [ fmnr_tot ] ] . in the present work
we extend the analysis of ref .
[ [ fmnr_tot ] ] ( to which we refer the reader for a general introduction and for notation ) by considering single inclusive distributions for heavy flavour production at hera . our analysis is based on the next - to - leading order calculation of heavy - quark photoproduction and hadroproduction cross sections performed in refs .
[ [ ellisnason],[nde ] ] , as implemented in a computer program developed in refs . [
[ mnr],[fmnrphoto ] ] .
we will consider cross sections at fixed photon energies .
when necessary , a discussion of the energy dependence of the distributions will be given .
we will also consider electroproduction for small photon virtuality , in the weizscker - williams approximation , which is appropriate for the bulk of the photoproduction cross section .
we will not consider the production of heavy flavours in deeply inelastic events , i.e. events in which high photon virtuality is required ( see refs . [ [ virtualphoton ] ] ) .
the paper is organized as follows . in section 2
we study the point - like contribution of the transverse momentum and pseudorapidity distributions in photoproduction .
section 3 is devoted to the study of the hadronic component , considering the possibility of separating it from the point - like component with appropriate cuts . in sections 4 and 5
we consider electroproduction of charm and bottom , in the weizscker - williams approximation . in section 6
we compare our fixed - order results with those of the monte carlo herwig [ [ herwig ] ] , and in section 7 we give our conclusions .
some technical details on the factorization schemes , the scale dependence , and the modified weizscker - williams approximation used here are given in the appendix .
unless specifically stated , we will use in the following the set of parton densities mrsa , ref . [ [ mrsa ] ] , with @xmath5mev .
this set of distribution functions has been recently updated ( mrsg , ref . [ [ mrsg ] ] ) , to include new hera deep inelastic scattering data , allowing a different small-@xmath6 behaviour of the gluon and sea quark densities .
we have checked that the shapes of the single - inclusive distributions we are considering are not significantly different when the new parametrization is used .
the default value of the charm quark mass will be @xmath7gev .
the renormalization scale will be taken as @xmath8 , and the factorization scale for the proton and for the photon will be taken as @xmath9 , where @xmath10 .
we begin by showing the transverse momentum distribution for the charm quark in fig .
[ pt286pnt ] , for @xmath11 gev ( the proton energy @xmath12 will be fixed in the following to be 820 gev ) .
we also show the effect of applying the peterson fragmentation function [ [ pete ] ] d(x)= to the final state quark .
we use the value @xmath13 , which is the central value quoted in ref .
[ [ chrin ] ] for charm quarks .
as can be expected this softens considerably the @xmath14 spectrum . as discussed in ref .
[ [ fmnrft ] ] , in the case of fixed target photoproduction experiments , the inclusion of peterson fragmentation correctly reproduces the shape of the measured @xmath14 distribution .
the effect of a pseudorapidity cut similar to the one applied by the zeus collaboration [ [ zeuscharm ] ] is also shown . at low to moderate @xmath14
this cut considerably lowers the cross section , while at higher @xmath14 it has a negligible effect . in fig .
[ ptkickpnt ] we compare the @xmath14 distributions at different photon energies .
we also illustrate the effect of applying an intrinsic transverse momentum @xmath15 to the incoming parton ( see ref .
[ [ fmnrft ] ] ) .
we find that even with the very large value of @xmath16gev@xmath17 the effect is small .
therefore , in the following , we will neglect the effect of an intrinsic transverse momentum of the incoming partons .
pseudorapidity distributions for charm are shown in fig .
[ eta286pnt ] for @xmath11 gev .
we observe that the point - like contribution to the cross section at this energy strongly favours large negative pseudorapidities .
a cut on the transverse momentum of the produced quark tends to move the distribution towards the central region .
the dotted curve shows the effect of fragmentation in the presence of a transverse momentum cut .
the fragmentation has little effect on the pseudorapidity of the quark , but degrades its transverse momentum , so that the @xmath14 cut has a stronger effect .
the effect of fragmentation without transverse momentum cut is not shown in the figure , since it is not well defined ( see ref .
[ [ fmnrft ] ] ) . at lower energies ( fig .
[ eta99pnt ] ) the pseudorapidity distributions become more central .
we now turn to the sensitivity of our distributions to the various parameters that enter the computation .
we have studied the dependence of our distributions upon the charm quark mass , by varying it between @xmath18 and @xmath19 gev , on the renormalization scale , which was varied by a factor of 2 below and above its default value , and upon the distribution functions , by considering the mrsd@xmath20 [ [ mrsdmeno ] ] and the cteq2mf [ [ cteq2mf ] ] sets .
the proton parton densities were chosen in order to span the allowed range for the small-@xmath6 behaviour of the gluon density , mrsd@xmath20 being the most singular , and cteq2mf the least singular one ( see ref .
[ [ fmnr_tot ] ] for a discussion of this point ) . studying the variation of our results with respect to @xmath21 is a difficult problem , since the values of @xmath21 extracted from fits to deep inelastic scattering data are not in good agreement with the lep value .
the cteq group has provided a parton density set in which the value of @xmath21 was pushed as high as possible ( see also ref .
[ [ grv_lam ] ] ) .
they use a value of @xmath22mev , which is still below the lep value , but adequate for a study of the sensitivity of the distributions .
the set of parton densities obtained in this fit ( cteq2ml ) was used in association with this value of @xmath21 .
we found that the only cases in which there is a noticeable shape variation is in the @xmath14 distributions when we vary the mass and the parton densities .
the relevant plots are given in figs .
[ pt286pnt_mvar ] and [ pt286pnt_strfvar ] . from the figures we can see a comparable variation of the @xmath14 spectrum in the low momentum region in the two cases .
the mass dependence is easily understood .
we expect nearly no mass dependence at high transverse momenta , while in the massless limit the cross section diverges at small momenta .
therefore , the smaller the mass , the higher the cross section at low @xmath14 .
as far as the distribution function dependence is concerned , the small-@xmath6 uncertainty in the densities plays here a major rle . the more singular the small-@xmath6 behaviour , the higher the cross section at small @xmath23 . as a last point
, we observe that , when measuring the total cross section , if a transverse momentum cut on the @xmath14 spectrum is being applied , the extrapolated total cross section will depend upon the assumptions we make about the value of the mass and the small-@xmath6 behaviour of the distribution functions .
this has to be properly taken into account when inferring properties of the parton density at small @xmath6 from the energy behaviour of the total charm cross section .
the point - like contribution of the charm cross sections will be contaminated by the hadronic component .
depending upon the chosen distribution functions for the photon , the contribution of the hadronic component to the total cross section may dominate over the point - like one . whatever the choice of the distribution functions , however , the two components differ remarkably in the pseudorapidity distribution . the point - like component , as we have seen , favours negative pseudorapidities , while the hadronic component favours the positive direction .
the separation of the two components will therefore require an analysis of a large pseudorapidity range . in the following
, we will show a study of the full ( i.e. point - like plus hadronic ) pseudorapidity distribution for different choices of photon and proton distribution functions .
we will see that the shape and magnitude of the @xmath24 distribution will allow us to distinguish between the most extreme parametrizations of photon distribution functions . in fig .
[ pdf_lac1_sum_286 ] we show the @xmath24 distribution , obtained with the lac1 [ [ lac1 ] ] set of photon parton densities , and three different sets of proton parton densities ; both the point - like and the full results are presented . a similar plot for the grv [ [ grvph ] ] set for the photon is given in fig . [ pdf_grv_sum_286 ] .
we have chosen the lac1 and grv sets because , for our purposes , they represent the two extreme possibilities , as discussed in ref .
[ [ fmnr_tot ] ] . in figs .
[ pdf_lac1_sum_286 ] and [ pdf_grv_sum_286 ] we have imposed a realistic transverse momentum cut , which in general will make the pseudorapidity distribution even narrower . in spite of the cut
, the separation of the two contributions is quite apparent .
the study of the pseudorapidity distribution could help in distinguishing among different proton parton density sets , especially if the large negative pseudorapidity region could be explored . on the other hand ,
the large difference in shape induced by the two photon sets considered has measurable effects even in the central region .
the full transverse momentum distribution is presented in fig .
[ pt286full ] .
the point - like and hadronic contributions are also separately shown . from the figure
, we can see that the hadronic contribution is dominant in the low-@xmath14 region . at @xmath25 gev , it amounts to 50% of the full cross section . therefore , even without investigating the very low-@xmath14 region , the softer behaviour of the full cross section with respect to the point - like one should be visible . obviously , this behaviour is strongly influenced by the photon parton densities used .
if the grv set were used , instead of the lac1 set , the full distribution would differ only sligthly from the point - like one already at moderate @xmath14 .
we can conclude that the charm transverse momentum spectrum at large values of the photon energy could help in further constraining the parton densities in the photon . from fig .
[ pdf_lac1_sum_286 ] , we also learn that asymmetric pseudorapidity cuts may enhance or suppress the effect of the hadronic component in the @xmath14 spectrum , thus providing another handle for its study .
in this section we discuss single - inclusive distributions for charm production in electron - proton collisions in the weizscker - williams approximation .
we take the electron and proton energies to be @xmath26 gev and @xmath27 gev , respectively . since the weizscker - williams function peaks in the small-@xmath6 region
, the bulk of the contribution to the cross section will be due to photons of relatively low energy .
therefore , the rle of the hadronic component is less important with respect to the monochromatic photon case . by applying a small-@xmath14 cut ,
the contribution of the hadronic component is even more suppressed , especially for soft photon parton densities .
therefore , the qcd predictions at next - to - leading order are less affected by the uncertainty originating from the photon distribution functions . because of the softness of the incoming photons , the pseudorapidity distribution of the point - like component is concentrated in the central region .
this can be seen in fig .
[ el_eta_all ] , where we show the pseudorapidity distribution for charm quarks , supplemented with peterson fragmentation and a transverse momentum cut .
the hadronic component is computed using our two extreme sets of photon parton densities .
it is apparent that its effect is small in the central pseudorapidity region explored at hera . the curves in fig .
[ el_eta_all ] are obtained with @xmath28 gev , @xmath8 , @xmath9 and the mrsa set for proton parton densities . while the total cross section is quite sensitive to the value of these parameters , as shown in ref .
[ [ fmnr_tot ] ] , the shape of the single - inclusive distributions is rather stable .
the pseudorapidity distribution shape is only mildly dependent upon the choice of proton parton densities .
this is due to the fact that the various parametrizations differ in the small-@xmath6 region , which is not deeply probed in the electroproduction process .
this is shown , for the point - like component , in fig .
[ el_eta_pdf ] .
the proton parton density dependence of the hadronic component is completely negligible , due to the softness of the partons in the electron .
we now consider the transverse momentum distribution . in fig .
[ el_pt_pnt ] we show the point - like contribution to this differential cross section , with and without peterson fragmentation
. the softening of the distribution due to fragmentation should be observable at hera .
the effect of applying a pseudorapidity cut is also shown : it affects the shape of the distribution less dramatically than in the case discussed in section 2 .
we have checked that the hadronic component contribution to this distribution is remarkably softer than the point - like one . in practice ,
its effect is less than 10% for @xmath29 gev when fragmentation is included and the pseudorapidity cut @xmath30 is applied .
we have also computed the same distribution for different values of the renormalization and factorization scales , and we found that the corresponding shape variations are small .
all the distributions presented so far were also evaluated in the case when an antitag condition on the outgoing electron is applied , that is to say when the electron scattering angle is below a given value , which we chose to be 5 mrad . for this computation , the improved form of the weizscker - williams distribution presented in ref .
[ [ fmnrww ] ] is appropriate ( see the appendix ) .
differences in the shapes of the distributions are found to be small . as an example , we show in fig . [ el_eta_cut_pnt ] the pseudorapidity distribution with and without the antitag condition described above . the antitag condition has the effect of decreasing the total cross section , as described in ref .
[ [ fmnr_tot ] ] , of slightly enhancing the contribution of positive pseudorapidities , and of mildly softening the transverse momentum spectrum .
we can conclude that charm electroproduction distributions are rather insensitive to the choice of the parameters entering the perturbative calculation .
therefore , single - inclusive charm electroproduction will be of little help in constraining the parton distribution functions of both the proton and the photon
. on the other hand , the comparison between data and theoretical predictions will be useful for the study of the production mechanism .
due to the higher value of the quark mass , perturbative qcd predictions for bottom production are more reliable than those for charm . furthermore , we are able to study the scale dependence in an exhaustive manner .
in fact , the factorization scale , when varied by a factor of two below and above its default value , is always higher than the mininum allowed by the distribution function parametrizations . in fig .
[ b_el_eta_var ] we show the point - like component of the pseudorapidity distribution for bottom electroproduction .
the solid curve represents our central prediction , which corresponds to choosing @xmath31 gev , @xmath32 , where @xmath33 .
the qcd result is supplemented with peterson fragmentation with @xmath34 , the central value of ref .
[ [ chrin ] ] for bottom quarks . in the case of bottom
the effect of fragmentation is less important than in the case of charm .
the band between the two dashed curves has been obtained by varying the bottom quark mass between 4.5 and 5 gev and the renormalization and factorization scales between @xmath35 and @xmath36 .
the proton parton density set chosen is mrsa .
we also show ( dotted curve ) the result obtained with a parton density set fitted at a higher value of @xmath37 , namely the set cteq2ml with @xmath38 mev , and the central values of mass and scales .
it is difficult to foresee what a realistic transverse momentum cut could be for bottom production , and we have therefore kept the same cut @xmath39 gev we used in the charm case .
figure [ b_el_eta_var ] does not describe the uncertainties on the theoretical prediction completely , because the cteq2ml set was used keeping the scales and the mass at their default value , and because the hadronic contribution was not included .
the full uncertainty on the @xmath24 distribution is presented in fig .
[ b_el_eta_all ] , in which mass and scales are varied also when using the cteq2ml set , and the possibility is considered that the effect of the hadronic component is as large as the lac1 set implies , or is completely absent .
the dramatic effect of the inclusion of the hadronic component is apparent in the positive pseudorapidity region .
this is due to the very soft behaviour of the lac1 partons and to the stronger sensitivity of the hadronic component to the choice of mass and scales with respect to the one of the point - like component .
we now turn to the transverse momentum distribution , shown in fig .
[ b_el_pt_var ] , which is analogous to the corresponding figure for the @xmath24 distribution , fig .
[ b_el_eta_var ] .
as expected , in the large-@xmath14 region the sensitivity to mass and scales choice is strongly reduced .
therefore , the perturbative prediction becomes more reliable there . the full uncertainty on the @xmath14 distribution is given in fig .
[ b_el_pt_all ] , which is analogous to fig .
[ b_el_eta_all ] . from the figure , it is quite clear that even with the lac1 set the hadronic component affects the prediction only marginally .
this fact is a consequence of the applied pseudorapidity cut , as can be inferred from fig .
[ b_el_eta_all ] .
we can therefore regard fig .
[ b_el_pt_all ] as a reliable prediction of qcd for the @xmath14 spectrum of @xmath40 hadrons at hera .
in order to assess the influence of higher - order and non - perturbative contributions on the shape of single - inclusive distributions , we have compared our fixed - order calculations with the results obtained with the parton - shower monte carlo program herwig [ [ herwig ] ] . when using herwig we have always excluded the flavour excitation processes .
this is because comparable processes , such as gluon splitting , are difficult to include .
we have therefore preferred to use the monte carlo at a consistent level of accuracy . as discussed in section 2 ,
the effect of an intrinsic transverse momentum of the incoming partons is negligible in photoproduction at hera energies .
this leads us to expect that higher - order and non - perturbative corrections are not as important here as in the case of hadroproduction [ [ fmnrft ] ] . on the other hand , it is interesting to investigate how peterson fragmentation compares with the model of cluster hadronization implemented in herwig .
we limit our discussion to the point - like component of the photoproduction cross section for fixed photon energy . in fig .
[ pt_vs_hrwg ] we show the transverse momentum distribution computed in fixed - order perturbation theory and with herwig .
all herwig curves have been normalized to give the same total cross section of the corresponding next - to - leading order qcd predictions .
we observe that before the hadronization process is switched on , the qcd prediction and the herwig result agree almost perfectly in all the presented @xmath14 range ( upper curves ) .
we then consider the effect of hadronization in the central pseudorapidity region ( @xmath30 ) by comparing the perturbative prediction supplemented with peterson fragmentation and the herwig results for charmed hadrons or @xmath41 mesons only ( lower curves ) . in this case
the two predictions display a different behaviour only in the low-@xmath14 region .
this is also the only part of the @xmath14 spectrum that is significantly affected by the pseudorapidity cut .
the results in fig .
[ pt_vs_hrwg ] allow us to conclude that peterson fragmentation gives a description of the hadronization process , which is consistent with the one suggested by cluster models only at sufficiently large transverse momenta . a similar study for the pseudorapidity distribution is presented in fig .
[ eta_vs_hrwg ] . in this case
, some difference is observed between the two distributions before hadronization , the fixed - order qcd prediction being slightly broader .
however , they are peaked around the same value of @xmath24 , which is rather far from the central region , where the experimental acceptance is larger . applying a small transverse momentum cut ,
the use of peterson fragmentation becomes possible also for the pseudorapidity distribution .
we are therefore able to compare the qcd next - to - leading order distribution , convoluted with the peterson fragmentation function , and the herwig curves for charmed hadrons .
we find a complete agreement between the two .
we have performed the same kind of analysis on the hadronic component , and we found that the qualitative behaviour is completely analogous to what has been presented for the point - like component .
we have studied single - inclusive cross sections for charm and bottom production at hera .
we considered charm production in monochromatic photon - proton collisions , for energies spanning the range accessible at hera .
we found that the shape of the @xmath14 and pseudorapidity distributions are quite stable with respect to the variation of the input parameters entering the calculation . on the other hand , the total cross section value ( see ref .
[ [ fmnr_tot ] ] for a discussion on this point ) is strongly affected by the choice of parameters .
we found that from the study of the pseudorapidity distributions in the central @xmath24 region ( @xmath30 ) it is difficult to distinguish between the various proton parton densities with different small-@xmath6 behaviour . on the other hand ,
the striking difference between the predictions obtained by using the lac1 and grv set for photon distribution functions is clearly visible .
this fact is almost completely due to the very soft behaviour of the lac1 parton densities . the transverse momentum distribution for large @xmath14 is reliably predicted by qcd .
measurements in this region should constitute a good test of the production mechanism . at moderate @xmath14 ,
the point - like result is strongly influenced by the hadronic contribution when the lac1 set for photon parton densities is used .
experimental investigations in this region of the spectrum could therefore help to distinguish among the various parton distribution functions in the photon .
charm electroproduction is a less clean test of the production mechanism with respect to monochromatic photon production , because of the convolution with the weizscker - williams function . on the other hand ,
the softness of the latter implies a strongly reduced dependence of the total cross section upon the phenomenological parameters entering the calculation .
furthermore , when a small transverse momentum cut is applied , the hadronic contribution is more suppressed than the point - like one , whatever the photon densitites used .
therefore , electroproduction is less sensitive to the contamination of the hadronic component than photoproduction with a monochromatic photon , and can be used to further investigate the point - like production mechanism .
unfortunately , a sizeable dependence upon charm mass and renormalization scale remains in the total cross section prediction .
the case in which an antitag condition is applied on the outgoing electron was also studied . aside from the effect on the total cross section value ( which has already been discussed in ref .
[ [ fmnr_tot ] ] ) , the angular cut slightly enhances the importance of the central pseudorapidity region and mildly softens the transverse momentum spectrum .
we also considered the production of bottom quarks in @xmath42 collisions . due to the higher value of the quark mass , perturbative qcd predictions for bottom production
are more reliable than those for charm .
it turns out that the point - like component is quite stable in shape .
the hadronic component has a dramatic effect on the pseudorapidity distribution , although less important than in the case of charm .
its effect in the central @xmath24 region is however marginal .
we are therefore able to give a reliable prediction of the @xmath14 spectrum for @xmath40 hadrons in the central @xmath24 region . finally , we compared our results with those obtained using the parton shower monte carlo herwig .
we limited our discussion to the point - like component for monochromatic photon - proton collisions . as far as open charm quark
is concerned , we found perfect agreement for the @xmath14 distribution , in the whole range considered .
the next - to - leading order qcd prediction for pseudorapidity distribution is instead slightly broader than the one given by herwig in the region around the peak .
this difference is however immaterial in the central pseudorapidity region , which is explored by the experiments .
we also point out that , without any small transverse momentum cut , differential cross sections in longitudinal " quantities such as @xmath24 are indeed not characterized by a true hard scale . in this case ,
differences between perturbative qcd and parton shower approach are expected .
coming to charmed hadron production , we compared herwig prediction with qcd calculation supplemented with the peterson fragmentation function . a small-@xmath14 cut was applied .
the agreement is surprisingly good in the whole @xmath24 range , while the two predictions for transverse momentum distribution differ in shape only in the small-@xmath14 region .
we wish to thank g. iacobucci , m. mangano , j. roldan , m. seymour and b. webber for useful discussions .
a differential photoproduction cross section can be written as d^()(p_,p_h)= d^()_point(p_,p_h)+ d^()_hadr(p_,p_h ) , [ fullxsec ] where the quantities in the right - hand side of this equation are the so - called point - like ( or direct ) and hadronic ( or resolved ) photon cross sections . in qcd [ [ css ] ] , d_point^()(p_,p_h)&=&_jdx f^(h)_j(x,_f ) d_j(xp_h,_s(_r),_r,_f , _ ) [ pointcomp ] + d_hadr^()(p_,p_h ) & = & _ ijdx dy f^()_i(x , _ )
f^(h)_j(y,_f ) d_ij(xp_,yp_h,_s(_r),_r,_f , _ ) .
+ [ hadrcomp ] in the case of heavy quark production at next - to - leading order , the short distance partonic cross sections are given by d_j(p_1,p_2,_s(_r),_r,_f,_)&= & _ em_s(_r)d_j^(0)(p_1,p_2 ) + & & + _ em_s^2(_r)d_j^(1)(p_1,p_2,_r,_f , _ ) [ pointxsec ] + d_ij(p_1,p_2,_s(_r),_r,_f,_)&= & _ s^2(_r)d_ij^(0)(p_1,p_2 ) + & & + _ s^3(_r)d_ij^(1)(p_1,p_2,_r,_f , _ ) .
[ hadrxsec ] in dimensional regularization , we have d_j^(1)(p_1,p_2)&= & d_j^(1)(p_1,p_2 , ) + _ kdx ( p_k(x ) -h_k(x))d_kj^(0)(xp_1,p_2 ) + & & + _ kdx ( p_kj(x)- k^(h)_kj(x ) ) d_k^(0)(p_1,xp_2 ) ; [ subpoint ] + d_ij^(1)(p_1,p_2)&= & d_ij^(1)(p_1,p_2 , ) + _ kdx ( p_ki(x)- k_ki^()(x ) ) d_kj^(0)(xp_1,p_2 ) + & & + _ kdx ( p_kj(x)- k_kj^(h)(x ) ) d_ik^(0)(p_1,xp_2 ) , [ subhadr ] where @xmath43 ( @xmath44 ) and @xmath45 ( @xmath46 ) are the full , @xmath47-dimensional regulated partonic cross sections at the next - to - leading ( leading ) order for the point - like and hadronic contributions respectively . the @xmath48 singularities in @xmath45 and @xmath43
are appropriately subtracted on the right - hand side of eqs .
( [ subpoint ] ) , ( [ subhadr ] ) .
the @xmath49 are the usual altarelli - parisi splitting function ; @xmath50 is the kernel for the splitting process @xmath51 , which is equal to @xmath52 up to a colour factor .
the functions @xmath53 , @xmath54 and @xmath55 are completely arbitrary , in that they define an extra finite part of the subtraction ; different choices correspond to different subtraction schemes . in eqs .
( [ subpoint ] ) and ( [ subhadr ] ) the @xmath56 scheme is equivalent to @xmath57 . for greater generality ,
we have admitted the possibility to have different subtraction schemes on photon and hadron legs .
a change in the subtraction scheme implies that parton distribution functions are modified as follows f_i^=f_i+h_i+_jk_ijf_j .
the term @xmath58 is present only in the photon case , and it is a direct consequence of the inhomogeneous term in the modified altarelli - parisi equations . in ref . [ [ grvdis ] ] a factorization scheme ( @xmath59 ) for the photon densities is introduced , which uses @xmath60 and @xmath61 . as can be seen for eqs .
( [ subpoint ] ) and ( [ subhadr ] ) , the @xmath59 scheme is therefore equivalent to @xmath56 as far as the hadronic component is concerned .
it does however modify the point - like component through the second term in the right - hand side of eq .
( [ subpoint ] ) , which is also responsible for a subtraction on the photon leg .
this clearly shows that the point - like and the hadronic component are closely related , and that they should not be considered separately . in principle , when considering the variation with respect to @xmath62 , cancellations occur between the hadronic and the point - like components , resulting in renormalization group invariance of the physical cross section @xmath63 up to terms of order @xmath64 .
photoproduction cross sections are usually linked to electroproduction ones by a convolution with the weizscker - williams function [ [ ww ] ] : d^(eh)(p_e , p_h)=dx f(x,_ww ) d^()(xp_e , p_h ) .
[ etogamma ] at hera , the h1 and zeus experiments can directly investigate the virtuality of the exchanged photon , by tagging the emitted electron and retaining only those events in which the electron scattering angle @xmath65 satisfies the condition @xmath66 , with @xmath67 typically of the order of few mrad . in ref .
[ [ fmnrww ] ] a functional form especially suited for this experimental set - up was proposed , namely f(x , e_e)&=&\{2(1-x ) + & & + } , where @xmath68 is the incoming electron energy in the laboratory frame . when no angular cut is applied , eq .
( [ etogamma ] ) is too restrictive ; we relax it by writing d^(eh)(p_e ,
p_h)=dx f(x,_ww ) d^()_point(xp_e , p_h)+ dx f(x,_ww^ ) d^()_hadr(xp_e , p_h ) , [ etogamma2 ] with @xmath69 ( for a discussion of the @xmath70 and @xmath71 values , see ref .
[ [ fmnr_tot ] ] and references therein ) . strictly speaking , the fact that @xmath69 gives up the renormalization group invariance of the physical cross section in eq .
( [ etogamma2 ] ) , in that no cancellation occurs between the point - like and the hadronic component when variation with respect to @xmath62 is considered . in practice
, we have verified that this effect is numerically negligible , the point - like component being almost invariant with respect to @xmath62 variations .
[ fixedtargetphotoproduction ] p. frabetti _ et al .
_ , e687 coll .
, ; j. c. anjos _ et al .
_ , e691 coll .
, ; m. p. alvarez _ et al .
_ , na14/2 coll .
, ; g. bellini , proceedings of `` les rencontres de physique de la valle daoste '' , la thuile , aosta valley , 6 - 12 march , 1994 .
|
# 6 plus 1000pt minus 1000pt # 1 # 1= to # 1#2#1#2 # 1#2#3 ^ 2 # 1#2 # 3 # 1#1| # 1| # 1 # 1#1 # 1/ # 1| # 1| # 1#1^ = # 1#2#3 @xmath0 # 1#2#3#4 @xmath1 # 1#2 @xmath2 # 1#2#3#4#5 @xmath3 # 1 # 1@xmath4 = 1.5ex plus 1pt = 11 addtoresetequationsection normalsizesetsizexiipt plus3pt minus3pt@ plus3pt plus3.5pt minus0pt twosidetrue
@=12 0.0 in 0.0 in 6.0 in 8.5 in .7 cm -1 cm .75 in = 11
|
plasmonics is now a very rapidly developing field of activity , and advancement to the mid- and far - infrared ( terahertz ) ranges of frequencies is one of the main streams in photonics @xcite .
our interest in this subject has arisen from discussions of the results of experiments on propagation of surface plasmon polaritons ( spps ) along gold - zns - air interfaces @xcite , which were carried out at novosibirsk free electron laser facility @xcite in the terahertz spectral range . for spps not to be disturbed by any material probe , plasmon - polariton characteristics were studied indirectly via sensing of the electromagnetic field in the space behind the tail facet of samples . in the experiments , the profile of radiated intensity is non - monotonic and does not resemble a surface wave profile which exponentially evanesces with the distance from the wedge face .
a maximum of the radiation intensity is observed at some distance from the surface plane whereas a maximum of the surface wave field should be located at the wedge surface .
these discrepancies had initiated search for theoretical explanation which has led us to the theory developed by g.d .
malyuzhinets in the 1950s .
the sommerfeld - malyuzhinets theory is known to provide a powerful method for exact solution to specific optical problems @xcite . in particular , g.d .
malyuzhinets solved the problem of diffraction of a surface wave by an impedance wedge , considering it as a special case of a plane wave propagation at the brewster angle @xcite .
he derived some general relations but he did not investigate the properties of integrand functions near the saddle points which is necessary for computation of the integrals . a few books and review papers @xcite of varying elaboration
provide a systematic introduction to the theory by g.d .
some earlier approaches to the problem of diffraction by wedges and screens are reviewed in @xcite .
we would like to add a few references to this list @xcite .
v. zon in ref .
@xcite applied the sommerfeld - malyuzhinets technique to the diffraction of a surface wave on conducting wedge with small surface impedance but she did not succeed in obtaining the final solution .
we do that in this paper .
we also conclude that her calculations of the reflection and transmission coefficients for the surface wave are incorrect . to obtain correct results we introduce a sort of landau s bypass rule known in the theory of plasma waves @xcite . in section [ s2 ]
, we briefly recall the properties of surface waves to a minimal extent necessary for understanding of the subsequent calculations and justify the use of the leontovich boundary conditions at the wedge faces ; in particular , we give ( without derivation ) a formula for the surface impedance of a metal substrate coated with a thin dielectric film .
the foundations of the sommerfeld - malyuzhinets technique are expounded in section [ s3 ] .
the diffracted field is calculated in sections [ s4 ] , [ s5 ] , and [ s6 ] .
a brief description of the experimental techniques and comparison of experimental results with theoretical predictions are given in section [ s7 ] .
section [ s9 ] concludes the paper and summarizes our results .
let a metal with the complex permittivity @xmath0 occupy the lower half - space @xmath1 . since a te wave ( the electric field is transverse to the plane of incidence ) has no surface branch , we consider a tm wave in the upper half - space @xmath2 , localized near the metal - air interface and propagating in the @xmath3-direction : @xmath4 , \qquad { \mathbf{e}}_{0 } = \{e_{0x},e_{0y } , 0\ } , \\ { \mathbf{b}}({\mathbf{r}},t ) & = { \mathbf{b}}_{0 } \exp\left[- \varkappa y + i(kx - \omega t)\right ] , \qquad { \mathbf{b}}_{0 } = \{0 , 0 , b_{0z}\ } .
\end{split}\ ] ] it is characterized by the circular frequency @xmath5 , wavenumber @xmath6 , and attenuation constant @xmath7 with a positive real part , @xmath8 .
similar fields ( with a negative attenuation constant , @xmath9 ) could be written for the lower half - space but they are needed only for derivation of the dispersion law for the surface wave : @xmath10 here @xmath11 the gaussian system of units is used , and the derivation can be found elsewhere ( see @xcite ) .
eq . can be interpreted as an evanescent plane wave ( also known as the zenneck wave @xcite ) propagating at a complex angle @xmath12 such that @xmath13 using dispersion relation of eq .
yields @xmath14 in a fictitious case of real @xmath0 , the surface wave exists provided that @xmath15 . for complex values of @xmath0 there is always a solution with @xmath16 which exponentially decreases with the distance from the boundary of the metal .
however , speaking about surface waves it makes sense only if the damping length @xmath17 is sufficiently long , i.e. @xmath18 .
this condition is certainly satisfied if @xmath19then @xmath20 and the imaginary part of @xmath21 is automatically small no matter how large or small the imaginary part of @xmath0 is . keeping the terahertz radiation at novosibirsk free electron laser @xcite in mind
, we will take as an example the value of permittivity @xmath22 for gold at a frequency corresponding to the wavelength @xmath23 @xcite .
then @xmath24 the components of the electric and magnetic field at the metal surface are related through the boundary condition , which can be derived from the maxwell equation @xmath25 its @xmath3-component @xmath26 together with eq .
yields the leontovich boundary condition @xmath27 in the general case , the leontovich boundary condition relates the tangential component of the electric field @xmath28 with that of the magnetic field .
it is usually written in the form @xmath29 , $ ] where the unit vector @xmath30 is directed along the outward normal to the surface of the metal , and the parameter @xmath31 is called ( dimensionless ) surface impedance . in leontovich
s theory , @xmath31 is assumed to be a function of frequency , which depends only on the conductor material but neither on the incident angle nor the type of incident wave . using fresnel s formulae
one can readily check that @xmath32 for a plane tm wave impinging against the metal surface under the angle @xmath33 to the normal .
a propagating te wave is characterized by a different surface impedance @xmath34 and @xmath35 for the surface wave . in any case
, @xmath36 provided that inequation @xmath37 holds .
we see that for large @xmath38 the surface impedance @xmath31 is approximately independent on parameters of the waves , which justifies using the leontovich boundary condition on a metal surface instead of looking for complete solution for wave propagation inside a metal .
below we will write the leontovich boundary condition in the following form @xmath39 , \ ] ] assuming that the formal parameter @xmath12 is related with the surface impedance by eq . .
as shown above , @xmath12 can be considered as a complex angle of propagation of surface wave . for the above cited parameters of golden surface , @xmath40 coating of metallic surfaces
strongly affects the magnitude of @xmath31 . for a metal coated with a thin film with a small width @xmath41 and permittivity @xmath42 , we derived the following expression for the surface impedance : @xmath43 properties of the film becomes dominating for @xmath44 .
for example , zns film with @xmath45 and @xmath46 over a gold substrate has the impedance @xmath47 with an imaginary part more than @xmath48 times larger than that of pure gold .
and @xmath49 are measured from the bisector of the wedge .
, scaledwidth=50.0% ] to begin with , let consider a plane wave of unit amplitude incident on a wedge at the angle @xmath50 , as shown in fig . [ fig:1 ]
. in polar coordinates @xmath51 and @xmath52 perpendicular to the edge of the wedge , the wave is written as @xmath53 .\ ] ] let the wedge occupy the region @xmath54 . within the framework of geometrical optics , the upper face of the wedge reflects the incident wave at the angle @xmath55 .
one more @xmath56 should be subtracted from this value for taking the reversal of the direction of propagation into account , thus the field of the reflected wave is derived from eq . with the substitution @xmath57 : @xmath58 .\ ] ] below we restrict ourselves to the case of tm waves with the transversal component of magnetic field @xmath59 .
the reason for such choice is that a surface wave may have no te polarization . in the vacuum region , @xmath60
, the tm wave obeys the helmholtz equation @xmath61 the non - zero components of the electric field in the tm wave , @xmath62 are expressed through the derivatives of @xmath63 .
the leontovich boundary conditions in eq . with the surface impedance of eq .
on the conducting faces of the wedge at @xmath64 now take the following form @xmath65 where the @xmath66 and @xmath67 signs label the quantities related to the faces of the wedge located at the angle @xmath68 and @xmath69 , respectively .
heuristic considerations enable construction of a solution to eq . with given boundary conditions @xcite .
according to sommerfeld @xcite , the solution is sought in the form of a superposition of plane waves @xmath70 where the integration is carried out over the contour @xmath71 in the complex @xmath72-plane , @xmath73 and the kernel @xmath74 is analytic outside the real axis .
the shadowed areas in figure [ fig:2 ] indicate the regions of the complex @xmath72-plane where @xmath75 and the exponent of eq
. tends to zero as @xmath76 .
following sommerfeld , we choose @xmath71 in the form of two loops .
then , the integral in eq .
is a solution to helmholtz s equation , provided that the two loops , @xmath77 and @xmath78 , are shifted upwards and downwards correspondingly by the distance @xmath79 ; thus any singularities ( see below ) of the function @xmath80 are located between these loops within the strip @xmath81 .
choosing the integration contour in the form of two symmetrical loops guaranties that the asymptote of the integral in eq . at @xmath82
contains no converging cylindrical waves except for the incident wave of eq . , and the incident wave is introduced into @xmath83 through given a pole at @xmath84 as explained in @xcite . with the sommerfeld double loops @xmath85 ( solid line ) and the steepest descent paths @xmath86 ( dashed line ) . , scaledwidth=50.0% ]
let now write functional equations for @xmath83 .
for the leontovich boundary conditions of eq . , we have @xmath87 \operatorname{d\!}p = 0 , \ ] ] where @xmath88 . integrating the first term by parts yields the following equations @xmath89
s(\pm\phi - p ) \operatorname{d\!}p = 0 .\ ] ] since the contour @xmath90 consists of two loops situated symmetrically to the point @xmath91 , these equations are true if the kernel @xmath83 is an even function of @xmath72 .
this leads us to the malyuzhinets functional equations @xcite : @xmath92 s ( p + \phi ) & = \left [ -\sin p + \sin\chi_{+ } \right ] s(-p + \phi ) , \\ \left [ \sin p - \sin\chi_{- } \right ] s ( p - \phi ) & = \left [ -\sin p - \sin\chi_{- } \right ] s(-p - \phi ) .
\end{split}\ ] ] with rather non - trivial calculations one can verify @xcite that the function @xmath93 , \end{gathered}\ ] ] supplies a particular solution for @xmath94 to the functional equations .
the malyuzhinets function @xmath95 is regular in the strip @xmath96 , where the integral in eq .
is even and satisfies the following functional relation @xmath97 with eq .
true this function extends beyond the indicated band . for @xmath98 , where @xmath99 is an irreducible rational number ,
the derivative @xmath100 of @xmath101 can be written as a finite sum of trigonometric functions @xcite .
in particular , @xmath102 for @xmath103 , which corresponds to a right angle wedge . looking for a general solution in the form @xmath104 we deduce from eqs . that the function @xmath105 obeys the following equations @xmath106 for description of an incident wave with unit amplitude
, @xmath107 should have a simple pole at @xmath108 with unit residue . by virtue of eqs .
, @xmath109 is symmetric about points @xmath110 , and therefore its system of poles should satisfy the same symmetry relationships . if the pole @xmath111 is successively reflected about points @xmath112 and @xmath113 , one obtains for @xmath80 a lattice of poles with the period @xmath114 at @xmath115 ( @xmath116 , etc . ) and a similar grid at points @xmath117 .
it is possible to guess a function with these properties .
it was found by g.d .
malyuzhinets @xcite : @xmath118 where @xmath119 . for @xmath120 @xmath121
at a first glance , the sommerfeld - malyuzhinets theory might be sought of as a purely mathematical trick .
however , it indeed describes a real physical phenomenon . to prove that
, in this section we show how an approximation of geometrical optics can be retrieved from this theory which provides an _ exact _ solution of the problem of vector field diffraction , although for a very limited set of problems related to the diffraction of electromagnetic waves on a conducting wedge . to avoid possible misunderstanding regarding the role of geometrical optics in the theory to be presented in the next three sections ,
it is worthy to emphasize that the field which will be derived in the current section is merely a part of a complete solution . from mathematical point of view
, it represents a contribution of the poles of the integrand core @xmath80 in eq . whereas a complete solution includes also a contribution of the saddle points @xmath122 .
the separation of the contribution of the poles and saddle points can be approximately made for the distances @xmath51 from the wedge top larger than the wavelength , @xmath123 .
following g.d .
malyuzhinets @xcite , we replace the contour @xmath90 with the 3-loop contour @xmath124 shown in fig . [ fig:2 ] with the dashed line , in order to compute the sommerfeld integral in eq . :
@xmath125 the contours @xmath126 , @xmath127 and @xmath128 are obtained by deformation of the contours @xmath77 and @xmath78 into the steepest descent paths through the saddle points @xmath129 and @xmath130 of the exponent @xmath131 in the integrand .
the new contour @xmath126 is formed from the middle parts of the contours @xmath77 and @xmath78 , whereas the contours @xmath132 and @xmath133 are composed by remaining end parts of @xmath77 and @xmath78 . as will be shown in section [ s5 ] , the contribution of the saddle point describes the diffracted field at large distances ( in the wave zone ) .
the integral over the contour @xmath126 is reduced to the sum of the residues in the poles of the integrand encircled by @xmath126 within the strip @xmath134 these are the poles to be crossed during the deformation of the original contour of integration @xmath135 to the steepest descent path contour @xmath136 .
the number of such poles depends on the values of @xmath49 , @xmath52 , and @xmath137 .
let consider first the poles of the function @xmath109 in eq . .
there are an infinite number of such poles located at the points @xmath138 however only three of them , namely @xmath139 have chance to fall into the region if @xmath140 . the residue at the pole @xmath141 stands for the wave @xmath142 $ ] , which is readily recognized as the incident wave .
it enters into the result of integration if @xmath143 .
values of @xmath52 outside the interval @xmath144 and within the interval @xmath145 represent the regions of geometrical shadow for the incident wave .
the pole @xmath146 gives rise to the wave @xmath147 = \exp[ik_{0}r\cos(\theta+\theta_0 - 2\phi+\pi)]$ ] , reflected from the upper face @xmath148 of the wedge ( see fig .
[ fig:1 ] ) .
it enters into the result of integration if @xmath149 , i.e. @xmath150 .
finally , the pole @xmath151 describes the wave @xmath152 $ ] , reflected from the lower face @xmath153 of the wedge .
it enters into the result if @xmath154 .
thus , the geometrical optics yields a solution constituted by ` pieces ' of plane waves . at the ends of these pieces represented by light - shadow boundaries the solution vanishes jumpwise to zero in the shadow region , which means that the geometric - optical solution is discontinuous .
these discontinuities of the optical fields will be eliminated with the addition of the diffracted fields in section [ s5 ] .
the surface wave , grazing towards the wedge top along the upper face , as shown in fig . [ fig:1 ] , has a complex propagation angle : @xmath155 it can not be reflected from the lower face of the wedge if @xmath140 as can be readily deduced from the above treatment , and only two poles , namely @xmath156 can fall into the region given by inequation .
this occurs if @xmath157 where the upper and the lower signs stand for @xmath141 and @xmath146 , correspondingly . due to the factor @xmath158 in eq .
, the residue about the pole @xmath159 is evaluated as a wave of unit amplitude : @xmath160 .\ ] ] it describes the incident surface wave propagating along the upper face to the top edge of the wedge . according to the inequation
, this wave does not penetrate into the shadow region @xmath161 if @xmath162 is sufficiently small .
we will see in section [ s5 ] that the transitional region near the formal boundary @xmath163 of the shadow region is described by a simple function , which includes the contributions from both the pole @xmath141 and the saddle point @xmath129 .
the contribution of the pole @xmath146 is exactly zero , which can be seen from the fact that formally calculated residue @xmath164 contains the multiplier @xmath165 , which is zero .
since the contributions of both the @xmath146 and @xmath151 poles are zero , the wedge does reflect surface waves ( this assertion will be clarified in section [ s6 ] ) .
let us now proceed to the poles of the function @xmath166 . to simplify our task , below we restrict ourselves to the case of a right angle wedge with @xmath167 .
then , @xmath166 is expressed through trigonometric functions with the aid of eq . , and
the poles can be found from the following equation @xmath168 \cos\left [ \frac{1}{6}\left(\theta + \chi _ { - } - p-\frac{5\pi}{4}\right ) \right ] \\
\times \cos\left [ \frac{1}{6 } \left(\theta -\chi_{+ } - p+\frac{5\pi}{4}\right ) \right ] \cos\left [ \frac{1}{6 } \left(\theta + \chi_{+ } - p+\frac{\pi } { 4}\right ) \right ] = 0 .
\end{gathered}\ ] ] they obey the conditions of eq . if @xmath169 for bare metals , @xmath170 is positive in the optical and infrared ranges of frequencies .
hence , the inequations can not be satisfied , and the function @xmath171 can not have poles inside the contour @xmath172 .
however the zero point @xmath173 of the factor @xmath174 $ ] approaches the first saddle point @xmath129 at @xmath175 .
similarly , the zero point @xmath176 of the multiplier @xmath177 $ ] is located near the second saddle point @xmath178 at @xmath179 . although one might expect that these poles strongly affect the diffracted field near the wedge faces at @xmath180 , calculations in section [ s6 ] do not confirm these fears .
evaluating the integral in eq . along the contours @xmath127 and @xmath128 in fig .
[ fig:2 ] yields radiated ( freely propagating ) electromagnetic fields . those contours pass through the saddle points @xmath129 and @xmath130 , respectively . to use the saddle - point method of integration
, we expand the exponent in the integrand of eq . about the saddle points so that @xmath181 putting the expansion @xmath182\ ] ] in eq
. reveals that , due to the factor @xmath183 , the main contribution to the integral over the contours @xmath127 and @xmath128 comes from the neighborhood of the saddle points .
turning to the integration in the variable @xmath184 and taking into consideration the direction of integration over the contours @xmath127 and @xmath128 , we transform the respective integrals to the following form : [ 5:2 ] @xmath185 in the wave zone that corresponds to the limit @xmath186 , the pre - exponential factor @xmath80 can be substituted with @xmath187 , which yields @xmath188 for the contour @xmath132 and @xmath189 for the contour @xmath133 .
( solid line ) and @xmath190 ( dashed line ) for @xmath191 .
the width of the narrow peak at @xmath192 in the plot of @xmath193 is of the order of @xmath194 . for values of @xmath195 relevant to the thz range of frequencies ,
the peak is much narrower and higher .
, scaledwidth=80.0% ] since the pole @xmath196 approaches the saddle point @xmath129 when @xmath52 tends to @xmath197 , the function @xmath198 has a narrow peak near @xmath199 , which is much bigger than @xmath200 as shown in fig . [ fig:3 ] .
hence , the field @xmath201 can be neglected near the peak , where the amplitude of the diffracted field is @xmath202 the angular distribution of the intensity of the radiation field in eq .
has the lorentzian profile with the halfwidth @xmath203 similar calculations were done by v. zon in @xcite . note . however . that the range of applicability of eqs
. is limited to very large distances , such as @xmath204 for smaller distances , the factor @xmath80 in eqs .
can not be considered smooth as compared with the exponent @xmath183 since the former has a pole at @xmath205 , which approaches the saddle point @xmath129 at @xmath199 .
the absolute magnitudes of the functions @xmath187 are plotted in fig .
[ fig:3 ] .
more accurate calculation can be performed using the following approximate expression for the kernel function @xmath206 it is derived by expanding @xmath80 in the laurent series about the pole @xmath159 but it provides a good approximation for @xmath80 in the entire range of integration as fig .
[ fig:4 ] proves .
( solid line ) and laurent expansion @xmath207 ( dashed line ) ; the real part is blue , the imaginary is black ; @xmath208 . ]
putting @xmath209 in eq .
yields @xmath210 where @xmath211 a method of calculating the integral @xmath212 in eq . is elaborated in the theory of plasma waves @xcite .
the result is expressed in terms of the imaginary error function @xmath213 and has two branches : @xmath214 .\ ] ] the first branch @xmath215 is originally computed in the assumption that the imaginary part of @xmath216 is positive , @xmath217 , and then analytically continued to the lower half of the complex plane @xmath216 , where @xmath218 , using eq . .
the second branch @xmath219 , on the contrary , is originally computed for @xmath218 and then analytically continued to @xmath217 . in terms of the contour integration in eq .
, the analytical continuation implies that the pole @xmath220 never crosses the contour of the integration ( which originally goes along the axis @xmath221 ) , and the contour is deformed when the pole @xmath220 crosses the axis @xmath221 to bypass the pole .
since the pole can be bypassed either from above or from below , there appear two branches of the function @xmath222 .
they differ in the residue of the integrand at the pole @xmath220 : @xmath223 for large @xmath224 , the functions @xmath225 are evaluated as @xmath226 respectively , if @xmath217 and as @xmath227 if @xmath218 .
note that the exponent @xmath228 becomes very big in sectors where @xmath229 .
the sign of @xmath230 is reversed as the observation point at the angle @xmath52 crosses the boundary @xmath231 of the shadow region . together with the discontinuous field @xmath232 , which represents the contribution of eq .
of the pole @xmath141 in the approximation of geometrical optics , the diffracted field @xmath233 forms a continuous field . this total continuous field
is given by the integral of eq .
computed for @xmath234 .
indeed , the pole @xmath141 goes away from the interior of the contour @xmath172 through the contour @xmath132 if the angle @xmath52 takes a sufficiently large negative value ( which corresponds to a deep shadow region ) thus making @xmath234 . in that case ,
the integral over the contour @xmath172 gives no contribution to the result of calculations so @xmath233 stands for the total field . taking the lower sign in eq .
, we obtain the final expression @xmath235 for the radiation field in the proximity of @xmath199 . in plasma physics
, our choice of the branch @xmath219 of the multivalued function @xmath222 is known as the landau rule @xcite . by analytic continuation ,
. represents the total field for any values of @xmath52 .
it is valid for any @xmath51 larger than the wavelength in a free space , @xmath236 , both in the near - field and wave zones .
the near field was computed numerically in @xcite by evaluating the integral in eq . .
eq . gives the desired result with a better accuracy , especially in case where @xmath237 . and
@xmath238 ( blue ) , @xmath239 ( magenta ) , and @xmath240 ( dark yellow ) .
the @xmath3 and @xmath241 axes go along the upper and side faces of the wedge as shown in fig .
[ fig:1 ] .
the distance to the curve is proportional to the intensity of scattered radiation at the given angle . ] for a gold wedge coated with @xmath242 zns film : @xmath238 ( blue ) , @xmath239 ( magenta , scaled @xmath243 ) , @xmath240 and ( dark yellow , scaled @xmath244 ) . ]
angular patterns of diffracted waves computed by formula in eq . for a pure gold substrate and a gold wedge coated with zns film are presented in figs .
[ fig:5 ] and [ fig:6 ] , respectively , where the intensity @xmath245 is depicted for several distances from the edge of the wedge for parameters relevant to ongoing experiments at novosibirsk free electron laser .
[ fig:5 ] is typical for the near - field zone , where the angular distribution has many peaks .
[ fig:6 ] , on the contrary , shows angular distributions with a single peak characterized by the lorentzian profile of eq . .
in the first case of a pure gold wedge the far wave zone begins after the distance @xmath246 from the wedge edge whereas for zns - coated gold the wave zone is located at @xmath247 .
in contrast to the lorentzian profile eq . , the total diffracted field in eq .
is not symmetric about the boundary of the shadow region @xmath248 .
it extends into the illuminated region @xmath249 much further than into the shadow region @xmath250 .
this assertion has found a convincing evidence in experimental data cited in @xcite .
another important conclusion is that the radiation intensity in the illuminated area does not copy the profile of the surface wave : instead of monotonic decreasing with the distance from the face of the wedge it rises at first .
again , this fact has been confirmed experimentally @xcite .
( blue ) , @xmath251 ( magenta ) , @xmath252 ( dark yellow ) , and @xmath253 ( green ) .
pure gold wedge . ] ( blue ) , @xmath251 ( magenta , scaled @xmath254 ) , @xmath252 ( dark yellow , scaled @xmath243 ) , and , @xmath253 ( green , scaled @xmath255 ) .
gold wedge coated with a @xmath242 zns film . ]
intensity of diffracted waves vs. distance from the upper face is shown in figs .
[ fig:7 ] and [ fig:8 ] , which illustrate that incident surface wave is mainly scattered into the upper hemisphere .
near the face surfaces of the wedge at @xmath180 the integrals @xmath233 and @xmath201 have commensurate magnitude and both should be computed .
moreover , the integrands in eqs . obey the equality @xmath256 at @xmath180 as can be seen from fig .
[ fig:3 ] .
the approximation of eq . provides a good fit for the kernel @xmath80 near the saddle point @xmath129 for almost entire interval @xmath257 of the angles @xmath52 .
however it does not satisfy eq . at the ends and could be improved there .
near the upper face at @xmath258 the pole @xmath259 approaches the saddle point @xmath129 at @xmath258 , as mentioned at the end of section [ s4 ] .
expansion of the kernel @xmath80 in eq . into the laurent series about the pole @xmath260 near the saddle point @xmath129 has the following form : @xmath261 the kernel @xmath262 in eq .
is regular at @xmath258 near the second saddle point @xmath130 . combining eqs . , , and , we obtain @xmath263 . \end{gathered}\ ] ] the second term in eq .
, containing @xmath264 , is small and could be neglected .
. gives a similar result at @xmath258 , which differs by a numerical coefficient of the order of unity .
similarly , the pole @xmath265 approaches the saddle point @xmath130 at @xmath266 .
the expansion of the kernel of the integral in eq .
about @xmath267 has the following form : @xmath268 and the kernel of eq .
is @xmath269 combining eq . with eqs . and yields @xmath270
\end{gathered}\ ] ] again , the second term with @xmath271 is small and can be dropped .
. describes a wave propagating along the upper face of the wedge back from the wedge edge , and eq .
is a wave propagating along the lower face .
they can be thought of as reflected and refracted waves , respectively , although they are not surface waves .
it is interesting that the amplitude of the refracted wave is @xmath272 times larger than the amplitude of the reflected one .
the experimental configurations are shown in fig . [ fig:9 ] .
monochromatic radiation of novosibirsk free electron laser ( fel ) at a frequency of @xmath273 ( @xmath274 ) entered the user station through a 16-m long beamline as a gaussian beam @xmath275 with a waist of @xmath276 . after passing a circular aperture with a diameter of @xmath277 ,
the beam was focused with a cylindrical mirror into the input mouth of a plane waveguide , formed by the gold - covered facets of two glass prisms . at the output mouth of the waveguide
, the radiation passed through the slit , transformed into a surface plasmon - polariton travelling along the metal - dielectric interface . since a portion of the radiation could be emitted as a free wave , to separate the spp and the bulk wave we made the input facet of the large prism tapered with an angle of 13 degrees to the upper sample plane .
the latter was @xmath278 long and @xmath279 wide .
the edge between these two facets was smoothed for the purpose of decreasing the spp radiation loss .
the large and small facets of glass slabs were covered with a @xmath280 thick gold layer , which was considered within this problem as a bulk metal since the skin - depth for the gold is much smaller than @xmath280 .
the spps , which were launched with the help of the waveguide , travelled along the large facet of the samples .
bare gold and gold covered with zns layers @xmath281 to @xmath272 @xmath282 thick were employed in the experiments . to study spp characteristics we applied two non - invasive techniques for detection of electromagnetic ( em ) radiation in the free space behind the end facet of the samples .
an optical system consisting of a tpx lens with @xmath283 @xcite and a microbolometer focal plane array @xcite was used for imaging of the em radiation wavefronts .
intensity of the em radiation at different distances @xmath284 was scanned along the @xmath241-axis using an opto - acustic goley cell @xcite with an input slit @xmath285 thick directed along the @xmath216-axis .
zns layer ( below ) , recorded with the microbolometer focal plane array ( configuration of fig .
[ fig:9],a ) . ] , b for a bare gold surface : ( a ) goley cell signal for three distances from the sample end ; ( b ) simulations . ]
the experimental results were presented in brief in the conference proceedings @xcite and will be published in detail elsewhere @xcite . in this paper
we present only the data relevant to the contents of the theory , namely we describe characteristics of the electromagnetic field ( emf ) that arose when an spp reached the end of the surface .
the main features of the emf were as follows ( see figs .
[ fig:10 ] and [ fig:11 ] ) .
( a ) the intensity of emf , in contrast to the spp intensity , reached its maximum at a distance of 1 to 2 mm above the sample surface , depending on the zns layer thickness , and only then sloped down .
( b ) oscillations were observed on the descending part of the distribution .
( c ) the width of the distribution decreased with the distance and the angular distribution became narrower .
all these features are in reasonable agreement with the theoretical predictions .
some discrepancies between the distribution shapes on the slope recorded with the mbfpa and the goley cell may be caused , for example , by field disturbance by the metal diaphragm at the cell input window .
in this paper , we have considered diffraction of a surface wave by a rectangular wedge with impedance facets using the sommerfeld - malyuzhinets technique .
we have derived eq .
that uniformly describes the diffracted field both in the near - field and far - field zones .
it was used in section [ s5 ] for analysis of diffracted radiation at various distances from the wedge .
main conclusions from this formula are confirmed by available data from ongoing experiments at novosibirsk free electron laser @xcite .
first , we have shown that the total diffracted field expressed by eq .
is not symmetric about the boundary of the shadow region and that the surface wave is scattered mainly into the upper hemisphere .
it extends into the illuminated region much further than into the shadow region as depicted in figs .
[ fig:7 ] and [ fig:8 ] .
another important conclusion is that the radiation intensity in the illuminated area does not follow a surface wave profile and at first increases instead of monotonic decreasing with the distance from the face of the wedge , as one might expect for the surface wave , which exponentially decays with the distance from the metal - air interface .
again , this fact has been confirmed experimentally in @xcite .
we have confirmed the conclusion made in @xcite that in the wave zone the angular distribution of the scattered wave has the lorentzian form with a width determined by impedance .
however , we have noted that the wave zone for a wedge with a small surface impedance begins at very large distances , specified by eq . .
this work was supported by the russian government , grant 11.g34.31.0033 , the russian foundation for basic research , grants 11 - 02 - 12252-ofi - m , 11 - 02 - 12171-ofi - m , and the russian ministry of education and science , state contracts no 14.b37.21.0732 , 14.b37.21.0784 and 14.b37.21.0750 .
the experiments mentioned in the paper were carried out using equipment belonging to the scstr .
33ifxundefined [ 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 * * , ( ) in @noop _ _ ( , , ) @noop * * , ( ) @noop _ _ , , vol .
( , ) link:\doibase 10.1016/s0165 - 2125(98)00042 - 0 [ * * , ( ) ] @noop _ _ , alpha science series on wave phenomena ( , , ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1088/1464 - 4258/9/9/s32 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/0041 - 5553(70)90121 - 7 [ * * , ( ) ] link:\doibase 10.1016/0041 - 5553(71)90120 - 0 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1016/s0165 - 2125(99)00005 - 0 [ * * , ( ) ] link:\doibase 10.1016/j.wavemoti.2004.05.009 [ * * , ( ) ] , link:\doibase 10.1016/j.wavemoti.2006.06.005 [ * * , ( ) ] link:\doibase 10.1016/s0165 - 2125(98)00049 - 3 [ * * , ( ) ] http://search.ieice.org/bin/summary.php?id=e84-c_7_994 [ * * , ( ) ] in @noop _ _ , vol . , ( , , ) pp
. @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1007/s10762 - 012 - 9878-x [ * * , ( ) ] \doibase http://dx.doi.org/10.1364/ao.24.004493 [ * * , ( ) ] http://www.springerlink.com/content/q68824p861u7271g/ [ * * , ( ) ] @noop * * , ( ) http://www.tydexoptics.com [ `` , '' ] @noop * * , ( ) @noop * * , ( ) @noop ( )
|
diffraction of a surface wave on a rectangular wedge with impedance faces is studied using the sommerfeld - malyuzhinets technique .
an analog of landau s bypass rule in the theory of plasma waves is introduced for selection of a correct branch of the sommerfeld integral , and the exact solution is given in terms of imaginary error function .
the formula derived is valid both in the near - field and far - wave zones .
it is shown that a diffracted surface wave is completely scattered into freely propagating electromagnetic waves and neither reflected nor transmitted surface waves are generated in case of bare metals which have positive real part of surface impedance .
the scattered waves propagate predominantly at a grazing angle along the direction of propagation of the incident surface wave and mainly in the upper hemisphere regarding the wedge face .
the profile of radiated intensity is nonmonotonic and does not resemble the surface wave profile which exponentially evanesces with the distance from the wedge face .
comparison with experiments carried out in the terahertz spectral range at novosibirsk free electron laser has shown a good agreement of the theory and the experiments .
|
[ cols="<,^,^,^,^,^,^,^,^ " , ]
the nonthermal spectral index @xmath3 , is the main parameter of interest because it is used to model cre generation and propagation .
however , it is not possible to measure it directly .
the quantity which is measured is the total spectral index @xmath4 which is contaminated by thermal free - free emission .
this contamination is significant for spatially resolved regions of the galaxy where the thermal fraction , where , @xmath5 is the flux density of the thermal emission . in the text
we express @xmath6 as percentages . ]
( henceforth , @xmath6 ) is high . in this section
, we assess the effect of @xmath6 on @xmath3 obtained between 333 mhz and near 1.4 ghz in spatially resolved regions of the galaxy . to do this
, we use the method of thermal - nonthermal separation developed by @xcite and apply it to five of our galaxies ( except ngc 3034 ) .
the method uses dust extinction corrected h@xmath4 maps to get a template for the free - free emission across the galaxy , and extrapolates it to the desired radio frequency to obtain spatially resolved thermal radio emission maps .
the extinction correction is achieved by estimating the dust temperature and optical depth using the far infra - red maps at @xmath770@xmath8 m and @xmath7160@xmath8 m ( see appendix [ appendixa1 ] for details ) .
the far infrared @xmath7160@xmath8 m map has an angular resolution of @xmath9 , and hence the final nonthermal spectral index maps have this angular resolution , which is significantly coarser than the total spectral index ( hereafter @xmath4 ) maps .
the usefulness of this method lies in the fact that it can be used in spatially resolved parts of the galaxy .
further , @xcite demonstrate the robustness of their method over several other existing thermal - nonthermal separation techniques ( see references therein and @xcite ) .
the mean @xmath6 for all our sample galaxies was found to be less than 5% at 333 mhz , although the value can go up to 10% in some specific bright regions and in the spiral arms . applying the same method to the available higher frequency data at or close to 1.4 ghz
, we found the mean @xmath6 to be less than 12% .
the distribution of the @xmath6 at 333 mhz and near 1.4 ghz for five galaxies is shown in figure [ figurea1 ] of the appendix . the nonthermal 333 mhz and higher frequency maps
were further used to obtain @xmath3 variation for the five galaxies .
the left panels of figure [ figure3 ] , show the @xmath4 maps and the right panels show the @xmath3 maps .
figure [ figure4 ] shows the variation of @xmath4 and @xmath3 with radius for the galaxies . here ,
the spectral index was estimated by azimuthally averaging over annuli of one beam width .
the radial profiles clearly reveal that the nonthermal spectral index is nominally steeper than the uncorrected values , and further steepens towards the outer parts of the galaxy .
however , we note that towards the outer parts of the galaxies , the steepening of both the total and nonthermal spectral indices may be caused due to `` missing flux density '' problems . as a function of @xmath6 at 1465 mhz for ngc 6946 obtained at spatially resolved regions of the galaxy .
the black circles corresponds to @xmath10 between 333 mhz and 1465 mhz , while the red triangles shows between 1465 mhz and 4860 mhz .
the top panel shows the flux density normalized to the peak value at 333 mhz ( in black ) and 1465 mhz ( in red ) respectively .
note that generally higher flux density regions have high thermal fractions.,width=302,height=302 ] [ figure5 ] we are now in a position to quantify the importance of thermal - nonthermal separation , for robust estimates of @xmath3 between 333 mhz and near 1 ghz .
the uncertainty in @xmath3 can be estimated by propagating the various sources of errors in obtaining the nonthermal maps .
three primary sources of error affect the @xmath3 measurement , namely : the rms in the two radio maps ( @xmath11 ) , the error associated with the uncalibrated system temperature of the instrument @xmath12 , and the error in thermal fraction , @xmath13 .
@xmath12 is about 5% at 333 mhz for the gmrt and assumed about 2% for vla and wsrt near or at 1.4 ghz maps .
the error @xmath13 is estimated to be about 10% at these frequencies by @xcite ( also see appendix a ) . incorporating these errors , we have computed the ratio @xmath14 in regions by dividing the galaxy into rectangular grids of size of approximately 40@xmath15 .
we find that for thermal fraction less that 5% at 333 mhz and less than about 10% near 1.4 ghz , @xmath16 .
this effect is illustrated for ngc 6946 in the bottom panel of figure 6 , where the @xmath14 is plotted ( in black ) as a function of thermal fraction near 1.4 ghz . within
the error bars the ratio is consistent with unity for thermal fraction less than 10% . for higher thermal fraction ( predominantly in the arms or regions
) the ratio drops below unity , and correction to @xmath4 , to obtain @xmath3 , could be as high as 20% .
the thermal - nonthermal separation is even more important for @xmath3 calculated between higher frequency pairs like 1.4 ghz and 4.8 ghz .
the ratio @xmath14 for ngc 6946 obtained between 1465 mhz and 4860 mhz ( using archival map downloaded from ned , @xcite ) is plotted in red in figure 6 , where we clearly see that the ratio is significantly below unity for the whole range of thermal fractions . note that since there is also a systematic uncertainty associated with the absolute calibration scale for calibrating radio flux densities , the absolute value of the spectral index can vary by about 10% .
* ngc 1097 : * our 333 mhz map ( see figure [ figure1 ] ) has an angular resolution of 16@xmath1711@xmath15 , which corresponds to a spatial resolution of @xmath2 1.1 kpc , if the galaxy is at a distance of 14 mpc .
the only other low frequency radio continuum study of this galaxy was done by @xcite and @xcite at 1465 mhz .
they reported strong radio emission coincident with the narrow dust lanes in the bar .
the central bar region is also clearly visible in our 333 mhz map , and is coincident with dust emission inferred from the far infra - red @xmath724@xmath8 m images from _ spitzer _ , as seen in figure 7 .
the bar pattern rotates in the clockwise direction ( see @xcite ) , and as expected prominent shock fronts are visible on the leading side of the bar with respect to the sense of rotation .
the @xmath6 in the bar region at 333 mhz is estimated to be @xmath181% , hence the radio emission at this frequency is entirely nonthermal in origin .
these observations fit well with the explanation that the radio emission from the bar results due to shock compression of gas , dust and magnetic field , as was concluded by @xcite .
the total emission from the bar and the central component ( the region shown in figure 7 ) is 835@xmath1957 mjy , out of which 750@xmath1952 mjy arises from the central compact region ( 25@xmath1721@xmath15 ) . for a similar sized central region @xcite quote a flux density of 260 mjy at 1465 mhz , based
on which we get a @xmath4 of @xmath20 .
the @xmath3 within a 40@xmath15 region based on our analysis is @xmath21 .
this @xmath3 close to the nuclear region is steeper when compared with other galaxies in our sample .
this indicates possibility of radiative cooling by inverse compton / synchrotron processes .
m mips image ( resolution @xmath26@xmath15 ) , overlaid is the high resolution 333 mhz map made using baselines more than 1 kilo@xmath7 with a resolution @xmath22 .
the contours start from 1 mjy , and subsequent contours are in multiples of @xmath23 .
the highest intensity contour is at 150 mjy / beam.,width=302 ] the southeastern ( se ) side of the bar has a flux density of @xmath24 mjy and the northwestern ( nw ) side has a flux density of @xmath25 mjy , measured within the 2 mjy closed contours .
the red lines in figure [ figure6 ] show the se bar and nw bar . in the optical
, the bar region continues into two prominent arms however the radio emission is highly diffuse .
we find the thermal fraction of the interarm region at 333 mhz and 1465 mhz to be 0.2% and 2.5% respectively whereas in the spiral arm regions it is 1% and 4.5% respectively . in the arm regions ,
the @xmath26 is @xmath27 while it gets significantly steeper ( @xmath28 ) in the interarm regions which may be caused due to missing flux density problems .
the radial profile of the @xmath4 and @xmath3 in figure [ figure4 ] , gets steep towards the outer parts of the galaxy possibly due to missing flux density .
we emphasize that the overall thermal fraction in this galaxy is very low , with @xmath29 of @xmath20.6% at 333 mhz and @xmath184.8% at 1465 mhz ( see figure [ figurea1 ] ) . at 333 mhz , we detect the radio galaxy discovered by @xcite , located @xmath290@xmath15 from the nucleus of ngc 1097 towards the southwest at a pa of @xmath30 ( see figure 7 ) .
the radio galaxy has a total flux density of @xmath31 mjy , with @xmath32 mjy in the northern component ( @xmath33 , @xmath34 ( j2000 ) ) and @xmath35 mjy in the southern component ( @xmath36 , @xmath37 ( j2000 ) ) . *
ngc 3034 ( m82 ) * : the 333 mhz map ( see figure [ figure1 ] ) of this galaxy has an angular resolution of 22@xmath38 , corresponding to a spatial resolution of about 0.5 kpc for a distance of 3.9 mpc .
we find the integrated flux density at 333 mhz to be 14 jy which is consistent with flux densities measured at other frequencies .
the rms of the image , which is 3 mjy / beam , is dynamic range limited since the peak flux density in the galaxy is 3 jy / beam .
morphologically the galaxy is featureless at our resolution , although the northern extension ( @xmath39 ) is coincident with the base of the h@xmath4/optical horn ( see figure [ figure1 ] ) , which is likely due to synchrotron emitting particles convected outwards by the nuclear wind , tracing the h@xmath4 horns .
the galaxy was observed between 330 mhz to 4835 mhz by @xcite , and a similar horn was seen at @xmath21515 mhz .
their study also revealed a nonthermal radio halo surrounding the galaxy , however our observations do not show this due to dynamic range limitations . for this galaxy ,
the thermal - nonthermal separation was not possible , but we have computed the @xmath4 map and radial profile as shown in figure [ figure3 ] and [ figure4 ] .
the 333 mhz gmrt map was convolved to the resolution of vla map at @xmath40 .
this is a good approximation for @xmath3 index variation across the galaxy since the thermal fraction of this galaxy has been estimated to be 15% at 32 ghz @xcite , corresponding to @xmath41% and @xmath42% at 1.4 ghz and 333 mhz respectively .
the spectral index is seen to be very flat with @xmath43 towards the center and steepens to @xmath44 at a distance of @xmath22 kpc from the center .
similar radial variation of the spectral index was also reported by @xcite .
9.9@xmath15 and the grey scale shows the h@xmath4 image ( in @xmath45 ) from the jacobi kapteyn telescope with the 6570 filter @xcite , smoothed to the same resolution as the 333 mhz map .
the lowest contour is at 3 mjy increasing in steps of 1 mjy.,width=302 ] * ngc 4736 ( m94 ) * : the 333 mhz map ( see figure [ figure1 ] ) of this galaxy has an angular resolution of 13@xmath46 , corresponding to an spatial scale of about 0.3 kpc for a distance of 4.7 mpc
. the other low frequency study of this galaxy exist at 610 mhz @xcite , observed with much lower angular resolution of 56@xmath47 . at our angular resolution
we clearly detect emission from the inner star forming circumnuclear ring of the galaxy as shown in figure 8 overlaid on the h@xmath4 map .
the ring is located at about 50@xmath15 ( major axis ) , corresponding to 1.1 kpc , from the center , and is @xmath219@xmath15 ( 420 pc ) wide ( up to the half power ) having a flux density of 240@xmath1913 mjy .
the estimated thermal fraction of the ring at 333 mhz and 1374.5 mhz is @xmath48% and @xmath49% respectively , and hence the ring at 333 mhz is largely nonthermal in origin . the ring is seen to be coincident with the h@xmath4 ring , which is shown in figure [ figure7 ] after convolving to the same radio resolution .
radio emission from the ring is also seen in the higher frequency 8.46 ghz observations made by @xcite , however at these frequencies the emission is a mixture of thermal and nonthermal components .
the ring has relatively high thermal fraction which leads to a noticeable change in spectral index ( see figure [ figure4 ] ) : @xmath50 is @xmath51 , while the @xmath26 is @xmath52 .
the ring has massive on going star formation @xcite and @xmath3 value indicates that both cre generation and escape happens on timescales shorter than the radiative timescale . the compact nucleus lies within the central 0.5 kpc ( 25@xmath1722@xmath15 ) with integrated flux density of 110@xmath196 mjy and a peak flux of 45 mjy / beam and @xmath53 . outside the ring extending to the edge of the galaxy there is no thermal emission observed leading to @xmath16 , * ngc 5055 ( m63 ) * : the 333 mhz observation ( see figure [ figure1 ] ) has an angular resolution of 17@xmath1710@xmath15 corresponding to a spatial scale of 0.7 kpc at a distance of 9 mpc .
as is evident from the figure , the extent of the galaxy is larger at 333 mhz than the optical . in an earlier study by @xcite at 610 mhz and 1410 mhz at the much lower angular resolution of 58@xmath1787@xmath15
, the galaxy appears featureless which is also the case in our higher resolution 333 mhz map .
however , 10.55 ghz polarization observations by @xcite show regular spiral structure .
the overall radio emission at 333 mhz at a level of 4@xmath54 is @[email protected]@xmath56 in extent , which has an integrated flux density of [email protected] jy .
this value is higher than what one would obtain using a spectral index of @xmath57 estimated between 610 and 1417 mhz @xcite .
the reason for this is @xcite estimate the integrated flux density within a region [email protected]@xmath56 at 610 mhz , which is smaller than our detected size at 333 mhz .
however , if a similar size region is considered , we obtain a flux density of [email protected] jy which is consistent with a spectral index of @xmath57 as seen in figure [ figure2](d ) .
ngc 5055 is highly inclined , with inclination angle , @xmath58 , and the extended low frequency radio emission can result due to escape of low energy ( @xmath18 2 gev ) cre from the disc which travel larger distances ( @xmath2 4.5 kpc ) into the galactic halo .
the mean thermal fraction @xmath29 at 1696 mhz ( @xmath718 cm ) and 333 mhz is 12% and 1.3% respectively .
the @xmath3 in the central region is @xmath2 @xmath59 which steepens to @xmath60 towards the outer parts of the galaxy to @xmath59 between the central and outer parts of the galaxy , while we find much steeper values .
the wsrt 1696 mhz map has a flux density of 284 mjy within the same region of [email protected]@xmath56 , which when extrapolated using a spectral index of @xmath57 , has a much lower value ( see figure [ figure2](d ) ) .
this may be due to missing flux density issues as discussed in section [ section2 ] .
thus the true spectral index should be flatter than what is observed by us . ]
( see figure [ figure4 ] ) .
this result supports the conclusion drawn by @xcite that the spectral index steepening is mainly due to energy losses of cre and decrease in number of relativistic electrons with increasing galactic radius .
, overlaid are the 333 mhz contours at resolution @xmath61 .
the marked regions are as per ondrechen ( 1985 ) .
the flux densities of the sources are as follows ; a : @xmath62 mjy , d : @xmath63 mjy , f : @xmath64 mjy , g : @xmath65 mjy , h : @xmath66 mjy , i : @xmath67 , x : @xmath68 mjy , width=340,height=340 ] * ngc 5236 ( m83 ) : * the 333 mhz observation ( see figure [ figure1 ] ) has an angular resolution of 16@xmath1712@xmath15 , which correspond to a linear scale of about 0.4 kpc at a distance of 4.2 mpc . an early study of this galaxy by @xcite at 327 mhz using the ooty synthesis radio telescope ( osrt ) quotes integrated flux density of [email protected] jy at a resolution of @xmath69 , while our gmrt observations has a slightly higher integrated flux density of [email protected] jy .
@xcite has reported several point sources in this galaxy . to verify the presence of these sources at 333 mhz
, we made a high resolution image of @xmath61 as shown in figure [ figure8 ] . comparing this with @xcite we detect the central bar ( shown in red lines ) and the regions named as a , f , h and i , and nonthermal polarized point sources d , g and x. note that source g
is identified as a background galaxy by @xcite .
the region e was identified as a shocked region in the arm , and has a flux density of @xmath70 mjy , while the point source q is a region and has flux density of 5 mjy / beam .
however , we were unable to detect point sources b and c. the flux densities of the point sources are given in figure [ figure8 ] .
in addition to this we see an extended radio arm , which runs along the faint narrow dust lane visible in the infrared @xmath770@xmath8 m map .
the galaxy has a low mean thermal fraction of about 3% at 333 mhz and about 7% at 1452 mhz , and hence @xmath16 as seen in the radial spectral index profile of the galaxy in figure [ figure4 ] . in the arms , at 333 mhz , the thermal fraction is estimated to be @xmath71% , where @xmath4 ranges from @xmath72 to @xmath73 which gets modified to @xmath3 as @xmath74 to @xmath75 .
@xmath3 along the arms is seen to change significantly from @xmath76 within the central 3 kpc to @xmath77 beyond 3 kpc .
thermal fraction at 333 mhz in the interarm regions is @xmath78% and beyond the central 4.5 kpc is less than 1% , indicating purely nonthermal emission . in the interarm regions ,
@xmath3 values lie in the range @xmath79 to @xmath80 . *
ngc 6946 : * the 333 mhz observation ( see figure [ figure1 ] ) has an angular resolution of 12@xmath1711@xmath15 which corresponds to a spatial resolution of 0.4 kpc at a galaxy distance of 6.8 mpc .
the integrated flux density at 333 mhz is [email protected] jy . using our 333 mhz and archival 1465 mhz vla maps
, we estimate the mean thermal fraction to be about 4.3% and 11% respectively . spectral index study by @xcite between 610 mhz and 10.7 ghz showed variation from about @xmath81 near the center to about @xmath82 towards the edge . the spectral index between gmrt 333 mhz map and vla 1465 mhz map shows a similar trend , with spectral index of @xmath83 at the center and @xmath59 towards the edge .
recently , @xcite studied the spectral index between 4860 mhz and 1465 mhz which showed flatter spectral index in the arm than the interarm due to higher thermal fraction . in our studies ,
thermal fraction is higher in the arm ( 510% ) than in the interarm ( 15% ) . in the arm
, @xmath4 lies in the range @xmath81 to @xmath77 , while the @xmath3 is steeper @xmath84 to @xmath85 . in the interarm regions the average spectral index is steeper with @xmath86 , while the @xmath3 does not change significantly , @xmath87 .
the 333 mhz map prominently shows the giant regions identified by @xcite .
the region at @xmath88 , @xmath89 ( j2000 ) ( region c in @xcite ) , is about 35@xmath15 ( 1.2 kpc ) in size , and can be seen in the @xmath90 low resolution nonthermal maps .
this region is a suitable candidate to investigate the cre injection near high star forming sites .
we find the thermal fraction to be @xmath91% at 333 mhz and @xmath92% at 1465 mhz .
the @xmath3 in this region is @xmath93 , whereas before the separation the spectral index is @xmath74 .
the central 35@xmath15 region ( at @xmath94 , @xmath95 ( j2000 ) ) has a flux density of 226@xmath1913 mjy at 333 mhz while at 1465 mhz the same region has a flux density of 117@xmath196 mjy .
this corresponds to an @xmath4 of @xmath96 at a resolution of @xmath97 .
however , after the separation , similar region has a @xmath3 of @xmath98 ( see figure [ figure3 ] ) . in the nuclear region about 15% of the total emission is due to thermal emission at 333 mhz , higher than the rest of the galaxy , where the thermal emission lies within 10% .
the 333 mhz gmrt observations of six nearby galaxies ( greater than 10@xmath56 in size ) presented in this paper are by far the highest angular resolution maps for these galaxies at this frequency which are sensitive to both small ( @xmath99 kpc ) and large ( @xmath100 kpc ) scale structures .
the hybrid array of gmrt enables us to recover emission from both the diffuse and small scale emission in these sources , which we have adequately verified by comparing our estimated integrated flux densities with those available in the literature .
our observations have comparable resolution as that attained at higher frequencies , and probe linear scales of about 0.41 kpc . at 333 mhz
the galaxies appear smoother than at higher frequencies , where spiral arms are often discernable .
the thermal emission ( which is primarily associated with ionized structures observed in h@xmath4 maps ) for all the galaxies was found to be less than 5% of the total emission , and hence at these frequencies , the emission is essentially nonthermal .
the smooth appearance is consistent with the conjecture that nonthermal emission at 333 mhz primarily results from a population of old ( @xmath101 yrs ) cre , which diffuses away from their formation sites filling up a volume of radius @xmath102 1 kpc , without loosing much energy . this scale is larger than the width of the gaseous spiral arms and comparable to the width of the inter - arm region ( see e.g. @xcite ) , hence reducing any contrast in intensity across the galaxy .
the prominent radio structures observed at higher frequencies ( @xmath103 1 ghz ) are visible solely due to the increased thermal fraction . we have determined robust estimates of spatially resolved @xmath3 ( see section [ section3 ] ) for five galaxies in our sample , with a resolution of 40@xmath15 , corresponding to a linear scale of 13 kpc , which also corresponds to the diffusion scale of the cre . the observed @xmath3 ranges from @xmath104 to @xmath105 in various regions of the galaxies .
generally cre are thought to be accelerated at snr where both the theoretical prediction @xcite and observation based on galactic snr suggest mean @xmath106 @xcite .
the snr are short lived ( @xmath107 yrs ) , while the cre diffuse away from their acceleration sites losing their energy through several physical mechanisms for typically 10@xmath108 yrs , and hence their initial energy spectrum gets distorted leading to a change in @xmath3 .
if the energy losses are due to ionization , @xmath3 flattens @xcite .
synchrotron radiation and inverse compton scattering loss leads to a steepening of the spectra , while adiabatic cooling and bremsstrahlung keep the spectra unchanged . to assess the behaviour of @xmath3 in spatially resolved parts of the galaxy we plot @xmath3 distribution ( as shown in figure [ figure9 ] ) for two regimes of thermal fraction : @xmath109 ( filled grey histogram ) and @xmath110 ( unfilled histogram ) ,
using data from 4 galaxies and @xmath3 might be steeper than the actual value . ] . to do this ,
we have computed @xmath3 by dividing the galaxy in rectangular grids of 40@xmath15 .
the histograms are binned with step size 0.05 in @xmath3 .
the high thermal fraction corresponds to bright regions in h@xmath4 which traces gas ionized by massive ob stars in star forming regions .
these regions host the cre generation and acceleration sites .
the distribution of @xmath3 for such high thermal fraction regions has a gaussian shape , with mean @xmath111 and narrow spread of 0.16 .
it is interesting that this narrow distribution arises from multiple spatially resolved regions from an assorted set of galaxies with very different star formation histories .
this is clearly indicative of a generic process of energy loss of cre by synchrotron radiation and inverse compton scattering as they propagate from their acceleration sites .
the flatter part of @xmath3 ( @xmath112 ) primarily arises from ngc 5236 , indicating that ionization losses are dominant @xcite .
@xcite , based on integrated flux density of 74 galaxies found the average @xmath113 which is comparable to the @xmath3 value that we have obtained .
the galaxy integrated spectral index is an average over tens of kpc where the flux density is highly dominated by bright regions which are mostly spiral arms or regions .
our analysis shows that the contribution to the average spectral index primarily arises at least on kpc scales .
the unfilled histogram , corresponding to low thermal fraction ( @xmath114 ) regions , has a much larger spread ranging from @xmath115 .
the steepening in the range @xmath116 arises from the interarm regions , and indicates energy losses through synchrotron emission and inverse compton scattering .
@xmath117 , is mostly from low flux density regions , primarily in the outer parts of the galaxies , and may be affected due to missing flux density in the higher radio frequency maps .
note that we have averaged in a region 40@xmath15 , which for all these galaxies corresponds to @xmath2 1 kpc to 3 kpc , and hence the flatter part of @xmath3 distribution may be affected due to bright regions .
as mentioned earlier , the absolute value of @xmath3 can change by @xmath210% due to uncertainty in flux scale at 333 mhz . a recent study by @xcite for ngc 628 ,
ngc 3627 and ngc 7331 , using infrared emission ( @xmath770@xmath8 m ) as a tracer of star forming regions , found that @xmath4 obtained using vla observations at 325 mhz and near 1.4 ghz , is flatter for infrared bright regions .
this is similar to what we observe in our set of galaxies .
the high angular resolution 333 mhz observations presented in this paper can be used to study in detail the spatially resolved fir radio and co fir correlations , which will be reported in a future paper .
the @xcite method uses an h@xmath4 map of the galaxy as a template of thermal emission .
since h@xmath4 emission is affected by extinction , far - infrared data at @xmath770@xmath8 m and @xmath7160@xmath8 m are used to obtain a model for extinction , which is then used to correct the h@xmath4 data .
the corrected h@xmath4 image is then used to predict the thermal emission at radio frequencies of interest ( the whole process is described in detail in sections 3 to 6 of @xcite ) . in this section ,
we summarize the various steps and the intermediate results that we obtained in the separation process . *
* data sets : * the far infrared images at @xmath118 m and @xmath119 m were obtained from the publicly available data from the _ spitzer infrared nearby galaxy survey _ ( sings ; @xcite ) . all @xmath118 m images have a pixel size of 4.5@xmath15 and a _ point spread function _ ( psf ) of about 16@xmath15 . at @xmath119 m ,
each pixel is 9@xmath15 in size and has a psf of 40@xmath15 .
both the @xmath118 m and @xmath119 m maps are calibrated in surface brightness units of @xmath120 .
+ the continuum subtracted h@xmath4 images for ngc 1097 ( 1.5 m ctio , filter : ct6586 ) and ngc 6946 ( kpno 2 m , filter : kp1563 ) were obtained from the ancillary data at the sings website .
the maps were in units of @xmath121 , which was converted to @xmath122 using the calibration provided in the sings fifth data delivery documentation .
+ the continuum subtracted h@xmath4 images for ngc 4736 , observed with the 1 m jkt at la palma , filter : ha6570 , @xcite , ngc 5055 observed with the 2.3 m telescope at kpno , filter : 6580 and ngc 5236 observed with the 0.9 m telescope at ctio , filter:6563 were downloaded from the ned . + the counts per second ( cps ) , in these maps were converted to apparent magnitude , @xmath123 , using the zero - point given in the fits header .
the apparent magnitude @xmath123 was further converted to specific intensity , using @xmath124 .
the flux in units of @xmath122 was computed using , @xmath125 . +
all the @xmath118 m and continuum subtracted h@xmath4 maps were convolved to the coarser 40@xmath15 psf of @xmath119 m , and pixels were re - gridded to 9@xmath15 .
all the subsequent calculations were done on a pixel by pixel basis . * * dust temperature and optical depth : * + following @xcite , as a first step for extinction correction of h@xmath4 maps we need to estimate the optical depth due to obscuring dust . for this
the colour temperature of the dust ( @xmath126 ) is found by fitting a black body spectrum incorporating dust absorption efficiency ( see equation 1 of @xcite ) to the far infrared maps at @xmath770@xmath8 m and @xmath7160@xmath8 m .
the distribution of the dust temperature for all the five galaxies is shown in the left panel of figure [ figurea1 ] .
the mean temperature estimated for all the galaxies is between 19@xmath12722 k. note that a single dust temperature model has been assumed .
the optical depth at 160@xmath8 m ( @xmath128 ) was then obtained by using equation 2 of @xcite using the estimated @xmath126 .
the h@xmath4 optical depth was obtained as @xmath129 @xcite with h@xmath4 filling factor @xmath130 @xcite .
for all the five galaxies the @xmath131 lies in the range 0.03 0.6 at linear scales @xmath103 1 kpc , with the highest values being towards the center of the galaxies . * * emission measure : * + extinction corrected h@xmath4 maps was used to estimate the emission measure ( @xmath132 pc @xmath133 , where @xmath134 is the thermal electron density ) using equation 3 and 4 of @xcite ( also see @xcite ) and assuming an electron temperature of @xmath135 k. the emission measure radial profile for all the five galaxies is shown in the middle panel of figure [ figurea1 ] .
the emission measure towards the center of the galaxy lies in the range of @xmath136 pc @xmath133 and falls off to @xmath137 pc @xmath133 towards the edge of the galaxy .
these range of emission measure values are consistent with those observed in the milky way ( see e.g. @xcite ) . * * thermal emission at radio frequencies : * + equation 5 and 6 of @xcite , were used to determine the radio continuum optical depth and the brightness temperature ( @xmath138 ) using the rayleigh - jeans approximation . the thermal flux density ( @xmath139 ) at a radio frequency @xmath140 was obtained from the @xmath138 using , @xmath141 here , @xmath142 and @xmath143 size of the restoring beam in the radio map ( in this case 40@xmath15 ) .
+ for all the five galaxies , the pixel - wise distribution of the thermal fraction ( @xmath144 ) at 333 mhz and near 1 ghz are shown in the rightmost panel of figure [ figurea1 ] .
the mean @xmath6 at 333 mhz and near 1 ghz for all the galaxies were found to be less than 5% and less than 12% respectively .
however , in certain giant regions , the thermal fraction can go up to 10% at 333 mhz and 30% near 1 ghz .
@xcite pointed out that the primary source of error in the thermal fraction arises from the unknown value of the @xmath145 . using similar arguments
, we estimate the uncertainty in the thermal fraction at 333 mhz and near 1 ghz to be @xmath2 10% and @xmath146% respectively .
we thank the referee , elly berkhuijsen and the editor , for insightful comments that have improved both the content and presentation of this paper .
we thank the gmrt staff for technical support during the observations .
gmrt is run by the national centre for radio astrophysics of the tata institute of fundamental research .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
the digitized sky surveys were produced at the space telescope science institute under u.s .
government grant nag w-2166 .
the images of these surveys are based on photographic data obtained using the oschin schmidt telescope on palomar mountain and the uk schmidt telescope .
this work is based [ in part ] on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
99 arp , h. , 1973 , apj , 183 , 791 baars j. w. m. , genzel r. , pauliny - toth i. i. k. , et al . , 1977 , a&a , 61 , 99 beck , r. , grave , r. , 1982 , a&a , 105 , 192 beck , r. , 1991 , a&a , 251 , 15 beck , r. , shoutenkov , v. , ehle , e. , harnett , j. i. , haynes , r. f. , shukurov , a. , sokoloff , d. d. , thierbach , m. , 2002 , a&a , 391 , 83 beck , r. , fletcher , a. , shukurov , a. , et al .
, 2005 , a&a , 444 , 739 beck , r. 2007 , a&a , 470 , 539 bell , r. , 1978 , mnras , 182 , 443 berkhuijsen , e. m. , beck , r. , hoernes , p. , 2003 ,
a&a , 398 , 937 berkhuijsen , e. m. , mitra , d. , & mueller , p. 2006
, astronomische nachrichten , 327 , 82 biermann , p. l. , strom , r. g. , 1993 , a&a , 275 , 659 bogdan , t. j. , v@xmath147lk , h. j. , 1983 , a&a , 122 , 129 braun , r. , oosterloo , t. a. , morganti , r. , klein , u. , beck , r. , 2007 , a&a , 461 , 455 briggs , d. s. 1995 , ph.d .
thesis , new mexico institute of mining technology , socorro , new mexico , usa broadbent , a. , osborne , j. l. , haslam , c. g. t. 1989 , mnras , 237 , 381 brown , r. h. , hazard , c. , 1961 , mnras , 122 , 479 de bruyn , a. g. , 1977 , a&a , 54 , 491 carilli , c. l. , holdaway , m. a. , ho , p. t. p. , de pree , c. g. , 1992 , apj , l59 , 399 caswell , j. l. , wills , d. , 1967 , mnras , 135 , 231 chy@xmath148y , k. t. , buta , r. j. , 2008 , apj , 677 , l17 condon , j. j. , 1987 , apj suppl .
, 65 , 485 condon , j. j. , 1992 , ara&a , 30 , 575 condon , et al .
1998 , aj , 115 , 1693 cornwell , t. j. , perley , r. a. , 1992 , a&a , 261 , 353 dickinson , c. , davies , r. d. , davis , r. j. , 2003 , mnras , 341 , 369 duric , n. , dittmar , m. r. , 1988 , apj , 332 , l67 gioia , i. m. , gregorini , l. , 1980 , a&a suppl . , 41 , 329 gioia , i. m. , gregorini , l. , klein , u. , 1982 , a&a , 116 , 164 green d.a .
, 1998 , a catalogue of galactic supernova remnants ( 1998 september version ) .
mullard radio astronomy observatory , cambridge , united kingdom ( available on the world - wide - web at http://www.mrao.cam.ac.uk/surveys/snrs/ ) heesen , v. , beck , r. , krause , m. , dettmar , r. j. , 2009 , a&a , 494 , 563 hummel , e. , 1980 , a&as , 41 , 15 hummel , e. , bosma , a. , 1982 , aj , 87 , 242 de jong , m. l. , 1966 , apj , 144 , 153 de jong , m. l. , 1967 , apj , 150 , 1 karachentsev , i. d. , sharina , m. e. , huchtmeier , w. k. , 2000 , a&a , 362 , 544 karachentsev , i. d. , sharina , m. e. , dolphin , a. e. , et al . , 2002 , a&a , 385 , 21 karachentsev , i. d. , sharina , m. e. , dolphin , a. e. , et al . , 2003 , a&a , 398 , 467 kellermann , k. i. , pauliny - toth , i. i. k. , williams , p. j. s. , 1969 , apj , 157 , 1 kennicutt , jr . , r. c. , armus , l. , bendo , g. , et al .
2003 , pasp , 115 , 928 klein , u. , emerson , d. t. , 1981 , a&a , 94 , 29 klein , u. , beck , r. , buczilowski , u. r. , wielebinski , r. , 1982 , a&a , 108 , 176 klein , u. , grave , r. , wielebinski , r. , 1983 , a&a , 117 , 332 klein , u. , wielebinski , r. , morsi , h. w. , 1988 , a&a , 190 , 41 knapen , j. h. , stedman , s. , bramich , d .- m . , folkes , s. l. , bradley , t. r. , 2004 , a&a , 426 , 1135 knapik , j. , soida , m.,dettmar , r .- j . , beck ,
r. , urbanik , m. , 2000 , a&a , 362 , 910 kothes , r. , fedotov , k. , foster , t. j. , uyaniker , b. , 2006 , a&a , 457 , 1081 kr@xmath149gel , e. 2003 , the physics of interstellar dust , the physics of interstellar dust , ed .
e. kruegel ( iop series in astronomy and astrophysics , bristol , uk : the institute of physics ) .
kurilchik , v. n. , onishchenko , l. v. , turyvskii , v. m. , 1967 , sov .
j. , 11 , 528 kurilchik , v. n. , andrievskii , a. e. , ivanov , v. n. , spangenberg , e. e. , 1970 , sov .
j. , 13 , 881 van der kruit , p. c. , allen , r. j. , rots , a. h. , 1977 , a&a , 55 , 421 longair , m. s. , 2011 , high energy astrophysics 3-ed , cambridge university press .
maddox , l. a. , cowan , j. j. , kilgard , r. e. , lacey , c. k. , prestwich , a. h. , stockdale , c. j. , wolfing , e. , 2006 , aj , 132 , 310 mayya , y. d. , carrasco , l. , luna , a. , 2005 , apj , 628 , l33 murgia , m. , helfer , t. t. , ekers , r. , blitz , l. , moscadelli , l. , wong , t. , paladino , r. , 2005 , a&a , 437 , 389 murphy , e. j. , helou , g. , condon , j. j. , schinnerer , e. , turner , j. l. , beck , r. , mason , b. s. , chary , r .- r . , armus , l. , 2010 , apj , 709 , l108 nagar , n. m. , falcke , h. , wilson , a. s. , 2005 , a&a , 435 , 521 niklas , s. , beck , r. , 1997 , a&a , 320 , 54 niklas , s. , klein , u. , wielebinski , r. , 1997 , a&a , 322 , 19 ondrechen , m. p. , van der hulst , j. m. , 1983 , apj , 269 , l47 ondrechen , m. p. , van der hulst , j. m. , hummel , e. , 1985 , aj , 90 , 1474 ondrechen , m. p. , 1989 ,
apj , 342 , 39 paladino , r. , murgia , m. , orr@xmath150 , e. , 2009 , a&a , 503 , 747 rieke , g. , et al . , 2004 , apjs , 154 , 25 roy , s. , rao p. a. , 2004 , mnras , 349 , l25 schommer , r. a. , sullivan , w. t. , 1976 , ap . letters , 17 , 191 seaquist , e. r. , odegard , n. , 1991 , apj , 369 , 320 segalovitz , a. , 1977a , a&a , 61 , 59 segalovitz , a. , 1977b , a&a , 54 , 703 smith , b. j. , lester , d. f. , harvey , p. m. , pogge , r. w. , 1991 , apj , 380 , 677 steer , d. g. , dewdney , p. e. , ito , m. r. , 1984 , a&a , 137 , 159 sukumar , s. , klein , u. , grave , r. , 1987 , a&a , 184 , 71 swarup , g. , ananthakrishnan , s. , kapahi , v. k. , et al .
1991 , current science , 60 , 95 tabatabaei , f. , beck , r. , krugel , e. , , krause , m. , berkhuijsen , e. m. , gordon , k. d. , menten , k. m. 2007 , a&a , 475 , 133 valls - gabaud , d. , 1998 , pasa , 15 , 111 de vaucouleurs , g. , et al . 1991 , third reference catalogue of bright galaxies ( berlin : cambridge university press ) wong , t. , blitz , l. , 2001 , apj , 540 , 771 wright , a. e. , wark , r. m. , troup , e. , otrupcek , r. , hunt , a. , cooke , d. j. , 1990 , pasa , 8 , 261 .
|
we report giant meterwave radio telescope ( gmrt ) continuum observations of six nearby normal galaxies at 333 mhz .
the galaxies are observed with angular resolutions better than @xmath0 ( corresponding to a linear scale of about 0.4 1 kpc ) .
these observations are sensitive to all the angular scales of interest , since the resolution of the shortest baseline in gmrt is greater than the angular size of the galaxies .
further , for five of these galaxies we show that at 333 mhz , the mean thermal fraction is less than 5% . using archival data at about 1 ghz , we estimate the mean thermal fraction to be about 10% at that frequency .
we also find that the nonthermal spectral index is generally steeper in regions with low thermal fraction and/or located in the outer parts of the galaxy . in regions of high thermal fraction , the nonthermal spectral index is flatter , and has a narrow distribution peaking at @xmath1 with a spread of 0.16 , putting stringent constraints on the physical conditions for generation , diffusion and energy losses of cosmic ray electrons at scales of @xmath2 1 kpc .
[ firstpage ] galaxies : spiral galaxies : individual ( ngc 1097 , ngc 3034 , ngc 4736 , ngc 5055 , ngc 5236 , ngc 6946 ) ( ism : ) cosmic rays ( ism : ) regions techniques : interferometric techniques : image processing
|
pain is a common health problem that causes over $ 60 billion a year in productivity losses in the united states @xcite .
research on pain sensing ( also known as nociception ) with human subjects is difficult for a number of reasons including ethical constraints , difficulties in quantifying a psychophysical response , and subjectivity in self - reporting @xcite .
since the molecular mechanisms of pain transduction are believed to be partially conserved among many different species @xcite , some of these problems are solved by using animal models in nociception research .
however , animal subjects can not communicate their perceived pain levels to us in an obvious fashion .
thus progress in using animal models depends crucially on our ability to quantitatively and objectively infer the perceived level of noxious stimuli from animal behavior .
historically , studies of pain primarily used mammalian models @xcite . for rodents , the tail flick test @xcite , the hot - plate test @xcite and the hargreaves method @xcite correlate pain perception with the reaction latency of different body parts to noxious stimuli . for larger animals such as canines @xcite and primates @xcite , similar nociceptive assays have also been developed . in recent years
, new approaches started incorporating facial expressions in nociception quantification @xcite .
although these mammalian models are extensively studied , several drawbacks hinder their use .
first , ethical issues and risks arise for certain experiments .
second , compared to lower - level organisms , mammalian subjects require more time and resources to maintain .
therefore , much effort has been devoted to investigations of the possibility of using invertebrate models in nociception research @xcite . in experiments involving @xmath0 larva ,
measures such as the response percentage of the total population @xcite and the time to response @xcite have been used to investigate changes in the ability of the animals to sense noxious stimuli . in experiments on _ caenorhabditis elegans _ , behavioral features such as the turning rate @xcite and the percentage of escape response @xcite have been used to characterize nociception .
all of these models share some common problems .
first , the nociceptive assays focus on one particular coarse behavioral feature of the subject , such as avoidance behavior , orientation , or turning rate .
such features are selected in an ad hoc fashion , subject to a particular design of an experiment .
this makes it difficult to compare results across different labs and experiments .
further , the behavior may be providing additional information about the perceived pain that is not being captured by the coarse measures .
second , some assays report measurements as a percentage of a population , so that these measurements can not be made for individuals . to overcome these problems ,
an ideal assay would infer a perceived pain level of an individual animal on a continuous scale , using comprehensive , objective measurements of its behavioral profile .
an even more important problem is that one would like to use the assays to calibrate the perceived pain level , and maybe even reductions in such levels due to effects of analgesic drugs , or mutations in the nociceptive pathways . at the same time ,
drugs or mutations can affect the motor response , rather than the nociception per se .
thus traditional pain assays mentioned above may convolve the perceived pain reduction , if any , with behavioral changes .
for example , a mutant defective for turning behavior will register a strong reduction in the turning rate , but it would be a mistake to interprete this as a reduction in the pain perception .
such concerns are very real , as is illustrated by a known fact that opiods can cause large behavioral changes , and thus they have effects beyond analgesic ones @xcite . to attribute a behavioral response difference to reduced nociception and not to motor changes , the response must be stereotyped and reflexive , which is often the case @xcite .
further , only the response amplitude or frequency , but not the detailed temporal structure , should change in response to a drug or a mutation . establishing the stability of the stereotypic response pattern
requires analysis of the _ entire response behavior _ , rather than of its few selected features , as is done by most behavioral assays . in this work ,
we address these issues in the context of the nematode _ c. elegans _ , developing it further as an animal model system for nociception research .
the worm is a great model organism for such studies for a number of reasons .
first , the behavioral dynamics of freely moving _ c. elegans _ is intrinsically low dimensional @xcite .
this makes quantification of its behavioral response relatively straightforward , providing an opportunity to use the entire motile behavior as a basis for assays .
second , the worms show a noxious response to a wide range of sensations including certain types of chemical @xcite , mechanical @xcite , and thermal @xcite stimuli , and such nociceptive response is different from and is transduced independently of the related taxis behaviors @xcite .
third , at the molecular level , many details of thermal nociception in the worm may be similar to vertebrate animals @xcite .
fourth , there are powerful genetic and optical tools to reveal mechanisms of nociception in _ c. elegans_. finally , the low cost , small size , and absence of ethical constraints make the animal amenable to large scale pharmacological screens for new analgesics @xcite .
we present combined experimental and modeling studies that show that the entire temporal behavioral profile during the thermal nociceptive response in _ c. elegans _ is highly stereotypical , with the frequency of the escape response and the amplitude of the escape velocity profile scaling with the stimulus level . by verifying the ability of the behavioral template to capture the response following a nociceptive stimulus
, the model we develop distinguishes changes in nociception from changes to the motor program . when a change is attributed to nociception
, the model can infer the reduction in the perceived thermal nociception level following pharmacological or genetic treatments from the behavior of an individual worm .
this quantification requires only a handful of worms ( about 60 ) to show statistically significant nociception reduction for a common analgesic , and its statistical power quickly improves with an increasing number of subjects .
the paper is organized as follows .
first we discuss the structure of the dataset and the model .
then we evaluate the performance of the model and discuss the stereotypical behavior we discovered .
further , we use the model to infer the nociceptive stimulus level of worms in three different conditions : wild - type untreated , wild - type treated with ibuprofen , and a mutant with defects in thermal nociception .
we argue that , while effects of the ibuprofen treatment can be attributed largely to a reduced nociception , the mutant s response shows changes to the behavior beyond nociceptive effects .
finally , we use the statistical model to discuss how the nociception experiments should be designed to achieve the highest statistical power to quantify analgesic effects .
we aim to infer the perceived worm nociception level from the temporal dynamics of the worm response .
the noxious signal is administered by thermal stimulation using an infrared laser while the worm crawls on an agar plate . the worm motion is captured by video microscopy and analyzed using custom image analysis software .
the worm postures are very stereotypical , adding up to simple sinusoidal motions forwards or backwards , and to turns @xcite .
thus without much loss to the statistical power , we characterize the entire nociceptive behavior of the animal by a time series of its center of mass velocity ( see _ materials and methods _ ) . our task is then to verify if such responses are stereotypical , scaling in frequency and amplitude with the applied laser current .
if they are stereotypical and thus can be used to characterize the perceived nociception , the next task is to infer the applied laser current from the velocity data . for each noxious stimulus trial
a random worm is selected on the plate and its motion is sampled at 60 hz for 15 s. an infrared laser pulse with a randomly chosen current between 0 to 200 ma and a duration of 0.1 s is then directed to the head of the worm 1 s after the start of the video recording .
the worm s center of mass motion in a typical trial consists of a forward motion before the stimulus , a stop and/or backwards motion after the stimulus , then followed by an `` omega '' turn , after which the worm emerges with a forward motion in a different direction , cf .
[ fig : cm_velocity ] .
this is a typical response to many noxious stimuli in _ c. elegans _ , and not just to thermal noxious stimulus @xcite
. such nociceptive response , at least at large stimuli levels , is different from the more commonly studied thermotactic behavior , and is mediated by different neural and genetic pathways @xcite .
s. the paused state is characterized by the near zero velocity after the laser stimulus .
( b ) actual trajectories of these two worms .
worm changes its direction of motion in two ways : in a `` reversal '' , it stops and backtracks along its previous path ; in an `` omega '' turn the worm s head curls back and crosses the tail , setting then a new direction of forward motion .
[ fig : cm_velocity ] ] to understand the effects of pharmacological and genetic interventions on nociception transduction , we collected three distinct datasets . in the first , the stimulus was applied to wild - type worms ( n2 ) . in the second ,
the wild - type worms were pre - treated with an ibuprofen solution ( see _ materials and methods _ ) . in the third
, we applied the stimulus to an untreated triple mutant ( ocr-2(ak47 ) osm-9(ky10 ) iv ; ocr-1(ak46 ) ) .
hereafter we refer to these datasets as `` control '' , `` ibuprofen '' , and `` mutant '' , respectively .
we collectively refer to ibuprofen and mutant worms as `` treated '' worms . forward
motion in all three data sets was similar ( typical forward velocity of @xmath1 , @xmath2 , @xmath3 pixel / s respectively , where the error denoted the standard deviation of the velocity distribution ) , so that there are no drastic defects in motility .
s. e. m. ) of the maximum reverse speed for the three worm types is plotted against the stimulus laser current , partitioned into five bins .
negative values correspond to backward motion .
( b ) maximum reverse speeds for individual control worms .
we define an active worm ( dot ) as having the maximum reverse speed of over 10 pixels / s ( about 0.2 body length per second ) ; otherwise the worm is paused ( plus sign ) .
( c ) probability of the paused state vs. the laser current .
dots represent the actual data ( @xmath4 s. e. m. ) , and lines are the fitted model .
the datasets are divided into equally sampled bins for each worm type .
[ fig : p and a state demo ] ] we bin the laser current of the nociceptive stimulus into five distinct levels ( bins ) , defined to have an equal number of control worms in each bin ( 40 per bin ) .
the maximum reverse escape velocity is indistinguishable among all three worm types for the largest stimulus level , indicating again no gross defects to motility or noxious response . at the same time , the velocity at smaller stimuli levels shows substantial differences ( @xmath5 scores of up to 3.8 ) between the control and the treated worms , especially in the vicinity of @xmath6 ma laser stimulation , see fig .
[ fig : p and a state demo](a ) .
mutant worms are especially different from the control over a wider stimulus range . however , as discussed above , it is unclear if such simple observed behavioral differences are indicative of the reduction of the perceived nociception level or of other changes to the motor response . to address this , instead of subjectively segmenting the complex escape behavior or choosing _ ad hoc _ metrics of the worm s movement , we choose to infer the applied stimulus strength from a comprehensive model of the entire worm s response velocity profile
due to the considerable randomness and individual variability of responses , we choose to model them probabilistically .
thus we are interested in estimating @xmath7 , the probability distribution of the applied laser current @xmath8 conditional on the observed velocity of the escape response @xmath9 . using bayes theorem ,
we write @xmath10 where @xmath5 is the normalization factor , and @xmath11 is the prior distribution of stimuli . when we characterize the velocity profiles of the noxious response of _ c. elegans _ , we notice that the worm can react to the stimulus in two different ways .
some worms pause after the thermal stimulus , even at large laser currents ( see fig . [
fig : p and a state demo ] ( b ) ) .
these worms remains largely immobile for a few seconds , sometimes as long as the recording duration .
other worms actively reverse and follow the classic escape behavior as shown in fig .
[ fig : cm_velocity ] . we choose to separate the active vs. the paused worms with a cutoff of 10 pixels / s , where 50 pixels is about one body length of the worm . to account for this heterogeneity in the behavior
, we introduce the state variable @xmath12 , which can take one of two values , @xmath13 or @xmath14 , for each individual worm .
then @xmath15 we model the probability of the paused state @xmath16 by a sigmoid function , cf .
[ fig : p and a state demo](c ) , @xmath17 where @xmath18 is the _ pause current _ threshold .
then the probability of the active state is @xmath19 we infer @xmath18 from data by maximizing @xmath20 , where @xmath21 is the number of trials with worms of the analyzed type , and @xmath22 is the actual laser current for a particular trial .
note that each of the three data sets has its own pause current ( @xmath23 , @xmath24 , and @xmath25 ma for the control , ibuprofen , and mutant worms , respectively ) . changes in this threshold , like that for the mutant , will result in different numbers of worms responding to the same stimulus , which can be consistent with the changes in the nociception level , depending on whether the response profiles themselves stay stable .
this is what we investigate next .
parenthetically , we note that the fraction of active worms is essentially the same as the percentage of the escape response , which has been used previously to quantify worm nociception @xcite . here
we go further and additionally analyze the behavioral profiles of the responding worms . _
c. elegans _ locomotion consists of a series of stereotyped postures and behavioral states @xcite .
further , in other animals , nociceptive responses are stereotyped as well @xcite .
therefore , it is natural to explore if the nociceptive response of _ c. elegans _ is also stereotyped , separately for the paused and the active states . for paused worms , the escape velocity is small and independent of the laser current , and we model it as a multivariate normal variable , @xmath26,\ ] ] where @xmath27 is the mean velocity profile of the paused worms measured from data , which we call the _ paused template _ velocity .
@xmath28 is the empirical covariance of the paused velocity , and @xmath29 is the total number of effectively independent time points in the velocity profile time series , determined using the autocorrelation structure of the profile ( see _ materials and methods _ ) .
we expect that , in the active state , the worm escape is laser current dependent . specifically , we seek to represent it by a current - dependent rescaling of a stereotypical escape velocity , @xmath30 , where @xmath31 is the scaling function , and @xmath32 is the _ active template _ velocity . since various features of the worm escapes ( the maximum reverse speed , the maximum reverse acceleration , and the time to the omega turn ) , scale non - linearly and saturate with the laser current ( cf .
[ fig : p and a state demo](a ) ) , the rescaling , @xmath33 , must be sigmoidal .
further , some worms have nonzero velocities even at zero laser current , so that @xmath34 may be nonzero .
finally , the overall scale of the template can be absorbed in the definition of @xmath32 .
the simplest scaling function obeying these constraints has only two parameters @xmath35 where @xmath36 and @xmath37 are again constants , different for the three different worm types . with this , we write the probability of a velocity profile given the laser current @xmath8 for the worm in an active state as a multivariate normal distribution @xmath38 , \label{pva}\ ] ] where @xmath39 is the covariance of the average velocity profile .
we find the constants @xmath36 and @xmath37 , @xmath32 and @xmath39 by maximizing the likelihood of the observed data ( see _ materials and methods _ ) . in summary ,
the probability of a velocity profile given the laser current in a certain trial is @xmath40\right\}\\ + \left\{\frac{(i/{{\mathcal i}}_0)^{2}}{1+(i/{{\mathcal i}}_0)^{2}}\times\frac{1}{(2\pi)^{\frac{t}{2}}|\sigma_{a}|^{\frac{1}{2 } } } \exp\left[-\frac{1}{2}({{\bf v}}-f_{{{\mathcal i}}_1,{{\mathcal i}}_{2}}(i){{\bf u}}_{a})^{t}\sigma_{a}^{-1}({{\bf v}}-f_{{{\mathcal i}}_1,{{\mathcal i}}_{2}}(i){{\bf u}}_{a})\right]\right\}. \label{eq : fullmodel}\end{gathered}\ ] ] the overall model of the experiment , eq .
( [ eq : bayes ] ) , also includes @xmath11 . to a large extent , this is controlled by the experimentalist , and details are described in _ materials and methods_. the model we built assumes a stereotypical nociceptive response .
is this assumption justified ?
velocities in the paused state are very small ( worms barely move ) .
thus whether the stereotypy assumption provides a good model of the data is determined largely by the stereotypy of the active worms .
if the active stereotypical response template exists , then it should be possible to collapse the average velocity profiles onto a single curve by the following transformation @xmath41 figure [ fig : collapse ] confirms this : the means of different bins collapse relatively compactly , providing evidence for the existence of the stereotypy in active nociceptive responses .
we illustrate the template velocities and the non - linear scaling function @xmath31 inferred from the control , ibuprofen , and mutant in fig .
[ fig : the - normalized - active ] . worms each , based on the stimulus current .
( b ) the mean and standard deviation of velocities in the same time period rescaled by @xmath42 .
rescaled mean velocities nearly collapse .
note that the parameters @xmath36 and @xmath43 are optimized as in _ materials and methods _ to collapse individual profiles , and not the five mean profiles illustrated here .
[ fig : collapse ] ] s corresponds to the moment of the stimulus application .
( b ) normalized active template velocity .
each template velocity is normalized by the absolute value of its minimum @xmath44 .
the error bars represent the model standard deviation , estimated by bootstrapping ( see _ materials and methods _ ) .
the subfigure shows the template velocities adjusted by the time of maximum reverse velocity , illustrating that the three templates nearly match .
( c ) the rescaling function @xmath45 for the three different worm types , normalized by multiplying by the absolute value of the minimum of the active template velocity @xmath46 .
the optimized parameter values are @xmath47 and @xmath48 for the control , ibuprofen , and mutant worms , respectively .
we deliberately do not report error bars on individual parameters , but rather show the one standard deviation confidence intervals for the entire rescaling curve as shaded regions on the figure .
the confidence region was again estimated using bootstrapping.[fig : the - normalized - active],scaledwidth=100.0% ] note that the three active template velocity profiles in fig .
[ fig : the - normalized - active](a ) are very similar , but the mutant template velocity profile shows a response lag of @xmath49 ms ( one time frame at 12 hz ) compared to the control or ibuprofen data set . in other words , the pharmacological treatments and the mutations weakly affect the templated response , save for possibly delaying it .
this bodes well for the assumption of the stereotypical response , definitely for ibuprofen and , to a somewhat lesser extent , for the mutant .
while the existence of the stereotypical patterns and their similarity across treatments is encouraging , we still need to quantify how good the statistical models are . in the ideal case ,
the variance @xmath50 of the collapsed velocities , eq .
( [ eq : collapse ] ) , calculated over individual trials , would be zero .
however , there are a number of expected sources of variance in the velocity , such as the individual variability and the model inaccuracies . to establish how good the stereotypical model fits are
, we need to disambiguate these contributions . for this
, we again partition all velocity profiles into five current bins .
we then write the total variance of all responses as @xmath51 where @xmath52 is the variance due to individual responses within each bin , and @xmath53 is the current - driven variance of the mean responses across the bins .
since the individuality of the worms is not accounted for in our model , @xmath53 represents the maximum potentially explainable variance in the data .
the stereotypy - based model would be nearly perfect if @xmath54 were to drop to zero after the @xmath55 rescaling . to explore this , in fig .
[ fig : var ] , we plot the total variance of the active response @xmath56 , panel ( a ) , and the fraction of the potentially explainable variance , @xmath57 , panel ( b ) .
the latter varies from 20% to 40% of the total variance , depending on the time post - stimulus and on the treatment . in both panels , the mutant and the control dataset
are nearly indistinguishable , while the ibuprofen worms show a smaller variance , and a smaller fraction of the explainable variance .
this is consistent with a smaller stimulus - driven nociceptive response for this analgesic treatment . at the same time , these figures suggest that the decrease in the maximum reverse speed in the mutant worm , fig . [ fig : p and a state demo ] ( a ) , should not be attributed entirely to the reduced nociception .
indeed , the similarity of the variance and the explainable variance in the control and the mutant worms , which have very different mean maximum reverse velocities , suggests existence of an additional ( explainable , non - templated ) component in the nociceptive response of the mutant , which is not present in the control . .
note that these numbers depend strongly on the preprocessing of the data in particular , smoothing of the velocity ( with the 500 ms filter , see _ materials and methods _ ) decreases the total variance . in ( b )
we show the fraction of potentially explainable variance @xmath58 .
finally , ( c ) shows the part of the explainable variance that was not explained by our statistical model @xmath59 .
see the main text for the discussion of the differences between the three worm types on this plot .
[ fig : var],scaledwidth=100.0% ] the explainable variance @xmath53 is further split into the variance explained by the model , @xmath60 , and the residual variance , @xmath61 , which the model fails to explain : @xmath62 in fig .
[ fig : var ] ( c ) , we plot @xmath63 , which is the fraction of the variance not captured by our model . of the explainable variance ,
about 80% is captured by our model for the control and the ibuprofen worms in the window between 1 s and 3.3 s since the start of the trial , on average .
this is a relatively large fraction for behavioral data , and provides an additional validation for our choice of a stereotypy - based model for representing nociceptive escape behavior in these worms .
stability of the template itself between these two conditions , and the stability of the fraction of the explained variance suggest that much of the effect of ibuprofen can be attributed to scaling of the templated response ( and also the fraction of active worms ) . in other words
, ibuprofen decreases the sensed nociception .
in contrast , the unexplained variance for the mutant is about twice as large as that for the control , and approaches 100% at @xmath64 s. this again illustrates that the templated response model is not very good for this treatment .
thus the mutations introduce changes in the behavior that are not consistent with a simple rescaling
mutations affect the fine motor behavior in addition to nociception per se .
one of the goals of our study is to develop methods for quantitative assessment of the efficacy of pharmacological interventions to decreases sensed nociception , at least in those cases where their action can be specifically interpreted as a change in nociception transduction .
we can use the developed statistical model for this . specifically , taking the model derived from the control worms , we can infer the laser current from behavior of all three different worm types . to the extent that the current inferred for the treated worms
is smaller than that for the control worms at the same applied current , the nociception level perceived by the treated worms is smaller .
figure [ fig : pvsa prob map = 000026 exp value ] shows the overall structure of the inference done with the model . in the first row , we plot the conditional distribution of the inferred laser current given the actual applied current @xmath8 for the three worm types , @xmath65 . for presentation purposes , we again bin the trials using @xmath22 into five bins @xmath66 , @xmath67 as before , and plot @xmath68 here @xmath69 is 1 if the stimulus on the @xmath70th trial for this worm type was in the @xmath66 bin , and zero otherwise .
further , @xmath71 is given by the full model , eq .
( [ eq : fullmodel ] ) , with the parameters inferred for the control worm , and with the empirically observed velocities @xmath72 in trial @xmath70 for each worm type .
we see that there is more probability concentrated at small @xmath73 for the ibuprofen and the mutant worms , suggesting a reduction in the perceived stimulus level . similarly , in the second row in fig .
[ fig : pvsa prob map = 000026 exp value ] , we plot the expected value , @xmath74 , of the distribution of the current inferred using the control model , @xmath71 , for each of the individual trials in each of the three worm types . to the extent that the dots for the ibuprofen and the mutant worms are again somewhat lower than for the control worms , there is some reduction in the perceived current by this measure as well . however , these population averaged results wash out important differences in the structure of the nociception response . to quantify these small effects more accurately , we now look at the perceived stimulus changes for individual worms in the datasets .
specifically , for each trial @xmath70 in the control dataset , a trial @xmath75 with the closest value of the applied laser current is found in the ibuprofen / mutant dataset ( the mean magnitude of the current mismatch is @xmath76 ma for both the ibuprofen and the mutant worms ) .
we then use the control model to calculate the expected value of the inferred current for the @xmath77th trial in the ibuprofen / mutant datasets .
this expectation is subtracted from the expectation value of the inferred current for the matched trial @xmath70 for the control dataset .
the difference of the expectation values , averaged over all control worms , is our measure of the reduction in the perceived nociception level @xmath78 we evaluate @xmath79 for different worm types and for controls worms binned into the five usual current bins , and show the results in fig .
[ fig : delta_ipred ] . to estimate the error of @xmath79
, we bootstrap the whole analysis pipeline , see _ materials and methods_. there is a statistically significant difference in nociception perception between the ibuprofen and the control worms .
the difference is most significant when the actual laser current is around 100 - 110 ma .
this coincide with our observation from fig .
[ fig : p and a state demo ] ( a ) that the most sensitive region of maximum reverse speed is around 100ma .
indeed , at smaller currents , the nociception level is small , many worms pause , and the behavior can not be used to reliably estimate the stimulus level . at high current ,
the noxious stimulus perception saturates , and all worms behave similarly , again reducing the ability to disambiguate the applied current level .
this analysis of ibuprofen worms achieves one of our main goals .
it proves our ability to reconstruct stimulus from the behavior , and shows that analgesic effects of pharmacological perturbations can be quantified from the behavior . at the same time
, @xmath80 turns out to be insignificant in fig .
[ fig : delta_ipred ] ( b ) , even though fig .
[ fig : p and a state demo ] ( a ) suggest that a large statistically significant difference exists between the mutant and the control behaviors . this failure to detect a significant nociception reduction is because the templated response model is not very good for the mutant worm ; thus our analysis can not reliably assign a mutant trajectory on a given trial to a specific nociception level .
in other words , the large error bars in fig . [
fig : delta_ipred ] ( b ) serve as yet another check for self consistency : effects of the mutations can not be attributed just to changes in nociception perception . .
inference is done with the model of the control worm behavior .
panels ( a ) , ( b ) , ( c ) show the results for the control , ibuprofen , and mutant worms , respectively .
the second row shows the expected inferred laser current for each trial , @xmath81 vs. the applied current @xmath22 .
the inference is again done using the control model , and panels ( d ) , ( e ) , ( f ) show the three worm types .
[ fig : pvsa prob map = 000026 exp value],scaledwidth=100.0% ] .
errors represent standard deviations , estimated with bootstrapping .
( b ) similar differences for the mutant worm , @xmath80 .
error bars are estimated by bootstrapping .
( c ) dependence of the statistical significance of the nociception level reduction ( measured by the @xmath5 score ) for ibuprofen at 110 ma on the number of ibuprofen trials .
[ fig : delta_ipred],scaledwidth=100.0% ]
we expect our analysis to be useable for screening large numbers of chemicals for analgesic action .
since our approach targets one individual worm at a time , we need to estimate the number of worms needed to achieve statistical significance in such screening experiments . for this
, we fix the number of control worms , arguing that these must be only analyzed once , and hence a relatively large number of them can be tested .
we then focus on ibuprofen , whose action is analgesic in our experiments , and on the bin at 110 ma , where the worms experience the most significant nociception reduction .
there are @xmath82 worms in this bin .
we randomly sample with replacements @xmath83 worms from among these ibuprofen - treated worms and repeat our analysis pipeline , estimating the @xmath84 . resampling 1000 times ( both the ibuprofen and the control datasets )
, we also estimate the variance of @xmath84 , and hence the @xmath5 score as a function of @xmath85 , which is plotted in fig . [
fig : delta_ipred ] ( c ) .
the plot here is an _ underestimate _ of the true @xmath5 score since resampling with replacements removes some stimulus values from the dataset , hence increasing the mismatch between control worms and the paired treated worms .
even with this , @xmath86 is achieved at @xmath87 ibuprofen worms . in other words , in a typical screening experiment , one would need to test @xmath88 or more worms to build the control model , and then at least @xmath89 worms additionally for each treatment condition .
typically the goal of sensory - response experiments is to develop a model that can predict the behavior in response to the stimuli .
here we wanted to do this in reverse .
our goal was to build a statistical model of the thermal stimulus from careful measurements of the nociceptive behavior of _ c. elegans _ , and to use this model to infer the changes in the perceived level of `` pain '' felt by the organism due to perturbation in the nociception transduction pathway . given this model , we could then measure changes in nociception perception due to effects of analgesics and mutations , and use this as a basis to study the mechanism of nociception transduction in a genetically tractable organism amenable to high - throughput screens . as a representative data set , we choose to study the standard laboratory _ c. elegans _
strain n2 , n2 treated with ibuprofen , and a mutant with defects in trpv function .
for the model to be successful , we had to meet a number of challenges .
since the worm could not communicate its nociception level to us directly we had to infer this level by reading the `` body language '' of the worm s `` pain '' response .
the difficulty with quantifying a behavioral response as a measure of nociception level is that drugs or mutations can affect locomotory behavior in addition to perturbing sensory transduction .
so in an attempt to deconvolve these effects , we used the entire behavioral profile instead of making ad hoc measurements and leveraged the fact that nociception responses are typically highly stereotyped .
this resulted in the discovery that the response can be modeled with a velocity profile template that scales non - linearly in response to an applied laser current .
the success of the template in modeling the stereotyped wild - type nociception response was confirmed by a functional collapse of the velocity profiles across different nociception levels .
this discovery of invariance is important since it not only allowed us to effectively correlate nociceptive behavior to the stimulus level , but it also allowed us to determine if the locomotory changes in our thermal nociception assay was due to changes specifically in the sensory transduction pathway or due to other general locomotory factors . by carefully accounting for the variation in our data and quantifying
how much of this variation is captured by the model , we showed that the stereotypical behavior is unaffected by ibuprofen , save for changing the amplitude of the response .
thus this drug application likely reduced the perceived nociception level in the worm .
in contrast , a trpv mutation changes locomotion in a way that is not as well captured by the template model .
thus we can be objectively critical about any nociception inference made with this strain .
the model was also useful in determining key experimental parameters for future measurements .
after verification that the model works well with the native and ibuprofen treated stimulus - response data , we quantified the changes in thermal nociception perception due to ibuprofen treatment .
our modeling and experimental assessment of thermal nociception identified the optimal stimulus range and required number of trials to determine statistically significant differences between the inferred current of n2 in the untreated and treated conditions . in conclusion
, we have built a general model that connects stereotyped behavior to stimulus . with a language to describe this relationship
, it is now possible to study quantitatively the effects of genetics and chemicals on this sensorimotor behavior .
we believe that the utility of the model is quite general and could be applied to different model system .
however , we particularly hope that this work helps further establish _ c. elegans _ as a model for nociceptive research .
all worms were grown and maintained under standard conditions @xcite , incubated with food at 20 @xmath90c .
well fed worms were washed twice then gently spun down for 1 minute and the supernatant discarded by aspiration .
for the drug application , 100 @xmath91l of ibuprofen in m9 at 100 @xmath91 m was added to the eppendorf tube . for the wild - type and mutant data set ,
m9 was used instead of the drug solution .
worms were then placed in an incubator for 30 minutes at 20@xmath90c .
after that worms were poured onto a seeded agar plate and transferred to agar assay plates by a platinum wire pick .
these assay plates were incubated at 20@xmath90c for 30 minutes , and then the experimental trials were done within the next 30 minutes . in total @xmath92 worms for the control group , @xmath93 worms for the ibuprofen group , and @xmath94 worms for the mutant group ( ocr-2(ak47 ) osm-9(ky10 ) iv ; ocr-1(ak46 ) )
group were tested .
the mutant strain was obtained from the caenorhabditis genetics center .
the thermal nociceptive stimulation instrument has been described previously @xcite . in summary ,
an infrared laser is directed to heat the head of a freely crawling worm ( @xmath95 mm fwhm ) on an agar plate .
the laser pulse is generated with a randomly chosen laser current between 0 to 200 ma , with a duration of 0.1 s. the heating of the worm is nearly instantaneous , and it is directly proportional to the current , between 0 and 2 @xmath90c for the the current range used in our experiments .
worms were stimulated only once and not reused .
the movements of the worms are imaged using a standard stereomicroscope with video capture and laser control software written in labview . for each nociception stimulus trial
, a random worm is selected on the plate and its motion is sampled at 60 hz for 15 s , and the laser is engaged 1 s after the start of the video recording .
the recorded nociceptive response videos were then processed with matlab to calculate the time series of the worm centroid motion as described previously @xcite .
all the worms that were not stimulated near the head or were not moving forward in the beginning of the video were discarded .
numerical derivatives of the centroid times series were then taken and filtered with a custom 500 ms gaussian filter , which was a one - sided gaussian at the edges of the recorded time period , becoming a symmetric gaussian away from the edges .
this removed the noise due to numerical differentiation and also averaged out the spurious fluctuations in the forward velocity due to the imperfect sinusoidal shapes of the moving worm .
we verified that different choices of the filter duration had little effect on the subsequent analysis pipeline .
the direction of the velocity was determined by projecting the derivative of the centroid time series onto the head - to - tail vector for each worm , with the positive and negative velocity values denoting forward / backward motion , respectively .
the filtered velocity profiles needed to be subsampled additionally .
this was because the statistical model of the data , eq . ( [ eq : fullmodel ] ) , involved covariance matrices of the active and paused velocity profiles , @xmath96 and @xmath97 ( note that velocity profiles are not temporally translationally invariant due to the presence of the stimulus , thus the full covariance matrix is needed , and not a simpler correlation function ) . to have a full rank covariance matrix
, the number of trials must be larger than the number of time points .
alternatively , regularization is needed for covariance calculation .
the autocorrelation function for all three worm types showed a natural correlation time scale of @xmath98 s , whether the data was pre - filtered or not .
thus subsampling at frequency @xmath99 hz would not result in data loss .
therefore , instead of an arbitrary regularization , we chose to subsample the data at 12 hz , leaving us with 37 data points to characterize the first 3 s of the worm velocity trace after the stimulus application , @xmath100 s since the start of the trial .
equation ( [ eq : fullmodel ] ) additionally needs knowledge of @xmath29 , the number of effectively independent velocity measurements in the profile .
this is obtained by dividing the duration of the profile by the velocity correlation time .
an uncertainty of such procedure has a minimal effect on the model of the experiment since it simply changes log likelihoods of models by the same factor , not changing which model has the maximum likelihood .
we then considered limiting the duration of the velocity profile used in model building : if velocities at certain time points do not contribute to identification of @xmath8 , they should be removed to decrease the number of unknowns in the model that must be determined from data ( values of the templates at different time points ) .
the first candidate for removal was the period of about 10 frames ( 0.16 s ) after the laser stimulation since worms do not respond to the stimulus so quickly . however , removal of this time period had a negligible effect on the model performance , and we chose to leave it intact .
in contrast , starting from 3.3 s ( 2.3 s after the stimulus ) the fraction of explainable variance drops to nearly zero ( cf .
[ fig : var ] ) since many worms already had turned by this time and resumed forward motion .
therefore , we eventually settled on the time in the @xmath101 s range for building the model . whenever needed , we estimated variance of our predictions by bootstrapping the whole analysis pipeline @xcite .
for this , we created 1000 different datasets by resampling with replacement from the original control dataset and the mutant / ibuprofen datasets .
control statistical models ( the scaling function @xmath31 and the velocity templates ) were estimated for each resampled control dataset .
standard deviations of these models were used as estimates of error bars in fig .
[ fig : the - normalized - active ] . for fig .
[ fig : delta_ipred ] , we additionally needed to form the closest control / treatment worm pairs .
these were formed between the _ resampled _ data sets for all worm types as well .
standard deviations of @xmath79 evaluated by such resampling were then plotted in fig . [
fig : delta_ipred ] and used to estimate @xmath5 scores .
note that such resampling produces data were control / treatment paired worms have slightly larger current differences than in the actual data ; this leads to our error bars being _
overestimates_. model in eq .
( [ eq : bayes ] ) requires knowing @xmath11 . in principle , this is controlled by the experimentalist , and thus should be known . in our experiments , @xmath11 was set to be uniform . however , as described above , some of the worms were discarded in preprocessing , and this resulted in non - uniformly distributed current samples . to account for this , we used the empirical @xmath102 in our analysis instead of @xmath103 . in turn
, @xmath102 was inferred using a well - established algorithm for estimation of one - dimensional continuous probability distributions from data @xcite .
[ [ calculating - the - template - velocities - the - covariances - and - the - scaling - function ] ] calculating the template velocities , the covariances , and the scaling function ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ the template for the paused state @xmath27 is calculated by taking the average of all paused velocity profiles for each of the three worm datasets . the covariance @xmath28 is then the covariance of the set of the paused velocity profiles . for active worms ,
we start with fixed putative parameter values @xmath36 and @xmath43 .
we then calculate the active template @xmath104 and the covariance matrix @xmath105 by maximizing the likelihood in eq .
( [ pva ] ) @xmath106=0 , \\
\frac{\partial \sum_{i}^{n_{{\rm type},a } } \log p({{\bf v}}_i|a , i_i ) } { \partial
\sigma_a}&\propto \sum_{i}^{n_{{\rm type},a}}\left[{{\bf v}}_i-{{\bf u}}_{a}f_{{{\mathcal i}}_1,{{\mathcal i}}_2}(i_i)\right]^2-(\sigma_a)^{-1}=0,\end{aligned}\ ] ] where @xmath107 is the number of active worms of the analyzed type .
this gives : @xmath108 ^ 2.\end{aligned}\ ] ] having thus estimated @xmath32 and @xmath97 at fixed parameter values @xmath36
, @xmath43 , we maximize @xmath109 over the parameters using standard optimization algorithms provided by matlab . we perform optimization from ten different initial conditions to increase the possibility that we find a global , rather than the local maximum .
we thank the granting agencies nsf , hfsp ( human frontier science program ) , and nserc ( natural sciences and engineering council of canada ) for funding , and the aspen center for physics for hosting the program on `` physics of behavior '' , where some of this work was done .
tobin dm , madsen dm , kahn - kirby a , peckol el , moulder g , barstead r , et al .
combinatorial expression of trpv channel proteins defines their sensory functions and subcellular localization in _ c. elegans _ neurons .
|
experiments of pain with human subjects are difficult , subjective , and ethically constrained .
since the molecular mechanisms of pain transduction are reasonably conserved among different species , these problems are partially solved by the use of animal models .
however , animals can not easily communicate to us their own pain levels .
thus progress depends crucially on our ability to quantitatively and objectively infer the perceived level of noxious stimuli from the behavior of animals .
here we develop a quantitative model to infer the perceived level of thermal nociception from the stereotyped nociceptive response of individual nematodes _
caenorhabditis elegans _ stimulated by an ir laser .
the model provides a method for quantification of analgesic effects of chemical stimuli or genetic mutations in _ c. elegans_. we test the nociception of ibuprofen - treated worms and a trpv ( transient receptor potential ) mutant , and we show that the perception of thermal nociception for the ibuprofen treated worms is lower than the wild - type . at the same time
, our model shows that the mutant changes the worm s behavior beyond affecting nociception . finally , we determine the stimulus level that best distinguishes the analgesic effects and the minimum number of worms that allow for a statistically significant identification of these effects .
|
in their role as proton - conducting membranes , ionomers are an important component of many hydrogen fuel cells . in these materials ,
the interplay among the short - range interactions between the hydrophobic backbone polymer and the hydrophilic terminal groups , and the long - range coulomb interactions between the electrostatic charges on the terminal groups and the protons induces a nanophase separation into proton - rich and proton - poor domains .
a general model for the phase morphology of ionomers has been proposed by eisenberg _
et al . _
@xcite , according to which a few head groups combine to form multiplets that restrict the mobility of the backbone chain segments directly attached to them .
a spherical geometry is often assumed for these multiplets , whose sizes are typically less than a nanometer @xcite .
the average distance between these multiplets is mostly dictated by the concentration of head groups relative to that of backbone monomers .
these multiplets then form microdomains , and one observes a microphase separation that can serve to facilitate proton diffusion .
recent experimental data on the morphology of ionomers describes the aggregations as elongated objects embedded in a continuous ionic medium @xcite .
a goal of much ionomer research is to increase the proton conductivity , and hence to make membranes that can operate under very low humidity conditions , as a higher conductivity results in a higher output power density in fuel cells containing these membranes .
proton conduction itself is a complex process , which strongly depends on the thermal and mechanical history of the membrane .
the mechanical history involves the manufacture of the membrane , which is usually achieved by one of two common procedures : solution casting or extrusion .
the former method is used to make membranes from a solution of dissolved ionomer by allowing the solvent to evaporate from the solution . while this technique is suitable only for small - scale laboratory production , it has the advantage producing isotropic membranes with no residual preferential orientation of their backbone within the plane of the membrane .
most commercially available ionomeric membranes are fabricated by extrusion of a molded sample .
this leads to a preferred orientation of the ionomer backbone @xcite , the extent of which depends on the draw speed of the extrusion .
this inherent structural anisotropy of the backbone matrix is believed to be a reason for the susceptibility to tearing or cracking of the membrane in any swelling or drying processes .
this creates technical problems in keeping the membrane taut in the changing temperature - humidity conditions encountered in fuel cells .
an equally important issue concerns the effect of this anisotropy on proton conductivity . while it has generally been assumed that such effects are significant , the problem of predicting the consequences of mechanical strain remains largely unsolved .
there have been several experimental studies in which membranes were uniaxially stretched in order to probe the effects of strain on internal morphology .
et al . _
@xcite analyzed the form of ionic domains in unstretched and stretched membranes , and showed that mechanical stretching induces ordering in the ionomer backbones .
et al . _
@xcite detected the anisotropy in stretched membranes by performing scattering experiments .
et al . _
@xcite used membrane elongation measurements to confirm the classical ionomer domain model , in which inverted micelles are interconnected by channels . as expected , uniaxial stretching of recast nafion causes a preferential orientation of the nafion backbone in the direction of stretching , and this is morphologically similar to the anisotropy in extruded membranes .
et al . _
@xcite , for example , stretched nafion and found the proton conductivity to be higher in the plane of the membrane than normal to it .
a slightly different result was found by lin _
et al . _
@xcite , who noted little change in transverse conductivity on stretching , but found an improved fuel cell performance , as compared to nafion 117 and unstretched recast nafion , as a consequence of a lowered methanol permeation rate . in a related experiment ,
elabd _ et al . _
@xcite measured the conductivities of an ionic block copolymer , and observed appreciable anosotropy .
et al . _
@xcite found that the conductivities of membranes could be rendered anisotropic through the alignment of suspended particles , with higher conductivity in the alignment direction .
these studies reveal the significant impact that organized and oriented structures can have in increasing proton conductivity . while there have been several intensive simulational studies of the effects of naturally occurring anistropies on the transport properties of nafion - like membranes @xcite , there are no studies of which we are aware of the effects of mechanical stretching on an equilibrated sample . in this paper
we investigate numerically to what extent induced orientation in the backbone polymer affects proton diffusion and conductivity in nafion - like ionomers .
the nature of our united - atom model for the ionomer is described in section ii , and the details of our simulation procedure are given in section iii .
section iv reports our results for stretch - induced morphological changes in the ionomer , while results for the proton conductivity are given in section v. in section vi we discuss our conclusion that stretching of solvated membranes leads to anisotropy of the conductivity , with the stretch - induced ordering of the backbone increasing the longitudinal protonic conductivity while decreasing the transverse conductivity .
because it is necessary to model quite large volumes of material in order to see the types of morphological changes in which we are interested , it is unfortunately not feasible to perform fully atomistic simulations .
we thus employ the same united - atom model for nafion @xcite as was used in our previous work @xcite . within this approach
the cf@xmath0 and cf@xmath1 groups of the backbone and sidechains , the ether oxygens of the sidechain , the sulfur atom and the o@xmath2 oxygen group of the sulfonates are modeled as lennard - jones ( lj ) monomers of diameter @xmath3 .
all the ether oxygens and fluorocarbon groups are assigned zero partial charges .
the electrostatic charges are located entirely on the so@xmath4 sulphonate head groups and on the h@xmath5 protons .
the total potential energy of the system is then given as @xmath6 here the first two terms on the right - hand side are pair interactions between nonbonded monomers .
the first term is the lennard - jones interaction @xmath7 where @xmath8 kcal / mol , and @xmath9 is the separation distance between monomers @xmath10 and @xmath11 . at a temperature of 300k , this value of @xmath12 would correspond to an energy of about @xmath13 .
the parameter @xmath14 is 1 for hydrophobic - hydrophobic ( hh ) interactions , and 0.5 for both hydrophobic - hydrophilic ( hp ) and hydrophilic - hydrophilic ( pp ) interactions between monomers . in the pp case ,
the lj potential was modified to be purely repulsive by truncating it at its minimum , where @xmath15 , and raising it by the addition of an amount @xmath12 .
these interaction parameters were chosen to agree in most instances with the nafion model of paddison _ et al . _
@xcite .
the electrostatic interaction between charged sulphonate groups and protons is @xmath16 here @xmath17 ( @xmath18 ) is + 1 for protons , + 1.1 for sulfur atoms , and @xmath19 for the combined oxygens of the sulfonate groups , thus ensuring that the total charge of a sulfonate head group so@xmath4 is @xmath20 .
the distance - dependent dielectric permittivity @xmath21 in eq .
( [ eq:2 ] ) reflects the fact that , at small separations , there is no intervening material between any two charges .
there is then no screening of the coulomb interaction , and the effective permittivity is close to unity . at modest distances
the charges will be separated by matter that may be water , ionic groups , or backbone polymer , and the energy of interaction will be reduced by the shielding effect of this material . the interaction between sulphonates ,
both inside the sulphonate head group clusters surrounded by polymeric matrix , and between two different sulphonate clusters separated by polymeric matrix , is thus a complicated function that includes the effects of image charge distributions at the boundaries @xcite .
when water is present , dielectric saturation and water immobilization effects in the hydrophilic phases of the ionomeric membrane @xcite add complications . while the best approximation for @xmath21 is still a matter for debate @xcite , it is clear that the permittivity must increase with distance .
we choose the form @xmath22 , which smoothly increases from @xmath23 near a chosen charge to being the dielectric permittivity @xmath24 of the bulk at large distances .
varying the value of @xmath24 between 8 and 20 has only a negligible effect on all the observed nanophase separations in the system under consideration .
the effective dielectric permittivity @xmath25 , calculated as a volume average of @xmath21 over a typical multiplet domain , appears to be in good agreement with reported dielectric constants for ionomers in ref .
@xcite .
the last three terms on the right side of eq.([eq:1 ] ) represent the potential energy of the bonded segments of the molecules .
the two - body bond - stretching potential @xmath26 has a simple hookean form with stretching constant @xmath27 kcal / mol @xmath28 ( roughly 1200 @xmath29/@xmath28 at room temperature ) and unstretched bond length @xmath30=1.54 ( 0.44@xmath31 ) .
for the three - body angle - bending potential , @xmath32 we use an equilibrium bending angle @xmath33 and a bending force constant @xmath34 kcal / mol deg@xmath28 ( or about 200 @xmath29/deg@xmath28 ) . finally , the dihedral ( four - body ) component of the total energy is written as @xmath35 with the parameters having values @xmath36 @xmath37 and @xmath38=18.1 kcal / mol @xmath39 10.8 @xmath29 ( 6.2 kcal / mol @xmath39 3.7@xmath29 ) for backbone ( sidechain ) segments , respectively .
we consider two models for the ionomer . in the first model , which is a highly - simplified representation of a completely dry membrane , each proton
is permanently attached to a sulfonate group .
each side - chain head - group is thus electrostatically neutral , and consists of two dipoles , connected head to tail .
we refer to this as the bound - proton model . in the second , and more realistic , model ,
the proton is more loosely bound to its sulfonate counter - ion by a coulomb attraction .
we refer to this as the solvated model . the activation energy needed to overcome the electrostatic attraction between the oppositely charged h@xmath40 and so@xmath41 ions , and
thus detach the proton , depends on the average distance between sulfonates and on the water content @xmath42 in the membrane .
this parameter @xmath42 is defined as the number of water molecules per sulfonate head group , and varies between 0 and 25 for solvated membranes .
it is believed that @xmath43 is sufficient to loosen the bonds tying the protons to the head groups and to achieve full dissociation .
these free protons contribute to the conductivity principally through diffusion and through the grotthus mechanism .
when @xmath44 , hopping diffusion , in which protons make a transition from one sulfonate to an adjacent one , becomes the main contributor to the conductivity .
we use a simple point charge ( spc ) fluid @xcite to model the water molecules in our simulations .
these move in a medium having a distance - dependent dielectric permittivity @xmath21 that we assume to be independent of @xmath42 .
this includes the case of @xmath45 , where in the absence of spc water molecules the proton conductivity occurs only through the hopping mechanism .
the simulations were performed in stages . the first step was to grow a continuous backbone with attached pendant sidechains by using monte - carlo techniques . the bond length , and the bending and dihedral angles for this initial polymer were fixed to be @xmath30 , @xmath46 and @xmath47 respectively .
after molecular - dynamics ( md ) runs of a few picoseconds , which were used to equilibrate the system , as this usually needed to respond to the strong steric repulsion of overlapping monomers , the constraints on angles and bonds were removed . following this ,
the side chains were detached from the backbone skeleton in an approach previously adopted by vishnyakov @xcite , and the backbone skeleton was cut into segments @xcite .
this greatly increases the relaxation rate of the ionomer , and allows a rapid equilibration of the system .
the system then consists of @xmath48 sidechain segments and @xmath49 backbone segments .
taking into account that the sulfonic acid groups are hydrophilic while the ethers and fluorocarbon groups are hydrophobic @xcite , we adopted the following coarse - grained representation .
the side chain architecture is written as 7@xmath50 + 3@xmath51 for a dry membrane and 7@xmath50 + 2@xmath51 for a humidified membrane . here the letters @xmath50 and @xmath51 stand for hydrophobic and hydrophilic monomers correspondingly .
the backbone segments are fully hydrophobic with a 14@xmath50 architecture .
molecular - dynamics runs with a langevin thermostat were performed for the segmented polymer for time periods up to 50ps in a constant-@xmath52 ensemble .
we imposed periodic boundary conditions for a cube of side @xmath53 and used the lekner summation method @xcite to handle the long - range electrostatic interactions between charged particles . in order to verify that the system had not become trapped into a metastable glassy state
, we repeated each run with several different initial configurations . in the next stage of the simulations ,
the segments were reassembled back into the branched chain characterizing the original nafion .
this was achieved by the simultaneous introduction of the bonds and angular constraints described by eqs .
( [ eq:3])([eq:5 ] ) between the ends of each backbone segment , which united them into a single chain .
similar bond and angular constraints between the tail monomer of each detached side chain and the median monomer of every backbone segment connected the side chains to the backbone polymer . to avoid the formation of star - like branched polymers , only a single occupancy of the backbone attachment sites was permitted .
the unequal numbers of backbone segments and side chain segments resulted in the formation of a polymer in which there were varying numbers of backbone monomers between the points of attachment of the side chains .
the simulation was then resumed , and run until a new equilibrium was achieved .
once the system was fully equilibrated , the statistically averaged quantities of interest were gathered during the next 23ns of run time for the initial case of unstretched membranes .
the stretched samples were constructed by the application of a set of forces designed to emulate the effects of a longitudinal stress .
the continuous backbone polymer was considered as a set of @xmath54 connected segments , and a weak force @xmath55 was applied to the end monomer of each segment .
the direction of this force was in the positive @xmath56 direction for the end with the greater value of its @xmath56 coordinate , and in the negative @xmath56 direction for the end with the lesser value of its @xmath56 coordinate in an extended - zone scheme where the periodic boundary conditions were not applied .
this force tends to align the backbone segments in the @xmath56 direction without having significant influence on the side - chain dynamics .
the strength of @xmath55 was restricted to values much smaller than the forces arising from the bond , angle , and dihedral potentials .
it usually took 50ps to equilibrate the stretched membrane . following this ,
runs of 23ns were performed to gather statistically significant results for the morphology of the stretched material . for studies of the proton conductivity , a weak electric field
was applied to the equilibrated stretched membrane , and additional 50ps runs were performed to reach steady - state conditions before non - equilibrium md ( nemd ) simulation runs of 23ns were undertaken . in this way
the effects of both stretching and applied fields could be determined .
a snapshot of an equilibrated membrane with a water content of @xmath57 molecules per head group and in the absence of an applied stretching force is given in fig .
[ fig-1](a ) . the hydrophobic backbone and the side chains
are shown as lines ( red and blue respectively in the online version ) .
the spheres indicate the sulfonate head groups of the sidechains , and show cluster - like aggregations of different sizes and shapes , in agreement with experimental observations @xcite . a moderately stretched membrane with dimensionless stretching force @xmath58=1 is shown in fig . [ fig-1](b ) .
the polymer backbone shows ordering in the direction of the applied force , which is parallel to the @xmath56 axis .
we determine the induced backbone ordering @xmath59 and sidechain ordering @xmath60 by calculating the order parameter @xmath61 an approach commonly used in liquid crystal systems . here
@xmath62 for the backbone polymer and sidechains respectively , the angular brackets @xmath63 denoting statistical averaging over the @xmath52 ensemble . the orientational angle @xmath64 for each segment is defined as the angle between the vector connecting the first and last monomers of the segment and the applied force @xmath65 , as illustrated in fig .
[ fig-2 ] .
the backbone matrix is considered as a set of @xmath54 segments of 14 monomers each .
if @xmath64 is larger than @xmath66 , the order of monomer numbering along the segment is reversed for a proper accounting its orientation . in a similar manner
we define the orientational angle @xmath64 for sidechains as also shown in fig .
[ fig-2 ] . here
@xmath64 is the angle between the stretching force @xmath65 and the vector pointing from the first monomer of the side chain towards its terminal group .
the order parameters @xmath59 and @xmath60 , calculated from eq . ( [ order ] ) , are shown in fig .
[ fig-3 ] .
as expected , membrane stretching leads to ordering in the ionomer conformation .
the backbone ordering , shown in fig .
[ fig-3](a ) , is greatest in the bound - proton model and decreases in the solvated case as the water content @xmath42 increases for a fixed force @xmath65 .
figure [ fig-3](b ) reveals the completely different response of the side chains to the stretching force .
when the bound - proton membrane is stretched , the sidechains tend to orient themselves parallel to the stretching force .
however when a solvated membrane is stretched , the sidechains are more likely to be perpendicular to the stretching force , in accord with previous experimental results @xcite . a possible explanation for these different responses to stress lies in the effect of the humidity on multiplet formation . in the absence of water ,
the dipolar interactions between sulfonates leads to the formation of dense multiplets containing 1020 terminal groups , as seen in fig .
[ fig-5](a ) .
the coulomb forces of attraction between these dipoles limit the amount of molecular relaxation of the sidechains in response to stretching .
the backbone deformation is thus followed by the rearrangement of multiplets via elongation in one direction and shrinking in the other direction .
this is evident from a comparison of the curves for stretched and unstretched membranes in fig .
[ fig-5]a .
in contrast , when the membrane is solvated , the water content lowers the densities of the multiplets .
these then coalesce to form larger clusters in which a mixture of solvent molecules and conducting protons gives the head groups more dynamic flexibility . as a result
, the sidechains can achieve a gain in entropy by rearranging themselves perpendicular to the direction of the applied stress .
this is the reason why only a slight deformation of the multiplet geometry for the solvated membrane is seen in fig .
[ fig-5](b ) . in summary , the order parameter of the membrane depends on its hydration level , the absorbed solvent being an obstacle for the backbone ordering @xmath59 , but facilitating the side - chain ordering @xmath60 . the most important aspect of the effect of stress on morphology concerns the formation of channels through which proton transport might occur .
an indication of this is shown in fig . [ fig-6 ] , in which the water distribution in a stretched ionomer is displayed .
the water molecules congregate in elongated clusters in association with the distribution of side - chain head - groups , and are oriented in the direction of the applied stress . during the course of the simulation , the clusters create temporary bridging connections allowing the water to diffuse easily back and forth along the stretching direction .
this morphology and dynamics of the water undoubtedly facilitates proton diffusion parallel to the stretching direction . in order for proton conduction to occur , it is necessary for there to be a continuous network of pathways through the hydrophilic material in a membrane .
this makes it useful to have a quantitative measure of the connectivity of the hydrophilic channels . to this end
we have developed a method @xcite aimed at detecting hydrophilic pathways in a network of hydrophilic clusters .
the only input parameter for this method is the maximum separation distance parameter @xmath67 between neighboring particles in the pathway .
our method calculates the distribution function @xmath68 of the hydrophilic channel length @xmath69 , which is defined as the distance between the initial and end points of the pathway .
we applied this method to determine how the stretched - membrane morphology alters the channeling in the hydrophilic network of clusters .
the distribution function @xmath68 when @xmath67 is chosen to be @xmath70 ( which is 7 , the average separation between sulfonates in nafion ) is shown in fig .
[ fig-8 ] , both for stretched and unstretched membranes .
both curves have a similar short - range behavior and show a maximum at a most probable channel length equal to about 3@xmath31 .
this is roughly the average size of the sulfonate multiplets in solvated membranes , as was seen in fig .
[ fig-5](b ) .
the long - range behavior of @xmath68 , however , shows a dependence on the mechanical treatment of the membrane . in a stretched membrane
the hydrophilic - sulfonic channels in the membrane are 50% longer than in an unstretched membrane .
we note that the calculated @xmath71 is an instantaneous measure of the pathways in a network of clusters .
because of the constantly changing patterns in the distribution , which are caused by the diffusion of ions and water and the motion of head groups , the real pathways are effectively longer , and can even traverse the whole system .
consequently one expects that the protons in a stretched membrane can continuously travel along the stretching axis @xmath56 .
we explore the resulting protonic conductivities in the next section .
in this section we analyze the effect of membrane stretching on the protonic current .
the electrical conductivity @xmath72 due to proton current in a solvated membrane can be obtained by performing nemd simulations to find the current @xmath73 in the presence of an external electrostatic field @xmath74 through the relation @xmath75 here the induced current density @xmath73 is given by @xmath76 where @xmath77 is the e - field component of the velocity of the @xmath78 ion , and @xmath79 is the total number of protons in the system . because a strong applied electric field may itself have an effect on the morphology , we isolate the effects of stretching by examining only the limit @xmath80 of @xmath81 as @xmath82 .
the longitudinal and transverse conductivities were evaluated by applying fields parallel and perpendicular to the stretching direction , and extrapolating the resultant conductivities to find their values at zero applied field .
the calculated values of @xmath80 for the case where @xmath83 is along the membrane stretching axis are shown in fig .
[ fig-9 ] as a function of the membrane order parameter @xmath59 for different solvation parameters @xmath42 .
it is seen that ordering in the membrane backbone increases the conductivity , with a dependence of @xmath84 on @xmath59 that is close to being linear for all water contents @xmath42 .
conversely , when we examine the transverse protonic conductivity @xmath80 across the stretched membrane , which is shown in fig .
[ fig-10 ] , we find that the conductivity decreases with increasing backbone ordering .
these two trends can be understood as a direct consequence of the channel - like nanophase separation along the stretching direction . while the protons can freely travel along these channels as a consequence of the water cluster bridgings , there are no continuous pathways for them to travel across the stretching direction .
this anisotropy in the conductivity is most pronounced in the dryer samples .
the simulation results presented in the previous sections show that the uniaxial stretching of a nafion - like ionomer should be expected to have a significant effect on its proton conductivity .
the conductivity along the stretching direction is considerably higher than its value in an unstretched membrane , while the proton diffusion perpendicular to the stretching axis is strongly reduced .
when a water - containing sample is stretched , the perfluorinated side chains exhibit a tendency to be oriented perpendicular to the stretching force direction .
the sulphonate clusters deform as the stretching is increased , and show a long - range ordering in their distribution across the membrane .
this affects the distribution of absorbed water , which forms a continuous network with narrow bridges .
the effect of the water is different for the backbone and sidechain ordering parameters . as seen from fig .
[ fig-3 ] , water is an obstacle to backbone ordering @xmath59 parallel to the direction of elongation .
however , water enhances the sidechain ordering @xmath60 perpendicular to the stretching direction . for a given stretching force @xmath65 , the order parameter @xmath59 of the membrane backbone depends on the solvation parameter @xmath42 .
the simulations reported here were carried out at room temperature , and under conditions of constant stress , and we made no attempt to investigate how the anisotropy in the protonic conductivity depends on the temperature at which the membrane was stretched .
experimental measurements of crystallinity , however , indicate that membrane stretching at elevated temperatures can lead to different results from ambient - temperature stretching @xcite .
it appears that uniaxial extension of membranes at temperatures above the boiling temperature of water results in a material with an oriented morphology that persists after the samples are removed from the extensional stress . also , higher - temperature compression - molded ionomers exhibit a backbone crystallization when stretched @xcite
. it would thus be of interest to pursue further studies to enlarge on those presented here .
finally , we note that with regard to the issue of improving fuel - cell performance by using stretched membranes , our results indicate that the observed increased proton conductivity occurs only in the direction of stretching . as a practical matter , the desired direction of proton transport is across the membrane , and so the production of a membrane with enhanced conductivity due to stress may be difficult to arrange .
we acknowledge stimulating discussions with p. pintauro and r. wycisk , who introduced us to their membrane stretching experiments .
e. a. thanks m. litt for valuable comments on equilibrium ionomer morphologies .
this work was supported by the us department of energy under grant de - fg02 - 05er46244 , and was made possible by use of facilities at the case its high performance computing cluster and the ohio supercomputing center .
98 eisenberg , a. ; hird , b. ; moore , r. b. _ macromolecules _ * 1990 * , _ 23 _ 4098 .
gohy , j. f. ; jerome , r. ; van den bossche , g. ; sobry , r. _ macromol . chem .
phys . _ * 1998 * , _ 199 _ 2205 .
rubatat , l. ; rollet , a. ; gebel , g. ; diat , o. _ macromolecules _ * 2002 * , _ 35 _ 4050 .
barbi , v. ; funari , s. s. ; gehrke , r. ; scharnagl , n. ; stribeck , n. _ polymer _ * 2003 * , _ 44 _ 4853 .
gebel , g. ; diat , o. _ fuel cells _ * 2005 * , _ 5 _ 261 .
elliott , j. a. ; hanna , s. ; elliott , a. m. s. ; cooley , g. e. _ macromolecules _ * 2000 * , _ 33 _ 4161 .
elliott , j. a. ; hanna , s. ; elliott , a. m. s. ; cooley , g. e. _ polymer engineering and science _ * 2006 * , _ 46 _ 228 .
cable , k. m. ; mauritz , k. a. ; moore , r. b. _ chem .
mater . _ * 1995 * , _ 7 _ 1601 .
lin , j. ; wycisk , r. ; pintauro , p. n. ; kellner , m. _ electrochemical and solid - state letters _
* 2007*,_10 _ b19 .
elabd , y.a . ; walker , c. w. ; beyer , f. l. _ journal of membrane science _ * 2004 * , _ 231 _ 181 .
oren , y. ; freger , v. ; linder , c. _ journal of membrane science _ * 2004 * , _ 239 _ 17 .
cui , s. ; liu , j. ; selvan , m. e. ; keffer , d. j. ; edwards , b. j. ; steele , w. v. _ j. phys .
b _ * 2007 * , _ 111 _ 2208 .
blake , n. p. ; petersen , m. k. ; voth , g. a. ; metiu , h. _ j. phys .
b _ * 2005 * , _ 109 _ 24244 . jang , s. s. ; molinero , v. ; cagin , t. ; goddard iii , w. a. _ j. phys
b _ * 2004 * , _ 108 _ 3149 .
wescott , j. t. ; qi , y. ; subramanian , l. ; capehart , t. w. _ j. chem .
_ * 2006 * , _ 124 _ 134702 .
vishnyakov , a. ; neimark , a. v. _ j. phys . chem .
b _ * 2001 * , _ 105 _ 7830 ; * 2001 * , _ 105 _ 9586 .
yamamoto , s. ; jinnouchi , r. ; yamakawa , sh . ; hyodo , sh . _ 14th international conference on the properties of water and steam , kyoto , japan _ * 2004 * , 411 .
allahyarov , e. ; taylor , p. l. _ j. chem .
phys . _ * 2007 * , _ 127 _ 154901 .
paddison , s. j. ; zawodzinski , t. a. _ solid state ionics _ * 1998 * , _ 113 - 115 _ 333 .
iversen , g. ; kharkats , yu . i. ; ulstrup , j. _ molecular physics _ * 1998 * , _ 94 _ 297 .
kreuer , k. d. ; paddison , s. j. ; spohr , e. ; sschuster , m. _ chem .
rev . _ * 2004 * , _ 104 _ 4637 .
paul , r. ; paddison , s. _ mat .
proc . _ * 2001 * , _ 677 _ aa7.16.1 .
taylor , p. l. ; xu , b. c. ; oliveira , f. a. ; doerr , t. p. _ macromolecules _ * 1992 * , _ 25 _ , 1694 .
l . ; yu , t. l. ; huang , c .- h . ; lin , t .- l . _ j. polymer science b : polymer physics _ * 2005 * , _ 43 _ 3044 .
chan , k. ; tang , y. w. ; szalai , i. _ molecular simulations_*2004 * , _ 30 _ 81 .
vishnyakov , a. ; neimark , a. v. ; _ j. phys .
b _ * 2000 * , _ 104 _ 4471 ; rivin , d. ; meermeier , g. ; schneider , n. s. ; vishnyakov , a. ; neimark , a. v. _ j. phys
b _ * 2004 * , _ 108 _ 8900 .
glotzer , s. c. ; paul , w. _ annu .
res . _ * 2002 * , _ 32 _ 401 .
mazars , m. _ j. chem
. phys . _ * 2001 * , _ 115 _ 2955 .
mauritz , k. a. ; moore , r. b. _ chem .
rev . _ * 2004 * , _ 104 _ 4535 .
chourdakis , n. ; voyiatzis , g. a. _ j polym .
part b : polym .
phys . _ * 2007 * , _ 45 _ 2509 .
allahyarov , e. ; taylor , p. l. _ submitted to phys .
review e _ * 2008*. trevino , s. f. ; young , s. k. _ army research labarotory report _
* 2002*. bagrodia , s. ; mohajer , y. ; wilkes , g. l. _ polymer bulletin _ * 1982 * , _ 8 _ 281 .
bauer , f. ; denneler , s. ; willert - porada , m. _ j. pol .
part b : pol .
phys . _ * 2005 * , _ 43 _ 786 .
= 0 ) and ( b ) stretched ( @xmath58=1 ) nafion membrane with water content @xmath57 .
the stretching force @xmath58 in ( b ) is directed along the @xmath56 axis .
colored beads represent the end - group oxygens of side chains .
pendant side chains and neutral backbone polymer are drawn by lines ( blue and red , respectively , in online version ) .
different bead colors correspond to different bead altitudes , with blue for beads at the bottom of the simulation box and red for beads at the top of the box .
the size of all structural elements is schematic rather than space filling .
[ fig-1],title="fig:",scaledwidth=54.0% ] = 0 ) and ( b ) stretched ( @xmath58=1 ) nafion membrane with water content @xmath57 .
the stretching force @xmath58 in ( b ) is directed along the @xmath56 axis .
colored beads represent the end - group oxygens of side chains .
pendant side chains and neutral backbone polymer are drawn by lines ( blue and red , respectively , in online version ) .
different bead colors correspond to different bead altitudes , with blue for beads at the bottom of the simulation box and red for beads at the top of the box .
the size of all structural elements is schematic rather than space filling .
[ fig-1],title="fig:",scaledwidth=54.0% ] for backbones ( left picture ) , and @xmath85 for sidechains ( right picture ) .
the continuous black line represents the backbone polymer of the ionomer . shorter gray lines ( green in online version )
represent the sidechains .
the dashed lines are vectors constructed to show the orientation of backbone segments and sidechains , with @xmath64 being the angle between the orientation vector and stretching direction .
[ fig-2 ] ] ( a ) and sidechain order parameter @xmath60 ( b ) as functions of applied stretching force @xmath58 for the bound - proton model and for different water contents @xmath42 .
[ fig-3 ] , title="fig:",scaledwidth=70.0% ] ( a ) and sidechain order parameter @xmath60 ( b ) as functions of applied stretching force @xmath58 for the bound - proton model and for different water contents @xmath42 .
[ fig-3 ] , title="fig:",scaledwidth=70.0% ] of a stretched membrane along the stretching direction as a function of the backbone order parameter @xmath59 for different membrane solvations .
the water - content parameters @xmath42 are indicated below the corresponding lines . [ fig-9],scaledwidth=100.0% ] of a stretched membrane perpendicular to the stretching direction as a function of the backbone order parameter @xmath59 for different membrane solvations .
the water - content parameters @xmath42 are indicated below the corresponding lines .
[ fig-10],scaledwidth=100.0% ]
|
we have used coarse - grained simulation methods to investigate the effect of stretching - induced structure orientation on the proton conductivity of nafion - like polyelectrolyte membranes .
recent experimental data on the morphology of ionomers describe nafion as an aggregation of polymeric backbone chains forming elongated objects embedded in a continuous ionic medium .
uniaxial stretching of a recast nafion film causes a preferential orientation of these objects in the direction of stretching .
our simulations of humid nafion show that this has a strong effect on the proton conductivity , which is enhanced along the stretching direction , while the conductivity perpendicular to the stretched polymer backbone is reduced .
stretching also causes the perfluorinated side chains to orient perpendicular to the stretching axis .
this in turn affects the distribution of water at low water contents
. the water forms a continuous network with narrow bridges between small water clusters absorbed in head - group multiplets .
ionomers , proton diffusion , morphology
|
the models in extra dimensions have been proposed for solving the hierarchy between the electroweak and planck scales @xcite . among these
the randall - sundrum ( rs ) model assumes the existence of two 3-branes connected by one extra spatial dimension and the metric in this model is non - factorizable @xcite .
the extra spatial coordinate has been orbifolded by the symmetry @xmath3 and one of the 3-branes can be identified as the visible brane and the other as the planck brane .
the warp factor in this model suppresses any planck scale quantities into electroweak scale on the visible brane . in the rs model , all the standard model fields
are assumed to be confined on the visible brane and only gravity propagates in the bulk of the five dimensions .
subsequently , the rs model was generalized to include other bulk fields in order to explain physical quantities such as neutrino masses and mixing pattern which normally can not be explained within the standard model of elementary particles .
the deficit in the solar and atmospheric neutrino flux has given evidence for non - zero masses to neutrinos @xcite .
fitting to the data of solar and atmospheric neutrino experiments , the following set of mass - square differences are obtained : @xmath4 ev@xmath5 and @xmath6 ev@xmath5 @xcite . here , @xmath7 are the mass eigenvalues of the 3 active neutrinos . apart from the mass - square differences an upper limit on the neutrino masses have been found through other experiments .
tritium @xmath8-decay puts an upper limit on the neutrino mass scale to be of the order of 2 ev @xcite .
whereas from the cosmological observations , the sum of the three neutrino masses needs to be less than about 1 ev @xcite .
since all the above experiments suggest tiny values for neutrino masses , perhaps a different mechanism should be operational for neutrino mass generation as compared to other fermion masses .
to explain the smallness of neutrino masses in the framework of rs model , an additional singlet neutrino field has been proposed , which like gravity field , propagates in the entire bulk of space - time @xcite .
the wave function of the bulk neutrino field is extended in the extra spatial dimension .
the boundary conditions for this wave function can be chosen in such a way that it will have a very small overlap on our visible brane , resulting in tiny masses for neutrinos .
the phenomenology of this model is determined through the kaluza - klein ( kk ) modes of the bulk neutrino field .
since the bulk neutrino field is singlet under the standard model gauge group , detection of kk modes of this field is challenging in the collider experiments .
however , the loop effects due to these fields to any physically observable quantity can give us some hints about its existence .
here , we study one of such observable quantities , namely the anomalous magnetic moment of muon . at the tree level ,
the anomalous magnetic moment of muon , @xmath9-factor , has a value of 2 and radiative contributions give some corrections to it .
hence , it can be quantified as @xmath10 . for a review on @xmath0 ,
see @xcite . the world average value of @xmath0
after the experiment e821 at the brookhaven national laboratory is as given below @xcite @xmath11 which is obtained with a precision of 0.54 parts per million .
various groups have computed the theoretical value for @xmath0 in the standard model .
most of the groups have found a discrepancy between the experiment and the corresponding standard model value of @xmath0 at about @xmath12 level @xcite . here
, we take this difference as follows @xcite @xmath13 the above difference would indicate existence of new physics .
moreover , since this difference is positive , the contribution due to new physics to @xmath0 should yield a net positive value . as explained before , in the model of ref .
@xcite , kk neutrinos can give some contribution to @xmath0 , which we have computed in this work .
previously , some work in this direction has been done in ref .
@xcite , where the authors have obtained a negative contribution to the @xmath0 by adopting mass insertion approximation in the unitary gauge . in the present work , by carrying out an exact analysis in the mass eigenstate basis , we have computed the contribution from the bulk neutrino field to the @xmath0 in the t hooft - feynman gauge and scan the parameter space in the region consistent with experimental bounds .
we have found that the bulk neutrino contribution to @xmath0 has the right sign to fit the above mentioned discrepancy in @xmath0 .
we have also incorporated contributions from other sources such as gravitons to the @xmath0 .
finally , we have studied the constraints that may arise from these various sources of @xmath0 in the parameter space defined in @xcite .
recently , in @xcite , constraints have been obtained from lepton flavor violation by considering a set of models in the rs frame work .
the paper is organized as follows . in the next section ,
we give a brief description of the warped extra dimensional model for neutrino masses @xcite . in sec .
3 , we have computed the contribution of bulk neutrino field to @xmath0 in the t hooft - feynman gauge . in this section , we have also performed a detailed phenomenological study of our obtained expression for @xmath0 . in sec . 4 , we describe possible constraints on the model parameters of @xcite by including contributions to @xmath0 from other sources such as gravitons .
we conclude in sec .
as described before , the model @xcite is based on the rs model , where an additional singlet neutrino field is introduced in the bulk of the 5-dimensional space - time .
the metric in this model is @xmath14 where @xmath15 , @xmath16 is the compactification radius of the fifth dimension and @xmath17 is an energy scale of the order of planck scale , @xmath18 . here , @xmath19 is an angular coordinate representing the fifth dimension . due to the @xmath3 orbifold symmetry
, @xmath19 varies from 0 to @xmath20 and the 3-branes located at these points are called planck and visible branes , respectively . in this model ,
the invariant action for a singlet bulk fermion @xmath21 is as given below .
@xmath22 -m_b{\rm sgn}(\phi)\bar{\psi}\psi\right\},\ ] ] where @xmath23 is the determinant of 5-dimensional metric @xmath24 , @xmath25 is the inverse vierbein and @xmath26 is the spin connection . the small case letters , @xmath27 , run over flat 5-dimensions and upper case roman letters run over curved 5-dimensional space - time . here
, @xmath28 is a bulk mass parameter which is @xmath29 .
the above action is invariant under @xmath30 , which should follow due to the @xmath31 orbifolding of the rs model .
this symmetry is known as @xmath19-parity which sets some boundary conditions on the wave functions of @xmath21 .
the bulk field @xmath21 decomposes into kk modes in the 4-dimensional world , which can be written as @xmath32 where @xmath33 , @xmath34 are the left- and right - handed kk modes in the 4-dimensions and @xmath35 are its corresponding wave functions . in order to have the following canonically normalized action for the kk modes in 4-dimensions @xmath36 the wave functions @xmath37 should satisfy the following conditions @xmath38 here , @xmath39 are the masses of kk modes . in the above equations , the following change of variables have been done : @xmath40 , @xmath41 and @xmath42 .
this change of variables imply that @xmath43 corresponds to the visible brane .
the unknown quantities in the above equation are : @xmath44 the @xmath19-parity , which is described above , imposes the following boundary conditions on the wave functions : @xmath45 .
the coupled differential equations of eq .
( [ e : coneqs ] ) can be exactly solved and for @xmath46 , the wave functions can be expressed in the form of bessel functions .
the wave function for the zeroth mode can be written as @xmath47 the zeroth mode wave function of the right - handed field is highly suppressed on the visible brane , for @xmath48 .
because of this reason , by choosing the boundary condition as @xmath49 , we can achieve a very small overlap of the singlet bulk field on the visible brane , for @xmath48 . using this mechanism , in the next paragraph
, we describe the neutrino masses in this model .
the invariant action for neutrino yukawa interaction is @xmath50 where @xmath51 is a left - handed lepton doublet and @xmath52 is the conjugate of the higgs doublet . here
@xmath53 has mass dimensions of @xmath54 and the metric on the visible brane is @xmath55 and @xmath56 . both the lepton and higgs doublets need to have the following rescaling : @xmath57 , @xmath58 , in order to have canonical kinetic terms for these fields . after substituting the kk mode expansion , eq .
( [ e : kkmode ] ) , the invariant action becomes @xmath59 here , @xmath60 is an order one parameter . in the basis ,
@xmath61 and @xmath62 , we have @xmath63 here , the vacuum expectation value of the higgs field is @xmath64 174 gev .
now , define @xmath65 then the physical masses are given by @xmath66 while diagonalizing the matrix @xmath67 , the masses @xmath39 of kk modes are determined by the equation @xmath68 .
however , the masses @xmath39 are not the physical masses of kk modes , since they get correction due to mixing with the left - handed lepton doublet after the electroweak symmetry breaking .
after diagonalizing the matrix @xmath67 , the lowest mass eigenvalue can be shown to be proportional to @xmath69 which yields a very small value due to small overlap of the wave function @xmath70 .
the mass of light neutrino can be shown to be @xcite @xmath71 so far we have assumed the presence of only one bulk neutrino field and it generates one light dirac neutrino mass . as explained in the previous section , to fit the neutrino oscillation data , we need at least two non - zero neutrino mass eigenstates . hence , in the model of ref .
@xcite we have to propose more than one bulk neutrino field .
however , it has been argued in @xcite that to cancel the anomalies related to @xmath19-parity , only even number of singlet bulk fields can be introduced into the model .
so , by introducing two bulk neutrino fields with slightly different bulk mass parameters , @xmath28 , we can fit both the solar and atmospheric neutrino mass scales . in this picture
the third neutrino has exactly zero mass .
however , by introducing 4 bulk neutrino fields in an analogous way , we can generate all the three light neutrino masses .
we have shown in the previous section that to generate neutrino masses , the kk modes of bulk neutrino field have some mixing with the left - handed lepton doublet .
this mixing would lead to gauge as well as yukawa type interactions of muon with the physical kk modes of neutrinos . as a result of this ,
the kk modes of neutrinos give some contribution to @xmath0 , which we will describe shortly .
it is to be noticed here that we have computed the contribution of bulk neutrino to @xmath0 in the t hooft - feynman gauge , hence the interactions of muon with nambu - goldstone states are necessary . for the sake of completeness , below we describe interaction terms involving the physical kk modes of bulk neutrino field .
the interaction terms of other fields are unchanged from that of the standard model .
+ : + in this model , the interaction term of @xmath72-boson with charged lepton and neutrino would come from the 5-dimensional invariant action of the kinetic energy of the lepton doublet . for our particular case of muon field , these interactions are as given below . @xmath73.\ ] ] after using eq .
( [ e : kk ] ) we get @xmath74.\ ] ] : + without loss of generality , we can go to a basis where charged lepton yukawa couplings are diagonalized . in this basis ,
the necessary lagrangian for our purpose is @xmath75 here @xmath76 is the left - handed lepton doublet containing muon . the second term in the above equation is from eq .
( [ e : sy ] )
. we can use @xmath77 where @xmath78 is the mass of muon . after using eq .
( [ e : kk ] ) , we can show that @xmath79 the interaction terms , as described above , generate some contribution to @xmath0 at one loop level , which are shown in fig . 1 . .
the momentum convention is that the outgoing muon has @xmath80 , the incoming muon has @xmath81 and the momentum on the photon line is @xmath82 , which is incoming . ]
the amplitude of any of the fig .
1 is of the form @xmath83 , where @xmath84 is the polarization of photon and @xmath85 is the positron charge . after using the gordon identity , @xmath86u(p)$ ] , where @xmath82
, we can put the interesting part of the amplitude as @xmath87 the contribution to @xmath0 is @xmath88 . the contribution from fig .
1(a ) to @xmath0 is @xmath89.\end{aligned}\ ] ] the sum of the contributions from figs .
1(b ) and 1(c ) to @xmath0 is as follows @xmath90.\end{aligned}\ ] ] the contribution to @xmath0 from fig .
1(d ) is @xmath91\frac{1}{(m^{\rm kk}_j)^2}f_c(x_j ) \right .
\nonumber \\ & & \left .
+ \frac{1}{vm^{\rm kk}_j}\left[u^l_{1j}y^*_k\left(v^r_{kj}\right)^*+\left(u^l_{1j}\right)^*y_kv^r_{kj}\right ] f_d(x_j)\right\ } , \quad x_j=\frac{m_w^2}{(m^{\rm kk}_j)^2 } , \nonumber \\
f_c(x ) & = & \frac{1}{(1-x)^4}\left[\frac{1}{3}+\frac{1}{2}x - x^2+\frac{1}{6}x^3+x\ln(x)\right ] , \nonumber \\
f_d(x ) & = & \frac{1}{(1-x)^3}\left[\frac{1}{2}-\frac{1}{2}x^2+x\ln(x)\right].\end{aligned}\ ] ] it is to be noticed that in the contribution to @xmath0 from all the plots of fig .
1 , the summation in indices over @xmath92 is from 2 , since we have to subtract the light neutrino contribution which exists in the standard model . here
we comment on our results on the contribution of bulk neutrino to @xmath0 . as already explained before that we have computed the contribution from fig . 1 in the t hooft - feynman gauge .
the first three diagrams of fig . 1 have given positive contributions , while the fig .
1(d ) has given both positive as well as negative contributions . in our numerical analysis , which we present below
, we have found that in most of the parameter space the contribution from fig .
1(d ) is dominant and it gives a net positive contribution .
specifically , we have seen that the negative and positive contributions are comparable to each other but the magnitude of the former one is at least an @xmath93 less than the later one in @xmath94 .
another comment is that , since there are two insertions of yukawa couplings in fig .
1(d ) we would expect the contribution to @xmath0 to increase with @xmath60 . in our numerical analysis
we have found that this is true but there are some exceptions to this , which we will explain in our numerical results .
the total contribution of one bulk neutrino field to the @xmath0 is @xmath95 the above contribution is mainly dependent on @xmath96 , @xmath97 and the 5-dimensional bulk mass parameter @xmath28 , which altogether determine the kk masses of the bulk neutrino field .
apart from this , @xmath98 also depends on the dimensionless parameter @xmath60 which determine the elements of the unitary matrices @xmath99 and @xmath100 . on the other hand , the light neutrino mass eigenvalue due to one bulk neutrino field , depends on @xmath96 , @xmath97 and @xmath60 . as described in sec .
1 , from the neutrino oscillation data we have some idea on the magnitude of neutrino masses . for instance , in the hierarchical pattern of neutrinos at least two mass eigenvalues should be : @xmath101 ev and @xmath102 ev . in this case , the third neutrino can have either zero mass or its mass should be less than of the order of @xmath103 . in the case of degenerate neutrinos ,
all the three neutrinos should have a nearly equal mass and the common mass eigenvalue should be sufficiently larger than @xmath104 . since an upper bound from the cosmological observations indicate that the sum of the three neutrino masses should be less than about 1 ev @xcite , we take the common mass to be as @xmath105 ev . by fixing the mass eigenvalues of light neutrinos , we may eliminate @xmath96 as independent variable .
the parameter @xmath60 should be @xmath93 and here we take its value in the range 0.1 to 2.5 . here
the lower limit of 0.1 is due to the naturalness argument and the upper limit is due to perturbativity constraints . in order not to hit the landau pole , the yukawa couplings @xmath106 should be less than @xmath107 . for @xmath46 , @xmath108 and
hence @xmath60 should be less than @xmath109 2.5 .
as for the bulk mass parameter @xmath28 , we take its value close to the planck scale . we take the values of @xmath97 to be around 12 , since for this set of values we can produce tev scale masses on the visible brane from the exponential warping . finally , in our numerical analysis , we have allowed a 50@xmath11050 mixing mass matrix of eq .
( [ e : mixmas ] ) .
after diagonalizing this matrix we get 49 kk neutrino mass modes which we have summed in the @xmath98 .
we have checked that the above mentioned number of kk modes in @xmath98 is sufficient and its value vary insignificantly by further increasing the number of kk modes . in fig .
2 we have plotted the contribution due to one bulk neutrino field to @xmath0 . has been plotted as a variable of @xmath60 . in these plots
the three curves which are closely stacked together represents whether the light neutrino mass fits the solar , or atmospheric , or degenerate mass .
the values of @xmath97 and @xmath28 have been varied in both these plots.,title="fig:",width=240,height=240 ] has been plotted as a variable of @xmath60 . in these plots
the three curves which are closely stacked together represents whether the light neutrino mass fits the solar , or atmospheric , or degenerate mass .
the values of @xmath97 and @xmath28 have been varied in both these plots.,title="fig:",width=240,height=240 ] in fig .
2(a ) we have fixed @xmath97 = 12.5 . in this plot ,
the upper three curves are due to a bulk mass of @xmath111 gev and the remaining three lower curves are for @xmath112 gev .
the three curves are arising depending on whether the light neutrino mass eigenvalue fits the solar , or atmospheric , or degenerate mass eigenvalue .
the value of @xmath96 is not shown in these plots , since as explained before it is not an independent parameter .
the meaning of curves in fig .
2(b ) are same as that of fig .
2(a ) , except that @xmath97 has taken a value 12.0 in fig .
for one particular value of @xmath60 , the contribution due to a bulk neutrino field which fits the solar , or atmospheric , or degenerate neutrino is nearly the same .
this is evident in both the figs .
2(a ) and 2(b ) , where the three curves due to different neutrino mass eigenvalues are close to each other .
we can understand the reason for this as follows . for fixed values of @xmath60 , @xmath97 and @xmath28
, the change in the value of @xmath113 from one curve to the other curve can happen only due to the change in the neutrino mass eigenvalue .
the ratio of neutrino mass eigenvalues between degenerate and solar cases could be at most by a factor of @xmath114 . from eq .
( [ e : mnu ] ) , it can be seen that the neutrino mass eigenvalue is related to @xmath96 dominantly through exponential factor , so we need only a change of @xmath1150.1 in @xmath96 to produce a change of factor @xmath114 in the mass eigenvalue .
these exact numerical values in the change of @xmath96 due to change in the neutrino mass eigenvalue can be seen in tab . 1 , where ranges of @xmath96 are given for extreme values of @xmath60 .
such a small change in @xmath96 produces a slight change in the kk masses of neutrino fields , and hence a slight change to the @xmath98 .
however , a closer examination would reveal that the amount of @xmath113 where solar mass is fitted is slightly more than that for atmospheric mass which is even slightly more than that for degenerate mass .
the change in these three cases could be at most in the first decimal place of @xmath113 value .
the values of @xmath96 in these three cases would be different even though @xmath60 could be same .
we have presented the ranges of @xmath96 for the cases of @xmath97 = 12.5 and 12.0 in tab .
1 . .ranges of @xmath96 for different values of @xmath97 and for different masses of light neutrino mass eigenvalues . in a particular range of @xmath96 ,
the left- and right - end corresponds to @xmath116 and @xmath117 , respectively . [ cols="^,^,^",options="header " , ] in the case of @xmath118 , two kk gravitons can exist to fit the @xmath0 .
the above analysis done for @xmath119 can be repeated for @xmath118 and we can get bounds on the parameters @xmath60 and @xmath96 accordingly . however , if we include the contribution of the order of 10 kk gravitons then @xmath97 should be close to 11.7 . in this case , by comparing the curves for @xmath120 of fig .
3 , for @xmath28 close to @xmath121 gev , a single bulk neutrino can contribute an order of @xmath122 to @xmath0 .
so for this choice of @xmath28 we can get upper limits on @xmath60 and @xmath96 . however , if the number of kk gravitons is of the order of 100 then @xmath97 should be around 11.3 .
for this low value of @xmath97 , a single bulk neutrino contribution would be at most @xmath123 even for @xmath124 gev .
hence the constraints on the bulk neutrino parameters would be less stringent in this case .
finally , to comment on the radion contribution to @xmath0 , the constraints on the bulk neutrinos which are described above may become even severe if the radion mass is less than about 200 gev . in the previous paragraph
we have described constraints on the bulk neutrino parameters which are arising purely from the @xmath125 deviation of @xmath126 . however , these constraints can be even stringent by including other observable quantities such as @xmath127 .
the current experiments have not found the decay @xmath128 and put an upper limit on its branching ratio as @xmath129 at 90@xmath130 cl @xcite . in the context of the warped model @xcite , this decay channel has been studied in @xcite . to satisfy the upper bound on @xmath127 and to get an appreciable contribution from @xmath113
, we may have to fine - tune the yukawa couplings .
it is interesting to study these effects but this is out of the scope of the current work .
finally , we comment that in this work we have studied correlation between neutrino mass eigenstate and @xmath0 for a single generation .
we hope to extend this to include all the generations in a future work .
some related work in the context of different models has been studied in @xcite .
the rs model , which is based on the warp geometry , is elegant to explain the hierarchy between the planck and electroweak scales @xcite .
subsequently it was shown that such model can offer possible explanation of the origin of small neutrino masses when the right handed component of the neutrino is allowed to propagate in the bulk @xcite .
the signals of such a model lies in the detection of kk modes of the bulk neutrino field . in this work ,
we have studied an indirect signal of these kk modes , _
i.e. _ the virtual effects of these kk modes to the anomalous magnetic moment of muon , @xmath0 .
we have computed the contribution of bulk neutrino to @xmath0 at one loop level in the t hooft - feynman gauge .
comparing our result with the experimental limits , we have found that this contribution comes with the right sign so as to fit with experimental data .
moreover , in some of the parameter space of the model in @xcite , the @xmath125 discrepancy of @xmath0 can be accommodated .
we have also studied how the bulk neutrino contribution would be restricted by including the graviton contribution to @xmath0 .
sr acknowledges the hospitality provided by prof .
alejandro ibarra and the cluster of excellence for fundamental physics origin and structure of the universe " at the technical university in munich ( tum ) during the final stages of this work .
y. fukuda _ et .
( kamiokande collaboration ) , phys .
lett . * 77 * , 1683 ( 1996 ) ; w. hampel _ et .
( gallex collaboration ) , phys .
b * 447 * , 127 ( 1999 ) ; j.n .
abdurashitov _ et . al . _
( sage collaboration ) , phys .
c * 60 * , 055801 ( 1999 ) ; q.r .
( sno collaboration ) , phys .
lett . * 87 * , 071301 ( 2001 ) ; k. eguchi _ et .
( kamland collaboration ) , phys .
lett . * 90 * , 021802 ( 2003 ) ; j. hosaka _ et .
( super - kamkiokande collaboration ) , phys .
d * 73 * , 112001 ( 2006 ) ; y. fukuda _ et .
( super - kamkiokande collaboration ) , phys .
81 * , 1562 ( 1998 ) ; m. ambrosio _ et .
( macro collaboration ) , phys .
b * 434 * , 451 ( 1998 ) .
s. hannestad , nucl .
* 145 * , 313 ( 2005 ) ; j. dunkley _ et .
_ , ( wmap collaboration ) , astrophys . j. suppl . * 180 * , 306 ( 2009 ) ; e. komatsu _ et .
al . _ , ( wmap collaboration ) , astrophys . j. suppl . * 180 * , 330 ( 2009 ) .
t. teubner , k. hagiwara , r. liao , a.d .
martin and d. nomura , arxiv:1001.5401 [ hep - ph ] ; m. davier , a. hoecker , b. malaescu and z. zhang , eur .
j. c * 71 * , 1515 ( 2011 ) [ erratum - ibid .
c * 72 * , 1874 ( 2012 ) ] ; see also , references cited by @xcite . m.l .
brooks _ et .
( mega collaboration ) , phys .
lett . * 83 * , 1521 ( 1999 ) ; m. ahmed _ et .
( mega collaboration ) , phys .
d * 65 * , 112002 ( 2002 ) ; j. adam _
[ meg collaboration ] , [ arxiv:1107.5547 [ hep - ex ] ] .
|
in the randall - sundrum model , a bulk neutrino field in the 5-dimensional space - time can give rise to tiny dirac masses to neutrinos .
in such a scenario , we have computed the contribution of the bulk neutrino field to the anomalous magnetic moment @xmath0 of muon .
we have computed this contribution in the t hooft - feynman gauge and have found that the contribution has the right sign to fit the current discrepancy between the experiment and the standard model value of @xmath0 .
we have also studied possible constraints on the model parameters by including contributions to @xmath0 from other sources such as bulk gravitons . * muon @xmath1 from the bulk neutrino field in a warped extra dimensional model * + r. s. hundi , sourov roy and soumitra sengupta + department of theoretical physics , + indian association for the cultivation of science , + 2a @xmath2 2b raja s.c .
mullick road , + kolkata - 700 032 , india .
|
strongly - correlated one - dimensional ( 1d ) and quasi-1d systems can be simulated accurately using the density matrix renormalization group ( dmrg ) @xcite , which is a set of algorithms operating on matrix product states ( mps ) @xcite .
real systems can usually not be isolated completely from their environment and experiments are hence necessarily done at finite ( nonzero ) temperatures . to allow for a direct investigation of experimental results or for theoretical investigations on the influence of temperature ,
dmrg techniques have been generalized to describe thermal states .
the historically first dmrg algorithm for finite temperatures is the quantum transfer - matrix renormalization group @xcite , which is however complicated by the non - hermiticity of the quantum transfer matrix and does not allow for a direct evaluation of non - local observables .
a recent approach combines time - dependent dmrg @xcite with monte carlo to evaluate thermal observables by sampling over so - called minimally entangled typical thermal states @xcite and is also applicable for the study response functions @xcite .
the arguably most successful method , which is at the focus of this paper , is based on a purification of the density matrix @xcite , i.e. , the representation through a pure state in an enlarged hilbert space . is called a purification of the density matrix @xmath2 on @xmath3 if @xmath4 . ] for numerical purposes , these are approximated very precisely in matrix product form , to obtain so - called matrix product purifications ( mpp ) @xcite .
first applications to static phenomena @xcite showed the potential of the approach which is also known as the ancilla method .
it has been extended for the computation of response functions @xcite and allows it also to work directly in the frequency domain @xcite .
the various applications comprise , for example , the phase diagrams of spin - ladder systems @xcite , thermometry for ultracold bosons in optical lattices @xcite , the drude weight in spin-@xmath0 and hubbard chains @xcite , as well as spectral properties of quantum magnets @xcite , bosonic systems @xcite , and spin - incoherent luttinger liquids @xcite .
it is straightforward to compute an mpp of a grand - canonical ensemble .
we consider lattice systems ( 1d or quasi-1d ) , such that the hilbert space has tensor - product structure .
the grand - canonical infinite - temperature state is then a tensor product of on - site identities and the corresponding mpp has bond dimension one .
a subsequent imaginary - time evolution yields the mpp of the grand - canonical ensemble at finite temperatures .
when a system features symmetries , these can be exploited to reduce computation costs substantially . for a conserved quantity @xmath5 , one can use that the imaginary - time evolution does not mix components with different @xmath5 quantum numbers .
however , the grand - canonical ensemble contains of course contributions from all quantum number sectors such that the benefit of symmetries is reduced .
this paper introduces and demonstrates efficient methods to compute mpps of canonical ensembles under utilization of symmetries .
in particular , we give exact matrix product representations for canonical infinite - temperature states ( sect . [
sec : ce_beta0_exact ] ) , and discuss how they can be constructed alternatively by applying matrix product operators to vacuum - type states ( sect .
[ sec : ce_beta0_mpo ] ) .
we also discuss an earlier variational approach using entangler hamiltonians @xcite and present a generalized version of it ( sect .
[ sec : entangler ] ) .
furthermore , we introduce a method for the evaluation of global quantum number probability distributions using matrix product density operators ( mpdo ) @xcite ( sect .
[ sec : mpdo_qn - distr ] ) .
the techniques are exemplified for antiferromagnetic spin-@xmath0 heisenberg chains , explaining why the difference in the energy densities of canonical and grand - canonical ensembles with the same magnetization decays as @xmath1 with the system size @xmath6 ( sect .
[ sec : example ] ) .
for simplicity , let us restrict our considerations to an abelian symmetry with a single conserved quantity @xmath7 such as the total number of particles or magnetization
. however , everything generalizes in a very similar manner to the cases with multiple conserved quantities and non - abelian symmetries @xcite . for the latter one exploits that dependencies inside each multiplet are given by clebsh - gordan coefficients as exemplified by the wigner - eckart theorem .
let us consider a lattice system with @xmath6 sites , orthonormal on - site basis states @xmath8 , and associated quantum numbers @xmath9 , i.e. , @xmath10 , where @xmath11 is the conserved quantity .
a _ matrix product state _ has the form @xmath12 with @xmath13 and @xmath14 matrices @xmath15 .
the @xmath16 are referred to as _ bond dimensions_. for the matrix product to yield a scalar , we require @xmath17
. one can enforce the mps @xmath18 to have quantum number @xmath19 by assigning quantum numbers to the bond indices @xmath20 and @xmath21 of the matrices , imposing that matrix elements @xmath22_{a_i , b_i}$ ] can be nonzero only if the quantum number constraint @xmath23 is obeyed , and enforcing @xmath24 and @xmath25 . noting that , in the matrix product , @xmath26 , the quantum number constraint leads indeed to @xmath27 for all nonzero contributions in the mps .
this is in agreement with the traditional interpretation of the mps tensors @xmath28 in the dmrg @xcite . from that point of view ,
the tensors take the role of projectors from a reduced hilbert space of dimension @xmath29 with basis @xmath30 for the block of sites @xmath31 $ ] to another reduced @xmath16-dimensional hilbert space with basis @xmath32 for the same block , such that @xmath33_{a_i , b_i}|a_i{\rangle}\otimes|n_i{\rangle}$ ] .
the constraint then means that mixing of quantum numbers is prohibited in these projections . for systems with multiple conserved quantities ,
@xmath34 , the only difference to the above is that quantum number labels become vectors such that @xmath35 , @xmath36 and @xmath37 .
the explicit implementation of symmetries can reduce computation costs considerably .
typical costly operations for matrix product states , e.g. , in a time - evolution or ground - state calculation , are singular value decompositions of the @xmath38-tensors like @xmath39 , where @xmath40 and @xmath41 are isometric according to @xmath42 and @xmath43 , and @xmath44 is a diagonal matrix containing the singular values .
the cost of such a singular value decomposition scales as @xmath45 with the onsite hilbert space dimension @xmath46 and the bond dimension @xmath47 .
now , with the implementation of symmetries according to the quantum number constraint , the @xmath38-tensors assume a block structure with nonzero blocks corresponding to groups of states obeying eq . .
a singular value decomposition can then be done block by block .
often , the total computation cost is then dominated by the biggest block , the dimensions of which can be much smaller than the total bond dimensions @xmath16 .
the hamiltonian @xmath48 commutes with the conserved quantity @xmath7 . in a situation where the system exchanges energy with a bath but these interactions commute with @xmath7 , according to jaynes maximum entropy principle from statistical mechanics , the equilibrium state of the system
is given by the ( here , unnormalized ) _ canonical ensemble _
@xmath49 with @xmath50 being the projection of the hamiltonian onto the quantum number @xmath19 subspace @xmath51 of the full hilbert space @xmath52 . according to the same principle , the equilibrium state is the _ grand - canonical ensemble _
@xmath53 if system and bath also interchange the quantity associated with @xmath7 . here , @xmath54 is the associated lagrange multiplier that fixes the expectation value of @xmath7 .
two examples are that @xmath7 is the total magnetization @xmath55 and @xmath56 the magnetic field or that @xmath7 is the total particle number @xmath57 and @xmath56 the chemical potential . in more complex cases with multiple conserved quantities
@xmath58 , one can also consider ensembles like @xmath59 $ ] , where @xmath60 is fixed to @xmath61 and the expectation value @xmath62 is fixed by the lagrange multiplier @xmath63 .
an example is the fermi - hubbard model , where the local basis states for site @xmath64 are @xmath65 with electron spin @xmath66 and particle number @xmath67 .
the total number of electrons @xmath68 could be fixed , but , for example due to a transverse magnetic field , only the expectation value of the magnetization @xmath69 might be fixed by a magnetic field in @xmath70-direction . in passing
, we will also comment on such ensembles .
we want to compute finite - temperature expectation values @xmath71 with respect to the grand - canonical ensemble using matrix product purifications and exploiting the @xmath7-conservation .
the partition function @xmath72 is needed for normalization . as a purification of the density matrix @xmath73 on @xmath3
, we can employ the vectorization @xmath74 of the density matrix at twice the temperature @xcite .
the vectorization @xmath75 of any operator @xmath76 on @xmath3 with respect to the orthonormal basis @xmath77 is defined as @xmath78 the vectorization @xmath79 of @xmath80 is a purification of the density matrix @xmath73 in the sense that @xmath81 where the partial trace is taken over the second ( _ auxiliary _ or _ ancilla _ ) part of the tensor product space @xmath82 . with the purification , one can compute expectation values as @xmath83 an accurate matrix product approximation of the purification in the form @xmath84 can be constructed by starting at infinite temperature ( @xmath85 ) with [ eq : gce_beta0 ] @xmath86 where @xmath87 is a basis state for site @xmath64 of the second ( auxiliary ) part in @xmath82 . an exact matrix product representation with bond dimensions @xmath88
is given by @xmath89_{1,1}=\delta_{n_i , n_i'}$ ] . with @xmath90 as the initial state , one can time - evolve the mpp in imaginary - time using the time - dependent dmrg algorithm ( tdmrg ) @xcite or time - evolved block decimation ( tebd ) @xcite to obtain an mpp for the finite - temperature state @xmath73 , @xmath91 as the purification @xmath79 of @xmath73 is simply the vectorization of @xmath80 , eq .
yields at the same time an approximation of @xmath80 in mpdo form @xcite . while the evolution operator in eq . commutes with both @xmath92 and @xmath93 ,
we can not utilize the corresponding conservation laws , because the initial state @xmath90 is not an eigenstate of either of them .
all quantum number sectors contribute in the infinite - temperature state .
however , it is an eigenstate of @xmath94 with eigenvalue @xmath95 . in analogy to what was described in section [ sec : symmetries ] , this can be exploited to reduce computation costs . with local quantum numbers @xmath96 ,
the quantum number constraint for the tensors @xmath89_{a_i , b_i}$ ] in the matrix product reads @xmath97 , in analogy to eq . .
with @xmath98 , it guarantees that @xmath99 is conserved in all operations .
an mpp or mpdo for the canonical ensemble can be computed quite similarly the biggest difference being the initial state at @xmath85 .
the restriction of the identity to the quantum number @xmath19 subspace @xmath51 can be written as @xmath100 clearly , this is an eigenstate of both , @xmath101 and @xmath93 such that @xmath102 . in a matrix product representation of it
, we can hence exploit the two conservation laws by assigning tuples @xmath103 of quantum numbers to the bond indices . for bond @xmath104 , these are the eigenvalues of @xmath105 and @xmath106 , respectively .
the quantum number constraints for the tensors @xmath89_{a_i , b_i}$ ] in the mpp then read @xmath107 in analogy to eq . .
with @xmath108 and @xmath109 , it is guaranteed that @xmath110 is conserved in all operations .
mps methods to construct the infinite - temperature state are described in the next section . in the subsequent imaginary - time evolution @xmath111 which can be implemented using tdmrg , the growth of the mps bond dimensions @xmath16 is controlled by truncations .
details on the truncation scheme that we employ can be found in refs .
@xcite . finally , in analogy to the grand - canonical case
, one can compute thermal expectation values with the formula @xmath112
an explicit matrix product representation of the canonical infinite - temperature state is provided in section [ sec : ce_beta0_exact ] , but it may be inconvenient to incorporate in existing mps codes .
therefore , let us first describe an alternative scheme .
the state can be generated by starting from a projection onto the minimum quantum number state ( vacuum ) and to then repeatedly apply a matrix product ( super-)operator that increases the quantum number until reaching quantum number @xmath19 and the state .
let us shortly describe the procedure for systems of bosons , spins , and fermions . for bosons ,
let us consider the case where the local bases have been truncated to allow for at most @xmath113 bosons per site .
the ( projected ) ladder operators are @xmath114 and , hence , @xmath115 .
the purification @xmath116 of the identity in the @xmath117-particle subspace can be written as @xmath118 where @xmath119 is the vacuum state with @xmath120 @xmath121 and @xmath122 ( the first @xmath123 acting in the primary hilbert space and the second @xmath123 acting in the auxiliary hilbert space ) . to this purpose ,
one first builds an mps representation of the vacuum state @xmath124 by choosing @xmath125 matrices @xmath126 .
alternatively , it can also be generated variationally as the ground state of the hamiltonian @xmath127 .
the operator @xmath128 can be written as an mpo with bond dimension 2 ; @xmath129 according to , we apply this mpo @xmath117 times to the mps representation of the vacuum state @xcite . for a spin-@xmath130 system with local bases
@xmath131 , one can proceed in exactly the same way , starting from the fully polarized state @xmath132 , i.e. , the state with eigenvalues @xmath133 for all operators @xmath134 and @xmath135 . in order to really obtain an equal weight superposition in the sector with total magnetization @xmath136 , for general @xmath130 , one should however not apply ( @xmath117 times ) the operator @xmath137 but the analog of the bosonic operator @xmath138 . to this purpose
, one can define the ladder operators as @xmath139 . for spin-@xmath0 systems , having single - site hilbert space dimensions @xmath140 , @xmath128 coincides with @xmath141 . for fermionic systems ,
the approach works just as well by replacing the bosonic ladder operators @xmath123 with fermionic ladder operators @xmath142 . in the construction of the infinite - temperature purification , one does not need to worry about fermionic sign factors , as operators always occur in pairs like @xmath143 .
two such products always commute , @xmath144=0 $ ] @xmath145 .
let us shortly discuss the bond dimensions required for matrix product representations of the canonical infinite - temperature states .
this is most transparently done in an operator - space representation . in the following
, we will consider a bipartition of the lattice into a left part @xmath146 $ ] and a right part @xmath147 $ ] .
the required matrix dimensions @xmath16 of an accurate matrix product representation of @xmath148 , or the corresponding purification @xmath149 , can be determined by writing it as a sum of tensor product operators ; @xmath150 where @xmath151 is the identity in the quantum number @xmath152 subspace of subsystem @xmath153 .
analogously , @xmath154 is the identity in the quantum number @xmath155 subspace in subsystem @xmath156 . for a bosonic or fermionic system with @xmath157 particles ( or a spin-@xmath130 system with @xmath158 ) and a bipartition at a bond @xmath104 sufficiently far away from the system boundaries ,
the sum in the decomposition would contain @xmath159 orthogonal terms , implying that we need dimensions @xmath160 for the matrix product representations of @xmath161 or the corresponding purification @xmath116 .
for a cut at a bond @xmath104 close to the left boundary , with @xmath162 , the sum would only contain @xmath163 terms ( the left subsystem can contain @xmath164 to @xmath165 particles ) , and a bond dimension @xmath166 would hence suffice at such bonds . see fig . [
fig : ce_nrange ] .
in essence , for an exact matrix product representation of canonical infinite - temperature states , the maximum bond dimension @xmath167 increases linearly in the system size if we keep the particle number density @xmath168 fixed .
in contrast , grand - canonical infinite - temperature states are product states such that @xmath88 @xmath121 .
if one exploits symmetries as explained in section [ sec : ce ] , the computation costs for the canonical ensemble at high temperatures are actually lower than what one might expect form the growth of the mps bond dimensions @xmath16 . according to the decomposition of the infinite - temperature state , every symmetry sector has dimension one such that the matrices @xmath169 in the matrix product representation consist of @xmath16 blocks of size @xmath125 .
hence , at high temperatures , computation costs do not scale as @xmath170 but are instead linear in @xmath16 . based on the decomposition
@xmath171 we can given an explicit matrix product representation of the canonical infinite - temperature state .
here , @xmath172 denotes the identity in the quantum number @xmath152 subspace of the subsystem @xmath173 .
a valid matrix product representation is given by [ eq : mpp_ce_beta0 ] @xmath174_{q',q''}=\delta_{n_i , n_i'}\delta_{q'+q(n_i),q''}.\label{eq : mpp_ce_beta0_mat}\end{gathered}\ ] ] here , we have used @xmath152 and @xmath155 as bond indices , which label at the same time the total quantum numbers for subsystems @xmath175 and @xmath176 , respectively .
they are hence eigenvalues of @xmath177 and @xmath178 . using the matrices of the mpdo in eq .
gives the infinite - temperature mpp @xmath149 .
the tuples @xmath103 of quantum numbers , assigned to bond indices of the mpp as discussed in section [ sec : ce ] , are then simply @xmath179 and @xmath180 .
the possible ranges of the indices and hence the matrix dimensions are easy to determine . as an example
, consider again a bosonic or fermionic system with @xmath157 particles or a spin-@xmath130 system with @xmath158 . in this case ,
@xmath181 and @xmath182 denote subsystem particle numbers , obeying the constraints @xmath183 where @xmath184 is the maximum ( allowed ) number of particles per site .
see fig .
[ fig : ce_nrange ] . the mpp of the grand - canonical infinite - temperature state , given in section [ sec : gce ] , can be considered as a special case of eq . through assigning to all states the same trivial quantum number @xmath185 , such that all bond dimensions are @xmath186 .
in the formulation above , the approach is generically applicable . as a specific example , consider ,
the fermi - hubbard model which has local basis states @xmath65 with occupation numbers @xmath67 .
assume the ensemble , where the total number of electrons @xmath68 is fixed to @xmath117 and the magnetization expectation value @xmath69 is fixed by a magnetic field .
the @xmath38-tensors of the corresponding infinite - temperature state are then @xmath187_{n',n''}=\delta_{n_{{\uparrow}},n_{{\uparrow}}'}\delta_{n_{{\downarrow}},n_{{\downarrow}}'}\delta_{n'+n_{{\uparrow}}+n_{{\downarrow}},n''}$ ] , where @xmath188 and @xmath189 take integer values as specified in eq . with @xmath190 .
earlier , it has been suggested to obtain the infinite - temperature purifications @xmath191 and @xmath192 [ eqs . and ] by dmrg ground - state computations for so - called entangler hamiltonians @xcite .
let us shortly explain this approach and point out why , especially for the canonical ensembles , it should be less favorable than the more direct generating procedures described in sections [ sec : ce_beta0_mpo ] and [ sec : ce_beta0_exact ] . for grand - canonical infinite - temperature ensembles
a variational computation is not really necessary . as described in section [ sec : gce ] , there is an explicit simple matrix product representation with bond dimension @xmath88 @xmath121 .
if however , for some practical reason , one needs to prepare this infinite - temperature purification by a variational computation , this can be done by computing @xmath191 as the zero - energy ground state of the hamiltonian @xmath193 where @xmath194 projects sites @xmath64 of the primary system and the auxiliary system onto the maximally entangled state @xmath195 occurring in eq . .
canonical infinite - temperature states are less trivial . in sections [ sec :
ce_beta0_mpo ] and [ sec : ce_beta0_exact ] , we have discussed two direct approaches to generate @xmath192 . as suggested in refs .
@xcite , an alternative procedure is to compute @xmath149 variationally . in a somewhat modified and generalized form , applicable to arbitrary systems ,
this approach consists in computing the ground state of the hamiltonian @xmath196 in @xmath197 .
the positive semi - definite operators @xmath198 act on sites @xmath64 and @xmath199 of both the primary and the auxiliary systems .
specifically , @xmath200 denotes what we call a _ shuffle operator _
, @xmath201 where @xmath202 is the dimension of the two - site subspace with quantum number @xmath203 . as will be shown below
, @xmath149 is the unique ground state of the hamiltonian in the diagonal subspace @xmath204 however , besides the complication of long - range interactions and comparatively strong entanglement in @xmath149 which can cause a slow and computationally costly convergence , there is another possible pitfall .
if the variational mps optimization to compute the ground state of @xmath205 is started with a state that happens to have no component in the diagonal subspace @xmath206 , the optimization would not converge to @xmath149 , but to a state in the orthogonal complement of @xmath206 .
the latter problem can in fact be resolved by restricting the optimization to the diagonal subspace as follows . instead of computing the ground state @xmath149 of @xmath205 in @xmath197
, we can compute the closely related equal - weight superposition @xmath207 ( @xmath51 is isomorphic to @xmath206 ) , which , in the following , will be shown to be the unique ground state of @xmath208 this is the restriction of @xmath205 to the diagonal subspace @xmath206 with shuffle operators @xmath209 after computing the ground state @xmath210 of this hamiltonian in mps form @xmath211 , the mpp @xmath149 in the form is obtained by defining @xmath212 @xmath213 .
let us discuss why the equal - weight superposition @xmath210 is indeed the unique ground state of @xmath214 in @xmath51 . as @xmath214 is a sum of positive semi - definite operators @xmath215 , are block - diagonal .
each block corresponds to a fixed quantum number @xmath216 .
in the two - site basis @xmath217 the block matrix is @xmath218 , where @xmath219 denotes the @xmath220 all - ones matrix .
@xmath221 has one eigenvalue 1 and all others are zero . consequently , the operators @xmath222 of eq . have , in each quantum number sector of the two - site hilbert space , one eigenvalue 0 and all others 1 .
and thus , @xmath215 , and also @xmath198 in eq . .
] its smallest possible eigenvalue is zero , i.e. , @xmath223 .
additionally , @xmath224 , such that @xmath210 is _ a _ ground state of @xmath214 .
furthermore , it is _ the unique _ ground state according to the perron - frobenius theorem , applied to @xmath225 in the basis @xmath77 , which is a hermitian irreducible non - negative matrix .
[ fig : ce_hentang ] shows the convergence behavior of the variational computation of @xmath210 . for the spin-@xmath0 chains
, one sees that the required number of dmrg sweeps increases substantially with the system size @xmath6 .
for the bosonic systems , the required number of sweeps does not change as strongly with the system size .
the exact matrix product representations of canonical infinite - temperature states @xmath226 allow us to compute the weights @xmath227 of quantum numbers @xmath19 in any other mpdo @xmath2 . with @xmath228 ,
the probabilities @xmath227 can be obtained by evaluating hilbert - schmidt inner products @xmath229 as an example , let us consider the grand - canonical ensemble @xmath230 as described above , dmrg computations are often formulated in terms of purifications instead of mpdos . with mpps @xmath231 , @xmath232 , and @xmath233 , the quantum number probability distribution
is then given by @xmath234 here , we have used the grand - canonical infinite - temperature state to obtain the normalization factor @xmath235 .
note that , whereas one employs the purifications @xmath79 for the evaluation of expectation values at inverse temperature @xmath236 [ eq .
] , one needs to go twice as far in imaginary time and use @xmath232 for the evaluation of probabilities @xmath227 at inverse temperature @xmath236 .
let us exemplify the described matrix product techniques for antiferromagnetic spin-@xmath0 heisenberg chains @xmath237 where @xmath238 .
we consider zero magnetization , i.e. , magnetic field @xmath239 for the grand - canonical ensemble and @xmath240-eigenvalue @xmath241 for the canonical ensemble .
we do quasi - exact mpp / mpdo simulations of both ensembles with imaginary - time step @xmath242 and a truncation threshold @xmath243 , which fixes the matrix product bond dimensions dynamically as detailed refs .
@xcite .
the upper panel of fig .
[ fig : ensembleequivalence ] shows the difference of the energy densities for the two ensembles .
it decays as @xmath1 , indicating that the two ensembles are equivalent in the thermodynamic limit .
the lower panel shows the standard deviation of the energy density in the grand - canonical ensemble .
it decays as @xmath244 .
this @xmath244 decay is well - known from the theory of thermodynamics .
it follows from the fact that the energy is extensive and that its derivative with respect to the inverse temperature @xmath236 , which is intensive , yields the energy variance .
@xmath245 such that @xmath246
. follows , e.g. , from @xmath247 . ]
the decay of the energy density difference @xmath248 shown in the upper panel of fig .
[ fig : ensembleequivalence ] can be explained as follows . according to thermodynamics , the standard deviation of the magnetization @xmath249 in the grand - canonical ensemble decays as @xmath244 , just like @xmath250 in eq . .
beyond that , the appropriately rescaled magnetization probability density converges in the thermodynamic limit such that @xmath251 . using the technique described in section [ sec : mpdo_qn - distr ] ,
we have computed the magnetization probabilities @xmath252 in the grand - canonical ensemble .
the results are displayed in fig .
[ fig : gce_szdistribution ] .
the scaling function @xmath253 is approximately of gaussian shape for the considered temperatures .
we have also computed the energies @xmath254 in canonical ensembles for total magnetizations @xmath255 . as shown in fig .
[ fig : ce_energy ] , the energy density @xmath256 as a function of the magnetization @xmath257 converges with increasing @xmath6 .
because the global ground state of the isotropic heisenberg antiferromagnet has zero magnetization , the energy density has a minimum at @xmath258 for the considered temperatures .
due to the global spin - flip symmetry , the scaling function is symmetric around @xmath258 and is found to be of the form @xmath259 . combining these observations , we find for the difference of the canonical and grand - canonical energy densities that @xmath260 which explains the @xmath1 scaling in the upper panel of fig .
[ fig : ensembleequivalence ] .
the same @xmath1 scaling of the energy density difference , @xmath261 occurs for nonzero magnetizations , i.e. , if we compare the canonical ensemble @xmath262 for total magnetization @xmath263 to the grand - canonical ensemble with magnetic field @xmath264 chosen such that @xmath265 . in that case
, @xmath266 has its minimum at @xmath267 and the probability distribution converges like @xmath268 .
it is straightforward to compute an mpp / mpdo of a grand - canonical ensemble , as the infinite - temperature state is simply a product state and has hence a matrix product representation with bond dimension one ( sect .
[ sec : gce ] ) . for systems with symmetries
it can be computationally beneficial to work instead with canonical ensembles . while , for a system with conserved quantity @xmath7 , grand - canonical ensembles obey @xmath269=0 $ ] , canonical ensembles obey @xmath270 .
this can be used to reduce computation costs substantially ( sect .
[ sec : ce ] ) . sometimes , canonical ensembles also better reflect experimental conditions ( e.g. , fixed particle number in experiments with ultracold atoms ) .
to compute matrix - product representations of canonical ensembles is somewhat more complicated as the corresponding infinite - temperature states are nontrivial . we have described and discussed different techniques to generate these states in matrix product form ( sect .
[ sec : ce_beta0 ] ) .
the cleanest way is certainly to directly use the exact matrix product representation .
if this is not feasible for practical reasons , one can alternatively generate canonical infinite - temperature states by applying mpos multiple times to vacuum - type states ( sect .
[ sec : ce_beta0_mpo ] ) or by doing variational computations with entangler hamiltonians ( sect .
[ sec : entangler ] ) . in the latter case
, one should either work in the diagonal subspace or make at least sure that the initial state for the variation has components in the diagonal subspace .
another useful application of the constructed canonical infinite - temperature states is to study probability distributions of global quantum numbers in mixed states ( mpdos ) . to this purpose
one needs to evaluate corresponding hilbert - schmidt inner products ( sect .
[ sec : mpdo_qn - distr ] ) . c. regg , k. kiefer , b. thielemann , d. f. mcmorrow , v. zapf , b. normand , m. b. zvonarev , p. bouillot , c. kollath , t. giamarchi , s. capponi , d. poilblanc , d. biner , and k. w. krmer , phys .
* 101 * , 247202 ( 2008 ) .
|
matrix product purifications ( mpp ) are a very efficient tool for the simulation of strongly correlated quantum many - body systems at finite temperatures .
when a system features symmetries , these can be used to reduce computation costs substantially .
it is straightforward to compute an mpp of a grand - canonical ensemble , also when symmetries are exploited .
this paper provides and demonstrates efficient methods to compute mpps of canonical ensembles under utilization of symmetries .
furthermore , we present a scheme for the evaluation of global quantum number distributions using matrix product density operators ( mpdo ) .
we give exact matrix product representations for canonical infinite - temperature states , and discuss how they can be constructed alternatively by applying matrix product operators to vacuum - type states or by using entangler hamiltonians .
a demonstration of the techniques for heisenberg spin-@xmath0 chains explains why the difference in the energy densities of canonical and grand - canonical ensembles decays as @xmath1 .
|
the most common type of functional data are curves , typically measured over time .
also spectral data , measured over a range of wavelengths , can be considered as functional data and can be analyzed as such @xcite .
when several measurements are taken at each time point , we obtain multivariate functional data .
bivariate examples are height and weight curves of children , temperature and dewpoint temperature at several weather stations during several days , or measurements of the human heart activity at two different places on the body @xcite .
whereas the domain of these data is always univariate , the response can thus be multivariate .
surface data are an increasingly common data type in several fields of research . in chemistry for example ,
fluorescence spectroscopy is a well known technique that yields surface data with an excitation and an emission dimension . in geography ,
surface data is obtained by measuring characteristics such as daily precipitation on a certain area of land . also digital image and video data
are represented by a grid of pixels and typically contain grayscale values or three - dimensional rgb values that define the color intensities .
all these advanced data types can thus be seen as multivariate functional data with a bivariate domain . to detect outlying surfaces or to flag outlying parts of a surface ,
it is well known that classical statistical techniques are not trustworthy as the analysis itself may be distorted by the outliers .
outlier detection for multivariate functional data with a univariate domain has been studied in depth in @xcite . in that paper a taxonomy of outlying curves
was proposed , and distribution - free statistical methods were introduced to measure their degree of outlyingness .
@xcite illustrated this on a raw excitation - emission fluorescence dataset , and proposed the functional outlier map ( fom ) as a graphical display to visualize the functions according to their degree and type of outlyingness . in this paper , we build on these methods with a threefold objective .
first , we propose a cutoff to distinguish between the outlying and the regular surfaces , which makes the fom more informative .
secondly , we improve the outlier detection ability for surface data by first modeling the data according to a multiway model and applying the fom to the residuals instead of the raw data .
finally , we illustrate the extension of these techniques to multivariate image and video data .
the next section introduces the methodology behind the fom as well as our new outlier detection rule . in section [ sec : surface ]
we apply the new rule to excitation - emission matrices before and after fitting a parafac model to them .
section [ sec : image ] demonstrates the performance of our method on mri image data and introduces the idea of including gradients in the analysis . in section [ sec : video ]
we apply the methodology to a video consisting of 633 color images .
the main concept for measuring the degree of outlyingness is the _ adjusted outlyingness _ ( ao ) of @xcite .
its construction is recalled in the appendix . for univariate data ,
the ao is a robust version of the absolute @xmath0-score as it measures an observation s deviation from the median .
it allows for skewness in the data by estimating scale separately on either side of the median . for multivariate data ,
the ao of a point is defined as its maximal univariate ao when projecting the data on many directions . to extend the ao to multivariate functional data , @xcite defined the _ functional adjusted outlyingness _ ( fao ) of a @xmath1-dimensional curve @xmath2 as the ( weighted ) average of its ao values over all points of its domain .
more precisely , let @xmath3 be a sample of @xmath1-dimensional curves , recorded at time points @xmath4 .
the fao of @xmath2 with respect to @xmath5 is then defined as @xmath6 where @xmath7 is a sample in @xmath1 dimensions and @xmath8 is a weight function for which @xmath9 .
this weight function allows to assign a different importance to the outlyingness of a curve at different time points .
one could for example downweight time points near the boundaries if measurements are recorded less precisely at the beginning and the end of the process .
definition for functional data with a univariate domain can easily be extended to functions with a bivariate domain such as surfaces and images .
we then assume that the outcome of the experiment is recorded at a grid of discrete points , e.g. discrete excitation and emission wavelengths .
therefore , it is convenient to use two indices @xmath10 and @xmath11 , one for each dimension of the grid , to characterize these points . a surface or image
can then be represented by a @xmath12 matrix where each cell @xmath13 contains the height of the surface or the color intensity at that grid point .
we then define the functional adjusted outlyingness of a multivariate function @xmath2 with bivariate domain relative to a dataset @xmath5 of functions recorded at grid points @xmath14 as @xmath15 with @xmath16 . for illustrative purposes we will use a uniform weight function throughout this paper , i.e. @xmath17 for all @xmath10 and @xmath11 .
definition can trivially be extended to functions with a multivariate domain of dimension @xmath18 , such as three - dimensional images consisting of voxels .
the fao can be applied to raw data , but in other applications it might be interesting to first fit a parametric model to the data , after which the fao can be computed on the residuals . for multiway data , parafac is a commonly used method to fit a trilinear model .
the concept of applying fao to the residuals of a parafac model is illustrated on a real data example in section [ sec : surface ] . when analyzing curves , it is often informative to consider the derivative of the curves as well .
in particular , this is helpful when the goal is to detect curves with a deviating shape , see @xcite for some examples . for surfaces or images
, we can augment the available raw data by including gradients instead .
this idea will be illustrated in section [ sec : image ] .
the functional outlier map ( fom ) was introduced in @xcite as a graphical tool to visualize the degree of outlyingness of curves .
a similar definition can be applied to surfaces or images .
the fom is a scatter plot which displays each function s fao on the horizontal axis and vao , a measure of the variability of its ao over all points of the domain , on the vertical axis . more precisely , for each @xmath19 the vao is defined as @xmath20 so the fom is a scatter plot of the points @xmath21 for @xmath22 .
we divide by fao in equation in order to measure relative instead of absolute variability .
this can be understood as follows .
suppose that the functions are centered around zero and that @xmath23 for all @xmath24 and @xmath25 .
then @xmath26 but their relative variability is the same . because @xmath27 , dividing by fao solves this problem
the objective of the fom is to reveal outliers in the data , and its interpretation is fairly straightforward .
we use the taxonomy of functional outliers presented by @xcite to distinguish between different types of outliers .
points in the lower left part of the fom represent regular surfaces which hold a central position in the dataset .
points in the lower right part are surfaces with a high fao but a low variability of their ao values .
this happens for shift outliers , i.e. surfaces which have the same shape as the majority but are shifted on the whole domain .
points in the upper left part have a low fao but a high vao .
typical examples are isolated outliers , i.e. surfaces which only display outlyingness over a small part of their domain .
the points in the upper right part of the fom have both a high fao and a high vao .
these correspond to surfaces which are strongly outlying on a substantial part of their domain .
in this paper we add an additional feature to the fom , namely a rule to flag the outliers . for this purpose
we define a summary measure , the _ combined functional outlyingness _ ( cfo ) of a surface @xmath19 as @xmath28 where @xmath29 , @xmath30 and similarly forvao .
note that the cfo characterizes the points on the fom through their euclidean distance to the origin , after scaling .
we expect outliers to have a large cfo . in general
, the distribution of the cfo is unknown but skewed to the right . to define a cutoff
, we first symmetrize the cfo by the logarithmic transformation .
then we center and scale the resulting values in a robust way and compare them with a high quantile of the gaussian distribution .
more precisely , let @xmath31 for all @xmath22 .
then we flag function @xmath19 as an outlier if @xmath32 in the denominator , mad denotes the median absolute deviation defined as@xmath33 .
in addition to the fom , it is often instructive to plot the ao values themselves over their domain . for a function with a bivariate domain ,
this yields a two - dimensional heatmap .
these plots can reveal particularly outlying regions and provide more insight into why certain surfaces are flagged as outliers by the fom .
this will be illustrated in the following sections .
the dorrit dataset contains excitation - emission landscapes of 27 mixtures of 4 fluorophores in an aqueous solution and has been studied extensively by @xcite , @xcite and @xcite .
such landscapes are typically stored in excitation - emission matrices ( eem ) .
these matrices contain measurements of fluorescence intensity for 116 emission spectra ranging from 250 to 482 nm at 18 excitation wavelengths ranging from 230 to 315 nm .
this yields 27 samples @xmath19 , each containing measurements @xmath34 for @xmath35 and @xmath36 .
raman and rayleigh scattering , showing up in figure [ fig : l8](a ) as two diagonal ridges in landscape 8 , are a common problem in this type of experiment and can severely distort the analysis . to resolve this we used the method of @xcite which automatically identifies the scattering .
after setting the scattering to missing values , we imputed them by interpolating each excitation profile .
the result of this process is illustrated in figure [ fig : l8](b ) . using definition with a uniform weight function ,
we obtain the functional outlier map depicted in figure [ fig : dorritfom ] .
the dashed black line is determined by the cutoff value and shows the boundary between the regular data and the outliers .
the outliers , plotted as red squares , are landscapes 3 and 5 .
we can improve our analysis by incorporating more information , for example by applying parallel factor analysis ( parafac ) to the data .
parafac is known to be a powerful tool for analyzing multiway data and can in particular be used to decompose three - way eem data into a trilinear model .
this trilinear model states that the value of a landscape @xmath19 at ee - point @xmath13 can be written as the product of three corresponding values of the score matrix @xmath37 and the loading matrices @xmath38 and @xmath39 , plus an error term : @xmath40 here @xmath41 denotes the number of components which we take as @xmath42 , corresponding to the 4 known fluorophores present in the mixtures . applying parafac to the data results in 27 residual landscapes @xmath43 . with the objective of outlier detection in mind ,
it is important to use a robust parafac algorithm as it is well known that the presence of outliers can severely impact the parafac decomposition .
we have used the robust parafac algorithm proposed by @xcite in which we set the robustness parameter @xmath44 .
this parameter indicates that the parafac model is based on the @xmath45 of surfaces which yield the smallest residuals . by computing the fao on the residuals instead of the original data , the trilinear structure of the dataset
is taken into account .
the resulting functional outlier map is shown in figure [ fig : dorritfom_parafac ] and flags landscapes 2 , 3 and 5 .
these outliers were also detected in @xcite .
note that points 3 and 5 lie much further from the boundary ( the dashed black curve ) than in the fom of figure [ fig : dorritfom ] .
this indicates that surfaces 3 and 5 have outlying values but mostly expose an outlying structure .
landscape 2 barely contains abnormally low or high values but has large residuals with respect to the parafac model fitted to the majority of eem landscapes .
a heatmap of the ao values over the domain allows us to inspect surface 2 in more detail .
figure [ fig : aomap_dorrit](a ) plots the aos of the original landscape 2 , whereas figure [ fig : aomap_dorrit](b ) shows the aos of its parafac residuals . to improve visibility ,
all ao values of 15 or higher received the darkest color .
the left plot indicates that landscape 2 indeed does not have large outlying regions , as the vast majority of its cells are light - colored .
however , the right plot shows that its residuals are significantly outlying in the emission spectra between 275 and 350 nm .
this indicates that the shape of landscape 2 differs strongly from the shape predicted by the parafac model .
therefore , this landscape can be classified as a shape outlier .
the mri dataset contains brain imaging data of 416 persons , aged between 18 and 96 years @xcite and can be freely accessed at _ www.oasis - brains.org_. for each person , multiple mri scans are available including 3 to 4 raw mri images and several processed scans such as an averaged image of the raw scans , the atlas - registered gain field - corrected image and a masked version of the latter scan resampled to 1 mm isotropic pixels . masking sets all non - brain pixels to an intensity value of zero .
moreover these images are normalized , meaning that the size of the head is exactly the same in each image .
the masked images have 176 by 208 pixels with grayscale values between 0 and 255 .
all together we thus have 416 observed images @xmath19 containing univariate intensity values @xmath34 , where @xmath46 and @xmath47 .
there is more information in the image than just the raw values .
we can incorporate part of the shape information by computing the gradient in every pixel of the image .
this gradient is composed of the derivatives in the horizontal and the vertical direction .
more precisely , the gradient in pixel @xmath13 is defined as the 2-dimensional vector @xmath48 . in practice
we do not have the exact derivatives of the image at the recorded pixels , so we have to rely on numerical approximations .
we use forward and backward finite differences to approximate the derivatives in the pixels at the boundary of the brain . for the other pixels
, we have used central differences . in the horizontal direction
these are given by @xmath49 the derivatives in the vertical direction are computed analogously . adding these derivatives to the data yields a multiway dataset of dimensions @xmath50 , so each @xmath34 is trivariate . for each person
we thus have three data matrices which represent the original mri image and its derivatives in both directions .
figure [ fig : mridata ] represents these three matrices for person 144 .
the functional adjusted outlyingness of an mri image @xmath19 relative to the sample @xmath5 is given by : @xmath51 where @xmath52 is the adjusted outlyingness of the trivariate pixel @xmath13 .
the resulting fom is presented in figure [ fig : mrifom ] .
this plot indicates the presence of 8 outliers of several types .
image 126 has a remarkably high fao combined with a relatively low vao .
this suggests a shift outlier , i.e. a function whose values are all shifted relative to the majority of the data .
images 29 and 92 have a large fao in combination with a high vao , indicating that they have strongly outlying subdomains .
images 108 , 188 and 234 have an fao which is on the high end relative to the dataset but which by itself does not make them outlying . only in combination with their large vao
are they flagged as outliers .
images 69 and 211 have a cfo value just slightly above the cutoff , meaning they are borderline cases .
a heatmap of the ao values can help to understand the root cause of an image s outlyingness . in figure
[ fig : mrigrid ] we compare the mri images ( left ) and the ao maps ( right ) of persons 144 , 92 , and 126 .
( ao values of 8 or higher received the darkest color . ) image 144 has the smallest cfo value , and as such can be thought of as the least outlying image in the dataset .
as expected , the ao heatmap of image 144 shows very few outlying pixels . for person 92 , the ao heatmap nicely marks the region in which the mri image deviates most from the majority of the images .
note that the boundaries of this region have the highest outlyingness .
this is the result of including the derivatives in the analysis , as they emphasize the pixels at which the grayscale intensity changes .
the ao heatmap of person 126 does not show any extremely outlying region but a rather high outlyingness over the whole domain , which explains its large fao and regular vao value .
the actual mri image to its left is globally lighter than the others , which confirms that it is a shift outlier .
this dataset consists of 633 images which together form a surveillance video of a beach , filmed with a static camera @xcite .
the data and the original video can be found at _ http://perception.i2r.a - star.edu.sg / bk_model / bk_index.html_. the video first shows a beach with a tree during 8 seconds , shown in the leftmost frame of figure [ fig : videodata ] .
then a man enters the screen from the left ( second frame ) , disappears behind the tree ( third frame ) , and then reappears again on the right side of the tree and stays on screen until the end of the video .
the aim of applying our method to this dataset is to detect the man in this video .
the images have @xmath53 pixels and are stored using the rgb ( red , green and blue ) color model , so each image corresponds to three matrices , each of dimensions @xmath53 .
overall we have 633 images @xmath19 containing trivariate @xmath34 for @xmath54 and @xmath55 . computing the fao according to yields the fom in figure [ fig : videofom ] .
it is very instructive as the path of the man in the video can be traced in it .
the first 480 images , which depict the beach and the tree and where only the water surface causes slight variations , are found on the bottom left side of the fom inside the dashed curve that separates the regular frames from the outliers . around frame 483
the man enters and as a result the standard deviation of the adjusted outlyingness rises slightly .
the fao itself increases more slowly as the fraction of the pixels covered by the man is still low .
these frames can thus be classified as isolated outliers .
their cfo barely exceeds the cutoff so on the fom they are still close to the dashed line .
frames 484487 have a very high fao and vao . in these images ,
the man is clearly visible between the left border of the frame and the tree .
consequently these images have outlying pixels in a substantial part of their domain .
frames 488491 see the man disappear behind the tree . in these frames
the fao goes down as the fraction of outlying pixels decreases . from frame 492
onward the man reappears on the right side of the tree and stays in the screen until the end , as is reflected in the fom since these frames contain many outlying pixels .
in addition to the fom we can construct ao heatmaps of individual frames . for frames 100 , 487 , 491 and 500 , figure [ fig : videosum ] shows the raw image on the left , the ao heatmap in the middle and the fom on the right . on the fom
we have drawn a blue circle around the relevant frame to indicate its location .
this sequence of plots clearly shows that the proposed method works very well for this surveillance video data : not only can the man s path be followed on the fom , it is also clear from the ao heatmaps where exactly the man is in those frames .
we have created a video in which the raw image , the ao heatmap and the fom evolve over time alongside each other .
it can be downloaded from _ http://wis.kuleuven.be
/ stat / robust / publ_.
mh , jr and pr acknowledge the financial support of grants c16/15/068 andstrt/13/006 of the internal fund ku leuven .
marcus , d. , wang , t. , parker , j. , csernansky , j. , morris , j. , and buckner , r. ( 2007 ) .
pen access series of imaging studies ( oasis ) : cross - sectional mri data in young , middle aged , nondemented , and demented older adults .
, 19:14981507 .
the adjusted outlyingness of multivariate data has been introduced by @xcite and was studied in detail by @xcite .
we first recall the adjusted outlyingness of a point @xmath56 relative to a univariate dataset @xmath57 .
it can be seen as a robust version of the absolute @xmath0-score since it uses robust measures of location and scale ( instead of the usual mean and standard deviation ) .
it also accounts for skewness by computing the scale on each side of the median separately . as a robust measure of location
we take the median of @xmath58 .
the robust estimates of scale are based on the adjusted boxplot @xcite .
this modified boxplot is well - suited for skewed distributions .
first , an interval called the fence is computed . for right - skewed data
it is given by @xmath59\ ] ] with @xmath60 and @xmath61 the first and third quartile of @xmath58 , @xmath62 and @xmath63 the medcouple , a robust measure of skewness @xcite .
for left - skewed data a similar definition holds . note that for perfectly symmetric data the medcouple is zero , and then the fence is identical to that of the standard boxplot .
the whiskers of the adjusted boxplot are lines from the box to the most remote data points inside the fence .
figure [ fig : adjbox ] shows the adjusted boxplot of a right - skewed dataset @xmath58 .
the points @xmath64 and @xmath65 have the same distance @xmath66 to the median of @xmath58 , but @xmath64 lies outside the lower whisker @xmath67 whereas @xmath65 lies inside the upper whisker .
the adjusted outlyingness of @xmath68 is now defined as @xmath69 with @xmath70 and @xmath71 . for @xmath72
the ao is given by @xmath73 with @xmath74 and @xmath75 .
in this example @xmath76 .
+ ( 170,50 ) ( 5.00,20.00)(0,1)20.00 ( 5.00,30.00)(1,0)30.00 ( 35.00,20.00)(1,0)40.00 ( 35.00,20.00)(0,1)20.00 ( 35.00,40.00)(1,0)40.00 ( 75.00,20.00)(0,1)20.00 ( 75.00,30.00)(1,0)70.00 ( 145.00,20.00)(0,1)20.00 ( 50.00,20.00)(0,1)20.00 ( 0,30.00 ) ( 5,30.00 ) ( 100,30 ) ( 160,30.00 ) ( 145,30.00 ) ( 3.00,25.00)(0,0)[rr]@xmath64 ( 103.00,25.00)(0,0)[rr]@xmath65 ( 8.00,44.00)(0,0)[rr]@xmath67 ( 147.00,44.00)(0,0)[rr]@xmath77 ( 25.00,16.00)(0,0)[rr]@xmath78 ( 77.00,16.00)(0,0)[rr]@xmath79 ( 25.00,5.00)(0,0)[rr]@xmath80 ( 100.00,5.00)(0,0)[rr]@xmath81 ( 54.00,44.00)(0,0)[rr]@xmath82 ( 38.00,44.00)(0,0)[rr]@xmath83 ( 77.00,44.00)(0,0)[rr]@xmath84 to compute the ao of a @xmath1-dimensional point @xmath56 relative to an @xmath86 data matrix @xmath58 , the following projection pursuit procedure is applied .
the point @xmath56 and all data points are projected on a direction @xmath87 , and we compute the univariate ao of the projected point @xmath88 relative to the projected dataset @xmath89 as in .
this is repeated for many directions @xmath87 , and the multivariate ao is defined as the maximum over all corresponding univariate ao s : @xmath90 this definition is based on the principle that a multivariate point is outlying with respect to a dataset if it is outlying in at least one direction of the multivariate space .
note that this outlying direction does not have to correspond to one of the coordinate axes . to compute the multivariate ao we have to rely on approximate algorithms , as it is impossible to consider the projections on _ all _ directions @xmath87 in @xmath1-dimensional space .
@xcite suggest using @xmath91 directions . a possible procedure to generate a direction
is to randomly draw @xmath1 data points , compute the hyperplane passing through them , and then to take the direction orthogonal to it .
this approach guarantees that the multivariate ao is affine invariant .
this means that the ao does not change when we add a constant vector to the data , or multiply the data by a nonsingular @xmath92 matrix .
|
surface , image and video data can be considered as functional data with a bivariate domain . to detect outlying surfaces or images ,
a new method is proposed based on the mean and the variability of the degree of outlyingness at each grid point .
a rule is constructed to flag the outliers in the resulting functional outlier map .
heatmaps of their outlyingness indicate the regions which are most deviating from the regular surfaces . the method is applied to fluorescence excitation - emission spectra after fitting a parafac model , to mri image data which are augmented with their gradients , and to video surveillance data .
* keywords : * adjusted outlyingness , functional data , image data , multiway , robustness .
|
the local group consists of three spiral galaxies and a large number of dwarf galaxies ( @xmath190% of its 40 known members ) .
this proportion of galaxy types may be typical of the local universe as similar distributions are known to exist in nearby groups @xcite and clusters @xcite .
the dwarf galaxies ( irregular , spheroidal or elliptical ) of the lg are of particular interest as their proximity allows us to study them in detail , testing predictions concerning their formation and evolution , as , for instance , the evolution of different kind of dwarfs @xcite or their star formation histories @xcite .
furthermore , their study is relevant because , according to the hierarchical scenario of galaxy formation , dwarf galaxies are the first structures to form and from their merging larger galaxies are built @xcite .
moreover , low - luminosity dwarf galaxies , as those discussed in this paper , are metal poor and are expected to have abundances close to primordial values @xcite . thus the abundances derived there can be useful in extrapolating primordial he / h ( @xcite ; @xcite ) .
it is also interesting to compare the overall star formation history ( sfh ) of the universe predicted by @xcite with the sfhs of galaxies of different morphological type that exist in the lg ( @xcite , @xcite ) .
this is fundamental in order to understand the variation of sfh from the global average in galaxies of differing morphological type ( @xcite , @xcite ) .
a quantitative sfh can be obtained with the colour - magnitude diagram ( cmd ) technique ( cf .
this technique involves three main ingredients : i ) the data , from which a deep observational cmd can be plotted ; ii ) stellar evolution and a bolometric correction libraries providing colors and magnitudes of stars as a function of age , mass , and metallicity ; and iii ) a method to relate the number of stars populating different regions of the observational cmd with the density distribution of stars as a function of age , mass , and metallicity , as predicted by stellar evolution theory ( see @xcite for a complete review ) .
this method can , in principle , constrain the entire star formation and chemical history of a galaxy , but , of course , it depends on the precision of the input models and on the assumptions @xcite .
moreover , since the overall results of this method are model dependent , complementary specific evolutionary phases can be used as _ age - tracers _ to mark different populations and to infer the presence of a given age component @xcite . in this paper
, we present new results from the latest study of individual group members as part of the continuing lg census .
this census is a narrow- and broad - band imaging survey of the galaxies of the lg , visible from the isaac newton group of telescopes ( ing ) , whose aim is to find , catalogue and study old and young emission - line populations , such as planetary nebulae , supernova remnants and regions @xcite .
we are interested in identifying planetary nebulae , which are good indicators of the presence of intermediate - age stars .
lgc observations were obtained with the wide field camera ( wfc ) at the 2.5 m isaac newton telescope ( int ) .
the detector of the wfc is a mosaic composed of four ccds , with 2048@xmath24096 pixels each .
it covers a field of view of @xmath3 .
the pixel scale is [email protected]/lgc for the description of the project ] the results obtained until now in the search for pne in the lg galaxies are described in the following papers : @xcite , @xcite , hereafter m03 ; @xcite ; @xcite ; @xcite . here we describe the results of the observations of the next three targets : the dwarf irregular galaxies ic 1613 ( morphological type ir v according to @xcite , hereafter vdb00 ) , ( ir iv - v ; @xcite ) , and 8 ( dirr ; @xcite ) .
the gas - rich dwarf irregular galaxy ic 1613 was first discovered by @xcite and because of its proximity ( 730 kpc , @xcite ) , its high galactic latitude ( -60.6 ) and consequently small galactic extinction , has provided an excellent opportunity to observe stellar populations in a low metallicity environment . in fact , its oxygen abundance as derived from regions is approximately 12 + log ( o / h)=7.7 ( @xcite ; @xcite ) .
moreover , ic 1613 is a relatively isolated non - interacting irregular galaxy , as proven by the lack of stellar clusters generated by interactions with other galaxies @xcite . from an evolutionary point of view , it is a very primitive galaxy because the stellar mass is comparable to the gas mass @xcite .
@xcite presented b , v , r and i stellar photometry pointing out the presence of stars in a wide range of ages , from young to intermediate and old population .
recent hubble space telescope observations by @xcite found an enhanced star formation rate ( sfr ) from 3 to 6 gry ago .
a survey for emission - line objects was done in this galaxy by @xcite who found one candidate pn . the wolf - lundmark - melotte ( ) galaxy was discovered by @xcite and independently by @xcite and by @xcite .
( called also ddo 221 , @xcite ) is a local group metal - poor dwarf irregular ( 12 + log(o / h ) = 7.7 ; @xcite ) , in a part of the local group relatively free of galaxies , at a distance of 925 kpc ( vdb00 ) .
@xcite studied with the hubble space telescope the sfh in a portion of wlm .
it appears to have begun no more than 12 gyr ago , forming more than half of its stellar population by 9 gyr ago .
the sfr has subsequently gradually decreased until a recent increase in activity starting between 1 and 2.5 gyr ago , but is still continuing to the present time .
recent star formation is indicated by several regions around o - type stars , as observed by @xcite . to date no candidate pn
is known in this galaxy .
in fact @xcite found the two candidate pne proposed by @xcite to be ordinary stars .
the other candidate pn identified by minniti & zijlstra ( 1996;1997 ) was found to be a compact region ( zijlstra , private communication ) .
8 was first discovered by @xcite and it was catalogued as ddo 155 by @xcite .
it is a suspected member of the local group ( vdb00 ) , but observations of a single cepheid by @xcite and of the tip of the rgb by @xcite placed this galaxy at the distance of 2.2 mpc , beyond the usually accepted limit of the lg .
8 is one of the intrinsically smallest irregular galaxies known , with an effective radius of only 50 pc and mass of 10@xmath5 @xcite .
several regions are known in this galaxy @xcite .
their abundances have been measured by @xcite , finding one of the lowest known oxygen abundances ( 12 + log ( o / h)=7.4 ) .
no pn have been identified in 8@xcite .
basic data on the three galaxies are reported in table [ tab_gal ] .
[ cols="<,<,>,<,<,<,<,^ , < " , ] [ tab_pos ]
given the mass range covered by pn progenitors , assumed to be classicaly between 0.8m@xmath6 ( or a bit higher , 1 m@xmath6 , cf .
@xcite ) and 8m@xmath6 , the stellar population from which pne derive is formed at intermediate ages ( roughly 1 to 8 gyr ago ) during the history of a galaxy .
one can evaluate the total mass of such a stellar population using the theoretical relation between the total luminosity of the population of pne progenitors and the pne number , obtained under the hypothesis of coeval , chemically homogeneous stars @xcite .
the total luminosity of the population of pne progenitors , @xmath7 , is related to the number of pne , @xmath8 , to their lifetime , @xmath9 , and to the so - called evolutionary flux , @xmath10 . the number of stars with initial mass @xmath11 per unit luminosity leaving the main sequence each year is then given by the expression @xmath12 thus , counting planetary nebulae and estimating the mean mass of the pne progenitors we have a measure of the total mass of the intermediate - age stellar population , or at least a lower limit to it , when the pne survey is not complete . using a lifetime of @xmath110 000 yrs for the pn phase , a mean specific evolutionary flux of 2@xmath13 yr@xmath14 @xmath15 ( @xcite , @xcite , @xcite ) , a mean progenitor mass of @xmath11.5 m@xmath16 ( computed using the initial mass function by @xcite , corresponding to a luminosity of @xmath15@xmath17 ) , the lower limit to the intermediate - age stellar mass in the lg galaxies can be estimated .
the intermediate - age stellar mass is proportional to the number of pne , and with the approximations described above , we found that about 2@xmath210@xmath18 m@xmath6 were formed for each pn observed .
we used all the _ candidate _ pne discovered in the local group by several authors ( see references below ) with the on - band / off - band or similar techniques , and the criteria described in sect . 4 , considered as _ bona fide _
pne for the following reasons .
the first is that these criteria give a good confidence on the pn nature of the candidate .
for instance , a follow - up spectroscopic study of pn candidates detected in m 33 by @xcite confirmed more than 70% as pne @xcite , even if the criterion of [email protected] was not used . among the remaining 30% objects
, some indetermination remains between low - excitation pne and compact regions , which have similar spectroscopic and photometric features .
the second reason is that it is impossible to study the statistical properties of pn population in the lg if considering only spectroscopically confirmed pne since spectroscopic observations have been obtained for only @xmath110% of the total number of lg pne .
in addition , once obtained , spectroscopy can not provide complete confidence in the determination of the nature of the object , in many cases because of the faintness of the objects @xcite . in figure [ fig - inter ]
, we show the result , plotting the total stellar mass of the galaxies ( values from vdb00 ) vs. the observed number of pne ( left vertical axis ) .
the latter is proportional to the the mass of the intermediate - age stellar population ( right vertical axis ) with a constant of proportionality of 2@xmath210@xmath18 .
errors amount to @xmath130% in the estimate of the total stellar mass of galaxies , and to @xmath150% in the estimate of the intermediate - age mass for the galaxies where we obtained the number of pne within 4 mag from the cutoff of the pnlf ( upper panel ) . for the galaxies where we have only the number of observed pne ( lower panel ) ,
only a lower limit to the intermediate - age mass can be given .
the references on the number of pne in the local group and also on their spectroscopic confirmation , if any , are : m 31 2764 pne identified with the planetary nebulae spectrograph by @xcite , @xmath130 confirmed with spectroscopy by @xcite , @xcite ; mw @xmath12400 ( 1143 + 242 true and probable pne in @xcite , and @xmath11000 in @xcite ) ; m 33 152 pne identified by @xcite ( @xmath130 confirmed by @xcite ) ; lmc @xmath11000 estimated , 350 discovered by @xcite and 700 newly discovered by @xcite ( @xmath1 200 confirmed by @xcite ) ; smc 132 estimated and 101 discovered @xcite ( @xmath1 70 confirmed by @xcite ) ; m32 30 @xcite ( 14 confirmed by @xcite ) ; ngc 205 35 @xcite ( 13 confirmed by @xcite ) ; ic 10 16 @xcite ; ngc 6822 17 @xcite ( 6 confirmed , leisy , private communication ) ; ngc 185 5 @xcite ( 5 confirmed by @xcite ) ; ngc 147 9 @xcite ; sagittarius 3 @xcite ( confirmed ) ; fornax 1 @xcite ( confirmed ) ; pegasus 1 @xcite ; leo a 1 @xcite ( confirmed , leisy , private communication ) ; ngc 3109 18 @xcite ; sextans b 5 @xcite ( confirmed by @xcite ) ; sextans a 1 @xcite ( confirmed by @xcite ) . despite the large errors , a linear trend can be traced ( the continuous line in the upper panel of figure [ fig - inter ] ) , where the total stellar mass of the galaxies is plotted versus the number of pne , for all cases in which we could estimate their population size corrected to a completeness limit of 4 magnitudes below the pnlf cutoff .
this number is extrapolated from the empirical formula of the luminosity function @xcite for the lg galaxies where the completeness limit in the search for pne was known , and where pne above this limit were discovered . from this plot
we can infer which galaxies had relatively strong star formation during the past @xmath118 gyrs , i.e. in the period corresponding to the formation of the pne progenitors .
in particular , we can compare the location of dwarf galaxies in figure [ fig - inter ] with their sfhs reviewed by @xcite .
we find that the galaxies which show little star formation during the past 18 gyrs , i.e. the galaxies with a relative star - formation rate @xmath200.2 during that peroid of time ( cf .
figure 8 of @xcite ) , generally lie in the diagram below the continuous line ( filled squares ) . on the contrary ,
those with strong intermediate - age star formation , namely [email protected] , are located above the least squares fit line ( filled circles).we note that galaxies which had a conspicuous star formation during the past @xmath118 gyrs lie above the continuous line , with the only exception being ngc 6822 .
it has a smaller number of pne than expected for its mass
. it might be due to the determination of its stellar mass which is particularly difficult because of the large extent of its halo @xcite .
another aspect which might be suggested from figure [ fig - inter ] is that all systems associated with m 31 ( m 32 , ngc 205 , ngc 147 , ic 10 , with exception of ngc 185 ) and with the mw ( smc , lmc ) have enhanced recent star formation , while several isolated systems have lower rates ( sextans a , sextans b , leo a , ngc 6822 ) , suggesting that star formation might be enhanced by interaction with the giant galaxies . in the lower panel , the total stellar mass of the galaxies is plotted versus the observed number of candidate pne . again
a linear trend can be traced , but we note that only a lower limit to the intermediate - age mass can be estimated from this plot because the number of pne might be not complete .
the location in this diagram of ic 1613 and wlm , is in agreement with what we know about their star formation histories , showing a higher star formation during intermediate ages in ic 1613 than in wlm .
pne are therefore confirmed to be useful evolutionary age tracers of the intermediate - age population .
the presence of pne is enough evidence for an intermediate - age population @xcite , and , in addition , the relationship of their number to the host galaxy mass gives information on the relative star formation history between different galaxies , where the same completeness in the search of pne has been reached .
these results agree with the those obtained by accurate sfh analysis @xcite .
in this paper we presented the search for pne in three dwarf irregular galaxies belonging to the local group , or close to it : ic 1613 , wlm , and 8 .
we discovered two new candidate pne in ic 1613 , one in wlm , and none in 8 .
their number and their absolute magnitude were analyzed as a function of the sfh of the hosting galaxies .
the number of pne of these galaxies together with the number of pne in the other lg galaxies are used to estimate the mass of intermediate - age population of each galaxy .
this is compared with the sfr of the lg dwarf galaxies from 1 to 8 gry ago , finding that the number of pne agrees with the results obtained by accurate sfh analysis and can therefore be used to constrain synthetic models of sfhs .
we are very grateful to the anonymous referee whose comments resulted in a significant improvement of the work reported in this paper .
the int data have been made publically available through the local group census programme ( pi : n. a. walton ) of the isaac newton group s wide field camera survey programme .
this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration , the apm sky catalogue and usno - a2.0 sky catalogues , and the eso online digitized sky survey .
dm is supported by fondap center for astrophysics 15010003 .
99 ables h.d . , 1971 , _ publ
u. s. naval obs .
_ , second series , 20 , part .
4 acker a. , ochsenbein f. , stenholm b. , tylenda r. , marcout j. , schohn c. , 1992 , 1996 , strasbourg eso catalogue of galactic planetary nebulae , eso aparicio a. , 2001 , a&ss , 276 , 545 aparicio a. , gallart c. , 1994 , in the local group : comparative and global properties , eso conference and workshop proceedings , eds andrew layden , r. chris smith , and jesper storm , p.115 aparicio a. , gallart c. , 2004 , aj , 128 , 1465 buzzoni a. , arnaboldi , m. , 2004 , to appear in the proceeding of eso - workshop `` planetary nebulae beyond the milky way '' , garching , may 19 - 21 , 2004 .
ciardullo r. , jacoby g.h .
, ford h.c . , neill j.d . 1989 , apj , 339 , 53 ciardullo r. , feldmeier j.j . ,
jacoby g.h . , kuzio de naray r. , laychak m.b . , durrell p.r . , 2002 , apj , 577 , 31 ciardullo r. , durrell p. r. , laychak m. b. , herrmann k. a. , moody k. , jacoby g. h. , feldmeier j. j. , 2004 , apj , 614 , 167 corradi r.l.m .
, magrini l. , 2005 , to appear in the proceedings of the eso - workshop `` planetary nebulae beyond the milky way '' , garching , may 19 - 21 , 2004 .
corradi r.l.m . , magrini l. , greimel r. , irwin m. , leisy p. , lennon d.j . , mampaso a. , perinotto m. , pollacco d.l . , walsh j.r
. , walton n.a . , zijlstra a.a . , 2005 ,
a&a in press danziger i.j . ,
webster b.l . , dopita m.a . , hawarden t.g , 1978 , apj , 220 , 458 dohm - palmer r.c . ,
skillman e.d . , gallagher j. , 1998 , aj , 116 , 1227 dolphin a. e. , 2000 , apj , 531 , 804 dolphin a. e. , saha a. , skillman e.d . ,
tolstoy e. , cole a.a .
, dohm - palmer r.c . ,
gallagher j.s . , mateo m. , hoessel j.g . , 2001 , apj , 550 , 554 freedman w. l. , 1988 , aj , 96 , 1248 gallouet , l. , heidmann , n. , dampierre , f. , 1975 a&as , 19 , 1 hodge p. , 1974 , apjs , 27 , 113 hodge p. , miller b.w . , 1995 ,
apj , 451 , 176 hodge p.w . ,
lee m.g . , kennicutt r. c. jr . , 1989 , pasp , 101 , 640 hodge p.w . , smith t.r . , eskridge p.b
, macgillivray h.t . , beard s. m. , 1991 , apj , 369 , 372 irwin m. , lewis j. , 2001 , new ar 45 , issue 1 - 2 , 105 jacoby g. h. , 1989 , apj , 339 , 39 jacoby g. h. , 2005 , to appear in the proceedings of the eso - workshop `` planetary nebulae beyond the milky way '' , garching , may 19 - 21 , 2004 .
jacoby g. h. , ciardullo r. , 1999 , apj , 515 , 169 jacoby g. h. , lesser m. p. , 1981 ,
aj 86 , 185 kunth d. , stlin g. , 2000 , ara&a , 10 , 1 lake g. , skillman e. d. , 1989 , aj , 98 , 1274l lee h. , grebel e.k . , hodge p.w . , 2003 , a&a , 401 , 141 leisy p. , corradi r.l.m . , magrini l. , greimel r. , mampaso a. , dennefeld m. , 2004 , a&a accepted leisy p. & dennefeld m. , 2005 , in preparation lequeux j. , meyssonnier n. , azzopardi m. , 1987 , a&as , 67 , 169 lequeux j. , peimbert m. , rayo j.f .
, serrano a. , torres - peimbert s. , 1979 , a&a , 80 , 155 lundmark k. , 1923 , pasp , 35 , 95 madau p. , pozzetti l. , dickinson , m. , 1998 , aj , 498 , 106 magrini l. , corradi r.l.m .
, mampaso a. , perinotto m. , 2000 , a&a 355 , 713 magrini l. , perinotto m. , corradi r. l. m. , mampaso a. , 2003a , a&a , 400 , 511 magrini l. , corradi r.l.m . , greimel r. , leisy p. , lennon d.j .
, mampaso a. , perinotto m. , pollacco d.l . ,
walsh j.r . ,
walton n.a . ,
zijlstra a.a . , 2003b , a&a , 407 , 51 magrini l. , corradi r.l.m .
, walton n.a . ,
zijlstra a.a . ,
pollacco d.l . ,
walsh j.r . , perinotto m. , lennon d.j . , greimel r. , 2002 , a&a , 386 , 869 magrini l. , corradi r.l.m .
, greimel r. , leisy p. , lennon d.j .
, mampaso a. , perinotto m. , pollacco d.l . ,
walsh j.r . ,
walton n.a .
, zijlstra a.a .
( m03 ) 2003 , a&a , 407 , 51 magrini l. , leisy p. , corradi r.l.m . , perinotto m. , mampaso a. , vlchez j. , 2005 , a&a submitted mateo m.l .
, 1998 , ara&a , 36 , 435 mcgaugh s. s. , bothun g. d. , 1994 , aj , 107 , 530 melotte p.j .
, 1926 , mnras , 86 , 636 mndez r.h .
, kudritzki r.p . , ciardullo r. , jacoby g.h . , 1993 , a&a 275 , 534 merrett h. et al . , 2005 ,
to appear in the proceedings of the eso - workshop `` planetary nebulae beyond the milky way '' , garching , may 19 - 21 , 2004 .
miller b.w . , 1996 ,
aj , 112 , 991 minniti , d. , zijlstra , a.a . , 1996 , apj , 467 , l13 minniti d. , zijlstra a.a . , 1997 , aj , 114 , 147 moles m. , aparicio a. , masegosa j. , 1990 , a&a , 228 , 310 monet , d. , bird , a. , canzian , b. , et al . , 1998 , usno - a2.0 , ( u.s . naval observatory , washington dc ) nagashima m. , yoshii y. , 2004 , apj , 610 , 23 oke j.b .
, 1990 , aj , 99 , 1621 parker q.a . , hartley m. , russeil d. , acker a. , ochsenbein f. , morgan d.h . , beaulieu s. , morris r. , phillipps s. , cohen m. 2003 , iau symp .
209 , s. kwok , m. dopita , & r. sutherland eds , asp / iau , p.41 peimbert m. , torres - peimbert s. , 1974 , apj , 193 , 327 phillips s. , 2001 , mnras , 326 , 1041 phillips s. , parker q. a. , schwartzenberg j. m. , jones j.b . , 1998 ,
apj , 493 , l59 prada , f. , et al . , 2005 , in preparation reaves g. , 1956 , aj , 61 , 69 reid w. a. , parker q. a. , 2005 , to appear in the proceedings of the eso - workshop `` planetary nebulae beyond the milky way '' , garching , may 19 - 21 , 2004 .
renzini a. , 1995 , in stellar populations , ed .
p. c. van der kruit & g. gilmore ( dordrecht : kluwer ) , 325 renzini a. , buzzoni a. , 1986 , in _ spectral evolution of galaxies _ , eds .
chiosi c. and renzini a. , reidel , ap .
space sci . lib .
122 , p.195 richer m. g. , mccall m. l. , 1995 , apj , 445 , 642 richer m. g. , mccall m. l. , 2002 , rmxac , 12 , 173 scalo j.m . , 1998 , in the stellar initial mass function , ed , g. gilmore , & d. howell ( san francisco : asp ) , asp conf .
ser . , 142 , 201 skillman e.d . ,
kennicutt r.c .
, hodge p.w . , 1989 , apj , 347 , 875 skillman e.d . , melnick j. , terlevich r. , moles m. , 1988 , a&a , 196 , 31 skillman e.d . , tolstoy e. , cole a.a . ,
dolphin a.e .
, saha a. , gallagher j.s .
, dohm - palmer r.c . , mateo m. , 2003 , apj , 596 , 253 stasinska g. , richer m. g. , mccall m. l. , 1998 , a&a , 336 , 667 stetson p. b. , 1987 , pasp , 99 , 191 tolstoy e. , saha a. , hoessel j.g .
, danielson g.e . , 1995 , aj , 109 , 579 van den bergh s. , 1959 , publications of the david dunlap observatory , university of toronto , v. 2 , no .
5 , toronto , canada : university of toronto press p. 147
van den bergh s. , 1966 , observatory , 115 , 103 van den bergh s. ( vdb00 ) , 2000a , in _ the galaxies of the local group _ , cambridge university press van den bergh s. , 2000b , pasp , 112 , 932 weldrake d.t.f .
, de blok w.j.g . ,
walter f. , 2003 , mnras , 340 , 12 wolf m. , 1906 , mnras , 67 , 91 wolf m. , 1923 , astron .
nachr . , 217 , 476 zijlstra , a.a . , et al .
, 2005 , to appear in the proceeding of eso - workshop `` planetary nebulae beyond the milky way '' , garching , may 19 - 21 , 2004 .
zijlstra a. a. , pottasch s. r. , 1991 , a&a , 243 , 478
|
in the framework of the local group census ( lgc ) , a survey of the local group ( lg ) galaxies above dec=@xmath0 , which is aimed at surveying their populations with strong emission lines , we have searched for planetary nebulae ( pne ) in the low - metallicity dwarf irregular galaxies ic 1613 , , 8 .
two new candidate pne have been found in ic 1613 , one in wlm and none in 8 .
the observations presented in this paper , together with the previous results from the lgc , represent the first step in the study of the pn population in low - metallicity , dwarf irregular galaxies of the local group .
they will be followed by deep spectroscopy to confirm their nature and to study their physical - chemical properties .
we used the observed number of pne in each lg galaxy to estimate a lower limit to the mass of the intermediate - age population which was compared to the star formation rate ( sfr ) of lg dwarf galaxies .
these results are in agreement with those from accurate star formation history ( sfh ) analysis for these small galaxy systems .
[ firstpage ] planetary nebulae : individual : ic 1613 , wlm galaxies : individual : ic 1613 , wlm , gr 8
|
almost all astrophysical objects have significant non - thermal processes that result in emission across a broad energy band .
even stars exhibit flares which are non - thermal processes .
multi - wavelength observations are therefore key to have complete understanding of a variety of astrophysical objects and astrophysical processes .
x - ray binaries and active galactic nuclei ( agn ) are prime examples of objects with emission over a wide band .
often the most dominant part of the electromagnetic emission is within the optical to hard x - ray band .
moreover , these accretion powered sources show intensity variation over a wide range of timescales , from milliseconds which is the time scale for material to go around a low magnetic field neutron star , to years which is the time scale around supermassive black holes in the most powerful of agns . the x - ray variation can be periodic , quasi - periodic , random with flares , bursts and outbursts , and often a mixture of two or more of these patterns .
the optical and uv emission also show intensity variations , though not quite as sharp and strong as x - rays . while in some sources
one type of emission is known to be reprocessed into another type and therefore , the causal relation between the multiple components are understood to some extent , there are types of sources in which the source of variations , their causes and effects on the other components are not yet understood very well .
simultaneous multi - wavelength observations are therefore key to our understanding of the broadband emission mechanism and interrelation between multiple emission processes in x - ray binaries and agns .
simultaneous multi - wavelength x - ray and uv - optical observations are however very difficult to carry out .
apart from the requirement of coordination between the space - based x - ray observatories and ground - based ( or on a different space telescope like the hst ) optical telescopes , the typical duty cycle for simultaneous multi - wavelength data from such observations is very low , owing to the different nature of good time interval ( gti ) of space and ground based observatories .
simultaneous x - ray and uv - optical observations from the same satellite platform can mitigate many of these problems and yield rich dataset .
astrosat , with its three co - aligned x - ray instruments and two uv - optical telescopes with a total of three bands is designed to bring in this new capability of quality simultaneous multi - wavelength observation capability .
two other space observatories , the xmm - newton@xcite and swift@xcite have limited optical and uv capabilities along with very powerful x - ray telescopes and the new potential of single platform simultaneous multi - wavelength observations have been demonstrated to quite a large extent with these instruments .
in addition to the multi - wavelength aspect , the three co - aligned x - ray instruments will also provide very good capability for broadband x - ray measurements , with particular improvement in capabilities in the hard x - ray band of 15 - 80 kev .
the large area x - ray proportional counter ( laxpc ) instrument of astrosat has an effective area that is in this range several times larger than any other x - ray instrument , which will enable certain unique measurements like quasi periodic oscillations ( qpo ) in the hard x - ray band , pulse phase - resolved study of the cyclotron resonance scattering fearure ( crsf ) in accretion - powered high magnetic field pulsars etc .
the astrosat@xcite is being developed under a large collaboration including several indian organisations and some participation from the united kingdom and canada .
though the mission has been delayed , it is expected to be launched in 2013 . in the following sections we give a brief description of the instruments onboard astrosat and discuss some of the main science topics that will be addressed with the laxpc , and with simultaneous observations with the uvit telescopes , with particular emphasis on x - ray binaries and compact objects .
there are five scientific payloads on astrosat .
four of the instruments , including three in the x - ray band are co - aligned while the fifth instrument is an x - ray sky monitor .
the sky monitor will scan about half of the sky in a given orientation of the spacecraft and as the spacecraft is pointed towards different sources , it will obtain all sky x - ray images to monitor and detect brightness variations in previously known bright x - ray sources and also to detect outbursts of new and known transient x - ray sources .
the uvit consists of a pair of telescopes each with a primary aperture of 38 cm@xcite .
one telescope is equipped with a far - uv camera in the wavelength band of 1300 - 1800 @xmath0 while the other telescope has a beam splitter and two cameras in the near - uv range of 1800 - 3000 @xmath0 and visible range of 3200 - 5300 @xmath0 .
uvit therefore provides three band simultaneous imaging over a field of view of 28 arcmin with an angular resolution of 1.8 arcsec .
all the cameras are equipped with multiple narrow and broadband filters .
uvit is built to provide the highest angular resolution images in broad band uv among all space uv telescopes .
the sxt is an imaging x - ray instrument with a grazing incidence x - ray telescope made with two sets of conical foil mirrors@xcite . at the focal plane
is a ccd camera that has an energy resolution of about 200 - 250 ev at 6.4 kev while the telescope has an angular resolution in the range of 3 - 4 arcmin and field of view of 41 arcmin . as a supporting instrument
, the sxt broadens the energy range of the x - ray instruments of astrosat at the lower end to about 0.35 kev .
the sxt will also be useful to determine positions of new transient x - ray sources detected with the sky monitor to within a fraction of an arcmin depending on the brightness .
the czti is a coded aperture mask instrument made of pixelated cadmium - zinc - telluride ( czt ) detectors of area 1024 @xmath1 in the image plane and a coded mask over it at a height of about 70 cm .
the czti has a field of view of 6x6 square degree in the low energy band and a wider field of view of 17x17 square degree above 100 kev .
it is designed to provide an angular resolution of eight arcmin and an energy resolution of @xmath2 in the energy band of 10 - 150 kev . due to a coded aperture mask design , the sensitivity of the czti depends on the presence of other bright sources in the field of view and it is different in crowded fields .
the laxpc is a set of three xenon gas proportional counter detectors , with a total geometrical area of 10,800 @xmath1 and open area in excess of 6,000 @xmath1@xcite .
it is designed to work in the energy band of 2 - 80 kev with an effective area in excess of 6000 @xmath1 over a large energy range of 3 - 20 kev and in excess of 2,400 @xmath1 upto 80 kev ( figure 1 ) . with such large effective area over a wide energy band
, the laxpc will be the workhorse of astrosat for x - ray timing .
along with the sxt in the lower energy side and the czt in the higher energy side , it will also be very useful for broad band x - ray spectroscopy .
the energy resolution of laxpc is moderate , in the range of 14%-18% in the entire energy band of laxpc .
it is designed to have a timing resolution and absolute timing accuracy of 10 micro seconds , suitable for study of rapid variability in x - ray binaries .
the ssm consists of three one dimensional position sensitive proportional counters with one dimensional coded aperture mask@xcite .
the instantaneous sky coverage of each detector - mask set is in the form of a long strip and the sky coverage is obtained by rotating the detectors around an axis perpendicular to the viewing axis of the four co - aligned instruments of astrosat .
the ssm works in the energy band of 2 - 12 kev and has a sky coverage of about 50% for a given pointing of the spacecraft .
the key features of astrosat are : * x - ray and uv - optical instruments on the same satellite . * high angular resolution , simultaneous three band imaging with the uvit . * wide energy band in the x - rays , suitable for broad band x - ray spectroscopy . * large effective area in the hard x - rays , suitable for hard x - ray timing studies . * all sky monitoring from the same satellite , opportunity for quick follow - up observations . * near equatorial orbit of the satellite resulting in low and stable x - ray background , suitable for accurate estimation of background rate .
other space observatories with multi - wavelength capabilities are xmm - newton@xcite and swift@xcite .
both these observatories have major x - ray instruments : the xmm - newton observatory with high throughput telescopes and multipurpose focal plane instruments including x - ray imagers and reflection grating spectroscopy ( rgs ) and the swift observatory with wide field burst alert telescope ( bat ) and x - ray telescope ( xrt ) for narrow field imaging and spectroscopy . however , the optical and the uv telescopes on these missions have limited sensitivity .
the ultraviolet imaging observatory galex@xcite has sensitivity comparable to the uvit .
however , the uvit , with an angular resolution of 1.8 arcsec compared to about 5 arcsec of galex has advantage in crowded fields and simultaneous three band imaging is another advantage with the uvit . in the x - ray band alone , there are two aspects of astrosat that can be compared with the other space observatories : broad band spectroscopy and fast timing .
the widest spectral coverage is available in the 5th japanese x - ray astronomy mission suzaku@xcite .
the hard x - ray instrument of suzaku , having a small area but very low background is more suitable for spectroscopy of faint objects than timing . in comparison ,
the astrosat - laxpc , with large collective area and relatively higher background is more suitable for timing studies and fast spectroscopy for bright sources .
another wide band x - ray observatory was the dutch - italian mission beppo - sax@xcite and similar comparison is valid with the astrosat - laxpc . in x - ray timing studies ,
the major milestones are exosat@xcite , ginga@xcite and the rxte@xcite .
the rxte , with an effective area of 6,000 @xmath1 and a high time resolution has pioneered fast timing studies .
the astrosat - laxpc , being the only timing instrument in space in its timeframe has very high potential for all types of fast timing studies .
of particular benefit will be the several times larger effective area of laxpc compared to the rxte - pca in the hard x - ray band ( figure 1 ) .
the laxpc instrument will provide some unique scientific prospects for the following reasons : * after the very successful operation of the rxte - pca during 1996 - 2012 , laxpc is the only x - ray timing instrument in its time frame : some of the new outbursts of known transient sources and new transients to be discovered with the ssm or other x - ray monitors will be followed up with the laxpc . *
the laxpc has a wider energy band compared to all previous large area instruments : all types of timing features studied with the previous observatories will now be investigated over wider energy band leading to better / tighter constraints for models . *
the effective area in the hard x - ray band is several times larger than all other hard x - ray detectors : physical processes that dominate in the hard x - ray band will be better probed with the laxpc , especially processes with some temporal characteristics .
* there are other soft and hard x - ray detectors extending the energy band for broad band spectroscopic studies : this will lead to better constraints on the continuum spectral model components .
* a near equatorial orbit will give low and stable background rate : this will allow observations of fainter sources compared to what is possible with an instrument with large and variable background .
* opportunity for simultaneous multi - savelength observations : this is an unique and almost new feature for a space observatory .
many astrophysical objects have complex and multiple physical processes ongoing simultaneously . among the x - ray sources ,
there are very few types of objects which produce x - rays by a single emission mechanism .
the relative importance of the difference processes and their interplay , often manifests in a wide energy band and measuring correlation of different spectral components and constraining their relative strengths is crucial .
the multiple , complex , and interrelated high energy processes are simultanelusly observed very well by broad band spectroscopy in the binary black holes@xcite and accretion powered x - ray pulsars@xcite .
the broadband spectral capability of astrosat will therefore be very suitable for detailed investigations of a wide range of x - ray sources .
the accreting , high magnetic field neutron stars called binary x - ray pulsars show strong hard x - ray emission@xcite , with a power - law spectral shape and a photon index of about 1.0 .
the x - ray spectrum usually has a high energy cutoff at around 20 kev above which the spectrum shows exponential fall with an e - folding energy of about 15 kev .
the hard x - ray spectrum is understood to be produced by inverse compton scattering of soft x - ray photons by energetic electrons in the accretion column / slab on the magnetic poles of the neutron star .
an interesting feature of the x - ray spectrum that has been observed in about 20 such sources is that there are broad absorption features@xcite , which are understood to be due to cyclotron resonance scattering of photons by elctrons in presence of the strong surface magnetic field ( a few times 10@xmath3 gauss ) of the neutron star .
the crsf often shows harmonics and several harmonics have been simultaneously detected in the source 4u 0115 + 63@xcite , which of course requires broad band x - ray detectors .
the crsf is a measure of the strength of the magnetic field of the neutron star ( crsf at 12 kev corresponds to b @xmath410@xmath3 gauss ) and is the only direct method .
moreover , a study of the crsf provides various ways to understand the accretion phenomena onto high magnetic field neutron stars and complex physical processes and geometrical aspects of the accretion column / slab .
the transient x - ray pulsars provide us a way to study the dependence of the processes on the mass accretion rate , and a study of the crsf energy against the x - ray luminosity has shown interesting results : anticorrelation between the crsf energy and x - ray luminosity in some sources and positive correlation between the two in some other sources@xcite .
the pulsars are interesting in another way as the rotation allows us to look at the neutron star and the magnetic field structure from different angles , a novelty in astronomy .
thus pulse phase resolved measurements of the crsf is a very useful way to investigate the accretion process onto the highly magnetized neutron stars .
it has barely been possible to study pulse phase dependence of the crsf in a few of the brighest pulsars@xcite , thanks to the broad band spectral coverage with the suzaku observatory .
the broad band x - ray spectrum including the crsf feature and its harmonic detected in the x - ray pulsar vela x-1 is shown in figure 2 . from the same observation with the suzaku observatory , strong pulse phase dependence of the crsf parameters
have been found including a variation in the ratio of energies of the fundamental and harmonics of the crsf@xcite .
the pulsar cyclotron line studies , both the luminosity dependendence and the pulse phase dependence can be carried out very efficiently with the astrosat - laxpc owing to its large effective area in the hard x - rays for these very bright sources .
a simulation of the crsf , as would be detected in observations of several bright x - ray pulsars are shown in figure 3 .
the qpo phenomena in x - ray binaries has been known for almost three decades@xcite .
the qpos are found in all types of x - ray binaries and they appear in different varities in sources with different types of compact stars ( black holes , low magnetic field neutron stars , high magnetic field neutron stars ) and even in one type of source , there are different distinct types of qpos . the most discussed , and perhaps scientifically most valuable are the kilohertz qpos in low mass x - ray binaries@xcite with low magnetic field neutron stars and the rarely occuring high frequency qpos in black hole x - ray binaries@xcite .
qpos have now also been observed in intermediate mass black hole systems and some agns@xcite .
however , inspite of being very rich in observable features , qpos are one of the least understood aspect of x - ray binaries .
the laxpc provides a new way of investigate the qpos , i.e. in the hard x - ray band .
the qpos show a trend of increasing rms with energy , at least upto 25 kev .
in spite of falling number of photons with energy in all x - ray binaries , the large effective area of the laxpc in hard x - rays , combined with the large rms variation of the source photons in hard x - rays give a high signal to noise ratio for qpo detection at hard x - rays . in figure 4 , a simulation shows the qpo features in a black hole x - ray binary as would be detected with laxpc in different energy bands .
some of the accreting low magnetic field neutron stars show thermonuclear x - ray bursts .
measurement of the luminosity and x - ray spectral evolution during the bursts is a very useful way to invesigate the mass - radius relation of neutron stars , a very important topic in astrophysics .
kilohertz oscillations have also been detected during the thermonuclear bursts , the study of which gives clues to the neutron star equation of state and also to the process of burst ignition and propagation@xcite .
the thermonuclear bursts are also very useful to investigate properties of the x - ray binaries and the physics of reprocessing of x - rays in plasma@xcite . in the low mass x - ray binaries
the companion star is often not dominant in the optical .
the uv and optical emission is partly from energy dissipation in the accretion disk , and partly from reprocessing of x - rays from the compact star .
the thermonuclear x - ray bursts provide a unique time varying input , and simultaneous measurement of the x - ray bursts and the reprocessed burst in uv and optical is a very powerful tool to understand the x - ray reprocessing mechanisms and parameters of the binary
. known orbital parameters of some x - ray binaries , and opportunity to study some sources in different orbital phases are additional advantages . however , the difficulties in carrying out simultaneous multiwavelength observations and relative rarity of the thermonuclear bursts ( a few hours interval on average ) has limited the growth of such studies .
the laxpc is well suited to carry out different types of studies of the lmxbs .
along with measurements of the high frequency periodic and aperiodic timing features , the laxpc will be very useful to tightly constrain the bolometric luminosity , especially near the peak of the bursts when a temperature in excess of 2 kev is reached .
the simultaneous uv - optical data will also be very useful for study of the x - ray reprocessing .
detection of bursts from the x - ray source exo 0748 - 676 with the xmm - newton optical monitor@xcite shown in figure 5 , demonstrates the suitability of astrosat for such observations , but with three bands in the uv - optical . in the last five years or
so a new class of binary x - ray sources , named sfxts , have been of great interest@xcite .
as the name suggests , in these x - ray binaries the companion star is a supergiant star and these sources show fast x - ray transient phenomena , ranging from a few minutes to several hours .
models proposed to explain the sfxt phenomena include gated accretion onto neutron stars with strong magnetic field , accretion from dense clumps in the winds of the companion star etc . during the non - burst periods , the sfxts have an x - ray intensity that is about two orders of magnitude lower than the typical high mass x - ray binaries with supergiant companions like vela x-1 .
the low average luminosity is understood to be due to the wide orbits and corresponding large orbital periods in these binaries . however , some exceptions like igr j16479 - 4514 which has a small orbital period@xcite of only 3.2 days ( figure 6 ) raises doubts about the proposed scenarios . though the compact objects in most of the sfxts are expected to be high magnetic field neutron stars ,
pulsations have been detected in only a few of these sources .
observations with the laxpc will be useful to make careful and sensitive searches for pulsations in more of these objects .
if found , the pulsations will be very useful to measure the binary parameters of the sfxts .
[ [ gravitational - wave - gw - radiation - from - accreting - neutron - stars - in - x - ray - binaries ] ] gravitational wave ( gw ) radiation from accreting neutron stars in x - ray binaries ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ accreting neutron stars are one of the different types of sources of gw emission that could be detected with the ground based gw detectors .
for example , an accretion mound on a rapidly rotating neutron star can be a reasonably strong source and candidate for detection .
however , to search for gw at a few hundred hz from a long data base of a few months or years , it is essential to know the spin and orbital parameters of the neutron star and its changes over the entire period . otherwise ,
the parameters space for search is very large@xcite . naturally , the sources with high accretion rate are likely to be stronger sources of gw but the spin and orbital parameters are least constrained for such sources . the sources with best constrained solutions for spin and orbital parameters , the millisecond accreting pulsars , on the other hand have low long term averge accretion rate .
if laxpc observations with improved sensitivity over rxte - pca leads to measurement of the spin and orbital solution for some of the higher accretion rate lmxbs ( see jain & paul 2011 for example ) , this could be a useful input for search of gw from such sources .
|
astrosat is an astronomy satellite designed for simultaneous multi - wavelength studies in the optical / uv and a broad x - ray energy range . with four x - ray instruments and a pair of uv - optical telescopes
, astrosat will provide unprecendented opportunity for simultaneous multi - wavelength observations , which is of immense value in study of highly variable sources , especially x - ray binaries and active galactic nuclei .
the large area x - ray proportional counters ( laxpc ) of astrosat , which has the largest effective area in the hard x - ray band compared to all previous x - ray missions , will enable high time resolution x - ray measurements in the 2 - 80 kev band with moderate energy resolution . here
we give a brief summary of the payload characteristics of astrosat and discuss some of the main science topics that will be addressed with the laxpc , and with simultaneous observations with the uvit telescopes , with particular emphasis on x - ray binaries and compact objects .
the possibility of aiding gravitational wave experiments is also briefly mentioned .
|
the pair correlation is one of the fundamental nucleon many - body correlations , which governs many aspects of the nuclear structure@xcite .
recently attentions are paid to the properties of the pair correlation in neutron - rich nuclei , which exhibit new features such as the neutron halo , the neutron skin and the presence of weakly bound neutrons .
for example , a spatial two - neutron correlation or the di - neutron correlation , between two weakly - bound neutrons forming the halo in @xmath3li and @xmath4he has been discussed extensively @xcite , and its experimental signatures are reported recently @xcite .
there exists also several theoretical studies demonstrating similar spatial two - neutron correlation in many other nuclear systems .
the spatial two - particle correlation was originally discussed in the closed - shell + two - particle systems @xcite , and recent selfconsistent hartree - fock - bogoliubov mean - field calculations suggest that the correlation is generically enhanced around the nuclear surface of neutron - rich nuclei covering also medium- and heavy - mass regions @xcite . furthermore it is found that neutron matter at low densities may exhibit features of the strong coupling pairing@xcite , characterized by the small size ( @xmath5 fm at minimum ) of the cooper pair comparable with the average inter - neutron distance @xcite , reflecting the large scattering length @xmath6fm of the bare nuclear force .
this may also suggest a possible enhancement of the pair correlation in the surface and exterior regions of neutron - rich nuclei .
the spatial two - neutron correlation and the surface enhancement of the neutron pairing is also discussed in the cases of a slab and a semi - infinite nuclear matter@xcite .
also the pairing interaction induced by the coupling to the phonons is expected to increase the surface enhancement@xcite .
it is important to ask how one can probe the surface enhancement of the pair correlation in neutron - rich nuclei .
we consider here the transfer of a neutron pair since the matrix element of the pair transfer has direct relation to the degree of the pair correlation@xcite .
recently the neutron pair transfer process in light neutron - rich nuclei @xmath3li and @xmath7he has been studied experimentally @xcite . in the theoretical side ,
a possible link between the pair transfer between the @xmath8 ground states and the surface enhancement of the pairing in medium and heavy neutron - rich nuclei was suggested by dobaczewski _
et al._@xcite .
recently the pair transfer feeding the excited @xmath8 states in neutron - rich nuclei has been investigated @xcite from the viewpoint of the pair vibrational mode , which are collective vibrational modes associated with the pair correlation@xcite .
et al._@xcite has argued that the pair transfer can be a possible probe of different models of the pairing interactions with differences in density dependence and surface enhancement . in the present work , we shed light on the neutron pair transfer with the _ quadrupole _
multipolarity@xcite , i.e . , the transitions from the ground state of a nucleus with neutron number @xmath9 feeding excited @xmath10 states , in particular the first @xmath10 state , in the neighboring @xmath11 isotopes .
it is noted that the first @xmath10 state in open shell nuclei is a collective state of the surface vibrational type with a strong influence by the pair correlation@xcite . with this surface character of the mode
, we expect that the pair transfer feeding the @xmath12 state may provide information on the pair correlation possibly enhanced in the surface and exterior regions of neutron - rich nuclei .
the low - lying quadrupole vibration is revealed recently to show new features in some neutron - rich nuclei .
for instance , the ratio between the neutron and proton amplitude becomes significantly different from the liquid drop ratio @xmath13@xcite .
new aspects of the quadrupole vibration may emerge also in the pair transfer transitions feeding the @xmath10 states . in our previous studies on the dipole and octupole responses
, we have shown that soft modes accompanying a large pair transition amplitude in the surface and exterior regions emerge above the threshold energy in nuclei with small neutron separation energy@xcite .
we also found that these soft dipole and octupole modes are sensitive to the surface - enhancement of the pair correlation .
the low - lying quadrupole vibration is different from these soft modes in many respects : it exists systematically from stable nuclei to nuclei far from the stability line , and the excitation energy is located below the separation energy in most of the cases . nevertheless , as we shall show below , the low - lying quadrupole mode also has sensitivity to the surface pair correlation .
the sensitivity is systematically seen in the pair - addition transfer feeding the lowest - lying quadrupole state , and it becomes significant in neutron - rich nuclei beyond the shell closure , i.e. in sn isotopes with @xmath1 .
we shall perform systematic numerical study by adopting the sn isotopes as a representative case since its long isotopic chain enables us to see systematic trends .
in particular we focus on isotopes beyond the @xmath14 shell gap , above which the neutron separation energy suddenly decreases .
our study is based on the skyrme - hartree - fock - bogoliubov mean - field theory in describing the pair correlated ground state , and the quasiparticle random phase approximation ( qrpa)@xcite ( and references in @xcite ) to describe the @xmath10 states and the associated pair transfer modes . in the present work ,
we concentrate on the nuclear structure aspects , and leave the cross section and the reaction mechanism of the neutron pair transfer@xcite for future works .
the formalism is described in section 2 .
to elucidate effects of the surface - enhanced pairing , we compare different models of the effective pairing interaction , which have and do not have the property of the surface enhancement . after presenting
the adopted effective pairing interactions in section 3 , we discuss numerical results in section 4 , and the conclusions are drawn in section 5 .
we adopt the hartree - fock - bogoliubov mean - field model and the quasiparticle random phase approximation ( qrpa ) to describe the pair transfer modes with the quadrupole multipolarity .
the basic framework of the present hfb+qrpa formalism is described in refs.@xcite . here
we recapitulate it briefly in order to explain the procedure to describe the pair transfer modes . in performing the hfb calculation
, we assume the spherical symmetry for the ground state and the associated mean - fields . the skyrme energy density functional ( the energy density functional associated with the skyrme interaction ) is adopted to derive the hartree - fock potential and the position - dependent effective mass@xcite .
the parameter set sly4 is adopted@xcite . as the residual interaction responsible for the pairing correlation we use the density - dependent delta interaction ( dddi ) @xcite @xmath15 where the the pairing interaction strength @xmath16 is a function of the neutron and proton densities @xmath17 ( @xmath18 ) , and
hence it depends on the position @xmath19 .
we solve the hartree - fock - bogoliubov equation in the radial coordinate space to obtain the wave functions of the quasiparticle states .
we assume the box boundary condition at @xmath20 with @xmath21 fm .
we truncate the quasiparticle states by the maximum quasiparticle energy @xmath22 mev .
we also truncate the quasiparticle states by the maximum orbital angular quantum number @xmath23 , which is sufficiently large to obtain the converged hfb solutions of the sn isotopes .
the qrpa calculation is then performed on the basis of the self - consistent hfb mean - fields .
as the residual interaction in the particle - particle channel , we adopt the dddi , eq .
( [ dddi ] ) , the same effective pairing interaction used for the hfb calculation .
concerning the residual interaction in the particle - hole channel , we derive it also from the skyrme functional used for the hfb calculation , but we apply the landau - migdal approximation to the residual interaction@xcite .
namely , we approximate the particle - hole residual interaction by a position - dependent contact interaction @xmath24 where @xmath25 and @xmath26 are the landau - migdal parameters derived from the skyrme functional . the position dependence is taken into account in the landau parameters by the local density approximation to the fermi momentum . as a prescription commonly adopted in the qrpa studies using the landau - migdal approximation@xcite
, we introduce a renormalization factor @xmath27 multiplying @xmath28 , and it is fixed to impose the zero excitation energy for the lowest - energy isoscalar dipole solution corresponding to the spurious center - of - mass motion .
we adopt the linear response formalism for the qrpa calculation@xcite .
given the above residual interactions , we have coupled linear response equations for three kinds of fluctuating densities @xmath29 , @xmath30 and @xmath31 , which correspond respectively to fluctuating parts of the nucleon density @xmath32 and of two kinds of the pair densities @xmath33 and @xmath34 in the frequency domain arise from a response of the system caused by external fields with multipolarity @xmath35 .
it is convenient to introduce operators @xmath36 here @xmath37 and @xmath38 represent respectively addition and removal at position @xmath19 of a pair of like - nucleons whose spins are coupled to the singlet @xmath39 .
the pair densities are written compactly as @xmath40 for a response with the multipolarity @xmath35 and the natural parity @xmath41 , the coupled linear response equation is written in the radial coordinate system as @xmath42 where @xmath43 is the unperturbed response function representing non - interacting two - quasiparticle excitations . @xmath44 and @xmath45 are the residual interactions in the particle - hole and the particle - particle channels , respectively . in the present studies we consider external fields of the pair addition and removal types
@xmath46 in addition to the usual external field of the particle - hole type @xmath47 note that the linear response equation ( [ rpa ] ) formulated in the present work contains @xmath48 representing the pair - addition / removal fields . for the unperturbed response function @xmath49 appearing in the linear response equation ( [ rpa ] )
, we adopt the spectral representation @xmath50 with @xmath51 using the discretized quasiparticle eigenstates of the hfb mean - field hamiltonian . here
@xmath52 is the two - component radial wave function of a quasiparticle state with the quantum number @xmath53 , and @xmath54 and @xmath55 are @xmath56 matrices defined for the three kinds of density @xmath57 and @xmath58 , eqs.([rhofl ] ) and ( [ pairpm ] ) ( see ref.@xcite , for details ) .
@xmath59 is a smearing parameter , for which in the present work we adopt @xmath60 mev corresponding to a convolution of the strength function with the lorentzian function of fwhm=100 kev . in the previous works@xcite we used the formalism which implements the continuum quasiparticle states satisfying the proper boundary conditions connecting to out - going waves .
the continuum qrpa calculation based on this formalism is useful for the states above the particle threshold energy , but , on the other hand , it costs quite severely a large computation time , especially when we choose a small value for the smearing parameter @xmath59 , and it is too demanding for systematic studies . the calculated lowest @xmath10 mode obtained in the present study
are all located below the neutron separation energy , and we use a large box size @xmath21 fm . in the present study , therefore , we adopt the spectrum representation using the discretized quasiparticle states obtained in the hfb calculation . an argumentation on this point will be given later .
the hfb+qrpa model enables us to calculate the transition matrix elements for pair transfer operators . here by the pair transfer operator we mean a generalized one - body operator which either add or remove a pair of singlet neutrons .
we consider the following pair - addition operator with multipolarity @xmath35 , @xmath61 which is essentially the same as the multipole pair operators used in the traditional works@xcite . here and
hereafter we omit the isospin index @xmath62 for simplicity since we discuss only the neutron pair transfer and the neutron pairing in the present paper .
note that @xmath63 is a form factor of the pair - addition operator , and we choose @xmath64 in the present analysis .
this is a simplification , but even this simple choice is useful for discussions on isotopic trends and roles of the surface enhanced pair interactions as we show below .
the strength function for the multipole pair - addition operator @xmath65 is given by @xmath66 where @xmath67 and @xmath68 are the ground and the excited states obtained in the qrpa , and @xmath69 is the qrpa excitation energy . in the linear response formalism of qrpa
, this quantity can be calculated as @xmath70 using the solutions of eq.([rpa ] ) obtained with the external field @xmath71 which corresponds to an external field identified to the pair - addition operator , i.e. , @xmath72 .
we evaluate the strength @xmath73 associated with an individual qrpa solution by integrating the strength function @xmath74 in an energy interval @xmath75 $ ] , where @xmath76 is the peak energy of a peak in @xmath74 , corresponding to a qrpa solution , and @xmath77 is the smearing parameter .
an isolated discrete peak has the fwhm of @xmath78 kev .
similarly , we can consider the multipole pair - removal operator @xmath79 the associated strength function @xmath80 and the strength of the pair - removal operator for the individual qrpa solution @xmath81 the former is calculated as @xmath82 under the external field of the multipole pair - removal operator @xmath83 : @xmath84 when we calculate the strength @xmath85 for the multipole operator of the particle - hole type @xmath86 , we choose the external field as @xmath87 .
we also calculate the transition density associated with the pair - addition @xmath88 and pair - removal transfer @xmath89 .
the transition density associated with the addition / removal of a neutron singlet pair at the position @xmath19 is defined and factorized as @xmath90 here the radial transition densities are expressed using the solution of the linear response equation as @xmath91 where @xmath92 is the excitation energy of the qrpa solution , and @xmath93 is a normalization constant .
note that the conventional transition density of the particle - hole type is given by @xmath94 the normalization factor @xmath93 is fixed so that it is consistent with the strengths @xmath95 or @xmath96 of the qrpa solution under consideration .
the pairing interaction strength @xmath97 for neutrons in @xmath98sn , plotted as a function of the radial coordinate @xmath99 for the dddi - bare , the mix and the volume pairing interactions .
the position of the half density radius is marked with the arrow . ]
the average neutron pairing gap @xmath100 in the sn isotopes .
the one calculated with the dddi - bare ( @xmath101 ) is plotted with the solid line . for @xmath102 ,
results obtained with the dddi - bare but using @xmath103 is also plotted with the thin solid line and the diamond symbols .
the ones calculated with the mix pairing and the volume pairing are plotted with the dotted and the dashed lines , respectively .
the experimental neutron gap estimated using the three - point formula of the odd - even mass difference@xcite and the mass compilation@xcite is also plotted with the circles . ]
in the present paper we intend to disclose influences of the surface enhanced pairing on the low - lying quadrupole vibrational mode . for this purpose ,
we compare three versions of dddi , eq.([rpa ] ) , which have different properties concerning the surface enhancement .
they are 1 ) the density - dependent pairing interaction which has a connection to the bare nuclear force via the scattering length , 2 ) the density - independent delta interaction , and 3 ) the so - called mix pairing interaction@xcite .
\1 ) the first is the one which we adopt mainly in the following analysis , and it is a version of dddi which reproduces properties of the bare nuclear force relevant for the pairing , in particular , the scattering length of the bare force@xcite . among such dddi s , we adopt the parametrization @xmath104 for the interaction of neutrons@xcite . here
the parameter @xmath105 is chosen to reproduce the scattering length @xmath6fm in the @xmath106 channel .
the density - dependence characterized by the power @xmath107 and and the prefactor @xmath108 is adjusted to reproduce the neutron pairing gap in dilute and uniform matter , more specifically the pairing gap obtained with the bare force and the bcs approximation .
we find a parameter set @xmath109 mev @xmath110 , @xmath111@xmath110 , @xmath112 , and @xmath113 for the quasiparticle energy cut - off with @xmath114mev@xcite .
however this parameter set leads to neutron pairing gaps in sn isotopes which are systematically smaller than the empirical values estimated from the odd - even mass difference using the three - point formula@xcite .
we therefore vary @xmath108 and fix to @xmath101 so as to reproduce the empirical gap in @xmath115sn@xcite while keeping the scattering length relevant to the low - density limit .
we call this parameter set dddi - bare since it is refereed to the bare nucleon force , but with a small modification .
\2 ) the second is the simplest density - independent delta interaction : @xmath116 the only relevant parameter @xmath105 is fixed to @xmath117 mev@xmath118 to reproduce the gap in @xmath115sn , as is done in ref.@xcite .
the density - independent delta interaction is often called the volume pairing interaction@xcite .
\3 ) the third is the so - called mix pairing interaction@xcite , which has a linear dependence on the total nucleon density @xmath119 , and given by @xmath120 with @xmath121@xmath110 . here
the parameter @xmath105 is fixed to reproduce the gap in @xmath115sn , leading to @xmath122 mev@xmath118 . the pairing force strength @xmath123
is compared among the three kinds of interaction in fig.[fig : dddi ] , which plots @xmath97 as a function of the radial coordinate in @xmath98sn .
it is seen that the dddi - bare exhibits the strongest radial dependence , i.e. , the strongest density dependence , and that the interaction strength in the region outer than the nuclear surface ( corresponding to the low density limit @xmath124 ) is more than twice larger than that of the volume pairing . the mix pairing exhibits an intermediate radial- ( density- ) dependence .
since these three versions of dddi have very different force strength in the external region we expect to see the difference in physical observables related to the surface region and/or in the exterior of nuclei .
we compare in fig.[fig : gap ] the calculated average neutron pairing gap @xmath125 obtained with the three interactions .
it is seen that they give the average pairing gaps which are almost identical for @xmath126 . in neutron - rich nuclei
@xmath127 beyond the @xmath14 shell closure , a small difference of the order of @xmath128 kev is seen ( except the subshell nucleus @xmath129sn where the volume pairing gives the vanishing gap ) .
it is pointed out in the literature@xcite that the surface enhancement of the pairing interaction influences the average paring gap only to a small extent unless one sees the trend in a very wide range of mass@xcite , and the difference may grow only in very neutron - rich nuclei close to the drip - line@xcite . on the other hand
, it is seen that the dddi - bare gives an overestimate of the average pairing gap for the isotopes with @xmath130 , by about @xmath131 kev compared with the other two interactions . for a fair comparison in the following analysis ,
we modify slightly the value of @xmath108 , i.e. , @xmath103 , only for @xmath132 so that it gives the average pairing gap comparable with those of the volume and the mixed pairing ( cf . the thin solid line in fig.[fig : gap ] ) .
in the first two subsections we will mainly discuss results which we obtain with the effective pairing interaction dddi - bare having the strong increase of the interaction strength in the surface and the external regions .
a comparative analysis using different pairing interactions will be presented in the subsection [ sec : sensitivity ] .
let us first discuss the basic properties of the calculated lowest - lying quadrupole state , on which we focus in the present paper .
figure [ fig : exstrength](a ) and ( b ) shows the excitation energy @xmath133 and @xmath134 of the @xmath0 qrpa solution in the sn isotopes .
the results are similar to those in the full self - consistent qrpa calculation @xcite using the same skyrme interaction sly4 .
it is seen in fig .
[ fig : exstrength](a ) and ( b ) that the calculated @xmath133 and @xmath135 reproduce reasonably well the experimental values . in the present qrpa calculation
the lowest quadrupole mode shows an instability in the @xmath136 and @xmath137 isotopes as is also seen in ref.@xcite .
we do not discuss these cases and the @xmath138 cases where the qrpa excitation energy is less than 0.5 mev since the quadrupole mode then needs to be described as an anharmonic vibration , which is beyond the scope of the present work .
note that there exist a few other qrpa calculations for the neutron - rich sn@xcite .
the transition densities associated with the @xmath0 qrpa solution ( a ) in @xmath98sn and ( b ) in @xmath115sn .
the solid and the dotted curves are the particle - hole transition density @xmath139 for neutrons and protons respectively while the dashed and the dotdashed curves represent the pair - addition transition density @xmath140 and the pair - removal transition density @xmath141 of neutrons , respectively .
the arrow indicates the half density radius @xmath142 .
the effective pairing interaction is the dddi - bare. , title="fig : " ] the transition densities associated with the @xmath0 qrpa solution ( a ) in @xmath98sn and ( b ) in @xmath115sn .
the solid and the dotted curves are the particle - hole transition density @xmath139 for neutrons and protons respectively while the dashed and the dotdashed curves represent the pair - addition transition density @xmath140 and the pair - removal transition density @xmath141 of neutrons , respectively . the arrow indicates the half density radius @xmath142 .
the effective pairing interaction is the dddi - bare. , title="fig : " ] the lowest - lying quadrupole state in isotopes beyond the @xmath14 shell closure are of special interest .
figure [ fig : exstrength](c ) shows the neutron quadrupole strength @xmath143 calculated in the same way as @xmath144 , but for neutrons instead of protons .
the isotopic dependences of @xmath144 and @xmath145 look similar for @xmath146 , but the values of @xmath145 are apparently larger than the proton strength @xmath147 as the neutron number exceeds the magic number @xmath14 .
this behavior is seen clearly in the ratio @xmath148 of the neutron and proton amplitudes , plotted in fig .
[ fig : exstrength](d ) . if the low - lying quadrupole mode is a collective surface oscillation of liquid drop of uniform mixture of neutrons and protons , the ratio @xmath149 would be @xmath13 .
the calculated double ratio @xmath150 is around 1.2 - 1.3 for @xmath151 , being not very different from the liquid drop limit 1 , and it is consistent with the typical values ( @xmath152 ) evaluated experimentally in stable isotopes @xmath153sn@xcite .
the double ratio , however , increases suddenly in @xmath154 , and amounts to 1.6 - 1.7 .
the ratio @xmath149 itself reaches 2.7 - 3.1 . in the isotopes beyond @xmath14 ,
the amplitude @xmath155 of neutrons dominate over that @xmath156 of protons in the first @xmath10 state .
note that the @xmath150 ratio obtained in ref.@xcite is somewhat smaller than our values by about 0.2 .
a characteristic feature of the lowest - lying quadrupole states beyond the @xmath14 shell gap can be demonstrated more explicitly in terms of the transition densities , which are shown in fig.[fig : trans - sn120-sn134 ] for ( a ) @xmath98sn and ( b ) @xmath115sn . in @xmath115sn
, the transition densities exhibit the features typical of the low - lying quadrupole surface vibration .
namely , the particle - hole transition densities for neutrons and protons , and the pair - addition and -removal transition densities of neutrons exhibit similar radial profile all of which have peaks around the nuclear surface @xmath157 fm ( @xmath142 denotes the half - density radius defined by @xmath158 @xmath110 ) . in @xmath98sn , in contrast , the peak positions of the neutron and proton particle - hole transition density @xmath159 deviates significantly with each other : the neutron transition density has amplitude which extends toward the exterior of the nucleus while the proton transition density does not .
the transition density @xmath160 of the neutron pair - addition mode is most characteristic in @xmath98sn .
it has the largest amplitude in the external region @xmath161 ( @xmath162 fm ) among the three kinds of transition densities .
the position where the amplitude of @xmath160 is maximum , @xmath163 fm , deviates by about 1fm from the nuclear surface ( @xmath164 fm ) , and it is even outside of the peak position of the neutron particle - hole transition density .
in addition , it is extended significantly toward the outside of the nucleus , and the tail reaches far to 12 - 13 fm ( in the scale of the figure ) .
the extension of the amplitude in the external region is much more significant than that in the neutron particle - hole transition density @xmath165 .
we note also that the transition density @xmath166 of the neutron pair - removal mode is very small in comparison with the pair - addition transition density @xmath160 , and the dominance of the pair - addition amplitude is commonly seen in all the cases of @xmath167sn .
( a ) the calculated transition strength @xmath168 of the neutron pair - addition transfer feeding the @xmath0 states in sn isotopes .
the diamonds connected with the solid line are the results obtained with dddi - bare while the dotted and the dashed lines ( the circles and the crosses ) are for the mix and the volume pairing interactions , respectively .
( b ) the same but for the strength @xmath169 of the quadrupole pair removal transfer .
, title="fig : " ] ( a ) the calculated transition strength @xmath168 of the neutron pair - addition transfer feeding the @xmath0 states in sn isotopes .
the diamonds connected with the solid line are the results obtained with dddi - bare while the dotted and the dashed lines ( the circles and the crosses ) are for the mix and the volume pairing interactions , respectively .
( b ) the same but for the strength @xmath169 of the quadrupole pair removal transfer .
, title="fig : " ] figure [ fig : pstr - sn120-sn134 ] shows the strength functions @xmath170 and @xmath171 for the quadrupole pair - addition and -removal modes in @xmath98sn and @xmath115sn . the pair - addition strength associated with the first @xmath10 mode in @xmath98sn is significantly larger than that in @xmath115sn , and it reflects the large pair - addition amplitude seen in fig.[fig : trans - sn120-sn134 ] .
the unperturbed strength functions are also plotted for comparison ( the dotted curve ) .
the first peak in the unperturbed calculation of @xmath98sn corresponds to the transition to the two - quasi - neutron configuration @xmath172 ^ 2_{i=2}$ ] .
the pair - addition strength for the @xmath0 state in the full calculation is about ten times larger than that in the unperturbed calculation , thus it indicates a significant enhancement due to a collectivity ( the rpa correlation in the other words ) in the pair - addition mode .
another feature observed in @xmath98sn is that the rpa correlation effect is large not only for the lowest mode but also for higher lying modes : the second peak of @xmath170 also has a large pair - addition strength comparable with the first peak , and there exist several peaks in the excitation energy region @xmath173 mev with significant pair - addition strength . the peaks above the one - neutron separation energy @xmath174 are discretized continuum states .
the continuum qrpa calculation@xcite is appropriate for the excitation energies above @xmath174 .
in fact , the result of the continuum qrpa calculation , shown with the thin dashed curve in fig . [
fig : pstr - sn120-sn134 ] , indicates that the proper treatment of the continuum modifies the strength function for @xmath175 . on the other hand
, it has only a minor effect on the strengths below @xmath174 , and it does not affect the discussions below . note also that the pair - removal strength in @xmath98sn is extremely small in the displayed energy range .
we shall discuss properties of the pair - addition mode in detail .
figure [ fig : pairstrength](a ) is the systematics of the pair - addition strength @xmath168 associated with the @xmath0 mode .
a feature immediately noticed in the figure is a strong variation along the isotopic chain .
@xmath168 takes a rather small value for @xmath151 , but it shows a big jump at @xmath176 , where the strength is larger than those in any other isotopes with @xmath151 . proceeding from @xmath177sn to@xmath98sn , @xmath178sn and @xmath179sn ,
the strength decreases gradually while keeping relatively large values .
it then turns to increase in @xmath129sn . and takes a very large value in @xmath180sn .
figure [ fig : transitiondensity](a ) shows the pair - addition transition densities @xmath160 associated with the lowest quadrupole mode in the isotopes @xmath181sn beyond the @xmath14 shell closure .
it is seen that they all exhibit sizable amplitudes extending to @xmath182 fm as is found in fig.[fig : trans - sn120-sn134 ] in the case of @xmath98sn .
the spatial extension of the pair - addition transition density develops further in @xmath129sn and @xmath180sn , for which we see significant amplitude even around @xmath183 fm , which is about @xmath184 fm apart from the half density surface @xmath185 fm .
this is contrasted with the pair - addition transition density for the isotopes @xmath186sn below the shell closure ( displayed in fig.[fig : transitiondensity](b ) ) . in these nuclei
spatial extension is hardly seen in the pair - addition transition density , and the transition densities is mostly concentrated in the internal and surface region @xmath187 fm @xmath188 fm .
the spatial extension emerges suddenly at @xmath176 . comparing the pair - addition transition density in @xmath189sn ( the largest one in fig.[fig : transitiondensity](b ) ) with that in @xmath178sn ( the third largest one among the solid curves in fig.[fig : transitiondensity](a ) ) , the maximum amplitude is almost the same , but the pair - addition strength @xmath168 in @xmath178sn ( @xmath190 ) is about twice as large as that in @xmath189sn ( @xmath191 ) because of the spatial extension in @xmath178sn .
the sudden change across the @xmath14 shell closure may be ascribed to the fact that the isotopes with @xmath154 have a feature of weakly bound nuclei although they are located far from the neutron drip - line .
the calculated one - neutron separation energy suddenly change from @xmath192 mev in @xmath177sn to @xmath193 mev in @xmath98sn .
this arises from the shell gap at @xmath14 and from the fact that the neutron single - particle orbits @xmath194 and @xmath195 lying above the @xmath14 shell gap are bound only weakly ( the hartree - fock single - particle energies in @xmath177sn are @xmath196 mev and @xmath197 mev ) .
these orbits have significantly small binding energies compared with the orbit @xmath198 just below the shell gap ( @xmath199 mev ) .
since the quadrupole pair - addition mode in @xmath177sn and heavier isotopes necessarily contain two - quasiparticle components consisting of these weakly bound orbits , the pair - addition transition density should inherit the long tail arising from the weak binding .
in particular , approaching @xmath129sn and @xmath180sn , neutrons start to occupy the @xmath195 orbit which has a very small binding energy .
we remark that the closing of shells has a strong influence on the pair transfer in general@xcite .
for instance , the neutron shell is closed at @xmath176 ( @xmath14 ) and the calculated neutron pairing gap vanishes there . in this case , the qrpa equation ( [ rpa ] ) is decomposed into the particle - hole , the particle - particle , and the hole - hole sectors .
the pair - addition mode in @xmath177sn belongs to the particle - particle sector , and the lowest - lying qrpa solution in this sector has a character of a correlated pair of neutrons built upon the @xmath177sn core .
it is nothing but the quadrupole pair vibrational mode in closed shell nuclei , and in this case a large collectivity in the pair - addition mode is known , with a typical example of @xmath200pb@xcite .
this is , however , only one of the mechanisms of the large pair - addition strength @xmath168 seen at @xmath176 ( @xmath14 ) .
we note that the strength in @xmath177sn for the dddi - bare interaction is more than twice larger than that in @xmath201sn ( another shell closure at @xmath202 ) , and this relative enhancement at @xmath177sn is a consequence of the characteristic spatial extension of the transition density in the isotopes beyond the @xmath14 shell closure .
figure [ fig : pairstrength ] ( b ) shows the systematics of the pair - removal strength @xmath169 for transition from the ground state of an isotope @xmath203 to the @xmath0 state in the isotope @xmath204 .
it is seen that the pair - removal strength is generally small , compared with the pair - addition strength .
another trend seen here is that the pair - removal strength is very small at the beginning of shells ( at @xmath205 ) and of a subshell ( at @xmath206 ) ; it takes a value comparable with the pair - addition strength @xmath168 around the mid - shell ( @xmath207 ) , and it is relatively large at @xmath208 because of the shell closures at @xmath202 and @xmath14 in parallel to the situation in @xmath200pb.@xcite figure [ fig : transitiondensity](c ) displays the pair - removal transition densities @xmath166 associated with the @xmath0 mode in the isotopes from @xmath115sn ( located in the middle of the shell ) to @xmath177sn ( at the end of the shell ) .
it is seen that the transition densities @xmath166 of the pair - removal mode in these isotopes does not exhibit the spatial extension toward the outside , rather it is concentrated only around and inside the surface region @xmath209 fm @xmath210 fm .
this feature originates from the fact that the pair - removal transition consists mainly of the obits situated near the fermi energy and those bound more deeply . in the isotopes with @xmath211 , these are strongly bound orbits with the binding energy @xmath212 mev ( e.g , @xmath199 mev in @xmath177sn ) and they do not have tails extending far outside the nuclear surface .
comparison of the transition strength functions @xmath170 for the neutron quadrupole pair - addition transfer , calculated with the volume pairing interaction ( solid curve ) and with the dddi - bare ( dotted curve ) .
( a ) and ( b ) are for @xmath98sn and @xmath115sn , respectively . , title="fig : " ] comparison of the transition strength functions @xmath170 for the neutron quadrupole pair - addition transfer , calculated with the volume pairing interaction ( solid curve ) and with the dddi - bare ( dotted curve ) .
( a ) and ( b ) are for @xmath98sn and @xmath115sn , respectively .
, title="fig : " ] we shall now discuss the role of the residual pairing interaction and the sensitivity of the quadrupole pair transfer to the enhancement of the pairing interaction in the surface and the external regions of a nucleus . in fig.[fig : pdd - sn120-sn134 ] we compare the pair - addition strength function in @xmath98sn and @xmath115sn , calculated with two pairing interactions , the dddi - bare and the volume pairing , which have very different density - dependence in the pairing interaction strength ( cf .
fig.[fig : dddi ] ) . in the case of @xmath98sn
we see a large difference in the quadrupole pair - addition strength in @xmath98sn while in the case of @xmath115sn the difference is small .
a systematic comparison is made in fig.[fig : pairstrength ] for the pair - addition and -removal strengths @xmath168 and @xmath169 of the @xmath0 mode calculated with the three kinds of the pairing interactions , including the mix pairing as well .
it is seen that the pair - addition strength exhibits a sensitivity in all the nuclei to some extent , and the sensitivity becomes prominent in @xmath181sn .
the ratio of @xmath168 obtained with the dddi - bare and that with the volume pairing is 2.3 , 2.8 , 4.0 , 3.5 , 2.8 for @xmath213,136,138 , 140 , 142 , respectively , while for @xmath214 it is in the range 1.4 - 2.0 .
note , on the other hand , that the pair - removal strength @xmath169 is much less sensitive in all the isotopes .
the difference in @xmath169 for the different pairing interactions is negligible in many isotopes .
let us analyze the mechanism of the sensitivity of the pair - addition strength by looking into the transition densities .
we show in fig .
[ fig : transitionint ] ( a)-(c ) the transition density @xmath160 of the pair - addition mode calculated with different pairing interactions in the three representative sn isotopes . in this figure
we see an intimate relation between the degree of sensitivity and the radial profile of the transition density .
namely a large sensitivity observed in @xmath98sn and @xmath180sn can be linked to the pair - addition transition density extending far outside the nuclear surface , @xmath215fm .
the sensitivity is small in @xmath115sn where the transition density is confined mostly within the surface region @xmath209fm . for more detailed discussion let us examine the values of the pair - addition transition density @xmath160 in @xmath98sn at @xmath216 fm , corresponding approximately to the half density radius @xmath217 ( @xmath218 fm ) , and @xmath219 fm , and @xmath220 fm .
the ratio of @xmath160 obtained with the dddi - bare and that with the volume pairing is 1.45 , 1.82 and 2.71 for @xmath216 fm , respectively , indicating an increase of the sensitivity as moving from the surface toward the outside .
it is also useful to compare @xmath160 with the transition density @xmath221 associated with the unperturbed pair - addition transfer ( i.e. the transition associated with the lowest - energy two - quasiparticle excitation obtained by neglecting the residual pairing interaction ) , which is also shown in fig.[fig : transitionint ] with the dotted curve .
the ratio between the transition density @xmath160 obtained with the dddi - bare ( or the volume pairing ) and the unperturbed transition density @xmath221 is a measure of the influence of the rpa correlation ( the configuration mixing ) caused by the residual pairing interaction . in the case of the dddi - bare , the ratio @xmath222 reads 2.7 at the nuclear surface @xmath223 , and it increases to 3.8 - 6.2 at @xmath224 .
this is in contrast to the volume pairing causing much smaller ratio 1.9 - 2.1 - 2.3 at the same positions .
we see even stronger sensitivity to the density - dependence of the pair interaction in @xmath180sn , for which the tail extends further outside .
we thus see that the strong sensitivity of the pair - addition transfer originates from the rpa correlation taking place in the surface and exterior regions in @xmath98sn and @xmath180sn .
we recall the reader here that the three pairing interactions have interaction strengths different by a factor of more than 2 ( cf .
fig.[fig : dddi ] ) at the positions @xmath215 fm corresponding to low densities @xmath225 . a contrasting case is @xmath115sn shown in fig .
[ fig : transitionint](c ) , in which we see very little the rpa correlation effect and the difference between the dddi - bare and the volume pairing interactions .
this is what we can expect from the discussions above as there is essentially no significant spatial extension of the pair - addition transition density in this tightly bound stable isotope , as seen in fig .
[ fig : transitionint](c ) . in fig .
[ fig : transitionint](d ) we show the transition density @xmath166 of the pair - removal mode in @xmath177sn . in this case
the pair - removal amplitude is concentrated only in the vicinity and the inside of the surface @xmath209 fm , exhibiting a very different radial profile from that of the pair - addition mode in @xmath226sn ( fig .
[ fig : transitionint](a ) and ( b ) ) .
it is also seen that the pair - removal transition density @xmath166 does not depend on which we adopt the dddi - bare or the volume pairing .
we can interpret this insensitivity in terms of fig .
[ fig : dddi ] : the effective pairing strengths of the three interactions are comparable in size on average in the region @xmath227 fm where the transition strength is concentrated .
finally we remark that the three pairing interactions give essentially identical results for @xmath133 and @xmath144 except for @xmath228 , as is seen in fig .
[ fig : exstrength ] ( a ) and ( b ) . the excitation energy and the @xmath144 of the @xmath0 state in superfluid nuclei
is known to depend on the pairing correlation@xcite .
however , the results indicate that these quantities are not influenced very much by the density - dependence of the pairing interaction as far as the average pairing gap is constrained to the same value .
we have investigated the neutron pair transfer modes with the quadrupole multipolarity feeding the low - lying quadrupole states in the neutron - rich sn isotopes by means of the qrpa based on the skyrme - hartree - fock - bogoliubov mean - fields . as a model of the surface - enhanced paring
, we used the density - dependent delta interaction ( dddi ) whose strength increases at low nucleon densities reflecting the large scattering length of the @xmath106 bare interaction in the free space .
the numerical calculations indicate that the lowest - lying quadrupole state exhibits a sudden change of its character as the neutron number increases across the magic number @xmath14 , beyond which features of weakly binding emerge in the neutron motion .
the most noticeable feature is found in the pair - addition mode .
namely the transition density of the pair - addition mode feeding the @xmath0 state extends spatially far outside the nuclear surface with a long tail reaching @xmath2 fm in the isotopes @xmath229sn .
accordingly the strength of the quadrupole neutron pair - addition transition is enhanced for @xmath154 .
the enhanced pair - addition transfer is a consequence of the rpa correlation caused by the pairing interaction acting strongly in the surface and external regions of a nucleus .
comparing with other dddi s having different density dependence , i.e. the density - independent volume pairing and the mixed pairing having an intermediate density dependence , it is found that the quadrupole pair - addition mode exhibits generally sensitivity to the increase of the pairing interaction strength in the surface and the external regions of a nucleus .
the sensitivity of the pair - addition mode to the three interactions is about a factor of 1.4 - 2.0 in the isotopes with @xmath151 , and it is magnified to a factor of 2.0 - 4.0 in the isotopes @xmath154 . therefore one can expect that the neutron pair - addition transfer feeding the @xmath0 state in the neighboring @xmath230 isotope , especially those feeding the isotopes @xmath231 , provides a probe to the enhancement of the pairing interaction strength expected in the surface and the external regions . on the other hand , the neutron pair - removal transfer feeding the @xmath0 state in the @xmath204 isotope does not have the sensitivity , and neither @xmath135 nor the quadrupole neutron strength @xmath232 does , as far as the sn isotopes analyzed in the present work are concerned .
one of the authors ( m.m . )
thanks prof .
m. kawai , prof .
m. ichimura and prof .
a. vitturi for useful discussions .
a part of the numerical calculations were performed on the nec sx-8 supercomputer systems at yukawa institute for theoretical physics , kyoto university , and at research center for nuclear physics , osaka university .
the work was supported by the grant - in - aid for scientific research ( nos .
20540259 , 21105507 and 21340073 ) from the japan society for the promotion of science , and also by the jsps core - to - core program , international research network for exotic femto systems(efes ) .
r. a. broglia , o. hansen , and c. riedel , `` two - neutron transfer reactions and the pairing model '' in _ advances in nuclear physics _
vol.6 , ( plenum , new york , 1973 ) ed . by m. baranger and e. vogt , pp.287 - 457 .
|
we investigate the neutron pair transfer modes associated with the low - lying quadrupole states in neutron - rich sn isotopes by means of the quasiparticle random phase approximation based on the skyrme - hartree - fock - bogoliubov mean field model .
the transition strength of the quadrupole pair - addition mode feeding the @xmath0 state is enhanced in the sn isotopes with @xmath1 .
the transition density of the pair - addition mode has a large spatial extension in the exterior of nucleus , reaching far to @xmath2 fm .
the quadrupole pair - addition mode reflects sensitively a possible increase of the effective pairing interaction strength in the surface and exterior regions of neutron - rich nuclei .
|
microwave conductivity measurements have provided a powerful method for investigating the low energy charge dynamics of high temperature cuprate superconductors .
early on , the observation of a linear temperature dependence of the magnetic penetration depth @xmath6 in yba@xmath0cu@xmath1o@xmath7 provided some of the first evidence of @xmath3-wave pairing in these materials.@xcite over a decade later , the @xmath3-wave superconducting phase remains one of the cornerstones of our understanding of the cuprate problem .
surprisingly , the physics of the superconducting state seems to be well described as a bcs instability of a two - dimensional fermi liquid . however , it is essential that this conjecture be thoroughly tested , and one of the best ways to do this is through detailed measurements of charge transport in high quality single crystals . in this work
we report on microwave spectroscopy of two gold - standard cuprate materials ortho - i yba@xmath0cu@xmath1o@xmath8 and ortho - ii yba@xmath0cu@xmath1o@xmath9 and develop a simple phenomenology that describes their low - energy charge dynamics . understanding the role of impurities in the cuprate superconductors
has taken many years of great effort .
the materials have complicated chemistry containing at least four elements and have much greater sensitivity to disorder than conventional superconductors due to finite - angular - momentum cooper pairing .
early measurements on the temperature dependence of the magnetic penetration depth of yba@xmath0cu@xmath1o@xmath2 showed deviations from the linear temperature dependence expected for a clean @xmath3-wave superconductor .
it was later suggested that these effects could be the result of pair - breaking disorder,@xcite a conjecture that was confirmed by showing that the controlled addition of zn impurities to high quality yba@xmath0cu@xmath1o@xmath2 crystals caused a crossover in @xmath10 from @xmath11 to @xmath12 behaviour.@xcite ongoing efforts to uncover the intrinsic physics of the cuprates depend on continued improvements in sample quality .
the samples used in this work are the most well - ordered cuprates available , and are of such exceedingly high purity and crystallinity that even the off - plane dopant oxygen atoms are highly - ordered.@xcite some of the earliest measurements of the low energy electromagnetic response of the cuprates were made by nuss _
et al . _ , who measured the conductivity of yba@xmath0cu@xmath1o@xmath13 thin films using terahertz spectroscopy over the frequency range 0.5 thz to 2.5 thz.@xcite the observation of a peak in @xmath14 below the critical temperature @xmath15 was attributed to a strong decrease in inelastic scattering , rather than the conductivity coherence peak that occurs in @xmath16-wave bcs superconductors . shortly afterward ,
measured a similar peak in @xmath14 in the microwave range.@xcite they proposed that this was the result of a competition between the rapid increase in quasiparticle lifetime @xmath17 and the decaying quasiparticle density @xmath18 as the temperature was reduced .
this analysis was carried out within the context of a drude model characterized by an energy - independent transport relaxation rate @xmath19 .
other transport measurements including thermal conductivity @xcite and higher frequency electromagnetic measurements @xcite have also confirmed the rapid collapse in scattering on entering the superconducting state . in a more recent publication ,
our group presented a series of microwave measurements that confirmed and extended these earlier findings.@xcite this was made possible by two experimental developments : the advent of higher purity yba@xmath0cu@xmath1o@xmath2 crystals grown in bazro@xmath1 crucibles and the development of a sufficient number of fixed - frequency cavity perturbation systems to map out a coarse microwave conductivity spectrum .
five superconducting resonators , operating between 1 ghz and 75 ghz , were used to examine the surface impedance of _ the same single crystal sample _ of fully - doped yba@xmath0cu@xmath1o@xmath8 over the temperature range 2 k to 100 k ( the conductivity inferred from these data is shown in fig . [
fig : figure6 ] of the present article ) . at each temperature
the conductivity spectrum was fit to a simple drude model , which captured the features of the data well , thus supporting the assumptions used in the earlier work .
the temperature dependent quasiparticle scattering rate @xmath19 derived from these fits was seen to decrease rapidly below @xmath15 and saturate at a constant value of 6@xmath20 s@xmath21 below 20 k , corresponding to a spectral width of approximately 9 ghz . assuming a fermi velocity @xmath22cm / s,@xcite the residual scattering rate implies a quasiparticle mean free path of 4 @xmath23 m , a very large distance for a complicated quaternary oxide . however , the observation of a temperature independent transport relaxation rate is at odds with standard models for the electrical conductivity of a @xmath3-wave superconductor@xcite and resolving this elastic scattering problem is a major goal of the work presented in this article .
the strong temperature dependence of the transport relaxation rate at higher temperatures , where the scattering processes are inelastic , has motivated models in which the opening of a gap in the spectrum of electronic excitations in the system leads to a collapse of the inelastic transport relaxation rate.@xcite finally , the same set of fixed frequency resonators was used to probe the conductivity for currents along the crystal @xmath5-axis of ortho - i yba@xmath0cu@xmath1o@xmath8 . in these measurements , currents are parallel to the cuo chains and reveal an anisotropy in charge conduction that can be attributed to the sum of a narrow cuo@xmath0 plane quasiparticle spectrum and a very broad spectral feature arising from the one - dimensional cuo chains.@xcite greater insight into the low temperature charge dynamics of yba@xmath0cu@xmath1o@xmath2 was gained following the development of a novel low temperature microwave spectrometer that could measure the conductivity spectrum as a _ continuous _ function of frequency.@xcite spectroscopy of ortho - ii yba@xmath0cu@xmath1o@xmath9 revealed that the nodal quasiparticle spectrum _ did _ in fact contain most of the features expected for weak - limit impurity scattering in a clean @xmath3-wave superconductor , namely a sharp cusp - like spectrum and a linear - in - temperature spectral width .
in contrast , similar spectroscopy of ortho - i yba@xmath0cu@xmath1o@xmath8 material displayed much more drude - like spectra , except at the very lowest temperatures .
we return to examining these features in the present article , presenting a complete body of microwave data on and ortho - ii yba@xmath0cu@xmath1o@xmath2 , and show how these data can be described by a simple phenomenlogy that captures the physics of energy - dependent charge dynamics .
the paper is organized as follows .
section [ sec : experiment ] begins with a brief discussion of the two main experimental techniques , fixed - frequency microwave cavity perturbation and broadband bolometric surface resistance spectroscopy . in sec .
[ sec : phenomenology ] we use the broadband bolometric surface resistance data to extract detailed conductivity spectra and present a simple phenomenological model of the conductivity . with this in hand , in sec . [
sec : conductivity ] we process data from the five fixed - frequency experiments , which provide a means of capturing the conductivity spectrum at higher temperatures than can be reached with the bolometric experiment . finally , we examine the two key parameters that emerge from fitting to the experimental data : the width of the conductivity spectrum , which measures the thermally - averaged transport relaxation rate ( sec .
[ subsec : spectralwidth ] ) and the total integrated spectral weight or ` quasiparticle oscillator strength ' ( sec . [
subsec : spectralweight ] ) . throughout this article
we use recently revised absolute values of the @xmath24 penetration depths obtained using a newly developed zero - field esr absorption method.@xcite the revised penetration depths are considerably shorter than existing values in the literature and result in substantial quantitative changes to the previously published quasiparticle conductivity and superfluid density . however ,
we emphasize that neither the qualitative form of the conductivity data nor the main conclusions drawn from it are affected .
in this paper we focus on the microwave conductivity , @xmath25 , in the long wavelength limit and in the linear response regime . in experiments that probe the conductivity , the quantity that is inferred from measurements is the microwave surface impedance @xmath26 .
when the applied microwave fields are screened on a length scale that is large compared to the characteristic size of the wavefunctions of both the cooper pairs and the single - particle excitations ( the coherence length @xmath27 and mean free path @xmath28 , respectively ) it is possible to describe the charge response by local electrodynamics , in which case : @xmath29 in very clean cuprate samples the mean free path will exceed the penetration depth and the electrodynamics will not necessarily remain in the local limit .
however , it is possible to take advantage of the two dimensional electronic structure of the cuprates , which gives rise to a short coherence length and mean free path along the crystal @xmath30 direction , to ensure that the charge response _ is _
local.@xcite in the work reported here we have exclusively used such geometries . as pairs condense in a superconductor , the reactive response of the superfluid quickly overwhelms the dissipative response of the quasiparticle excitations , and over most of the temperature range @xmath31 .
the local electrodynamic relation then simplifies to the following approximate forms : @xmath32 from these expressions it is clear that a measurement of the surface resistance @xmath33 contains information about the real part of the conductivity @xmath34 , while a measurement of the surface reactance @xmath35 is a direct probe of the penetration depth @xmath36 . as with nearly all forms of spectroscopy ,
technical limitations require one to work with incomplete spectroscopic information and subsequently to rely on causality to extract the quantities of interest .
our present work is no exception .
however , in this article we present a wide range of measurements that not only allow us to extract the frequency - dependent conductivity , but also provide a set of checks and balances through the oscillator - strength sum rule that let us verify our results in detail . in these measurements ,
a well - defined electromagnetic geometry must be used for three reasons : to separate the individual components of the conductivity tensor , to ensure that demagnetization effects are well controlled , and to enforce local electrodynamics . a particularly clean approach that has been widely used
is to place the sample in a microwave enclosure near a position of high symmetry , where there is an electric node and the microwave magnetic field is quasi - homogeneous .
single crystal samples of many of the cuprate superconductors grow naturally as platelets having a broad @xmath37@xmath38 plane crystal face and thin @xmath39-axis dimension . in this case ,
demagnetization effects are minimized if the broad face of the sample is aligned parallel to the microwave magnetic field . in response to the microwave fields , screening currents flow near the surface of the sample , across the broad @xmath37@xmath38 face , closing along the @xmath39 direction to complete a loop .
all measurements presented in this article employ thin crystals with negligible @xmath30-axis contribution , and were carried out in this low - demagnetizing factor orientation .
an experimental feature common to each of the seven sets of apparatus employed in this work is the use of a sapphire hot finger.@xcite this arrangement allows a sample to be held at an elevated temperature in the microwave fields , while the resonator remains at base temperature . for the field configuration we utilize ,
the sapphire plate is transparent to microwaves and makes a negligible contribution to the dissipation .
the sample is affixed to the sapphire plate ( with a very small amount of silicone grease ) that is then inserted into the microwave field through a hole in the metal end - wall that is small enough to cut off microwave propagation at the operating frequency .
the thermometer and heater are located outside the cut - off hole , in intimate thermal contact with the sapphire isothermal stage , which is in turn weakly linked to the 1.2 k pumped helium bath . resolving the microwave surface impedance of small single - crystal cuprate samples over a wide temperature range
has been achieved by the use of cavity - perturbation techniques.@xcite in these experiments the sample under test is brought into the microwave fields of a high quality - factor ( high-@xmath40 ) resonant structure such as a superconducting cavity or a loop gap resonator . in cavity perturbation ,
shifts in the resonant frequency @xmath41 and quality factor @xmath40 are related to the surface impedance via:@xcite @xmath42 the quantity @xmath43 , where @xmath44 is the quality factor of the resonator when it contains the sample at temperature @xmath11 , and @xmath45 is the quality factor of the empty resonator .
@xmath46 is the shift in frequency of the loaded resonator on warming the sample from base temperature to temperature @xmath11 .
similarly , @xmath47 .
@xmath33 is the absolute surface resistance , and the geometric factor @xmath48 depends on the dimensions of both the cavity and the sample , as well as the geometry of the fields that are being perturbed , and is determined empirically . the direct relationship in eq .
[ eqn : cavitypert ] assumes that the sample is much thicker than the length scale of the electromagnetic screening either the superconducting penetration depth @xmath36 or the normal - state skin depth . in situations when the screening currents penetrate to a depth approaching the sample thickness , eq . [ eqn : cavitypert ] must be modified . for our yba@xmath0cu@xmath1o@xmath2 samples , which have a thickness @xmath49 ,
this becomes important only in a very narrow temperature region just below @xmath15 , and in the normal state where the skin depth is large . to account for this effect
, we use a model that solves the electromagnetic field equations in our one - dimensional geometry with the microwave field applied parallel to the broad face of the sample.@xcite the result is a more complete version of eq .
[ eqn : cavitypert ] that explicitly involves the sample thickness @xmath50 : @xmath51\ ] ] where @xmath52 is the sample volume , @xmath53 is the effective cavity volume , @xmath54 is a constant set by the type of cavity and its field configuration , and @xmath55 is the complex propagation constant for the electromagnetic field in the sample . for this expression , @xmath56 is modified and is now the shift in frequency on inserting the sample into the empty resonator , @xmath57 .
the physical quantities of interest , including @xmath33 , @xmath35 , @xmath36 , @xmath34 and @xmath58 are determined from measurements of @xmath56 and @xmath59 and the relationships @xmath60 and @xmath61 .
while it is possible to calibrate @xmath33 measurements to very good accuracy , cavity - perturbation techniques do not easily yield an absolute value for @xmath62 since a measurement of the absolute penetration depth amounts to a comparison of the total volume of the sample to the volume of the sample minus the small field - penetrated region at the surface of the sample .
it is impossible to measure the dimensions of a single crystal sample precisely enough to achieve an absolute measure of the in - plane @xmath36 this way .
recent gd esr measurements have provided a breakthrough in absolute penetration depth measurements by using the characteristic microwave absorption of gd ions lightly doped through the crystal as a magnetic tracer to measure the field - penetrated volume at the surface of the sample _
directly_.@xcite the calibration procedure for the superconducting resonators relies on measurements of well - characterized pbsn samples of similar dimensions as the yba@xmath0cu@xmath1o@xmath2 crystal , as well as runs without any sample at all .
these measurements allow for the small corrections necessary to account for background losses and the finite volume of the crystal .
each resonator experiment must be calibrated independently .
as a result , there is up to a 10% systematic uncertainty in the overall scale factor at each frequency .
stochastic noise is typically much less than 5% for measurements well below @xmath15 . in cavity perturbation experiments ,
the resonator is generally restricted to operate at a single fixed frequency , requiring the use of many separate sets of experimental apparatus to capture a frequency spectrum .
furthermore , the cavity perturbation method requires that the dissipation of the unknown sample be comparable to the dissipation of the cavity itself in order for the absorption to be measured with high precision .
this is a strong requirement on the cavity when the sample is a high quality superconductor in the @xmath63 limit .
the measurement of the residual absorption in superconductors is challenging at any frequency : in the case of infra - red spectroscopy the problem becomes that of measuring values of reflectance that are very close to unity . in microwave cavity perturbation measurements
there is a similar problem of discerning the small difference in cavity q with the sample in and out .
both techniques have problems with calibration and discerning small differences between two large numbers .
these sensitivity problems are circumvented in a direct absorption measurement . we have recently developed an apparatus capable of obtaining continuous - frequency measurements of the absolute microwave surface resistance in low - loss single crystal samples , which has been described in detail elsewhere.@xcite the instrument employs a bolometric method of detection in which the sample of interest is exposed to a microwave frequency magnetic field @xmath64 whose frequency can be varied , and the corresponding temperature rise is measured .
this method relies on the fact that the power absorption in a microwave magnetic field is directly proportional to the surface resistance @xmath33 : @xmath65 where @xmath64 is the root - mean - square magnitude of the magnetic field at the surface @xmath66 .
measurements of the temperature rise of the sample as a function of frequency directly give @xmath67 .
to enhance rejection of spurious temperature variations , the rf power is amplitude modulated at low frequency and the resulting temperature oscillations of the sample are detected synchronously . to provide a frequency - tunable microwave magnetic field ,
the sample is placed in a custom - made rectangular coaxial transmission line that supports a tem mode in which the @xmath68 fields lie in the transverse plane and form closed loops around the centre conductor . the line is terminated by shorting the center conductor and outer conductor with a flat , metallic endwall , thus enforcing an electric field node at the end of the waveguide .
the sample is positioned with its flat face parallel and very close to the endwall such that it experiences spatially uniform fields over its dimensions .
a second metallic reference sample of known absorption is placed in an electrically equivalent position that serves to calibrate the absolute field strength , which varies considerably across the frequency range due to standing waves in the transmission line .
this technique has provided a very sensitive means of measuring the broadband surface resistance spectrum in superconducting crystals over the range 0.5 ghz to 21 ghz , permitting the first exploration of the detailed low temperature conductivity line shape in a number of compounds.@xcite the conductivity spectrum @xmath69 can be extracted very simply from @xmath67 using eq .
[ eqn : rx ] .
our full analysis is slightly more detailed , allowing for the small frequency dependence of the penetration depth that comes from thermally excited quasiparticles . a very convincing verification of the results from multiple sets of experimental apparatus is provided by a comparison of the broadband @xmath70 data with measurements of the _ same sample _ in five different superconducting resonators .
as seen in fig .
[ fig : figure1 ] , the agreement between the experiments is excellent . in each of the six experiments the absolute value of @xmath33
is determined to better than 10% , with the principle uncertainty coming from the calibration procedure .
comparison of surface resistance measurements made on the same sample of yba@xmath0cu@xmath1o@xmath9 using the broadband experiment ( solid symbols ) with those from five microwave resonators ( open symbols ) .
the agreement between methods is excellent .
the data are plotted as @xmath71 to remove the frequency dependence associated with the superfluid screening . ] the surface resistance @xmath72 for ortho - ii ordered yba@xmath0cu@xmath1o@xmath9 measured with currents in the @xmath37-direction ( top panel ) and @xmath38-direction ( bottom panel ) . ]
the bilayer cuprate compound yba@xmath0cu@xmath1o@xmath2 has the distinct advantage that it has a stoichiometric cation composition with a very low level of cross - substitution , less than one part in @xmath73 .
tunable hole doping of the cuo@xmath0 planes is achieved by varying the oxygen content of cuo chain layers , which act as a reservoir for oxygen dopants . however , the chains add the complication of an extra one - dimensional conduction channel that acts in parallel with the two - dimensional channel associated with the cuo@xmath0 planes .
varying the oxygen content of yba@xmath0cu@xmath1o@xmath2 usually occurs at the expense of introducing off - plane disorder .
fortunately there are several highly - ordered phases for the chain oxygen atoms that are stable at room temperature , and we exploit two of these here .
the first is the fully - oxygenated yba@xmath0cu@xmath1o@xmath8 , with ortho - i order in which _ every _ cuo chain is nearly completely filled and @xmath74 k. optimal doping for yba@xmath0cu@xmath1o@xmath2 occurs at @xmath75 , corresponding to @xmath76 k. increasing the oxygen content beyond 6.93 reduces @xmath15 . since adding oxygen to the cuo chains is a diffusion - limited process , filling a millimetre - sized crystal beyond @xmath77 is impractical . for this publication , we have combined measurements on two different ortho - i crystals to provide a complete data set : sample # 1 has @xmath15=89 k and sample # 2 has @xmath15=91 k. the two samples provided very similar results but with some interesting differences that we attribute to the small difference in oxygen contained in the cuo chains .
the other ordered phase we have studied is ortho - ii , in which the cuo chains _ alternate _ between full and empty .
ortho - ii order is seen to exist over a wide range of doping , from @xmath78 to @xmath79 , however the most complete chain filling has been found to occur at @xmath80.@xcite the particular sample used for the present microwave study is very close to this , having @xmath81 and @xmath82 k. figure [ fig : figure2 ] presents fixed - frequency microwave absorption measurements for a high quality crystal of yba@xmath0cu@xmath1o@xmath9 made in a geometry where the screening currents are along either the @xmath37 or the @xmath38 crystal direction .
the surface resistance is seen to drop rapidly , by up to four orders of magnitude , on cooling below @xmath15 , and displays a strong frequency dependence in the microwave range .
much of the frequency dependence of @xmath33 is due to the factor of @xmath83 in eq .
[ eqn : rx ] that results from screening by the superfluid .
the remaining frequency dependence is due to the quasiparticle conductivity @xmath34 itself .
the raw surface impedance data for fully doped yba@xmath0cu@xmath1o@xmath8 have already been published and are not reproduced here.@xcite
for a superconductor , the conductivity can be expressed as the sum of two conduction mechanisms acting in parallel : one due to charge carriers in the superconducting condensate ( denoted by @xmath84 ) and another due to thermally excited quasiparticles ( denoted by @xmath85 ) : @xmath86 the superfluid conductivity can be parameterized using the london phenomenology:@xcite @xmath87 where the temperature dependent superfluid spectral weight is represented by @xmath88 .
this is an experimentally accessible quantity that is related to the london penetration depth in the following manner : @xmath89 the goal of this work is to explore the phenomenology of the superconducting state without assuming a microscopic model for the conductivity of thermally excited _
d_-wave quasiparticles .
we extract @xmath90 from our measurements of the real and imaginary components of @xmath91 using eq .
[ eqn : zs ] , to obtain both the spectral lineshape and its temperature dependence
. this process would be straightforward if we could measure both @xmath33 and @xmath35 at each frequency .
however , measurements of the frequency dependence of @xmath35 are very difficult to carry out , for reasons that are closely related to the difficulty in measuring _ absolute _ penetration depth at a fixed frequency . to carry out the analysis of our broadband data , we therefore use measurements of @xmath35 at one low frequency only ( 1 ghz ) and extend this to higher frequencies using a self - consistent analysis of the surface resistance data .
we are helped greatly by the fact that the frequency dependence of @xmath36 is very slight , due to the dominant contribution from the superfluid .
the small frequency dependence of @xmath36 is due solely to the weak screening contributions from the thermally excited quasiparticles .
note that we distinguish between the magnetic penetration depth @xmath92 and the london penetration depth @xmath93 that appears in eq .
[ eqn : phenomenology3 ] .
broadband absorption measurements of @xmath33 for ortho - ii ordered yba@xmath0cu@xmath1o@xmath9 with screening currents parallel to the crystal @xmath37-axis ( top panel ) and @xmath38-axis ( bottom panel ) . plotting the data as @xmath94 gives a quick means of estimating the conductivity spectrum from the raw data ( see eq .
[ eqn : rx ] ) . here
we emphasize that the frequency dependence does not follow @xmath95 as it would for a drude conductivity , but rather the power is closer to @xmath96 over most of the range . ] we begin by analyzing the microwave conductivity for currents parallel to the crystal @xmath37-axis .
this is perpendicular to the cuo chains , so these should make a negligible contribution to the @xmath37-axis conductivity .
in addition , we expect that the cuo@xmath0 bilayer should lead to two nearly identical contributions to @xmath97 , since the splitting between bonding and antibonding combinations of the planar wavefunctions is expected to be weak.@xcite we will therefore model the @xmath37-axis conductivity by a _
conductivity spectrum due to _ one _ effective two - dimensional band .
there is currently no accepted phenomenology to describe the detailed quasiparticle conductivity of a cuo@xmath0 plane .
several authors @xcite have assumed a drude - like form : @xmath98 where @xmath99 is the temperature dependent quasiparticle
spectral weight ( often referred to as ` normal fluid ' ) and @xmath100 is the thermally - weighted and energy - averaged quasiparticle transport relaxation rate .
while this particular model has proven useful for interpreting qualitative features of the electrodynamics , there is no _ a priori _ reason why it should model the quasiparticle conductivity correctly .
in fact , broadband measurements demonstrate that the lineshape is distinctly _ non_-drude - like.@xcite a very clear and simple demonstration of this is obtained by plotting @xmath94 at fixed temperature @xmath11 , as shown in figure fig .
[ fig : figure3 ] for yba@xmath0cu@xmath1o@xmath9 . within the low temperature approximation of eq .
[ eqn : rx ] , @xmath94 is proportional to @xmath101 .
the key point is that @xmath94 does not follow a @xmath95 power law at high frequencies , as would a drude spectrum .
the data are much closer to following a @xmath96 power law .
consequently , we have proposed a more realistic phenomenological lineshape to describe the @xmath37-axis quasiparticle conductivity spectrum in yba@xmath0cu@xmath1o@xmath2:@xcite @xmath102^{y(t)}}\ ] ] where @xmath103 represents the zero frequency limit of the quasiparticle conductivity , @xmath104 the inverse spectral width , and @xmath105 the anomalous power .
note that this is equivalent to a drude model if we set @xmath106 . for the spectrum with @xmath107 ,
several conditions apply to the use of eq .
[ eqn : phenomenology1].@xcite the expression applies to positive frequencies only , and @xmath108 must be greater than one or the oscillator - strength sum rule will not be satisfied ( _ i.e. _ @xmath101 will not be integrable ) .
a more subtle point is that the exponent @xmath108 must always eventually reach a high frequency value of 2 since the energy - dependent relaxation rate is cut off by the thermal windowing .
however , eq . [ eqn : phenomenology1 ] provides excellent fits to the spectra presented in this article .
furthermore , eq . [ eqn : phenomenology1 ] is very useful in modeling the quasiparticle contribution to the imaginary part of the conductivity , because it can be extrapolated and integrated _ beyond _ the frequency range of our measurements .
it will be shown below that success in preserving the oscillator strength sum rule in the superconducting state supports the use of this phenomenological form as a means of accounting for the normal fluid contribution to screening .
since the electrical conductivity is a causal response function , the real part of the conductivity determines the imaginary part via a kramers - kronig relation:@xcite @xmath109 where @xmath110 denotes the principal part of the integral . using this in conjunction with the measured superfluid response ( via eq .
[ eqn : london1 ] ) yields a complete expression for the imaginary part of @xmath111 in eq .
[ eqn : zs ] : @xmath112 the strategy is to fit broadband @xmath70 data at a given @xmath11 using the phenomenological model for the quasiparticle conductivity ( eqs . [ eqn : phenomenology1 ] and [ eqn : phenomenology2 ] ) along with measurements of the magnetic penetration depth to parameterize the superfluid conductivity ( eq . [ eqn : london1 ] ) . for the spectrum at each temperature ,
we obtain best - fit parameters for @xmath103 , @xmath104 and @xmath105 .
finally , we generate experimental estimates for the real part of the quasiparticle conductivity by solving eq .
[ eqn : zs ] for @xmath113 and assuming @xmath114 as given by the best - fit parameters . in many respects
this procedure is akin to that used in infrared spectroscopy studies of superconductors , where one needs to extrapolate the measured reflectance to zero frequency in order to utilize kramers - kronig relations.@xcite the difference is that we need a model lineshape to effectively extrapolate our measurements beyond our maximum measurement frequency @xmath115 , but we do so ultimately for the same purpose .
this technique requires sufficient curvature within the experimental bandwidth ( _ i.e. _ @xmath116 ) for a reliable extrapolation .
this approach will be best for low @xmath11 data , where @xmath117 will be governed by elastic scattering of _
quasiparticles from very dilute defects in our high purity samples .
as we turn to modeling the data at higher @xmath11 , inelastic scattering processes ought to push @xmath117 well beyond @xmath115 . fortunately , as the @xmath113
spectrum broadens , the quasiparticle contribution to @xmath118 becomes negligible throughout our bandwidth . in this case , one can set @xmath119 and extract @xmath113 from @xmath70 and @xmath120 without requiring any of the corrections discussed above .
figure [ fig : figure3 ] shows the broadband @xmath38-axis surface resistance data for yba@xmath0cu@xmath1o@xmath9 plotted as @xmath94 versus @xmath41 . while the @xmath37-axis data seem to asymptotically follow @xmath121 at high frequencies , the @xmath38-axis data show a change in @xmath108 beyond 10 ghz . in a previous publication we argued that yba@xmath0cu@xmath1o@xmath2 should have a quasi - one - dimensional band whose low energy excitations in the superconducting state are primarily chain - like , and will therefore only contribute to charge conduction along the @xmath38-axis.@xcite the chain scattering rate is expected to be much larger than that for the cuo@xmath0 planes , and should therefore give rise to a very broad contribution to @xmath122 that can be approximated as a temperature dependent constant over the experimental bandwidth in our analysis .
the change in @xmath108 observed above 10 ghz in the @xmath38-axis spectra , but not in the @xmath37-axis , is due to this very broad component .
in this case the @xmath38-axis microwave conductivity can then be expressed as a sum of contributions from the two - dimensional bands ( a term identical in form to eq .
[ eqn : phenomenology1 ] ) plus a constant attributed to the quasi - one - dimensional band : @xmath123^{y(t ) } } + \sigma_1^{1d } ( \omega\rightarrow 0,t ) .\ ] ] in general , the parameters @xmath124 , @xmath125 and @xmath108 for @xmath38-axis currents will not be required to match those used to fit @xmath37-axis data .
we extract @xmath122 from @xmath33 data using measurements of @xmath93 and the modeled @xmath126 as described in the previous section ( a constant in eq .
[ eqn : phenomenology4 ] will give no contribution to @xmath127 ) .
the procedure will have limitations similar to those for the @xmath37-axis case , but the addition of a fourth parameter @xmath128 renders the analysis less reliable as @xmath117 approaches the experimental bandwidth at higher @xmath11 .
we show the real part of the quasiparticle conductivity spectrum for both in - plane principle directions at representative temperatures for both dopings in figs . [
fig : figure4][fig : figure6 ] , as obtained from our surface impedance measurements using the procedures described above .
the figures are comprised of data from the broadband experiment ( open symbols ) and the microwave resonators ( solid symbols ) .
we include results from two different ortho - i samples , denoted sample # 1(@xmath15=89 k ) and sample # 2 ( @xmath15=91 k ) , because we did not have a complete set of data ( both crystal directions , all sets of apparatus ) for a single sample . as will be shown below , the two samples provided very similar results but with some interesting differences that we attribute to the small difference in oxygen contained in the cuo chains . in figs .
[ fig : figure4][fig : figure6 ] the smooth lines represent the best fit lineshapes obtained using eqs .
[ eqn : phenomenology1 ] and [ eqn : phenomenology4 ] to describe the @xmath37 and @xmath38-axis results , respectively . at each temperature
our model lineshape appears to reliably capture the high frequency behaviour , however our expressions do not accurately follow the data at the very lowest frequencies , particularly at @xmath129 k. note that our model for the two - dimensional conductivity term is necessarily concave downwards at low frequency while the lowest temperature bolometry data clearly suggest a cusp .
nonetheless , we argue that our model lineshape provides a reliable means of extrapolating our measurements to high frequency and provides a reasonable measure of the width and spectral weight of the conductivity peaks .
representative curves of the quasiparticle conductivity @xmath130 for ortho - ii ordered yba@xmath0cu@xmath1o@xmath9measured with currents in the @xmath37-direction ( top panel ) and @xmath38-direction ( bottom panel ) .
the fits shown use the phenomenological model of eq .
[ eqn : phenomenology1 ] .
the low temperature cusp - like shape and high frequency power law ( @xmath131 with @xmath132 ) are features expected for a @xmath3-wave superconductor in the presence of weak - limit impurity scattering , and the spectrum broadens quickly as other scattering mechanisms are introduced at higher temperatures . ]
representative curves of the quasiparticle conductivity @xmath130 for sample # 1 of ortho - i ordered yba@xmath0cu@xmath1o@xmath8 measured with currents in the @xmath37-direction ( top panel ) and @xmath38-direction ( bottom panel ) .
the solid lines are fits to the data with eq .
[ eqn : phenomenology1 ] and eq .
[ eqn : phenomenology4 ] for the @xmath37 and @xmath38-axis respectively . at the lowest temperature
the cusp - like spectrum is evident , however a more drude - like spectrum is recovered at higher temperatures , consistent with our earlier findings for this ortho - i ordered yba@xmath0cu@xmath1o@xmath8 . ]
the low temperature quasiparticle conductivity @xmath130 for sample # 2 of ortho - i ordered yba@xmath0cu@xmath1o@xmath8measured using the bolometric method with currents in the @xmath37-direction ( top panel ) and @xmath38-direction ( bottom panel ) .
the spectrum is very similar to that of sample # 1 , although some differences in the fit parameters will be discussed . ]
a key point made in the preceding section is that the quasiparticle conductivity spectra we have measured are distinctly more cusp - like than the lorentzian line shape of a drude model . in fig .
[ fig : figure8 ] we present the best - fit values of @xmath105 that were used in eq .
[ eqn : phenomenology1 ] to describe the conductivity data extracted from the bolometry data .
error bars indicate the standard deviation obtained from fitting to @xmath70 as a function of @xmath133 at each given temperature .
two key features are important to observe : first , @xmath105 appears to be crudely @xmath134 isotropic for each doping .
second , @xmath105 displays a relatively weak temperature dependence characterized by a slow increase with increasing temperature , and appears to saturate at approximately @xmath135 and @xmath136 for the overdoped and underdoped samples , respectively .
the exponent @xmath108 of eqs .
[ eqn : phenomenology1 ] and [ eqn : phenomenology4 ] obtained from fitting to broadband measurements of the absorption spectrum for yba@xmath0cu@xmath1o@xmath8 and yba@xmath0cu@xmath1o@xmath9 samples .
the exponent @xmath108 describes the high frequency decay of the quasiparticle conductivity curves shown in figs . [
fig : figure4][fig : figure6 ] .
the dashed lines indicate the asymptotic values of 1.45 ( ortho - ii ) and 1.75 ( sample # 1 , ortho - i ) subsequently used for fitting to the higher temperature spectra obtained with the fixed - frequency resonators .
sample # 2 of the ortho - i material shows larger values of @xmath108 , corresponding to broader @xmath101 curves , which we attribute to a slightly increased level of oxygen disorder in that sample . despite this ,
the value of @xmath108 is still inconsistent with the drude value of @xmath106 . ]
the temperature evolution of the quasiparticle conductivity @xmath14 for sample # 1 of overdoped yba@xmath0cu@xmath1o@xmath8measured with currents in the @xmath37-direction ( top panel ) and @xmath38-direction ( bottom panel ) .
this data has been processed using the new absolute values of @xmath137 given in table [ table : lambda ] resulting in a modification of the absolute magnitude of the curves from their previously published@xcite form . ]
the temperature evolution of the quasiparticle conductivity @xmath14 for ortho - ii ordered yba@xmath0cu@xmath1o@xmath9 measured with currents in the @xmath37-direction ( top panel ) and @xmath38-direction ( bottom panel ) . ]
we now turn to the problem of extracting the real part of the quasiparticle conductivity from our cavity perturbation experiments . in this case
we only have measurements at discrete frequencies , from independently calibrated experiments . to help constrain the fits when using such a sparse set of frequencies , we choose to fix @xmath108 at the asymptotic high - temperature values of fig .
[ fig : figure8 ] before proceeding to extract the real part of the microwave conductivity from @xmath138 .
we note that changing @xmath108 ultimately leads to negligible changes in the extracted values of @xmath130 within our experimental bandwidth .
however , we will show below that the choice of @xmath108 has implications for the integrated quasiparticle spectral weight and that the values obtained are consistent with the oscillator strength sum rule.@xcite representative spectra from the resonator measurements with fits are included in figs .
[ fig : figure4][fig : figure6 ] .
we show @xmath130 extracted from all of the cavity perturbation measurements in figs .
[ fig : figure8 ] and [ fig : figure9 ] .
note in each plot there is a sharp peak in the curves at @xmath15 due to superconducting fluctuations,@xcite in addition to the large , broad peak which shifts to higher temperatures as the measurement frequency is increased .
these results can be interpreted in the same qualitative manner as in previous work.@xcite neglecting the sharp peaks at @xmath15 , @xmath130 increases smoothly on cooling from the normal state into the superconducting state .
the large increase in @xmath130 below @xmath15 can be attributed to a rapid decline in the quasiparticle transport relaxation rate below @xmath15 , combined with an increase in the condensate fraction.@xcite the frequency dependence of @xmath130 becomes apparent at our measurement frequencies for temperatures below 55 k for the ortho - i sample and 24 k for the ortho - ii sample .
this is in accord with the idea that the peak in the temperature dependence comes from the development of long quasiparticle transport lifetimes .
.experimental values of the anisotropic magnetic penetration depth measured by a zero field esr technique.@xcite the absolute value of @xmath36 sets the absolute value of the conductivity . [ cols="^,^,^,^,^,^,^",options="header " , ]
the similarities and differences between the data shown for yba@xmath0cu@xmath1o@xmath9and yba@xmath0cu@xmath1o@xmath8 offer insight into the nature of the in - plane anisotropy in this system , the doping dependence of the inelastic scattering , and the elastic scattering of nodal quasiparticles by defects .
both materials show a very broad component in @xmath69 in the direction parallel to the cuo chains , which is presumably due to the quasi - one - dimensional fermi surface sheet derived from a cuo chain band hybridized with the cuo@xmath0 plane bands .
atkinson showed that such a mechanism is needed to explain the anisotropy of the in - plane superfluid density.@xcite he also needed to invoke strong pair - breaking by o defects to explain the lack of a low temperature upturn in the @xmath38-axis superfluid density , an upturn expected if the superconductivity on the chains arises from the proximity effect .
this strong scattering would likely explain the broad @xmath69 spectrum of quasiparticle excitations on this fermi surface sheet .
an important and puzzling issue arises by examining the anisotropy of the transport relaxation rate and temperature dependent spectral weight . in the cuprates ,
the cuo@xmath0 plane currents are carried by nodal quasiparticles .
since the gap nodes lie close to the diagonals of the brilloiun zone , these excitations carry both @xmath37 and @xmath38-axis currents .
we might therefore expect the transport relaxation rate , as measured by the width of the @xmath139 spectra , to be @xmath37@xmath38 plane isotropic .
this is indeed the case in the ortho - i sample : the conductivity spectra have approximately the same width in the @xmath37 and @xmath38 directions , and the temperature - dependent spectral weights agree to within experimental uncertainty .
the latter conclusion is sensitive to the choice of @xmath140 , but we conclude that ortho - i yba@xmath0cu@xmath1o@xmath8 is well described by a 3-band model consisting of a cuo chain - like band acting in parallel with the bonding / antibonding cuo@xmath0 plane bands . however , the ortho - ii results present a surprising contrast .
the widths of the conductivity spectra and the temperature - dependent quasiparticle spectral weights differ by a factor of two between the @xmath37 and @xmath38 directions .
one of the possible explanations for this anisotropy may be the effect of ortho - ii ordering on the electronic structure of the material .
ortho - ii ordering doubles the unit cell of yba@xmath0cu@xmath1o@xmath2 in the @xmath37-axis direction , and can therefore substantially alter the fermi surface topology with respect to the fully - doped compound .
in particular , bascones _
et al . _ have predicted that ortho - ii ordering can generate two new quasi - one - dimensional bands that are derived entirely from the cuo@xmath0 plane state.@xcite this might generate additional nodal quasiparticles derived from the new quasi - one - dimensional bands , increasing the @xmath134 anisotropy of the in - plane conductivity .
we leave this issue as an open question for future work in this area , and as a feature that must be kept in mind for any theoretical treatment of this phase .
figure [ fig : figure10 ] demonstrates that the rapid collapse of the inelastic scattering below @xmath15 occurs in both of these highly - ordered phases of yba@xmath0cu@xmath1o@xmath2 .
one theoretical proposal treats the opening of a gap in the spectrum of excitations that scatters the nodal quasiparticles , including the key detail that umklapp processes are needed to relax charge currents.@xcite the @xmath11 dependence of the damping @xmath141 is consistent with exponentially - activated behaviour , the expectation for umklapp scattering .
the activated behaviour arises because there is a minimum threshold quasiparticle wave vector that satisfies the umklapp requirement .
the new results for the ortho - ii phase yba@xmath0cu@xmath1o@xmath9 are qualitatively consistent with this picture .
the damping rises even more quickly , broadening @xmath101 out of the microwave range by 20 k. this could result not only from the lower @xmath15 , but also because changes in the fermi surface at this lower doping would enhance umklapp processes by moving the required excitation closer to the nodes as compared to yba@xmath0cu@xmath1o@xmath8 .
unfortunately , a quantitative fit is difficult because of the strong linear term in the elastic scattering in this sample .
quasiparticle scattering in yba@xmath0cu@xmath1o@xmath8 and yba@xmath0cu@xmath1o@xmath9 samples is likely dominated by elastic scattering below about 20 and 10 k , respectively . in both cases ,
the lowest temperature spectrum at 1.3 k has the cusp - like shape expected for weak scattering.@xcite however , the yba@xmath0cu@xmath1o@xmath9 spectra retain the weak - scattering cusp - like form up to higher temperatures where inelastic scattering takes over , but the yba@xmath0cu@xmath1o@xmath8 spectra take on a more drude like shape and do not show the temperature independent @xmath142 expected for weak limit scattering .
one possible explanation is that the scatterers are neither in the born nor the unitary limit , something also suggested by recent thermal conductivity measurements.@xcite in such a case , the shape and temperature evolution depend on the strength of the scattering , the density of scatterers and the size of the superconducting energy gap .
detailed fits to the spectra will be the subject of future work , but one can qualitatively state that the scattering in yba@xmath0cu@xmath1o@xmath8 appears to be slightly farther from the born limit than it is in yba@xmath0cu@xmath1o@xmath9 .
it has also been suggested that the nearly linear temperature dependence of @xmath143 for any given fixed frequency @xmath144 at low t is also a consequence of intermediate scattering phase shifts.@xcite in attempting to generate a microscopic description of conductivity data many researchers have focused on models of quasiparticle scattering from point - like defects .
this is a logical starting point , but it has become apparent that such models are incapable of simultaneously accounting for the observed spectral lineshapes and the significant amount of oscillator strength that remains uncondensed at 1.2 k , more than would be expected as the conductivity evolves towards the @xmath145 limit .
et al . _ tried to invoke intermediate scattering phase shifts for point scatterers to explain experimental results , but did not present a quantitative set of parameters that could simultaneously describe the spectral shape , its temperature dependence , and the uncondensed oscillator strength at 1.2 k.@xcite this problem is widespread and far more severe in other cuprates . in @xmath146 for instance there is much greater uncondensed oscillator strength observed in both terahertz measurements of films @xcite and microwave measurements of single crystals.@xcite in these cases , residual oscillator strength as high as 30 % of the superfluid density was observed , compared to the few percent seen in the yba@xmath0cu@xmath1o@xmath2 samples discussed here .
motivated by the spectroscopic inhomogeneity observed in scanning tunneling spectroscopy,@xcite orenstein @xcite has suggested that this oscillator strength was due to a collective mode in the presence of inhomogeneity .
more recently , nunner _
_ have been driven by tunneling results to suggest that randomly distributed interstitial o dopants may be the source of the electronic inhomogeneity in @xmath146.@xcite this has suggested that rather than focussing on point scatterers , a better approach is to study extended scatterers , which give rise to slowly varying potentials on the cuo@xmath0 planes created by off - plane disorder .
such models start from knowledge of the actual defects in the system and have had considerable success in fitting both the @xmath146 microwave data and the yba@xmath0cu@xmath1o@xmath8 data presented here .
the basic suggestion is that off - plane disorder is a controlling feature of all cuprates and that even though the relative amount disorder is much smaller in the oxygen - ordered phases of yba@xmath0cu@xmath1o@xmath2 , it still plays an important role in understanding its electronic properties .
the new spectroscopic data presented here , together with that on other materials@xcite provides a testing ground for the different models : collective modes in an inhomogeneous superconductor , point scatterers with intermediate phase shifts , and scattering by off - plane disorder .
|
extensive measurements of the microwave conductivity of highly pure and oxygen - ordered yba@xmath0cu@xmath1o@xmath2 single crystals have been performed as a means of exploring the intrinsic charge dynamics of a @xmath3-wave superconductor .
broadband and fixed - frequency microwave apparatus together provide a very clear picture of the electrodynamics of the superconducting condensate and its thermally excited nodal quasiparticles .
the measurements reveal the existence of very long - lived excitations deep in the superconducting state , as evidenced by sharp cusp - like conductivity spectra with widths that fall well within our experimental bandwidth .
we present a phenomenological model of the microwave conductivity that captures the physics of energy - dependent quasiparticle dynamics in a @xmath3-wave superconductor which , in turn , allows us to examine the scattering rate and oscillator strength of the thermally excited quasiparticles as functions of temperature .
our results are in close agreement with the ferrell - glover - tinkham sum rule , giving confidence in both our experiments and the phenomenological model .
separate experiments for currents along the @xmath4 and @xmath5 directions of detwinned crystals allow us to isolate the role of the cuo chain layers in yba@xmath0cu@xmath1o@xmath2 , and a model is presented that incorporates both one - dimensional conduction from the chain electrons and two - dimensional transport associated with the cuo@xmath0 plane layers .
|
the spectroscopy is a serious task for any model of hadron structure .
it can yield a detailed information on the source of hyperfine interaction , as well as on effective degrees of freedom in the description of hadron dynamics .
the so - called `` three - body spin - orbit puzzle '' has a long history . in the original papers of isgur and karl @xcite
it was shown that the whole su(3 ) baryon spectrum can be reproduced qualitatively and ( with some restrictions ) quantitatively in a constituent quark model based on a non - relativistic hamiltonian with a central confining potential plus effective color - magnetic one - gluon exchange ( oge ) forces .
the hyperfine interaction was assumed to be a sum of the spin - spin , tensor and spin - orbit interaction potentials between the constituent quarks . however , the spin - orbit term yields a very large ( of order 500 mev ) matrix elements in the three - body system ( typical baryons ) , while in mesons and meson - like baryons @xcite the matrix elements are small as a result of the strong cancelation between the dynamical one - gluon exchange spin - orbit forces and a pure kinematical spin - orbit forces due - to the thomas precession of the quark spin in the central confining potential .
contrary , in ordinary baryons ( like @xmath3 and @xmath4 ) , there are not enough cancelation effects between the two origin spin - orbit forces . showing these possibilities and assuming that in baryons like in mesons must be a consistent cancelation , isgur and karl , as a first approximation , excluded all spin - orbit forces from the hyperfine interaction potential .
since the beginning the problem was called a `` three body spin - orbit puzzle '' in baryons and up to now it was not resolved in a consistent way , although in the relativized version of the constituent quark models ( see @xcite ) some reduction of the size of the spin - orbit interactions to acceptable levels was obtained .
there are other strong candidates for the hyperfine interaction in baryons such as goldstone - boson exchange ( gbe ) forces @xcite and the instanton induced exchange forces ( iie ) @xcite . the gbe based constituent quark model was proposed as a realization of the spontaneous breaking of the chiral symmetry of qcd , which plays a crucial role for the hadron structure .
the debates between the oge and gbe based quark models over the last decade @xcite were focused on the `` three - body spin - orbit puzzle '' , as well as the level ordering problem and the mixing angles in excited baryon sector .
while reproducing the correct level ordering ( for example @xmath5 - @xmath6 , @xmath7 ) , the gbe based model does not contain spin - orbit terms .
the main issue of these debates likely supports the idea that the origin of hyperfine interaction between valence quarks is a sum of the gbe and oge forces ( and possible two - pion exchange @xcite , and additionally iie forces for the nucleon sector ) in consistence with the theory of the low - energy nn - interaction @xcite .
the question is , however , how much the each exchange mechanism does contribute to the baryon spectrum ?
the solving of this complex problem requires many efforts to be done both theoretically and experimentally . on the other hand , from the beginning of the proton spin crisis
@xcite it became clear that the non - relativistic constituent quark models are not adequate and complete picture for the hadron structure , since only about 30 percent of the proton spin is described by the valence quarks !
the possible contribution comes from the coupling to the lower component of the dirac spinor , one - gluon exchange and pion cloud effects .
the problem can be resolved in the frame of the cloudy bag model ( cbm ) @xcite , as was claimed in the recent papers @xcite . a relativistic picture of the hadron structure , while playing a decisive role for the understanding of the proton spin crisis ,
also should yield a consistent description of the hadron spectrum including excited baryons .
the questions , concerning the source of the hyperfine splitting in baryons , the relativistic structure of the one - meson and one - gluon exchange mechanisms , the role of the lower component of the dirac spinor of valence quarks and the contribution of the sea - quarks to baryon spectrum are still being open ! the use of the bag models in baryon spectroscopy was restricted to the octet and decuplet baryons @xcite while having some specific problems concerning the convergence of the self - energy and the sharp surface pion - quark coupling .
the recent progress in the frame of the chiral quark - meson coupling model ( cqmcm)@xcite was done for the description of the n and @xmath0 spectrum and the nuclear matter with inclusion both gluon and meson one - loop diagrams while restricting the intermediate baryon states to the lowest mode when considering the self - energy diagrams . the aim of present paper is to show that a description of the baryon structure in a field - theoretical based chiral quark model @xcite can open many windows to the long - standing problems of excited baryon spectroscopy . in the last work we have calculated the lower excitation spectrum of the su(2 ) baryons on the basis of one pion - loop diagrams .
further we will show that relativistic one - meson and color - magnetic one - gluon exchange forces between valence quarks generate spin 0 , spin 1 , spin 2 etc .
operators , which can be considered as analogy of the non - relativistic spin - spin , spin - orbit , tensor etc .
operators , respectively .
however , all these relativistic operators couple the upper and the lower dirac components of the two interacting valence quarks , hence have a strongly different nature .
for instance , the two valence quarks in s - waves , interacting via one - gluon or one pion exchange forces are coupled only by the spin 1 operators ( the analogy of the spin - orbit forces ) , while in the non - relativistic models the corresponding terms have a spin - spin structure , i.e. are spin 0 operators .
this finding has an important consequence for the proton spin problem , since the possible contribution to the proton spin comes not only from one - gluon exchange , but also from one - meson exchange mechanism .
additionally , in recent paper @xcite meson - cloud effects were shown to be important for the description of the flavor asymmetry of the nucleon sea .
we will show in a phenomenological way that a reasonable description of the lower excitation spectrum of the @xmath3 and @xmath8 can be obtained in a relativistic chiral quark model , including all second - order corrections induced by pion fields ( corresponding to the self energy and exchange diagrams ) to the zeroth - order quark - core results with further possible improvement by including the two - pion exchange and iie ( for the nucleon sector ) forces .
it will be shown also that the contribution of the color - electric ( coulomb ) component of the gluon fields to the baryon spectrum is enormously large at reasonable values of the strong coupling constant . to have an acceptable contribution of the gluon one - loop diagrams one needs a small value of the strong coupling constant which results the gluon - exchange contribution to the nucleon-@xmath0 mass splitting to be negligible .
this result confirm the prediction of the non - relativistic constituent quark model @xcite , that the one - gluon exchange mechanism does not play an important role in the description of the su(2 ) baryon spectrum .
we will show also that the restriction of the intermediate baryon states to the lowest mode when calculating the contributions of the self - energy diagrams is not a good approximation and in order to have convergent results , one has to include all intermediate excited quark and antiquark states with the total angular momentum up to j=25/2 .
the relativistic quark model is based on an effective chiral lagrangian describing quarks as relativistic fermions moving in a confining static potential .
the modification of the model , the so - called perturbative chiral quark model recently was applied to the ground- state baryon masses @xcite .
the potential is described by a lorentz scalar and the time component of a vector potential , where the latter term is responsible for short - range fluctuations of the gluon field configurations @xcite .
the model potential defines unperturbed wave functions of the quarks which are subsequently used in the calculations of baryon properties .
the baryons are considered as bound states of valence quarks , surrounded by a pion cloud as required by the chiral symmetry and by gluons .
interaction of quarks with a pion is introduced on the basis of the linearized @xmath9-model @xcite .
the residual color - magnetic quark - gluon interaction is introduced on the field - theoretical basis as prescribed by qcd .
calculations are performed perturbatively to second order in the quark - pion and quark - gluon interaction .
all calculations are performed at one loop or at order of accuracy @xmath10 . in the following we proceed as follows : we first describe the basic formalism of our approach .
then we indicate the main derivations relevant to the problem and finally present the numerical results .
the effective lagrangian of our model @xmath11 contains the quark core part @xmath12 the quark - pion @xmath13 and the quark - gluon @xmath14 interaction parts , and the kinetic parts for the pion @xmath15 and gluon @xmath16 : @xmath17\psi(x ) - 1/f_{\pi } \bar\psi[s(r ) i \gamma^5 \tau^i \phi_i]\psi- \\
-g_s \bar\psi a_{\mu}^a\gamma^{\mu}\frac{\lambda^a}{2 } \psi + \frac{1}{2}(\!\partial_{\mu}\phi_i)^2 -{1\over 4 } g^a_{\mu\nu } g_a^{\mu\nu}.\end{aligned}\ ] ] here , @xmath18 , @xmath19 and @xmath20 are the quark , the pion and the gluon fields , respectively .
the matrices @xmath21 and @xmath22 are the isospin and color matrices , correspondingly . the pion decay constant @xmath2393 mev .
the scalar part of the static confinement potential is given by @xmath24 where c and m are constants .
at short distances , transverse fluctuations of the string are dominating @xcite , with some indication that they transform like the time component of the lorentz vector .
they are given by a coulomb type vector potential as @xmath25 where @xmath26 is approximated by a constant .
the quark fields are obtained from solving the dirac equation with the corresponding scalar plus vector potentials @xmath27\psi(x)=0\ ] ] the respective positive and negative energy eigenstates as solutions to the dirac equation with a spherically symmetric mean field , are given in a general form as @xmath28 @xmath29 the quark and anti - quark eigenstates @xmath30 and @xmath31 are labeled by the radial , angular , azimuthal , isospin and color quantum numbers @xmath32 and @xmath33 , which are collectively denoted by @xmath26 and @xmath34 , respectively .
the spin - angular part of the quark field operators @xmath35_{jm_j } \ , \ , j=|\kappa|-1/2.\ ] ] the quark fields @xmath36 are expanded over the basis of positive and negative energy eigenstates as @xmath37 the expansion coefficients @xmath38 and @xmath39 are operators , which annihilate a quark and create an anti - quark in the orbits @xmath26 and @xmath34 , respectively .
the free pion field operator is expanded over plane wave solutions as @xmath40\ ] ] with the usual destruction and creation operators @xmath41 and @xmath42 respectively .
the pion energy is defined as + @xmath43 the expansion of the free zero mass gluon field operators is of the same form . in denoting the three - quark vacuum state by @xmath44 ,
the corresponding noninteracting many - body quark green s function ( propagator ) is given by the customary vacuum feynman propagator for a binding potential @xcite : @xmath45 since the three - quark vacuum state @xmath46 does not contain any pion or gluon , the pion and gluon green s functions are given by the usual free feynman propagator for a boson field : @xmath47 \ , , \ ] ] @xmath48 \,,\ ] ] ( in the coulomb gauge ) , where we choose @xmath49 , @xmath50 .
using the effective lagrangian and the time - ordered perturbation theory one can develop a calculation scheme for the lower excitation spectrum of the nucleon and delta . in the model the quark core result ( @xmath51 ) is obtained by solving eq.([dirac ] ) for the single quark system numerically . since we work in the independent particle model and limited with the lower excitation spectrum of the nucleon and delta , the bare three - quark state of the @xmath52-flavor baryons corresponds to the structure @xmath53 in the non - relativistic spectroscopic notation .
the corresponding quark core energy is evaluated as the sum of single quark energies with : @xmath54 the result for @xmath51 still contains the contribution of the center of mass motion . to remove this additional term we resort to three approximate methods , which correct for the center of mass motion : the @xmath55 @xcite , @xmath56 @xcite and lho @xcite methods . the second order perturbative corrections to the energy spectrum of the su(2 )
baryons due to the pion field ( @xmath57 ) and the gluon fields ( @xmath58 ) are calculated on the basis of the gell - mann and low theorem : @xmath59 \ , |\phi_0>_{c}\end{aligned}\ ] ] with @xmath60 , where the relevant quark - pion and quark - gluon interaction hamiltonian densities are @xmath61 @xmath62 the stationary bare three - quark state @xmath63 is constructed from the vacuum state using the usual creation operators : @xmath64 where @xmath65 and @xmath66 represent the quantum numbers of the single quark states , which are coupled to the respective baryon configuration . the energy shift of eq.([energy_shift ] )
is evaluated up to second order in the quark - pion and quark - gluon interaction , and generates self - energy and exchange diagrams contributions .
the self - energy terms contain contribution both from intermediate quark @xmath67 and anti - quark @xmath68 states .
these diagrams correspond to the case when a pion or gluon is emitted and absorbed by the same valence quark which is excited to the intermediate quark and anti - quark states . in this way one can estimate the contribution of the sea - quarks to the hadron spectrum that can not be done in non - relativistic quark models .
the pion part of the self energy term ( pion cloud contribution ) ( see fig.[fig1a ] ) is evaluated as @xmath69 with @xmath70 .
the @xmath71 transition form factors are defined as : @xmath72 the vertex function of the @xmath73 and @xmath74 transition is @xmath75 where @xmath76 is the color unity matrix .
the estimations of the @xmath77 and @xmath78 transition form factors are given in the appendix a and b , respectively .
meson fields [ fig1a],width=566 ] after integration over the angular part in eq .
( 17 ) , the self - energy diagrams contribution to the baryon spectrum induced by pion fields ( pion cloud contribution ) is evaluated as : @xmath79 ^ 2}{e_{\alpha}-e_{\alpha ' } + p_0 } q_{s.e.}(l , l',l_n , j , j ' ) - \\ \sum\limits_{\beta}\frac{[\int dr r^2 g_{\beta \alpha ' } ( r)s(r)j_{l_n}(pr)]^2}{e_{\beta}+e_{\alpha ' } + p_0 } q_{s.e.}(l , l',l_n , j , j ' ) \biggr \ } , \end{aligned}\ ] ] where @xmath80 is the bessel function . the radial overlap of the single quark states with quantum numbers @xmath81 and @xmath82 is defined as @xmath83 the angular momentum coefficients @xmath84 are evaluated for all su(2 ) baryons as @xmath85[l_n][j ] \biggl [ c^{l'0}_{l^{\pm}0l_n 0 } w ( j \frac{1}{2}l_nl ' ; l^{\pm}j')\biggr ] ^2 \sum\limits_{m_j}\sum \limits_{m_j ' \leq \alpha_f } \biggl [ c^{j'm_j'}_{jm_jl_n(m_j'-m_j ) } \biggr ] ^2 , \end{aligned}\ ] ] where @xmath86 and @xmath87 are the clebsch - gordan and wigner coefficients , respectively . the gluon part of the second order self - energy diagrams ( gluon cloud ) contribution is estimated in a similar fashion as @xmath88 where the transition form factor is evaluated with the corresponding vertex matrix @xmath89 with the isospin unity matrix @xmath90 .
the corresponding feynman diagrams are given in fig.[fig1b ] , where the contribution from intermediate quark and anti - quark levels have opposite signs .
after evaluation of the transition form - factors and integration over angular variables , the self - energy term induced by gluon fields can be written as a sum of color - electric ( coulomb ) and color - magnetic parts : @xmath91\left ( \frac{[l][l^*]}{[l'][l'^ * ] } \right ) ^{1/2}c^{l'0}_{l0l_n0 } c^{l'^*0}_{l^*0l_n0 } \ ] ] @xmath92 ^ 2 } { e_{\alpha}-e_{\alpha ' } + p}pdp - \int \frac { [ r_{\beta \alpha'l_n}(p)+f_{\beta \alpha'l_n}(p)]^2 } { e_{\beta}+e_{\alpha ' } + p}pdp \right ] \ ] ] @xmath93\ ] ] @xmath94 \biggr \ } , \ ] ] where we define function @xmath95 @xmath96 and radial integrals @xmath97,\ ] ] @xmath98,\ ] ] @xmath99.\ ] ] , width=566 ] the angular momentum coefficients @xmath100 and @xmath101 are given in the appendix c. we note also that the sum in eq.(26 ) does not depend on the orientation of the full momentum of the valence quark @xmath102 .
the pion exchange contribution to the baryon energy - shift ( see fig.[fig2a ] ) is evaluated as : @xmath103 by using the wick s theorem we can write a more convenient expression for the energy shift of the su(2 ) baryons from the second order pion exchange diagrams : @xmath104 where @xmath105 and the operators @xmath106 and @xmath107 are summed over single quark levels @xmath108 of the su(2 ) baryon . in the quark model , the baryon wave function @xmath109 is presented as a bound state of three valence quarks , and it can be written down commonly as @xmath110 and @xmath111 are intermediate spin and isospin couplings , respectively .
the states @xmath36 are the single particle states , labeled by a set of quantum numbers @xmath26 , @xmath34 and @xmath112 , excluding the color degree of freedom .
meson exchange diagrams [ fig2a],width=566 ] the operator @xmath113 in equation ( [ pionex ] ) is the radial integration operator : @xmath114.\end{aligned}\ ] ] with @xmath83 where @xmath81 and @xmath82 are two sets of the single quark quantum numbers .
the matrix elements of the operator @xmath115 are given by @xmath116[l_n][j(\alpha)]\biggr)^{1/2}c^{l(\beta)0}_{l^{\pm } ( \alpha)0l_n 0 } \\
\nonumber w ( j(\alpha)\frac{1}{2}l_n , l(\beta ) ; l^{\pm(\alpha)},j(\beta ) ) c^{j(\beta)m(\beta)}_{j(\alpha)m_j(\alpha)l_n(m(\beta)-m(\alpha ) ) } , \nonumber\end{aligned}\ ] ] and the hermitian conjunction @xmath117 where @xmath118 are the quantum numbers of the single quark state @xmath119 .
the contribution of the second - order gluon - exchange terms to the baryon spectrum ( see fig.[fig2b ] ) is given by @xmath120 by using the wick s theorem we can write more convenient expression for this equation @xmath121 with the corresponding color - electric ( coulomb ) and color - magnetic parts : @xmath122 , width=566 ] the operator @xmath123 is the radial integration operator with the factor @xmath124 .
the operators @xmath125 and @xmath126 are the angular integration operator with the factors @xmath127 and @xmath128 respectively .
all these operators are summed over single quark levels @xmath129 of the su(2 ) baryon .
we note that the relativistic one - pion and one - gluon ( color - magnetic ) exchange forces have strongly different structure from corresponding non - relativistic exchange forces : the exchanged pion or gluon in the @xmath130-orbit couple the upper ( lower ) component of i - valence quark with the lower ( upper ) component of another j - valence quark ( see above equations ) and generate spin 0 , 1 , 2 , etc .
operators . as we see below , the only contribution to the g.s .
n and @xmath0 comes from the pion exchange and gluon exchange ( color - magnetic part ) via the spin 1 operator ( an analog of the spin - orbit operator ) , while in the non - relativistic case the corresponding terms are spin 0 ( spin - spin ) operators .
another important point is that the both exchange operators are symmetric under the permutation of the upper and the lower components of the interacting valence quarks in i- and j - orbits .
this fact indicates the chiral invariance of the lagrangian in the matrix elements level .
in order to account for the finite size effect of the pion , we introduce a one - pion vertex regularization function in the momentum space , parameterized in the dipole form as @xmath131 from the flux - tube study @xcite we fix @xmath132 . the strong coupling constant @xmath133 with the value @xmath2 . the parameters of the confining potential ( @xmath134 and @xmath135 ) are chosen to reproduce the correct axial charge @xmath136 of the proton ( and the empirical pion - nucleon coupling constant @xmath137 via the goldberger - treiman relation ) and a normal value for the quark core rms radius of the proton ( see @xcite ) .
in table 1 we indicate the model parameters together with the corresponding single - quark energies for the models @xmath138 and @xmath139 . by examining the two model parameters we can check the sensitivity of the baryon spectrum on the description of the static properties of the proton .
table 2 contains the quark core results @xcite for the static properties of the proton with the correction on the center of mass motion ( cm ) .
one can note from the table that a larger value of the strength parameter @xmath134 ( model @xmath139 ) of the confining potential yields a smaller value for the proton rms radius .
further we will see that the calculated spectrum is quite sensitive on the choice of the the strength parameter , hence on the proton rms radius . in table 3
we give the perturbative corrections to the valence quarks energy shifts in the @xmath140 orbits from the self - energy diagrams induced by pion and gluon fields .
contributions from intermediate quark and anti - quark states are included with the full momentum up to @xmath1 which enables convergent results opposite to the bag models ( see @xcite ) .
the columns with @xmath141 corresponds to the special cases where an intermediate quark is in the same orbit as the initial and final quark .
these cases are useful to check the limitation of the self energy in the literature with the intermediate ground state of the baryon .
as one can note from the table , the self - energy results with the restriction @xmath141 are negative for the both pion and gluon ( color - magnetic part ) fields , while are positive without the restriction ( due to the intermediate anti - quark states ) .
thus the restriction @xmath141 decreases the baryon energy contrary to the complete sum of intermediate quark and anti - quark states .
the contributions of the exchange diagrams to the energy - shift of the lower su(2 ) baryon states are given in tables 4 and 5 . as was pointed out above , the relativistic structures of the pion and gluon exchange mechanisms are different from the corresponding non - relativistic operators . in first case
the both pion and gluon exchange operators couple the upper ( lower ) and lower ( upper ) dirac components of the two interacting valence quarks .
one can see from table 4 that in the ground and radially excited states of the @xmath142 and @xmath0 where all three valence quarks are in the @xmath143-orbits , the exchanged pion or gluon can be only in the @xmath144 orbit , i.e. generate spin 1 operators ( analogy of the spin - orbit operator ) .
these operators , naturally , can contribute to the proton spin like the lower component ( p - wave ) of the dirac wave function . the non - relativistic pion- and gluon - exchange operators between valence quarks in the ground state have a spin - spin structure , i.e. are spin 0 operators and as a result do not contribute to the proton spin . for the baryon states with the last excited quark in the @xmath145-orbit , the contribution comes from the both @xmath146 ( spin - spin ) and @xmath144 ( spin - orbit ) terms . for baryons with a structure
@xmath147 the corresponding exchange forces generate the @xmath144 ( spin - orbit ) and @xmath148 ( tensor ) operators . from the table 5
one can note that the color - electric forces yield large spin - independent and small spin - dependent matrix elements . in the case when the intermediate quark states are restricted to the ground state in the @xmath149 and @xmath150 , the color - electric component of the gluon - cloud contribution is exactly canceled by the corresponding exchange term in consistence with the mit bag - model result .
the analysis of the results presented for the excited baryon states with the structure @xmath151 and @xmath152 indicates that the spin - orbit matrix elements generated by the both pion- and gluon - exchange ( color - magnetic ) forces are of acceptable order contrary to the non - relativistic one - gluon exchange spin - orbit matrix elements @xcite . in table 6
we give the mass values for the g.s .
n(939 ) with and without cm correction for the both model a and b. the restriction to the intermediate ground state when estimating the self - energy ( i=0 ) yields too small values for the nucleon energy ( and also for other baryons ) .
contrary , the inclusion of all excited intermediate states results large mass values .
model b yields even larger value for the mass of the nucleon .
this means that larger values of the strength parameter @xmath134 of the confining potential is not likely .
another important point is that the inclusion of the gluon - exchange and self - energy diagrams do not improve the situation . for the model a
, the contribution of the pion fields to the @xmath149 is about 200 mev , while the gluon fields yield more than 500 mev shift with the strong coupling constant @xmath2 .
one can note that the three methods for the correction of center of mass motion agree within 50 mev which seems too large .
however , these three methods always give corrections with systematic differences .
for example , the lho method yields correction larger than the @xmath56 method , but smaller than the @xmath55 method .
thus , we can fix one of these methods and go to the excited sector . and
finally in last table 7 we compare the theoretical energy values including the cm correction in the lho method with experimental data from pdg @xcite .
we assume that the lower negative parity excited n and @xmath0 states correspond to the excited last valence quark in the @xmath153-orbit , while the radially excited nucleon @xmath154 ( roper ) and delta @xmath155 resonances are assigned with the radially ( in 2s - state ) excited valence quark .
we note , that one can recalculate the whole gluon - loop corrections to the su(2 ) baryon spectrum by changing the value of the strong coupling constant ( we use @xmath2 ) in order to have smaller contribution to the mass - shift values .
however , this way yields smaller values for the splittings between su(2 ) baryon states .
the situation is not helpful for the @xmath156 splitting , since we have only 63 mev with the above value of the strong coupling constant .
thus , we come to the conclusion that the gluon field corrections to the mass - shift values of the su(2 ) baryon states must be small and confirm the results of the non - relativistic cqm results @xcite . when ignoring the gluon loop corrections , we have an overall good description of the su(2 ) baryon spectrum . the nucleon g.s . and the roper resonance @xmath157 are overestimated by more than 200 mev , while the orbital excitations of the nucleon , @xmath158 and @xmath159 are slightly underestimated by about 40 - 60 mev .
the situation in the delta sector is quite different . unlike the nucleon sector , the @xmath160 resonance and it s first radial excitation @xmath161
are overestimated slightly by 50 - 90 mev in the model a. but , the orbitally excited delta resonances @xmath162 and @xmath163 are also underestimated by small values ( like the nucleon sector ) . in order to check whether the gluon field contributions can be replaced by the multiple pion exchange mechanism , we repeat the calculations of the ground state nucleon and @xmath160 with smaller and larger values of the pion decay constant . for the @xmath164 1.5 @xmath165 we obtained @xmath166 mev and @xmath1671118 mev which have to be compared with the numbers 1166 mev and 1310 mev from the table 7 , respectively . for the value
@xmath164 0.5 @xmath165 we have @xmath168 mev and @xmath1672342 mev , which are very large comparing the corresponding estimations calculated with the experimental pion decay constant @xmath165 .
the different level of the estimations for the nucleon and delta sectors indicates on the necessity of the strong instanton induced exchange mechanism , which does not change the delta spectrum , while effecting the nucleons essentially .
additionally , the two - pion loop corrections are needed for the both sector , which would yield small corrections .
we have developed a relativistic chiral quark model with the inclusion of the gluon and pion fields contribution to the mass spectrum of the su(2 ) baryons .
these fields were shown to generate the second - order self - energy and exchange corrections .
the overall good description of the lower excitation spectrum of the @xmath142 and @xmath0 states are obtained when restricting to the pion fields corrections .
the color - electric components of the gluon fields give enormously large contributions to the baryon spectrum ( about 500 mev with @xmath2 for the nucleon g.s . ) .
the exchanged pion or gluon ( color - magnetic component ) generates the spin 0 ( spin - spin ) , spin 1 ( spin - orbit ) and spin 2 ( tensor ) operators with reasonable values of the matrix elements for the su(2 ) baryons energy shifts , thus indicating no any `` spin - orbit problem '' in the excited nucleon and delta sectors . the only contribution to the g.s .
nucleon and delta comes from the spin 1 ( spin - orbit ) operator which can contribute also to the baryon spin .
when using a small value of the strong coupling constant or completely ignoring the gluon loop corrections , the model yields a quite good description of the delta resonances mass - spectrum within 50 - 60 mev .
however , the calculated nucleon sector differs from the experimental values essentially . in conclusion
, the developed model still needs an additional mechanism for the description of the excited nucleon and @xmath0 spectrum .
possible candidates are the two - pion exchange and the instanton induced exchange ( for the nucleon sector ) forces . _
acknowledgements_. the work is supported in part by the daad research fellowship ( germany ) and by the scientific fund for the support of fundamental research , uzbekistan academy of sciences .
the author acknowledges the institut fr theoretische physik , universitt tbingen for kind hospitality and thanks d.baye , a.faessler , th .
gutsche , a. rakhimov , f. stancu , c. semay for useful discussions .
99 isgur n and karl g 1977 phys . lett . *
72b * 109 isgur n and karl g 1978 phys . rev .
* d18 * 4187 isgur n 1999 phys . rev .
* d60 * 054013 capstick s and isgur n 1986 phys . rev .
* d34 * 2809 stancu f and stassart p 1990 phys . rev .
* d41 * 916 stancu f and stassart p 1991 phys .
b269 * 243 glozman l y and riska d o 1996 phys . rept .
* 268 * 263 glozman l y , plessas w , varga k and wagenbrunn r f 1998 phys .
* d58 * 094030 lring u , kretzschmar k , metsch b c and petry h r 2001 euro.phys.jour . * a10 * 309 + lring u , metsch b c and petry h r 2001 euro.phys.jour . * a10 * 395 + lring u , metsch b c and petry h r 2001 euro.phys.jour . * a10 * 447 glozman l y , papp z , plessas w , varga k and wagenbrunn r f 1998 phys
. rev . * c57 * 3406 capstick s and roberts w 2000 prog . part .
* 45 * s241 riska d o and brown g e 1999 nucl . phys . *
a653 * 251 holinde k 1992 nucl . phys . *
a543 * 143c ashman j et al 1988 phys .
* b206 * 364 thomas a w , theberge s and miller g a 1980 phys . rev . * d22 * 2838 ; 1981 phys . rev .
* d24 * 216 thomas a w 2008 prog.part.nucl.phys . * 61 * 219 + myhrer f and thomas a w 2008 phys.lett .
* b663 * 302 saito k 1984 prog . of theor .
* v71 * 775 nagai s , miyatsu t and saito k 2008 phys . lett . * b666 * 239 oset e , tegen r and weise w 1984 nucl . phys .
a * 426 * 456 gutsche t 1987 ph .
d. thesis , florida state university , 249 p. ( unpublished ) gutsche t and robson d 1989 phys .
b229 * 333 tursunov e m 2005 j. phys .
g : nucl . part .
31 * 617 bijker r and santopinto e 2008 arxiv:0809.2299 ( nucl - th ) inoue t , lyubovitskij v e , gutsche t and faessler a 2006 int .
e * 15 * 121 lscher m 1981 nucl .
b180 * 317 gell - mann m and levy m 1960 nuovo cim .
* 16 * 1729 fetter a i and waleska j d 1971 quantum theory of many particle systems ( mcgraw - hill , new york ) lu d h , thomas a w and williams a g 1998 phys.rev . *
c57 * 2628 tegen r , brockmann r and weise w 1982 z.phys .
* a307 * 339 wilets l 1989 " non - topological solitons ( world scientific , singapoure ) yao w m et al ( particle data group ) 2006 j.phys.*g33 * 1 putting explicit expression of the vertex matrix @xmath170 from eq.(19 ) into eq.(17 ) we receive next equation : @xmath171s(r ) \\ & & \int d\hat{r}\big [ { \cal y}_{jl}^{m_j^+}(\hat{r})(\vec{\sigma}\hat{r } ) { \cal y}_{j'l'}^{m_j'}(\hat{r})e^{-i\vec{p}\vec{r}}\big ] < m_t|\tau^a|m_t'><m_c|i_c|m_c'>\end{aligned}\ ] ] now using @xmath172 which couples the lower orbital momentum to the spin , and expanding the exponential function over spherical bessel functions and integrating over angular part of the variable @xmath173 , we get next equation for the integral @xmath174= \sum_{l_n}(-i)^{l_n } j_{l_n}(pr ) y_{l_n}^{m_j'-m_j}(\hat{p } ) { \cal f}(l^{\pm},l',l_n , j , j',m_j , m_j'),\ ] ] where coefficients @xmath175 are defined as @xmath176[l_n][j]\biggr)^{1/2}c^{l'0}_{l^{\pm}0l_n0 } w ( j\frac{1}{2}l_n , l ' ; l^{\pm},j ' ) c^{j'm'_j}_{jm_jl_n(m'_j - m_j)}.\ ] ] for the transition form - factor now it is easy to write the next expression : @xmath177s(r ) j_{l_n}(pr ) \\ & & y_{l_n}^{m_j'-m_j}(\hat{p } ) { \cal f}(l^{\pm},l',l_n , j , j',m_j , m_j ' ) < m_t|\tau^a|m_t'><m_c|i_c|m_c'>.\end{aligned}\ ] ] the hermitian conjunction of the transition form factor @xmath178s(r ) j_{l_n}(pr ) \\ & & y_{l_n}^{(m_j'-m_j)*}(\hat{p } ) { \cal f}(l^{\pm},l',l_n , j , j',m_j , m_j ' ) < m_t'|\tau^a|m_t><m_c'|i_c|m_c>.\end{aligned}\ ] ] for the self - energy diagrams we use an alternative expression of the transition form - factors .
putting the quark wave functions with further integration over the radial part of the spatial coordinate one can write for the transition form - factor next equation : @xmath183[l_n](4\pi ) } { [ l']}\big)^{\frac{1}{2}}(-i)^{l_{n } } y_{l_{n}m_{n}}(\hat{p})m_{m_{s}m_{s}'}^{\mu}c_{\mathrm{l0l_{n}0}}^{\mathrm{l'0 } } c_{\mathrm{lm_{l}\frac{1}{2}m_{s}}}^{\mathrm{jm_{j } } } c_{\mathrm{l'm_{l}'\frac{1}{2}m_{s}'}}^{\mathrm{j'm_{j } ' } } c_{\mathrm{lm_{l}l_{n}m_{n}}}^{\mathrm{l'm_{l}'}}\cdot \\ \cdot\int{r^{2}r_{\mathrm{\mu_{ll'}}}^{\mathrm{\alpha\alpha'}}(r)j_{\mathrm{l_{n}}}(pr)dr } < m_{t}|i_{t}|m_{t}'><m_{c}|\frac{\lambda_{a}}{2}|m_c ' > , \end{aligned}\ ] ] the coefficients @xmath190 are sums of the clebsch - gordan coefficients with the corresponding delta factors ( instead of the delta factors of the coefficient @xmath191 ) : @xmath192 , @xmath193 , and @xmath194 , respectively .
|
the lower excitation spectrum of the nucleon and @xmath0 is calculated in a relativistic chiral quark model .
corrections to the baryon mass spectrum from the second order self - energy and exchange diagrams induced by pion and gluon fields are estimated in the field -theoretical framework .
convergent results for the self - energy terms are obtained when including the intermediate quark and antiquark states with a total momentum up to @xmath1 .
relativistic one - meson and color - magnetic one - gluon exchange forces are shown to generate spin 0 , 1 , 2 , etc .
operators , which couple the lower and the upper components of the two interacting valence quarks and yield reasonable matrix elements for the lower excitation spectrum of the nucleon and delta .
the only contribution to the ground state nucleon and @xmath0 comes from the spin 1 operators , which correspond to the exchanged pion or gluon in the l=1 orbit , thus indicating , that the both pion exchange and color - magnetic gluon exchange forces can contribute to the spin of baryons .
is is shown also that the contribution of the color - electric component of the gluon fields to the baryon spectrum is enormously large ( more than 500 mev with a value @xmath2 ) and one needs to restrict to very small values of the strong coupling constant or to exclude completely the gluon - loop corrections to the baryon spectrum . with this restriction
, the calculated spectrum reproduces the main properties of the data , however needs further contribution from the two - pion exchange and instanton induced exchange ( for the nucleon sector ) forces in consistence with the realistic nn - interaction models .
|
suppose that @xmath0 is a closed connected oriented surface @xmath1 of genus @xmath2 with a set of @xmath3 labelled marked points @xmath4 . by a _ translation surface _
@xmath5 we mean a flat riemannian metric and a parallel vector field on @xmath0 .
the metric has cone type singularities at all of the points @xmath6 of @xmath7 where the total angle is of the form @xmath8 for some integer @xmath9 .
each geodesic on a translation surface moves in a constant direction , so geodesics do not have self intersections and a regular geodesic that connects a non - singular point to itself comes back with the same angle , so it is a periodic geodesic .
a periodic geodesic is always part of a maximal connected periodic region : a maximal cylinder of parallel periodic geodesics of the same length .
we refer to such a maximal cylinder as a _ periodic cylinder _ or , for short , a _ cylinder _ , and we say that the common length of the periodic geodesics that make up the cylinder is the _ width _ of the cylinder .
the number @xmath10 of cylinders of width less than @xmath11 grows like @xmath12 : a first fundamental result of masur @xcite states that there exist two positive constants @xmath13 and @xmath14 such that @xmath15 . using
this fact , eskin and masur @xcite prove the deep theorem that for almost every surface in a stratum there is an exact asymptotics @xmath16 .
the constant @xmath17 is called a _ siegel
veech constant_. the main tool to study siegel
veech constants is the method of veech @xcite that also showed the exact asymptotics in a more general setting , but for a weaker form of convergence .
the main object of this paper is the computation of various siegel
veech constants related to counting periodic regions in different ways . siegel
veech constants are particularly interesting because they appear in various fields . in arithmetic
, they give the asymptotics of the number of primitive points in certain lattices . in dynamics
they are related to the sum of lyapunov exponents for the teichmller geodesic flow in any invariant suborbifold of a stratum of abelian differentials by the main formula of @xcite . for a teichmller curve
, they have an algebro - geometric interpretation involving degrees of certain line bundles given by a formula of bouw - mller @xcite .
furthermore they are related to slopes of effective divisors and intersection theory in the moduli space of curves : this aspect is studied by chen ( see @xcite for example ) and chen - mller ( see @xcite for example ) . by identifying @xmath18 and @xmath19 , a translation surface inherits from @xmath19 a complex structure on @xmath20 and a holomorphic one form ( _ abelian differential _ ) @xmath21 .
a zero of @xmath21 of order @xmath22 corresponds to a conical singularity of angle @xmath23 .
there is a one to one correspondence between translation surfaces endowed with a distinguished direction and abelian differentials .
if we denote by @xmath24 the orders of the zeros of @xmath21 then we have @xmath25 . for a given @xmath26 such that @xmath25 and @xmath9 , for @xmath27
, we consider the stratum @xmath28 of the moduli space of abelian differentials on @xmath0 that have zeros at the points of @xmath7 of orders @xmath29 , or equivalently the moduli space of translation surfaces @xmath5 with singularities at the points of @xmath7 of angles @xmath30 .
the dimension of @xmath28 is @xmath31 .
the stratum @xmath28 may be non connected @xcite but contains at most three connected components @xcite .
the stratum @xmath28 admits a volume element @xcite that induces a finite @xmath32-invariant measure on the hyperspace @xmath33 of area one surfaces in @xmath28 .
the volumes of the connected components of the strata of abelian differentials were effectively computed by a. eskin and a. okounkov @xcite .
the boundary of a periodic cylinder contains singularities .
generically , each boundary component contains exactly one singularity so it is a _ closed saddle connection _
, i.e. a geodesic that joins a singularity to itself ( and contains no other singularity ) .
consider a translation surface @xmath5 .
for each positive real number @xmath34 , denote by @xmath35 the number of periodic cylinders in @xmath5 of width at most @xmath11 and by @xmath36 the total area of these cylinders .
it was shown in @xcite that let @xmath28 be a stratum , where @xmath26 with @xmath9 , for @xmath27 . for every connected component @xmath37 of @xmath33 , there exist constants @xmath38 and @xmath39 such that for almost every translation surface @xmath5 in @xmath37 one has @xmath40 the _ siegel veech constants _ @xmath38 and @xmath39 only depend on @xmath37 . an earlier version of this result ( in a more general setting ) where convergence is replaced by convergence in @xmath41 was proved in @xcite .
the results in @xcite assure the existence of quadratic asymptotics ( siegel veech constants ) if one counts cylinders with weights ( under certain conditions ) .
the existence of all the siegel
veech constants we consider in this paper is justified by these results ( see section [ sec : notation ] ) . the function @xmath42 for example counts cylinders with weight the area of the cylinder .
one can also only count cylinders with sufficiently big area : for @xmath43 , denote by @xmath44 the number of cylinders in @xmath5 of width at most @xmath11 and of area at least @xmath45 .
we denote the corresponding siegel
veech constant by @xmath46 .
it is shown in @xcite that [ th : vorobets ] let @xmath28 be a stratum , where @xmath26 with @xmath9 , for @xmath27 . then for any connected component @xmath37 of @xmath33 @xmath47 the ratio @xmath48 of part a ) of the theorem can be interpreted as the ( asymptotic ) mean area of a cylinder on a generic surface in @xmath37 in the following sense : for almost any surface @xmath49 in @xmath37 one has @xmath50 .
part b ) of the theorem is an answer to a question of veech in @xcite where the author asks if there is a simple formula for @xmath51 ( which is the same as @xmath52 ) . as a by - product of our results using the methods from @xcite we get an alternative proof of the theorem of vorobets by evaluating an explicit integral that is a simplified version of the integral used in @xcite .
the methods from @xcite allow for the computation of various other siegel veech constants associated to counting cylinders for more specific data that we allude to next .
it might happen that the geodesic flow in a given direction on @xmath49 contains several ( maximal ) periodic cylinders in that direction .
generically , this only happens if the boundary saddle connections of the cylinders are homologous .
the boundary saddle connections might be part of a larger family of homologous saddle connections , where the extra saddle connections do not bound a cylinder .
we refer to such a family as a _ configuration of homologous saddle connections _ or simply as a _ configuration_. a topological representation of such configurations can be obtained by taking blocks of surfaces as in figures [ fig : type i ] , [ fig : type ii ] , and [ fig : type iii ] ( the surfaces are drawn as tori , but might have arbitrary genus ) , arranging them in a cyclic order and then identifying the boundary components .
note that the fact that the saddle connections are homologous implies that they are all of the same length . call this the _ length _ of the configuration .
it also implies that a configuration persists under small deformations of the translation structure .
one says that two configurations of homologous saddle connections correspond to the same _ topological type _ if the saddle connections are based at the same singularities , if the complementary regions are of the same topological type and have the same number and type of singularities , e.t.c .
in particular , they always have the same number of complementary cylinders .
the length of the configuration is the common width of the cylinders coming from the configuration . for each connected component @xmath37 of a stratum
there are only a finite number of _ admissible _ topological types of configurations , i.e. topological types of configurations that are realized on at least one surface in @xmath37 .
in fact almost all surfaces share the same admissible topological types .
we will consider only configurations of this special type .
their distinguishing feature is that they persist under any small deformation . for related counting problems in the case of general periodic components ( so not configurations ) ,
see @xcite . for @xmath5 in @xmath37 , denote by @xmath53 the number of configurations of homologous saddle connections on @xmath5 of type @xmath54 and whose length is at most @xmath11 . for a given connected component @xmath37 of a stratum @xmath33 and an admissible topological type @xmath54 of configurations
there exists a siegel veech constant @xmath55 such that for almost any surface @xmath5 in @xmath37 one has @xmath56 the authors of @xcite give a method to compute these siegel veech constants .
we denote by @xmath57 the number of cylinders of length less than @xmath11 coming from a configuration of type @xmath54 . for a real number @xmath58
we denote by @xmath59 the sum of the @xmath60-th power of the area of each of these cylinders .
we denote the corresponding siegel
veech constants by @xmath61 , resp .
@xmath62 . for @xmath63
we write @xmath64 .
note that if @xmath54 comes with @xmath65 cylinders then @xmath66 . it follows from the general result of @xcite that there is a siegel veech constant @xmath62 such that for almost any surface @xmath5 in @xmath37 one has @xmath67 the methods from @xcite can be applied to compute @xmath62 in a way similar to the computation of @xmath55 .
the expression for @xmath62 contains a constant @xmath49 that depends only on combinatorial data such as the dimension of the stratum , the order of the singularities and the possible symmetries .
it is given by an explicit formula in
13.3 . of @xcite .
it also contains the `` principal boundary stratum '' @xmath68 determined by @xmath37 and @xmath54 ( the stratum of possibly disconnected surfaces we get by contracting the closed saddle connections of the configuration ) .
the constant @xmath69 always denotes the complex dimension of @xmath68 .
if @xmath54 comes with @xmath65 cylinders then we have @xmath70 . with this notation
we show in section [ sec : mean - area ] , [ th : area - p ] given a real number @xmath71 .
let @xmath28 be a stratum , where @xmath26 and @xmath9 , for @xmath27 .
let @xmath37 be a connected component of @xmath33 and @xmath54 an admissible topological type of configuration containing @xmath72 cylinders .
then @xmath73 where @xmath70 .
@xmath49 denotes the combinatorial constant given in 13.3 .
of @xcite and @xmath68 denotes the principal boundary stratum .
note that @xmath74 and @xmath75 , so the second and third equation of the previous theorem follow from the first .
1 . evaluation of @xmath76 is motivated by the question of m. mller related to the study of quasimodular properties of the related counting function .
we recall that the saddle connections in a configuration of given type @xmath54 can be named .
choose and fix one of the saddle connections that bounds a cylinder .
in the proof of theorem [ th : area - p ] we show that if we only consider the area of this cylinder then we get the same formulas for @xmath62 and @xmath64 except that the factor @xmath65 is missing .
we get as a corollary : [ cor : mean - area - p ] given a real number @xmath58 and let @xmath28 be a stratum of abelian differentials on a surface @xmath0 , where @xmath26 with @xmath9 , for @xmath27 . then for any connected component @xmath37 of @xmath33 and any admissible type @xmath54 of configuration containing at least one cylinder , @xmath77 where @xmath78 .
\(a ) the quotient of the corollary can be interpreted as the ( asymptotic ) mean area of a cylinder coming from a configuration of type @xmath54 , where the area is counted with a power @xmath60 .
\(b ) for a natural number @xmath79 we obtain @xmath80 define @xmath81 in the same way as we defined @xmath36 in section [ sec : known - results ] , except that the area of each cylinders is counted with a power @xmath60 . if @xmath82 denotes the corresponding siegel - veech constant , then we have @xmath83 where the sum is taken over all admissible topological types of configurations for @xmath37 with at least one cylinder .
this implies [ cor : vorobets - p ] given a real number @xmath58 and let @xmath28 be a stratum , where @xmath26 with @xmath9 , for @xmath27 . then for any connected component @xmath37 of @xmath33 @xmath84 where @xmath78 .
for @xmath63 , corollary [ cor : vorobets - p ] becomes part ( a ) of the theorem of vorobets as stated in section [ sec : known - results ] .
corollary [ cor : mean - area - p ] gives more detailed information as corollary [ cor : vorobets - p ] .
for example , the mean area @xmath85 of a cylinder coming from a configuration of type @xmath54 does not depend on the number of cylinders .
this means in particular that the mean area of a cylinder is the same if the cylinder makes up the whole periodic region of a configuration or if the periodic region is made up of several cylinders . as a variation of the above ,
we denote by @xmath86 the @xmath60-th power of the total area of the periodic region ( union of the cylinders ) on @xmath5 coming from a configuration of topological type @xmath54 whose length is at most @xmath11 .
the corresponding siegel - veech constant is denoted by @xmath87 .
we show in section [ sec : mean - area - conf ] [ th : area - p - conf ] given a real number @xmath71 and let @xmath28 be a stratum , where @xmath26 and @xmath9 , for @xmath27 .
let @xmath37 be a connected component of @xmath33 and @xmath54 an admissible topological type of configuration containing @xmath72 cylinders .
then @xmath88 @xmath49 denotes the combinatorial constant given in 13.3 .
of @xcite and @xmath68 denotes the principal boundary stratum .
\(a ) for a natural number @xmath79 we obtain @xmath89 where @xmath90 .
\(b ) the quotient in part ( b ) of the preceeding theorem can be interpreted as the asymptotic mean area of the periodic part ( taking the @xmath60-th power of the area ) .
for @xmath63 we get @xmath91 , which is consistent , as @xmath92 is the mean area of a cylinder .
\(c ) for @xmath93 we have @xmath94 .
we next count configurations with cylinders of large area .
we recall that the saddle connections in a configuration of given type @xmath54 can be named .
choose and fix one of the saddle connections that bound a cylinder and call this the first cylinder .
given @xmath43 .
denote by @xmath95 the number of configurations of type @xmath54 of length at most @xmath11 and such that the area of the first cylinder is at least @xmath45 ( of the area one surface @xmath5 ) .
we denote by @xmath96 the corresponding siegel
veech constant .
note that we have @xmath97 .
we show in section [ sec : onecylindersp ] [ th : vorobets : constant ] given @xmath43 .
let @xmath28 be a stratum of abelian differentials on a surface @xmath0 , where @xmath26 with @xmath9 , for @xmath27 .
then for any connected component @xmath37 of @xmath33 and any admissible topological type @xmath54 of configuration for @xmath37 containing at least one cylinder , @xmath98 summing over all configurations we get as a corollary part ( b ) of the theorem of vorobets but our result contains more detailed information .
we next count configurations with periodic regions of large area .
let @xmath54 be a topological type of configuration that comes with @xmath72 cylinders . for @xmath43 , denote by @xmath99 the number of configurations on @xmath5 of type @xmath54 of length at most @xmath11 and such that the total area of the @xmath65 cylinders is at least @xmath45 ( of the area one surface @xmath5 ) .
we denote the corresponding siegel
veech constant by @xmath100 .
the _ incomplete beta function _ @xmath101 and the _ regularized incomplete beta function _
@xmath102 are defined for @xmath103 $ ] by @xmath104 for more details about the incomplete beta function see section [ sec : toolbox ] : figure [ fig : graphesbis ] represents the density function @xmath105 for various values for @xmath106 and @xmath107 . with this notation
we show in section [ sec : allcylindersp ] : [ th : big - area ] given @xmath43 , and let @xmath28 be a stratum , where @xmath26 with @xmath9 , for @xmath27 .
suppose that @xmath37 is a connected component of @xmath33 and that @xmath54 is an admissible topological type of configurations for @xmath37 containing exactly @xmath65 cylinders .
@xmath108 where @xmath109 .
see figure [ fig : graphes ] for the graph of @xmath110 for various values for @xmath69 and @xmath65 .
graphs of the function @xmath111 1 . the fraction @xmath112 can be interpreted as the mean proportion of configurations of type @xmath54 whose periodic complementary region is big , that is , the total area of the cylinders is at least @xmath45 of the area of the surface .
2 . @xmath113 and @xmath114 so @xmath115 in this form we can compare the result with the previous one for one cylinder , given in theorem [ th : vorobets : constant ] . in section [ sec : correlation ]
we consider the problem of correlation between the area of two cylinders .
let @xmath54 be an admissible configuration for a connected component @xmath37 that comes with at least two cylinders .
choose ( and fix ) two cylinders and let @xmath116 .
we denote by @xmath117 the number of configurations of width length at least @xmath11 such that the area @xmath118 of the first cylinder is at least @xmath119 and such that the area @xmath120 of the second cylinder is at least @xmath45 of the remaining surface , i.e. it is at least @xmath121 .
we denote by @xmath122 the corresponding siegel
veech constant . to simplify notation we will write @xmath123 instead of @xmath124 .
we show in section [ sec : correlation ] , [ th : correlation ] for any connected component @xmath37 of a stratum @xmath33 , where @xmath26 with @xmath9 , for @xmath27 , and any admissible topological type @xmath54 of configurations containing at least two cylinders , @xmath125 the result does not depend on @xmath119 .
morally , we compute the asymptotic probability that , among configurations whose first cylinder has area @xmath119 , we have a second cylinder with area at least @xmath126 . comparing theorems [ th : correlation ] and
[ th : vorobets : constant ] we see that this is the probability that , among all configurations , the area of the second cylinder is at least @xmath45 , except that the parameter space has one fewer dimension .
so , in some sense , except for the fact that the area of the first cylinder gives a restriction on the range for the area of the second cylinder , the area of the second cylinder is independent of the area of the first cylinder . in the results presented above we studied individual configurations . in the remaining part of the paper we study extremal properties of configurations among all configurations in a given stratum or even among all strata for a fixed genus . in section [ sec : maximal - abelian ]
we address the question of finding topological types of configurations @xmath54 ( admissible for some connected component @xmath37 ) that maximizes @xmath127 the constant @xmath128 can be interpreted as the asymptotic mean area of the periodic part ( union of the cylinders ) of the complementary region of a configuration of topological type @xmath54 .
each stratum has at most three connected components that are classified by the invariants `` hyperellipticity '' , and `` parity of spin structure '' .
( we recall in section [ sec : simple - abelian ] the classification of connected components from @xcite . ) the quantity @xmath129 varies considerably among strata . for the connected stratum @xmath130 ,
the maximal value of @xmath129 over all the configurations is @xmath131 . for the connected component @xmath132 it is equal to @xmath133 .
the following proposition gives an uniform bound on the ratio @xmath129 .
[ prop : qmax : dim : ab ] let @xmath37 be any connected component of a stratum @xmath28 and @xmath54 be any admissible topological type of configuration for @xmath37 , then the asymptotic mean area of the periodic complementary regions satisfies @xmath134 the maximum is attained for any genus @xmath135 : for each @xmath136 there is a topological type of configuration @xmath137 that is admissible for the component @xmath138 ( @xmath139 zeros of order @xmath140 ) of abelian differentials on a surface of genus @xmath2 such that the corresponding constant @xmath141 is @xmath142 . to prove this result we need to determine the maximal number of cylinders that can come from any configuration which is admissible for a fixed stratum @xmath28 .
we insist on the fact that we compute here the number of cylinders in rigid collections of saddle connections , which is different from the studies of naveh @xcite and lindsey @xcite where they count the number of parallel cylinders .
in section [ sec : simple - abelian ] we answer a question of a. eskin and a. wright : is it possible to find in each connected component of each stratum a topological type of configuration whose complementary regions are tori with boundary and cylinders .
the answer depends on the connected component .
we show ( proposition [ prop : alex ] ) : for hyperelliptic components this is not possible ; for the components with even spin structure when the genus is even this is not possible unless we allow one of the complementary regions to be a genus two surface ; in all other cases this is possible .
we introduce here most of the notation we need for the computation of siegel veech constants . for survey material on abelian differentials and translation surfaces
see @xcite . for the exact definition of configurations and related results see @xcite .
let @xmath143 denote a closed oriented surface @xmath20 of genus @xmath2 on which there are @xmath3 marked points @xmath4 .
by @xmath144 we denote a riemann surface structure @xmath145 on @xmath20 together with an abelian differential @xmath21 . if @xmath21 is not identically zero then @xmath7 is the set of zeros of @xmath21 . we usually only write @xmath21 for @xmath144 .
if we denote by @xmath146 the orders of the zeros of @xmath21 then we have @xmath25 .
the form @xmath21 can be used to define an atlas of adapted coordinates on @xmath145 . in these
adapted coordinates the abelian differential @xmath21 becomes @xmath147 in a neighborhood of any point of @xmath148 and is @xmath149 in a neighborhood of a point @xmath150 , where @xmath151 is the order of the zero @xmath6 .
transition functions away from the zeros for these adapted coordinates are translations .
we refer to a surface together with an atlas whose transition functions are translations as _ translation surfaces_. using such an atlas , @xmath148 inherits from the complex plane a flat ( zero curvature ) riemannian metric .
a zero of order @xmath151 of the abelian differential ( that is a regular point of the riemann surface structure ) corresponds to a conical singularity of the flat metric of total angle @xmath8 .
these points are also called _
saddles_. the horizontal unit vector field on @xmath19 pulls back by adapted coordinates to a horizontal unit vector field on @xmath148 .
away from the singularities , the leaves of the corresponding foliation are geodesics with respect to the flat metric .
in fact , for each @xmath152 we have a unit vector field and so a foliation in that direction that comes from the unit vector field in direction @xmath153 of the complex plane .
the converse construction is also possible : suppose that @xmath5 is a translation surface structure on @xmath0 , i.e. an atlas on @xmath0 whose transition functions are translations .
a translation surface inherits from @xmath18 a flat riemannian metric and a parallel vector field on @xmath0 .
we assume that @xmath20 is the metric completion of @xmath0 .
the points of @xmath7 that are not regular points for the metric , are cone type singularities where the total angle is of the form @xmath8 for some @xmath154 . by identifying @xmath18 and @xmath19
, a translation surface inherits from @xmath19 a complex ( riemann surface ) structure on @xmath20 and a holomorphic one form ( _ abelian differential _ ) @xmath21 .
the zeros of @xmath21 are contained in @xmath7 .
each translation surfaces can be represented by a polygon in the plane whose edges come in pairs of parallel sides of the same length . identifying
each pair by a translation one obtains a translation surface .
the vertices give rise to singularities ( or regular points if the total angle is @xmath155 . )
we identify two translation surface structures if there is a bijection of the underlying topological surface that is in local coordinates a translation .
we identify two abelian differentials ( for some complex structures ) if they are biholomorphically equivalent .
there is then a one to one correspondance between abelian differentials and translation surfaces .
we denote by @xmath156 the moduli space of abelian differentials @xmath21 or equivalently of translation surfaces @xmath5 .
the moduli space @xmath156 is an algebraic variety . given @xmath24 such that @xmath157 .
the set @xmath28 of abelian differentials that share the same zero structure @xmath158 is called a stratum .
the stratum @xmath28 is an algebraic subvariety that admits a natural affine structure and a natural `` lebesgue '' volume element induced by this affine structure @xcite .
the dimension of @xmath28 is @xmath31 .
the area of the translation surface @xmath5 defined by the abelian differential @xmath159 is @xmath160 .
we denote by @xmath33 the hyperspace of @xmath28 of area one surfaces .
masur @xcite and veech @xcite showed that @xmath33 with the measure induced by the measure on @xmath28 is of finite volume .
there is a natural @xmath161 action on @xmath28 .
an element @xmath162 acts on local coordinates by postcomposition of @xmath2 .
the measure on @xmath28 and @xmath33 is @xmath161 invariant .
if we represent a translation surfaces by a polygon in @xmath18 , then the action of @xmath2 is the usual action on @xmath18 .
geodesic segments in the flat metric have constant angle with respect to the flat metric , so a geodesic segment @xmath163 can be represented by a holonomy vector @xmath164 in @xmath18 .
the angle and length of the vector is given by the direction and length of the geodesic segment .
we will often use @xmath163 both for the geodesic segment and for the holonomy vector . as we said above , a closed geodesic
is always contained in a maximal cylinder of closed geodesics and each boundary component of the cylinder is generically a closed saddle connection .
as all of the geodesics in the cylinder are represented by the same vector we say that this vector represents the cylinder .
its length is the _ width _ of the cylinder .
suppose that on a translation surface we find in some direction a _ configuration of homologous saddle connections _ , meaning a maximal family of saddle connections , that are homologous relative to the singularities .
all of the saddle connections are parallel and of the same length , so they share the same holonomy vector .
its length is the _ length _ of the configuration .
a configuration defines the following data : the named singularities the saddle connections are based at ; the topological type of the complementary regions ; the knowledge of which saddle connection bounds which complementary region , so in particular the cyclic order of the complementary regions ; the order of the singularities ( if there are any ) in the interior of each complementary region ; the singularity structure on the boundary ( the type of boundary and angles at the boundary singularities to be described below ) .
we say that two configurations are of the same _ topological type _ if they define the same data .
almost all surfaces in a given connected component @xmath37 share the same topological types of configurations that can be realized on the surface .
we talk about an _ admissible topological type _ for @xmath37 .
denote by @xmath165 the ( discrete ) set of holonomy vectors associated to configurations on @xmath5 of type @xmath54 .
we are allowed to associate weights to the holonomy vectors .
denote by @xmath166 the disk of radius @xmath11 centered at the origin and by @xmath167 the cardinality of @xmath168 , where the elements of @xmath168 are counted with their weights .
so if we write an element of @xmath165 as @xmath169 where @xmath170 is the weight of the holonomy vector @xmath171 , then @xmath172 . by using appropriate weights on the holonomy vectors the counting function @xmath167 becomes the counting functions @xmath53 , @xmath57 , @xmath173 e.t.c .
introduced in section [ sec : results ] .
if for example we count holonomy vectors associated to a configuration with weight one ( resp . the number of cylinders , total area of the cylinders ) then @xmath167 equals @xmath53 ( resp .
@xmath57 , @xmath173 ) .
given a connected component @xmath37 of some stratum @xmath33 and an admissible topological type @xmath54 for @xmath37 .
it follows from @xcite that the set of ( weighted ) holonomy vectors @xmath165 and the associated counting functions @xmath167 we consider in this paper verify the following conditions : 1 . for every @xmath174 , @xmath175 .
2 . for every @xmath176
there exists a constant @xmath177 such that @xmath178 .
the constant @xmath179 can be chosen uniformly on compact sets of @xmath37 .
3 . there exist constants @xmath34 and @xmath180 such that @xmath181 is @xmath182 as a function of @xmath5 .
in fact , the authors of @xcite show that the above conditions are verified for the set @xmath183 of holonomy vectors of closed saddle connections on @xmath5 where the weight associated to a holonomy vector is the number of saddle connections on @xmath5 that share this vector .
the weights we use in this paper are `` invariant under @xmath161 '' , so the sets @xmath165 we use verify condition ( a ) .
the number of configurations is bounded by the number of saddle connections , and as we use bounded weights , the counting functions @xmath167 we consider also satisfy conditions ( b ) and ( c ) . for @xmath184 one defines the function @xmath185 by @xmath186 the fact that the sets @xmath165 and associated counting functions @xmath167 we consider in this paper satisfy the conditions ( a ) , ( b ) , ( c ) , implies , using @xcite , that : let @xmath37 be a connected component of some stratum @xmath33 of abelian differentials and @xmath54 an admissible topological type of saddle connections for @xmath37 .
then for any of the sets @xmath187 and associated counting functions @xmath188 we consider in this paper there is a constant @xmath189 such that the following holds : 1 . for almost any translation surface @xmath5 in @xmath37 , @xmath190 2 . for any @xmath184 ,
@xmath191 if the convergence in ( a ) is replaced by convergence in @xmath41 then this theorem follows from a more general theorem under similar hypotheses in @xcite . in that paper , ( b )
is proved for integrable functions @xmath192 of compact support .
we will only use ( b ) for the characteristic function on a disk .
the paper @xcite explains how this last theorem can be used to compute the siegel veech constant @xmath55 .
the same method can be used to compute the other siegel veech constants we consider in this paper .
the strategy used in @xcite is as follows : if we apply ( b ) for the characteristic function @xmath193 of the disc @xmath194 of radius @xmath195 centered at the origin then the integral of the right hand side becomes @xmath196 and we have @xmath197 so @xmath198 denote by @xmath199 the subset of translation surfaces in @xmath37 that contain at least one configuration of type @xmath54 of length smaller than @xmath195 .
so @xmath200 is zero outside @xmath199 .
denote by @xmath201 the subset of @xmath199 of translation surfaces that contain exactly one configuration of type @xmath54 of length smaller than @xmath195 but contain no other closed saddle connection of length smaller than @xmath195 . using a result from @xcite
the autors of @xcite show that @xmath202 so @xmath203 consider a translation surface @xmath5 in @xmath201 .
on @xmath5 we have a configuration of closed saddle connections of topological type @xmath54 of length smaller than @xmath195 and no other short closed saddle connection .
cutting along the closed saddle connections we decompose @xmath5 into several pieces .
there will be some , say @xmath79 , surfaces @xmath204 with boundary and some , say @xmath205 , periodic cylinders @xmath206 , all having the same width . the pieces are arranged in a cyclic order .
the boundary of each @xmath207 is made up of two closed saddle connections .
for each connected component @xmath207 , by taking out the boundary and then taking the compactification we get a surface with either one boundary component , a _ figure eight _ , as in the right part of figure [ fig : figureeight ] , or we get a surface with two boundary components , a _ pair of holes _ , as in the right part of figure [ fig : twoholes ] . in the first case ,
the figure eight boundary describes two interior sectors of the surface of angles @xmath208 and @xmath209 , for some integers @xmath210 .
( figure [ fig : figureeight ] illustrates the case @xmath211 and @xmath212 . ) by shrinking the figure eight boundary to a point we produce a singularity of order @xmath213 if @xmath214 , or a regular point if @xmath215 .
( see the left part of figure [ fig : figureeight ] . )
figure eight construction ] in the second case , each of the boundary components comes with a boundary singularity of angles @xmath216 and @xmath217 for some integers @xmath218 .
( figure [ fig : twoholes ] illustrates the case @xmath219 and @xmath220 . ) by shrinking the two boundary components we produce singularities ( or regular points ) of orders @xmath221 and @xmath222 .
( see the left part of figure [ fig : twoholes ] . )
creating a pair of holes ] the type of boundary , figure eight or pair of holes , and the associated angles is what we referred to above as `` singularity structure on the boundary '' .
we get in this way closed surfaces @xmath223 that belong to some stratum @xmath224 , for @xmath225 .
we write @xmath226 and @xmath227 .
we will say that the surface @xmath228 whose connected components are @xmath229 belongs to @xmath230 .
we say that @xmath228 belongs to the _ principal boundary _ of @xmath28 determined by @xmath54 and that @xmath230 is the corresponding _ principal boundary stratum_. all this data ( topological type of complementary regions , cyclic order , boundary stratum , e.t.c . )
are the same for any configuration of the same topological type @xmath54 .
in fact the topological type is characterized by this data . by shrinking a closed saddle connection ,
the singular point it is based at might become a regular point with total angle @xmath155 . in this case
the regular point will be considered as a marked point of order @xmath231 .
so @xmath232 can contain one or two @xmath231 .
the procedure of shrinking saddle connections can be reversed .
we summarize the description form @xcite in case where @xmath28 has only one connected component @xmath37 . start with a ( maybe disconnected ) surface @xmath233 in @xmath230 and
call the connected components @xmath234 .
choose some holonomy vector @xmath163 in @xmath194 .
there are two types of surgery .
figure eight surgery _ where we start with a singularity or a marked point of order @xmath235 , choose @xmath236 such that @xmath237 and then metrically create a figure eight boundary that consists of two saddle connections in direction @xmath163 that are of length @xmath238 and are based at the same singularity .
there will be two sectors of angles @xmath208 and @xmath209 ( see figure [ fig : figureeight ] ) .
pair of holes surgery _ where we start with two points that are either a singularity or a marked point of orders @xmath239 and @xmath240 and then create a boundary saddle connection at each singularity .
the pair of holes boundary has two boundary singularities of angles @xmath216 and @xmath217 ( see figure [ fig : twoholes ] ) . by performing an appropriate surgery on each @xmath234
we obtain surfaces @xmath241 that are homeomorphic to the surfaces @xmath207 and have the same type of singularities in the interior and on the boundary .
we then take @xmath65 cylinders @xmath242 with a marked point on each boundary .
we finally combine the surfaces @xmath241 and the cylinders @xmath242 in the way prescribed by the topological type @xmath54 to produce a surface @xmath243 in @xmath201 .
we do this by identifying pairs of boundary components by an isometry that identifies boundary singularities .
the boundary components give rise to a configuration of closed saddle connections of topological type @xmath54 .
in fact each surface in @xmath201 can be produced in this way .
the parameters used to produce surfaces in @xmath201 are the following : * a maybe disconnected surface in @xmath230 . * a holonomy vector @xmath163 in @xmath194 . * a combinatorial constant @xmath49 that only depends on the configuration .
there is a @xmath244 correspondence between the surfaces in @xmath201 and the surfaces in @xmath68 .
this is mainly due to the facts that at a zero of order @xmath245 there are @xmath246 sectors of angle @xmath155 where we can produce a saddle connection in the direction of @xmath163 and to possible symmetries of the surface in @xmath201 ( see 13.3 . of @xcite ) . * the heights @xmath247 of the @xmath65 cylinders @xmath248 ( the width
is given by @xmath238 ) .
* for each cylinder @xmath248 a twist parameter @xmath249 that describes the relative position of the marked points on the two boundary components .
if @xmath28 has more than one connected component than the correspondence between the thick part @xmath201 of a connected component @xmath37 of @xmath33 and the principal boundary @xmath230 is slightly more complicated .
for example , to construct a surface @xmath243 in @xmath201 , where @xmath37 is a hyperelliptic component , we must start with a surface @xmath233 in the principal boundary such that all connected components @xmath234 of @xmath233 are hyperelliptic surfaces .
so only the hyperelliptic components of some strata are in the principal boundary of @xmath37 .
it might even happen that only part of a connected component of a stratum is in the principal boundary .
we still denote the principal boundary of @xmath201 by @xmath230 , although @xmath230 might be the union of ( parts of ) connected components of some strata .
we remark in passing that there are also some parameters used to describe the surgeries but this does not affect our computations . if @xmath250 , then to parametrize the @xmath65 tori we will always replace @xmath251 by @xmath252 .
so the heights @xmath253 are in the cone @xmath254 given by the conditions @xmath255 , for @xmath256 , and @xmath257 . for a given @xmath180
we define @xmath258^q\}.\qquad\end{gathered}\ ] ] in what follows , @xmath238 will always be small , but if no specific restriction on @xmath238 is needed we will write @xmath259
. we write @xmath260 for the measure @xmath261 we refer to the elements of @xmath262 as tori as we obtain a torus with @xmath65 marked points by joining the @xmath65 cylinders .
we write @xmath263 for the subset of @xmath262 of area @xmath264 tori , meaning that they satisfy the condition @xmath265 .
we remark in passing that @xmath262 can be interpreted as the @xmath195 -neighborhood of the `` cusp '' of the moduli space of flat tori with @xmath65 marked points except that the marked points are already named by the way we parametrize them .
this is in contrast to the cusp of the usual moduli space @xmath266 where the marked points can be arbitrarily named .
we denote by @xmath267 ( resp .
@xmath268 ) the measure on @xmath28 ( resp .
@xmath230 ) .
it is shown in @xcite that @xmath269 let @xmath5 be a translation surface in @xmath33 . for @xmath270 a positive real number we denote by @xmath271 the surface we get by multiplying the flat metric on @xmath5 by @xmath270 .
equivalently , if we represent @xmath5 by an abelian differential @xmath21 with respect to some complex structure , then @xmath272 corresponds to the abelian differential @xmath273 with respect to the same complex structure .
note that we have in particular @xmath274 .
if @xmath275 is a subset of @xmath33 then the cone @xmath276 is defined to be @xmath277 if we denote by @xmath278 the measure on @xmath33 induced by the measure @xmath267 then we have @xmath279 so in particular one has @xmath280 one has analogous statements for @xmath68 and @xmath281 with their induced measures denoted by @xmath282 and @xmath283 .
note that there are @xmath284 complex parameters for @xmath262 and that one has @xmath285 .
so by writing @xmath286 we have @xmath287 we recall for completeness that if @xmath26 with @xmath9 for @xmath27 , then @xmath288
we will describe in this section the method used to compute the siegel
veech constants introduced in section [ sec : results ] .
consider a stratum @xmath28 of abelian differentials where @xmath26 and @xmath9 for @xmath27 .
suppose that @xmath37 is a connected component of @xmath33 and that @xmath54 is an admissible topological type of configurations for @xmath37 containing exactly @xmath72 cylinders .
we denote the corresponding principal boundary stratum by @xmath230 .
recall that @xmath201 is the set of translation surfaces in @xmath37 that contain exactly one configuration of type @xmath54 of length smaller than @xmath195 but do not contain another closed saddle connection of length smaller than @xmath195 . to simplify notation
we will write from now on @xmath289 for @xmath201 .
suppose that @xmath290 is one of the above mentioned counting functions that counts configurations on @xmath5 of length smaller than @xmath195 and of topological type @xmath19 .
we want to evaluate @xmath291 .
the decomposition of a surface @xmath292 as @xmath65 tori with marked points on the boundary and a surface @xmath228 in @xmath230 on which one performs surgeries ( @xmath49 choices ) gives the following parametrisation of the cone @xmath293 : before performing the surgeries , a surface @xmath5 in @xmath293 is made up of @xmath294 and @xmath295 for some scalars @xmath296 where @xmath297 and @xmath298 .
the conditions are the following : + ( i ) the total area @xmath299 of @xmath5 satisfies @xmath300 ; + the area @xmath264
surface @xmath301 is in @xmath289 so the waist curve of @xmath302 is smaller than @xmath195 which means that the waist curve of @xmath303 is smaller than @xmath304 .
so the second condition is : + ( ii ) @xmath303 has to lie in @xmath305 , where @xmath306 .
we have @xmath307 where @xmath308 .
note that if we construct @xmath309 as described above then @xmath5 contains exactly one configuration of type @xmath54 and of width less than @xmath195 , so @xmath290 can be replaced by the weight with which we count configurations .
this integral is a simplification of the integral from @xcite , page 133 . in
@xcite the statement is about the volume of @xmath293 ( in our notaton ) as the authors only count configurations with weight one , so @xmath310 on @xmath289 .
fix an admissible configuration @xmath54 in a connected component @xmath37 of a stratum @xmath33 that comes with @xmath72 cylinders and let @xmath60 be a real number @xmath58 .
we choose and fix one of the named closed saddle connection of @xmath54 that bounds a cylinder and call this cylinder the first one .
we denote by @xmath311 the number of configurations on @xmath5 of type @xmath54 of length at most @xmath195 , counted with weight the p - th power of the area of the first cylinder .
we denote the corresponding siegel - veech constant by @xmath312 .
as described above , we decompose a surface @xmath5 in @xmath293 as @xmath313 and @xmath295 .
we use the usual parameters @xmath314 to parametrize @xmath315 .
we assume the notation chosen so that the first cylinder of @xmath5 corresponds to the first cylinder of @xmath303 .
so the area of the first cylinder of @xmath5 is @xmath316 , where @xmath317 .
we use the following weight : @xmath318 this weight is invariant under scaling of @xmath5 and if the area @xmath299 of @xmath5 equals @xmath264 then it reduces to @xmath319 which is the @xmath60-th power of the area of the first cylinder of @xmath5 . a surface @xmath309 contains exactly one configuration of type @xmath54 of width at most @xmath195 , so we need to evaluate the integral of equation for @xmath320 the first step to do this is to show that [ lem : cusp - p ] @xmath321 where for @xmath93 the denominator is by convention equal to @xmath264 .
for @xmath93 we have @xmath322 , so the computation is a simplified version of the general case .
assume that @xmath250 .
we will first integrate over the domain @xmath323 , still using the parameters @xmath314 for @xmath324 . to have a weight that is invariant under scaling of @xmath243 and that becomes @xmath325 for an area one torus @xmath303
, we use the weight @xmath326 the domain of integration for @xmath324 is as follows : the area @xmath327 of @xmath243 satisfies @xmath328 .
we have @xmath329 , so the length @xmath330 of the waist curve of @xmath331 is smaller than @xmath195 , which gives @xmath332 .
( see figure [ fig : domain : of : integration:2 ] . )
domain of integration for @xmath333 @xmath334^q}{\mathrm{d}}t_1\dots { \mathrm{d}}t_{q } .\end{gathered}\ ] ] the volume of the cone @xmath335 is @xmath336 and the volume of the cube @xmath337^q$ ] is @xmath338 . using the change of variables @xmath339
, the right hand side of equation then evaluates to @xmath340 where @xmath341 is the beta function defined by .
we get the right hand side using relations and .
the weight @xmath342 satisfies @xmath343 and , if @xmath344 , @xmath345 , so @xmath346 which completes the proof of the lemma . to continue the proof of theorem [ th : area - p ] , we evaluate equation .
we use lemma [ lem : cusp - p ] with @xmath347 and integrate the function @xmath348 that is the remaining part of @xmath311 .
so equation becomes @xmath349 where @xmath350 using polar coordinates @xmath351 , @xmath352 we get @xmath353 using @xmath354 we get @xmath355 using the fact that for @xmath309 , @xmath356 is invariant under scaling of @xmath5 and reduces to the initial definition of @xmath357 if @xmath358 , we can show as in the proof of lemma [ lem : cusp - p ] that @xmath359 where @xmath360 .
and so @xmath361 using equation we conclude that @xmath362 note that if we define @xmath363 in the same way as @xmath311 , except that we use the area of the @xmath364-th cylinder , for @xmath365 , then we have @xmath366 .
it follows that @xmath367 , which completes the proof of theorem [ th : area - p ] .
\(a ) we note that the volume of the stratum @xmath68 of disconnected surfaces was computed in @xcite , equation ( 12 ) ; writing @xmath227 and @xmath368 , we have @xmath369 ( b ) as an example we consider the moduli space of tori . to have a configuration we need to mark one regular point .
the only possible topological type @xmath54 of configuration is a closed saddle connection based at this regular point .
we get , using lemma [ lem : cusp - p ] for @xmath370 and @xmath93 , @xmath371 which is the well - known factor for the proportion of coprime lattice points in @xmath372 .
fix an admissible topological type of configuration @xmath54 for a connected component @xmath37 of a stratum @xmath33 that comes with @xmath72 cylinders . recall that @xmath86 denotes the @xmath60-th power of the total area of the periodic region ( union of the cylinders ) on @xmath5 coming from a configuration of topological type @xmath54
whose length is at most @xmath11 .
the argument is a special case of the argument in section [ sec : mean - area ] . by decomposing as before a surface @xmath5 in @xmath293 as @xmath313 and @xmath295
we use the weight @xmath373 so by taking @xmath370 in lemma [ lem : cusp - p ] , @xmath374 we then have @xmath375 where @xmath376 satisfies relation .
we conclude as in section [ sec : mean - area ] that @xmath377 it suffices to apply relation .
fix an admissible configuration @xmath54 in a connected component @xmath37 of a stratum @xmath33 that comes with @xmath72 cylinders and let @xmath43 be a real parameter . recall that @xmath378 denotes the number of configurations on @xmath5 of type @xmath54 of length smaller than @xmath195 and such that the total area of the @xmath65 cylinders is at least @xmath45 ( of the area one surface @xmath5 ) .
we use a modification of the argument from section [ sec : mean - area ] . here
we consider the subset of @xmath289 consisting of surfaces @xmath5 whose area of the periodic part is at least @xmath45 ( of the area @xmath264 surface @xmath5 ) .
construct @xmath5 in the cone of this set using @xmath294 and @xmath295 for some scalars @xmath296 and @xmath297 and @xmath379 .
we need to evaluate the integral in equation with an additional condition : when scaling @xmath5 by @xmath380 we get a surface of area one that satisfies @xmath381 we count a configuration that satisfies this additional constraint with weight @xmath264 , so equation becomes @xmath382 where @xmath347 .
using relation , the right hand side becomes @xmath383 where @xmath384 the domain of integration for @xmath385 and @xmath386 is described in figure [ fig : domaininte ] .
domain of integration . ] using polar coordinates @xmath351 , @xmath352 and setting @xmath387 we get @xmath388 using the change of variables @xmath354 we get @xmath389 where @xmath390 is the incomplete beta function as defined in equation .
note that we have @xmath391 , so up to the same constant , the integral used to compute @xmath392 , resp @xmath55 is given by @xmath393 , resp @xmath394 ( or equivalently @xmath395 , see equation ) , so @xmath396 we proved theorem [ th : big - area ] of section [ sec : results ] , where , to expand @xmath110 , we use lemma [ lem : combi4 ] . fix an admissible configuration @xmath54 in a connected component @xmath37 of a stratum @xmath33 that comes with @xmath72 cylinders and let @xmath45 be a parameter that satisfies @xmath397 .
we choose and fix one of the named closed saddle connection of @xmath54 that bounds a cylinder and call this cylinder the first one .
recall that @xmath398 is the number of configurations of type @xmath54 of length at most @xmath11 and such that the area of the first cylinder is at least @xmath45 ( of the area one surface @xmath5 ) .
suppose that we have @xmath72 cylinders .
the argument is as in section [ sec : allcylindersp ] with the following modification : we replace `` @xmath399 '' by `` area of the first cylinder '' in the condition @xmath400 .
denote by @xmath401 the subset of @xmath305 of tori @xmath303 such that the area of the first cylinder of @xmath303 is at least @xmath402 . using the usual parameters @xmath403 , a torus @xmath243 is in the cone of @xmath401 if the area one torus @xmath404 is in @xmath401 , so the parameters for @xmath243 have the additional constraint @xmath405 . to compute the volume of @xmath401 we proceed as in the proof of lemma [ lem : cusp - p ] for @xmath406 .
the only difference is that we replace the condition @xmath407 by @xmath408 .
so @xmath253 is in a cone whose volume is @xmath409 .
so the volume is as in equation except that we have an extra factor @xmath410 .
so we need to integrate @xmath411 using polar coordinates @xmath351 , @xmath352 followed by the change of variables @xmath412 we get @xmath413 note that @xmath414 , so @xmath415 as claimed .
( recall that @xmath416 . ) for @xmath93 we have @xmath417 ( see theorem [ th : big - area ] ) .
let @xmath54 be an admissible configuration for a connected component @xmath37 that comes with at least two cylinders .
choose ( and fix ) two cylinders and let @xmath116 .
recall that we denote by @xmath117 the number of configurations of length at least @xmath11 such that the area @xmath118 of the first cylinder is at least @xmath119 and such that the area @xmath120 of the second cylinder is at least @xmath121 .
we denote by @xmath122 the corresponding siegel
veech constant . to simplify notation we will write @xmath123 instead of @xmath124 .
we proceed as in the proof of theorem [ th : vorobets : constant ] ( see section [ sec : onecylindersp ] ) . using the same notation , we have the condition that the area @xmath418 of the first cylinder of @xmath419 is at least @xmath119 and the area @xmath420 of the second cylinder of @xmath419 is at least @xmath421 .
this means that the area @xmath118 of the first cylinder of @xmath303 satisfies @xmath422 and the area @xmath120 of the second cylinder is at least @xmath423 which gives @xmath424 , where @xmath402
. denote by @xmath425 the subset of tori in @xmath305 that satisfies these conditions . using the usual parameters @xmath403 , a torus @xmath243 is in the cone @xmath426 if and only if @xmath427 is in @xmath425 . so the areas @xmath118 and @xmath120 of the first and second cylinders of @xmath303 must satisfy @xmath428 and @xmath429 .
so we get @xmath430 equation @xmath431 has a solution if and only if @xmath432 which can be written as @xmath433 so we need to modify @xmath434 : @xmath435 equation @xmath436 has a solution if and only if @xmath437 .
this translates into @xmath438 , where @xmath439 , which in turn becomes @xmath440 we also have @xmath407 and @xmath441 .
but for all possible @xmath442 we have @xmath443 , @xmath444 , and @xmath445 , so @xmath446 can take all values between @xmath447 and @xmath448 and @xmath449 can take all values between @xmath450 and @xmath451 .
the computation of the volume of @xmath425 is as in the proof of lemma [ lem : cusp - p ] for @xmath406 , except that we replace @xmath452 by @xmath453^{q-1}}{1-x } \frac{h^{q-1}}{(q-1 ) ! } = \frac{(1-a_2)^{q-1}}{1-x } \frac{h^{q-1}}{(q-1)!},\end{gathered}\ ] ] where @xmath454 .
so the volume is as in equation except that we have an extra factor @xmath455 .
so we need to integrate @xmath456 note that up to the factor @xmath457 this is @xmath458 as defined in .
the integral used to compute @xmath459 is @xmath460 , so we conclude @xmath461 using equation and @xmath462 we find @xmath463
consider an admissible topological type @xmath54 of configuration for some connected component @xmath37 of some stratum @xmath28 , where @xmath26 satisfies @xmath9 , for @xmath27 .
we denote by @xmath464 the number of cylinders that come with @xmath54 . in this section
we prove theorem [ prop : qmax : dim : ab ] , so we look for a topological type of configuration @xmath54 ( that is admissible for some connected component @xmath37 ) that maximizes @xmath465 where for the last equality we used corollary [ cor : mean - area - p ] and @xmath466 . for a given stratum @xmath28 , we start by determining the maximal possible number of cylinders @xmath467 that can come from an admissible topological type of configuration for @xmath28 : @xmath468 it is shown in @xcite that topological types of configurations can be constructed by creating singular points of the following three types : 1 . a cylinder , followed by @xmath469 surfaces @xmath207 of genus @xmath470 with figure eight boundary , followed by a cylinder
. see figure [ fig : type i ] for @xmath471 and @xmath472 , @xmath473 .
we say that the newborn singularity is of type @xmath474 .
2 . a cylinder , followed by @xmath475 surfaces @xmath207 of genus @xmath470 with figure eight boundary , followed by a surface @xmath476 of genus @xmath477 with a pair of holes boundary . see figure [ fig : type ii ] for @xmath478 and @xmath472 , @xmath479 .
we say that the newborn singularity is of type @xmath480 . for @xmath481
we just have a cylinder followed by a surface with a pair of holes boundary .
we might also reverse the order : a pair of holes torus followed by @xmath475 figure eight tori followed by a cylinder .
3 . a singularity of type @xmath482 , which is obtained from a singularity of type @xmath480 by replacing the cylinder by a surface with a pair of holes boundary . as for surfaces of type @xmath480 there might be no surface with a figure eight boundary . see figure [ fig : type iii ] where all of the surfaces are of genus @xmath264 and where we have two surfaces with figure eight boundary . ] ] ] counting angles , it is shown in @xcite that the order of the newborn zeros is as follows : 1 . to create a zero of type @xmath474 one uses @xmath469 figure eight boundaries that were created at zeros of orders @xmath443, ,@xmath483 ( a zero of order @xmath231 being a regular point ) .
the zero then has order @xmath484 .
all orders bigger or equal to @xmath140 are possible .
2 . to create a zero of type @xmath480 one uses a figure eight boundary that was created at a zero of order @xmath485 .
if there is no figure eight boundary involved then the newborn zero has order @xmath486 . if there are @xmath469 figure eight boundaries involved that were created at zeros of orders @xmath443,
,@xmath483 then the newborn zero has order @xmath487 .
all orders bigger or equal to @xmath264 are possible .
3 . to create a zero of type @xmath482 we use two pair of holes boundaries created at zeros of orders @xmath488 and @xmath489 .
if there are no figure eight boundaries involved then the order of the newborn zero is @xmath490 . if there are @xmath469 figure eight boundaries involved that were created at zeros of orders @xmath443,
,@xmath483 then the newborn zero has order @xmath491 .
all orders bigger or equal to @xmath140 are possible . for a given small @xmath163 one can create all of the boundaries of the surfaces involved in the above construction by an appropriate figure eight surgery or a pair of holes surgery ( see figures [ fig : figureeight ] and [ fig : twoholes ] ) such that the boundaries all have holonomy vector @xmath163 . by arranging the blocks in a cyclic order and
identify boundary components we create an admissible topological type of configuration of homologous saddle connections where each saddle connection is based at a newborn singularity of one of the three types .
there is only the following obstruction : one needs to either use at least one surface with a pair of holes boundary or , if one only uses surfaces with figure eight boundaries , then one needs to use at least one cylinder .
each cylinder is bounded by two saddle connections , so to have a one to one correspondence we think of the cylinder as being cut into two parts by the central waist curve .
each half cylinder then has a saddle connection @xmath163 on the boundary ( that is not the waist curve ) that joins a saddle @xmath492 to itself .
we say that @xmath492 accounts for this half - cylinder .
a zero of type @xmath474 ( see figure [ fig : type i ] ) accounts for two half - cylinders ( so for one cylinder ) , a zero of type @xmath480 ( see figure [ fig : type ii ] ) accounts for one half of a cylinder , and a zero of type @xmath482 ( see figure [ fig : type iii ] ) does not account for any half cylinder .
note that only singularities of type @xmath480 can have order @xmath264 .
it follows that if for @xmath493 we define @xmath494 then we have @xmath495 . to maximize the number of cylinders we construct the zeros of orders greater than or equal to @xmath140 by zeros of type @xmath474 .
the other zeros of order @xmath264 must be created by zeros of type @xmath480 ( coming from a cylinder followed by a torus with a pair of holes boundary ) .
this works well if there is an even number of zeros of order @xmath264 in which case we constructed a configuration with @xmath496 cylinders .
if there is an odd number of zeros of order @xmath264 then the construction of the zeros of order @xmath264 ends with a surface with a pair of holes boundary . in this case
we are obliged to construct one of the zeros of order greater than @xmath264 by a surface of type @xmath480 , so we constructed a configuration with @xmath497 cylinders .
we showed [ prop : bound : for : q : max ] consider a stratum @xmath28 , where @xmath26 satisfies @xmath9 , for @xmath27 .
we have @xmath498,\ ] ] where the square brackets denote the integer part of a number .
the proposition says that it is possible to find in each stratum @xmath28 a topological type of configuration @xmath54 such that @xmath499 is as stated .
it is not possible to find a topological type @xmath54 in each _ connected component _ of @xmath28 with @xmath467 cylinders .
we next show that given any connected component @xmath37 of a stratum @xmath28 and any admissible topological type @xmath54 of configuration for @xmath37 , @xmath134 let @xmath136 . denoting by @xmath500 the length of @xmath158 we have @xmath501 we determine @xmath502 where @xmath503 denotes the set of permutations of @xmath504 .
first let us prove that a partition @xmath158 containing at least one entry @xmath264 is not maximizing .
indeed , if there is at least one _ pair _ of entries @xmath264 we can modify the initial partition @xmath158 by replacing the elements @xmath505 with a single entry @xmath140 .
this does not change the value of @xmath467 , but decreases the denominator in the ratio @xmath506 . if there is a _ single _
entry @xmath264 in the partition @xmath158 , then @xmath496 is not an integer .
we can modify @xmath158 by deleting the entry @xmath264 and increasing some other entry by @xmath264 .
this operation does not change the value of @xmath467 , but decreases the denominator in the ratio @xmath506 .
thus we have proved that all of the entries of the partition maximizing the ratio @xmath507 are greater then or equal to @xmath140 .
the formula for such partitions simplifies to @xmath508 .
now note that for any @xmath135 the function @xmath509 is strictly decreasing .
thus , among all partitions of @xmath504 with entries strictly greater than one we have to chose the one maximizing @xmath500 .
this is the partition @xmath510 where the order @xmath140 appears @xmath139 times . for this partition
we get @xmath511 and so @xmath128 is bounded from above by @xmath512 as claimed .
it suffices to show that for each genus @xmath2 there is an admissible topological type of configuration @xmath54 for a connected component @xmath37 of the stratum @xmath513 ( @xmath139 zeros of order @xmath264 ) such that @xmath128 attains the upper bound @xmath512 .
note that for the upper bound @xmath512 for @xmath128 we used @xmath514 .
as we explained in the proof of the relation , the only way to obtain @xmath515 is to only use zeros of type @xmath474 ( a cylinder followed by a torus with a figure eight boundary that was created at a regular point followed by a cylinder ) . doing this we obtain a surface in @xmath513 with a configuration that comes with @xmath139 cylinders
. by lemma 14.2 in @xcite the surface in @xmath513 we constructed that comes with a maximizing configuration has odd parity of spin structure , so this surface is in @xmath516 .
( we recall in the next section the classification of connected components . )
proposition [ prop : qmax : dim : ab ] is proved .
this section provides an answer to the following question of alex eskin and alex wright : is it possible to find in each connected component of a stratum an admissible topological type of configuration whose complementary regions are tori ( with boundary ) and cylinders .
motivations for this problem can be found in @xcite .
the answer depends on the connected component , so we need to recall the classification of connected components for strata @xmath28 of abelian differentials from @xcite .
some connected components are characterized by the fact that they only contain hyperelliptic surfaces . for a surface @xmath5 in a stratum @xmath517 where all @xmath151 are even one
has the notion of _ parity of spin structure _ that is either @xmath231 or @xmath264 .
we then have let @xmath518 be a stratum of abelian differentials on a surface of genus @xmath519 .
the strata @xmath520 and @xmath521 are the only strata to have a hyperelliptic component @xmath522 , resp . @xmath523 . apart from these hyperelliptic components we have : 1 .
if at least one of the @xmath151 is odd then there is only one non - hyperelliptic component .
2 . if all of the @xmath524 are even then there are two non - hyperelliptic components , @xmath525 with even and @xmath526 with odd parity of spin structure .
the classification in @xcite also covers @xmath527 but we will not need this . [ prop : alex ] let @xmath528 denote a connected component of a stratum of abelian differentials on a surface of genus @xmath529 . then
: 1 . if @xmath528 is hyperelliptic ( so @xmath522 or , if @xmath139 is even , @xmath523 ) then it is not possible to find an admissible topological type of configuration whose complementary regions are only tori ( with boundary ) and cylinders .
2 . if @xmath2 is even and @xmath530 then we can find a topological type of configuration whose complementary regions are tori , cylinders and one surface of genus two .
but it is not possible to only have tori and cylinders .
3 . in all remaining connected components one can explicitly construct an admissible topological type of configuration whose complementary regions are tori and cylinders . first note that a configuration containing only tori and cylinders , or tori , cylinders , and at most one genus two surface does not occur in hyperelliptic strata , since a hyperelliptic surface can contain at most two closed homologous saddle connections , which rules out @xmath529 .
lemma @xmath531 in @xcite describes precisely configurations in hyperelliptic components of strata .
we suppose for what follows that @xmath528 is not a hyperelliptic component .
recall the description of singularities of types @xmath474 , @xmath480 , and @xmath482 from the previous section .
if apart from cylinders , we only use surfaces of genus @xmath264 , then we must do figure eight surgeries and pair of holes surgeries at regular marked points .
so we have : 1 .
a zero of type @xmath474 has order @xmath532 , where @xmath469 is the number of tori with figure eight boundaries ( see figure [ fig : type i ] ) .
2 . a zero of type
@xmath480 has order @xmath533 , where @xmath475 is the number of tori with figure eight boundaries ( see figure [ fig : type ii ] ) .
3 . a zero of type @xmath482 has order @xmath534 , where @xmath475 is the number of tori with figure eight boundaries ( see figure [ fig : type iii ] ) .
assume that @xmath158 contains at least one odd @xmath151 , so @xmath535 with @xmath71 and @xmath536 ( @xmath370 corresponds to the case when all @xmath524 are odd ) .
we construct blocs of surfaces that contain zeros of type @xmath480 of orders @xmath537 , , @xmath538 .
note that since @xmath539 is even , the number @xmath270 of odd @xmath151 is even , so in our construction the first and last surface is a cylinder . if there is at least one zero of even order then we construct in addition blocs of surfaces that contain zeros of type @xmath474 of orders @xmath540, ,@xmath541 . for this construction
the first and last surface is also a cylinder . arranging the blocs in a cyclic order and identifying cylinders we get in each case an admissible topological type of configuration for @xmath28 whose complementary regions are cylinders and tori .
suppose now that all @xmath151 in @xmath28 are even , so we can not have a zero of type @xmath480 .
we then either have only zeros of type @xmath474 or only zeros of type @xmath482 as having zeros of types @xmath474 and @xmath482 necessarily implies that we have a zero of type @xmath480 .
it is easy to verify ( @xmath542 in @xcite ) that if we only have zeros of type @xmath474 then the corresponding surface has an odd parity of spin structure and if we only only have zeros of type @xmath482 then the parity of the resulting surface is the parity of @xmath139 .
so for @xmath2 odd we are done , but for even @xmath2 the parity of the spin structure is @xmath264 in both cases .
if we only have zeros of type @xmath482 but replace one of the pair of holes boundary tori with a genus @xmath140 surface with a pair of holes boundary then the parity of the spin structure of the resulting surface is the parity of @xmath2 .
so for even @xmath2 we constructed a surface of parity @xmath231 .
this completes the proof .
we recall some well known facts about the beta function and incomplete beta function .
we would like to thank anton zorich for having initiated the present work .
he asked most of the questions and initiated some answers .
we thank alex wright for the formulation of the problem of section [ sec : simple - abelian ] , and for some comments on typos .
we thank the anonymous referee(s ) for the careful reading of the manuscript .
both authors thank anr `` geodym '' for financial support .
eskin , a. , kontsevich , m. , zorich , a. : sum of lyapunov exponents of the hodge bundle with respect to the teichmller geodesic flow .
publications mathmatiques de lihs , springer berlin heidelberg ( 2013 ) vorobets , y. : periodic geodesics on generic translation surfaces algebraic and topological dynamics , s. kolyada , yu . i. manin and t. ward ( editors ) , contemporary math .
385 , amer .
soc . , providence , ri , ( 2005 ) zorich , a. : flat surfaces .
frontiers in number theory , physics and geometry , vol . 1 : on random matrices , zeta functions and dynamical systems .
p. cartier ; b. julia ; p. moussa ; p. vanhove ( editors ) , springer verlag ( 2006 )
|
an abelian differential gives rise to a flat structure ( translation surface ) on the underlying riemann surface . in some directions
the directional flow on the flat surface may contain a periodic region that is made up of maximal cylinders filled by parallel geodesics of the same length .
the growth rate of the number of such regions counted with weights , as a function of the length , is quadratic with a coefficient , called siegel veech constant , that is shared by almost all translation surfaces in the ambient stratum .
we evaluate various siegel
veech constants associated to the geometry of configurations of periodic cylinders and their area , and study extremal properties of such configurations in a fixed stratum and in all strata of a fixed genus .
|
very neutron - rich nuclei may emit one or more neutrons when they disintegrate via @xmath0 decay .
this is the so - called @xmath0-delayed neutron ( @xmath15 ) emission process , which is energetically allowed when the q@xmath16-value of the decay exceeds the neutron separation energy ( s@xmath17 ) of the daughter nucleus .
the @xmath15-emission has been experimentally determined for about 230 neutron rich nuclei , spanning from @xmath1he up to @xmath2la @xcite .
most of these measurements took advantage of the large fission yields around the two fission peaks at @xmath18 and @xmath19 .
however , @xmath0-delayed neutron emission has remained essentially inaccessible for nuclei heavier than @xmath4 , where only a minuscule value of 0.007% has been reported for the @xmath15-emission probability of @xmath3tl @xcite . because of the scarce or non - existent @xmath15 data in the heavy mass region , rapid neutron capture @xmath20-process @xcite calculations have to rely entirely on theoretical models @xcite .
however , the performance of such models for reproducing the features of the @xmath0-decay in @xmath20-process waiting - point nuclei has been tested with experimental data only in the two shell - closures at @xmath21 and @xmath22 in measurements such as reported in @xcite .
the comparison is much more limited in the @xmath6 region , where only half - lives in the neighborhood of the doubly magic @xmath23pb were available ( see e.g. ref .
@xcite ) .
the neutron - rich nuclei south " of @xmath23pb are difficult to measure experimentally because of the very small production cross - sections and the large background conditions induced by the heavy primary beam . in the present work it was possible to produce and identify reliably secondary neutron - rich nuclei in the region
south - east " of @xmath23pb in the chart of nuclides thanks to the high - energy ( 1 gev / u ) @xmath7u beams available at the gsi facility .
extended motivation and results for this experiment were recently published in @xcite .
here we present more details on the experimental apparatus in section [ sec : setup ] , the analysis methodology and results , which are reported in section [ sec : analysis ] and finally sections [ sec : discussion ] and [ sec : conclusions ] summarize the main results and conclusions .
the present measurements were carried out at the gsi helmholtz center for heavy ion research . a @xmath7u beam was accelerated to an energy of 1 gev / u by the unilac linear accelerator coupled to the sis 18 synchrotron .
the average beam intensity was @xmath24 ions / spill , with a pulsed beam structure of 1 s spill duration ( sis extraction ) and a repetition cycle of 4 s. the beam impinged onto a @xmath25be target with a thickness of 1.6 g/@xmath26 at the entrance of the fragment separator ( frs ) @xcite .
the selection of the ions of interest , from this point to the detection system , was done with the @xmath27-method using the frs as an achromatic spectrometer .
neutron - rich nuclei south - east " of @xmath23pb in the chart of nuclides were produced using two frs settings centered on @xmath28hg and @xmath29tl . in order to minimize the number of ionic charge states of the secondary beam a niobium layer with a thickness of 223 mg/@xmath26
was placed behind the be - target .
in addition , a homogeneous al degrader with a thickness of 2.5 g/@xmath26 was placed at the first frs focal plane ( s1 ) in combination with thick cu slits , which served to reduce the contribution of fission fragments and primary beam charge - states with initially similar magnetic rigidity ( @xmath30 ) as the setting isotope . a wedge - shaped al degrader with a thickness of 3874 mg/@xmath26 was employed as achromatic degrader at the intermediate focal plane ( s2 ) , ( see fig .
[ fig : frsscheme ] ) .
ion species were identified by means of standard frs tracking detectors .
two plastic scintillators located at s2 and at the final focal plane ( s4 ) were used to measure the time - of - flight ( @xmath31 ) of the ions .
two systems of time projection chambers ( tpcs ) @xcite placed at s4 and s2 allowed us to determine accurate @xmath30 values for each ion by measuring their trajectory with respect to the central fragment .
the measured @xmath31 in combination with the @xmath30 provided the necessary information to calculate the mass - to - charge ratio ( a / q ) on an event - by - event basis .
the resolution thus obtained in a / q was 2.5@xmath32 fwhm . in order to determine the atomic number ( @xmath33 ) , two fast multiple sampling ionization chambers ( musics ) @xcite were placed in the s4 experimental area .
although the detected nuclei were mainly bare , h- and he - like charge states were also detected in musics .
the latter events were corrected by combining the information of the two musics and calculating the energy loss in the s2 degrader following the method applied in previous studies in this mass region @xcite .
in addition , it was needed to treat the gain fluctuations in the music detectors caused by the variations of the temperature in the experimental hall , and consequent changes in gas pressure , which were also corrected numerically @xcite .
the final resolution obtained in z for the pb - bi region was of @xmath34 fwhm .
finally , the z versus a / q particle identification diagram ( pid ) was experimentally validated with a dedicated run .
@xmath35bi ions were implanted into a passive plastic stopper and the decay of well known isomeric transitions @xcite were measured with hpge detectors .
the pid obtained , including nuclei produced in both frs settings with all the statistics accumulated during the experiment , is shown in fig .
[ fig : pid ] .
all identified nuclei were already reported from previous experiments @xcite .
the secondary beam of neutron - rich isotopes was focused at the final focal plane ( s4 ) of the frs .
a third al degrader with variable thickness was adjusted in order to slow down the ions of interest and to implant them into the central region of an active stopper named silicon implantation beta absorber " ( simba ) @xcite .
simba enabled position and energy measurement of heavy charged ions as well as charged particles from @xmath36 and @xmath0 decays .
the main difference between the simba system employed here with respect to previous versions was the smaller number of si detectors required to stop and implant the ions of the present experiment , which had a higher atomic number ( @xmath37 ) than those measured in the past ( @xmath38 , @xcite ) .
in addition , the geometry and overall size of simba were also modified in order to optimize neutron detection ( see below ) . the present simba version consisted of a stack of nine highly segmented si detectors ( see figs . [
fig : simba2 ] and [ fig : simba ] ) .
the first two layers , called xy - tracking detectors , consisted of single - sided si - strip detectors ( sssd ) with their strips in orthogonal orientation with respect to each other . they were used for determining accurately the ion position in the transversal plane of the beam .
a center - of - gravity method applied to the charge shared over all the strips allowed us to determine the ion position with an accuracy of @xmath39 @xcite , corresponding to one pixel in a silicon layer of simba .
the implant and decay sensitive region consisted of two sssd layers ( front absorbers ) , three double sided silicon stripped detectors ( dsssd ) designated as implantation layers a , b and c , and two sssd layers ( rear absorbers ) .
the energy deposited by the ions along these seven si detectors was used to detect whether the ion was implanted or if it punched through , as well as to determine the corresponding implant layer or depth .
the segmentation of the dsssds was 60-fold in x and 40-fold in y direction , with a strip width of 1 mm .
[ f3-numimplants ] shows the total amount of implanted ions for each isotope in the dsssds of simba .
the energy deposited by @xmath0 decays in each si - electrode showed , as expected , a continuum spectrum which extended up to @xmath402 - 3 mev .
the energy deposited in the dsssds was determined from the response of the y - strips , which were readout via logarithmic preamplifiers .
the latter allowed to clearly separate the ion implants and decays in the spectrum ( see table [ t2-alpha ] and figs .
[ fig : dssd]a and [ fig : dssd]b ) .
a pulse generator was used to perform the gain - matching of the y - strips .
an accurate energy calibration of layers a and b was accomplished using well known @xmath36 decays from several at , bi , rn and po isotopes produced in the decay of implanted tl , pb and bi nuclei ( see figs . [
fig : dssd]a and [ fig : dssd]b ) .
no @xmath36 emitter was implanted in the layer c and , therefore , a coarse energy calibration of the latter was made using only the broad @xmath0 spectrum ( fig .
[ fig : dssd]d ) .
nevertheless , for implant-@xmath0 time - correlations only an energy window covering the broad @xmath0 spectrum is required and , therefore , an accurate energy calibration is of secondary relevance for this data analysis .
ccccc + @xmath36 energy & simba & @xmath36 emitter&precursor + ( kev ) & layer & & implanted + + 5304.3 & a , b & @xmath3po & ( implanted ) + 5869.5 & a&@xmath28at & ( implanted ) + 6002.4 & a&@xmath41po&@xmath41bi , @xmath41pb + 6050.8 & a&@xmath42bi&@xmath42tl + 6208.0 & a&@xmath43at&@xmath43bi + 6288.1 & a&@xmath44rn&@xmath44bi + 6300.0 & b&@xmath42bi&@xmath42tl + 6340.0 & b&@xmath42bi&@xmath42tl + 6537.0 & a&@xmath45po&@xmath45pb + 6622.9 & a&@xmath28bi&@xmath28tl + 6778.5 & a&@xmath46po&@xmath46pb + 7386.1 & a&@xmath29po&@xmath29pb + 7450.3 & a&@xmath28po&@xmath28at + 7686.8 & a , b&@xmath47po&@xmath47pb , @xmath47tl + 8375.9 & a , b&@xmath48po&@xmath48tl + simba was placed inside the cylindrical hole ( 23 cm diameter ) of the beta delayed neutron ( belen ) detector @xcite .
belen consisted of an array of 30 @xmath8he - counters of 2.54 cm diameter , embedded in a high - density polyethylene ( pe ) matrix ( fig .
[ fig : setup_picture ] ) .
the @xmath8he tubes were distributed in two rings , an inner one with a radius of 14.5 cm and 10 tubes of 10 atm , and the outer ring with a radius of 18.5 cm and 20 tubes of 20 atm .
this configuration was designed by means of geant4 @xcite and mcnpx @xcite monte carlo ( mc ) simulations in order to achieve a high and flat detection efficiency ( see fig . [
fig : efficiency ] ) . up to a neutron energy of 1 mev the detection efficiency was 40(2)% , and it decreased to 25% at 5 mev .
the mc codes were experimentally validated at e@xmath17=2.3 mev with a dedicated measurement of a well calibrated @xmath49cf source .
the q@xmath50 window of the exotic nuclei involved in the present measurement spans neutron energies up to 2.5 mev ( see table [ t5-pn ] ) . in this energy range , the average neutron detection efficiency is 38% ; this value was used in the data analysis ( see sec .
[ sec : analysis ] ) . using the calibrated @xmath49cf source a gain - matching of the response of all 30 counters was carried out before the experiment .
the stability of the overall detector response was checked regularly during the experiment .
the accumulated spectrum for all 30 tubes during the @xmath28hg setting is shown fig .
[ fig : response ] . the energy window considered in the data analysis as
neutron events comprises the range from a low threshold of the 191 kev peak up to the end of the main peak of the reaction ( see eq .
[ e3_nreac ] ) at 765 kev .
this peak corresponds to the kinetic energy realized by the two reaction products , a triton and a proton.@xmath51 in order to reduce the neutron background in belen two additional elements were implemented in the setup . on one hand , a pe wall with a thickness of 30 cm was installed upstream from belen in order to shield the detector from neutron background induced by the secondary beam ( see fig .
[ fig : setup_picture ] ) .
this wall had a central hole of 180 mm in @xmath52-direction and 70 mm in @xmath53-direction to let the beam into simba .
in addition , a layer of borated rubber was attached to the back side of the pe wall in order to absorb thermalized neutrons that could eventually reach the rear side of the wall . on the other hand ,
the belen detector was surrounded by 20 cm of pe shielding ( see fig . [
fig : setup_picture ] ) in order to moderate and absorb scattered neutrons from the surroundings .
the gsi multi branch system ( mbs ) @xcite was used to acquire data from the frs tracking detectors and simba .
this data acquisition system ( dacq ) was triggered by a scintillator at s4 with an efficiency of @xmath40100% for heavy ions .
mbs was also triggered by high - energy implant and low - energy @xmath0-decay events in simba .
on the other hand , the neutron data from belen were acquired using a digital self - triggered system @xcite based on sis3302 vme digitizers from struck innovative systems @xcite .
each module had 8 input channels running at 100 mhz sampling rate with an adc resolution of 16 bit .
a common clock was used for time - stamping the events acquired in the belen and mbs dacq systems with 20 ns resolution . in this way ,
ion - implant versus @xmath0-decay time - correlations and ion-@xmath0-neutron time - correlations could be built over an arbitrarily long time - window and in both , forward ( increasing ) and backward ( decreasing ) time directions .
the latter aspect was a key feature in the analysis stage in order to determine reliably the background level ( see sec .
[ sec : analysis ] ) .
the methodology followed here for the analysis of the @xmath0-decay half - lives is similar to the approach successfully applied in previous experiments at isol and fragmentation facilities , such as described in @xcite .
there are two fundamental aspects in this data analysis , namely a reliable background characterization of @xmath0- and neutron events and the spatial- and temporal- correlation approach .
we describe first the analysis methodology from a rather general perspective and afterwards we show in detail its application to one of the largest implant statistics case , which is @xmath48tl .
the latter is also used to illustrate the background treatment in the analysis .
let us consider one specific isotope @xmath55 .
the bateman analytical expression @xcite describes the time - evolution of its abundance @xmath56 . in this analysis we assumed that only the parent ( @xmath57 ) and daughter ( @xmath58 ) decays are contributing to the decay curve .
this assumption is justified because all granddaughters of the analyzed nuclei are either stable nuclei or have a half - life much longer than that of parent and daughter nuclei . the resulting expression after this assumption
is given by @xmath59 where @xmath60 is the decay constant for the implanted nucleus @xmath61 , with unknown half - life @xmath62 , and @xmath63 the decay constant of the daughter nucleus @xmath64 . as it is shown below , in many cases
even the contribution of the daughter nucleus was very small .
regarding the spatial correlation , we considered an implant and a decay event to be spatially associated when the @xmath0-decay position measured in layers a , b or c of simba was within a correlation region of 3 mm@xmath65 around the implant position measured in the same layer . extending this condition to neighboring layers
did not improve noticeably the statistics and was therefore disregarded . in general , smaller and larger correlation areas were not found to provide a better result in terms of statistics and signal - to - background ratio in the time - correlation diagrams @xcite .
regarding the time correlation , the methodology followed here consisted of building a correlation histogram for every implant of a certain species @xmath55 containing its time difference with respect to all subsequent @xmath0 events within a broad time window @xmath66 . in this analysis we used ten times the expected half - life , @xmath67 . as it is demonstrated below
, the uncorrelated background rate is a constant value , that can be referred to as @xmath68 . in this case , the probability density function describing the time - dependency of the correlation distribution is given by @xcite @xmath69 where @xmath70 is the @xmath0-detection efficiency in simba .
thus , for a certain number of implanted events @xmath71 , the total number of @xmath0 particles detected @xmath72 at a time @xmath73 with respect to the implantation time ( @xmath74 ) is given by @xmath75 where @xmath76 is the total number of detected decays at a time @xmath73 , @xmath68 is the @xmath0-background normalized and corrected by @xmath70 and @xmath77 corresponds to the bin time - width used in the implant-@xmath0 time - correlation histogram .
the time evolution of the parent abundance is described by @xmath78 , whereas the contribution of the daughter is given by @xmath79 , assuming @xmath80 . in the data analysis the common factor @xmath81 in eq .
[ e4-allbetas-2 ] is derived from the first bin in the correlation diagram @xcite . in this way
the quantity of interest ( @xmath62 ) can be reliably determined without knowing explicitly the @xmath0-detection efficiency .
the thallium isotope @xmath48tl was implanted with large statistics ( 1015 implants ) and therefore we used this case to establish the analysis methodology on a reliable statistical basis .
the @xmath0 background showed a dependency with the time structure of the pulsed primary beam .
indeed , during beam extraction from sis ( 1 s ) the overall background level of @xmath0-like events in simba was @xmath4040% higher than during the time - interval between spills @xcite .
this feature led to a better signal - to - background ratio in the implant-@xmath0 time - correlation histograms when only @xmath0-events outside of the spill time - intervals were considered in the correlations , when compared to the same diagram including @xmath0-events inside and outside of the spill . with this restriction in mind ,
the background level evaluation was based on time backward ( @xmath82 ) implant-@xmath0 correlations , i.e. time difference between each implant and all the @xmath0-events occurred before it , within a broad time window ( @xmath83 ) and in the same correlation area used in the forward analysis of 3@xmath843 mm@xmath65 .
the background level thus determined allows one to adjust the parameter @xmath68 in eq .
[ e4-allbetas ] .
this approach is illustrated in fig .
[ f4 - 213tl ] for the case of @xmath48tl , which shows backward ( negative ) and forward ( positive ) implant-@xmath0 correlations .
the contribution to the measured @xmath0-activity from decays of other nuclei can be assumed to be negligible due to the very low average implantation rate of @xmath85 ions / s / pixel .
using eq .
[ e4-allbetas-2 ] a binned maximum likelihood ( ml ) @xcite analysis of the time correlation histogram was carried out , which yielded a half - life for @xmath48tl of @xmath86 s. comparison with other literature values is detailed in table [ t5-hl ] .
the implant-@xmath0-neutron correlations were analyzed by selecting a correlation window of @xmath87400 @xmath88 forward and backward in time , following each @xmath0 detection .
this time interval was determined according to the expected neutron moderation time in polyethylene .
considering the @xmath0 efficiency @xmath70 as a constant value along the range of energies of interest , the @xmath89-value can be directly obtained from the subtraction of time - forward and time - backward @xmath0-neutron correlated events , @xmath90 where @xmath91 is the belen neutron efficiency , @xmath92 the number of forward correlated implant-@xmath0-neutron events and the @xmath72 the number of parent @xmath0-decays .
@xmath93 designates the backward @xmath0-neutron correlations , which were used to define the uncorrelated neutron background level .
the efficiency can be considered flat along the energy range of interest according to the @xmath94-values of the implanted isotopes ( 100 kev - 2.5 mev , see table [ t5-pn ] ) , with a constant value of 38% and with a relative uncertainty of 5% ( see fig .
[ fig : efficiency ] ) .
this overall uncertainty includes the contributions of statistical errors of @xmath92 , @xmath93 and @xmath72 together with the uncertainty for the belen detector efficiency from the fluctuations along the energy range .
for @xmath48tl we measured 5 forward and none backward correlated neutrons ( fig .
[ f4-neu-1 ] ) , which yielded a @xmath89-value of @xmath95 .
five more thallium isotopes were measured and their @xmath0-decay half - lives and neutron branching ratios were analyzed following the approach described above for @xmath48tl .
the implant-@xmath0 time - correlation histograms for all of them , @xmath96tl are shown in fig .
[ f4-tl ] . the binned ml - analysis ( dashed - dotted line on diagrams ) yields @xmath97
s for @xmath28tl , @xmath98 s for @xmath42tl and @xmath99 s for @xmath47tl .
an almost negligible contribution from the much slower decay of the daughter nuclei ( @xmath100pb ) can be observed in these three cases .
the most exotic thallium nuclei @xmath101tl were measured with rather limited statistics of only 281 and 99 implants , respectively . in this case
, we noticed a slight improvement of the signal - to - background ratio in the correlation histograms when the correlation area was enlarged from @xmath102 mm@xmath65 to @xmath103 mm@xmath65 .
the ml analysis also shown in diagrams of fig .
[ f4-tl ] yielded a half - life of @xmath104 s for @xmath29tl and @xmath105 s for @xmath46tl .
the neutron data analysis showed the presence of only one correlated implant-@xmath0-neutron event for each of @xmath106tl , which yields neutron branching ratios of 2.2(2.2)% , 1.8(1.8)% and 4.6(4.6)% , respectively .
for these three cases , given the low number of events compatible with the physical boundary , we have alternatively calculated a conservative upper - limit based on the bayesian approach @xcite , which yields upper limits of 10% , 8% and 20% at a confidence level ( cl ) of 95% .
on the other hand , with belen we were able to observe a rather large number ( 10 ) of implant-@xmath0-neutron correlated events for @xmath47tl ( histogram also shown in fig .
[ f4-tl ] ) , resulting in a @xmath89 value of @xmath107 .
no single correlated or uncorrelated event was detected for @xmath46tl , which according to the implantation statistics obtained led an upper limit of 11.5% and a bayesian upper limit of @xmath108% at the 95% cl .
concerning the other implanted species , @xmath109tl , the implantation statistics was not enough to determine either their half - lives nor the neutron branching ratios .
the lead isotopes @xmath110pb were identified in the frs and implanted in simba . out of them
, @xmath12pb were implanted with enough statistics for a reliable half - life analysis .
@xmath47pb was also implanted with large statistics ( see fig . [ f3-numimplants ] ) , but its half - life of @xmath111 s @xcite was too long for our analysis methodology and instrumentation .
the ml analysis for the implant-@xmath0 correlation diagrams of @xmath12pb are shown in fig .
[ f4-pb ] .
the @xmath29pb analysis yielded a half - life of @xmath112 s , as it can be observed on its diagram , contributions from @xmath29pb and his daughter @xmath29bi are present in the correlation histogram . for the case of the measurement of @xmath46pb , it allowed us to carry out a cross - check of the aforementioned analysis methodology , as its half - life can be determined by means of two different methods : implant-@xmath0 and implant-@xmath36 correlations .
the former is also illustrated in fig .
[ f4-pb ] , which shows the implant-@xmath0 correlation and the result of the ml analysis . on the other hand , the peak of @xmath36-particles at 6778.5 kev , clearly identified in the energy spectrum of simba ( see fig . [
fig : dssd]b ) , corresponds to the decay of its granddaughter nucleus @xmath46po .
taking into account that the half - life of @xmath46po , 145@xmath1132 ms @xcite , is much shorter than that of the direct daughter , @xmath46bi , 2.25@xmath1135 s @xcite , it was possible to apply the method described in @xcite to obtain the half - life of @xmath46pb . with this method we determined a half - life of @xmath114 s
( see diagram of fig .
[ f4 - 216pb - alpha ] ) which is in perfect agreement with the one obtained applying the conventional method described above for implant-@xmath0 time correlation , @xmath115 s. the accuracy in the analysis of @xmath45pb and @xmath41pb was mainly limited by the implantation statistics , which was of 436 and 235 implants , respectively . however , a reliable ml analysis was possible from their implant-@xmath0 correlation diagrams , as shown in the bottom diagrams of fig .
[ f4-pb ] .
the resulting half - lives were @xmath116 s for @xmath45pb and @xmath117 s for @xmath41pb . according the negative q@xmath118-values of all these lead isotopes
( see table [ t5-pn ] ) , no neutron branching emission is expected on them .
from the eight mercury isotopes identified , @xmath119hg , implant statistics were high enough to analyze reliably four of them , @xmath10hg .
the decay curve is strongly determined by the number of ion implants , as well as by the value of the half - life .
thus , for @xmath23hg , with a low implantation statistics ( 220 events ) and a relatively large half - life obtained from the ml analysis , @xmath120 s , the statistical uncertainty was 38% .
on the other hand , @xmath121hg has a factor of about two more implants ( 583 events ) and a shorter half - life ( @xmath122 s ) , which leads to a much lower uncertainty of 17% .
an intermediate situation is found for the remaining two mercury isotopes , @xmath14hg , for which their half - lives analysis yielded @xmath123 s and @xmath124 s , respectively . in the latter case ,
the value obtained in the present work for the half - life of @xmath28tl was employed in the analysis .
all aforementioned analyzed half - lives are shown in diagrams of fig . [ f4-hg ] .
regarding the neutron emission branching ratio , for @xmath3hg and @xmath28hg one implant-@xmath0-neutron event has been detected in the forward ( moderation ) time - window for each nucleus .
this measurement yields neutron branching ratios of 2.2(2.2)% and 6.3(6.3)% , respectively . for these cases
the calculated conservative upper - limit based on the bayesian approach @xcite yields upper constraints of 10% , and 28% at a cl of 95% . the @xmath125au isotopes were identified in this measurement , but the implantation statistics was only high enough to analyze three of them , @xmath9au . in the case of @xmath126au ( see fig
. [ f4 - 204au ] ) we also found a slight improvement of the decay curve when the correlation area was enlarged from 1 pixel ( 9 mm@xmath65 ) to 2 pixels ( 25 mm@xmath65 ) around the implant location , and included in the correlation those implant-@xmath0 events detected during the spill time .
these provide a better sensitivity for the analysis , which yields a half - life of @xmath127 s. for the cases of @xmath35au and @xmath128au , both with @xmath129 implants , the ml analysis of the implant-@xmath0 correlation diagrams yield half - lives of @xmath130 s and @xmath131 s , respectively .
[ f4 - 205 - 206au ] shows the analysis of these two gold isotopes .
bismuth was the heaviest element implanted in simba and we were able to determine the half - lives of three isotopes , @xmath13bi .
the ml analysis of @xmath41bi yields a half - life of @xmath132 s and , as it can be seen in its diagram on fig .
[ f4 - 218 - 219bi ] , the @xmath0 contribution comes only from its own decay , as its daughter ( @xmath41pb ) is an @xmath36 emitter .
the half - life analysis for @xmath43bi yields @xmath133 s ( see the diagram also on fig . [ f4 - 218 - 219bi ] ) and it includes the recent published half - life of its daughter , @xmath43po , @xmath134 s @xcite
. for the @xmath0-decay analysis of @xmath44bi one has to take into account that the half - life value of its daughter nucleus @xmath44po is still unknown .
thus , our analysis provides a range of possible half - life values for @xmath44bi , which spans between 4 s and 15 s. the bold marker in fig .
[ f4 - 220bi ] represents the @xmath44bi half - life using the theoretical prediction calculated by frdm+qrpa model @xcite for the half - life of @xmath44po , @xmath135 s.
this section provides a summary of the main results obtained in this work , as well as a comparison with all previous experimental results and the theoretical predictions of frdm+qrpa @xcite and df3+cqrpa @xcite models . a detailed comparison with the more recent calculations of refs .
@xcite can be found in ref.@xcite . in summary ,
half - live values for 20 neutron - rich isotopes of au , hg , tl , pb and bi have been determined experimentally , as well as neutron - branching ratios ( or upper limits ) for eight of them . as some of the analyzed isotopes are not expected to be neutron emitters according their q@xmath50-values ( see the last column of table [ t5-pn ] ) , due to the low statistics available , the neutron emission analysis has been focused in those with large enough q@xmath50 .
the results are displayed in fig .
[ f5-resultshl ] and fig .
[ f5-pn ] , respectively . for comparison purposes ,
previously published theoretical and experimental half - life values have been also included in fig .
[ f5-resultshl ] .
apart from the discrepancies found for the @xmath136hg isotopes , a rather good agreement is found between the present results and previously published experimental data . in the case of gold isotopes ,
all half - lives are fully compatible with those reported in previous measurements @xcite , which are in reasonable agreement with df3+cqrpa for @xmath137 . for @xmath128au ( @xmath6 )
the half - life obtained confirms the recent value published in ref .
@xcite and both differ from the df3+cqrpa value and the trends predicted by the available theoretical models .
whether this gold anomaly " is related to effects changing the occupation of the @xmath138 orbital , or due to a weakening of the spin - orbit field caused by the tensor force @xcite , or might be due to the recently proposed three - body force mechanism @xcite remains an open question which calls for more detailed theoretical studies and further specific experiments allowing for reconstruction of the decay scheme . concerning hg isotopes , as discussed in ref .
@xcite , recent measurements at cern - isolde @xcite indicate that the half - life of @xmath136hg nuclei are much shorter than the values reported in ref .
@xcite and , therefore , this discrepancy will not be discussed further .
theoretical predictions by the frdm+qrpa seem to agree better with the measured half - lives , with the only exception of @xmath121hg , whereas the opposite is found for the hg half - lives predicted with df3+cqrpa . regarding the thallium isotopes ,
previous half - lives were obtained in another experiment by using a similar experimental setup , but using a completely different analysis approach @xcite . in summary a good agreement is found for @xmath28tl , whereas the half - lives of @xmath42tl and @xmath48tl differ by factors of 2 - 3 .
it is worth emphasizing the overall good agreement for the thallium chain between the present results and frdm+qrpa predictions , including the case of the most exotic nuclei reported here for the first time , @xmath139tl .
this result seems to indicate the rather low relevance of first forbidden ( ff ) transitions in the @xmath140 mass region , as discussed in ref .
@xcite .
the analysis of the lead isotopes includes three new half - lives , @xmath141pb .
the half - life obtained for @xmath29pb is in reasonable agreement with the two previous measurements @xcite .
along the lead isotopic chain , the frdm+qrpa model overestimates the experimental values by factors between 3 - 9 .
the values predicted by df3+cqrpa are , on average , slightly closer to the measured half - lives , but the mass ( neutron number ) dependency of the half - life is not satisfactorily reproduced by any of these two models . for the bismuth chain
the frdm+qrpa predictions agree reasonably well with the heaviest measured nuclei @xmath142bi , whereas almost one order of magnitude difference is found for @xmath41bi .
the latter half - life is , however , rather well established experimentally @xcite .
the result quoted in this work for the half - life of @xmath44bi , 4 - 15 s ( see fig . [ f4 - 220bi ] ) , can be re - determined more accurately once the half - life of @xmath44po is measured .
the neutron - branching ratios determined in this work represent the first set of experimental data available in this mass region .
therefore , the values reported here can be only compared with theoretical predictions .
both theoretical and experimental @xmath89-values are listed in table [ t5-pn ] and displayed in fig .
[ f5-pn ] . in summary , the agreement between theory and experiment is rather good for the hg and tl isotopes with masses between 208 and 214 .
however , for the two heaviest thallium nuclei @xmath101tl , ktuy @xcite and rhb+rqrpa @xcite models are in agreement but the other theoretical predictions of the neutron emission overestimate substantially the obtained experimental results .
since both the half - life and the neutron emission probability are integral quantities of the @xmath0-decay , it is difficult to explain why the frdm+qrpa reproduces so well the average decay strength over the full q@xmath16-window ( see fig . [ f5-resultshl ] ) along the chain of measured tl isotopes , whereas it seems to fail dramatically in the upper energy range of @xmath101tl , beyond the neutron separation energy of the daughter nuclei , as deducted from their determined neutron branching ratios . at first sight ,
one is tempted to attribute such a discrepancy to the possible contribution of high - energy first forbidden ( ff ) transitions populating low - lying levels of the daughter nuclei and hindering therefore the emission of neutrons . nevertheless , this interpretation is at variance with the overall systematic found in the @xmath140 region @xcite , where gamow - teller ( gt ) transitions seem to play a dominant role in general . at present , the only plausible explanation for such a feature could be that the overall strength of the @xmath0 decay is overestimated in frdm+qrpa both , in the full q@xmath16 and in the upper energy window q@xmath50 . as reported in ref .
@xcite , more advanced microscopic models such as rhb+rqrpa @xcite and ktuy @xcite show an inverted behavior , yielding good predictions for the @xmath89-values , but discrepant values for the half - lives , thus not improving the situation .
recent calculations based on an improved qrpa and hf theory @xcite , included in fig
. [ f5-pn ] , also show large discrepancies . clearly more @xmath0-decay measurements and theoretical efforts are needed in this mass region , in order to gain a better understanding of the underlying nuclear structure effects , as well as to guide global theoretical models far - off stability .
this work was supported by the _ spanish ministerio de economia y competitividad _ under grants no .
fpa2011 - 28770-c03 - 03 , fpa2008 - 04972-c03 - 03 , aic - d-2011 - 0705 , fpa2011 - 24553 , fpa2008 - 6419 , fpa2010 - 17142 , fpa2014 - 52823-c2 - 1-p , fpa2014 - 52823-c2 - 2-p , cpan csd-2007 - 00042 ( ingenio2010 ) , and the program severo ochoa ( sev-2014 - 0398 ) . i.d and
m.m were supported by the german helmholtz association via the young investigators grant vh - ng 627 ( lisa- lifetime spectroscopy for astrophysics ) and the nuclear astrophysics virtual institute ( vh - vi-417 ) , and by the _
german bundesministerium fr bildung und forschung _ under no .
06mt7178 / 05p12wofnf .
r.c.f acknowledges the support of the spanish nuclear security council ( csn ) under a grant of catedra argos .
uk authors acknowledge the support of the uk science & technology facilities council ( stfc ) under grant no .
st / f012012/1 .
yu.a.l . acknowledges support from helmholtz - cas joint research group ( hcjrg-108 ) .
r.c.f . and i.d
. are also supported by the national research council of canada ( nserc ) discovery grants sapin-2014 - 00028 and rgpas 462257 - 2014 at triumf .
cccccccc + nuclei&@xmath143 & implanted & t@xmath144 ( s ) & previous & frdm+qrpa & df3+cqrpa + & & ions&(this work)&t@xmath144 ( s)&(s ) @xcite&(s ) @xcite + + @xmath126au & 125 & 54 & 33.7 @xmath11314.9 & 39.8 @xmath1130.9 @xcite & 455.3 & 42.4 + & & & & 37.2 @xmath1130.8 @xcite & & + @xmath35au & 126 & 103 & 35.4 @xmath11316.7 & 31.0 @xmath1130.2 @xcite & 222.0 & 18.7 + & & & & 32.5 @xmath11314.0 @xcite & & + @xmath128au & 127 & 106 & 55.7 @xmath11316.7 & 40.0 @xmath11315.0 @xcite&21.3 & 0.72 + + @xmath23hg & 128 & 220 & 132.2 @xmath11350.0 & 2460@xmath145 @xcite & 168.9 & 12.1 + @xmath121hg & 129 & 583 & 6.3 @xmath1131.1 & 35.0@xmath146 @xcite & 33.6 & 3.7 + @xmath3hg & 130 & 512 & 63.7 @xmath11311.6 & @xmath147300 ns & 41.2 & 5.2 + @xmath28hg & 131 & 253 & 26.4 @xmath1138.1 & @xmath147300 ns & 14.9 & 1.9 + + @xmath28tl & 130 & 483 & 76.5 @xmath11317.8 & 88@xmath148 @xcite & 70.9 & 114.9 + @xmath42tl & 131 & 1056 & 30.9 @xmath1138.0 & 96@xmath149 @xcite & 29.0 & - + @xmath48tl & 132 & 1015 & 23.8 @xmath1134.4 & 101@xmath150 @xcite & 32.4 & 70.4 + & & & & 46@xmath151 @xcite & & + @xmath47tl & 133 & 598 & 11.0 @xmath1132.4 & @xmath147300 ns & 14.4 & - + @xmath29tl & 134 & 281 & 9.7 @xmath1133.8 & @xmath147300 ns & 7.8 & - + @xmath46tl & 135 & 99 & 5.9 @xmath1133.3 & @xmath147300 ns & 2.8 & - + + @xmath29pb & 133 & 1079 & 98.4 @xmath11330.8 & 147 @xmath11312 @xcite & 282.5 & 27.1 + & & & & 160 @xmath11340 @xcite & & + @xmath46pb & 134 & 1005 & 99.4 @xmath11311.7 & @xmath147300 ns & 852.2 & 52.0 + @xmath45pb & 135 & 436 & 19.9 @xmath1135.3 & @xmath147300 ns & 104.9 & 8.5 + @xmath41pb & 136 & 235 & 14.8 @xmath1136.8 & @xmath147300 ns & 66.3 & 28.4 + + @xmath41bi & 135 & 294 & 38.5 @xmath11321.6 & 33 @xmath1131 @xcite & 2.92 & - + & & & & 36 @xmath11314 @xcite & & + @xmath43bi & 136 & 306 & 8.7 @xmath1132.9 & 22 @xmath1137 @xcite & 26.54 & - + @xmath44bi & 137 & 176 & 9.5 @xmath1135.7 & @xmath147300 ns & 5.17 & - + ccccccccccc + nuclei & @xmath143 & @xmath152(%)&frdm+qrpa & df3+cqrpa&rhb+rqrpa&ktuy&arxiv:1608.01956 & @xmath94(kev ) @xcite + & & ( this work)&(% ) @xcite&(% ) @xcite&(% ) @xcite&(% ) @xcite&(% ) @xcite & ( extr . = extrapolated ) + + @xmath126au & 125 & - & 0.0 & 0.0&0.1&0.0&- & @xmath1533453 @xmath113200 ( extr . )
+ @xmath35au & 126 & - & 0.0 & 0.0&0.2&0.0&- & @xmath1532151 @xmath113196 ( extr . )
+ @xmath128au & 127 & - & 0.0 & 0.0&0.5&0.0&0.0 & 2 @xmath113298 ( extr . )
+ + @xmath23hg & 128 & - & 0.0 & 3.2&0.3&0.0&- & @xmath153303.32 @xmath11331.23 + @xmath121hg & 129 & - & 0.0 & 2.8&0.5&0.0&- & 34 @xmath113149 ( extr . ) + @xmath3hg & 130 & 2.2 @xmath1132.2 & 0.0 & 9.3&0.6&0.0&71 & 201 @xmath113196 ( extr . )
+ @xmath28hg & 131 & 6.3 @xmath1136.3 & 0.81 & 7.5&0.8&0.0&11 & 551 @xmath113196 ( extr . )
+ + @xmath28tl & 130 & 2.2 @xmath1132.2 & 0.04 & 0.95&1.2&0.28&- & 578.67 @xmath11341.95 + @xmath42tl & 131 & 1.8 @xmath1131.8 & 0.56 & - & 1.3&0.23&0.0 & 869 @xmath113200 ( extr . )
+ @xmath48tl & 132 & 7.6 @xmath1133.4 & 13.26 & 2.93&6.8&1.93&100 & 1259.73 @xmath11327.10 + @xmath47tl & 133 & 34.3 @xmath11312.2 & 10.38 & - & 10&1.56&66 & 1595 @xmath113196 ( extr . ) + @xmath29tl & 134 & 4.6 @xmath1134.6 & 55.24 & - & 14.2&5.54&100 & 2021 @xmath113298 ( extr . ) + @xmath46tl & 135 & @xmath15411.5&55.36 & - & 17&4.45&96 & 2230 @xmath113315 ( extr . ) + + @xmath29pb & 133 & - & 0.0 & 0.0&0.3&0.0&- & @xmath1532455 @xmath113102 ( extr . ) + @xmath46pb & 134 & - & 0.0 & 0.0&0.3&0.0&0.0 & @xmath1532240 @xmath113196 ( extr . ) + @xmath45pb & 135 & - & 0.0 & 0.0&0.4&0.0&0.0 & @xmath1531705 @xmath113298 ( extr . )
+ @xmath41pb & 136 & - & 0.0 & 0.0&0.4&0.0&0.0 & @xmath1531348 @xmath113299 ( extr . )
+ + @xmath41bi & 135 & - & 0.0 & - & 1.2&0.0&0.0 & @xmath153740.61 @xmath11327.73 + @xmath43bi & 136 & - & 0.06 & - & 0.3&0.0&0.0 & @xmath153148 @xmath113196 ( extr . ) .
+ @xmath44bi & 137 & - & 0.01 & - & 0.4&0.0&0.0 & 66.0 @xmath113298 ( extr . )
+ 71ifxundefined [ 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 * * , ( ) in @noop _
_ , vol .
( ) p. @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ , vol .
( , ) p. @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.117.012501 [ * * ( ) , 10.1103/physrevlett.117.012501 ] @noop * * , ( ) @noop * * , ( ) http://www-w2k.gsi.de/frs/technical/frssetup/detectors/music80/music80_manual.pdf[__ ] , ( , ) _ _ , @noop ph.d
. thesis , ( ) _ _ , @noop ph.d .
thesis , ( ) _ _ , http://geb.uni-giessen.de/geb/volltexte/2012/8767[ph.d .
thesis ] , ( ) @noop * * , ( ) _ _ , @noop ph.d . thesis ,
( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) _ _ , @noop ph.d . thesis , ( ) _ _ , @noop ph.d .
thesis , ( ) _ _ , @noop ph.d .
thesis , ( ) link:\doibase 10.1007/s10751 - 012 - 0617 - 4 [ * * , ( ) ] @noop * * , ( ) _ _ , @noop ph.d .
thesis , ( ) @noop _ _ , ( , ) _ _ , @noop ph.d .
thesis , ( ) @noop * * , ( ) @noop _ _ , ( , ) @noop _ _ ( ) @noop _ _ ( ) http://web-docs.gsi.de/~mbs/v51/gm_mbs_rel_51.pdf [ `` , '' ] ( ) @noop * * , ( ) @noop `` , '' @noop * * , ( ) @noop * * , ( ) @noop _ _ , vol .
( , ) @noop _ _ , vol .
( , ) http://www.nndc.bnl.gov/chart/ and http://www.nndc.bnl.gov/nudat2/indx_dec.jsp [ ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) , @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) ,
|
* background : * previous measurements of @xmath0-delayed neutron emitters comprise around 230 nuclei , spanning from the @xmath1he up to @xmath2la . apart from @xmath3tl , with a minuscule branching ratio of 0.007% , no other neutron emitter is measured yet beyond @xmath4 . therefore new data are needed , particularly in the heavy mass region around n=126 , in order to guide theoretical models and to understand the formation of the third r - process peak at @xmath5 . + * purpose : * to measure both , @xmath0-decay half - lives and neutron branching ratios of several neutron - rich au , hg , tl , pb and bi isotopes beyond @xmath6 . + * method : * ions of interest are produced by fragmentation of a @xmath7u beam , selected and identified via the gsi - frs fragment separator . a stack of segmented silicon detectors ( simba ) is used to measure ion - implants and @xmath0-decays .
an array of 30 @xmath8he tubes embedded in a polyethylene matrix ( belen ) is used to detect neutrons with high efficiency and selectivity .
a self - triggered digital system is employed to acquire data and to enable time - correlations .
the latter are analyzed with an analytical model and results for the half - lives and neutron - branching ratios are derived using the binned maximum - likelihood method .
+ * results : * twenty new @xmath0-decay half - lives are reported for @xmath9au , @xmath10hg,@xmath11tl,@xmath12pb and @xmath13bi , nine of them for the first time .
neutron emission probabilities are reported for @xmath14hg and @xmath11tl . + * conclusions : * the new @xmath0-decay half - lives are in good agreement with previous measurements in this region .
the measured neutron emission probabilities are comparable or smaller than values predicted by global models like rhb+rqrpa .
|
triangular matrix factorization is a main building block in computational linear algebra . driven by a large range of applications in computational sciences ,
parallel numerical dense lu factorization has been intensively studied since several decades which results in software of great maturity ( e.g. , linpack is used for benchmarking the efficiency of the top 500 supercomputers @xcite ) .
more recently , efficient sequential exact linear algebra routines were developed @xcite .
they are used in algebraic cryptanalysis , computational number theory , or integer linear programming and they benefit from the experience in numerical linear algebra . in particular , a key point there is to embed the finite field elements in integers stored as floating point numbers , and then rely on the efficiency of the floating point matrix multiplication ` dgemm ` of the blas .
the conversion back to the finite field , done by costly modular reductions , is delayed as much as possible .
hence a natural ingredient in the design of efficient dense linear algebra routines is the use of block algorithms that result in gathering arithmetic operations in matrix - matrix multiplications .
those can take full advantage of vector instructions and has a high computation per memory access rate , allowing to fully overlap the data accesses by computations and hence delivers peak performance efficiency . in order to exploit the power of multi - core and many - core architectures
, we now investigate the parallelization of the finite field linear algebra routines .
we report in this paper the conclusions of our experience in parallelizing exact lu decomposition for shared memory parallel computers .
we try to emphasize which specificities of exact computation domains led us to use different approaches than that of numerical linear algebra . in short
, we will illustrate that numerical and exact lu factorization mainly differ in the following aspects : * the pivoting strategies , * the cost of the arithmetic ( of scalars and matrices ) , * the treatment of rank deficiencies .
those have a direct impact on the shape and granularity of the block decomposition of the matrix used in the computation . [
[ types - of - block - algorithms . ] ] types of block algorithms .
+ + + + + + + + + + + + + + + + + + + + + + + + + + several schemes are used to design block linear algebra algorithms : the splitting can occur on one dimension only , producing row or column slabs @xcite , or both dimensions , producing tiles @xcite . note that , here , we denote by tiles a partition of the matrix into sub - matrices in the mathematical sense regardless what the underlying data storage is .
algorithms processing blocks can be either iterative or recursive .
figure [ fig : blockalg ] summarizes some of the various existing block splitting obtained by combining these two aspects .
most numerical dense gaussian elimination algorithms , like in @xcite , use tiled iterative block algorithms . in @xcite the classic tiled iterative algorithm
is combined with a slab recursive one for the panel elimination . over exact domains ,
recursive algorithms are preferred to benefit from fast matrix arithmetic ( see below ) .
slab recursive exact algorithms can be found in @xcite and references therein and @xcite presents a tiled recursive algorithm .
[ [ the - granularity ] ] the granularity + + + + + + + + + + + + + + + is the block dimension ( or the dimension of the smallest blocks in recursive splittings )
. matrices with dimensions below this threshold are treated by a base - case variant ( often referred to as the panel factorization ) .
it is an important parameter for optimizing efficiency : a finer grain allows more flexibility in the scheduling when running numerous cores , but it also challenges the efficiency of the scheduler and can increase the bus traffic . [ [ the - cost - of - the - arithmetic . ] ] the cost of the arithmetic .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + in numerical linear algebra , the cost of arithmetic operations is more or less associative : with dimensions above a rather low threshold ( typically a few hundreds ) , sequential matrix multiplication of the blas reaches the peak efficiency of the processor .
hence the granularity has very little impact on the efficiency of a block algorithm run in sequential . on the contrary , over a finite field
, a small granularity can imply a larger number of costly modular reductions , as we will show in section [ sec : modcount ] .
moreover , numerical stability is not an issue over a finite field , and asymptotically fast matrix multiplication algorithms , like winograd s variant of strassen algorithm @xcite can be used on top of the blas .
their speed - up increases with matrix dimension .
the cost of sequential matrix multiplication over finite field is therefore not associative : a larger granularity delivers better sequential efficiency .
[ [ pivoting - strategies - and - rank - deficiencies . ] ] pivoting strategies and rank deficiencies .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in dense numerical linear algebra , pivoting is used to ensure good numerical stability and good data locality @xcite . in the context of dense _ exact _ linear algebra ,
stability is no longer an issue . instead
, only certain pivoting strategies will reveal the echelon form or , equivalently , the rank profile of the matrix @xcite .
this is a key invariant used in many applications using exact gaussian elimination , such as grbner basis computations @xcite and computational number theory @xcite . over exact domains
, the rank deficiency of some blocks also leads to unpredictable dimensions of the tiles or slabs , as will be illustrated in section [ sec : rankdef ] .
this makes the block splitting necessarily dynamic contrarily to the case of numerical lu factorization where all panel blocks usually have full rank and the splitting is done statically according to a granularity parameter .
consequently the design of a parallel exact matrix factorization necessarily differs from the numerical algorithms as follows : * granularity should be as large as possible , to reduce modular reductions and benefit from fast matrix multiplication ; * algorithms should preferably be recursive , to group arithmetic operations in matrix products as large as possible .
* block splitting and pivoting strategies must preserve and reveal the rank profile of the matrix it also implies several requirements on the parallel run - time being used : * the block splitting has to be dynamically computed ; * the computing load for each task is not known in advance ( some panel blocks may have high rank deficiency ) , making the tasks very heterogeneous .
this motivated us to look into parallel execution runtimes using tasks with work - stealing based scheduling .
all experiments have been conducted on a 32 cores intel xeon e5 - 4620 2.2ghz ( sandy bridge ) with l3 cache(16384 kb ) .
the numerical blas is atlas v3.11.4 , lapack v3.4.2 and plasma v2.5.0 .
we used x - kaapi-2.1 version with last git commit : xkaapi_2.1 - 30-g263c19c638788249 .
the gcc compiler version used is gcc 4.8.2 that supports openmp 3.1 .
we introduce in section [ sec : prelim ] the algorithmic building blocks on which our algorithms will rely and the parallel programming models and runtimes that we used in our experiments . in order to handle each problem separately , we focus in section [ sec : fullrank ] on the simpler case where no rank deficiency occur . in particular section [ sec : modcount ] presents detailed analysis of the number of modular reductions required by various block algorithms including the tiled and slab recursive , the left - looking , right - looking and crout variants of the tiled iterative algorithm .
lastly section [ sec : rankdef ] deals with elimination with rank deficiencies .
we there present and compare new slab iterative , tiled iterative and tiled recursive parallel algorithms that preserve rank profiles .
we then show that the latter can match state of the art numerical routines , even when taking rank deficiencies into account .
all block algorithms that we will describe rely on four type of operations that we denote using the blas / lapack naming convention : ` gemm ` : : : general matrix multiplication , computing @xmath0 , ` trsm ` : : : solving upper / lower triang . syst . with matrix
right / left h.s @xmath1 . `
laswp ` : : : permuting rows or columns by sequence of swaps . `
getrf ` : : : computing @xmath2 , @xmath3 and @xmath4 stored in place of @xmath5 , s.t . @xmath6 .
a first prefix letter ` d ` or ` f ` is specifies if the routine works over double precision floating point numbers or finite field coefficients and an optional prefix ` p ` stands for parallel implementation .
our implementations use the sequential routines of the ` fflas - ffpack`library @xcite . there ,
the elements of a finite @xmath7 for a prime @xmath8 of size about 20 bits are integers stored in a double precision floating point number .
the sequential ` fgemm`routine combines recursive steps of winograd s algorithm calls to numerical blas ` dgemm`and reductions modulo @xmath8 when necessary .
the ` ftrsm`and ` fgetrf`routines use block recursive algorithms to reduce most arithmetic operations to ` fgemm ` .
more precisely ` fgetrf`is either done by a slab recursive algorithm @xcite or a tile recursive algorithm @xcite .
we base our implementation on the task based parallel features of openmp standard .
this is motivated by the use of recursive algorithms where tasks are mandatory .
now in tile iterative algorithms , loop with tasks happen to perform at least as good as parallel loops . `
libgomp`is the gnu implementation of the openmp api for multi - platform shared - memory parallel programming in c / c++ and fortran .
alternatively , we also used ` libkomp ` @xcite , an optimized version of ` libgomp ` , based on the ` xkaapi`runtime , that reduces the overhead of the openmp directives and handles more efficiently threads creation , synchronization and management . in the experiments of the next sections
, we will compare efficiency of the same code linked against each of these two libraries . in the iterative block algorithms , all matrix product tasks are sequential , on the contrary the recursive block algorithms must call parallel matrix products `
pfgemm ` , which we describe here .
operation ` pfgemm`is of the form @xmath9 . in order to split the computation into independent tasks , only the row dimension of @xmath5 and the column dimension of @xmath10 only are split .
the granularity of the split can be chosen in two different ways : either as a fixed value , or by a ratio of the input dimension ( e.g. the total number of cores ) .
we chose the second option that maximizes the size of the blocks while ensuring a large enough number tasks to for the computing ressources .
all our experiments showed that this option performs better than the first one .
when used as a subroutine in a parallel factorization , it will create more tasks than the number of available cores , but this heuristic happen to be a good compromise in terms of efficiency .
figure [ fig : pfgemmtime ] shows the computation time on 32 cores of various matrix multiplications : the numerical ` dgemm`implementation of ` plasma - quark ` , the implementation of ` pfgemm`of ` fflas - ffpack`using openmp tasks , linked against the ` libkomp`library .
this implementation is run over the finite field @xmath11 or over field of real double floating point numbers , with or without fast strassen - winograd s matrix product .
one first notices that most routine perform very similarly .
more precisely , ` plasma - quark``dgemm`is faster on small matrices but the effect of strassen - winograd s algorithm makes ` pfgemm`faster on larger matrices , even on the finite field where additional modular reductions occur .
+ in terms of speed - up , the ` pfgemm`reaches a factor of approximately 27 ( using 32 cores ) whereas the numerical ` dgemm`of ` plasma - quark`reaches a factor of 29 , but this is mostly reflect the fact that ` dgemm`has a less efficient sequential reference timing since it does not use strassen - winograd s algorithm .
similarly , other basic routines used in the recursive block algorithms , such as ` ftrsm`(solving matrix triangular systems ) and ` flaswp`(permuting rows or columns ) , have been parallelized by splitting a dimension into a constant number of blocks ( typically the number of cores ) .
in this section , we make the assumption that no rank deficiency occur during the elimination of any of the diagonal block .
this hypothesis is satisfied by matrices with generic rank profile ( i.e. having all their leading principal minor non zero ) .
this assumption allows us to focus on the problem of reducing the modular reduction count . when computing over finite field , it is of paramount importance to reduce the number of modular reductions in the course of linear algebra algorithms .
the classical technique is to accumulate several multiplications before reducing , namely replacing @xmath12 with @xmath13 while keeping the result exact .
if @xmath14 and @xmath15 are integers between @xmath16 and @xmath17 this is possible with integer or floating point units if the result does not overflow , or in other words if @xmath18 , see , e.g. , @xcite for more details .
this induces a splitting of matrices in blocks of size the largest @xmath19 satisfying the latter condition .
now the use of block algorithms in parallel , introduces a second blocking parameter that interferes in the number of reductions
. we will therefore compare the number of modular reductions of three variants of the tile iterative algorithm ( left - looking , right - looking and crout , see @xcite ) , the slab recursive algorithm of @xcite , and the tile recursive algorithm of @xcite .
for the sake of simplicity , we will always assume that the block dimensions in the parallel algorithms are always below @xmath19 .
in other words operations are done with full delayed reduction for a single multiplication and any number of additions : operations of the form @xmath20 are reduced modulo @xmath8 only once at the end of the addition , but @xmath21 requires two reductions .
for instance , with this model , the number of reductions required by a classic multiplication of matrices of size @xmath22 by @xmath23 is simply : @xmath24 . from (
* theorem 3 ) , this extends also for triangular solving @xmath25 to @xmath26 : with unit diagonal , @xmath27 ( actually the computation of the lowest row of the solution requires no modulo as it is just a division by @xmath28 , we will therefore rather use @xmath29 ) and @xmath30 ( with the previous refinement for @xmath31 , this also reduces to @xmath32 ) .
table [ tab : blockvariants ] sketches the different shapes of the associated routine calls in the main loop of each variant .
ccc left looking & crout & right looking + ` utrsm`((i-1)k,(i-1)k , k ) ` gemm`(n-(i-1)k,(i-1)k , k ) ` pluq`(k , k ) ` trsm`(k , k , n - ik ) & ` gemm`(n-(i-1)k,(i-1)k , k ) ` gemm`(k,(i-1)k , n - ik ) ` pluq`(k , k ) ` utrsm`(k , k , n - ik ) ` trsm`(k , k , n - ik ) & ` pluq`(k , k ) ` utrsm`(k , k , n - ik ) ` trsm`(k , k , n - ik ) ` gemm`(n - ik , k , n - ik ) + then the number of modular reductions required for these different lu factorization strategies is given in table [ tab : modcount ] .
.counting modular reductions in full rank block lu factorization of an @xmath33 matrix modulo @xmath8 when @xmath34 , for a block size of @xmath35 dividing @xmath36 . [ cols="^,<,^ " , ] in figure [ fig : moduloblocklu ] we compare the tiled iterative variants with the tiled recursive algorithm .
the latter uses as a base case an iterative crout algorithm too which performs fewer modular operations , the tiled recursive algorithm performs better than all other tiled iterative versions .
this can be explained by a finer and more adaptive granularity and a better locality .
the left looking variant performs poorly for it uses an expensive sequential ` trsm`task . although crout and right - looking variant perform about the same number of matrix products , those of an iteration of the right - looking variant are independent , contrarily to those of the crout variant , which explains a better performance despite a larger number of modular reductions .
figure [ fig : doubleblocklu ] shows the performance without modular reductions , of the tiled recursive parallel implementation on full rank matrices compared to ` plasma - quark ` .
the best block size for the latter library was determined by hand for each matrix size .
the two possible data - storage for ` plasma - quark`are used : the collection of tiles or the row - major data - storage .
our tiled recursive parallel pluq implementation without modular reductions behaves better than the ` plasma - quark``getrf_tile ` .
this is mainly due to the bi - dimensional cutting which allows for a faster panel elimination , parallel ` trsm`computations , more balanced ` gemm`computations and some use of strassen - winograd s algorithm .
this explains why performance join again on more than 24 cores : the size of the sequential blocks get below the threshold where this algorithm speeds up computations ( typically 2400 on this machine ) .
we now consider the general case of matrices with arbitrary rank profile , that can lead to rank deficiencies in the panel eliminations .
algorithms computing the row rank profile ( or equivalently the column echelon form ) used to share a common pivoting strategy : to search for pivots in a row - major fashion and consider the next row only if no non - zero pivot was found ( see @xcite and references therein ) .
such an iterative algorithm can be translated into a slab recursive algorithm splitting the row dimension in halves ( as implemented in sequential in @xcite ) or into a slab iterative algorithm .
more recently , we presented in @xcite a more flexible pivoting strategy that results in a tile recursive algorithm , cutting both dimensions simultaneously .
as a by product , both row and column rank profiles are also computed simultaneously .
[ [ a - slab - iterative - algorithm . ] ] a slab iterative algorithm .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + in the slab iterative algorithm shown in figure [ fig : tilecup ] , each panel factorization has to be run by a sequential algorithm .
this sequential task is costly and therefore imposes a choice of a fine granularity , which , as we saw , on the other hand implies more modular reductions and a lesser speed - up of strassen - winograd s algorithm .
another difficulty is the fact that the starting column position of each panel is determined by the rank of the blocks computed so far .
it can only be determined dynamically upon the execution .
this implies in particular that no data - storage by tiles , that fit the tiles of the algorithm is possible here .
moreover , the workload of each block operation may strongly vary , depending on the rank of the corresponding slab .
such heterogeneous tasks lead us to opt for work - stealing based runtimes instead of static thread management .
[ [ tiled - iterative - elimination . ] ] tiled iterative elimination .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + in order to speed - up the panel computation , we can split it into column tiles .
thanks to the pivoting strategy of @xcite , it is still possible to recover the rank profiles afterwards . now with this splitting , the operations remain more local and updates can be parallelized .
this approach shares similarities with the recursive computation of the panel described in @xcite .
figure [ fig : panel - pluq ] illustrates this tile iterative factorization obtained by the combination of a row - slab iterative algorithm , and a column - slab iterative panel factorization .
this optimization used in the computation of the slab factorization improved the computation speed by a factor of 2 , to achieve a speed - up of 6.5 on 32 cores with ` libkomp ` .
[ [ tiled - recursive - elimination . ] ] tiled recursive elimination .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + +
lastly , the tile recursive algorithm described in @xcite can be run in parallel using recursive tasks and the ` pfgemm ` , ` ftrsm`and ` flaswp`routines .
contrarily to most recursive factorization algorithms , the recursive splitting is done in four quadrants
. it has the interesting feature that if the tile top - left tile is rank deficient , then the elmination of the bottom - left and top - right tiles can be parallelized .
figure [ fig : pluq - gflops ] shows performance obtained for the tiled recursive and the tiled iterative factorization .
both versions are tested using ` libgomp`and ` libkomp`libraries .
the input ` s16k ` is a @xmath37 matrix with low rank deficiency ( rank is 15500 ) .
linearly independant rows and columns of the generated matrix are uniformly distributed on the dimension . with rank @xmath38,title="fig : " ] + the implementation with openmp of the tiled recursive lu maintained high efficiency in the case of rank deficient matrices .
it attained a speed - up of 13.6 on 32 cores . besides the fact that it benefits from strassen - winograd implementation
, it is adapted to minimize memory accesses and optimize data placement .
using ` libkomp`instead of ` libgomp`library and numactl , for round and robin interleave memory placement , that helps reducing dependency on bus speed , we manage to obtain peak performance for our tiled recursive lu factorization .
we analyzed five different algorithms for the computation of gaussian elimination over a finite field .
the granularity surely optimizes the parallelization of these algorithms but at the cost of more modular operations .
algorithm optimizing modular reductions are unfortunately not the most efficient in parallel .
the best compromise is obtained with our recursive tiled algorithm that performs best in both aspects .
our future work focuses on two main issues .
first , the use of specific allocators that can be used for a better mapping of data in memory and reduce distant accesses .
second , parallel programming frameworks for multicore processors @xcite could be more effective than binding threads on each numa node .
dataflow based dependencies , like when using openmp 4.0 directives , can ensure more parallelism for recursive implementation using ` libkomp`@xcite library .
|
we propose efficient parallel algorithms and implementations on shared memory architectures of lu factorization over a finite field . compared to the corresponding numerical routines , we have identified three main difficulties specific to linear algebra over finite fields .
first , the arithmetic complexity could be dominated by modular reductions .
therefore , it is mandatory to delay as much as possible these reductions while mixing fine - grain parallelizations of tiled iterative and recursive algorithms .
second , fast linear algebra variants , e.g. , using strassen - winograd algorithm , never suffer from instability and can thus be widely used in cascade with the classical algorithms . there , trade - offs are to be made between size of blocks well suited to those fast variants or to load and communication balancing .
third , many applications over finite fields require the rank profile of the matrix ( quite often rank deficient ) rather than the solution to a linear system .
it is thus important to design parallel algorithms that preserve and compute this rank profile . moreover , as the rank profile is only discovered during the algorithm , block size has then to be dynamic .
we propose and compare several block decomposition : tile iterative with left - looking , right - looking and crout variants , slab and tile recursive .
experiments demonstrate that the tile recursive variant performs better and matches the performance of reference numerical software when no rank deficiency occur .
furthermore , even in the most heterogeneous case , namely when all pivot blocks are rank deficient , we show that it is possbile to maintain a high efficiency .
|
the coupling of the optic and electric response of liquid crystals has lead to their wide application in display devices in recent years .
liquid crystalline materials are often made up of long , thin , rod - like molecules@xcite .
this anisotropy is what leads to their useful optic properties .
the molecular geometry and interactions can lead to a wide range of equilibrium phases . in this paper
we are concerned with the nematic phase where the molecules tend to align along a preferred direction , referred to as the director , giving long - range orientational order .
the propensity to order , as well as the direction along which the system orders are conveniently described by a tensor order parameter @xmath0 @xcite .
an important feature of the nematic phase is the possibility of topological defects .
these are director configurations that can not relax to the nematic ground state by a continuous rotation of the molecules .
examples of such defects that we will consider are shown in fig .
[ fig_twodefects](b ) .
in addition , although widely recognized as an important factor in device performance , the flow of the liquid crystal in response to changes in the orientation of the molecules is both difficult to measure experimentally and to incorporate into a simulation of the device .
liquid crystal hydrodynamics are typically described by the ericksen - leslie - parodi equations of motion @xcite .
however these are restricted to an order parameter field of constant magnitude and do not include the hydrodynamics of topological defects . here
we consider the beris - edwards formulation of liquid crystal hydrodynamics@xcite .
this allows for variations in the magnitude of the nematic order parameter and correctly models both defect dynamics and the coupling between the velocity field and the motion of the order parameter .
the beris - edwards equations are complex and previous numerical work to explore different flow regimes is very limited . in ref.@xcite
these equations were used but the problem was simplified by imposing a velocity field and ignoring back - flow ( back - flow refers to the effect of the order parameter dynamics on the velocity field ) . in @xcite an euler solution of a somewhat simpler version of the beris - edwards model was used to study the effect of hydrodynamics on phase ordering in liquid crystals .
however it has recently been shown that the beris - edwards equations can be solved successfully using a lattice boltzmann algorithm @xcite . in this paper
we will extend the lattice boltzmann approach described in @xcite to include the electric field terms necessary for a description of electro - optic device physics . the model is also generalized to include the three elastic constant formula for the deformation free energy of the liquid crystal .
the equations of motion are presented in section ii and an outline of the numerical algorithm in section iii .
we then apply the algorithm to some examples of interest in device physics @xcite .
the nematic liquid crystal is confined between two plates a few @xmath1 apart . the director configuration on the plates
is fixed . when an electric field is switched on ,
the molecules align parallel or perpendicular to the field depending on the sign of the anisotropy of the dielectric constant .
after switching off the field , long - range elastic interactions ensure that the molecules reorient themselves in the direction preferred by the surfaces .
the device can be used as a display because different liquid crystal orientations have different optical properties .
switching between different director orientations can take place homogeneously throughout the bulk of the device or it may proceed through domain growth involving the motion of domain walls , together with associated defects @xcite
. the operational state of a device such as a pi - cell may be topologically distinct from its state at zero voltage .
before the device can be used , the operational state must be nucleated and must grow to fill the display .
it is important to examine how this depends on different parameters of the liquid crystal material . in particular liquid crystals
are complex fluids and the speed of domain motion will depend on their hydrodynamic properties . in section
iv we examine the global stability of two topologically distinct states as a function of the applied voltage and the tilt angle at the plates . in section v we discuss the relevance of defect motion to domain growth between the two states and show that hydrodynamic coupling significantly changes the defect velocities .
we summarize the formulation of liquid crystal hydrodynamics described by beris and edwards@xcite and extended to include electric fields .
the continuum equations of motion are written in terms of a tensor order parameter @xmath0 which is related to the direction of individual molecules @xmath2 by @xmath3 where the angular brackets denote a coarse - grained average .
( greek indices will be used to represent cartesian components of vectors and tensors and the usual summation over repeated indices will be assumed . )
@xmath0 is a traceless symmetric tensor .
its largest eigenvalue , @xmath4 , @xmath5 , describes the magnitude of the order .
we first write down a landau free energy which describes the equilibrium properties of the liquid crystal .
this appears in the equation of motion of the order parameter , which includes a cahn - hilliard - like term through which the system evolves towards thermodynamic equilibrium .
* free energy : * the equilibrium properties of a nematic liquid crystal can be described by the landau - de gennes free energy@xcite @xmath6 @xmath7 describes the bulk free energy@xmath8 for @xmath9 there is a first - order transition from the isotropic to the nematic phase .
@xmath10 is the analogue of the frank elastic free energy density @xmath11 the dielectric constant is usually measured along and perpendicular to the nematic axis , @xmath12 , and is usually assumed to give a relation between the electric displacement @xmath13 and field @xmath14 of the form @xmath15 for a uniaxial nematic , this is entirely equivalent to assuming that @xmath16 where @xmath17 the electric contribution to the thermodynamic potential @xmath18 is @xmath19 at the surfaces of the device we assume a pinning potential @xmath20 this corresponds to a director at the surface preferring to lie along the direction of the eigenvector of @xmath21 corresponding to the largest eigenvalue @xmath22 .
* equation of motion of the nematic order parameter : * the equation of motion for the nematic order parameter is@xcite @xmath23 where @xmath24 is a collective rotational diffusion constant . the first term on the left - hand side of equation ( [ qevolution ] ) is the material derivative describing the usual time dependence of a quantity advected by a fluid with velocity @xmath25 .
this is generalized by a second term @xmath26 where @xmath27 and @xmath28 are the symmetric part and the anti - symmetric part respectively of the velocity gradient tensor @xmath29 .
@xmath30 appears in the equation of motion because the order parameter distribution can be both rotated and stretched by flow gradients .
this is the consequence of the rod - like geometry of the liquid crystal molecules .
@xmath31 is a constant which will depend on the molecular details of a given liquid crystal .
the term on the right - hand side of equation ( [ qevolution ] ) describes the relaxation of the order parameter towards the minimum of the free energy .
the molecular field @xmath32 which provides the driving motion is related to the derivative of the free energy by @xmath33 where @xmath34 and @xmath35 is the kronecker delta . for ( [ h_elastic ] ) @xmath36 is assumed , and the symmetry and tracelessness of @xmath37 is also exploited for simplification . when computing the functional derivative in ( [ h(q ) ] )
surface terms arise from the integration by parts . for reference , these are @xmath38 + l_3 \sigma_\gamma q_{\gamma\epsilon}(\partial_\epsilon q_{\alpha\beta } ) \right\},\ ] ] where @xmath39 is the surface ( unit ) normal .
* continuity and navier - stokes equations : * the fluid momentum obeys the continuity @xmath40 and the navier - stokes equation @xmath41 where @xmath42 is the fluid density and @xmath43 is related to the viscosity .
the form of this equation is not dissimilar to that for a simple fluid .
however the details of the stress tensor reflect the additional complications of liquid crystal hydrodynamics .
there is a symmetric contribution @xmath44 and an antisymmetric contribution @xmath45 the background pressure @xmath46 is taken simply to be @xmath47 and is found to be a constant in our simulations to a very good approximation .
the maxwell stress tensor @xmath48 @xcite describes the stress due to the external electric field @xmath49
we now describe a lattice boltzmann algorithm which solves the hydrodynamic equations of motion of a liquid crystal ( [ qevolution ] ) , ( [ continuity ] ) , and ( [ ns ] ) .
lattice boltzmann algorithms are defined in terms of a set of continuous variables , usefully termed partial distribution functions , which move on a lattice in discrete space and time @xcite .
the simplest lattice boltzmann algorithm , which solves the navier - stokes equations of a simple fluid , is defined in terms of a single set of partial distribution functions which sum on each site to give the density . for liquid crystal hydrodynamics
this must be supplemented by a second set , which are tensor variables , and which are related to the tensor order parameter @xmath0 @xcite .
we define two distribution functions , the scalars @xmath50 and the symmetric traceless tensors @xmath51 on each lattice site @xmath52 .
each @xmath53 , @xmath54 is associated with a lattice vector @xmath55 .
we choose a nine - velocity model on a square lattice with velocity vectors @xmath56 .
physical variables are defined as moments of the distribution function @xmath57 the distribution functions evolve in a time step @xmath58 according to @xmath59 , \label{eq2}\\ & & { \bf g}_i({\vec x}+{\vec e}_i \delta t , t+\delta t)-{\bf g}_i({\vec x},t)= \nonumber\\ & & \qquad\qquad \qquad\qquad \qquad\qquad\frac{\delta t}{2}\left [ { \cal c}_{{\bf g}i}({\vec x},t,\left\{{\bf g}_i \right\})+ { \cal c}_{{\bf g}i}({\vec x}+{\vec e}_i \delta t , t+\delta t,\left\{{\bf g}_i^*\right\})\right ] . \label{eq3}\end{aligned}\ ] ] this represents free streaming with velocity @xmath55 and a collision step which allows the distribution to relax towards equilibrium . @xmath60 and @xmath61 are first order approximations to @xmath62 and @xmath63 respectively .
discretizing in this way , which is similar to a predictor - corrector scheme , means that lattice viscosity terms are eliminated . in a lattice boltzmann algorithm for a simple fluid these terms
can be added to the physical viscosity term . here
this is not the case : the lattice viscosity gives an additional , spurious term in the equations of motion .
an additional advantage of the predictor - corrector approach is that the stability of the lattice boltzmann algorithm is improved .
the collision operators are taken to have the form of a single relaxation time boltzmann equation@xcite , together with a forcing term @xmath64 the form of the equations of motion and thermodynamic equilibrium follow from the choice of the moments of the equilibrium distributions @xmath65 and @xmath66 and the driving terms @xmath67 and @xmath68 .
@xmath69 is constrained by @xmath70 where the zeroth and first moments are chosen to impose conservation of mass and momentum .
the second moment of @xmath71 controls the symmetric part of the stress tensor , whereas the moments of @xmath67 @xmath72 impose the antisymmetric part of the stress tensor . for the equilibrium of the order parameter distribution
we choose @xmath73 this ensures that the order parameter is convected with the flow .
finally the evolution of the order parameter is most conveniently modeled by choosing @xmath74 which ensures that the fluid minimizes its free energy at equilibrium .
conditions ( [ eq6])([eq9 ] ) can be satisfied as is usual in lattice boltzmann schemes by writing the equilibrium distribution functions and forcing terms as polynomial expansions in the velocity@xcite @xmath75 where @xmath76 identifies separate coefficients for different absolute values of the velocities .
a suitable choice is @xmath77 where any coefficients not listed are zero .
we now use the lattice boltzmann algorithm described in sections 2 and 3 to investigate a simple liquid crystal device .
the device consists of a nematic liquid crystal confined between two planes a distance @xmath78 apart . at the surface @xmath21
is chosen to give a surface tilt to the director of @xmath79 at @xmath80 and @xmath81 at @xmath82 .
at zero applied voltage these conditions result in a global minimum free energy state with a splayed director configuration , or h state as shown in figure [ fig_surf2](a ) . at high voltages , typically on order of @xmath83 ,
the h state is no longer the global minimum , and a bend configuration ( v state ) is obtained like the one shown in figure [ fig_surf2](b ) . at intermediate voltages ,
the v state is more relaxed , like the one shown in figure [ fig_surf2](c ) .
the v state may remain for some time even at zero voltage as it is metastable .
as the h and v states are topologically distinct , the transition from one to the other requires nucleation and the generation of defects .
this process is illustrated in figure [ fig_twodefects ] .
the simulations were performed on a @xmath84 grid .
we used @xmath85 , @xmath86 , @xmath87 , @xmath88 , @xmath89 , @xmath90 and @xmath91 . by a choice of appropriate length
, time and pressure scales these correspond to physical values values @xmath92 pn , @xmath93 pn , @xmath94 pn , @xmath95 @xmath1 which is consistent with the nematic liquid crystal e7 ( merck ltd .
uk ) at @xmath96 .
periodic boundary conditions were used in the @xmath97 direction and bounce - back in the @xmath98 direction . a phase diagram for the h and v states in the tilt angle / voltage plane
is shown in fig .
[ phasediag ] .
the system was started in a configuration like that shown figure [ fig_twodefects ] . for low fields
the h domain has a lower free energy density , thus it grows ( crosses ) . for high enough fields the h domain shrinks ( circles ) .
the boundary is defined as the voltage at which the two domains have the same free energy .
this diagram compares well with those measured experimentally for a system using e7 @xcite .
consider the dynamics of the domain growth in the system . rather than fix parameters to a specific material ,
we wish to explore how the dynamics are affected by different parameters .
in particular , we are interested in how hydrodynamics affects the speed of the domain growth .
this is motivated by the observation in ref .
@xcite that the domain growth was anisotropic and the speculation that this may be due to hydrodynamics . to isolate effects due to different elastic constants , for these simulations we set @xmath99 , the so - called one - elastic constant approximation .
the initial configuration , depicted in fig . [ fig_twodefects](a ) , is a horizontal ( i.e. , along the @xmath97-direction ) domain in an otherwise vertically aligned state .
this models a time shortly after the electric field has been switched off when small but macroscopic domains have formed in the device .
once the simulation commences the director configuration relaxes rapidly to that shown in fig . [
fig_twodefects](b ) .
the horizontal and vertical domains are both tilted slightly by the elastic coupling to the surface spins
. then defects are formed at the domain boundaries , which merge to form single defects with strengths @xmath100 at the centre of each domain wall respectively .
once the two defects have formed the vertical domain begins to grow and the defects move in opposite directions .
our aim is to investigate the effect of back - flow and surface director tilt on this motion .
most previous work on the dynamics of topological defects has ignored the back - flow @xcite .
we first investigate the effect of the surface tilt @xmath101 on the defect speed .
the horizontal domain grows because it decreases the free energy of the device .
it can be shown @xcite that the difference in free energy between the horizontal and vertical domains is proportional to @xmath102 , where @xmath101 is the angle of the surface tilt to the @xmath97-axis .
thus as @xmath101 increases , the free energy difference decreases , and the defects move more slowly . at @xmath103
the two domains have the same free energy and the defects stop moving .
for @xmath104 the horizontal domain begins to shrink .
a particular advantage of the simulation is that the back - flow can easily be switched off by setting @xmath105 and @xmath106 to zero in ( [ expansion_in_u ] ) and ( [ eq14 ] ) .
thus the effect of the back - flow can be unambiguously identified .
since no flow is imposed , this leads to a zero velocity field throughout the whole simulation .
the dynamical equation for the system in this case can be obtained from ( [ qevolution ] ) by setting @xmath25 to zero .
it corresponds to a simpler model which does not include hydrodynamics .
consider first the diamonds in fig .
[ fig_defectspeed_tilt ] .
this corresponds to the case with back - flow switched off . for this case
both defects move at the same speed ( but in opposite directions ) .
this can be explained by considering a local co - ordinate transformation @xmath107 which is equivalent to changing the sign of the off - diagonal elements in @xmath0 .
assuming @xmath108 , the dynamical equation ( [ qevolution ] ) is invariant under this transformation whereas the two defects in fig .
[ fig_twodefects](b ) transform into each other .
the triangles and circles in fig . [ fig_defectspeed_tilt ]
show the velocity of the defects when back - flow is included in the model .
now the two defects move with different speeds since the @xmath109 elements of the stress tensors ( [ bestress ] ) and ( [ as ] ) are not invariant under the @xmath107 transformation .
thus the stress fields are different for the two defects . due to this stress field
two different flow fields are formed around the defects ( see fig .
[ fig_velocityfield ] ) . the flow is stronger around the @xmath110 defect and
it is substantially accelerated .
the speed of the @xmath111 defect is affected much less by the flow . an experimental setup similar to fig .
[ fig_twodefects ] was considered in @xcite where the growth of a circular twist domain in a horizontal environment under the influence of an external field was studied for different surface tilt values .
it was found that the domain growth was anisotropic .
although this problem is three - dimensional , a vertical cross section through the growing domain gives a geometry similar to that considered here .
it seems extremely plausible , that the essential physics is captured by our model with the anisotropy in growth resulting from the difference in the flow fields corresponding to different defect topologies at the defect boundary .
quantitative mappings between the experiment and the simulation results must be treated with caution because of the different dimensionality . however , in both the simulation and the experiment the difference between the wall speeds was about @xmath112 .
the simulations were performed on a @xmath113 grid .
we used @xmath114 , @xmath115 , @xmath116 , @xmath91 , @xmath117 , @xmath118 and @xmath119 . the larger @xmath120 corresponds to increasing the resolution of the lattice in order to perform accurate defect velocity measurements .
in this paper we have described a lattice boltzmann algorithm to simulate the dynamics of a non - newtonian fluid , liquid crystals .
in particular we have considered the the presence of an external electric field and the possibility of a general frank free energy with three elastic constants . in the continuum limit
we recover the beris - edwards formulation within which the liquid crystal equations of motion are written in terms of a tensor order parameter @xcite .
the equations are applicable to the isotropic , uniaxial nematic , and biaxial nematic phases . working within the framework of a variable tensor order parameter it is possible to simulate variations in the magnitude of order and hence the dynamics of topological defects and non - equilibrium phase transitions between different flow regimes .
the algorithm was used to explore states in nematic liquid crystal devices .
we find that switching between topologically distinct states is , as expected , strongly inhibited by a free energy barrier between the initial and the final states . in this case
the state can only change once a domain of the final state has been nucleated .
defects form at the moving domain walls and hence this geometry allows an investigation of defect hydrodynamics .
we find that a @xmath121 defect is substantially speeded up by back - flow effects .
this is relevant to device physics where the speed of switching is an important design variable .
there are many directions for further research opened up by the rich physics inherent in liquid crystal hydrodynamics and the generality of the beris - edwards equations .
for example , another possible switching pathway would be via the escape of the director into a third dimension .
the addition of flexoelectric terms to the equations of motion will allow problems relevant to bistable liquid crystal displays to be addressed .
even cases where the switching is the consequence of back - flow @xcite can be modeled .
numerical investigations are vital to address these problems because of the complexity of the equations of motion . fig
. 1 alignment of molecules for a tilt angle @xmath79 on the top surface and @xmath81 on the bottom surface .
director configuration when @xmath122 and ( a ) the field is switched off and the system has had time to relax to its global minimum ; ( b ) the field is switched on the a fairly high voltage @xmath123 ; and ( c ) the field is at a low voltage @xmath124 or lower .
the system may remain in the metastable state ( c ) for some time even at zero voltage .
( a ) initial director configuration used to study domain growth .
after the vertical electric field is switched off , a horizontal domain is created in order to nucleate the growth process .
( b ) shortly thereafter two defects are formed at the boundary of the horizontal and vertical domains .
the left ( right ) defect has a topological strength @xmath125 ( @xmath126 ) .
both the horizontal and the vertical domains are slightly distorted due to the defects and the surface tilt .
3 . phase diagram for the h and v states as a function of tilt angle and voltage . at low fields ( crosses )
the h domain has a lower free energy density , thus it grows . if the field strong enough ( circles )
, the h domain shrinks . on the line
the two domains have equal free energy density .
4 . the velocity of the two defects as the function of surface tilt if back - flow is ignored ( diamonds ) and included .
notice that if back - flow is not included then the two defects move with the same speed .
hydrodynamics accelerates the @xmath126 defect ( triangles ) substantially , while it affects less the @xmath125 defect ( circles ) .
5 . ( a ) velocity field corresponding to the director configuration shown in fig .
[ fig_twodefects](b ) .
notice that there is a strong vortex structure around the @xmath126 defect and at the defect core it points in the direction of defect movement .
the flow at the core of the other defect is much weaker .
( b ) flow field corresponding to the dashed box . for better visibility the length of the vectors
is changed disproportionally .
|
we extend a lattice boltzmann algorithm of liquid crystal hydrodynamics to include an applied electric field .
the approach solves the equations of motion written in terms of a tensor order parameter .
back - flow effects and the hydrodynamics of topological defects are included .
we investigate some of the dynamics relevant to liquid crystal devices ; in particular defect - mediated motion of domain walls relevant to the nucleation of states useful in pi - cells .
an anisotropy in the domain wall velocity is seen because defects of different topology couple differently to the flow field .
+ * key words : * lattice boltzmann ; liquid crystals ; complex fluids .
|
about 25% of all gamma - ray bursts detected by batse were `` short '' ( duration less than 2 s , * ? ? ? * ) and had hard spectra . due to the non - detection of an afterglow from a short burst , rendering quick and accurate localization unfeasible ,
their nature have remained elusive .
grb @xmath0b was detected by the _
swift _ burst alert telescope ( bat ) on 2005 may 9 at 04:00:19.23 ( ut ) @xcite and upon slewing the _ swift _ x - ray telescope ( xrt ) started observations of the burst 62 s after the bat was triggered .
for the first time , an x - ray afterglow from a short burst was detected enabling a localization within @xmath3 @xcite .
imaging of the xrt error box showed that it overlapped with an elliptical galaxy with a redshift of @xmath4 @xcite .
based on the unlikeliness of a chance alignment between grb @xmath0b and such a galaxy it has been argued that this is the host galaxy of the burst @xcite .
however , several much fainter ( and presumably more distant ) galaxies were also detected within the xrt error circle ( see fig.1 in * ? ? ?
a unique feature of grb @xmath0b is that it is situated in the direction of a nearby cluster of galaxies , identified in the zwicky catalog ( zwcl 1234.0 + 02916 , * ? ? ?
* ) and in the northern sky optical cluster survey with a photometric redshift of @xmath5 ( nsc j123610 + 285901 , * ? ? ?
x - ray emission from a hot intracluster medium in zwcl 1234.0 + 02916 is clearly detected by the xrt from which @xcite determined the cluster x - ray centroid and temperature . if grb @xmath0b is hosted by one of the fainter galaxies beyond the cluster the emission from the burst is inevitably boosted by gravitational lensing .
this could enhance the probability of detecting the burst and imply a ( possibly large ) correction to the derived burst energetics .
here we ( i ) derive the x - ray flux , temperature and mass of the cluster zwcl 1234.0 + 02916and estimate the probability of a chance alignment of a gamma - ray burst and a cluster with an x - ray flux at least as large as that of zwcl 1234.0 + 02916 , and ( ii ) calculate the gravitational lensing magnification within the grb @xmath0b xrt error circle , constraining scenarios where grb @xmath0b is gravitationally lensed .
a cosmology with @xmath6 km s@xmath7 mpc@xmath7 , @xmath8 , @xmath9 is assumed throughout this letter .
all errors quoted are 90% confidence limits .
the _ swift _ xrt @xcite observed the grb @xmath0b region for a duration of 75 ks .
the x - ray afterglow of grb @xmath0b is detected during the first 400 s of the xrt observation .
we investigated the x - ray emission from an intracluster medium in the cluster zwcl 1234.0 + 02916 , by analyzing xrt data obtained after the first 400 s where the grb @xmath0b x - ray afterglow had faded below detection . only data taken in `` photon counting mode '' , providing full imaging and spectroscopic information , were included in the analysis yielding 31.9 ks of exposure time . using the _ xrtpipeline _ procedure version 0.8.8 in the heasoft 6.0 software package we applied standard screening parameters and the latest available calibration files ( 2005 june 1 ) to the level 1 data obtained from the _ swift _ quick - look data compilation to produce cleaned event lists of the full xrtfield of view .
subsequently , the _ xselect _ tool in the heasoft 6.0 package was used to produce images and spectra with the spatial and spectral filters given below .
an image was obtained in the 0.310 kev band and sources were detected using a wavelet algorithm ( _ wavdetect _ implemented in the chandra interactive analysis of observations ( ciao ) version 3.2 software package ) .
the astrometry was fixed in the following way : we measured the centroid of the x - ray afterglow in the _ swift _ frame of reference from the first 400 s of the xrt observations .
then the x - ray afterglow centroid was adjusted to match the position determined by @xcite and @xcite ( who registered a 50 ks _ chandra _ image to 2mass coordinates , and then registered a 30 ks xrt image to the _ chandra _ image ) .
hence we apply a shift to the _ swift _ reference frame of @xmath10r.a.(j2000)@xmath11 and @xmath10decl.(j2000)@xmath12 . in order to characterize the intracluster emission , two point sources near the cluster
were excluded and the source regions were filled with a poissonian pixel value distribution sampled from the surroundings ( using the _ dmfilth _ procedure in ciao 3.2 ) .
finally , we ran the _ wavdetect _ procedure on the resulting image for determining the centroid of the intracluster emission .
x - ray spectra were generated from events in the 0.57 kev range within a circular region centered on the cluster x - ray centroid and with a radius between @xmath13 and @xmath14 ( corresponding to the radius at which the intracluster emission is @xmath15 times above the background level or equals the background level , respectively ) .
the energy bounds were chosen so as to leave out the lowest energies where the spectral calibration is most uncertain , and to disregard the highest energies dominated by background .
background spectra were generated from circular regions devoid of sources , not overlapping with the cluster extraction region or each other , and with the same radius as used for the intracluster emission .
the average background level is @xmath16 of the intracluster emission .
we used the _ xspec _ package version 12.2.0 @xcite to fit background subtracted spectra of the intracluster emission , binned to a minimum of 20 counts per bin , with a hot plasma model ( mekal , * ? ? ?
* ; * ? ? ?
* ) absorbed by neutral matter along the line of sight .
the galactic absorbing column density in this direction is @xmath17 @xmath18 @xcite .
the intracluster x - ray emission centroid is r.a.(j2000)@xmath19 , + decl.(j2000)@xmath20 ( with an uncertainty of @xmath21 ) .
this is @xmath22 off - set from the center of the xrt error circle .
the diffuse emission extends to about @xmath23 corresponding to 430 kpc for a cluster redshift of @xmath4 @xcite .
327 net counts ( 0.310 kev ) are detected within @xmath14 of the x - ray centroid . with the spatial resolution achieved by the photon statistics the morphology of the intracluster emission is close to circular with a slight ne - sw elongation in the cluster outskirts ( see fig.[fig1 ] ) .
observations obtained with the _ chandra x - ray observatory _ can further reveal the structure of the intracluster medium @xcite .
the near coincidence between the x - ray centroid and several bright elliptical galaxies strongly suggests that this is the bottom of the cluster potential well rather than the optical center of zwcl 1234.0 + 2916 ( r.a.(j2000)@xmath24 , decl.(j2000)@xmath25 , * ? ? ? * ) or nsc j123610 + 285901 ( r.a.(j2000)=@xmath26 , decl.(j2000)=@xmath27 , * ? ? ?
* ) . at the cataloged optical cluster centers no enhancement in x - ray emission
is seen .
the spectrum of the intracluster medium within a radius of @xmath14 of the x - ray centroid is well fitted ( @xmath28 for 21 d.o.f . ) by a mekal model with absorption fixed at the galactic value and a typical abundance of @xmath29 @xmath30 @xcite , see fig .
the best fit temperature is @xmath31 kev and the 0.52 kev flux @xmath32 erg @xmath18 s@xmath7 .
the main systematic error in determining the temperature and flux is the background subtraction . using four different background regions results in best fit temperatures between @xmath33 kev and @xmath34 kev , and 0.52 kev fluxes between @xmath35 erg @xmath18 s@xmath7 and @xmath36 erg @xmath18 s@xmath7 .
we also investigated the influence on the best temperature from varying ( i ) the cluster emission extraction radius , ( ii ) the absorbing column density , and ( iii ) the intracluster abundance .
fitting the spectrum extracted within a radius of @xmath13 of the cluster x - ray centroid yields a temperature of @xmath37 kev . letting the absorbing column density
be a free parameter in the fit results in a best fit value of @xmath38 @xmath18 and the best fit temperature is @xmath39 kev .
however , given that only a small amount of extra - galactic absorption is expected @xcite , and that the galactic absorbing column density is 50% uncertain in worst case , a maximum column density of @xmath40 @xmath18 is expected . for this column density ,
the best fit temperature is @xmath41 kev .
the abundance is constrained to be @xmath42 @xmath30 and varying the abundance in the range @xmath43 @xmath30 ( corresponding to the observed abundance variation in clusters , * ? ? ?
* ) changes the best fit temperature by only @xmath44 . from the above analysis
we conclude that the global temperature of the zwcl 1234.0 + 02916intracluster medium is @xmath45 kev where the latter uncertainty is due to systematics from background subtraction . within the errors
this is consistent with the temperature derived by @xcite ( @xmath46 kev ) . for a cluster redshift of @xmath4
the 0.12 kev luminosity is @xmath47 erg s@xmath7 which is in good agreement with the expected luminosity from the empirical luminosity temperature relation of clusters @xcite .
the x - ray properties of zwcl 1234.0 + 02916 thus shows it to be a cluster intermediate to the virgo and coma clusters .
the cluster mass can be estimated from an empirical mass x - ray temperature relation . using the x - ray determined mass temperature relation of @xcite for their full sample we find @xmath48 m@xmath49 taking into account the uncertainties in the mass temperature relation and the uncertainties in the temperature .
@xmath50 is the mass within the radius at which the mean cluster density is 500 times the critical density of the universe . from the optically determined mass
temperature relation of @xcite we find @xmath51 m@xmath49 in excellent agreement with the above mass estimate .
the alignment of the grb @xmath0b xrt error circle and a bright elliptical galaxy , g1 , in the cluster zwcl 1234.0 + 02916 makes it relevant to consider the possibility that grb @xmath0b is a gravitationally lensed background source @xcite .
hence , we have calculated the gravitational lensing magnification from g1 and the cluster for several possible redshifts of grb @xmath0b .
the gravitational lensing calculations were carried out using the _ gravlens _ software package @xcite .
we model the galaxy as a singular isothermal ellipsoid with an axis ratio of 0.81 @xcite with the semimajor axis aligned along a position angle @xmath52 , a velocity dispersion of @xmath53 km s@xmath7 @xcite , and the cluster as a singular isothermal sphere with @xmath54 m@xmath49 ( see section 3 ) .
we take the center of the xrt error circle to be at r.a.(j2000)@xmath55 and decl.(j2000)@xmath56 @xcite and choose that as the center of our coordinate system .
we place the center of the galaxy g1 at r.a.(j2000)=@xmath57 and decl.(j2000)@xmath58 and the center of the cluster at the x - ray centroid of the diffuse cluster emission ( see section 3 ) .
below we give results for point sources at the expected lower and upper limit on the grb @xmath0b redshift @xmath59 and @xmath60 , respectively . due to the blue continuum and lack of emission lines in the faint galaxies in the xrt error box ,
the lower limit on their redshift is @xmath61 @xcite
. the upper limit on the redshift is motivated by models predicting an intrinsic duration of a grb longer than 8 ms @xcite .
we note that chosing a larger upper limit does not affect our conclusions .
taking uncertainties in the velocity dispersion of g1 into account this gives an einstein radius for g1 of @xmath62 arcsec for @xmath60 and @xmath63 arcsec for @xmath59 .
similarly we get for the cluster ( taking the uncertainty in the cluster mass into account ) that @xmath64 arcsec for @xmath60 and @xmath65 arcsec for @xmath59 .
we then calculate the median magnification within the xrt error circle of radius @xmath66 @xcite , see fig .
most of the magnification is provided by the galaxy g1 , and the main effect of the cluster is to increase the magnification in the east - west direction .
the xrt error circle crosses the critical curves ( paths in the image plane where a point source will be exposed to infinite magnification ) for nearly all mass models of g1 and the cluster .
the median of the magnification within the error circle is a factor of @xmath67 for @xmath60 and @xmath68 for @xmath59 .
we thus find that potential the magnification of grb @xmath0b is typically a factor 12 , fairly independent of the redshift of grb @xmath0b .
however , if grb @xmath0b is positioned within @xmath69 of g1 it could be more strongly magnified .
we note that the source revealed by subtracting away g1 @xcite , situated @xmath70 from the center of g1 , is not a strongly lensed background source .
from our gravitational lensing study we conclude that lensing is not strongly boosting the probability that grb @xmath0b is situated in a background galaxy . furthermore , @xcite estimate that the probability of a chance alignment between grb @xmath0b and g1 is @xmath71 ( using the apparent magnitude of g1 ) while @xcite arrive at probability of @xmath72 ( using the luminosity and distance of g1 ) .
interestingly , a correlation between error boxes of short bursts in the batse catalog and the positions of galaxies in the local universe has been claimed @xcite .
we now proceed to estimate the probability of a chance alignment between the xrt error circle and a cluster of galaxies with an x - ray flux and optical richness at least as large as that of zwcl 1234.0 + 02916 .
the sky surface density of clusters to a 0.52 kev flux limit of the zwcl 1234.0 + 02916 flux is well determined to 0.51 clusters deg@xmath73 from several x - ray cluster surveys @xcite .
the probability for a chance alignment within @xmath74 of the cluster center and the grb @xmath0b error circle center is equal to the sky coverage fraction of such clusters with a radius of @xmath74 , i.e. @xmath75 .
the sky surface density of clusters in the northern sky optical cluster survey with a richness at least as large as nsc j123610 + 285901 ( @xmath76 , corresponding to less than abell richness class 0 ) is @xmath77 deg@xmath73 @xcite , yielding a chance alignment probability of @xmath78 .
although a posteriori statistics is always uncertain grb @xmath0b comprises the first strong case for a short burst located in a galaxy cluster , based on either the x - ray flux or the optical richness of zwcl 1234.0 + 02916 .
the association of grb @xmath0b with zwcl 1234.0 + 02916 provides independent evidence that grb @xmath0b is also associated with the galaxy g1 since this is a cluster member .
interestingly , it has recently been suggested that the line - of - sight towards the short burst grb 050813 coincides with a cluster at redshift @xmath79 @xcite .
previously , studies of the few small error boxes of short bursts have not revealed any short burst associated with a cluster @xcite a statistical correlation between gamma - ray burst error boxes ( of short as well as long bursts ) and the position of abell clusters has been claimed @xcite , but subsequent refined analyses have failed to confirm this @xcite .
however , abell clusters are relatively nearby and rich so a correlation between short bursts and abell clusters is only expected if short bursts occur predominantly within the completeness redshift of abell clusters ( @xmath80 , * ? ? ?
the likely association between grb @xmath0b and the cluster zwcl 1234.0 + 02916 motivates further studies of the spatial correlation between galaxy clusters and short burst sky positions .
since zwcl 1234.0 + 02916 is less rich than abell clusters it would be beneficial to use cluster catalogs including poorer and more distant clusters .
such studies would provide interesting constraints on the environments of short bursts and hence on their origin . if , as suggested by the favoured models @xcite , short bursts originate from merging of compact stellar remnants their progenitors are @xmath81 yr old @xcite .
targeting the environment of short bursts is thus potentially a new independent way for tracing the old stellar population throughout the universe . up to now
, entire populations of old stars in galaxies have been identified through their integrated optical / near - infrared colors .
short bursts are promising pointers to sites of individual old stars , revealing the spatial distribution of old stars , be they located in cluster ellipticals , field galaxies , or even in intergalactic space @xcite .
the dark cosmology centre is supported by the dnrf .
kp , jp , and jmcc acknowledge support from instrument centre for danish astrophysics , and jmcc acknowledges support from nbi s international ph.d .
school of excellence .
jg acknowledges is partially supported by the spanish ministry of science and education through programmes esp2002 - 04124-c03 - 01 and aya2004 - 01515 ( including feder funds ) .
this work was supported by the european community s sixth framework marie curie research training network programme , contract no .
mrtn - ct-2004 - 505183 `` angles '' .
the authors acknowledge benefits from collaboration within the ec fp5 research training network `` gamma - ray bursts an enigma and a tool '' .
|
the first arcsecond localization of a short gamma - ray burst , grb @xmath0b , has enabled detailed studies of a short burst environment .
we here report on studies of the environment of grb @xmath0b using the _ swift _ x - ray telescope ( xrt ) .
the xrt error circle of the burst overlaps with an elliptical galaxy in the cluster of galaxies zwcl 1234.0 + 02916 .
based on the measured x - ray flux of the cluster we estimate that the probability for a chance superposition of grb @xmath0b and a cluster at least as x - ray bright as this cluster is @xmath1 , presenting the first strong case of a short burst located in a cluster of galaxies .
we also consider the case for grb @xmath0b being located behind zwcl 1234.0 + 02916 and gravitationally lensed . from the velocity dispersion of the elliptical galaxy and the temperature of hot intracluster gas , we model the mass distribution in the elliptical galaxy and the cluster , and calculate the gravitational lensing magnification within the xrt error circle .
we find that , if grb050509b would be positioned significantly behind the cluster , it is most likely magnified by a factor less than two , but that the burst could be strongly lensed if it is positioned within @xmath2 of the center of the bright elliptical galaxy .
further mapping of arcsecond size short burst error boxes is a new promising route to determine the spatial distribution of old stars throughout the universe .
|
the introduction to this paper is in two parts .
the first subsection defines resonance and gives our prototypical example on alternating sign matrices .
the second subsection describes our main results ; these include two instances of resonance ( on plane partitions and increasing tableaux ) , an equivariant bijection between these two sets with a number of new consequences , and a higher - dimensional analogue of n. williams and the third author s result on the equivariance of ( poset-)promotion and rowmotion @xcite .
we introduce the following concept of _ _
resonance__. suppose @xmath2 is a cyclic group acting on a set @xmath3 , @xmath4 a cyclic group of order @xmath5 acting nontrivially on a set @xmath6 , and @xmath7 a surjection .
we say the triple @xmath8 exhibits * resonance with frequency @xmath5 * if , for all @xmath9 , @xmath10 , that is , the following diagram commutes : \(a ) @xmath3 ; ( b ) @xmath3 ; ( c ) @xmath6 ; ( d ) @xmath6 ; ( a ) edge[pil ] node[below]@xmath11 ( b ) ; ( a ) edge[pil ] node[right]@xmath12 ( c ) ; ( b ) edge[pil ] node[right]@xmath12 ( d ) ; ( c ) edge[pil ] node[below]@xmath13 ( d ) ; in our examples , @xmath6 will be either a set of planar diagrams with @xmath14 acting by rotation or a set of words with @xmath14 acting by a cyclic shift .
resonance is a _ pseudo - periodicity _ property of the @xmath15-action , in that the resonant frequency @xmath5 is generally less than the order of the @xmath15-action .
note that , in general , @xmath16 satisfies the definition of resonance with frequency @xmath17 ; we call this an instance of _ trivial resonance_. our prototypical example of resonance is the action of _ gyration _ on _ alternating sign matrices_. an * alternating sign matrix ( asm ) * is a square matrix with the following properties : the entries are in the set @xmath18 , each row and each column sums to one , and the nonzero entries along each row or column alternate in sign .
let @xmath19 denote the set of @xmath20 alternating sign matrices .
alternating sign matrices were introduced by d. robbins h . rumsey @xcite as part of their study of the lambda - determinant . with w. mills
@xcite , they then conjectured an enumeration for @xmath19 , which was proved by d. zeilberger @xcite and g. kuperberg @xcite ( cf .
@xcite for a detailed exposition of this history ) .
alternating sign matrices are known to be in bijection with _ fully - packed loop configurations _
@xcite ; see figures [ fig : fplex ] and [ fig : fplex2 ] .
[ def : fpl ] consider an @xmath21 \times [ n]$ ] grid of dots in @xmath22 .
beginning with the dot at the upper left corner , draw an edge from that dot up one unit .
then go around the grid , drawing such an external edge at every second dot ( counting corner dots twice , since at the corners , external edges could go in either of two directions ) . a * fully - packed loop configuration ( fpl ) * of order
@xmath1 is a set of paths and loops on the @xmath21 \times [ n]$ ] grid with boundary conditions as described above , such that each of the @xmath23 vertices within the grid has exactly two incident edges .
there is a ( non - injective ) map from fully - packed loop configurations to their _ link patterns_. [ def : lp ] given a fully - packed loop configuration , number the external edges clockwise , starting with the upper left external edge .
each external edge will be connected by a path to another external edge , and these paths will never cross . this matching on the external edges is a noncrossing matching on @xmath24 points , and is called the * link pattern * of the fpl .
consider the following action on fully - packed loop configurations .
[ def : fplgyr ] given an @xmath21\times[n]$ ] grid of dots , color the interiors of the squares in a checkerboard pattern . given an fpl of order @xmath1 drawn on this grid , its * gyration * , gyr ,
is computed by first visiting all squares of one color then all squares of the other color , applying at each visited square the `` local move '' that swaps the edges around a square if the edges are parallel and otherwise leaves them fixed . the following theorem of b. wieland gives a remarkable property of gyration .
gyration of an fpl rotates the associated link pattern by an angle of @xmath25 .
we reformulate this theorem into a statement of resonance .
let @xmath12 be the map from an alternating sign matrix through its fpl to the link pattern . then , @xmath26 exhibits resonance with frequency @xmath24 . for example , consider gyration on @xmath27 alternating sign matrices . gyration has orbits of size 2 , 4 , 5 , and 10 .
so the order of gyration in this case is 20 , but @xmath28 , gyr , @xmath29 exhibits resonance with frequency @xmath30 .
consider the orbit of gyration in figure [ fig : fplex ] .
this orbit is of size 4 , while the link pattern orbit is of size 2 .
so even though @xmath31 for @xmath32 an asm in this orbit , @xmath33 ( since , in this case , @xmath34 ) . as another example , the asm in figure [ fig : fplex2 ] is in a gyration orbit of size @xmath35 ( @xmath36 ) , while @xmath37 , gyr , @xmath29 exhibits resonance with frequency @xmath38 .
so even though @xmath39 for @xmath32 an asm in this orbit , @xmath40 .
@xmath41 asm with gyration orbit of length 84 , with its corresponding fpl and link pattern ] we think of the property of resonance as somewhat analogous to the _ cyclic sieving phenomenon _ ( introduced by v. reiner d .
stanton d .
white @xcite , generalizing the @xmath42 phenomenon of j. stembridge @xcite ) and the _ homomesy property _
( isolated by j. propp t .
roby @xcite , inspired by observations of d. panyushev @xcite ) in being a somewhat subtle ` niceness ' property of a cyclic group action .
we suspect that the phenomenon of resonance , like those of cyclic sieving and homomesy , is significantly more common than previously realized .
heuristically , one is led to suspect the presence of resonance in a system by observing that many orbit cardinalities are multiples or divisors ( or multiples of divisors ) of @xmath5 . the remainder of this paper centers around two new examples of resonance on _ plane partitions _ under _ rowmotion _ and _ increasing tableaux _ under _ @xmath0-promotion _ , as well as a new equivariant bijection relating these phenomena . here
we summarize our main results ; see the referenced sections for relevant definitions .
rowmotion has attracted much attention since it was first studied ( under another name ) by a. brouwer a .
schrijver @xcite in 1974 ; see for example @xcite .
more recently , several authors have studied a _
birational _ lift of rowmotion @xcite , with some relations to zamolodchikov periodicity .
let @xmath43 denote the set of plane partitions inside an @xmath44 box and @xmath45 denote rowmotion ; see section [ sec : resonance_pp ] for the definitions of @xmath46 and d. our first main resonance result is the following .
@xmath47 exhibits resonance with frequency @xmath48 . to better study plane partitions , we introduce and develop the machinery of _ affine hyperplane toggles _ and _ @xmath1-dimensional lattice projections
_ , including a higher - dimensional analogue of n. williams and the third author s result on the equivariance of ( poset-)promotion and rowmotion @xcite .
we obtain a large family of toggling actions @xmath49 whose orbit structures are equivalent to that of rowmotion .
see sections [ subsec:3dprorow ] and [ subsec : provpi ] for further details .
let @xmath50 be a finite poset with an @xmath1-dimensional lattice projection @xmath51 .
let @xmath52 and @xmath53 , where @xmath54 .
finally suppose that @xmath55 and @xmath56 are bijections .
then there is an equivariant bijection between @xmath57 under @xmath58 and @xmath57 under @xmath59 .
our other main object of study is @xmath0-promotion on increasing tableaux .
this operator was first described by the second author @xcite .
it is an variant of m .-
schtzenberger s promotion operator @xcite built on the @xmath0-_jeu de taquin _ that was introduced by h. thomas a .
yong @xcite to study @xmath0-theoretic schubert calculus .
@xmath0-promotion has been further studied in @xcite .
let @xmath60 denote the set of increasing tableaux of shape @xmath61 and entries at most @xmath62 , and let con be the content map . in section 2.2 , we prove the following , our first main result on @xmath0-promotion .
@xmath63 exhibits resonance with frequency @xmath62 .
this similarity of theorems 2.2 and 3.10 leads us to establish an equivariant bijection between plane partitions under rowmotion and increasing tableaux under @xmath0-promotion .
@xmath64 under @xmath45 is in equivariant bijection with @xmath65 under @xmath66 .
part of our approach to establishing this equivariant bijection involves the reinterpretation of @xmath0-promotion in terms of @xmath0-bender - knuth involutions , which we introduce ; see proposition [ prop : kbk ] .
we also extend , in section [ sec : descent_cycling ] , a result of b. rhoades on _ descent cycling _ to the @xmath0-promotion setting .
we obtain a variety of corollaries of this equivariant bijection .
many of these corollaries are new proofs of previously discovered results on the order of @xmath45 and @xmath66 .
we highlight here only those results that are new .
the order of @xmath66 on @xmath67 is @xmath68 .
the order of @xmath66 on @xmath69 is @xmath70 .
we also obtain the following strengthening of a theorem of p. cameron d .
fon - der - flaass ( * ? ? ?
* theorem 6(a ) ) .
the original theorem had the more stringent hypothesis @xmath71 .
if @xmath48 is prime and @xmath72 , then the cardinality of every orbit of @xmath45 on @xmath64 is a multiple of @xmath48 . *
the rest of this paper is structured as follows . * in section [ sec : increasing_tableaux ] , we recall the @xmath0-promotion operator on increasing tableaux and establish a number of new properties ( including resonance ) that we will use . in section [ sec : toggle ] , we establish resonance of plane partitions under rowmotion , and extend machinery developed by n. williams and the third author @xcite , to introduce the family of toggle group actions @xmath73 and show that each @xmath74 acts with the same cycle structure as rowmotion .
in section [ sec : pp_inctab ] , we give an equivariant bijection between increasing tableaux under @xmath0-promotion and plane partitions under @xmath75 and @xmath45 .
we then extract a number of corollaries from this equivariant bijection , including new proofs of theorems of a. brouwer a .
schrijver @xcite and p. cameron d .
fon - der - flaass @xcite , a strengthening of a theorem of p. cameron d .
fon - der - flaass @xcite , and several new results on the order of @xmath0-promotion .
finally , we conjecture the order of rowmotion on plane partitions of height @xmath76 ( which we have shown to be also the order of @xmath0-promotion on certain classes of increasing tableaux ) . in section [ subsec :
conj ] , we propose additional instances of resonance related to alternating sign matrices and _ totally symmetric self - complementary plane partitions_.
in this section , we study increasing tableaux , the first of the objects in our main bijection ( theorem [ thm : mainbij ] ) . after recalling the basic concepts , we establish resonance of increasing tableaux under @xmath0-promotion in corollary [ cor : resonance_of_increasing_tableaux ] . in section [ sec : k - bk ] , we reinterpret @xmath0-promotion in terms of @xmath0-bender - knuth involutions , which we introduce ; this interpretation plays an important role in section [ sec : equivariant_bijection ] in establishing equivariance of our main bijection . in section [ sec : descent_cycling ] , we extend a _ descent cycling _ result of b. rhoades ( * ? ? ?
* lemma 3.3 ) from standard young tableaux to increasing tableaux ; this extension is used in theorem [ thm : our_cfdf ] to improve on a theorem of p. cameron d . fon - der - flaass @xcite .
identify a partition @xmath61 with its young diagram .
( throughout this paper , we use the english orientation on young diagrams . ) an * increasing tableau * of shape @xmath61 is a filling of @xmath61 with positive integers such that labels strictly increase from left to right across rows and from top to bottom down columns .
an example appears in figure [ fig : an_increasing_tableau ] .
we write @xmath60 for the set of all increasing tableaux of shape @xmath61 with all entries at most @xmath62 .
( in contrast to other definitions that have appeared in the literature , we do not assume here that every integer between @xmath77 and @xmath62 appears . )
increasing tableaux have appeared in various contexts within algebraic combinatorics .
most notably for our purposes , h. thomas a .
yong introduced @xcite a _ @xmath0-jeu de taquin _
algorithm for increasing tableaux , which they applied to _ @xmath0-theoretic schubert calculus _ , obtaining littlewood - richardson rules for the grothendieck rings of algebraic vector bundles over _
grassmannians_. this algorithm has been has been extended to the @xmath0-theory of a wider variety of spaces by @xcite , as well as to the _ torus - equivariant _ @xmath0-theory of grassmannians @xcite . in @xcite ,
the second author studied a @xmath0-promotion operator , analogous to that of m .-
schtzenberger for semistandard tableaux @xcite , but using @xmath0-_jeu de taquin _ in place of ordinary _ jeu de taquin_. @xmath0-promotion has been further studied by j. bloom d .
saracino and the second author @xcite , t. pressey a .
stokke t .
visentin @xcite and b. rhoades @xcite .
* @xmath0-promotion * is defined as follows .
let @xmath78 .
delete all labels @xmath77 from @xmath79 .
( note there is at most one such label . )
consider the set of boxes that are either empty or contain @xmath80 .
this set naturally decomposes into connected components that are * short ribbons * , i.e. connected sets of boxes containing no @xmath81 subshape and with each column and row of length at most @xmath80 . for each such short ribbon containing more than one box
, we delete each label @xmath80 , while simultaneously placing @xmath80 in each empty box .
we do not make any change to short ribbons consisting of a single box .
now consider the set of boxes that are either empty or contain @xmath76 , and repeat the above process .
continue until all empty boxes are located at outer corners of @xmath61 . finally , label those boxes @xmath82 and then subtract @xmath77 from each entry .
the result is @xmath83 .
\(a ) @xmath84 ; ( b ) ; ( c ) ; ( d ) ; ( e ) ; ( f ) ; ( g ) ; ( h ) ; ( i ) @xmath85 ; ( a ) edge[pil ] node[above]delete @xmath77 s ( b ) ; ( b ) edge[pil ] ( c ) ; ( c ) edge[pil ] ( d ) ; ( d ) edge[pil ] ( e ) ; ( e ) edge[pil ] ( f ) ; ( f ) edge[pil ] ( g ) ; ( g ) edge[pil ] ( h ) ; ( h ) edge[pil ] node[above]fill and node[below]decrement ( i ) ; define the * content * of an increasing tableau @xmath78 to be the binary sequence @xmath86 , where @xmath87 if @xmath88 is an entry of @xmath79 and @xmath89 if it is not
. that is , @xmath90 where @xmath91 denotes the indicator function for the label @xmath88 .
[ prop : content_cycling ] let @xmath78 .
if @xmath92 , then @xmath93 is the cyclic shift @xmath94 .
case 1 : ( @xmath95 ) : then @xmath79 has no labels @xmath77 .
hence the first step of @xmath0-promotion is trivial , deleting no labels .
the ribbon switching process is also trivial , as there are no empty boxes .
therefore , at the final step , there are no boxes to fill .
thus , in this case , the total effect of @xmath0-promotion is merely to subtract @xmath77 from each entry .
the lemma is then immediate in this case .
case 2 : ( @xmath96 ) : then the first step of @xmath0-promotion is to delete a nonempty collection of labels @xmath77 .
hence there are a nonzero number of empty boxes .
the ribbon switching process may change the number of empty boxes , but clearly preserves its nonzeroness .
hence in the final step of @xmath0-promotion , there will be a nonzero number of boxes filled with @xmath82 and then decremented by @xmath77 .
hence @xmath97 .
let @xmath98 and suppose @xmath99 .
then @xmath88 appears as an entry of @xmath79 .
the ribbon switching process preserves this property ( though not in general the number of entries @xmath88 ) .
hence after subtracting one from each entry , this yields @xmath100 .
if instead @xmath101 , then @xmath88 does not appear in @xmath79 .
hence during the ribbon switching process , when we consider the ribbons consisting of @xmath88 s and empty boxes , each is a single empty box and by definition we make no change .
hence the ribbon switching process preserves the absence of @xmath88 .
after decrementing , this yields @xmath102 .
the following instance of resonance follows directly from lemma [ prop : content_cycling ] .
[ cor : resonance_of_increasing_tableaux ] @xmath63 exhibits resonance with frequency @xmath62 .
this leads to the following corollary .
[ cor : weakest_cfdflike_statement ] suppose @xmath62 is prime and @xmath78 does not have full content .
then the size of the @xmath0-promotion orbit of @xmath79 is a multiple of @xmath62 . in this subsection , we reinterpret @xmath0-promotion as a product of involutions , which we will need in our proof of theorem [ thm : main ] .
we define operators @xmath103 on @xmath60 for each @xmath104 .
take @xmath78 .
we compute @xmath105 as follows : consider the set of boxes in @xmath79 that contain either @xmath88 or @xmath106 .
this set decomposes into connected components that are short ribbons . on each nontrivial such component
, we do nothing . on each component that is a single box , replace the symbol @xmath88 by @xmath106 or _
vice versa_. the result is @xmath105 .
that is , the action of @xmath103 on @xmath79 is to increment @xmath88 and/or decrement @xmath106 , _ wherever possible_. these operators are illustrated in figure [ fig : kbenderknuth ] .
clearly each @xmath103 is an involution .
we call it the @xmath88th * @xmath0-bender - knuth involution * because in the case @xmath79 is standard , @xmath103 coincides with the classical involution introduced by e. bender d .
knuth @xcite .
\(a ) ; ( b ) ; ( c ) ; ( a ) edge[pil , color = dandelion ] node[left]@xmath107 ( b ) ; ( a ) edge[pil , color = skyblue ] node[right]@xmath108 ( c ) ; [ prop : kbk ] for @xmath78 , @xmath109 . another way to think of @xmath103 is as the _ @xmath0-infusion _ ( * ? ? ?
* section 3 ) of the labels @xmath88 through the labels @xmath106 .
that is , treat the labels @xmath88 as empty boxes and swap the short ribbons of empty boxes and @xmath110 s as in the definition of @xmath0-promotion ; then relabel each @xmath106 as @xmath88 and each empty box as @xmath106 . from this characterization ,
it is clear that @xmath111 amounts to deleting the @xmath77 s and swapping the empty boxes successively through each other label in order , decrementing each other label as the empty boxes swap through it , and finally labeling the empty boxes at outer corners by @xmath62 .
this is transparently the same as @xmath0-promotion , except that the decrementing of labels happens throughout the process instead of all at the end . in this subsection , we restrict consideration to increasing tableaux of rectangular shape . we extend a result of b. rhoades ( * ? ?
* lemma 3.3 ) from standard young tableaux to increasing tableaux .
our proof is a elaboration of his argument .
we will use this result in theorem [ thm : our_cfdf ] to improve on a theorem of p. cameron d .
fon - der - flaass ( * ? ? ?
* theorem 6(a ) ) . throughout this section ,
we write `` east '' , `` east '' and `` southeast '' to mean `` strictly east '' , `` weakly east '' and `` weakly south and strictly east '' respectively , etc .
let @xmath112 .
for @xmath113 , the symbol @xmath88 is a * descent * of @xmath79 if some instance of @xmath88 appears in a higher row than some instance of @xmath106 .
additionally , @xmath62 is a * descent * of @xmath79 if @xmath114 is a descent of @xmath115 .
[ lem : descents ] suppose @xmath88 is a descent of @xmath112 .
then @xmath116 is a descent of @xmath115 . throughout this proof ,
we use the original definition of @xmath0-promotion involving empty boxes , instead of the @xmath0-bender - knuth alternative .
* case 1 : ( @xmath117 ) : * @xmath79 has an instance of @xmath88 in row @xmath118 and an instance of @xmath106 in row @xmath119 with @xmath120 . in @xmath115 ,
there is an @xmath121 in row @xmath118 or @xmath122 and there is an @xmath88 in row @xmath119 or @xmath123 .
hence @xmath121 is a descent in @xmath115 if @xmath124 .
thus assume @xmath125 .
restrict attention to rows @xmath118 and @xmath126 of @xmath79 .
@xmath79 has a unique @xmath88 in row @xmath118 and a unique @xmath106 in row @xmath126 .
by increasingness , this @xmath106 is not east of this @xmath88 .
suppose the @xmath106 is west of the @xmath88 .
then @xmath79 contains the local configuration . since @xmath127 , the @xmath106 can not move north during this application of @xmath0-promotion .
hence @xmath115 has @xmath88 in row @xmath126 , and @xmath121 is a descent of @xmath115 .
thus , it remains to consider the case that @xmath88 and @xmath106 are in the same column of @xmath79 .
the @xmath106 can only move north if the @xmath88 moves .
if the @xmath88 moves north , we are done , so assume @xmath88 moves west
. then @xmath79 has the local configuration where @xmath128 .
but by increasingness , @xmath129 .
hence @xmath130 , so @xmath79 has the local configuration .
therefore , @xmath115 has the local configuration and thus @xmath121 is a descent of @xmath115 .
* case 2 : ( @xmath131 ) : * we must show that @xmath62 is a descent of @xmath115 , that is , @xmath114 is a descent of @xmath132 . for @xmath133 , let @xmath134 be the * flow path * of @xmath135 , that is the set of pairs of adjacent boxes @xmath136 of @xmath135 such that @xmath137 and @xmath138 are at some point part of the same short ribbon during the application of @xmath66 to @xmath135 . for @xmath137 a box of @xmath139 , we write @xmath140 for the box immediately north of @xmath137 , @xmath141 for the box immediately east of @xmath137 , etc
. define the * upper flow path * @xmath142 to be those @xmath143 such that @xmath144 is northmost in its columns among @xmath134 together with those @xmath145 such that @xmath145 is eastmost in its rows among @xmath134 .
similarly define the * lower flow path * @xmath146 to be those pairs in @xmath134 that are southmost or westmost .
figure [ fig : flowpath ] shows an example of these flow paths .
( 0,0 ) grid ( 4,5 ) ; ( a ) at ( 0.5,4.5 ) @xmath77 ; ( b ) at ( 1.5,4.5 ) @xmath80 ; ( c ) at ( 2.5,4.5 ) @xmath147 ; ( d ) at ( 3.5,4.5 ) @xmath148 ; ( e ) at ( 0.5,3.5 ) @xmath80 ; ( f ) at ( 1.5,3.5 ) @xmath149 ; ( g ) at ( 2.5,3.5 ) @xmath150 ; ( h ) at ( 3.5,3.5 ) @xmath151 ; ( i ) at ( 0.5,2.5 ) @xmath76 ; ( j ) at ( 1.5,2.5 ) @xmath152 ; ( k ) at ( 2.5,2.5 ) @xmath148 ; ( l ) at ( 3.5,2.5 ) @xmath30 ; ( m ) at ( 0.5,1.5 ) @xmath152 ; ( n ) at ( 1.5,1.5 ) @xmath153 ; ( o ) at ( 2.5,1.5 ) @xmath154 ; ( p ) at ( 3.5,1.5 ) @xmath155 ; ( q ) at ( 0.5,0.5 ) @xmath150 ; ( r ) at ( 1.5,0.5 ) @xmath38 ; ( s ) at ( 2.5,0.5 ) @xmath156 ; ( t ) at ( 3.5,0.5 ) @xmath157 ; ( a ) ( b ) ; ( b ) ( c ) ; ( c ) ( g ) ; ( g )
( k ) ; ( k )
( l ) ; ( l ) ( p ) ; ( p ) ( t ) ; ( a ) ( e ) ; ( e ) ( i ) ; ( i ) ( m ) ; ( m ) ( q ) ; ( q ) ( r ) ; ( r ) ( s ) ; ( s ) ( t ) ; ( i ) ( j ) ; ( j )
( k ) ; let @xmath158 be the box in the lower right corner of @xmath139 . by proposition [ prop : content_cycling ] , @xmath62 appears in @xmath115 .
hence by increasingness , @xmath115 has @xmath159 .
thus it suffices to show that @xmath160 .
the proof proceeds by comparing @xmath161 and @xmath162 .
let @xmath163 .
it is clear that @xmath164 contains exactly one box from each column of @xmath139 , except the eastmost column .
if @xmath165 , then there is some @xmath166 such that @xmath167 . choose @xmath137 to be maximally west among such boxes .
since @xmath137 is chosen maximally west , @xmath168 .
suppose @xmath169 .
then in @xmath115 , the entry of @xmath137 is strictly less than the entry of @xmath170 .
that is , if @xmath118 is the entry of @xmath137 and @xmath119 is the entry of @xmath170 then @xmath120 .
however , in @xmath79 we have @xmath171 and @xmath172 ; this contradicts the increasingness of @xmath79 .
thus @xmath137 is the northwestmost box of @xmath139 . since @xmath77 is a descent of @xmath79 and @xmath167
, @xmath79 has @xmath173 , @xmath174 and @xmath175 .
let @xmath176 .
we claim that if @xmath177 , then there is a pair @xmath178 with @xmath32 north of @xmath137 in the same column . to see this , first observe by local analysis that if @xmath179 and @xmath180 , then @xmath181 .
now recall that @xmath182 contains exactly one box from each column of @xmath139 , except the eastmost column .
moreover since @xmath79 has @xmath174 , no box of @xmath182 is in the northmost row . the claim follows .
thus @xmath183 and we are done .
* case 3 : ( @xmath184 ) : * by definition .
[ prop : descent_cycling ] the symbol @xmath88 is a descent of @xmath112 if and only if @xmath116 is a descent of @xmath115 .
suppose @xmath88 is a descent of @xmath79 .
by lemma [ lem : descents ] , @xmath116 is a descent of @xmath115 .
since @xmath185 is finite , there is some @xmath186 such that @xmath187 .
hence by @xmath186 applications of lemma [ lem : descents ] , @xmath106 is a descent of @xmath188 .
let @xmath112 . for @xmath113 , @xmath88 is * transpose descent * of @xmath79 if some instance of @xmath88 appears in a lower indexed column than some instance of @xmath106 .
additionally @xmath62 is a * transpose descent * of @xmath79 if @xmath114 is a descent of @xmath115 .
equivalently , @xmath189 is a transpose descent of @xmath79 if and only if @xmath189 is a descent of the transpose of @xmath79 .
[ prop : transpose_descent_cycling ] the symbol @xmath88 is a transpose descent of @xmath112 if and only if @xmath116 is a transpose descent of @xmath115 . since clearly
@xmath0-promotion commutes with transposing , the proposition is immediate from proposition [ prop : descent_cycling ] .
the following is an enriched version of corollary [ cor : weakest_cfdflike_statement ] for rectangular tableaux .
[ prop : content_descents ] let @xmath112 with @xmath62 prime .
suppose at least one of the following is true : * @xmath79 does not have full content , * some @xmath104 is not a descent in @xmath79 , or * some @xmath104 is not a transpose descent in @xmath79 .
then , the @xmath0-promotion orbit of @xmath79 has cardinality a multiple of @xmath62 .
if @xmath79 does not have full content , corollary [ cor : weakest_cfdflike_statement ] applies . otherwise , some @xmath104 is not a ( transpose ) descent in @xmath79 .
the proposition is then immediate by proposition [ prop : descent_cycling ] or [ prop : transpose_descent_cycling ] . finally , we prove the following lemma , which we will use in section [ subsec : conseq ] . [
lem : three_for_two ] let @xmath112 and suppose that @xmath113 is both a descent and a transpose descent in @xmath79
. then the number of @xmath88 s in @xmath79 plus the number of @xmath110 s in @xmath79 is at least @xmath76 .
since @xmath88 is a descent , both @xmath88 and @xmath106 must appear in @xmath79 .
hence if @xmath88 appears at least twice in @xmath79 , we are done .
thus assume @xmath88 appears exactly once in @xmath79 .
since @xmath88 is a descent , some @xmath106 appears south of this @xmath88 .
since @xmath88 is a transpose descent , some @xmath106 appears east of this @xmath88 .
we claim these instances of @xmath106 are distinct , completing the proof of the lemma .
otherwise , we have @xmath106 southeast of @xmath88 . consider the label @xmath190 of the box that is in the row of the @xmath88 and in the column of the @xmath106 . by the increasingness conditions on @xmath79 , @xmath191 , contradicting that @xmath190 is an integer . in section [ subsec : conseq ]
, we will use proposition [ prop : content_descents ] , lemma [ lem : three_for_two ] , and our main results , theorems [ thm:3dprorow ] and [ cor : rowinc ] , to give in theorem [ thm : our_cfdf ] a strengthening of a theorem of p. cameron d .
fon - der - flaass on plane partitions in a box .
in this section , we switch our focus from increasing tableaux to our other main objects of study : _ plane partitions_. a * plane partition * is a stack of unit cubes in the positive orthant , justified toward the origin in all three directions .
plane partitions inside a box with side lengths @xmath192 , @xmath193 , and @xmath14 , are counted by p. macmahon s box formula : @xmath194 where the product is over all @xmath195 , @xmath196 , @xmath197 @xcite .
plane partitions inside an @xmath44 box can be seen as _ order ideals _ in the product of three chains poset @xmath198 . thus , most of our discussion in this section centers on posets and order ideals , keeping in mind that all such general results can be applied to plane partitions .
we begin in section [ subsec : row ] by discussing the _ rowmotion _
action on order ideals and some results on the order of this action on products of two and three chains . in section [ subsec :
toggle ] , we discuss the _ toggle group _ , first defined by p. cameron d .
fon - der - flaass @xcite and further studied by n. williams and the third author @xcite . in section [ sec : resonance_pp ] , we use the main theorem of @xcite to prove resonance of plane partitions under rowmotion
. the toggle group will be the algebraic structure underlying sections [ subsec:3dprorow ] and [ subsec : provpi ] , in which we revisit this main result of @xcite by proving , in theorem [ thm:3dprorow ] , a generalization in the setting of _
projections_. let @xmath50 be a finite partially ordered set ( poset ) .
@xmath50 is a * chain * if all its elements are mutually comparable .
let @xmath199 denote the @xmath1-element chain .
the * product of @xmath119 chains * poset , @xmath200 , has as elements ordered integer @xmath119-tuples @xmath201 such that @xmath202 with partial order given by componentwise comparison .
a subset @xmath203 is an * order ideal * if it is closed downward , i.e. if @xmath204 and @xmath205 , then @xmath206 . denote the set of order ideals of @xmath50 as @xmath57 .
an order ideal in @xmath50 is uniquely determined by its set of maximal elements , or alternatively by the set of minimal elements of its complement in @xmath50 .
we study the orbit structure of * rowmotion * , @xmath207 , defined as the order ideal whose maximal elements are the minimal elements of @xmath208 .
the function @xmath45 has a long history of rediscovery and has appeared under many names .
a partial summary of previous work follows ; for a more complete discussion , see @xcite .
a. brouwer a . schrijver @xcite studied @xmath45 for @xmath209 , the product of two chains .
they discovered that this action has much smaller orbits than one naively expects : [ thm : square ] the order of @xmath45 on @xmath210 is @xmath211 .
p. cameron d .
fon - der - flaass @xcite studied the same question on plane partitions , that is , @xmath64 .
[ thm : height2 ] the order of @xmath45 on @xmath212 is @xmath213 . extrapolating from theorems [ thm : square ] and [ thm : height2 ]
, one might speculate that @xmath45 has order @xmath48 on @xmath64 . in general
, the order is unknown but often significantly greater than this naive guess .
however , p. cameron d .
fon - der - flaass established the following related fact .
[ thm : cfdf ] if @xmath214 is prime and @xmath215 , then the cardinality of every orbit of @xmath45 on @xmath64 is a multiple of @xmath48 .
we will revisit theorems [ thm : square ] and [ thm : height2 ] in remark [ remark : p_pp ] . in section [ sec : resonance_pp ] , we give a new proof of theorem [ thm : cfdf ] .
furthermore , as a consequence of our main equivariant bijection between plane partitions and increasing tableaux ( theorem [ cor : rowinc ] ) , we will show , in theorem [ thm : our_cfdf ] , that in theorem [ thm : cfdf ] the condition @xmath216 may be relaxed to @xmath217 .
this is evidence toward the conjecture of p. cameron d .
fon - der - flaass @xcite that this condition may be dropped entirely . the approach of p. cameron d .
fon - der - flaass was to reinterpret rowmotion as a _ toggle group action_. we describe the toggle group in the next subsection .
the * toggle group * was first studied by p. cameron d .
fon - der - flaass @xcite and subsequently by j. striker
williams @xcite .
it is the subgroup of the symmetric group on all order ideals @xmath218 generated by certain involutions , called * toggles*. for each element @xmath219 define its * toggle * @xmath220 as follows .
@xmath221 [ remark : commute ] observe that @xmath222 commute whenever neither @xmath223 nor @xmath12 covers the other .
the following theorem interprets rowmotion as a toggle group action .
[ thm : row_tog ] given any poset @xmath50 , @xmath45 is the toggle group element that toggles the elements of @xmath50 in the reverse order of any linear extension . if @xmath50 is ranked , this is the same as toggling the ranks ( rows ) from top to bottom . in 2012
@xcite , n. williams and the third author built on the work of p. cameron d .
fon - der - flaass , showing that rowmotion is conjugate to the toggle group action they called * promotion * , or @xmath224 , defined as toggling the elements of the poset from left to right ( given a suitable notion of left - to - right , for which they used the term _ rc - poset _ ) .
[ thm : stwiprorow ] for any rc - poset @xmath50 , there is an equivariant bijection between @xmath57 under @xmath224 and @xmath57 under @xmath45 .
we discuss this result in further detail in sections [ subsec:3dprorow ] and [ subsec : provpi ] and give a multidimensional generalization of it in theorem [ thm:3dprorow ] .
[ remark : p_pp ] for many posets , the orbit structure of promotion is easier to study than that of rowmotion .
thus theorem [ thm : stwiprorow ] yielded many results on the orbit structure of rowmotion by translating from the analogous result on promotion .
theorem [ thm : stwiprorow ] was applied in @xcite to give simple new proofs of theorem [ thm : square ] of a. brouwer a .
schrijver and theorem [ thm : height2 ] of p. cameron d .
fon - der - flaass ( discussed in section [ subsec : row ] ) , as well as easy proofs of the _ cyclic sieving phenomenon _ of v. reiner , d. stanton , and d. white @xcite in these cases and a few others . in the next subsection
, we use theorem [ thm : stwiprorow ] to prove resonance of rowmotion on plane partitions . in this subsection
, we prove our second main resonance result , theorem [ cor : resonance_of_pp_directly_row ] .
we also give a new proof of theorem [ thm : cfdf ] . in (
* section 7.2 ) , n. williams and the third author applied their theory to plane partitions , that is , the order ideals @xmath43 .
they characterized @xmath43 in terms of _ boundary path matrices_. we give a sketch of this characterization here ; for futher details , see @xcite .
given an order ideal in a special kind of planar poset ( in the language of @xcite , an rc - poset of height 1 , or in the language of the next section , a poset with a 2-dimensional lattice projection ) , its * boundary path * is a binary sequence that encodes the path that separates the order ideal from the rest of the poset . given a plane partition @xmath225 ,
its * boundary path matrix * is a @xmath226 matrix @xmath227 with entries in @xmath228 such that the @xmath88th row consists of the boundary path of layer @xmath88 preceded by @xmath229 zeros and succeeded by @xmath230 zeros .
the rows of a boundary path matrix each sum to @xmath192 and the entries obey the condition @xmath231 it was noted in ( * ? ? ?
* section 7.2 ) that @xmath224 traces from left to right through the columns of the boundary path matrix , swapping each pair of entries in adjacent columns and the same row that result in a matrix still satisfying the condition above . given @xmath232 with boundary path matrix @xmath227 ,
define @xmath233 to be the vector of length @xmath48 whose @xmath189th entry is @xmath234 .
[ prop : zeroscol_cycling ] let @xmath232 .
if @xmath235 , then @xmath236 is the cyclic shift @xmath237 . for @xmath98 ,
if column @xmath88 of the boundary path matrix is all zeros , then in the application of @xmath224 , all of these entries swap with the entries of column @xmath121 , since the condition on the partial row sums is not violated .
if @xmath131 , the column of all zeros swaps all the way through the matrix , from the first column to the last column .
thus , under @xmath224 , a column of all zeros cyclically shifts to the left .
the following instance of resonance follows directly from lemma [ prop : zeroscol_cycling ] .
[ cor : resonance_of_pp_directly ] @xmath238 exhibits resonance with frequency @xmath48 .
let @xmath239 be the conjugating toggle group element between rowmotion and promotion given in ( * ? ? ?
* theorem 5.4 ) . by the equivariance of @xmath224 and @xmath45 in @xcite
, we have the following statement of resonance on rowmotion , whose proof follows directly from proposition [ cor : resonance_of_pp_directly ] and ( * ? ? ?
* theorem 5.4 ) .
[ cor : resonance_of_pp_directly_row ] @xmath47 exhibits resonance with frequency @xmath48 .
this leads to the following corollary .
[ cor : weakest_cfdflike_statement_pp ] suppose @xmath48 is prime and @xmath240 .
suppose there is a zero in @xmath233 .
then the size of the promotion orbit of @xmath50 is a multiple of @xmath48 .
using corollary [ cor : weakest_cfdflike_statement_pp ] , we have a new proof of theorem [ thm : cfdf ] of p. cameron d .
fon - der - flaass .
if @xmath48 is prime and @xmath241 , then there are a total of @xmath242 ones in the boundary path matrix , but a total of @xmath243 columns in the matrix , so there must be a column of all zeros . thus , there is a zero in @xmath233 for any plane partition @xmath50 in an @xmath44 box , and the promotion orbit is a multiple of @xmath48 by corollary [ cor : weakest_cfdflike_statement_pp ] .
then by theorem [ thm : stwiprorow ] , the orbits of rowmotion are also multiples of @xmath48 .
p. cameron d .
fon - der - flaass s original proof of theorem [ thm : cfdf ] is similar , though more complicated , since it analyzes rowmotion directly rather than conjugating to promotion . in this and
the next subsections , we adapt the proof of the conjugacy of promotion and rowmotion from @xcite to give a generalization in the setting of _
@xmath1-dimensional lattice projections _ , which we introduce in definition [ def : lattice_proj ] .
( this new perspective includes the original theorem as the case @xmath244 . )
we prove , in theorem [ thm:3dprorow ] , the equivariance of the @xmath245 toggle group actions given in definition [ def:3dpro ] .
we say that a poset @xmath50 is * ranked * if it admits a rank function @xmath246 satsifying @xmath247 when @xmath248 covers @xmath249 .
[ def : lattice_proj ] we say that an * ( @xmath1-dimensional ) lattice projection * of a ranked poset @xmath50 is an order and rank preserving map @xmath250 , where the rank function on @xmath251 is the sum of the coordinates and @xmath252 in @xmath251 if and only if the componentwise difference @xmath253 is in @xmath254 . in light of remark [ remark : commute ] , the key feature of @xmath51 is that it preserves cover relations .
that is , if @xmath248 covers @xmath249 in @xmath50 , then @xmath255 covers @xmath256 in @xmath251 .
however , since @xmath251 is ranked , @xmath51 being cover - relation preserving would make @xmath257 a rank function for @xmath50 . and if @xmath50 is ranked , then a map @xmath250 being cover - relation preserving is equivalent to it being order and rank preserving ( up to a shift of the rank functions ) . in @xcite , the definition of an rc - poset was a poset that had a 2-dimensional lattice projection ( albeit to a slightly different lattice )
however , e. sawin noted that every ranked poset @xmath50 with rank function @xmath258 has such an embedding given by @xmath259 for @xmath260 @xcite .
similarly , any poset @xmath50 with a lattice projection @xmath51 has a rank function given by the sum of the coordinates in @xmath256 for @xmath260 .
additionally , a ranked poset may have multiple distinct projections .
for example , in figure [ fig : cube ] , we have the boolean lattice on three elements , which we can think of a product of three chains of length 2 . in figure
[ fig:3dlattproj ] , we have the standard three - dimensional lattice projection of this poset we get by thinking of it as a product of three chains . in figure
[ fig : rcprojection ] , we show two different two - dimensional lattice projections of this poset . in the projection on the right , we assign every element of the same rank to the same point , but instead of doing so along the @xmath249-axis as in the previous paragraph , we do this diagonally in a zig - zag pattern .
therefore , instead of considering rc - posets , we consider any ranked poset , but with respect to a given lattice projection . at ( 0,0 ) a ; at ( -1,1 ) b ; at ( 0,1 ) c ; at ( 1,1 ) d ; at ( -1,2 ) e ; at ( 0,2 ) f ; at ( 1,2 ) g ; at ( 0,3 ) h ; ( 0,0 ) ( 1,1 )
( 0,3 ) ( -1,2 )
( -1,1 ) ( 0,0 ) ; ( -1,2 ) ( 0,1 ) ( 1,2 ) ; ( -1,1 ) ( 0,2 ) ( 1,1 ) ; ( 0,0 ) ( 0,1 ) ; ( 0,3 ) ( 0,2 ) ; at ( 0,0,0 ) ; at ( 0,0,1 ) ; at ( 0,1,0 ) ; at ( 1,0,0 ) ; at ( 0,1,1 ) ; at ( 1,0,1 ) ; at ( 1,1,0 ) ; at ( 1,1,1 ) ; ( 0,0,0 ) ( 0,0,1 ) ( 0,1,1 ) ( 1,1,1 ) ; ( 0,0,0 ) ( 1,0,0 )
( 1,1,0 )
( 1,1,1 ) ; ( 0,0,0 )
( 0,1,0 )
( 1,1,0 )
( 1,1,1 ) ; ( 1,0,0 )
( 1,0,1 )
( 1,0,0 ) ; ( 0,0,1 )
( 1,0,1 )
( 1,1,1 ) ; ( 0,1,0 )
( 0,1,1 ) ; ( 0,0,1 )
( 0,0,1.5 ) node[anchor = north ] @xmath249 ; ( 0,1,0 )
( 0,1.5,0 ) node[anchor = west ] @xmath248 ; ( 1,0,0 ) ( 1.5,0,0 ) node [ anchor = west ] @xmath190 ; ( 0,0 ) grid ( 1,2 ) ; at ( 0,0 ) ; at ( 0,1 ) ; at ( 1,0 ) ; at ( 1,1 ) ; at ( 0,2 ) ; at ( 1,2 ) ; ( 1,0 )
( 1.5,0 ) [ dotted,- > ] ; ( 0,2 ) ( 0,2.5 ) [ dotted,- > ] ; ( 0,0 ) grid ( 1,2 ) ; at ( 0,0 ) ; at ( 0,1 ) ; at ( 1,0 ) ; at ( 1,1 ) ; at ( 0,2 ) ; at ( 1,2 ) ; ( 1,0 ) ( 1.5,0 ) [ dotted,- > ] ; ( 0,2 )
( 0,2.5 ) [ dotted,- > ] ; we now define a toggling order on our poset with respect an @xmath1-dimensional lattice projection , and with respect to a distinguished direction .
[ def:3dpro ] let @xmath50 be a poset with an @xmath1-dimensional lattice projection @xmath51 , and let @xmath52 , where @xmath261 .
let @xmath262 be the product of toggles @xmath263 for all elements @xmath249 of @xmath50 that lie on the affine hyperplane @xmath264 .
if there is no such @xmath249 , then this is the empty product , considered to be the identity .
then define * promotion with respect to @xmath51 and @xmath265 * as the toggle product @xmath266 .
see figure [ fig : hyperplanes ] for an example .
note that @xmath267 , so we will generally only consider distinguished vectors with @xmath268 , as all promotion operators are either of this form , or the inverse of something of this form .
[ lem : hyperplane ] two elements of the poset that lie on the same affine hyperplane @xmath264 can not be part of a covering relation , so by remark [ remark : commute ] , the operator @xmath262 is well - defined and @xmath269 .
assume that @xmath248 covers @xmath249 , and they both lie on the same affine hyperplane ( @xmath270 ) .
then @xmath271 .
but since @xmath248 covers @xmath249 , @xmath272 for some @xmath88 . and since @xmath265 has all coordinates @xmath273 , then @xmath274 , a contradiction . in 2,2 in ( 0,0
) ( @xmath275 )
node[anchor = north ] @xmath248 ; ( 0,0 )
( @xmath276 ) node[anchor = north ] @xmath249 ; ( 0,0 )
( @xmath277 ) node[anchor = west ] @xmath190 ; ( a1 ) at ( @xmath278 ) ; at ( a1 ) ; at ( 0,-1 ) @xmath279 ; in 2,2 in ( 0,0 ) ( @xmath275 )
node[anchor = north ] @xmath248 ; ( 0,0 )
( @xmath276 ) node[anchor = north ] @xmath249 ; ( 0,0 )
( @xmath277 ) node[anchor = west ] @xmath190 ; ( a2 ) at ( @xmath280 ) ; ( b2 ) at ( @xmath281 ) ; ( c2 ) at ( @xmath282 ) ; at ( a2 ) ; at ( b2 ) ; at ( c2 ) ; ( a2 )
( b2 ) ( c2 ) cycle ; at ( 0,-1 ) @xmath283 ; in 2,2 in ( 0,0 ) ( @xmath275 )
node[anchor = north ] @xmath248 ; ( 0,0 )
( @xmath276 ) node[anchor = north ] @xmath249 ; ( 0,0 )
( @xmath277 ) node[anchor = west ] @xmath190 ; ( a3 ) at ( @xmath284 ) ; ( b3 ) at ( @xmath285 ) ; ( c3 ) at ( @xmath286 ) ; ( d3 ) at ( @xmath287 ) ; ( e3 ) at ( @xmath288 ) ; at ( a3 ) ; at ( b3 ) ; at ( c3 ) ; at ( d3 ) ; at ( e3 ) ; ( a3 ) ( c3 ) ( b3 ) ( d3 ) ( e3 ) cycle ; at ( 0,-1 ) @xmath289 ; in 2,2 in ( 0,0 ) ( @xmath275 ) node[anchor = north ] @xmath248 ; ( 0,0 ) ( @xmath276 )
node[anchor = north ] @xmath249 ; ( 0,0 ) ( @xmath277 )
node[anchor = west ] @xmath190 ; ( a4 ) at ( @xmath290 ) ; ( b4 ) at ( @xmath291 ) ; ( c4 ) at ( @xmath292 ) ; ( d4 ) at ( @xmath293 ) ; ( e4 ) at ( @xmath294 ) ; at ( a4 ) ; at ( b4 ) ; at ( c4 ) ; at ( d4 ) ; at ( e4 ) ; ( a4 ) ( c4 ) ( b4 ) ( d4 ) ( e4 ) cycle ; at ( 0,-1 ) @xmath295 ; in 2,2 in ( 0,0 )
( @xmath275 ) node[anchor = north ] @xmath248 ; ( 0,0 )
( @xmath276 ) node[anchor = north ] @xmath249 ; ( 0,0 )
( @xmath277 ) node[anchor = west ] @xmath190 ; ( a5 ) at ( @xmath296 ) ; ( b5 ) at ( @xmath297 ) ; ( c5 ) at ( @xmath298 ) ; at ( a5 ) ; at ( b5 ) ; at ( c5 ) ; ( a5 ) ( c5 ) ( b5 ) cycle ; at ( 0,-1 ) @xmath299 ; in 2,2 in ( 0,0 ) ( @xmath275 )
node[anchor = north ] @xmath248 ; ( 0,0 )
( @xmath276 ) node[anchor = north ] @xmath249 ; ( 0,0 )
( @xmath277 ) node[anchor = west ] @xmath190 ; ( a6 ) at ( @xmath300 ) ; at ( a6 ) ; at ( 0,-1 ) @xmath301 ; for ease of notation , we may suppress explicitly listing the lattice projection map @xmath51 or the direction @xmath265 when referring to the generalized promotion operator , if it is clear from context .
note that for a finite poset @xmath50 , @xmath262 will be the identity operator for all but finitely many @xmath88 . to compare with the notion of promotion and rowmotion given in @xcite , for a given 2-dimensional lattice projection @xmath51 of a finite poset @xmath50
, rowmotion corresponds to @xmath302 , and promotion corresponds to @xmath303 . for
any finite ranked poset @xmath50 and lattice projection @xmath51 , @xmath304 .
@xmath305 sweeps through @xmath50 from top to bottom ( in the reverse order of a linear extension ) , so by theorem [ thm : row_tog ] , this is rowmotion .
we give some further definitions and lemmas , in order to state and prove theorem [ thm:3dprorow ] in full generality .
let @xmath50 be a poset , and let @xmath51 , @xmath265 , and @xmath262 be as in definition [ def:3dpro ] .
define the * support * of @xmath306 , denoted @xmath307 , to be the smallest interval @xmath308\subseteq\mathbb{z}$ ] such that @xmath262 is the identity operator for all @xmath309 $ ] . if @xmath306 has finite support , that is , @xmath310\subset\mathbb{z}$ ] , let @xmath311\rightarrow[a , b]$ ] be a bijection .
then define * promotion with respect to @xmath50 , @xmath51 , @xmath265 , and @xmath312 * as the following product of hyperplane - toggles : @xmath313 .
we will use the following toggle group element in the proof of theorem [ thm:3dprorow ] .
[ def : gyr ] for a poset @xmath50 , define the * parity * of @xmath314 as even ( resp .
odd ) if the parity of @xmath315 is even ( resp .
define * gyration * @xmath316 as the toggle group element which first toggles all @xmath314 with even parity , then all @xmath317 with odd parity .
[ remark : asmposetgyr ] there is a connection between the toggle group element @xmath316 defined above and the gyration action on alternating sign matrices discussed in section [ sec : resonance ] .
namely , there is a poset @xmath318 whose order ideals are in bijection with alternating sign matrices , and for which the gyration action of definition [ def : fplgyr ] is the action of @xmath316 . for details ,
see ( * ? ? ?
* section 8) and @xcite . given a lattice projection @xmath51 ,
the rank of @xmath317 is the same as the rank of @xmath319 , which is @xmath320 .
since all the coordinates in @xmath265 are @xmath273 , the parity of @xmath321 will be the same as the parity of @xmath322 .
thus , all elements lying on the same affine hyperplane with respect to @xmath265 will have the same parity .
[ lem : gyr ] if @xmath306 has finite support @xmath308 $ ] , then for any bijection @xmath311\rightarrow[a , b]$ ] such that @xmath323 is odd if @xmath324 and even if @xmath325 , we have @xmath326 .
we are now nearly ready to state and prove the main theorem of this section .
we will need the following lemma , which appears as ( * ? ? ?
* lemma 5.1 ) .
[ lem : conjugatelem ] let @xmath15 be a group whose generators @xmath327 satisfy @xmath328 and @xmath329 if @xmath330 .
then for any @xmath331 , @xmath332 and @xmath333are conjugate .
the main theorem of this section is below , whose proof follows the proof of ( * ? ? ?
* theorem 5.2 ) .
[ thm:3dprorow ] let @xmath50 be a finite poset with an @xmath1-dimensional lattice projection @xmath51 .
let @xmath52 and @xmath53 , where @xmath54 .
finally suppose that @xmath55 and @xmath56 are bijections .
then there is an equivariant bijection between @xmath57 under @xmath58 and @xmath57 under @xmath59 .
suppose @xmath50 is a finite poset with an @xmath1-dimensional lattice projection @xmath51 .
let @xmath52 , where @xmath334 .
we claim the toggles @xmath262 for @xmath335 satisfy the conditions of lemma [ lem : conjugatelem ] . by lemma [ lem : hyperplane ] , @xmath269 . also , if @xmath264 and @xmath336 , then @xmath337
. so if @xmath338 , as all the coefficients in @xmath265 are @xmath273 , then @xmath339 can not be @xmath340 for any @xmath88 , and @xmath248 and @xmath249 can not be part of a covering relation .
thus , toggles on non - adjacent hyperplanes commute , and we have @xmath341 when @xmath338 .
so by lemma [ lem : conjugatelem ] , for any bijections @xmath342 , there is an equivariant bijection between @xmath57 under @xmath58 and @xmath57 under @xmath343 ( since such bijections can be considered as permutations in @xmath344 if @xmath310 $ ] ) .
consider @xmath316 of definition [ def : gyr ] . by lemma [ lem : gyr ] ,
for every @xmath265 there exists a @xmath345 such that @xmath316 can be realized as @xmath346 .
therefore , there is an equivariant bijection between @xmath57 under @xmath58 and under @xmath347 , from which the theorem follows .
after we see a bijection between increasing tableaux and plane partitions given in the next section , we will use theorem [ thm:3dprorow ] to give an improvement on theorem [ thm : cfdf ] of p. cameron d .
fon - der - flaass ( discussed in section [ subsec : row ] ) .
in this section , we introduce bijections between increasing tableaux and plane partitions .
the existence of these bijections should not be at all surprising
. however , these maps have amazing properties that will be key to many of our results . these maps are also fundamental to @xcite , where they are used to give the first bijective proofs of various results on plane partitions , including r. proctor s main result from @xcite .
we define a map @xmath348 as follows .
let @xmath232 .
thinking of @xmath50 in the standard way as a pile of small cubes in an @xmath349 box , project onto the @xmath139 face .
record in position @xmath350 the number of boxes of @xmath50 with coordinate @xmath351 for some @xmath352 .
the result is a standard plane partition representation of @xmath50 , as a filling of the young diagram @xmath139 with nonnegative integers such that rows weakly decrease from left to right and columns weakly decrease from top to bottom .
rotate this plane partition @xmath353 , so that rows and columns become weakly _ increasing_. now thinking of @xmath139 as a graded poset with the upper left corner box the unique element of rank @xmath354 , add to each label its rank plus 1 .
that is , increase each label by one more than its distance from the upper left corner box .
( this is just the standard way of converting a _ weakly _ increasing sequence into a _
strictly _ increasing one . )
the result is the increasing tableau @xmath355 .
for an example of this transformation , see figure [ fig : incretization ] .
\(a ) ; ( z ) @xmath356 ; ( b ) ; ( c ) ; ( d ) ; ; ( a ) edge[pil ] node[above]project to node[below]bottom face ( b ) ; ( b ) edge[pil ] node[right]rotate @xmath353 ( c ) ; ( c ) edge[pil ] node[above]add @xmath357rank ( d ) ; [ thm : mainbij ] @xmath348 gives a bijection between plane partitions inside an @xmath44 box and increasing tableaux of shape @xmath358 and entries at most @xmath48 .
the map is defined as the composition of a projection , a rotation , and entrywise addition , all of which are clearly invertible .
similarly , define bijections @xmath359 and @xmath360 projecting onto the @xmath361 and @xmath362 faces , respectively ( cf . figure [ fig:3projections ] ) .
\(a ) ; in 4,4,4,3,4,3,3,2,3,2,2,1,3,1 in ( 0,0 ) ( 0,-1 ) ( 210:1 ) (150:1)(0,0 ) ; ; ; ( 0,0 ) ( 30:1 ) ( 0,1 ) (150:1)(0,0 ) ; ; ; ( 0,0 ) ( 30:1 ) ( 0,1 ) (150:1)(0,0 ) ; ; ; ( 0,0 ) ( 30:1 ) ( -30:1 ) (0,-1)(0,0 ) ; ; ; ( 0,0 ) ( 30:1 ) ( -30:1 ) (0,-1)(0,0 ) ; ; ; \(l ) ; \(t ) ; ( r ) ; \(ll ) ; ( tt ) ; ( rr ) ; \(a ) ( t ) ; ( a )
( l ) ; ( a ) ( r ) ; ( t ) ( tt ) ; ( l ) ( ll ) ; ( r ) ( rr ) ; given the simplicity of the bijection of theorem [ thm : mainbij ] , one might wonder why it was previously overlooked .
the set of increasing tableaux in bijection with plane partitions includes those with gaps in the content .
however much previous research on increasing tableaux was motivated by @xmath0-theoretic geometry , and in this context there is little reason to consider increasing tableaux without full content . moreover , by restricting to tableaux of full content , one obtains some attractive enumerations @xcite ; for instance , the number of increasing tableaux with shape @xmath363 and full content is the @xmath1th _ small schrder number _
* theorem 1.1 ) .
it was the equivariance of the actions of @xmath66 and @xmath45 , discussed in the next section , which led us to observe the bijection of theorem [ thm : mainbij ] .
our first main result was theorem [ thm:3dprorow ] , that given a poset @xmath50 with lattice projection @xmath51 , there is an equivariant bijection between the order ideals @xmath57 under @xmath58 and @xmath59 , where @xmath364 are any permutations of the hyperplane toggles associated to the @xmath365-vectors @xmath366 . in this section ,
we use theorem [ thm:3dprorow ] in our proof of our second main result , theorem [ cor : rowinc ] , that @xmath66 and @xmath45 are in equivariant bijection .
[ thm : main ] @xmath367 intertwines @xmath368 and @xmath66 .
that is , the following diagram commutes : \(a ) @xmath64 ; ( b ) @xmath65 ; ( c ) @xmath64 ; ( d ) @xmath65 ; ( a ) edge[pil ] node[below]@xmath367 ( b ) ; ( a ) edge[pil ] node[right]@xmath368 ( c ) ; ( b ) edge[pil ] node[right]@xmath66 ( d ) ; ( c ) edge[pil ] node[below]@xmath367 ( d ) ;
let @xmath232 and let @xmath369 .
note that the poset @xmath370 has a @xmath76-dimensional lattice projection , in the sense of definition [ def : lattice_proj ] , given by the identity map . by proposition [ prop : kbk ] , @xmath371 .
similarly , @xmath372 .
thus , it suffices to show that @xmath373 by definition [ def:3dpro ] , @xmath374 is the product of the the toggles @xmath263 for all @xmath375 lying on the affine hyperplane determined by @xmath376 .
consider @xmath377 on this hyperplane .
then @xmath378 .
we have @xmath379 if and only if the @xmath380 entry of @xmath79 is at least @xmath381 .
since @xmath382 , we can rewrite this condition as the @xmath380 entry of @xmath79 being at least @xmath383 .
hence @xmath384 if and only if the @xmath380 entry of @xmath79 is at least @xmath385 .
( case 1 : @xmath384 ) : if @xmath386 , then @xmath249 is unaffected by the toggle and the @xmath380 entry of @xmath79 is at least @xmath387 and so unaffected by @xmath388 .
otherwise @xmath389 and the @xmath380 entry of @xmath79 equals @xmath390 .
@xmath388 will turn this @xmath390 into @xmath391 exactly when neither the @xmath392 nor the @xmath393 entry of @xmath79 equals @xmath391 . by increasingness of @xmath79 , neither entry is greater than @xmath391 .
the @xmath392 entry of @xmath79 is at least @xmath391 exactly when @xmath394 .
similarly the @xmath393 entry of @xmath79 is at least @xmath391 exactly when @xmath395 .
hence @xmath388 will turn this @xmath390 into @xmath391 exactly when neither @xmath396 nor @xmath397 is in @xmath50 .
but this is exactly when the hyperplane toggle removes @xmath249 from @xmath50 . since @xmath50 is an order ideal , @xmath398 ,
so if @xmath374 removes @xmath249 from @xmath50 , then the @xmath380 entry of @xmath399 equals @xmath391 , as desired .
( case 2 : @xmath400 ) : the @xmath380 entry of @xmath79 is at most @xmath391 . if it is less than @xmath391 , then @xmath401 .
hence @xmath249 is unaffected by the hyperplane toggle and the @xmath380 entry of @xmath79 is unaffected by @xmath388 .
otherwise , the @xmath380 entry of @xmath79 equals @xmath391 and @xmath398 .
@xmath388 will turn this @xmath391 into @xmath390 exactly when neither the @xmath402 nor the @xmath403 entry of @xmath79 equals @xmath390 .
this happens exactly when both @xmath404 and @xmath405 .
thus @xmath374 toggles @xmath249 into @xmath50 exactly when @xmath388 turns this @xmath391 into @xmath390 .
[ r : symm ] by symmetry of @xmath64 , we obtain analogous results for @xmath406 and @xmath407 . as a consequence of the above lemma and theorem [ thm:3dprorow ]
, we obtain the following .
[ cor : rowinc ] @xmath64 under @xmath45 is in equivariant bijection with @xmath65 under @xmath66 . in this subsection
, we give a number of consequences of theorem [ cor : rowinc ] .
we first give another statement of resonance on plane partitions in corollary [ cor : row_resonance ] ( cf . theorem [ cor : resonance_of_pp_directly_row ] ) .
in corollary [ cor : trifold_sym ] , we give @xmath66-equivariant bijections between various sets of increasing tableaux using the tri - fold symmetry of @xmath64 .
we exploit this symmetry to prove corollaries [ cor : kproab ] and [ cor : ab1kpro ] .
we make a conjecture about the order of @xmath45 on @xmath408 .
finally , in theorem [ thm : our_cfdf ] , we improve the bound of theorem [ thm : cfdf ] of p. cameron d .
fon - der - flaass .
we obtain the following statement of resonance of rowmotion on plane partitions as a consequence of theorems [ cor : rowinc ] and [ cor : resonance_of_increasing_tableaux ] .
[ cor : row_resonance ] @xmath409 exhibits resonance with frequency @xmath48 .
we furthermore obtain the following corollary via the tri - fold symmetry of @xmath64 .
[ cor : trifold_sym ] there are @xmath66-equivariant bijections between the sets @xmath410 , @xmath411 , and @xmath412 . by lemma [ thm : main ] and
remark [ r : symm ] , @xmath413 is a @xmath66-equivariant bijection between @xmath410 and @xmath411 .
similarly , @xmath414 is an equivariant bijection between @xmath410 and @xmath412 .
theorem [ cor : rowinc ] and corollary [ cor : trifold_sym ] allow us to obtain a number of results for small values of @xmath14 .
we obtain new proofs of known results theorems [ thm : square ] and [ thm : height2 ] , while corollaries [ cor : kproab ] and [ cor : ab1kpro ] are new .
we use the following trivial fact about the order of @xmath66 on increasing tableaux of one row .
[ fact:1row_pro ] the order of @xmath66 on @xmath415 is @xmath62 .
the following is a new proof of theorem [ thm : square ] of a. brouwer a .
schrijver @xcite , which we restate for convenience .
the order of @xmath45 on @xmath210 is @xmath68 .
the order of @xmath45 on @xmath210 is the same as the order of @xmath45 on @xmath416 . by corollary
[ cor : rowinc ] , the order of @xmath45 on @xmath416 equals the order of @xmath66 on @xmath417 .
by fact [ fact:1row_pro ] , the order of @xmath66 on @xmath417 is @xmath68 .
the following result is new .
[ cor : kproab ] the order of @xmath66 on @xmath418 is @xmath211 . by the tri - fold symmetry of corollary [ cor : trifold_sym
] , there is a @xmath66-equivariant bijection between the sets @xmath418 and @xmath419 .
the result is then immediate by fact [ fact:1row_pro ] .
we can also use theorem [ cor : rowinc ] and corollary [ cor : trifold_sym ] to show that the theorem [ thm : height2 ] of p. cameron d .
fon - der - flaass @xcite is equivalent to a theorem of the second author on increasing tableaux , thus providing a new proof of theorem [ thm : height2 ] .
alternatively , one may use theorem [ thm : height2 ] along with theorem [ cor : rowinc ] and corollary [ cor : trifold_sym ] to give a new proof of the second author s result . the order of @xmath45 on @xmath212
is @xmath70 . by theorem [ cor : rowinc ] and corollary [ cor : trifold_sym ] ,
the order of @xmath45 on @xmath212 equals the order of @xmath66 on @xmath420 . by (
* theorem 1.3 ) , the latter is @xmath70 . ( the cited paper only considers those increasing tableaux of complete content ; however its proof extends easily to the case of general content . )
the following result is new .
[ cor : ab1kpro ] the order of @xmath66 on @xmath421 is @xmath213 . by corollary [ cor : trifold_sym ] , there is a @xmath66-equivariant bijection between the sets @xmath421 and @xmath420 .
the result is then immediate from ( * ? ? ?
* theorem 1.3 ) . recall that for @xmath422 , the order of @xmath45 on @xmath64 is generally greater than @xmath48
nonetheless , we make the following conjecture .
[ conj : three_row ] the order of @xmath45 on @xmath408 is @xmath423 . as with the above corollaries , the results of
this paper show that conjecture [ conj : three_row ] is equivalent to the order of @xmath0-promotion being @xmath423 on either @xmath424 or @xmath425 .
we have verified conjecture [ conj : three_row ] for @xmath426 and @xmath193 arbitrary .
finally , we improve the bound in theorem [ thm : cfdf ] of p. cameron d .
fon - der - flaass @xcite by more than a factor of @xmath427 .
this is evidence toward the conjecture of p. cameron d .
fon - der - flaass @xcite that this condition may be dropped entirely .
[ thm : our_cfdf ] if @xmath214 is prime and @xmath217 , then the cardinality of every orbit of @xmath45 on @xmath64 is a multiple of @xmath48 .
let @xmath428 .
the case @xmath429 is trivial , so assume @xmath62 is odd .
consider @xmath232 and let @xmath430 .
if @xmath79 does not have full content , then by corollary [ cor : weakest_cfdflike_statement ] , the @xmath0-promotion orbit of @xmath79 has cardinality a multiple of @xmath62 .
hence by theorem [ cor : rowinc ] , the rowmotion orbit of @xmath50 has cardinality a multiple of @xmath62 , as claimed .
thus , we may assume @xmath79 has full content .
similarly , by proposition [ prop : content_descents ] , we may assume that every @xmath104 is both a descent and a transpose descent in @xmath79 .
hence by lemma [ lem : three_for_two ] , for @xmath431 , the number of @xmath432 s in @xmath79 plus the number of @xmath433 s in @xmath79 is at least @xmath76 . by the increasingness conditions on @xmath79
, there is exactly @xmath77 instance of @xmath62 in @xmath79 .
thus the total number of labels in @xmath79 is at least @xmath434 . since @xmath112 ,
this forces @xmath435
. thus @xmath436 , contradicting the assumed bound on @xmath14 .
we conclude by reformulating some observations from @xcite in terms of resonance ; for further details , see ( * ? ? ?
* sections 8.3 and 8.4 ) .
recall from remark [ remark : asmposetgyr ] that there is a poset @xmath318 whose order ideals are in bijection with @xmath20 alternating sign matrices such that gyration ( of definition [ def : fplgyr ] ) is equivalent to the action of the toggle group element @xmath316 of definition [ def : gyr ] .
another element , @xmath437 , of the toggle group on @xmath318 was introduced in ( * ? ? ?
* definition 8.14 ) .
it is shown in ( * ? ? ?
* theorem 8.15 ) that the orbit of the empty order ideal in @xmath438 under @xmath437 has cardinality @xmath439 .
further data contained in ( * ? ? ?
* figure 22 ) leads us to propose the following . construct a natural map @xmath12 such that @xmath440 exhibits resonance with frequency @xmath439 .
similarly , there is a poset @xmath441 whose order ideals are in bijection with _ totally symmetric self - complementary plane partitions _ inside a @xmath442 box ( denote this set as @xmath443 ) .
it is shown in ( * ? ? ?
* theorem 8.19 ) that the cardinality of the rowmotion - orbit of the empty order ideal in @xmath444 is @xmath439 .
further data contained in ( * ? ? ?
* figure 22 ) suggests the following .
construct a natural map @xmath12 such that @xmath445 exhibits resonance with frequency @xmath439 .
we suspect that a solution to the above problems would be a major step towards exhibiting an explicit bijection between @xmath19 and @xmath443 , which are known ( non - bijectively ) to be equinumerous .
this work began at the `` dynamical algebraic combinatorics '' workshop at the american institute of mathematics ( aim ) in march 2015 .
the authors are grateful for many inspiring conversations with the participants .
the authors would like to thank jim propp , tom roby , and nathan williams for their organizational efforts , as well as aim for funding this workshop and the aim staff for their hospitality .
kd was supported by rtg grant nsf / dms-1148634 . op was supported by an illinois distinguished fellowship and an nsf graduate research fellowship .
js was supported by a national security agency grant ( h98230 - 15 - 1 - 0041 ) , the north dakota epscor national science foundation grant ( iia-1355466 ) , and the ndsu advance forward program sponsored by national science foundation grant ( hrd-0811239 ) .
9999999999 d. armstrong , c. stump and h. thomas , _ a uniform bijection between nonnesting and noncrossing partitions _ , trans .
* 365 * ( 2013 ) , 41214151 .
e. bender and d. knuth , _ enumeration of plane partitions _
, j. combin .
theory ser .
a * 13 * ( 1972 ) , 4054 .
j. bloom , o. pechenik and d. saracino , _ proofs and generalizations of a homomesy conjecture of propp and roby _ , discrete math .
* 339 * ( 2016 ) , 194206 .
bressoud , _ proofs and confirmations : the story of the alternating sign matrix conjecture _ , maa spectrum , mathematical association of america , washington , dc , 1999 .
a. brouwer and a. schrijver , _ on the period of an operator , defined on antichains _ , math centrum report zw * 24/74 * ( 1974 ) .
a. buch and v. ravikumar , _ pieri rules for the @xmath0-theory of cominuscule grassmannians _ , j. reine angew .
( crelle s j. ) * 668 * ( 2012 ) , 109132 .
a. buch and m. samuel , _ @xmath0-theory of minuscule varieties _ , j. reine angew .
( crelle s j. ) , to appear , 2014 .
cameron and d. fon - der - flaass , _ orbits of antichains revisited _ , eur .
j. combin .
* 16 * ( 1995 ) , 545554 .
e. clifford , h. thomas and a. yong , _
@xmath0-theoretic schubert calculus for @xmath446 and jeu de taquin for shifted increasing tableaux _
, j. reine angew . math .
( crelle s j. ) * 690 * ( 2014 ) , 5163 .
d. einstein and j. propp , _ piecewise - linear and birational toggling _ , discrete math .
proc . * at * ( 2014 ) , 513524 .
d. fon - der - flaass , _ orbits of antichains in ranked posets _ , eur
. j. combin .
* 14 * ( 1993 ) , 1722 .
d. grinberg and t. roby , _ iterative properties of birational rowmotion i : gen- eralities and skeletal posets _ , to appear , electron .
j. combin . , 2015 .
d. grinberg and t. roby , _ iterative properties of birational rowmotion ii : rectangles and triangles _
, electron .
j. combin .
* 22 * ( 2015 ) , 149 .
z. hamaker , r. patrias , o. pechenik and n. williams , _ coincidental types and their minuscule doppelgngers _ , in preparation .
p. hoffman and j. humphreys , _ projective representations of the symmetric groups : @xmath158-functions and shifted tableaux _ ,
oxford mathematical monographs , oxford university press , oxford , 1992 .
g. kuperberg .
_ another proof of the alternating sign matrix conjecture _ , int
. math .
res . not .
* 3 * ( 1996 ) , 139150 .
macmahon , _ combinatory analysis _ , 2 vols , cambridge university press , 1915 - 16 . w. mills , d. robbins , and h. rumsey , jr . , _ alternating sign matrices and descending plane partitions _ , j. combin .
theory ser .
a * 34 * ( 1983 ) , no . 3 , 340359 .
d. panyushev , _ on orbits of antichains of positive roots _ , eur .
j. combin .
* 30 * ( 2009 ) , 586594 .
o. pechenik , _ cyclic sieving of increasing tableaux and small schrder paths _ , j. combin .
theory ser .
a * 125 * ( 2014 ) , 357378 .
o. pechenik and a. yong , _ equivariant @xmath0-theory of grassmannians _ , preprint , 2015 .
t. pressey , a. stokke and t. visentin , _ increasing tableaux and generalized narayana numbers _ , to appear , ann .
, arxiv:1409.2841 .
r. proctor , _ shifted plane partitions of trapezoidal shape _ , proc .
* 89 * ( 1983 ) , 553559 .
j. propp and t. roby , _ homomesy in products of two chains _ , electron . j. combin .
* 22 * ( 2015 ) , 129 .
j. propp , _ the many faces of alternating - sign matrices _ , discrete math .
. proc . *
aa * ( 2001 ) , 4358 .
v. reiner , d. stanton and d. white , _ the cyclic sieving phenomenon _ , j. combin .
theory ser .
a * 108 * ( 2004 ) , 1750 .
b. rhoades , _ cyclic sieving , promotion , and representation theory _
, j. combin .
theory ser .
a * 117 * ( 2010 ) , 3876 .
b. rhoades , _ a skein action of the symmetric group on noncrossing partitions _ , preprint , arxiv:1501.04680 .
d. robbins and h. rumsey , jr . , _ determinants and alternating sign matrices _ , adv . math .
* 62 * ( 1986 ) , 169184 .
d. rush and x. shi , _ on orbits of order ideals of minuscule posets _ , j. algebraic combin .
* 37 * ( 2013 ) , 545569 .
d. rush and k. wang , _ on orbits of order ideals of minuscule posets ii : homomesy _ , preprint , arxiv:1509.08047 .
schtzenberger , _ promotion des morphismes densembles ordonns _ , discrete math .
* 2 * ( 1972 ) , 7394 .
e. sawin , personal communication ( 2013 ) .
j. stembridge , _ some hidden relations involving the ten symmetry classes of plane partitions _ , j. combin . theory ser . a * 68 * ( 1994 ) , 372409 .
j. striker and n. williams , _ promotion and rowmotion _ , eur .
j. combin .
* 33 * ( 2012 ) , 19191942
. j. striker , _ the toggle group , homomesy , and the razumov - stroganov correspondence _
, electron .
j. combin .
* 22 * ( 2015 ) , no . 2 .
h. thomas and a. yong , _ a jeu de taquin theory for increasing tableaux , with applications to @xmath0-theoretic schubert calculus _ , algebra number theory * 3 * ( 2009 ) , 121148 .
h. thomas and a. yong , _ equivariant schubert calculus and jeu de taquin _ , ann . inst .
fourier ( grenoble ) , to appear , 2013 .
b. wieland , _ a large dihedral symmetry of the set of alternating sign matrices _ , electron .
j. combin .
* 7 * ( 2000 ) .
d. zeilberger , _ proof of the alternating sign matrix conjecture _ , electron .
j. combin .
* 3 * ( 1996 ) .
|
we introduce a new concept of resonance on discrete dynamical systems .
this concept formalizes the observation that , in various combinatorially - natural cyclic group actions , orbit cardinalities are all multiples of divisors of a fundamental frequency .
our prototypical example of this phenomenon is b. wieland s gyration action on alternating sign matrices .
our main result is an equivariant bijection between plane partitions in a box ( or order ideals in the product of three chains ) under rowmotion and increasing tableaux under @xmath0-promotion .
both of these actions were observed to have orbit sizes that were small multiples of divisors of an expected orbit size , and we show this is an instance of resonance , as @xmath0-promotion cyclically rotates the set of labels appearing in the increasing tableaux .
we extract a number of corollaries from this equivariant bijection , including a strengthening of a theorem of [ p. cameron d .
fon - der - flaass 95 ] and several new results on the order of @xmath0-promotion . along the way
, we adapt the proof of the conjugacy of promotion and rowmotion from [ j. striker
n .
williams 12 ] to give a generalization in the setting of @xmath1-dimensional lattice projections .
|
a global view of dynamical systems has been one of the leading problems considered by the dynamical systems community . based on the pioneering works of peixoto and smale ,
a conjecture proposed by palis in the nineties has constituted a route guide for a global description of the space of dynamical systems .
this program , that roughly describes complement of uniform hyperbolicity as the space of diffeomorphisms that are approximated by those exhibiting either homoclinic tangencies or heteroclinic cycles , has been completed with much success in the @xmath2-topology , where perturbation tools like the pugh closing lemma , franks lemma , hayashi s connecting lemma or ma ergodic closing lemma developed for the characterization of structural stability are available ( see e.g. @xcite and references therein ) .
although the uniform geometric structures of invariant manifolds present in hyperbolic diffeomorphisms are a basic ingredient in achieving several core results and are quite well established , some important questions on the regularity of finer dynamical properties still remain to be answer .
motivated by the analysis of the regularity of the lyapunov exponents similarly to the proof of the stability conjecture , periodic orbits and their eigenvalues should play a key role .
we say that a periodic point @xmath5 of period @xmath6 for a diffeomorphism
@xmath3 has _ simple spectrum _ if all the eigenvalues of @xmath7 are real and distinct .
since all periodic points for an axiom a diffeomorphism , in the same basic piece , have the same index and are homoclinically related , then we will attribute them a signature , which consists of an ordered list of the dimensions of their finest dominated splittings .
moreover , we will say that two periodic points @xmath8 have different signatures if the oseledets splittings for @xmath7 and @xmath9 , which consist of the finest dominated splittings at the points , are distinct .
our purpose in this paper is to revisit uniformly hyperbolic basic and explore the notion of domination of oseledets splittings and characterize the continuity points of lyapunov exponent functions in this context .
these results lie ultimately in the analysis of the eigenvalues and the finest dominated splitting over the periodic points . in the first part of the paper ( see theorem [ teo1 ] ) we prove that periodic points with simple lyapunov spectrum are dense in non - trivial basic pieces of @xmath0-residual diffeomorphisms on three - dimensional manifolds ( @xmath1 ) .
in particular , any diffeomorphism with a hyperbolic basic piece can be @xmath0-approximated by a diffeomorphism with a dense set of periodic points in the continuation of the hyperbolic set .
the proof of this result explores the hyperbolic structure of hyperbolic basic pieces , hence it does not rely on the classical @xmath2-perturbation lemmas . in the case of the @xmath2-topology
, one can describe the mechanisms to deduce simple lyapunov spectrum for all invariant measures assuming the same property at periodic points , in the spirit of the previous works @xcite .
on the one hand we prove that , if all periodic points of a hyperbolic basic piece for a diffeomorphism @xmath3 have simple spectrum and the same property holds in a @xmath2 neighborhood of @xmath3 ( c.f . definition [ def : simplestar ] ) then @xmath3 has a finest dominated splitting into one - dimensional sub - bundles ( see theorem [ thm:1fds ] ) . in this case
all lyapunov exponent functions of @xmath3 are continuous in the weak@xmath4-topology .
on the other hand , if periodic points miss to have simple spectrum robustly , then , by a @xmath2-arbitrarily small perturbation , one can create an equidimensional cycle associated to a pair of periodic points which robustly exhibit different signatures ( see theorem [ thm : dichotomy ] ) .
finally , we justify that these equidimensional cycles associated to periodic points with different signatures play a similar role to the one of tangencies and heterodimensional cycles by its instability character .
in fact , the existence of homoclinic tangencies or heterodimensional cycles is often associated to the so - called newhouse phenomenon of persistence of infinitely many sources or sinks ( see e.g. @xcite and references therein ) . in this
setting , the existence of a equidimensional cycle associated to periodic points with different signatures can be used to generate ( by perturbation ) infinitely many periodic points with any of the signatures of the generating periodic orbits ( see theorem [ thm : explo ] ) . using the existence of markov partitions
we deduce that similar results hold in the time - continuous setting ( see corollary [ cor : flows ] ) .
let @xmath10 be a compact riemannian manifold @xmath10 and @xmath11 .
let @xmath12 denote the set of periodic points for @xmath3 and @xmath13 denote the non - wandering set of @xmath3 .
given an @xmath3-invariant compact set @xmath14 we say that @xmath15 is a _ uniformly hyperbolic set _ if there is a @xmath16-invariant splitting @xmath17 and constants @xmath18 and @xmath19 so that @xmath20 for every @xmath21 and @xmath22 .
we can always change the metric in order to obtain @xmath23 .
we refer to @xmath17 as the _
hyperbolic splitting _ associated to @xmath3 .
we say that @xmath15 is an _ isolated set _ if there exists an open neighborhood @xmath24 such that @xmath25 .
finally , we say that @xmath15 is _ transitive _ if there exists @xmath26 such that the @xmath3-orbit of @xmath27 is dense in @xmath15 .
along the paper unless stated otherwise we always assume that @xmath15 is a uniformly hyperbolic , isolated and transitive set . for simplicity
we call such a set a _ hyperbolic basic piece _ or simply a _ basic piece_. it is well - know that hyperbolic sets admit analytic continuations , that is , there exists a @xmath2-neighborhood @xmath28 of @xmath3 and an open neighborhood @xmath29 of @xmath15 so that @xmath30 is a basic piece for @xmath31 and @xmath32 is topologically conjugated to @xmath33 .
finally , we say that a hyperbolic basic piece is _ non - trivial _ if it does not consist of a hyperbolic periodic point .
we say that a @xmath2-diffeomorphism @xmath3 is _ axiom a _ if @xmath34 and @xmath35 is a uniformly hyperbolic set .
we observe that the non - wandering set of an axiom a diffeomorphism can be decomposed in a finite number of basic pieces .
if @xmath36 is an @xmath3-invariant probability measure , then it follows from oseledets theorem @xcite that for @xmath36-almost every @xmath27 there exists a decomposition ( _ oseledets splitting _ ) @xmath37 and , for @xmath38 , there are well defined real numbers @xmath39 called the _
lyapunov exponents _ associated to @xmath3 and @xmath27 .
it is well known that , if @xmath36 is ergodic , then the lyapunov exponents are almost everywhere constant and @xmath40 is constant . in this case
the lyapunov exponents are denoted simply by @xmath41 .
the _ lyapunov spectrum _ of a probability measure is the collection of all its lyapunov exponents .
if @xmath15 is a hyperbolic set for @xmath3 and @xmath36 is an @xmath3-invariant and ergodic probability measure supported on @xmath15 , poincar recurrence theorem implies that @xmath42 and , consequently , the decomposition @xmath43 is finer than @xmath44 .
in particular @xmath36 is hyperbolic i.e. has only non - zero lyapunov exponents . given a @xmath16-invariant decomposition @xmath45 over a basic piece @xmath15 of a diffeomorphism @xmath3 we say @xmath46 is dominated by @xmath47 if there exists @xmath18 and @xmath48 so that @xmath49 it is known that we can change the metric in order to obtain @xmath23 .
it is clear from the definition that any hyperbolic splitting is necessarily a dominated splitting .
[ def : simple ] we say that @xmath3 has a _
one - dimensional finest dominated splitting _ over a basic piece @xmath15 if there exists a continuous @xmath16-invariant decomposition @xmath50 in one - dimensional subspaces at every @xmath21 such that @xmath51 is dominated by @xmath52 for every @xmath53 .
it is clear from the previous definition that having a one - dimensional finest dominated splitting is a @xmath2-open condition for @xmath0-diffeomorphisms @xmath54 .
since dominated splittings vary continuously with the base point on a basic set @xmath15 , if the diffeomorphism @xmath3 has a one - dimensional finest dominated splitting @xmath55 then there exists a continuous function @xmath56 such that @xmath57 for every @xmath58 and @xmath21 and , consequently , the limit superior @xmath59 is everywhere well defined and it does not depend on the choice of the vector @xmath60 . defining @xmath61 and @xmath62 by @xmath63
then it follows that each value @xmath64 can be written as @xmath65 using , for any @xmath3-invariant and ergodic probability measure @xmath36 each of the values @xmath64 coincide @xmath36-almost everywhere coincide with the lyapunov exponents @xmath41 , for every @xmath58 .
more generally , for any @xmath3-invariant probability measure @xmath36 we define their _ lyapunov exponents _ by @xmath66 for every @xmath58 .
if @xmath36 is an @xmath3-invariant probability measure , the ergodic decomposition theorem guarantees that @xmath67 where each @xmath68 is an @xmath3-invariant and ergodic probability measure defined for @xmath36-almost every @xmath69 , and @xmath70 converges to @xmath68 in the weak@xmath4-topology .
thus for @xmath36-almost every @xmath26 and every @xmath58 , @xmath71 equality in can also be written as @xmath72 for every @xmath58 . in this subsection
we discuss the simplicity of the lyapunov spectrum for @xmath0-diffeomorphisms ( @xmath1 ) restricted to hyperbolic basic pieces .
[ teo1 ] assume @xmath73 and @xmath1 .
let @xmath3 be a @xmath0-diffeomorphism and @xmath15 be a non - trivial hyperbolic basic set for @xmath3 and let @xmath74 be a @xmath0-open neighborhood of @xmath3 so that the hyperbolic continuation @xmath75 is well defined for all @xmath76 .
then , there exists a @xmath0-residual subset @xmath77 such that if @xmath78 then the set of periodic orbits with simple spectrum is dense in @xmath75 .
some comments are in order . in the case that @xmath79 there are some perturbation tools sufficient to guarantee the previous result ( e.g. the methods of @xcite used in the proof of theorem [ thm : dichotomy ] ) . in the case that @xmath80 , where there is a lack of perturbation tools , we use strongly both the robustness of the hyperbolic splitting and the persistence of the hyperbolic basic piece to prove that generic diffeomorphisms have a dense set of periodic orbits with simple spectrum on non - trivial basic pieces .
we also observe that the three - dimensional assumption on the manifold is crucial . in this subsection
we provide a criterium over the periodic points for a hyperbolic basic piece to admit a one - dimensional finest dominated splitting .
[ def : simplestar ] given a basic piece @xmath15 for a diffeomorphism @xmath3 we say that @xmath32 is _ simple star _ if there exists a @xmath2-open neighborhood @xmath81 of @xmath3 so that , for any @xmath82 all periodic points for @xmath33 have simple spectrum .
we observe that after @xcite , a @xmath2-diffeomorphism is called star if it admits a @xmath2-open neighborhood formed by diffeomorphisms whose periodic points are all hyperbolic .
since the non - wandering set is uniformly hyperbolic then it is clear that the set of axiom a diffeomorphisms forms a @xmath2-open subset of star diffeomorphisms .
it follows from @xcite that periodic points are enough to determine dominated splittings at the closure of periodic points . since periodic points are dense in basic pieces we obtain the following result .
[ thm:1fds ] if @xmath32 is simple star , then @xmath3 admits a one - dimensional finest dominated splitting over @xmath15 .
moreover , the lyapunov spectrum is simple for all @xmath3-invariant ergodic probability measures and the map @xmath83 is continuous , where @xmath84 denotes the space of @xmath3-invariant probability measures supported in @xmath15 and endowed with the weak@xmath4 topology . we observe that if the diffeomorphism @xmath3 has a one - dimensional finest dominated splitting over @xmath15 , then every @xmath3-invariant probability measure supported on @xmath15 admits @xmath85 distinct lyapunov exponents that differ by a definite amount .
thus , the following is a direct consequence of theorem [ thm:1fds ] .
[ tadinho ] let @xmath15 be a basic piece of a diffeomorphism @xmath3 .
the following properties are equivalent : * @xmath86 is simple star ; * @xmath86 admits a one - dimensional finest dominated splitting ; * there exists a constant @xmath87 so that any @xmath3-invariant probability measure @xmath36 supported in @xmath15 has @xmath88 distinct lyapunov exponents and @xmath89 for every @xmath90 .
the later corollary ultimately means that if the lyapunov spectrum of periodic points simple is @xmath2-robust , then there exists a uniform gap in the lyapunov spectrum of periodic points . in this subsection
we obtain a dichotomy of one - dimensional finest dominated splitting _ versus _ equidimensional cycles with robust different signatures for basic pieces .
let us introduce these notions more precisely .
let @xmath15 be a basic set of a diffeomorphism @xmath3 and @xmath91 be a hyperbolic periodic point of period @xmath92 .
given the finest @xmath7-invariant dominated splitting @xmath93 we define the _ unstable signature at @xmath5 _ ( and denote it by @xmath94 ) to be the @xmath95-uple @xmath96 .
stable signatures at periodic points are defined analogously and denoted by @xmath97 .
unstable ( resp .
stable ) signatures describe the existence of finer splittings in the unstable ( resp .
stable ) bundle of periodic points .
[ fig4 ] [ fig3 ] on the projective space in the case of two - dimensional unstable sub - bundle ordered as figure 1.,width=377,height=86 ] [ def : equidim ] we say that a @xmath2 diffeomorphism @xmath3 exhibits an _
equidimensional cycle with robust different signatures _ on a basic piece @xmath15 if there are two periodic points @xmath98 with different stable or unstable signatures , and this property holds under arbitrarily @xmath2-small perturbations .
[ fig1 ] and @xmath99 such that @xmath100 and @xmath101.,width=321,height=151 ] in dimension three , the robustness of the different signatures is associated to the existence of a pair of periodic points whose action of the dynamical cocycle on the projective space of one periodic point exhibits a dominated splitting while for the other point it presents a rotational effect ( c.f .
figures 1 - 3 ) .
this is in strong analogy with the lack of domination observed in the presence of homoclinic tangencies . in the space of @xmath2-diffeomorphisms
the presence of homoclinic tangencies and heterodimensional cycles appears as a mechanism to create infinitely many sources or sinks ( see e.g. the newhouse phenomenon in @xcite ) .
roughly , if there exists a homoclinic tangency associated to a dissipative periodic point of saddle type then the lack of domination allows to perturb the system to mix the directions and to create a sink / source .
the infinitely many sinks / sources arise associated to some phenomenon of persistence of tangencies .
[ thm : dichotomy ] assume that @xmath102 .
let @xmath15 be a non - trivial basic piece of a diffeomorphism @xmath3 and @xmath28 a @xmath2-open neighborhood of @xmath3 on which the analytic continuation @xmath75 of @xmath15 is well - defined for all @xmath103 .
either @xmath32 has a one - dimensional finest dominated splitting , or @xmath3 can be @xmath2-approximated by diffeomorphisms @xmath31 that exhibit a equidimensional cycle in @xmath75 with robust different signatures .
we notice that the property of having one - dimensional finest dominated splitting and the property of exhibiting a equidimensional cycle with robust different signatures are @xmath2-open .
it is not hard to construct examples that lie in the boundary of the diffeomorphisms with a one - dimensional finest dominated splitting in a basic piece but not on the closure of diffeomorphisms exhibiting a equidimensional cycle with robust different signatures , due to the existence of nilpotent part . in particular , there are examples of @xmath2 diffeomorphisms exhibiting periodic points ( of low period ) with equal eigenvalues that can not be perturbed to create a complex eigenvalue . after this work
was completed we were informed that , in the @xmath2-topology , bochi and bonatti @xcite describe all lyapunov spectra that can be obtained by perturbing the derivatives along periodic orbits of a diffeomorphism .
the next result explains the conceptual similarity between equidimensional cycles with robust different signatures and the notion of heterodimensional cycles for dynamics beyond uniform hyperbolicity .
[ thm : explo ] assume that @xmath102 and let @xmath15 be a non - trivial basic piece of a diffeomorphism @xmath3 . if @xmath32 admits two periodic points @xmath8 with robust different unstable signatures , then there exists a @xmath2-open neighborhood @xmath81 of @xmath3 and a residual subset @xmath104 such that for any @xmath105 we have that : 1 .
there exists a countable infinite set of points @xmath106 with @xmath107 for all @xmath108 and 2 .
there exists a countable infinite set of points @xmath109 with @xmath110 for all @xmath111 , where @xmath75 stands for the hyperbolic continuation of the set @xmath15 . in what follows
we apply our results to the case of hyperbolic basic pieces for @xmath2-flows . by bowen and ruelle @xcite , the restriction of a @xmath2-flow to a hyperbolic basic piece for a @xmath2-flow admits a finite markov partition and is semiconjugate to a suspension flows over a subshift of finite type and there exists a one - to - one correspondance between invariant measures of the discrete - time and continuous - time dynamical systems .
indeed , given a measurable space @xmath112 , a map @xmath113 , an @xmath114-invariant probability measure @xmath115 defined in @xmath112 and a ceiling function @xmath116 in @xmath117 and bounded away from zero , consider the space @xmath118 defined by @xmath119 where @xmath120 stands for the identification between the pairs @xmath121 and @xmath122 .
the semiflow defined on @xmath123 by @xmath124 , where @xmath125 is uniquely defined by @xmath126 is called a _ suspension semiflow_. if @xmath114 is invertible then @xmath127 is indeed a flow . since @xmath128 is bounded below , if @xmath129 denotes the one dimensional lebesgue measure , then @xmath130 is a one - to - one correspondence between @xmath114-invariant probability measures and @xmath131-invariant probability measures ( we refer the reader @xcite for more details ) .
hence , if @xmath132 for a @xmath133-generic point @xmath27 then @xmath134 is the oseledets splitting associated to the vector field @xmath135 for @xmath136-almost every @xmath137 , where @xmath138 is the subbundle generated by the vector field .
in particular , the lyapunov spectrum of @xmath139 is given by the lyapunov exponents of @xmath140 and zero ( corresponding to the flow direction ) .
given the previous characterization , a hyperbolic basic piece @xmath15 for a vector field @xmath135 , the vector field @xmath141 is @xmath2-simple star if and only if the poincar map @xmath142 is @xmath2-simple star .
we also stress that such a poincar map inherits the information of the lyapunov exponents for both periodic points and all invariant measures : the lyapunov exponents for the flow and for the poincar map differ from the integral of the return time function .
the definitions from the discrete time setting extend to this setting in an analogous way . since the @xmath2-perturbation tools used in the proofs of theorems [ thm:1fds ] and [ thm : dichotomy ]
are available for flows , then the following result is a consequence of the above mentioned theorems .
[ cor : flows ] if @xmath143 is a simple star structurally stable flow , then @xmath143 admits a one - dimensional finest dominated splitting and the lyapunov exponent functions vary continuously with the invariant probability measure .
moreover , if @xmath144 we have the following dichotomy : if @xmath143 is a structurally stable flow then either @xmath143 admits a one - dimensional finest dominated splitting , or @xmath143 can be @xmath2-approximated by flows that exhibit a equidimensional cycle associated points with robust different signatures .
in this section we prove our main results .
the arguments here use strongly that @xmath10 is a three - dimensional manifold . let @xmath1 , let @xmath3 be a @xmath0-diffeomorphism and @xmath15 be a non - trivial hyperbolic basic set for @xmath3 and
let @xmath74 be a @xmath0-open neighborhood of @xmath3 so that the hyperbolic continuation @xmath75 is well defined for all @xmath76 . in particular , for every @xmath82 there exists an homeomorphism @xmath145 close to the identity so that @xmath146 .
consequently , if @xmath147 are the periodic points of period @xmath148 for @xmath149 then the continuations @xmath150 are the periodic points of period @xmath148 for @xmath31 .
let @xmath151 denote the set of periodic points with simple lyapunov spectrum .
the set @xmath152 is clearly a @xmath0-open subset of @xmath74 .
we claim that @xmath153 is a @xmath0-dense subset of @xmath74 .
this will be enough to prove the theorem because the residual subset @xmath154 is so that the set of periodic points with simple lyapunov spectrum for any @xmath105 is dense in @xmath75 .
in other words , the residual subset @xmath155 satisfies the requirements of the theorem .
thus , we are left to prove that for every @xmath22 the set @xmath153 is a @xmath0-dense subset of @xmath74 . before that we need a simple and useful lemma to perform the local @xmath0-perturbations .
this lemma is quite different from the well - known franks lemma because the approximation on the derivative depends on a pre - fixed size of neighborhood of the support of the perturbation .
[ pert ] given a @xmath0 diffeomorphism @xmath156 , @xmath157 , @xmath158 and @xmath159 , there exists @xmath160 such that the following holds : if @xmath161 is @xmath162-close to @xmath163 in the uniform norm of operators then there exists a @xmath0 diffeomorphism @xmath31 such that : for simplicity we assume all the computations in local charts in @xmath167 and that @xmath5 is fixed .
take @xmath157 and @xmath159 and consider a @xmath168 _ bump function _
@xmath169 such that @xmath170 if @xmath171 and @xmath172 if @xmath173 .
it follows from fa di bruno s formula that , given @xmath174 , we have @xmath175 where the sum is over all vectors with nonnegative integers entries @xmath176 such that we have @xmath177 .
these derivatives are bounded above by a constant @xmath178 ( depending on @xmath114 ) .
let @xmath179 be a linear map @xmath162-close to @xmath163 , where @xmath180 is very small ( to be determined later on and depending on the derivatives in ( [ fdb ] ) ) .
take a @xmath0 smooth map @xmath181 defined by @xmath182^{-1}\cdot a(p)$ ] , take @xmath183 and define @xmath184 .
clearly , by construction , @xmath31 coincides with @xmath3 in the complement of @xmath165 and @xmath185^{-1}\cdot a(p ) = a(p).\ ] ] this proves ( 2 ) and ( 3 ) .
furthermore , since @xmath186 is constant and @xmath187 , then the @xmath188-distance between @xmath3 and @xmath31 is proportional to @xmath189 and can be taken smaller than @xmath164 provided that @xmath162 is small enough .
given @xmath157 , @xmath22 and a non - trivial basic piece @xmath15 of a @xmath0-diffeomorphism @xmath3 , there exists a @xmath0-diffeomorphism @xmath31 that is @xmath164-@xmath0-close to @xmath3 and displaying a @xmath31-periodic orbit @xmath190 that is @xmath191-dense in @xmath75 and has simple lyapunov spectrum .
let @xmath157 , @xmath3 be a @xmath0 diffeomorphism , and @xmath15 be a non - trivial basic piece for @xmath3 .
assume without loss of generality that @xmath17 with @xmath192 and @xmath193 ( otherwise just consider the diffeomorphism @xmath194 ) . in this three - dimensional setting , @xmath195 has simple lyapunov spectrum if and only if there are no complex eigenvalues for @xmath196 or @xmath196 has no equal real eigenvalues .
notice that this last mentioned situation has empty @xmath0-interior .
fix @xmath195 .
up to consider @xmath197 we just assume that @xmath198 and that the restriction of @xmath3 to the non - trivial hyperbolic homoclinic class @xmath15 associated to @xmath5 is topologically mixing . in particular , @xmath199 satisfies the specification property : for each @xmath200 , there is an integer @xmath201 for which the following is true : if @xmath202 are pairwise disjoint intervals of integers with @xmath203 for @xmath204 and @xmath205 then there is a point @xmath206 such that @xmath207 for @xmath208 and @xmath209 .
every topologically mixing hyperbolic elementary set of a diffeomorphism satisfies the specification property ( see @xcite ) .
consider a @xmath210-dense set @xmath211 in @xmath15 .
then there exists @xmath212 so that for any large @xmath148 there exists a periodic point @xmath213 of period equal to @xmath214 , such that the first @xmath148 iterates of @xmath215 belong to a small neighborhood @xmath216 of @xmath5 and the piece of orbit @xmath217 is @xmath191-dense .
observe that @xmath218 is fixed and depends only on @xmath219 .
defined in this way , as @xmath148 increases , the point @xmath215 is sufficiently close to @xmath5 and spends a large amount of time near @xmath5 , and the matrix @xmath220 inherits the dynamical behavior of @xmath221^m$ ] whereas the matrix @xmath222 , corresponding to the periodic orbit computed along the whole orbit of @xmath215 .
define the compact @xmath3-invariant hyperbolic sets @xmath223 the persistence of the hyperbolic splitting @xmath224 for the map @xmath3 under @xmath2 ( thus @xmath225 ) perturbations allows us to conclude that any @xmath31 sufficiently @xmath225-close to @xmath3 also has a hyperbolic splitting @xmath226 ( with a two - dimensional unstable and one - dimensional stable bundles @xmath227 and @xmath228 , respectively ) over the @xmath31-invariant compact set @xmath229 obtained as in by taking the continuation of the hyperbolic periodic points @xmath230 .
now we will show that some @xmath0-small perturbation @xmath31 can be made in order to find a periodic point @xmath215 in @xmath229 such that all the eigenvalues of @xmath231 are real and distinct , where @xmath232 .
such point @xmath215 is @xmath191-dense by construction . actually , although explicitly defined at a neighborhood of @xmath5 , the detailed construction of the perturbation map necessarily makes use of bump - functions . for shortness
we shall omit the full detailed construction .
for any @xmath22 , let @xmath233 stand for the rotation number of @xmath234 . given a continuous arc of maps @xmath235 near @xmath3 let @xmath236 denote the oscillation of the rotation number along the arc , that is , @xmath237 .
the next claim corresponds to ( * ? ? ?
* lemma 9.3 ) in our setting .
* claim : * there exists a continuous arc of maps @xmath238}$ ] satisfying @xmath239 and @xmath240 for all @xmath241 $ ] so that the following holds : for every @xmath242 there exists @xmath243 so that @xmath244})>1 $ ] for all @xmath245 .
_ proof of the claim : _ we consider from now on a basis adapted to the splitting @xmath246 ( for @xmath3 ) , a constant @xmath247 , @xmath248 and maps @xmath249 with support in a small ball @xmath250 such that : @xmath251 as in lemma [ pert ] , consider the @xmath0-diffeomorphisms @xmath252 and let @xmath253 be sufficiently small in order to have @xmath254 @xmath164-@xmath0-close to @xmath3 for all @xmath241 $ ] . since the uniformly hyperbolic splitting varies continuously with the point and the angle between stable and unstable bundles is bounded away from zero each fiber @xmath255 and @xmath256 is a graph over @xmath257 and @xmath258 ( for all maps in the family @xmath259}$ ] ) for all points @xmath27 in the hyperbolic isolated set . identifying all the stable and unstable subspaces with @xmath260 and @xmath261 by parallel transport ,
consider the decomposition @xmath262 where the @xmath263 s correspond to the derivative @xmath264 at iterates of @xmath215 by @xmath254 inside the neighborhood @xmath216 of @xmath5 and the @xmath265 s to the derivative @xmath264 at the iterates of @xmath215 by @xmath254 that are outside it .
since , for each @xmath148 we only allow @xmath266 iterates outside the neighborhood @xmath216 we obtain that , up to consider some subsequence if necessary , @xmath267 converges uniformly to some @xmath268 as @xmath269 ( i.e. the period of the @xmath215 s increases ) .
we need to prove that for any @xmath242 there exists @xmath243 so that @xmath244})>1 $ ] for all @xmath245 .
notice that each @xmath270 is uniformly close to the composition of an homothety by a rotation of angle @xmath271 and so all give a contribution to the increase of the original rotation number by some positive constant and the claim is proved .
@xmath272 by the claim , for any @xmath157 there exists a small @xmath242 and a large @xmath148 such that @xmath254 is @xmath164-@xmath0-close to @xmath3 and @xmath273 .
therefore , @xmath274 has two equal real eigenvalues of norm larger than 1
. then , we perturb once more to obtain simple spectrum by simply consider a ` diagonal ' perturbation at @xmath215 given for @xmath275 arbitrarily small and @xmath276 , in a suitable base of eigenvalues , by the matrix @xmath277 the argument to realize such perturbation of the derivative by a @xmath0-close diffeomoprhism follows the lines of lemma [ pert ] .
this completes the proof of the proposition .
assume that @xmath32 is simple star .
we explore the fact that the set of ( hyperbolic ) periodic points is dense in @xmath15 .
if @xmath15 consists of a hyperbolic periodic point the result is immediate .
so , we assume that @xmath15 is a non - trivial hyperbolic set . using that a dominated splitting extends to the closure of a set and that @xmath149 is simple star ( thus can not be @xmath2-approximated by a diffeomorphism so that the lyapunov spectrum at some periodic point is not simple ) then , by corollary [ bgv ] , there exist @xmath278 so that there exists an @xmath279-dominated splitting @xmath280 over the set @xmath281 . if @xmath282 has dimension one we take @xmath283
otherwise , repeating the procedure with @xmath284 there exists @xmath285 and @xmath286 ( multiple of @xmath279 ) and a @xmath287-dominated splitting @xmath288 with @xmath289 . if @xmath290 has dimension one we take @xmath291 . proceeding recursively with @xmath282 and @xmath211
it follows that there exist @xmath292 and an @xmath148-dominated splitting @xmath293 using that @xmath15 is a ( non - trivial ) homoclinic class and that every non - trivial homoclinic class contains infinitely many periodic points then @xmath294 is dense in @xmath15 .
thus @xmath149 has a one - dimensional finest dominated splitting , and the first claim in the theorem follows .
now , we shall prove that the lyapunov exponent function associated to @xmath149 is continuous .
in fact , let @xmath55 be the one - dimensional , @xmath16-invariant finest dominated splitting associated to @xmath3 .
since dominated splittings varies continuously with the point then there exists a continuous function @xmath56 such that @xmath57 and , by , the lyapunov exponents are given by the integrals of continuous functions @xmath295 for every @xmath58 and @xmath21 .
it is immediate that these vary continuously with respect to @xmath36 in the weak@xmath4 topology and , consequently , the lyapunov exponent map is continuous . by the dominated splitting property it follows that the lyapunov spectrum of every invariant measure @xmath36 is simple .
this finishes the proof of theorem [ thm:1fds ] .
we observe that an alternative argument for the simplicity of the lyapunov spectrum for all invariant measures follows from the denseness of periodic measures in the set of @xmath3-invariant probability measures , by specification @xcite .
since dominated splittings also vary continuously with the diffeomorphism , then every @xmath31 that is @xmath2-close to @xmath3 , then @xmath75 is a basic piece with a one - dimensional finest dominated splitting and there exists a continuous function @xmath296 such that @xmath57 . since lyapunov exponents are computed as integrals of continuous functions ( cf . )
then given a sequence @xmath297 convergent to @xmath3 in the @xmath2-topology , and invariant measures @xmath298 that converge to @xmath299 then we also obtain that @xmath300 for every @xmath58 .
assume that @xmath102 , @xmath15 is a non - trivial basic piece of a diffeomorphism @xmath3 and @xmath28 a @xmath2-open neighborhood of @xmath3 on which the analytic continuation @xmath75 of @xmath15 is well - defined for all @xmath103 . to prove the theorem we are reduced to prove that if @xmath32 can not be @xmath2-approximated by diffeomophisms @xmath31 so that @xmath301 has a one - dimensional finest dominated splitting then @xmath3 can be @xmath2-approximated by diffeomorphisms @xmath31 that exhibits a equidimensional cycle in @xmath75 with robust different signatures
assume that @xmath302 with @xmath192 and @xmath193 .
we use the following direct consequence of ( * ? ? ?
* corollary 2.18 ) for the cocycle @xmath303 together with franks lemma and , henceforth , is specific of the @xmath2-topology . 1 .
either @xmath305 admits an @xmath148-dominated splitting along the orbit of @xmath27 or else ; 2 . for any neighborhood @xmath29 of the orbit of @xmath27 ,
there exists an @xmath164-perturbation @xmath31 of @xmath3 in the @xmath2-topology , coinciding with @xmath3 outside @xmath29 and on the orbit of @xmath27 , and such that the differential @xmath306 has all eigenvalues real and with the same modulus .
since all periodic points are homoclinically related , we are reduced to prove first that there are ( up to a @xmath2-perturbation ) two periodic points with robust different signatures .
we organize our argument in three cases .
if this is the case then @xmath309 for every @xmath308 . by corollary [ bgv ] ,
item ( 2 ) , there is a diffeomorphism @xmath31 , @xmath2-arbitrarily close to @xmath3 , and a periodic point @xmath5 for which @xmath310 has two real eigenvalues with the same modulus . since this is a non - generic property , by a @xmath2-small perturbation @xmath311 of @xmath31 , the continuation @xmath312 of the periodic point @xmath5 is so that @xmath313 has real and simple spectrum . since @xmath311 can be obtained displaying some periodic point @xmath190 in the basic piece such that @xmath314 has a complex eigenvalue ( hence @xmath315 robustly ) , then @xmath316 .
if @xmath318 has simple real spectrum we are done .
otherwise , since there are periodic points whose derivative at the period has a complex eigenvalue in this case a simple and arbitrary small @xmath2-perturbation @xmath31 supported in a neighborhood of @xmath190 in such a way that @xmath319 has real simple spectrum and , consequently , an equidimensional cycle with robust different signatures .
\(a ) all periodic points @xmath308 are so that @xmath307 has real simple spectrum ( or , in other words , @xmath320 ) .
as @xmath32 can not be @xmath2-approximated by diffeomophisms @xmath31 so that @xmath301 has a one - dimensional finest dominated splitting , then it follows from corollary [ tadinho ] that @xmath3 is not a simple star diffeomorphism .
hence , it can be arbitrarily @xmath2-approximated by a diffeomorphism @xmath31 so that either there exists a periodic point @xmath5 such that @xmath310 has one complex eigenvalue or two real equal eigenvalues .
such perturbation can be performed in order to maintain a periodic point with real simple eigenvalues . moreover
, the period of @xmath5 can be taken as larger as necessary .
we are left to prove that a perturbation can be done in order to obtain complex eigenvalues ( along @xmath46 ) for some periodic point in the case when we have two real equal eigenvalues for @xmath310 .
we will deal with this situation in the next two cases ( b ) and ( c ) .
\(b ) if there exists a periodic point @xmath5 so that @xmath321 then @xmath307 has two real eigenvalues larger than one with no nilpotent part . hence , by a small @xmath2-perturbation whose support is contained in a neighborhood of @xmath5 one can obtain a @xmath2-diffeomorphism @xmath31 so that @xmath322 has a complex eigenvalue ( this can be obtained via franks lemma and which can be written locally by @xmath323 , where @xmath324 represents the rotation of angle @xmath325 on the plane @xmath46 and the identity in @xmath47 ) .
\(c ) if all periodic points @xmath5 have two real equal eigenvalues and are such that @xmath326 we shall borrow an idea from ( * ? ? ?
* section 9 ) and proceed as follows . in this case
all of these periodic points have a nilpotent part and , in a suitable coordinate system , can be written by @xmath327 where @xmath328 denotes the nilpotent part and @xmath329 . actually , since @xmath3 is axiom a then it admits a finite markov partition @xmath330 .
let @xmath331 be given such that @xmath332 .
fix @xmath333 as above and assume , for simplicity , that @xmath334 is a fixed point . given @xmath22 ,
let @xmath335 be a periodic point so that @xmath336 for every @xmath337 , @xmath338 and @xmath339 .
for any @xmath340 the @xmath3-invariant compact set @xmath341 inherits the behavior of @xmath163 . for periodic points @xmath342 with period @xmath343 one has @xmath344 where @xmath345 is a @xmath346 matrix corresponding to the @xmath347 iterates of transitions between @xmath348 and @xmath349 , and the nilpotent part @xmath350 grows linearly with @xmath351 . in particular , @xmath352 as @xmath353 .
given @xmath354 choose @xmath355 with large period so that @xmath356 .
first we perform a @xmath357-@xmath2-small perturbation along a piece of orbit of @xmath190 so that the resulting diffeomorphism @xmath31 satisfies , in appropriate coordinates , @xmath358 if @xmath345 have complex eigenvalues we are done .
otherwise we notice that @xmath359 and that @xmath360 is continuous .
there exists an appropriate base , given by the jordan canonical form , so that the matrix @xmath345 can be written of the form @xmath361 for some real numbers @xmath362 ( depending on the @xmath218 iterates @xmath363 of the orbit of the periodic point @xmath190 ) . by the continuity of the matrix @xmath345 with the periodic point @xmath190
the real numbers @xmath362 can be taken bounded from above and below by uniform constants independent on the periodic point @xmath190 and period @xmath351 .
the strategy is to perturb the derivative of the diffeomorphism @xmath31 along a piece of orbit @xmath364 ( @xmath365 to be determined below ) in such a way that the resulting @xmath2-diffeomorphism @xmath311 is so that @xmath314 has complex eigenvalues . given @xmath354 take @xmath366 ( more precisely , @xmath367 in case of matrix @xmath368 and @xmath369 in case of matrix @xmath370 ) and take @xmath342 with period @xmath371 . by franks lemma , in the case of a matrix of the form @xmath368 then we perturb the @xmath2-diffeomorphism over the piece of orbit @xmath372 by concatenations of perturbations of the derivative to get a @xmath2-diffeomorphism @xmath31 whose derivative is obtained from the one of @xmath3 by perturbations ( @xmath373-close to the identity ) of the form @xmath374 in an appropriate base . for the resulting diffeomorphism we get that @xmath375 in the case of a matrix of the form @xmath370
we proceed analogously perturbing the derivative of the diffeomorphism @xmath3 over the piece of orbit @xmath376 by perturbations of the form @xmath377 in an appropriate base .
thus we get @xmath378 hence there exists a periodic point @xmath379 such that @xmath380 has simple real spectrum .
this finishes the proof of the theorem .
let @xmath3 be a @xmath2 diffeomorphism and @xmath15 be a hyperbolic basic piece .
assume there are two periodic points @xmath381 with robust different unstable signatures .
let @xmath28 be a @xmath2-open neighborhood of @xmath3 so that for any @xmath103 the analytic continuation @xmath75 of @xmath15 is well defined and let @xmath382 and @xmath383 denote the continuations of @xmath5 and @xmath190 , respectively , and @xmath384 and @xmath385 . since @xmath102
we may assume without loss of generality that @xmath386 and @xmath387 .
for any @xmath388 we define @xmath389 to be the set of @xmath2 diffeomorphisms @xmath390 such that @xmath31 has @xmath219 distinct periodic points with robust unstable signature equal to the one of @xmath382 and has @xmath219 distinct periodic points with robust unstable signature equal to the one of @xmath383 .
observe that , by hypothesis , @xmath391 and so @xmath392 is @xmath2-open and dense in @xmath28 .
for any @xmath393 the set @xmath389 is clearly @xmath2-open . if one shows that each @xmath389 is @xmath2-dense then the residual set @xmath394 satisfies the statement of theorem [ thm : explo ] .
assume that @xmath395 is @xmath2-dense for any @xmath396 and fix @xmath157 and any @xmath397 .
we claim that there exists @xmath398 which is @xmath164-@xmath2-close to @xmath31
. the diffeomorphism @xmath399 will be obtained from @xmath31 by @xmath2-small perturbations at two periodic points in order to obtain one more point of each robust signature . by corollary [ bgv ]
we know that there exists @xmath400 such that , for any periodic point @xmath27 of period @xmath401 either there exists an @xmath148-dominated splitting along the orbit of @xmath27 or else for any neighborhood @xmath29 of the orbit of @xmath27 , there exists an @xmath402-perturbation @xmath399 of @xmath31 in the @xmath2-topology , coinciding with @xmath31 outside @xmath29 and on the orbit of @xmath27 , and for which the tangent map @xmath403 has a real eigenvalue with multiplicity two .
assume without loss of generality that @xmath404 and @xmath405 .
the existence of @xmath383 implies that the set of periodic points @xmath27 of period larger that @xmath218 and with an @xmath148-dominated splitting along the unstable fiber is not dense in @xmath75 .
indeed , if this was not the case , then @xmath383 would have an @xmath148-dominated splitting on @xmath406 . by the previous dichotomy it follows that @xmath383 is accumulated by open sets without periodic points displaying an @xmath148-dominated splitting on @xmath407 .
thus we can pick a periodic point @xmath27 distinct from the @xmath408 marked periodic point for @xmath31 in one of these open sets and with arbitrarily large period @xmath409 .
using the dichotomy there exists an @xmath402-perturbation @xmath399 of @xmath31 in the @xmath2-topology , coinciding with @xmath31 outside @xmath29 and on the orbit of @xmath27 , and for which the tangent map @xmath403 has a real eigenvalue with multiplicity two . finally , since the periodic point @xmath27 can be chosen with an arbitrarily large period we can proceed as in subsection [ second ] and perform an @xmath402-perturbation @xmath410 of @xmath399 in the @xmath2-topology so that @xmath411 has a complex eigenvalue .
clearly , @xmath27 has an unstable robust signature @xmath412 .
if @xmath410 already have @xmath413 distinct periodic points with robust unstable signature equal to the unstable signature of @xmath414 we are done .
otherwise , we are left to show that @xmath410 can be @xmath402-approximated in the @xmath2-topology by @xmath415 exhibiting one more distinct periodic point with robust unstable signature equal to the unstable signature of @xmath414 .
indeed , if all but @xmath219 periodic points have unstable signature @xmath416 and since there is no dominated splitting restricted to the unstable fiber and along these orbits we can perform an @xmath402-approximated in the @xmath2-topology in order to obtain a distinct periodic point @xmath27 with robust unstable signature equal to @xmath417 .
this completes the proof of the @xmath2-denseness of @xmath418 in @xmath81 and finishes the proof of the theorem .
the previous construction of infinitely many periodic points of a certain signature has the same flavor of the construction of a generic set with infinite sinks and sources for diffeomorphisms with homoclinic tangencies : the perturbation are localized in regions where there is a lack of domination among the stable and unstable bundles .
* acknowledgements : * this work was partially supported by cmup ( uid / mat/ 00144/2013 ) , which is funded by fct ( portugal ) with national ( mec ) and european structural funds through the programs feder , under the partnership agreement pt2020 .
mb was partially supported by national funds through fct fundao para a cincia e a tecnologia , project pest - oe / mat / ui0212/2011 .
pv was partially supported by a cnpq - brazil .
|
in this paper we revisit uniformly hyperbolic basic sets and the domination of oseledets splittings at periodic points . we prove that periodic points with simple lyapunov spectrum are dense in non - trivial basic pieces of @xmath0-residual diffeomorphisms on three - dimensional manifolds ( @xmath1 ) . in the case of the @xmath2-topology
we can prove that either all periodic points of a hyperbolic basic piece for a diffeomorphism @xmath3 have simple spectrum @xmath2-robustly ( in which case @xmath3 has a finest dominated splitting into one - dimensional sub - bundles and all lyapunov exponent functions of @xmath3 are continuous in the weak@xmath4-topology ) or it can be @xmath2-approximated by an equidimensional cycle associated to periodic points with robust different signatures . the later can be used as a mechanism to guarantee the coexistence of infinitely many periodic points with different signatures .
|
in the past decade considerable amount of attention have been drawn both towards the experimental and theoretical studies of the internal scattering of atoms with a graphene layer and with various carbon nanostructures @xcite .
owing to the fact that these nanostructures are endowed with exceptional electronic , optical , mechanical , thermal , and magnetic properties that are of vested interest to the modern communication engineering technologies @xcite , their applications are in huge demand both in the scientific and industrial laboratories .
graphene , in particular , manifest unique properties incurring its honey comb - lattice structure that could maximize the interaction of atom on the layer .
in fact , the knowledge of the atom - graphene interactions has been very useful in the construction of the hydrogen storage devices @xcite and also plays an important role in understanding different physical , chemical and biological processes @xcite .
moreover , these interactions are connected to the phenomenon of quantum reflections whose studies are of special interest today to many experimentalists and theoreticians for explaining their exact behavior @xcite .
in addition to this , gaining insights of atom - graphene contacts are crucial in the development of graphene based electronics . in particular ,
metals adsorbed on graphene can form different types of structures and can change graphene s electronic behavior instigating towards observation of interesting physical phenomenon @xcite . among the metals that can be adsorbed on graphene ,
study for the li atom is particularly very interesting for the applications in the storage of hydrogen gas @xcite , improving efficiencies of li - ion batteries @xcite , and making superconductors @xcite .
k atoms have also been used to tune the electronic structures of the graphene bilayers @xcite .
a graphene layer rolled to make a carbon nanotube ( cnt ) has some very peculiar properties and gained special attention by the researchers world - wide @xcite .
interaction of the alkali atoms with the single walled cnts have profound applications in the purification of the cnts @xcite .
adsorbed alkali atoms have been demonstrated to act as chemical dopants on the cnts and have been used to fabricate field effective transistors @xcite .
accurate experimental measurements of the @xmath0 coefficients of any atomic systems with a graphene layer and with a cnt is extremely difficult .
a number of theoretical methods have been performed , particularly using the density functional theories @xcite , lower order many - body methods @xcite , lifshitz theory @xcite etc .
, to uncover the nature of interactions of carbon nanostructures with various materials .
klimchitskaya and co - worker have used lifshitz theory to explain the interaction between graphene layer with different materials including metal plate @xcite , conduction cylinder @xcite , atoms such as h @xcite , na , rb , cs @xcite , h@xmath1 molecule @xcite , he * ion @xcite , etc .
interaction of cnt with the h atom and h@xmath1 molecule have also been explained by them in great detail @xcite .
however , these calculations for graphene layer have been basically carried out by employing single oscillator model ( som ) for the estimation of the dynamic polarizabilities . in this paper
, we verify the results by evaluating them for the alkali atoms with their accurate values of the dynamic polarizabilities and emphasis on the need to use such accurate values by comparing our results with the som results .
moreover , the interactions between cnt and alkali metal - atoms remain to be investigated thoroughly . in view of the fact that these interactions play crucial roles in a number of applications and keeping in mind their vast experimental use
, it would be expedient to carry out more accurate theoretical analysis of the interactions of the graphene layer and cnt with the alkali atoms .
the interaction between an atom and a wall is usually modeled by calculating the interaction between the atom and its image charge ( reflection ) in the wall .
the reflection coefficients required for such calculations are well described by the widely celebrated lifshitz theory which expresses these quantities as the functions of the dynamic dielectric permittivity of the wall and of the dynamic dipole polarizabilities of the atoms @xcite .
although accurate evaluation of the dynamic dipole polarizabilities in the atomic systems require sophisticated many - body methods , but their values for the alkali atoms , which are the utmost used atoms in the ultra - cold atomic experiments , are now known reliably at least with the sufficient precision at this stage of interest @xcite .
in contrast , the dynamic dielectric permittivity values are generally known to insufficient accuracy in the material mediums due to their strenuous procedure of evaluation and could remain to be a tedious task for a long to determine them precisely . in particular , the nanostructures with thickness of the size of an atom such as the considered graphene layer and single walled cnt do not have well defined dielectric permittivity .
this entails the need for adopting suitable models to estimate the @xmath0 coefficients by introducing some effective parameters that can substitute the role of the dielectric permittivity of the wall in the lifshitz theory . in this context , the two most popular models that are often employed in the theoretical determination of the dispersion coefficients are the hydrodynamic model @xcite and the dirac model @xcite . in this work ,
we intend to apply both the models and would like to compare the obtained results using accurate values of the dynamic dipole polarizabilities .
in addition , we plan to present a very handy functional form of the radial dependent dispersion coefficients so that they can be easily derived for any arbitrary values of the atom - wall distance and the cnt radius for their convenient use in the practical applications .
this paper is organized as follows : in sec .
[ sec2 ] , we present the modified lifshitz theory for the reflection coefficient on the graphene layer and cnt in the hydrodynamic and dirac model framework .
this follows with a brief description of the method of calculations of the dynamic dipole polarizabilities in sec .
[ sec3 ] which are later used in the evaluation of the dispersion coefficients .
calculated results for the @xmath0 coefficients using accurate values of the dynamic polarizabilities and using the som model are given in sec .
[ sec4 ] . in the same section we present the dispersion coefficients for the graphene layer and cnt determined by employing both the hydrodynamic and dirac models and compare them with the results obtained for an ideal conducting medium , au , and sio@xmath1 wall which were reported earlier .
unless stated explicitly , the results are given in atomic unit ( a.u . ) throughout the paper .
the general form of the interaction potential energy in the configuration of a micro - particle and a material planar structure interacting at a distance @xmath2 is described by the lifshitz theory which is expressed by @xcite @xmath3 where @xmath4 is the fine structure constant , @xmath5 is the frequency dependent dielectric constant of wall material , @xmath2 is the separation distance between the atom and the surface and @xmath6 is the dynamic polarizability of the atom with imaginary argument .
the function @xmath7 is given by @xmath8 with the matsubara frequencies denoted by @xmath9 . for small separation distances
, the above potential can be approximated to @xmath10 where @xmath0 is known as the dispersion coefficient for the corresponding atom - wall interaction . for a perfect conductor with @xmath11
, we have @xmath12 the dispersion coefficient for a cnt with radius @xmath13 is expressed using the proximity force approximation ( pfa ) by @xcite @xmath14 where @xmath15 and @xmath16 are the reflection coefficients of the electromagnetic oscillations on cnt for the transverse magnetic and transverse electric polarizations of the electromagnetic field .
it has been shown in ref .
@xcite that relative differences between the exact and pfa results could be within 4% for the condition @xmath17 . for a thin single layer graphene with limit @xmath18 , we can simplify the above expression to @xcite @xmath19 the separation distance dependent @xmath0 coefficients given above include both the retarded and nonretarded interaction energies which are applicable up to the separation distances where the thermal effects are not significant ( typically @xmath20 ) @xcite .
the most difficult part in the evaluation of the above expressions is to get the @xmath15 and @xmath16 reflection coefficients correctly .
two different models widely used to describe the electronic structure of graphene are the hydrodynamic model and the dirac model . within the framework of hydrodynamic model , the reflection coefficients for a graphene layer or cnt are given by @xcite @xmath21 with the wave number of graphene sheet @xmath22 and @xmath23 .
in this model , graphene is considered as an infinitesimally thin positively charged sheet carrying a homogeneous fluid with some mass and negative charge densities .
the energy of the quasi - particles in graphene is quadratic with respect to their momenta .
therefore , this model works well at large energies and fails at the low energies ( where actual energy of the quasi - particles is linear function of momentum ) .
this model is an approximate one and does not take into account the dirac character of the charge carriers in graphene . within the framework of the dirac model of the electronic structure of graphene ,
the quasi - particle fermion excitations in graphene are treated as massless dirac fermions moving with a fermi velocity .
it takes into account the properties of graphene which are valid at the low energies of the quasi - particles in graphene , specifically energies which are linear function of momentum .
the explicit relations for the reflection coefficients considering the electronic structure of the graphene or cnt according to the dirac model are given by @xcite @xmath24 where the function @xmath25 determines the polarization tensor in an external electromagnetic field in three dimension space - time coordinate and is give by @xcite @xmath26 where @xmath27 is known as the mass gap parameter .
the exact value of @xmath27 remains to be unknown however , its commonly accepted upper bound value quoted in the literature is 0.1 ev @xcite .
the parameter @xmath28 in the above equation is defined in terms of the fermi velocity @xmath29 as @xmath30^{1/2 } .\ ] ] in the next section , we shall briefly discuss the method of calculations for the dynamic polarizabilities which are required for evaluating @xmath0 coefficients as discussed above and would like to compare them with the results obtained considering the som results .
the dynamic dipole polarizabilities of the ground state @xmath31 of the alkali atoms for the corresponding principal quantum number @xmath32 due to the direct current electric field with the frequency @xmath33 are given by @xmath34 \nonumber \\ & = & \frac{2}{3(2j_n+1 ) } \sum_i \frac { ( e_i - e_n ) |\langle \psi_n || d || \psi_i\rangle |^2}{(e_i - e_n)^2 - \omega^2 } , \label{pol1}\end{aligned}\ ] ] where @xmath35 is the total angular momentum of the corresponding ground state , sum over @xmath36 represents all possible allowed intermediate states for the dipole transition , @xmath37 are the energies of the corresponding states and @xmath38 is the e1 reduced matrix element of the dipole operator @xmath39 between the states @xmath31 and @xmath40 .
alternatively , the above polarizability expression can be expressed as @xmath41 with @xmath42 by solving the equation @xmath43 lclc transition & e1 mat.el . & transition & e1 mat .
+ li & & na & + @xmath44 & 3.318(4 ) & @xmath45 & * 3.5246(23 ) * + @xmath46 & 0.182(2 ) & @xmath47 & 0.304(2 ) + @xmath48 & 0.159(2 ) & @xmath49 & 0.107(1 ) + @xmath50 & 0.119(4 ) & @xmath51 & 0.056(2 ) + @xmath52 & 0.092(2 ) & @xmath53 & 0.035(2 ) + @xmath54 & 0.072(1 ) & @xmath55 & 0.026(2 ) + @xmath56 & 4.692(5 ) & @xmath57 & * 4.9838(4 ) * + + @xmath58 & 0.257(2 ) & @xmath59 & 0.434(2 ) + @xmath60 & 0.225(2 ) & @xmath61 & 0.153(2 ) + @xmath62 & 0.169(4 ) & @xmath63 & 0.081(2 ) + @xmath64 & 0.130(2 ) & @xmath65 & 0.051(2 ) + @xmath66 & 0.102(1 ) & @xmath67 & 0.037(2 ) + k & & rb & + @xmath68 & * 4.131(20 ) * & @xmath69 & * 4.227(6 ) * + @xmath70 & 0.282(6 ) & @xmath71 & 0.342(2 ) + @xmath72 & 0.087(5 ) & @xmath73 & 0.118(1 ) + @xmath74 & 0.041(5 ) & @xmath75 & 0.061(5 ) + @xmath76 & 0.023(3 ) & @xmath77 & 0.046(3 ) + @xmath78 & 0.016(3 ) & & + @xmath79 & * 5.800(8 ) * & @xmath80 & * 5.977(9 )
* + @xmath81 & 0.416(6 ) & @xmath82 & 0.553(3 ) + @xmath83 & 0.132(6 ) & @xmath84 & 0.207(2 ) + @xmath85 & 0.064(5 ) & @xmath86 & 0.114(2 ) + @xmath87 & 0.038(3 ) & @xmath88 & 0.074(2 ) + @xmath89 & 0.027(3 ) & & + the advantage of using expression given by eq . ( [ pol1 ] ) to evaluate the dynamic polarizability is that the e1 matrix elements for many important transitions that are predominantly contributing to the polarizabilities are now well studied and their values are known to quite reasonable accuracy @xcite .
use of these matrix elements along with the experimental energies will certainly give more precise contributions from these matrix elements to @xmath90 .
however , the limitation of this sum - over - states approach is that it can not estimate contributions from the core electrons and can only take into account a few low - lying intermediate states .
it has been found in the previous studies that contributions with the core orbitals and high - lying intermediate states ( tail ) are small compared to the low - lying intermediate states ( e.g. see @xcite and _ references therein _ ) .
therefore , we employ a third order many - body perturbation theory ( mbpt(3 ) method ) as described in @xcite to determine the core and tail contributions . among the important e1 matrix elements between the low - lying states , the matrix elements for few primary transitions in the na , k , and rb atoms have been obtained using a fitting procedure from the precise measurements of the lifetimes and static dipole polarizabilities of the first few low - lying excited states as have been given in @xcite .
for instance , the e1 matrix elements of the @xmath91 transitions are taken from the complied data list of ref .
the other important matrix elements whose values were not deducible accurately from the measured quantities are evaluated by employing a relativistic coupled - cluster ( rcc ) theory . in our rcc method
, we express the atomic wave function with the valence electron @xmath92 as @xmath93 where @xmath94 is the dirac - fock ( df ) wave function and @xmath95 and @xmath96 operators account the correlation effects to all orders through the excitations of the electrons from the core orbitals alone and from the valence orbital together from the core orbitals , respectively .
we consider here the singly and doubly excited configurations with important triple excited configurations in the well - known ccsd(t ) method framework for calculating the atomic wave functions .
[ cols="<,^,^,^,^",options="header " , ] one of our motivations to carry out this study is also to find out the dependence of the atom - wall interactions on the radius of cnt .
for this purpose , we present the results computed for the @xmath0 coefficients as a function of distance @xmath2 and radius @xmath13 of cnt in table .
[ cnt ] . the range for @xmath13 and @xmath2
have been chosen in accordance with the validity range of pfa . from the table
, we notice that the @xmath0 coefficients increase slowly with the increase in the cnt radii , however the rate of increase is not very magnificent . with a three - fold increase in the radius
, it raises the @xmath0 coefficients only about one and half times .
as expected these coefficients get stronger as the size of the atom increases ; i.e. from li to rb for a given separation distance @xmath2. we were unable to find out any previous work to compare our results with ; however , we have exercised the cross - checking between our results for the h atom and h@xmath1 molecule independently with the results reported in @xcite for cnt to ascertain our calculation procedure .
a lot of research work is devoted to the experimental investigation of the behavior of the interactions between the trapped atoms with the graphene layers or cnts @xcite . for simplification of reproducing the surface interaction potentials from our reported @xmath0 coefficients and for any comparison of our results with theoretical values
, we give a logistic fit for the interaction potential of the atom - graphene layer interaction using the following form @xmath97 where @xmath98 ( in a.u . ) and @xmath99 ( in @xmath100 ) are the fitting parameters that depend on the properties of the atom .
a list of these fitting parameters for the li , na , k and rb atoms are given in table [ fit ] .
the above equation is a useful tool to predict the interaction between the alkali atoms and a graphene layer for any given separation distance @xmath2. we have used the mass gap parameter @xmath27 value as 0.1 ev in our calculations .
it has been observed that a change in @xmath27 value from 0.1 ev to @xmath101 ev causes a change of 13% in the fitting parameters .
our fitting parameters for interaction between graphene and na atoms are considerably different from those calculated in ref @xcite ( @xmath98=7.11 a.u . and @xmath99=9.77 @xmath100 ) .
as mentioned previously our results are more reliable keeping in mind the error in the code in ref .
@xcite and use of our fitting parameters is recommended in extrapolating the interaction potential for graphene - alkali atom interaction .
similarly , we also fit the @xmath102 results for the interaction of these atoms with cnt .
however , a logistic equation did nt serve as a suitable fit for cnt , instead we use a rational taylor equation to fit the results in the following functional form @xmath103 and present the respective fitting coefficients with units in table [ fit ] with best goodness . in table
[ cnt ] , we compare our fitted @xmath0 coefficient values ( column labeled ii ) with those calculated using eq .
( [ eq - c3cnt ] ) and eq .
( [ eq - d ] ) ( column labeled i ) .
we see a deviation of less then 4% at all the separation distances .
to summarize , we have investigated the dispersion coefficients for the atom - graphene and atom - carbon nanotube interactions for the li , na , k , and rb atoms in this work and compared our results with the previously reported results and against the results for the interaction of atoms with a perfect conductor .
the interaction potentials of the alkali atoms are studied using both the hydrodynamic and dirac models and their dependence on the distance between the atom and the nanotube or graphene layer and radius of the nanotube are investigated .
the importance of using high precision dynamic polarizability values for such calculations specially for the heavier atoms is highlighted .
readily usable functional forms for the interaction potentials are suggested for the easy extrapolation and comparison of the experimental results with the theoretical values .
the work of b.a . is supported by csir grant no .
03(1268)/13/emr - ii , india .
acknowledges the financial support from csir .
computations were carried out using 3tflop hpc cluster at physical research laboratory , ahmedabad .
|
we evaluate separation dependent van der waal dispersion ( @xmath0 ) coefficients for the interactions of the li , na , k and rb alkali atoms with a graphene layer and with a single walled carbon nanotube ( cnt ) using the hydrodynamic and dirac models .
the results from both the models are evaluated using accurate values of the dynamic polarizabilities of the above atoms .
accountability of these accurate values of dynamical polarizabilities of the alkali atoms in determination of the above @xmath0 coefficients are accentuated by comparing them with the coefficients evaluated using the dynamic dipole polarizabilities estimated from the single oscillator approximation which are typically employed in the earlier calculations .
for practical description of the atom - surface interaction potentials the radial dependent @xmath0 coefficients are given for a wide range of separation distances between the ground states of the considered atoms and the wall surfaces and also for different values of nanotube radii .
the coefficients for the graphene layer are fit to a logistic function dependent on the separation distance . for cnt ,
we have carried out a paraboloid kind of fit dependent on both the separation distances and radii of the cnt .
these fitted functions , with the list of fitting parameters , can be used to extrapolate the interaction potentials between the considered alkali atoms and the graphene layer or cnt surface conveniently at the given level of accuracy .
|
up to one percent of the mass of the interstellar cloud surrounding the heliosphere is carried by dust grains that interact with the heliosphere @xcite . the flow of interstellar material past the sun at 26.4 drives large interstellar dust grains into the heliosphere , while small grains are diverted around the heliosphere by the interstellar magnetic field ( ismf ) .
the heliospheric trajectories of intermediate sized grains can be complicated and depend on the solar wind magnetic field and thus solar cycle phase .
observations of isdgs in the solar system by the ulysses , galileo and cassini spacecraft show that the density of smallest grains are deficient in the inner heliosphere when compared to the nominal mathis et al .
`` mrn '' power law size distribution ( e.g. * ? ? ?
* ; * ? ? ?
for instance , the density of grains of mass @xmath0 g ( @xmath1 ) is reduced in the inner heliosphere by a factor of @xmath2 below the mrn predictions and @xmath3 g ( @xmath4 ) grains are deficient by three orders of magnitude . at the same time large ( @xmath5 ) grains are absent from the mrn distribution , but are abundant in the inflowing isdgs . while several alternative grain size distribution models exist ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ) , the constraints of staying within the limits of cosmic abundances and explaining the interstellar extinction curve lead in all cases to the presence in the models of small grains that are absent from the observed distribution and a cutoff on the large grain size end that is below that of the observed distribution .
the deficiencies of small grains may result from small grains being deflected around the heliosphere because their high charge - to - mass ratios cause them to couple tightly to the ismf . on the other hand
, the circumheliospheric interstellar medium ( chism ) may simply be deficient in small grains ( balancing the overabundance of large grains ) . to evaluate which is
the case requires calculations of the grain distribution as a function of grain size in the context of realistic heliosphere models . with recent data on the shape of the heliosphere from the crossing of the termination shock ( ts ) by voyagers 1 and 2 @xcite , and on the deflection of inflowing interstellar h@xmath6 relative to he@xmath6 @xcite ,
such models face tighter constraints than ever before
. the inferred asymmetry of the heliosphere can be explained by the orientation of the ismf relative to the direction of the cloud - sun relative motion .
the ismf required to fit the heliosphere data , in turn , has direct implications for isdg trajectories . in this paper
we present new calculations of grain trajectories and density distributions for small grains that are primarily excluded from the inner heliosphere .
small grains experience enhanced charging rates due to the ejection of secondary electrons in the hot @xmath7 k plasma between the termination shock and heliopause @xcite .
we calculate grain trajectories in 3-d , based on a self - consistent steady state mhd - kinetic heliosphere model with grain charge calculated at each location along the path according to the plasma conditions and radiation field .
( we note that the recent voyager 2 results @xcite on the lower than expected plasma temperature in the inner heliosheath are not incorporated in the heliosphere model we use .
we expect that this could result in somewhat lower dust charging in this region than we calculated , though the detailed effects of charging by thermal particles vs. puis has yet to be fully explored . )
the model of plasma density and magnetic field in the heliosphere at each point of space properly accounts for the asymmetries that result from the angle between the ismf and interstellar gas / dust inflow direction , and ismf and ecliptic plane @xcite . when combined with recent models of grain charging @xcite and a realistic ultraviolet ( uv ) radiation field , the lorentz force at each point in space
can be used to calculate the grain trajectory .
total 3-d grain densities are then calculated for a sample of interstellar grain sizes based on the total gas - to - dust mass ratio in the chism and isdg models .
many previous calculations have been made of grain interactions with the heliosphere ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) but ours is the first to use a realistic heliosphere model , including distortion due to an ismf orientation that is consistent with recent observational data .
the grain trajectory calculations require a 3-d model that self - consistently includes the interaction and charge exchange between interstellar h and the solar wind plasma , and that is based on realistic boundary conditions for the heliosphere as set by the surrounding interstellar cloud .
the 10 au difference in the termination shock distances seen by voyager 1 versus voyager 2 suggests an asymmetry in the outer heliosphere topology that can be explained by the tilt of the ismf with respect to the upwind direction .
we therefore select a heliosphere model that reproduces the ts asymmetry , the hydrogen - helium offset of @xmath8 , and the angle of the ismf with respect to the ecliptic plane as indicated by the h - he offset and data on the ismf near the sun .
we use the heliosphere of @xcite and @xcite , which is based on coupled 3-d mhd ( ion ) fluid and kinetic ( neutral ) particle code including the effects of charge exchange to calculate the plasma and neutral particle properties .
the model parameters of the solar wind and inflowing interstellar material are given in table 1 ( where is the neutral h density and is the proton density ) .
the solar wind in the model carries a parker spiral type magnetic field and has @xmath9 .
the ismf strength is @xmath10 with the direction 30away from the ism inflow direction and pointed toward the southern hemisphere , and lying in the h deflection plane for this model @xcite .
the heliosphere model uses a coordinate system with the x - axis perpendicular to the ecliptic plane , the z - axis in the ecliptic plane towards upwind and perpendicular to the x - axis , and the y - axis in the ecliptic plane to form a righthanded coordinate system .
the sun is located at @xmath11 .
[ fig : np ] shows the ismf lines bending over the heliopause .
.parameters of heliosphere model with positive magnetic polarity at the nep [ cols="<,^,^,^,^,^ " , ] we illustrate the deflection of small grains around the heliosphere by calculating the charge and trajectories for spherical grains of radii 0.1 , and 0.01 , composed of astronomical silicates .
the assumption that the small grains are silicates is supported by observations and analysis indicating that the gas phase carbon abundance in the chism is high ( supersolar in fact * ? ? ?
* ) leaving none to be locked up in grains .
grain charging is calculated using the model and grain - charging code of @xcite , combined with plasma densities and temperatures from the heliosphere model , and the modeled extreme - uv / far - uv photon fluxes as discussed below .
the primary mechanisms affecting the charge of grains near and in the heliosphere are the photoelectric emission caused by absorption of photons from the uv radiation field , and the sticking of electrons and ions that collide with the grains .
additional grain charging occurs in the hot subsonic solar wind plasma in the inner heliosheath between the solar wind termination shock and the heliopause because of secondary electron emission produced by collisions of high energy electrons with the grains .
the radiation field used in our calculations is comprised of the constant interstellar radiation field from @xcite , and the solar flux data taken from the timed / see project ( see http://lasp.colorado.edu/see/see_data.html ) .
the timed / see project creates spectra using a combination of measured and modeled fluxes .
the photon flux longward of 13.6 ev is dominated by the interstellar radiation field , while the extreme ultraviolet radiation field is dominated by the sun .
the interstellar and solar fluxes have equal effect in charging grains at @xmath12 au .
the grain trajectory calculations begin with the grains upstream of the heliosphere and traveling with the interstellar gas .
however , the grains are likely to have been accelerated by turbulence in the interstellar magnetic field and thus should have some initial velocity relative to the gas . taking guidance from @xcite , we give the grains an initial 3 speed ( roughly the gas turbulent velocity observed for the local interstellar cloud ) relative to the gas , and with a net direction that is perpendicular to the magnetic field but
is otherwise randomly directed .
we integrate the equations of motion for the grains , calculating at each point the charge on the grain depending on its position ( and properties of the surrounding plasma ) and including the lorentz forces , radiation pressure and gravity of the sun . for the density distribution
we take over @xmath13 grain trajectories for grains starting in a grid in a plane far ( @xmath14 au ) from the sun in the ism .
the forces acting on the grains include gravity , radiation pressure , and the lorentz force from the motion of charged grains relative to the magnetic field lines .
the equation of motion is given by @xmath15 where @xmath16 , @xmath17 , and @xmath18 are the velocity , charge and mass of the grain , @xmath19 is the ratio of radiation pressure to gravitational attraction , @xmath20 is magnetic field , and @xmath21 is gas ( solar wind or ism ) velocity .
although radiation pressure is not significant for spherical grains in the outer heliosheath that obey mie scattering , @xcite have shown that radiation pressure acting on fluffy grains results in non - radial torques that may significantly perturb the grain motion in the outer heliosphere .
this effect is not included here .
we show sample grain trajectories for silicate grains of radii 0.01 and 0.1 in fig .
[ fig : traj ] .
the resulting density distributions of grains ( ) are shown for a slice in the plane at @xmath22 that is perpendicular to the interstellar flow direction ( fig .
[ fig : zplane ] ) , and for a meridian slice through the heliosphere nose ( fig .
[ fig : xyplane ] ) . these simulations of grain locations in the extended 3-d volume were performed with the university of alabama , huntsville , cspar cluster .
this study shows that the distortion of the local interstellar magnetic field by the heliosphere causes large - scale asymmetries in the spatial distribution of tiny interstellar grains flowing past the sun and within 1000 au .
grains of radius @xmath23 appear to be completely excluded from the heliosphere .
the trajectories of these grains will be sensitive to the direction of the local interstellar magnetic field , and relatively insensitive to the solar 22-year magnetic activity cycle since the grains do not penetrate the heliopause . in fig .
[ fig : coldens ] we show the column densities of these grains within 400 au of the sun , and outside of the heliopause .
the ismf distortion by the heliosphere is imprinted on the distribution of these small grains .
grains as small as 0.1 can partially penetrate the termination shock .
the trajectories of 0.1 grains are initially influenced by the interstellar magnetic field .
once inside of the heliosphere , the solar wind magnetic field and solar cycle become significant ( e.g. * ? ? ?
the high charge - to - mass ratio for these grains is enhanced by secondary electron ejection in the inner heliosheath , increasing the coupling between 0.1 grains and the magnetic field in this region .
we thank joe weingartner for sharing his grain - charging code with us , and richard lieu , haru washimi , and xianzhi ao for helpful comments .
this work has been supported by nasa grants nnx08aj33 g , nnx09ah50 g , nnx07ah18 g , nnx08aj21 g , nnx09ab24 g , nnx09ag29 g , and nnx09ag62 g , and nasa contract nng05ec85c . p. c. frisch , j. m. dorschner , j. geiss , j. m. greenberg , e. grn , m. landgraf , p. hoppe , a. p. jones , w. krtschmer , t. j. linde , g. e. morfill , w. reach , j. d. slavin , j. svestka , a. n. witt , and g. p. zank , _ apj _ * 525 * , 492516 ( 1999 ) .
j. heerikhuisen , n. v. pogorelov , v. florinski , and g. p. zank , `` modeling kinetic neutral atoms in the solar - wind / interstellar - medium interaction region , '' in _ numerical modeling of space plasma flows _ , edited by n. v. pogorelov , e. audit , and g. p. zank , 2008 , vol .
385 of _ astronomical society of the pacific conference series _ , p. 204
|
the distribution of interstellar dust grains ( isdg ) observed in the solar system depends on the nature of the interstellar medium - solar wind interaction .
the charge of the grains couples them to the interstellar magnetic field ( ismf ) resulting in some fraction of grains being excluded from the heliosphere while grains on the larger end of the size distribution , with gyroradii comparable to the size of the heliosphere , penetrate the termination shock .
this results in a skewing the size distribution detected in the solar system .
we present new calculations of grain trajectories and the resultant grain density distribution for small isdgs propagating through the heliosphere .
we make use of detailed heliosphere model results , using three - dimensional ( 3-d ) magnetohydrodynamic / kinetic models designed to match data on the shape of the termination shock and the relative deflection of interstellar and flowing into the heliosphere .
we find that the necessary inclination of the ismf relative to the inflow direction results in an asymmetry in the distribution of the larger grains ( 0.1 ) that penetrate the heliopause .
smaller grains ( 0.01 ) are completely excluded from the solar system at the heliopause .
address = harvard - smithsonian center for astrophysics , 60 garden st .
, cambridge , ma 02138 ( usa ) address = dept .
astronomy & astrophysics , u. chicago , 5640 s. ellis ave . ,
chicago , il 60637 ( usa ) address = u .
alabama - huntsville , center for space plasma and aeronomic research , huntsville , al 35899 address = u .
alabama - huntsville , center for space plasma and aeronomic research , huntsville , al 35899 address = dept . of physics and astronomy , 6127 wilder lab , dartmouth college , hanover , nh 03755 address = infrared processing and analysis center , caltech , pasadena , ca
91125 address = u .
alabama - huntsville , center for space plasma and aeronomic research , huntsville , al 35899 address = u .
alabama - huntsville , center for space plasma and aeronomic research , huntsville , al 35899 address = u .
alabama - huntsville , center for space plasma and aeronomic research , huntsville , al 35899
|
modern semiconductor nanodevices have reached critical device dimensions in the sub-10 nanometer range .
these devices consist of complicated two or three dimensional geometries and are composed of multiple materials .
confined geometries such as ultra thin body ( utb)@xcite , finfets@xcite and nanowires@xcite structures are usually adopted in nanometer scale device designs to obtain desired performance characteristics.most of the electrically conducting devices are not arranged in infinite periodic arrays , but are of finite extent with contacts controlling the current injections and potential modulation . typically , there are about 10000 to 10 million atoms in the active device region with contacts controlling the current injection .
these finite sized structures suggest an atomistic , local and orbital - based electronic structure representation for device level simulation .
quantitative device design requires the reliable prediction of the materials band gaps and band offsets within a few mev and important effective masses within a few percent in the geometrically confined active device regions .
etb model is usually fitted to bulk dispersions without any definition of the spatial wave function details .
however , recent _ ab - initio _ study of utbs @xcite showed that the surface carrier distribution in confined systems is strongly geometry and material dependent .
this suggests that the charge distribution for realistic predictions of nanodevice performances should be resolved with subatomic resolution .
_ ab - initio _ methods offer atomistic representations with subatomic resolution for a variety of materials .
however , accurate _ ab - initio _ methods , such as hybrid functionals @xcite , gw @xcite and bse approximations @xcite are in general computationally too expensive to be applied to systems containing millions of atoms .
furthermore , those methods assume equilibrium and can not truly model out - of - equilibrium device conditions where e.g. a large voltage might have been applied to drive carriers .
the etb methods are numerically much more efficient than _ ab - initio _ methods .
etb has established itself as the standard state - of - the - art basis for realistic device simulations .
it has been successfully applied to electronic structures of millions of atoms @xcite as well as on non - equilibrium transport problems that even involve inelastic scattering .
@xcitethe accuracy of the etb methods depend critically on the careful calibration of the empirical parameters . the traditional way to determine the etb parameters is to fit etb band structures to experimental data of bulk materials .
@xcite the etb basis functions remain implicitly defined during traditional fitting processes .
the lack of explicit basis functions makes it difficult to predict wave function dependent quantities like optical matrix elements with high precision .
more importantly , etb models parameterized by traditional fitting processes suffer from potential ambiguity when applied to ultra small structures such as utbs , nanowires and more complicated geometries .
for instance , the existing etb parameters of gaas @xcite applied to a as terminated gaas utb with an implicit hydrogen passivation model @xcite results in unphysical top valence band states as shown in fig .
[ fig : unphysical_tb_states ] : the real space probability amplitudes of _ ab - initio _ topmost valence bands correspond to confined states with the probability amplitude peaking in the center of the utb rather than the surface of the utb as in etb.in fig .
[ fig : unphysical_tb_states ] , the hybrid functional calculations include hydrogen atoms explicitly whereas the etb calculations include only their impact implicitly .
@xcite the mismatch between the envelopes of etb and _ ab - initio _ wavefunctions suggests a calibration of wave functions in the etb parameterization process is necessary .
it is also found that the method of passivation ( i.e. implicit or explicit inclusion of hydrogen atoms ) has an effect on the nature of the valence band states . .
therefore , a more fundamental fitting process that relates both the band structure and the wave functions of etb models with _ ab - initio _ calculations is desirable .
existing approaches to construct localized basis functions and tightbinding - like hamiltonians from _ ab - initio _ results include maximally localized wannier functions(mlwf ) , @xcite quasi - atomic orbitals , @xcite or dft - tb analysis .
@xcitethe mlwfs are constructed using bloch states of either isolated bands @xcite or entangled bands .
@xcitethese methods typically include interatomic interactions beyond first nearest neighbors
. however , these methods do not eliminate the above discussed ambiguity of the commonly used orthogonal sp3d5s * etb models .
furthermore , these approaches usually disregard excited orbitals which are often needed to correctly parameterize semiconductors conduction bands . in previous work ,
it was already suggested how to generate etb parameters that are compatible with typical etb models and still reproduce _ ab - initio _ results .
@xcitethis previous method was already applied to several materials such as gaas , mgo @xcite and smse @xcite and yielded a good agreement between bulk etb and _ ab - initio _ band structures .
however , the resulting wave functions did not satisfactorily agree with the _ ab - initio _ wave functions . in this paper , an improved algorithm of ref .
is presented that `` maps '' _ ab - initio _ results ( i.e. eigenenergies and eigenfunctions ) to tight binding models .
compared with the previous work@xcite , the presented method allows much better agreement of the etb and _ ab - initio _ wave functions . in this present mapping algorithm ,
rigorous , wavefunction - derived etb parameters for the hamiltonian , for highly localized basis functions , and for explicit surface passivation are obtained .
it is important to mention that the etb hamiltonian of this method can be limited to first nearest neighbor interaction .
the mapping process is applied to both bulk si and gaas to generate etb parameters and explicit basis functions from corresponding hybrid functional calculations .
it is demonstrated in this work , that the wave - function derived etb hamiltonian does not yield the ambiguity discussed with fig .
[ fig : unphysical_tb_states ] . in the same way , the transferability of the etb model to nanostructures is improved .
this is demonstrated by a comparison of etb and hybrid functional results in gaas and si utbs .
this paper is organized as follows . in section [
sec : method ] , the algorithm of parameter mapping from _ ab - initio _ calculations to tight binding models is described .
section [ sec : results ] shows the application of the mapping algorithm to bulk and utb systems .
subsection [ sec : application_to_bulk_materials ] presents the application of the present algorithm to bulk si and gaas .
bulk band structures and realspace basis functions are shown and discussed there as well .
subsection [ sec : application_to_utb ] shows the application of the algorithm to utb systems and compares etb band structures and wave functions with corresponding _ ab - initio _ results .
the algorithm and its results are summarized in section [ sec : conclusion ] .
the algorithm of the parameter mapping from _ ab - initio _ results to etb models is shown in fig .
[ fig : process_flow].as will be shown in the following , the etb parameters and basis functions are obtained in an iterative fitting procedure that spans over 5 steps ( with steps 3 through 4 being iterated ) . the resulting 1st nearest neighbor hamiltonian @xmath0 is of slater koster table type .
@xcite the resulting basis @xmath1 is composed of orthonormal real space functions @xmath2 which have the shape ( vectors are given in bold type)@xmath3 here , @xmath4 labels the atom type , whereas the @xmath5 , @xmath6 and @xmath7 are principle , angular and magnetic quantum numbers , respectively .
all materials considered in this work contain no magnetic polarization .
therefore , the basis functions are spin independent .
the tesseral spherical harmonics @xmath8 describe the dependence of the basis functions on the angular coordinates @xmath9 and @xmath10 . the functions @xmath11 and @xmath12 define the radial @xmath13 dependence of the basis functions .
the contribution of @xmath14 to the basis functions is much smaller than the contribution of @xmath15 .
the detailed shapes of the radial functions @xmath16 and @xmath17 are subject to the fitting algorithm . .
* step 1 : * first , electronic band structures @xmath18 and wave functions @xmath19 are solved which serve as fitting targets to the overall mapping algorithm@xmath20 the index @xmath21 corresponds to the band index and @xmath22 represents a momentum vector in the first brillouin zone . in principle , any method that is capable of solving band diagrams and explicit basis functions can provide these fitting targets . throughout this work , however , hybrid functional calculations are performed for step 1 .
@xcite * step 2 : * in the second step , initial guesses for the etb basis functions and etb parameters are defined . during the fitting process
, the etb basis @xmath23 is spanned by non - orthogonal functions @xmath24 given by@xmath25 the @xmath26 in eq .
( [ eq : definition_atomic_orbitals ] ) differ from the @xmath27 of the final basis functions in eq .
( [ eq : definition_atomic_orbitals_final ] ) .
the details of the initial guesses for the diagonal and off - diagonal elements of the hamiltonian @xmath28 are not essential for the overall algorithm .
nevertheless , initial guesses that follow the framework of existing etb parameter sets improve the overall fitting convergence .
urban et al . and lu et al .
discuss that interactions up to third nearest neighbors might be needed to exactly reproduce _ ab - initio _ results .
@xcitein contrast , we find that the interatomic interaction elements of @xmath28 can be limited to first nearest neighbor interactions throughout this work while still reproducing _ ab - initio _ results very well .
* step 3 : * the nonorthogonal basis functions @xmath29 in position space are transformed into the bloch representation @xcite@xmath30@xmath31 \phi_{a , n , l , m}\left ( \mathbf{r}-\mathbf{r}-\boldsymbol{\tau}_{a}\right ) , \label{eq : bloch_basis_tight_binding}\end{aligned}\ ] ] where @xmath32 is the position of atom type @xmath4 in the unit cell and the sum runs over all unit cells of the system with @xmath33 , the position of the respective cell . to improve readability of all formulas in the dirac notation ,
the indices of atom type and quantum numbers are merged into greek indices @xmath34 . for the further steps , an orthogonal basis @xmath35
is created out of the basis @xmath23 with lwdin s symmetrical orthogonalization algorithm .
@xcitesince steps 4 and 5 are formulated in the basis @xmath36 , the wave functions @xmath37 of step 1 must be transformed into this basis@xmath38 where @xmath39 @xmath40with the projection operator@xmath41 equation ( [ eq : approx_abinit ] ) contains an approximation of the _ ab - initio _ wave functions in so far that the sum over @xmath42 extends only over those orbitals that are included in the tight binding basis @xmath36 .
this basis and @xmath36 of similar etb models have much fewer basis vectors than the input _ ab - initio _ calculation .
this rank reduction is a typical outcome of rectangular transformations such as @xmath43 and is well known in the field of low rank approximations @xcite .
* step 4 : * here , the quality of the etb fitting is assessed . in this step ,
the band structures of the current etb model @xmath44 and the _ ab - initio _ input @xmath45 are compared .
if these sufficiently agree , the phases of the etb wave functions are modulated to agree with the _ ab - initio _ ones and both wave functions are compared after that .
the etb hamiltonian of step 2 is diagonalized in the basis @xmath36 of step 3 to obtain etb band structures @xmath44 and eigen vectors @xmath46@xmath47 with @xmath48 to assess the quality of the etb results is assessed , different fitness functions @xmath49,@xmath50 and @xmath51 are defined for energies , masses and wave functions respectively .
the @xmath49 and @xmath50 are given by @xmath52 where @xmath53 and @xmath54 are weights defined for each target .
as a convention for wave functions phases , another set of etb wave functions @xmath55 is introduced@xmath56 the unitary transformation @xmath57 is defined by @xmath58 with @xmath59 here , the sum over @xmath60 and @xmath61 runs over all @xmath62 etb states @xmath63 and @xmath62 _ ab - initio _
states @xmath64 with equivalent energies @xmath65 . with this transformation
, the equation holds @xmath66 for equivalent states .
this phase adaption can only work if the etbband structure is close enough to the _ ab - initio _ result .
the etb wave function fittness is given by@xmath67the weights @xmath68 are varying depending on respective fitting focusses .
deviations of @xmath69 from @xmath70 have in general two reasons : inadequate basis functions and/or eigenfunctions of a poorly approximated etb hamiltonian .
therefore , @xmath51 can be estimated as @xmath71 \left\vert \psi_{j,\mathbf{k}}^{ab}\right\rangle \right\vert ^{2}\nonumber\\ & + 2\left\vert \hat{p}\left ( \mathbf{k}\right ) \left\vert \psi _ { j,\mathbf{k}}^{ab}\right\rangle -\left\vert \tilde{\psi}_{j,\mathbf{k}}^{tb}\right\rangle \right\vert ^{2}. \label{eq : inequality}\ ] ] the first right hand side term of the last equation describes the deviation of the low - rank approximated _ ab - initio _ wave functions .
this becomes obvious with the projector property @xmath72 @xmath73 \left\vert \psi_{j,\mathbf{k}}^{ab}\right\rangle \right\vert ^{2}=\left\langle \psi_{j,\mathbf{k}}^{ab}\left\vert \left [ \hat{i}-\hat{p}\left ( \mathbf{k}\right ) \right ] \right\vert \psi_{j,\mathbf{k}}^{ab}\right\rangle .\ ] ] the second term on the right hand side of eq .
( [ eq : inequality ] ) contains information about the quality of the eigenfunctions of the approximate etb hamiltonian @xmath74 .
this is understandable when eqs .
( [ eq : approx_abinit ] ) and ( [ eq : tb_transformed_wf ] ) are inserted into this term@xmath75 .\end{aligned}\ ] ] the fitness function @xmath51 represents the major improvement over the traditional etb eigenvalue fitting ( e.g. typically limited to energies and effective masses ) .
all fitness functions are minimized by iterating over the steps 3 and 4 : the slater koster type parameters for the etb hamiltonian @xmath76 and the parameters of the radial etb basis functions @xmath77 are adjusted for every iteration of step3 . *
step 5 : * once the fitness functions are small enough to cease the iterations , it is assumed that those eigenfunctions of the etb hamiltonian @xmath78 that were subject to the fitting are identical to the eigenfunctions of the _ ab - initio _
hamiltonian @xmath79 after a transformation @xmath80@xmath81 this transformation @xmath82 is determined by a singular value decomposition of the rectangular overlap matrix of _ ab - initio _ eigenstates with etb eigenstates@xmath83 the row index @xmath84 runs over all _ ab - initio _ eigenstates - exceeding those that served as fitting targets , whereas the column index @xmath21 covers all the etb eigenfunctions .
the @xmath85 and @xmath86 are square and @xmath87 is a rectangular matrix . the transformation @xmath82 is then defined as @xmath88 @xmath82 is constructed from relevant columns of a unitary transformation . combining eqs .
( [ eq : a ] ) and([eq : tb_eigenstates_1stnns ] ) allows to determine the bloch periodic final basis functions@xmath89 the real space counterpart of @xmath90 is given by@xmath91
in this work , _ ab - initio _ level calculations of si and gaas systems were performed with vasp @xcite .
the hse06 hybrid functional @xcite is used to produce reasonable band gaps in both the bulk and the utb cases . in all hse06 calculations ,
a cutoff energy of 350ev is used .
@xmath92-point centered monkhorst pack kspace grids are used for both bulk and utb systems .
the size of the kspace grid for bulk calculations is a @xmath93 , while one for utb is @xmath94 .
k - points with integration weights equal to zero are added to the original @xmath93 or @xmath94 grids in order to generate energy bands with higher k - space resolution . the spin orbit coupling is included in band structure calculations .
small hydrostatic strains up to @xmath95 are introduced to adjust the bulk band gaps in order to match experimental results .
the lattice const used in this work is given by table [ table : sktype_tb_parameters ] . for bulk si and gaas ,
fitting targets include the band structures of the lowest 16 bands ( with spin degeneracy ) along high symmetry directions , important effective masses and wave functions at high symmetry points such as @xmath92 , @xmath96 and @xmath97 points .
etb basis functions in real space is reconstructed on @xmath93 @xmath92 center k space grid using eq ( [ eq : real_space_tb_basis ] ) .
[ c]c|c & si + [ c]ctargets + @xmath98 + @xmath99 + @xmath100 + @xmath101 + + @xmath102 + @xmath103 + @xmath104 + @xmath105 + @xmath106 + @xmath107 + @xmath108 + @xmath109 + @xmath110 + + @xmath111 + @xmath112 + & [ c]cccctb ref & hse06 & tb & error @xmath113 + @xmath114 & @xmath115 & @xmath116 & @xmath117 + @xmath118 & @xmath119 & @xmath120 & @xmath121 + @xmath122 & @xmath123 & @xmath124 & @xmath125 + @xmath126 & @xmath127 & @xmath128 & @xmath129 + & & & + @xmath130 & @xmath131 & @xmath132 & @xmath133 + @xmath134 & @xmath135 & @xmath136 & @xmath137 + @xmath138 & @xmath139 & @xmath140 & @xmath141 + @xmath142 & @xmath143 & @xmath144 & @xmath145 + @xmath146 & @xmath147 & @xmath148 & @xmath149 + @xmath150 & @xmath151 & @xmath152 & @xmath153 + @xmath154 & @xmath155 & @xmath156 & @xmath157 + @xmath154 & @xmath155 & @xmath156 & @xmath158 + @xmath159 & @xmath155 & @xmath156 & @xmath160 + & & & + @xmath161 & @xmath162 & @xmath163 & @xmath164 + @xmath165 & @xmath166 & @xmath167 & @xmath168 + + the band structures and dos of bulk si and gaas ( hse06 vs etb ) are shown in fig .
[ fig : si ek and dos ] and [ fig : gaas ek and dos ] respectively .
the band structures using existing si and gaas etb parameters@xcite are also shown in corresponding figures .
the etb band structures and dos using parameters generated by this work show better agreement with the corresponding hybrid functional results compared with the existing parameterizations . for bulk si , the existing parameterization shows a unexpected low @xmath169 band around 5 ev above topmost valence bands . in the traditional fitting process
, the @xmath169 band shows a strong preference for moving downward @xcite . due to large number of parameters to be determined , traditional ( energy - gap and effective - mass ) based fitting procedures can find local minima in their fitness functions corresponding to wave functions significantly different from those predicted by _
ab - initio _ methods .
the present method has the important advantage that optimization involves not only masses and gaps but also wavefunctions .
thus the etb wavefunctions can be kept close to their _ ab - initio _ counterparts . for gaas ,
the existing parameterization shows 2 ev higher @xmath170-type low lying valence bands . the etb parameters of bulk si and gaas are listed in table [ table : sktype_tb_parameters ] .
it can be seen from tables [ tab : targets_comparison_si ] and [ tab : targets_comparison_gaas ] , the anisotropic hole masses by etb show a remarkable agreement with hse06 results .
the principal authors of the previous works@xcite explicitly pointed out that fitting hole masses had been very difficult with the previous methods .
[ c]c|c & gaas + [ c]ctargets + @xmath98 + @xmath99 + @xmath100 + @xmath101 + + @xmath102 + @xmath103 + @xmath104 + @xmath105 + @xmath106 + @xmath107 + @xmath108 + @xmath109 + @xmath110 + + @xmath171 + @xmath172 + @xmath173 + @xmath111 + @xmath112 + @xmath174 + @xmath175 + & [ c]cccctb ref & hse06 & tb & error@xmath113 + @xmath176 & @xmath177 & @xmath178 & @xmath121 + @xmath179 & @xmath180 & @xmath181 & @xmath182 + @xmath183 & @xmath184 & @xmath185 & @xmath186 + @xmath187 & @xmath188 & @xmath189 & @xmath190 + & & & + @xmath191 & @xmath192 & @xmath193 & @xmath194 + @xmath195 & @xmath196 & @xmath197 & @xmath198 + @xmath199 & @xmath200 & @xmath201 & @xmath202 + @xmath203 & @xmath204 & @xmath205 & @xmath206 + @xmath207 & @xmath208 & @xmath209 & @xmath210 + @xmath211 & @xmath212 & @xmath213 & @xmath214 + @xmath215 & @xmath216 & @xmath217 & @xmath218 + @xmath215 & @xmath216 & @xmath217 & @xmath219 + @xmath215 & @xmath216 & @xmath217 & @xmath220 + & & & + @xmath221 & @xmath222 & @xmath223 & @xmath224 + @xmath221 & @xmath225 & @xmath223 & @xmath226 + @xmath221 & @xmath222 & @xmath223 & @xmath227 + @xmath228 & @xmath229 & @xmath230 & @xmath231 + @xmath232 & @xmath167 & @xmath144 & @xmath233 + @xmath234 & @xmath235 & @xmath236 & @xmath237 + @xmath238 & @xmath239 & @xmath240 & @xmath241 + + the orthogonal etb basis functions @xmath1 of si , ga and as atoms are shown in fig . [ fig_basisfn_si_ga_as ] .
the etb basis functions are slightly environment dependent because they are orthogonal .
thus the etb basis functions are not invariant under arbitrary rotations but invariant under symmetry operations within @xmath242 group , as pointed out by slater and koster @xcite .
it can be seen from fig .
[ fig_basisfn_si_ga_as].(a ) to ( f ) that the @xmath170 and @xmath60 orbitals show @xmath170 and @xmath60 features near the atom .
more complicated patterns in the area further away from the atom can be observed .
these complicated patterns correspond to components with high angular momentums .
the feature of orthogonal etb basis function resembles the augmented basis functions used in _ ab - initio _ level calculations such as augmented plane waves(apws ) and muffin tin orbitals(mtos ) .
the orthogonal etb basis functions have multiple angular parts in each orbital as shown by fig . [ fig_basisfn_si_ga_as].(g),(h ) and ( i ) . the @xmath170 , @xmath60 and @xmath243 type etb basis functions are dominated by components with @xmath244 , @xmath245 and @xmath246 respectively .
more than @xmath247 for the @xmath170,@xmath60 and @xmath243 orbitals are comprised of their @xmath244 , @xmath245 and @xmath246 components respectively .
the excited @xmath169 type etb basis functions have higher angular momentum and the @xmath248 components have contributions of @xmath249 to @xmath250 . the second largest contribution in @xmath169 orbitals
is the @xmath251 component with @xmath252 .
the @xmath251 component attached to the @xmath169 orbitals have angular part equivalent to real space function @xmath253 .
this is a result of the existence of @xmath253-like crystal field near each atom in zincblende and diamond structures . to validate the transferability of the etb model , band structures and eigen functions of [ 001 ] utbs passivated by hydrogen atoms
are calculated by both hse06 and etb models .
the current calculations assume no strain in the utbs . in the hse06 calculations ,
charged hydrogen atoms are used to passivate the dangling bonds of the surface atoms in gaas utbs . the surface as and ga atoms are passivated by charged hydrogen atoms with 3/4 ( denoted by @xmath254 ) and 5/4 ( denoted by @xmath255 ) electron respectively .
the charged hydrogen atoms neutralize most of the surface induced electric field in the utbs . as a result , the charge distribution and local potential shows almost flat envelopes inside the utbs .
small deviation of potential can only be observed at the surface
si / ga / as atoms .
the nearly flat potential envelope suggests geometry dependent build - in potentials are needed only for surface atoms .
thus the comparisons between self - consistent hybrid functional calculations and single shot etb calculations are fair . ) . ]
the hse06 calculations show that the hydrogen orbitals contribute to the deep valence bands , thus hydrogen atoms are considered explicitly into the etb hamiltonian of utbs in this work .
1s orbital is used as the etb basis function for hydrogen atoms .
the explicit passivation model include extra slater koster type etb parameters @xmath256 , @xmath257 , @xmath258,@xmath259 and @xmath260 .
further more , a geometry and element dependent potential @xmath261 is included for surface atoms .
the onsite energies of the surface atoms are shifted by @xmath261 .
the onsite energy of the surface ga atoms are thus @xmath262 ; and for surface as atoms , the onsite energies are @xmath263 .
here the @xmath42 stands for @xmath170,@xmath60,@xmath243 and @xmath264 orbitals .
etb parameters of si / gaas in si / gaas utbs are identical with the parameters of unstrained bulk materials provided in section [ sec : application_to_bulk_materials ] . ) .
] to determine the etb parameters of h - passivation , band structures and real space wave functions of selected bands near the fermi level of the utbs are considered as fitting targets .
the inclusion of wave functions as targets serves the purpose of correcting possible problematic states .
the target si / gaas utbs contain 17 non - hydrogen atomic layers .
parameters for hydrogen atoms are also shown in table [ table : sktype_tb_parameters ] . in gaas
utbs , as and ga are passivated by hydrogen atoms with different charge , thus the hydrogen atoms have different onsite energies when different types of atoms are passivated .
the hydrogen atoms bonding with as atoms are charged positively while the ones bonding with ga atoms are charged negatively .
consequently , the @xmath254 which forms bond with as have a higher onsite energy than the @xmath255 which forms bond with ga .
band structures of si / gaas utbs are shown in fig .
[ fig_utb_ek ] .
the etb band structures match the hse06 band structures well for energies ranging from 1ev below the topmost valence bands to 1ev above the lowest conduction bands . using the explicit etb basis functions , etb wave functions of utbs with subatomic resolution are obtained and
can be compared with corresponding hse06 wave functions .
planar averaged probability amplitudes of wave functions of the lowest conduction band and top most valence bands in si / gaas utbs are shown in fig . [ fig_utb_wf ] .
it can be seen that not only the envelope but also details in subatomic resolution of the etb planar averaged @xmath265 show agreement with corresponding hse06 results . on the other hand , fig .
[ fig_utb_wf_tb ] compares the etb atom site resolved probability amplitudes among etb models in present and previous works ( ref . ) .
the cations and anions in gaas utbs form different envelopes for all of the presented states .
the lowest conduction and highest valence states turn out to be well confined states in si utbs in all of the calculations . while , in gaas utbs , the lowest conduction states has significant contribution from the surface atoms . in si
etb probability amplitudes by previous parametrizations show similar envelopes compared to the etb and hse06 probability amplitudes in this work .
[ fig_utb_wf_tb ] ( d ) shows the problematic valence states in as terminated gaas utb by parameters of previous work .
the corresponding valence states by this work turn out to be a well confined ones . to investigate this issue in more detail , in fig .
[ fig_gaas_wf_para_passivation_cmp ] , etb atom site resolved probability amplitudes for the topmost valence states of the four possible as - terminated gaas utbs are plotted : ( a ) previous parameters @xcite and implicit passivation @xcite ; ( b ) previous parameters and explicit passivation ; ( c ) new parameters and implicit passivation ; ( d ) new parameters and explicit passivation .
it is clear that , for a given set of bulk parameters , the implicit passivation model leads to wavefunctions that are less - confined than those of the explicit passivation model . on the other hand , with the same passivation model , the etb parameters by this works shows more confined top valence states than the existing etb parameters .
thus the un - confined etb state using the existing parameter set and implicit passivation model appears to be due to both the bulk gaas parameters and the passivation model .
the implicit model@xcite replaces the @xmath170- and @xmath60-orbitals of the surface atoms by sp3 hybrids and raises the energy of the dangling hybrids by @xmath266 .
the @xmath243- and @xmath264-orbitals are left completely un - passivated , and the unconfined states of fig .
[ fig_gaas_wf_para_passivation_cmp ] ( a ) are only slightly affected by changing the value of @xmath267 .
the impact of alternate implicit passivation model to explicit passivation model is obvious by comparing sub - figures ( a ) to ( b ) , as well as ( c ) to ( d ) .
to better understand the role of bulk parameters to this behavior , we experimented by reducing the magnitude of the nearest - neighbor @xmath268-@xmath269 coupling parameters in both sets as @xmath270 , @xmath271 .
remarkably , in both cases the topmost valence - band state became much more confined .
bulk valence band wave functions in modified and original parameter sets tell the story : the general trend is that bulk sets which generate more @xmath60-like top of vb states give better confinement under passivation ( and especially implicit passivation ) than do those with higher d - content .
the reduction of @xmath272 and @xmath273 lead to more @xmath60-like top vb states .
ga terminated case has less passivation problems because its top - of - vb bulk states have more contribution from the as atoms than from the ga atoms .
[ fig_utb_gap ] shows the band gaps of the si and gaas [ 001 ] utbs as functions of utb thickness . with the etb parameters by this work
, the etb bandgaps of si and gaas utbs with thickness from 0.5 nm to 4 nm agree well with the gaps by hse06 calculations .
the etb bandgaps of si utbs using parameters from previous work also show good agreement with the hse06 results .
however the etb bandgaps of gaas utbs using parameters from previous work and implicit passivation model are of around @xmath274 lower than the hybrid functional results .
the gaps of gaas utbs terminated with ga and as atoms are very close in value for both hybrid functional and etb results in this work , however the gaps of gaas utbs terminated with ga and as atoms by previous parameterizations and implicit passivation model show 0.1 to 0.2ev discrepancies .
the band gap change in si utbs thicker than 3 nm can be model by effective mass model ( assuming parabolic e - k relation ) .
while in the gaas utbs , the discrepancies between effective mass calculations and hse06 or tb calculations are obvious for all gaas utbs presented , suggesting the non - parabolic feature of the gaas valleys have significant impact to gaas nano structures .
the gaps by previous parameterization with implicit passivation model of as terminated gaas utbs has lower confined energies due to the unconfined valence states .
it has been shown that the existing etb parameterization together with the implicit passivation model gives unphysical states in as terminated gaas utb calculations . a more reliable technique of _ ab - initio _ mapping which generates etb parameters and basis functions from _ ab - initio _ is developed .
the _ ab - initio _ mapping process is applied to both bulk si and gaas .
slater koster type etb parameters within 1st nearest neighbour approximation and highly localized etb basis functions are obtained .
the etb parameters and basis functions of si and gaas are validated in corresponding utb systems with passivation models that consider hydrogen atom explicitly .
band gaps in si and gaas utbs with different thickness are also calculated by hse06 , etb and effective mass model .
compared with the existing etb parameterizations and implicit passivation model , the etb calculations in this work show good agreements with hse06 calculations in both band structures and wave functions .
this work shows that the etb parameters by _ ab - initio _ mapping have good transferability .
the mapping method developed here significantly reduces the uncertainty in both bulk and passivation models .
nanohub.org computational resources operated by the network for computational nanotechnology funded by nsf are utilized in this work .
evan wilson from network for computational nanotechnology , purdue university is acknowledged for improving the manuscript .
|
empirical tight binding(etb ) methods are widely used in atomistic device simulations .
traditional ways of generating the etb parameters rely on direct fitting to bulk experiments or theoretical electronic bands .
however , etb calculations based on existing parameters lead to unphysical results in ultra small structures like the as terminated gaas ultra thin bodies(utbs ) . in this work
, it is shown that more reliable parameterizations can be obtained by a process of mapping _
ab - initio _ bands and wave functions to tight binding models .
this process enables the calibration of not only the etb energy bands but also the etb wave functions with corresponding _ ab - initio _ calculations .
based on the mapping process , etb model of si and gaas are parameterized with respect to hybrid functional calculations .
highly localized etb basis functions are obtained .
both the etb energy bands and wave functions with subatomic resolution of utbs show good agreement with the corresponding hybrid functional calculations .
the etb methods can then be used to explain realistically extended devices in non - equilibrium that can not be tackled with _
ab - initio _ methods .
|
sensor array imaging in a scattering medium is limited because coherent signals recorded at the source - receiver array and coming from a reflector to be imaged are dominated by incoherent signals coming from multiple scattering by the medium . for instance , in a randomly perturbed waveguide , it is known that the field becomes completely incoherent when the propagation distance becomes larger than the equipartition distance , which corresponds to the distance beyond which the source energy has been shared equally among all the propagating modes ( * ? ? ?
* chapter 20 ) .
as we will see , if the distance between the source - receiver array and the reflector is larger than the equipartition distance , then classical migration of the signals recorded at the array can not give a good image .
sources can be expensive or difficult to implement but receivers can be cheap and easy to implement , so an imaging problem in which there are a few sources ( all of them being far from the reflector ) and many receivers ( some of them being close to the reflector ) is of theoretical and practical interest . if there is a unique source far from the reflector ( farther than the equipartition distance ) and if the receiver array is close to the reflector ( closer than the equipartition distance ) , then classical migration of the recorded signals fails again .
this was shown is various contexts and we will show it again in the waveguide geometry .
however , in such a situation , another kind of migration can be used : from the work devoted to coherent interferometry imaging @xcite and ambient noise imaging @xcite , it is known that migration of cross correlations of noisy signals can be more stable than migration of the signals themselves .
the migration of cross correlations of noisy signals recorded by auxiliary passive arrays was proposed by @xcite in geophysical contexts and analyzed recently in randomly scattering open media in @xcite , and we would like to address the same problem in the waveguide geometry . indeed
the number of propagating modes is finite in the waveguide geometry so that the statistical behavior of partially coherent fields in random waveguides is very different from the open medium case @xcite . in our paper , we show that , if a receiver array can be placed close to the reflector to be imaged , then the cross correlations of the incoherent signals on this array can be used to image the reflector
. we will give a detailed resolution and stability analysis .
we will show that the statistical stability requires a broadband source and that good resolution and stability properties do not require the receiver array to span the whole cross section of the waveguide , which is an effect specific to the waveguide geometry .
the paper is organized as follows . in section [ sec : prelim ] , we review the mathematical background of the imaging problem in a random waveguide . in section [ sec : imagclas ] we describe and analyze the classical migration functional using the recorded signals and show that it can not give a good image when the propagation distance is beyond the energy equipartition distance . in section [ sec : imagf ] , we introduce the correlation - based imaging functional ; it has two versions which correspond to the time - harmonic case and the broadband case . in section [ sec : resol ]
, we analyze the resolution of the proposed imaging functionals .
detailed analyses are provided for full aperture and limited aperture arrays .
these results are based on the statistical average of the imaging functionals .
the variances of these functionals are very important as well because they determine the statistical stability of the imaging functionals . in section
[ sec : stabi ] , we study the variances of the imaging functionals .
some concluding remarks are listed at the end of the paper .
we consider linear scalar ( acoustic ) waves propagating in a two - dimensional space .
the governing equation is @xmath0 here @xmath1 is the scalar field ( acoustic pressure ) ; @xmath2 is the speed of propagation in the medium ( sound speed ) ; @xmath3 models the forcing term .
we consider a waveguide geometry , and we decompose the spatial variable @xmath4 as @xmath5 .
that is , @xmath6 is along the axis of the waveguide while @xmath7 denotes the transverse coordinate , and @xmath8 is the transverse section of the waveguide .
we assume that the forcing term is localized in the plane @xmath9 : @xmath10 where @xmath11 for some @xmath12 .
we assume that the medium is quiescent before the pulse emission , that is @xmath13 we consider dirichlet boundary conditions at the boundary of the waveguide : @xmath14 using the fourier method , the scalar field can be written as a superposition of waveguide modes .
a waveguide mode is a time - harmonic wave of the form @xmath15 with frequency @xmath16 , where @xmath17 satisfies the time - harmonic form of the wave equation without a source term : @xmath18 here , @xmath19 is the transverse laplace operator in the transverse section @xmath20 with dirichlet boundary conditions ; @xmath21 is the homogeneous wavenumber .
consequently , can be solved using the eigenmodes of @xmath22 , that is , using the orthonormal basis @xmath23 of @xmath24 given by @xmath25 the eigenvalues are simple , satisfying @xmath26 . the eigenvalues and eigenvectors are given by @xmath27 using the method of separation of variables on , we see that the waveguide mode @xmath28 can be further written as superposition of @xmath29 , where @xmath30 for a given frequency @xmath16 , there exists a unique integer @xmath31 such that @xmath32 : @xmath33 here and in the sequel , @xmath34 means the integer part of a real number @xmath35 .
the modes @xmath36 are propagating waveguide modes and @xmath37 are called the modal wavenumbers . on the other hand ,
@xmath38 are evanescent modes because they decay as @xmath39 goes to @xmath40 . from now on
we assume that the waveguide is randomly perturbed and the scalar field satisfies the perturbed wave equation @xmath41 where @xmath42 is the randomly heterogeneous speed of propagation of the medium .
we consider the case where the typical amplitude of the fluctuations of the speed of propagation is small , which we call the weakly random regime .
when the correlation length of the fluctuations is of the same order as the typical wavelength the interactions between the waves and the random medium become nontrivial . due to the small amplitude of the fluctuations
, however , the effect of the interaction becomes important only after a long propagating distance .
more exactly we assume that a randomly heterogeneous section in @xmath43 $ ] is sandwiched in between two homogeneous waveguides : the speed of propagation is of the form @xmath44 \times [ 0,\widetilde{l}_0],\\ \frac{1}{c_0 ^ 2 } & \mbox{otherwise } \end{array } \right .
\label{eq : rhokdef}\ ] ] here , @xmath45 is a mean - zero , stationary and ergodic random processes with respect to the axis coordinate @xmath39 .
it is assumed to satisfy strong mixing conditions in @xmath39 .
the relative amplitude of the fluctuations of the speed of propagation is denoted by @xmath46 .
we assume that the correlation length of the random perturbation is of the same order as the typical wavelength @xmath47 , for @xmath48 the central frequency of the source .
we assume that the propagation distance @xmath49 is much larger than the typical wavelength .
we will see that the interesting regime is when the ratio @xmath50 is of order @xmath51 , so we introduce the normalized propagation distance @xmath52 : @xmath53 in this regime the cumulative effects of the interaction of the scalar wave with the small fluctuations of the speed of propagation become of order one . for a fixed frequency @xmath16 ,
the fourier transformed scalar field @xmath28 defined by @xmath54 satisfies the equation @xmath55 \hat{p}(\omega , x , z ) = \hat{f}(\omega ) \delta(x - x_{\rm s } ) \delta(z ) .
\label{eq : pbarwer}\ ] ] to solve this equation , we make the following two simplifications that are justified in ( * ? ? ?
* chapter 20 ) or @xcite .
_ ignoring the evanescent modes_. first , we only consider the propagating modes : @xmath56 this is valid because we are mainly concerned with the scalar field for @xmath57 and the evanescent modes decay exponentially fast .
furthermore , we parameterize the complex mode amplitude @xmath58 by the amplitudes of its right- and left - going components .
let @xmath59 and @xmath60 be the amplitudes of these components , defined by @xmath61 using these representations , one obtains a system of ordinary differential equations ( odes ) for @xmath62 ( * ? ? ?
* section 20.2.4 ) or ( * ? ? ?
* section 3.1 ) .
the coefficients of the system depend on the integrated quantities of the form @xmath63 this system of odes is closed by the boundary conditions at @xmath9 where the source @xmath64 is imposed and at @xmath65 where there is no left - going component . _
forward scattering approximation_. second , we neglect the left - going ( backward ) propagating mode , assuming that they do not interact with the right - going ones .
this is valid in the limit @xmath66 when the second - order moments of @xmath45 satisfy certain conditions ( * ? ? ?
* section 20.2.6 ) or ( * ? ? ?
* section 3.3 ) . in this case , the rescaled amplitude @xmath67 of the right - going wave satisfies @xmath68 where @xmath69 denotes the @xmath31-dimensional vector @xmath70 and @xmath71 is a @xmath72 complex matrix with components @xmath73 define the propagator matrix @xmath74 to be the fundamental solution of the system , _
i.e. _ , @xmath75 with @xmath76
. then @xmath77 , where the initial amplitude @xmath78 is determined by the source @xmath64 .
in fact , integrating ( [ eq : pbarwer ] ) across the plane @xmath9 and using ( [ eq : pbarwer],[eq : leftright ] ) , we find that @xmath79 consider an array of receivers located in the plane @xmath80 of the random waveguide section , where @xmath81 and @xmath82 .
let @xmath83 be the @xmath84-entry of the propagator matrix @xmath85 .
it is the rate of conversion of the initial @xmath86-mode into the @xmath87-mode in the plane @xmath88 of the random waveguide section . in particular , we have @xmath89 repeating the argument above
, we see that the field beyond @xmath90 , that is @xmath91 for @xmath92 , are related to @xmath93 as follows @xmath94 since the random waveguide is stationary , @xmath95 has the same distribution as @xmath96
. therefore , we can apply and in proposition [ prop : momentt ] and conclude that @xmath97 in probability provided that @xmath98 .
this is equivalent to say @xmath99 using this approximation and the expressions and , we can write the scalar field at @xmath100 with @xmath101 as : @xmath102 emits a short pulse that propagates through the random waveguide . the target ( square ) in the plane @xmath103 is a reflector . the receiver array ( triangles ) in the plane @xmath80
records the signals . , width=340 ] in the imaging problem to be investigated ( see figure [ fig0 ] ) , the goal is to locate a point reflector centered at @xmath104 from signals recorded at the receiver array in the plane @xmath88 .
the reflector is supposed to be at a relatively small distance ( compared to @xmath105 ) , that is to say @xmath106 note that we also assume that @xmath107 , i.e. the distance between the reflector and the receiver array is much larger than one ( the order of magnitude of the wavelength ) , to ignore the evanescent modes emitted by the reflector .
the reflector can be modeled as a local change in the density and/or the bulk modulus of the medium , so that the sound speed is locally modified as @xmath108 where @xmath109 is a small domain around @xmath110 which represents the center of the reflector ; @xmath111 is a parameter characterizing the contrast of the reflector . with this modification , the right - hand side of should have an additional term @xmath112 .
we assume that the diameter of the scattering region @xmath109 is small compared to the typical wavelength and that the velocity contrast is such that @xmath113 satisfies @xmath114
. then we can model the scattering region by a point reflector @xmath115 _ born approximation_. the above setting allows us to solve the scalar field with the presence of the point reflector using the born approximation for the reflector .
given a fixed frequency @xmath16 , we have @xmath116 here , @xmath117 is the primary field induced by the source @xmath118 propagating through the random waveguide and computed in the previous section ( eq . ( [ eq : ppfs0 ] ) ) , and @xmath119 is the secondary field , that is the first - order scattered field due to the additional source @xmath120 at the reflector : @xmath121 \hat{p}_{\rm s } ( \omega , x , z ) = -\omega^2{\sigma_{\rm r}}\delta(\mathbf{x } - \mathbf{x}_{\rm r } ) \hat{p}_{\rm p } ( \omega , x_{\rm r},\widetilde{z}_{\rm r } ) .
\label{eq : pbars}\ ] ] note that in the born approximation one replaces the full wave field at the reflector by the primary field in the right - hand side of ( [ eq : pbars ] ) .
the primary field is solved exactly as in the previous section . summarizing the results there ,
one obtains that for @xmath122 and @xmath123 , @xmath124 the secondary field satisfies ( [ eq : pbars ] ) .
again , we solve this equation using the orthonormal basis @xmath125 and we ignore the evanescent modes . since the reflector is within a distance smaller than @xmath105 from the receiver array , for @xmath126 , the propagator matrix from @xmath127 to @xmath39 can be approximated by the identity matrix in probability , and we only need to decompose the secondary source at the reflector into waveguide modes . using , the decomposition and the fact that there is no left - going wave from @xmath128 , we find that @xmath129 we note that there is no right - going secondary wave because we do not consider back - scattering of the left - going secondary wave near the receivers .
finally , recall the expression of the primary field at the reflector , we obtain for @xmath130 that @xmath131
in this section , we introduce the classical imaging functional to localize the point reflector using the scalar ( pressure ) field recorded at the receiver array at @xmath80 .
this imaging functional is based on the migration of the array data to a search point @xmath132 .
our goal is to show that classical kirchhoff migration functional does not give a good image when the medium between the source at @xmath9 and the receiver array at @xmath80 is scattering .
the data of scalar ( pressure ) field recorded by the receivers are @xmath133 note that we consider in this section the full aperture case : the receivers span the whole cross section of the waveguide and they record data at all time .
we consider the frequency- and mode - dependent data @xmath134 according to the analysis carried out section [ subsec : reflector ] , it can be decomposed as @xmath135 from ( [ eq : ppfs ] ) and ( [ eq : psfs ] ) the primary and secondary contributions are @xmath136 with @xmath137 which can be interpreted as an illumination of the reflector .
the secondary contribution @xmath138 contains the information about the reflector , and its form ( [ eq : psj ] ) motivates the definition of the kirchhoff migration imaging functional : @xmath139 where the search point is @xmath132 with @xmath140 .
a simple case is when the source term is time - harmonic , _
i.e. _ @xmath141 with @xmath142 and @xmath143 then the data set is reduced to @xmath144 and the kirchhoff migration functional has the form @xmath145 we need to compute the mean of the imaging functional in order to characterize its resolution properties and its variance in order to characterize its stability properties .
these statistical moments depend on the moments of the propagator matrix which were studied in ( * ? ? ?
* propositions 20.6 and 20.8 ) or ( * ? ? ? * propositions 6.1 and 6.3 ) .
[ prop : momentt ] the first - order moments of the transmission coefficients have limits as @xmath146 , which are given by @xmath147 & \xrightarrow[]{{\varepsilon}\to 0 } 0 , & & \text { if } j\ne l,\\ { \mathbb{e}}[t^{\varepsilon}_{jj}(\omega ) ] & \xrightarrow[]{{\varepsilon}\to 0 } e^ { - d_{j}(\omega)l } , & & \text { otherwise.}\end{aligned}\ ] ] the second - order moments of the transmission coefficients have limits as @xmath146 , which are given by @xmath148 & \xrightarrow[]{{\varepsilon}\to 0 } e^{- q_{jl}(\omega)l } , & & \text { if } j\ne l,\\ { \mathbb{e}}[t^{\varepsilon}_{jl}(\omega ) \overline{t^{\varepsilon}_{jl}}(\omega ) ] & \xrightarrow[]{{\varepsilon}\to 0 } \mathcal{t}^{(l)}_j(\omega , l ) , & & \label{eq : prop : t1 }
\\ { \mathbb{e}}[t^{\varepsilon}_{jl}(\omega ) \overline{t^{\varepsilon}_{mn}}(\omega ) ] & \xrightarrow[]{{\varepsilon}\to 0 } 0 , & & \text { otherwise.}\end{aligned}\ ] ] the functions @xmath149 are the solutions of the system of linear equations @xmath150 the positive coefficients @xmath151 and @xmath152 and the matrix @xmath153 depend on the correlation function of the random process @xmath45 .
furthermore , we have @xmath154 \right\rvert \le c
e^{-l / l_{\mathrm{equip } } } , \quad \quad \sup_{j , l } \left\lvert \mathcal{t}_j^{(l ) } ( \omega , l ) - \frac{1}{n } \right\rvert \le c e^{-l / l_{\mathrm{equip } } } , \label{eq : stjllim}\ ] ] where @xmath155 is the equipartition distance for the mean mode powers introduced at the end of section 20.3.3 in @xcite ( or at the end of section 4.2 in @xcite ) .
the results on the first - order moments describe how the wave loses its coherence as it propagates in the random waveguide .
the results on the second - order moments describe how the wave energy becomes equipartitioned on the waveguide modes . when @xmath156 is larger than the energy equipartition length @xmath155 , then the first - order moments of the transmission coefficients are vanishing .
based on this observation , we have @xmath157 \approx 0 .\ ] ] it turns out that the fluctuations of the imaging functional are much larger than its mean .
this can be seen by studying the standard deviation of the imaging functional .
when @xmath156 is larger than the energy equipartition length @xmath155 , then the second - order moments of the transmission coefficients are vanishing except @xmath158 $ ] which converge to @xmath159 . based on this observation ,
the second - order moment of the imaging functional for a time - harmonic source is : @xmath160 & = & \frac{|\hat{f}(\omega_0)|^2 \phi_{-1}(x_{\rm s})}{n } \big [ \frac{1}{4 } \phi_{1}(x^{\rm s } ) + \big ( \frac{\omega^2 \sigma_{\rm r } n}{4}\big)^2 \phi_{-1 } ( x_{\rm r } ) & & + \big ( \frac{\omega^2 \sigma_{\rm r } n}{4}\big ) \im m \big ( { \psi ( x^{\rm s},-z^{\rm s};{x_{\rm r}},-{z_{\mathrm{r } } } ) } \psi ( x^{\rm s},z^{\rm s};{x_{\rm r}},-{z_{\mathrm{r}}})\big ) \big ] , \label{eq : kmmean1}\end{aligned}\ ] ] where , for any integer @xmath87 , we have defined @xmath161 the first term in the right - hand side of ( [ eq : kmmean1 ] ) is the contribution of the primary field .
the second term is the contribution of the secondary field .
the third term is a crossed contribution .
these results show that , when the waveguide is randomly perturbed and long enough ( longer than the equipartition distance ) , then the illumination of the reflector becomes incoherent and kirchhoff migration , which is based on coherent effects , gives a completely unstable and noisy image .
the analysis is complete in the time - harmonic case .
the analysis of the broadband case ( when the support of the source spectrum is not reduced to a single carrier frequency ) goes along the same line although it is necessary to use the asymptotic expressions of the two - frequency second - order moments of the transmission coefficients ( see ( * ? ? ?
* proposition 20.7 ) or ( * ? ? ?
* proposition 6.3 ) ) : due to the loss of coherence , the mean of the imaging functional is zero while its variance is not .
in this section , we introduce a new imaging functional to localize the point reflector using the scalar field recorded at the receiver array at @xmath80 . this functional
is based on the correlation functions of the recorded signals , which we stress in the first subsection .
let @xmath162 denote the positions of the receivers in the plane @xmath80 .
the data of scalar ( pressure ) field recorded by the receivers are @xmath163 . for simplicity
, we have assumed that the receivers record data at all time . from these data
one can form the cross correlation of the recorded field : @xmath164 in fourier domain , it has the form : @xmath165 using the decomposition @xmath166 in , we can decompose the above cross correlation function into four parts .
let @xmath167 denote the cross correlation between the primary fields at the two receivers .
thanks to the formula , it admits the expression @xmath168 let @xmath169 denote the cross correlation between the primary field at the first receiver with the secondary field at the second receiver . recall that the secondary field contains information about the waves emitted from the reflector at @xmath170 , with @xmath171 .
due to and , it admits the expression @xmath172 similarly , let @xmath173 denote the cross correlation between the secondary field at the first receiver with the primary field at the second receiver .
one verifies that @xmath174 finally , the cross correlation between the secondary fields at the two receivers is much smaller than those above and its contribution is ignored .
we neglected also the contributions from the error terms of the decomposition .
these are justified because @xmath114 consistently with the born approximation .
recall that the source in the acoustic model is due to the force @xmath175 where @xmath176 indicates the location of the source . in the rest of the paper
, we will consider two special cases as follows .
we first consider the case where the source is given by @xmath177 with @xmath178 here , @xmath48 is the carrier frequency . in the fourier domain
, we have @xmath179 here , @xmath180 is assumed to be a function with compact support or fast decay .
when @xmath181 the bandwidth has no effect and the situation is equivalent to the time - harmonic case that we address in the next section .
when @xmath182 , the pulse is said to be broadband and the bandwidth plays a role in the propagation in the waveguide for a propagation distance of the order of @xmath183 . although the analysis can be carried out in general , we restrict ourselves to the case @xmath184
because when @xmath185 , the number of propagating modes @xmath31 varies with @xmath16 over the bandwidth and the analysis is a little bit more delicate .
nevertheless , the overall picture does not change in the latter case . henceforth , @xmath186 is a fixed number in the interval @xmath187 .
let @xmath188 .
then @xmath189 in terms of the new variable @xmath190 .
further , we have the following taylor expansions @xmath191 here , @xmath192 is the derivative of @xmath193 at the carrier frequency @xmath48 ; further , the reduced wavenumber @xmath193 is also evaluated at @xmath48 .
using these formulas , the cross correlation functions become @xmath194\widetilde{l } } e^{-i(\omega_0 + { \varepsilon}^\alpha h)\tau } dh , \end{aligned}\ ] ] @xmath195\widetilde{l } } e^{i(\beta_q + \beta_m){z_{\mathrm{r } } } } e^{i(\beta'_q + \beta'_m){\varepsilon}^\alpha h { z_{\mathrm{r } } } } e^{-i(\omega_0+{\varepsilon}^\alpha h ) \tau } dh , \end{aligned}\ ] ] @xmath196 \widetilde{l}}e^{-i(\beta_q + \beta_j){z_{\mathrm{r } } } } e^{-i(\beta'_q + \beta'_j ) { \varepsilon}^\alpha h { z_{\mathrm{r } } } } e^{-i(\omega_0+{\varepsilon}^\alpha h ) \tau } dh .
\end{aligned}\ ] ] a simple case is when the source term is time - harmonic , _
i.e. _ @xmath141 with @xmath142 and @xmath143 in this case , the wave field has the form @xmath197 .
the definition of the correlation function should be modified to @xmath198 the second equality holds because the integrand above is in fact independent of @xmath199 . using the decomposition and the expressions and , we obtain the following expressions for the cross correlations : @xmath200 @xmath201 @xmath202 we are now ready to present the imaging functionals , which consist in migrating the cross correlations of the recorded signals .
the imaging functionals are designed according to the settings of receiver arrays .
we consider two cases .
the ideal case is when the receiver array spans the whole cross section , _
i.e. _ , @xmath203 . then given the data , for any pair of modes @xmath204 and @xmath205
, we define @xmath206 due to the orthogonality of @xmath207 , the function @xmath208 is the @xmath84 mode of the cross correlation function @xmath209 . a search point for the reflector
will be denoted as @xmath132 where @xmath210 is its transversal coordinate and @xmath211 is its axial coordinate .
equivalently , @xmath212 is the axial coordinate starting from the receiver array .
we define the imaging functional @xmath213 as @xmath214 the choice of multiplication by @xmath215 is suggested from our analysis of the correlation functions in the next section .
a realistic situation is when the receiver array only covers part of the transversal section , _
i.e. _ , @xmath216 $ ] for @xmath217 .
consequently , the exact @xmath84 component of the cross correlation can not be extracted . in this case , we design the following imaging functional @xmath218 as @xmath219 again , these definitions of imaging functionals are based on the analysis of the correlation functions in the next section .
we remark also that it is possible to show that when @xmath220 , the second functional @xmath221 is very close to @xmath222 and we have in fact : @xmath223
in this section , we analyze the imaging functionals proposed above to search the reflectors in the waveguide . due to the random perturbations of the long section
@xmath224 $ ] of the waveguide , the values of the imaging functionals , which depend on the waveguide parameters through the data , are random . hence , we analyze the mean of the imaging functional and show that it achieves its maximum at the reflector location .
we also analyze how this mean decays from its maximum ; this information provides the resolution of the proposed imaging functionals .
we emphasize that the above observation on the mean of the imaging functional itself is not enough to claim that the functionals are effective , because it is not certain , _ a priori _ , that the one realization in practice is well reflected by the mean .
statistical stability ( _ i.e. _ , the analysis of the fluctuations of the imaging functionals ) is needed to secure this claim .
this will be investigated in the section [ sec : stabi ] .
we first consider the ideal case where the receiver array @xmath162 spans the whole cross section , so the imaging functional @xmath213 is chosen .
the key tool for analysis of the mean of @xmath213 is proposition [ prop : momentt ] . as shown by proposition [ prop : momentt ] ,
when @xmath156 is larger than the energy equipartition length @xmath155 , the main contribution from terms of @xmath225 comes from when @xmath226 and @xmath227 . following this observation , we have the following limits for the cross correlations . @xmath228 where @xmath229 is defined by ( [ def : phij ] ) and we have also used the fact that @xmath230 converges to @xmath159 in the regime @xmath231 .
following the same lines , we have @xmath232 and @xmath233 in fact , the migration imaging functional @xmath234 is designed from the above characterization of the cross correlation function . from the calculations before , we find @xmath235 where @xmath236 is the kronecker symbol .
we also find @xmath237 @xmath238 therefore , for a search point @xmath132 , with @xmath140 , we have the following . from now on
, @xmath239 denotes the carrier wavelength . if @xmath240 , @xmath241 , and @xmath242 , then @xmath243 \simeq \frac{\pi \omega_0 ^ 2 { \sigma_{\rm r}}}{32a^3 } \re e \big\ { \big [ \int_{-\frac{\pi}{2}}^{\frac{\pi}{2 } } \cos\theta e^{i(\tilde{\eta } \cos\theta + \tilde{\xi } \sin\theta)}d\theta \big]^2 \big\ } , \label{eq : prop1}\ ] ] where we have introduced
the normalized cross range offset @xmath244 , and the normalized range offset @xmath245 .
therefore , the imaging functional @xmath213 works well to detect the point reflector @xmath246 .
in particular we can see that both the range and cross - range resolutions are of the order of the wavelength @xmath247 .
note that @xmath240 means that @xmath248 .
calculations show that @xmath249 & \rightarrow & \frac{-i \phi_{-1}(x_{\rm s})}{4n } \big [ \frac{1}{n } \sum_{j=1}^n \beta_j \phi^2_j(x^{\rm s } ) e^{-i2\beta_j z^{\rm s } } \big ] \\ & & \hspace*{-0.5 in } + \frac{\omega_0 ^ 2 { \sigma_{\rm r}}\phi_{-1}(x_{\rm s})}{8 } \psi(x^{\rm s},z^{\rm s};x_{\rm r},z_{\rm r})^2 - \frac{\omega_0 ^ 2 { \sigma_{\rm r}}\phi_{-1}(x_{\rm s})}{8 } \psi(x^{\rm s},z^{\rm s};x_{\rm r},-z_{\rm r})^2 , \hspace*{0.2 in } \label{eq : espifap}\end{aligned}\ ] ] where @xmath250 and @xmath251 are defined by ( [ def : psi ] ) and ( [ def : phij ] ) . the first term does not contain information about the reflector and can be viewed as background field . in fact , its contribution is negligible because the fast oscillations in the complex exponential ( @xmath252 ) average out and there is an overall factor @xmath159 . for the second term
, we can use an integral approximation for the sum in the continuum limit ( @xmath248 ) .
that is , @xmath253 since the phase becomes zero when @xmath254 , this function peaks at @xmath255 and @xmath256 .
the integral considered above can be written as , with the second term neglected , @xmath257 similarly , the sum in the third term can be approximated by @xmath258 note that this function does not have a peak comparable with the previous function .
in fact , stationary phase calculation shows that it is of order @xmath259 where @xmath260 .
the evaluation of @xmath261 $ ] follows the same lines : @xmath262 & \rightarrow & \frac{i \phi_{-1}(x_{\rm s})}{4n } \big [ \frac{1}{n } \sum_{j=1}^n \beta_j \phi^2_j(x^{\rm s } ) e^{i2\beta_j z^{\rm s } } \big ] \\ & & \hspace*{-0.5 in } - \frac{\omega_0 ^ 2 { \sigma_{\rm r}}\phi_{-1}(x_{\rm s})}{8 } \psi(x^{\rm s},-z^{\rm s};x_{\rm r},z_{\rm r})^2 + \frac{\omega_0 ^ 2 { \sigma_{\rm r}}\phi_{-1}(x_{\rm s})}{8 } \psi(x^{\rm s},-z^{\rm s};x_{\rm r},-z_{\rm r})^2 , \hspace*{0.2 in } \label{eq : espifam}\end{aligned}\ ] ] and we find a peak at @xmath254 and @xmath256 .
finally we have @xmath263 which completes the proof of the proposition .
to eliminate the @xmath193 on the denominator , we multiplied by the exact @xmath193 factor in constructing the imaging function .
alternatively , we can differentiate the cross correlation function to remove the denominator ; see .
if we do so , the above approximation will lead to the integral @xmath264 in the square brackets of eq .
( [ eq : prop1 ] ) . this function should be compared with in the next section .
next , we consider the realistic setting where the receiver array only covers part of the transverse section , namely @xmath265 and @xmath266 . for a search point
@xmath132 , with @xmath267 , we have the following result .
[ prop2]if @xmath268 , @xmath269 , and @xmath242 , then @xmath270 \simeq \frac{\pi
\omega_0 ^ 2 k^2 { \sigma_{\rm r}}(a_2-a_1)^2}{32a^5 } \re e \big\ { \big [ \int_{-\frac{\pi}{2}}^{\frac{\pi}{2 } } \cos^2\theta e^{i(\tilde{\eta } \cos\theta + \tilde{\xi } \sin\theta)}d\theta \big]^2 \big\ } , \label{eq : eila}\ ] ] where we have introduced the normalized cross range offset @xmath271 , and the normalized range offset @xmath272 .
this proposition gives the form of the point spread function of the mean imaging functional , that is the normalized form of the peak centered at the reflector location .
the range and cross - range widths of the peak are the range and cross - range resolutions .
since the variables @xmath273 and @xmath274 are normalized with respect to the wavelength , we see that both the range and cross - range resolutions are of order of the wavelength , which is the diffraction limit .
the form of the peak can can be seen in figure [ fig1 ] which plots the transverse and longitudinal shapes of the point spread function @xmath275 ^ 2 \big\ } = \pi^2 \frac{j_1(\tilde{\xi})^2}{\tilde{\xi}^2 } , \\
g(\tilde{\eta})&= & \re e \big\ { \big [ \int_{-\frac{\pi}{2}}^{\frac{\pi}{2 } } \cos^2\theta e^{i\tilde{\eta } \cos\theta } d\theta\big]^2 \big\ } = \pi^2 j_1 ' ( \tilde{\eta})^2 - \big [ \int_{-\frac{\pi}{2}}^{\frac{\pi}{2 } } \cos^2\theta \sin ( \tilde{\eta } \cos\theta ) d\theta\big]^2 , \end{aligned}\ ] ] where we have used formula 9.1.20 @xcite to express @xmath190 and @xmath276 in terms of the bessel function @xmath277 . @xmath190 and @xmath276 are even functions maximal at @xmath278 and stationary phase calculations also show that @xmath279 proposition [ prop2 ] also shows that the resolution of the mean imaging functional does not depend on the aperture of the array @xmath280 .
this is a consequence of the waveguide geometry , since multiple reflections at the boundaries of the waveguide generate multiple replicas of the receiver array in the plane @xmath80 , which gives an effective aperture that is large enough to reach the diffraction limit .
and @xmath281 which give the normalized form of the point spread function in the range direction ( @xmath282 , left picture ) and in the cross range direction ( @xmath281 , right picture).,title="fig:",width=264 ] and @xmath281 which give the normalized form of the point spread function in the range direction ( @xmath282 , left picture ) and in the cross range direction ( @xmath281 , right picture).,title="fig:",width=264 ] since the two functions above can be analyzed in the same manner , we focus on one of them , namely @xmath283 .
as before , the cross correlation @xmath209 can be treated component - wise . for the function @xmath283 ,
the main contribution comes from the primary - secondary component @xmath169 , which we focus on for the moment .
let @xmath284 denote this main term , _
@xmath285 then from and the fact that @xmath286 and @xmath287 , we verify that @xmath288 \longrightarrow & \frac{\omega_0 ^ 2 { \sigma_{\rm r}}\phi_{-1}(x_{\rm s})}{8n^2 } \sum_{q , j , q',j'=1}^n \beta_j \beta_q \phi_j({x_{\rm r } } ) \phi_{j'}(x^{\rm s } ) \phi_q({x_{\rm r } } ) \phi_{q'}(x^{\rm s } ) e^{i(\beta_j + \beta_q){z_{\mathrm{r } } } } e^{-i(\beta_{j'}+\beta_{q ' } ) z^{\rm s } } \\ & \times \int_{\mathcal{a}^2 } \phi_j(x_1 ) \phi_{j'}(x_1 ) \phi_q(x_2 ) \phi_{q'}(x_2 ) dx_1 dx_2 . \end{aligned}\ ] ] if @xmath289 , the last integral will be @xmath290 and we are back in the case of full aperture array . in the current situation ,
this integral has to be dealt with more carefully .
the above limit can be further written as @xmath291 ^ 2 .
\label{eq : si1ps}\ ] ] introducing the difference index @xmath292 , the double sum above can written in terms of @xmath293 and @xmath86 .
the integral over @xmath294 can be calculated explicitly as follows : @xmath295 .
\end{aligned}\ ] ] for most @xmath293 , the first term above dominates .
hence , the number of indices @xmath86 s so that the above quantity is significant is roughly of order @xmath296 . using the explicit expression @xmath297 , we rewrite the function inside the brackets in as @xmath298 using the explicit expression @xmath299 and the fact that the total number of modes @xmath300 is @xmath301 and assuming that @xmath302 is an integer for simplicity , we see that for @xmath303 , @xmath304 using this expansion and some trigonometric identities , we approximate the integral above by @xmath305 \sin ( \frac{\pi j ' x^{\rm s } } { a } ) e^{-i\frac{k^2}{\beta_{j ' } } \frac{j'}{n } \frac{l}{n } { z_{\mathrm{r}}}}\\ \times & \left [ \sin^2 ( \frac{\pi j ' x_1}{a } ) \cos ( \frac{\pi l x_1}{a } ) + \cos \sin ( \frac{\pi j ' x_1}{a } ) \sin ( \frac{\pi l x_1}{a } ) \right ] e^{i 2\pi \sqrt{1-(\frac{j'}{n})^2 } \frac{{z_{\mathrm{r}}}- z^{\rm s } } { \lambda } } dx_1 . \end{aligned}\ ] ] defined @xmath306 and @xmath307 .
substitution of some further trigonometric identities transform the above integral into @xmath308 here , @xmath309 consists of products of trigonometric functions , and for each product , one of the trigonometric functions is evaluated at @xmath310 , @xmath311 or @xmath312 . assuming that @xmath313 and @xmath314 are of order one , we observe that comparing with @xmath315 , which is explicitly written above and in which @xmath316 can be of order much smaller than one , @xmath317 varies much faster as @xmath293 varies .
in other words , @xmath317 can be viewed as a fast variable , and @xmath318 with @xmath319 is a slow variable .
similarly , in the expression of the complex potentials and @xmath320 for fixed @xmath86 , the variable @xmath321 with @xmath248 involved is also a slow variable .
consequently , the contribution of @xmath322 to the sum over @xmath293 is negligible comparing with those of the terms explicitly written .
this two - scale analysis also shows that the above integral is much smaller for large @xmath316 comparing with @xmath319 .
we consider the regime @xmath248 , and use the continuum approximation to rewrite the sum over @xmath293 as an integral with respect to the variable @xmath323 .
we further assume @xmath324 , which is equivalent to say @xmath325 ; the second relation validates the linearization and allows us replacing @xmath326 by @xmath327 , while the first relation justifies the usage of poisson summation formula @xmath328 \end{aligned}\ ] ] and @xmath329 \end{aligned}\ ] ] note that we have used the fact that @xmath330 again . for each fixed @xmath331 , let @xmath332 , @xmath333 , denote the four dirac distributions .
define also @xmath334 , @xmath335 ; then we have @xmath336 ; the integral above becomes @xmath337 dt . \end{aligned}\ ] ] if we extend the domain of @xmath199 to @xmath338 , the above integral simplifies to @xmath339 dt.\ ] ] integrate over @xmath294 first .
the two dirac distributions restrict the range of integration of @xmath199 to the following intervals respectively : @xmath340 to interpret these conditions , imagine that the boundaries of the waveguide are two mirrors , then the array @xmath341 has an image @xmath342 in the lower mirror ; the two mirrors then generate a replica of images @xmath343 and @xmath344 .
if we define an angle @xmath345 by @xmath346 then the above restrictions of the dirac measures can be restated as @xmath347 the first one restrict the angle @xmath345 to those formed by the reflector and the array @xmath348 and the images of this array .
the second set restrict the angle @xmath345 to those formed by the reflector and the image array @xmath349 and the replicas . to analyze the resulting integral
, we consider the simplest set - up where : @xmath348 is centered in the cross section ; the reflector is also centered in the cross range direction , _
i.e. _ @xmath350 . further we assume the large distance regime @xmath351 .
in such a setting , with the notation @xmath352 and @xmath353 , the integral above boils down to @xmath354 since @xmath355 , for each fixed @xmath331 , the integral is over a very small angle section .
hence we can approximate the integral by the value at mean angle times the length of the angle section . the mean angle @xmath356 in the angle section is @xmath357 .
we further check that @xmath358 .
consequently , with @xmath359 set to @xmath360 , the sum becomes @xmath361 finally , let us verify that the primary - primary and secondary - primary cross correlations do not have significant contributions to the imaging functional @xmath283 .
let @xmath362 and @xmath363 denote these two terms respectively .
similar to , the expectation of @xmath363 converges , as @xmath66 , to @xmath364 ^ 2 .
\label{eq : si1sp}\ ] ] this function has the same form of and can be analyzed in the same way .
the key of these two functions is that the phase function in @xmath363 is a sum . as a result ,
the variable @xmath365 in can not be defined and have to be replaced by @xmath366 which is of order @xmath367 .
this renders @xmath368 fast varying no matter how close @xmath212 is to @xmath369 . due to the averaging of fast oscillations
, there is no significant contribution from @xmath363 . for the primary - primary component
, the analog to reads @xmath370 \longrightarrow \frac{-i\phi_{-1}(x_{\rm s})}{2a } \sum_{j } \beta_j^3 \big [ \int_{\mathcal{a } } \phi_j(x_1 ) \frac{1}{n } \sum_{j ' } \sin ( \frac{2\pi x_1}{\lambda } \frac{j}{n } ) e^{-2\pi \sqrt{1-(\frac{j'}{n})^2 } \frac{z^{\rm s } } { \lambda } } dx_1 \big]^2.\ ] ] again , in the sum over @xmath293 , only fast variables are involved . in the regime @xmath248 ,
the contribution of the function above is negligible .
the term @xmath371 can be analyzed similarly . combining the main contributions in @xmath283 and @xmath371
, we obtain the desired result . as we will see in section [ sec : stabi ] , the imaging functionals are not statistical stable if the source is time - harmonic . hence it is required to consider a broadband source ( [ broadbandsource1 ] ) .
we show that the results obtained above for the means of imaging functionals apply to the broadband setting as well .
using proposition [ prop : momentt ] , the main contribution of the two moment of mode coupling matrix at the same frequency comes from the case when @xmath226 and @xmath227 .
therefore , @xmath372 similarly , for the primary - secondary field , we have @xmath373 for the secondary - primary field , we have @xmath374 from these limits , we see that as long as @xmath369 , the distance between the reflector and the array is much smaller than @xmath375 , the integral in @xmath190 above can be approximated by the energy of the source ( square @xmath376 norm of @xmath377 ) .
the rest parts of the limiting expectation of the cross correlation functions are exactly the same as the time - harmonic case .
consequently , the resolution analyses in the previous subsections based on the mean value of the cross - correlation migration imaging functionals remain the same .
the key tool is the following proposition which analyzes the asymptotic behavior of the fourth - order moment of the transmission coefficients in the limit @xmath66 ; see ( * ? ? ?
* section 20.9.3 ) or ( * ? ? ?
* section 8.4 ) .
[ prop : fourmoment]the expectation of four transmission coefficients at the same frequency has a limit as @xmath146 . in the regime @xmath378
we have @xmath379 \stackrel{l \gg l_{\rm equip}}{\simeq } \left\ { \begin{array}{ll } \frac{2}{n(n+1 ) } & \mbox { if } ( j , l)=(m , n)=(j',l')=(m',n ' ) \ , , \\
\frac{1}{n(n+1 ) } & \mbox { if } ( j , l)=(m , n ) \neq ( j',l')=(m',n ' ) \ , , \\
\frac{1}{n(n+1 ) } & \mbox { if } ( j , l)=(m',n ' ) \neq ( j',l')=(m , n ) \ , , \\ 0 & \mbox { otherwise}\ , .
\end{array } \right.\ ] ] let @xmath380 and @xmath381 .
the expectation of four transmission coefficients at two frequencies @xmath16 and @xmath382 has a limit as @xmath383 . in the regime @xmath378
we have @xmath384 \stackrel{l \gg l_{\rm equip}}{\simeq } \left\ { \begin{array}{ll } \frac{1}{n^2 } & \mbox { if } ( j , l)=(m , n ) \mbox { and } ( j',l')=(m',n ' ) \ , , \\ 0 & \mbox { otherwise}\ , .
\end{array } \right.\ ] ] the previous section shows that the mean of the imaging functional has a peak centered at the reflector location .
the width of the peak is of the order of the wavelength .
however , the imaging functional will give the reflector location only if it is statistically stable , that is to say , if the standard deviation of the fluctuations of the imaging functional is smaller than the mean amplitude of the peak .
we address the full aperture case in which the imaging functional is defined by ( [ eq : ikmdef ] ) . by using ( [ eq : espifap]-[eq : espifam ] )
the mean of the imaging functional is @xmath385 = \frac{\omega_0 ^ 2 { \sigma_{\rm r}}}{4 } \phi_{-1}({x_{\mathrm{s } } } ) \big\ { \re e \big ( \psi(x^{\rm s},z^{\rm s};{x_{\rm r}},{z_{\mathrm{r } } } ) ^2 \big ) + o\big ( \frac{1}{kz^{\rm s } } \big ) + o \big ( \frac{1}{n \omega_0 ^ 2 { \sigma_{\rm r } } } \big ) \big\ } , \ ] ] where @xmath251 and @xmath250 are defined by ( [ def : phij ] ) and ( [ def : psi ] ) . here : + - the term with the real part comes from the contributions of the cross correlation of secondary ( reflected ) and primary waves @xmath386 and @xmath387 that contain @xmath388 ( second term in ( [ eq : espifap ] ) ) or @xmath389 ( third term in ( [ eq : espifam ] ) ) .
+ - the term @xmath390 comes from the contributions of the cross correlation of secondary ( reflected ) and primary waves @xmath386 and @xmath387 that contain @xmath391 or @xmath392 ( third term in ( [ eq : espifap ] ) and second term in ( [ eq : espifam ] ) ) .
in such a case there are at least a product of two these terms , which gives the @xmath393 decay . + - the term @xmath394 comes the contributions of the cross correlation of primary waves @xmath395 ( the first terms in ( [ eq : espifap ] ) and in ( [ eq : espifam ] ) ) .
the expression ( [ eq : meanif ] ) is valid provided @xmath300 is large enough so that @xmath396
. then it is true that the mean imaging functional is dominated by the first term in the right hand side , which is a peak centered at the reflector location .
the mean amplitude of the peak at the reflector location is @xmath397 in the continuum approximation @xmath398 , we have ( [ asy : phi-1 ] ) and @xmath399 and therefore @xmath400 the second moment of the imaging functional can be computed using proposition [ prop : fourmoment ] in the regime @xmath401 and @xmath378 .
@xmath402 & = & \frac{n}{n+1}\big| { \mathbb{e}}\big [ \mathcal{i}_{\rm fa } ( x^{\rm s},z^{\rm s})\big ] \big|^2\\ \nonumber & & + \frac{\omega_0 ^ 4 { \sigma_{\rm r}}^2 n}{32 ( n+1)}\phi_{-1}^2({x_{\mathrm{s } } } ) \big\ { \phi_1(x^{\rm s } ) \phi_{-1}({x_{\rm r } } ) | \psi(x^{\rm s},z^{\rm s};{x_{\rm r}},{z_{\mathrm{r}}})|^2 \\ & & \quad + \re e \big ( \psi(x^{\rm s},z^{\rm s};{x_{\rm r}},{z_{\mathrm{r}}})^4 \big ) + o\big ( \frac{1}{kz^{\rm s } } \big ) + o \big ( \frac{1}{n^2 \omega_0 ^ 4 { \sigma_{\rm r}}^2 } \big ) \big\ } .
\label{eq : varth1}\end{aligned}\ ] ] here : + - the term @xmath390 comes from the contributions of the cross correlation of secondary ( reflected ) and primary waves @xmath386 and @xmath387 that contain @xmath391 or @xmath392 ( in such a case there are at least a product of two these terms , which gives the @xmath393 decay ) .
+ - the term @xmath403 comes from the contribution of the cross correlation of primary waves @xmath395 that can be computed in a more quantitative way : @xmath404 the variance of the imaging functional at the reflector location is therefore @xmath405 in the continuum approximation @xmath398 , we have ( [ asy : phi-1 ] ) , ( [ asy : phi0 ] ) , and @xmath406 and therefore @xmath407 to summarize : + 1 ) the typical amplitude of the fluctuations of the imaging functional for @xmath408 ( _ i.e. _ at the reflector location ) is @xmath409 ( as shown by ( [ eq : varth2 ] ) ) .
+ 2 ) the typical amplitude of the fluctuations of the imaging functional for @xmath410 ( _ i.e. _ away from the reflector location ) is @xmath411 ( as shown by ( [ eq : varth1 ] ) ) .
the second result shows that the fluctuations of the image far from the main peak location are of the order of @xmath412 relatively to the amplitude of the main peak .
they are due to the contributions of the primary cross correlation . provided the number of modes is large enough @xmath413 , they are small .
the first result shows that the amplitude of the peak at the reflector location has relative fluctuations of order one .
this is due to the fact that the reflector is illuminated by a field whose amplitude is randomly spatially varying , so that the reflected energy is proportional to the squared amplitude of the primary field at the reflector location , which is a random quantity .
this is the origin of the statistical instability in the time - harmonic case . + we know that the frequency coherence radius @xmath414 in a waveguide with length @xmath415 is of the order of @xmath51 ( see ( * ? ? ?
* proposition 20.7 ) or ( * ? ? ?
* proposition 6.3 ) ) . as a result , as soon as a broadband source with a bandwidth larger than @xmath414 is used , then the field is the superposition of decorrelated frequency components . as a consequence
the field is self - averaging in the time domain .
more exactly , from the expressions of the cross correlations in the broadband case given in subsection [ subsec : crossbroad ] , the mean and the variance of the imaging functional are of the form @xmath416 & = & \int dh |\hat{f}_0(h)|^2 \sum_{j , l , m , n } { \mathbb{e}}\big [ \overline{t_{jl}^{\varepsilon } } t^{\varepsilon}_{mn}(\omega_0+{\varepsilon}^\alpha h)\big ] c_{j , l , m , n } ( x^{\rm s } , z^{\rm s } ) , \\ { \rm var } \big ( \mathcal{i}_{\rm fa } ( x^{\rm s } , z^{\rm s } ) \big ) & = & \iint dh dh ' |\hat{f}_0(h)|^2 |\hat{f}_0(h')|^2\sum_{j , l , m , n , j',l',m',n ' } \\ & & \big\ { { \mathbb{e}}\big [ \overline{t_{jl}^{\varepsilon } } t^{\varepsilon}_{mn}(\omega_0+{\varepsilon}^\alpha h ) t_{j'l'}^{\varepsilon}\overline{t^{\varepsilon}_{m'n'}}(\omega_0+{\varepsilon}^\alpha h ' ) \big ] \\ & & \hspace*{-0.9 in } - { \mathbb{e}}\big [ \overline{t_{jl}^{\varepsilon } } t^{\varepsilon}_{mn}(\omega_0+{\varepsilon}^\alpha h)\big ] { \mathbb{e}}\big [ t_{j'l'}^{\varepsilon}\overline{t^{\varepsilon}_{m'n'}}(\omega_0+{\varepsilon}^\alpha h ' ) \big ] \big\ } c_{j , l , m , n } ( x^{\rm s } , z^{\rm s } ) \overline{c_{j',l',m',n ' } } ( x^{\rm s } , z^{\rm s } ) , \end{aligned}\ ] ] where @xmath417 is a shorthand for the deterministic coefficient that contains the phase and mode amplitudes .
first , since @xmath418 $ ] is independent on @xmath190 to leading order ( because @xmath419 ) , the mean satisfies @xmath420_{\rm broadband } \sim { \mathbb{e}}\big [ \mathcal{i}_{\rm fa } ( x^{\rm s } , z^{\rm s } ) \big]_{\rm narrowband } , \ ] ] as already noticed .
second , the term in the curly brackets in the expression of the variance is vanishing if @xmath421 is larger than the frequency coherence radius @xmath414 .
so the double integral in @xmath422 is reduced to a domain that has the form of a thin diagonal band , whose thickness is limited by the frequency coherence radius @xmath414 . as a result we obtain that @xmath423 \\ & & \hspace*{-0.9 in } - { \mathbb{e}}\big [ \overline{t_{jl}^{\varepsilon } } t^{\varepsilon}_{mn}(\omega_0)\big ] { \mathbb{e}}\big [ t_{j'l'}^{\varepsilon}\overline{t^{\varepsilon}_{m'n'}}(\omega_0 ) \big ] \big\ } c_{j , l , m , n } ( x^{\rm s } , z^{\rm s } ) \overline{c_{j',l',m',n ' } } ( x^{\rm s } , z^{\rm s } ) , \end{aligned}\ ] ] or more simply @xmath424 where @xmath425 is the bandwidth of the source that is larger than the frequency coherence radius @xmath426 .
we had seen that the use of broadband sources does not affect the resolution of the imaging functional but it ensures its statistical stability .
provided the bandwidth is larger than the frequency coherence radius , the typical amplitude of the fluctuations of the imaging functional is smaller than the amplitude @xmath409 of the main peak at the reflector location , and therefore the reflector can be localized .
in this paper we have shown that migration of the cross correlations of the data recorded by a passive receiver array can allow for diffraction - limited imaging of the reflector in a random waveguide even though the sources are very far from the reflector , provided the receivers are close enough from it .
the statistical stability of the imaging functional is ensured by the use of broadband sources .
the resolution properties are ensured by the waveguide geometry : even when the receiver array does not span the whole cross section of the waveguide , the width of the point spread function of the imaging functional is of the order of the wavelength , provided the diameter of the array is larger than the wavelength .
this paper has addressed the case of a two - dimensional waveguide with dirichlet boundary conditions , but the conclusions should be qualitatively the same for fairly general situations , when addressing three - dimensional waveguides , with neumann , dirichlet or mixed boundary conditions , with random fluctuations of the index of refraction or of the boundaries as in @xcite .
the authors would like to thank the anonymous referees for their careful reading of the manuscript and helpful comments .
this work was supported by erc advanced grant project multimod-267184 .
|
we consider reflector imaging in a weakly random waveguide .
we address the situation in which the source is farther from the reflector to be imaged than the energy equipartition distance , but the receiver array is closer to the reflector to be imaged than the energy equipartition distance . as a consequence ,
the reflector is illuminated by a partially coherent field and the signals recorded by the receiver array are noisy .
this paper shows that migration of the recorded signals can not give a good image , but an appropriate migration of the cross correlations of the recorded signals can give a very good image .
the resolution and stability analysis of this original functional shows that the reflector can be localized with an accuracy of the order of the wavelength even when the receiver array has small aperture , and that broadband sources are necessary to ensure statistical stability , whatever the aperture of the array .
|
let @xmath3 be a linear connection on a real - analytic manifold @xmath4 which is identified with the zero section in the tangent bundle @xmath5 . a complex structure defined on a domain @xmath6 of @xmath7 containing @xmath8
is adapted to the connection if for any @xmath9-geodesic @xmath10 its complexification @xmath11 , given by @xmath12 , is holomorphic on @xmath13 .
in this situation we refer to @xmath6 as an adapted complexification of @xmath14 . in the case where @xmath3 is real - analytic , r. bielawski ( @xcite ) and
r. szke ( @xcite ) recently showed that the adapted complex structure exists in a neighborhood of @xmath15 . in @xcite
one also finds a uniqueness result . for the levi civita connection of a real - analytic riemannian manifolds , such results were known since the pioneering works of guillemin - stenzel ( @xcite ) and lempert - szke ( @xcite ) . in the presence of a large enough " lie group acting on @xmath8 and preserving the geodesic flow induced by @xmath9 , one can prove that there exists a maximal domain @xmath6 for the adapted complex structure , i.e.@xmath16every adapted complexification is necessarily contained in @xmath6 ( see@[email protected] , cf .
if @xmath8 is a non - compact , riemannian symmetric space such a maximal complexification is well - known under the name of akhiezer - gindikin domain ( see @xcite , cf .
@xcite ) .
recall that @xmath8 is the fixed point set of the anti - holomorphic involution on @xmath17 given by @xmath18 and in the case of the levi civita connection associated to a pseudo - riemannian manifold the metric @xmath19 appears as the restriction of a pseudo - khler metric @xmath20 with the same index as @xmath21 .
moreover @xmath20 admits a global potential whose properties give important geometric insights of @xmath6 ( [ sz3 ] , cf .
@xcite , @xcite , @xcite , @xcite , @xcite ) .
for a connected , non - compact , real semisimple lie group @xmath22 , let @xmath23 be the cartan decomposition of its lie algebra @xmath24 with respect to a maximal compact subalgebra @xmath25 .
denote by @xmath26 the killing form of @xmath0 and consider the distinguished one - parameter family of left - invariant metrics @xmath27 ( degenerate for @xmath28 ) uniquely defined by @xmath29 for any @xmath30 and @xmath31 in @xmath32 .
note that for all real @xmath33 , the action of the group @xmath34 , given by left and right multiplication on @xmath22 , is isometric . here
@xmath35 is the connected subgroup of @xmath0 generated by @xmath25 .
recall that for @xmath36 these metrics appear in the classification of all naturally reductive , left - invariant riemannian metrics on @xmath0 given by c. gordon in @xcite .
our main goal is to present new examples of maximal complexifications adapted , in the non - degenerate cases , to the levi civita connection associated to a metric of the above family . in the degenerate case one
finds a maximal complexification adapted to the unique real - analytic linear connection which is obtained as the limit of such levi civita connections . for @xmath37
we give a precise description of these complexifications and we determine their basic complex - geometric properties . for positive @xmath38
this gives , along with previous results ( see @xcite , @xcite , @xcite ) , examples among all classes of @xmath39-dimensional , naturally reductive , riemannian homogeneous spaces ( cf .
@xcite ) . for @xmath40
one obtains the symmetric pseudo - riemannian case which has been investigated , among others , by g. fels ( @xcite ) and r. bremigan ( @xcite ) .
the paper is organized as follows .
basic results and properties of the above metrics @xmath27 are recalled in section 2 . there
we also point out with an example that in order to perform ( pseudo ) khlerian reduction in this pseudo - riemannian context , one may need more conditions than those necessary in the riemannian case ( rem .
[ aguilar ] , cf .
@xcite ) . in section 3
we give a version of the characterization of maximal adapted complexifications given in @xcite which is suitable in our situation ( prop.@xmath16[polarmap ] ) .
in particular we show the existence of the maximal complexification @xmath41 adapted to @xmath42 .
this is realized as an @xmath43-equivariant riemann domain over the universal complexification @xmath44 of @xmath22 , with _ polar map _ @xmath45 . by considering the usual identification @xmath46
, such complexification can be described via a slice for the induced @xmath43-action by @xmath47 where @xmath48 is a semi - analytic subset of the product of @xmath49 and the closure of a weyl chamber in a maximal abelian subalgebra @xmath50 of @xmath51 .
that is , @xmath52 if and only if @xmath53 belongs to the intersection of sublevel sets of certain real - analytic functions of @xmath54 ( prop .
[ slice ] and [ domain ] ) .
the case @xmath1 is carried out in detail and the defining functions for @xmath55 as well as the polar map @xmath56 are explicitly determined in terms of a fixed basis of @xmath54 in sections 4 and 6 . for this
it is useful to have concrete realizations of slices and quotients of @xmath57 with respect to the involved actions , i.e. those of @xmath58 , @xmath59 and @xmath60 .
this is separately discussed in section 5 .
finally , in sections 7 and 8 we single out the following different situations , whose boundary cases are given by the symmetric pseudo - riemannian @xmath40 and the degenerate @xmath61 .
for @xmath62 all maximal adapted complexifications are biholomorphic via @xmath56 to a non - stein @xmath43-invariant domain .
namely @xmath57 without a single @xmath60-orbit ( thm.@xmath16[domain<-1 ] ) .
for @xmath63 the polar map @xmath56 remains injective but its non - stein image misses more and more @xmath60-orbits . if @xmath64 , the maximal adapted complexifications turn out to be neither holomorphically convex , nor holomorphically separable ( thm.@xmath16[domain>-1 ] ) . in all cases the envelope of holomorphy of @xmath65 is shown to be biholomorphic to @xmath57 ( cf .
@xcite , sect .
@xmath169 ) .
note that in the riemannian context @xmath36 all metrics @xmath27 have mixed sign sectional curvature .
a similar situation can be noticed in the examples discussed in @xcite . in these examples
certain left - invariant riemannian metrics on the generalized heisenberg group were considered .
their sectional curvature has mixed sign and the associated maximal adapted complexifications have similar complex - geometric properties .
it would be interesting to know if this is only a coincidence .
we wish to thank rbert szke for sharing with us a preliminary version of @xcite and the referee for his helpful suggestions .
let @xmath0 be a non - compact , semisimple lie group . here
we recall basic properties of the one - parameter family of left - invariant metrics on @xmath0 which will be considered in the sequel .
such a family contains degenerate , riemannian and pseudo - riemannian , naturally reductive metrics .
the riemannian ones appear in the classification given by c. gordon in @xcite .
more details and curvature computations can be found in @xcite .
we also recall those facts on adapted complex structures which are needed in the present paper .
finally we point out an example showing that the reduction procedure indicated by r. aguilar ( @xcite ) in the riemannian context does not apply automatically when dealing with pseudo - riemannian geometry .
[ natred ] ( cf .
@xcite ) a pseudo - riemannian metric @xmath19 on a homogeneous manifold @xmath8 is naturally reductive if there exist a connected lie subgroup @xmath66 of @xmath67 acting transitively on @xmath8 and a decomposition @xmath68 of @xmath69 , where @xmath70 is the lie algebra of the isotropy group @xmath71 at some point of @xmath15 , such that @xmath72 and @xmath73_\mathfrak m , \,z ) = \tilde \nu(x,\ , [ y,\,z]_\mathfrak m)\ ] ] for all @xmath74 . here
@xmath75_\mathfrak{m}\,$ ] denotes the @xmath76-component of @xmath75\,$ ] and @xmath77 is the pull - back of @xmath19 to @xmath78 via the natural projection @xmath79 . in this
setting we refer to @xmath80 as a naturally reductive decomposition and to @xmath81 as a naturally reductive realization of @xmath15 . for a naturally reductive realization @xmath81 every geodesic through the base point @xmath82 is the orbit of a one - parameter subgroup of @xmath66 generated by some @xmath83 ( see @xcite , p. 313 ) .
in fact for a riemannian homogeneous manifold @xmath81 with an @xmath84-invariant decomposition @xmath80 this property implies that @xmath81 is a naturally reductive realization ( see , e.g. @xcite ) .
let @xmath0 be a connected , non - compact , semisimple lie group and let @xmath85 be the cartan decomposition of its lie algebra with respect to a maximal compact lie subalgebra @xmath25 .
let @xmath26 denote the killing form on @xmath24 and , for every real @xmath33 , assign a left - invariant metric @xmath27 on @xmath0 by defining its restriction on @xmath86 as follows : @xmath87 for any @xmath30 and @xmath88 in @xmath32 .
since @xmath26 is negative definite on @xmath49 and positive definite on @xmath51 , these metrics are riemannian , degenerate or pseudo - riemannian when @xmath89 , @xmath28 or @xmath90 , respectively .
let @xmath35 be the connected subgroup of @xmath0 generated by @xmath49 ( which is compact if @xmath0 is a finite covering of a real form of a complex semisimple lie group ) .
since @xmath25 , @xmath91 and @xmath26 are @xmath92-invariant , @xmath27 is right @xmath93-invariant , i.e. the action of @xmath94 on @xmath0 defined by @xmath95 is by isometries . here
we allow discrete ineffectivity given by the diagonal in @xmath96 , where @xmath97 is the center of @xmath22 .
one has @xmath98 with @xmath71 the diagonal in @xmath99 .
note that a different choice of a maximal compact connected subalgebra @xmath100 induces an _ equivalent _ left - invariant riemannian structure , i.e. there exists an isometric isomorphism @xmath101 this is given by the internal conjugation transforming @xmath49 in @xmath102 .
we summarize the main properties of the above metrics in the following proposition where the degenerate metric @xmath103 can be regarded as a limit case of non - degenerate ones .
[ semisimple ] ( @xcite , sect.3 , cf .
@xcite , proof of thm.5.2 ) @xmath104let @xmath0 be a non - compact , semisimple lie group and , for @xmath105 , let @xmath27 be the above defined left - invariant metric
. then 1 .
the action of @xmath106 by left and right multiplication is by isometries and 2 .
the direct sum @xmath107 , with @xmath108 the isotropy lie algebra and @xmath109 is a naturally reductive decomposition of @xmath110 .
in particular for every @xmath111 in @xmath112 the unique geodesic through @xmath113 and tangent to @xmath53 is given by @xmath114 , @xmath115 from @xcite , section@xmath165 , it follows that in the riemannian cases @xmath36 the connected component of the isometry group is essentially given by @xmath116 ( here discrete ineffectivity is allowed ) while in the pseudo - riemannian symmetric case @xmath40 it coincides with @xmath117 . further information regarding the levi civita connections ( or their limit when @xmath38 vanishes ) , the curvature tensor , the scalar and ricci curvature can be found in @xcite , section 3 .
let @xmath8 be a complete real - analytic riemannian manifold .
following the results of guillemin - stenzel ( @xcite ) and lempert - szke ( @xcite ) one can introduce a complex structure on a subdomain of the tangent bundle @xmath7 which is _ canonically _ adapted to the given riemannian structure .
recently r. bielawski ( @xcite ) and r. szke ( @xcite ) have pointed out that for the existence of such a complex structure it is enough to have a real - analytic linear connection @xmath9 .
a real - analytic complex structure on a domain @xmath6 of @xmath7 is adapted to @xmath3 if all leaves of the induced foliation are complex submanifolds with their natural complex structure .
that is , for any @xmath9-geodesic @xmath118 the induced map @xmath119 defined by @xmath120 is holomorphic on @xmath121 with respect to the adapted complex structure . here
@xmath122 is the scalar multiplication in the vector space @xmath123 .
the adapted complex structure exists and it is unique on a sufficiently small neighborhood of @xmath15 , which is identified with the zero section in its tangent bundle @xmath5 . if @xmath6 is a domain of @xmath7 containing @xmath8 on which this structure is defined , then we refer to it as an _ adapted complexification_. associated to every non - degenerate metric @xmath27 of the above introduced one - parameter family one has the levi civita connection . for @xmath124 and @xmath125 in @xmath24
this is given by the formula ( cf .
@xcite ) @xmath126 note that this uniquely defines a left - invariant , real - analytic , linear connection also in the degenerate case @xmath61 .
therefore for all real @xmath38 one has an adapted complex structure at least in a neighborhood of @xmath8 in @xmath127 . in the next section
we will see that there exists an adapted complexification which is maximal in the sense of containing any other adapted complexification .
[ aguilar ] @xmath16 let @xmath8 be a real - analytic , riemannian manifold with a free action by isometries of a compact lie group @xmath35 and endow @xmath128 with the unique riemannian metric such that the natural projection @xmath129 becomes a riemannian submersion .
then , as a consequence of results in @xcite , if the adapted complex structure exists on all of @xmath5 , so it does on @xmath130 . in the pseudo - riemannian context
the analogous result does not hold in this generality .
for instance , let @xmath131 be a compact , semisimple lie group and denote by @xmath132 its universal complexification .
let @xmath0 be a non - compact , real form of @xmath132 and @xmath133 .
denote by @xmath49 and @xmath134 the lie algebras of @xmath35 and @xmath135 , respectively , and by @xmath26 the killing form on @xmath135 .
consider the unique left - invariant riemannian metric @xmath19 on @xmath131 defined for all @xmath136 by @xmath137 where @xmath138 .
endow @xmath139 with the unique bi - invariant pseudo - metric @xmath77 such that @xmath140 for all @xmath141 in @xmath142 .
then the projection @xmath143 turns out to be a pseudo - riemannian submersion .
moreover the adapted complex structure is defined on all of @xmath144 .
indeed @xmath145 is essentially ( up to the sign of the metric in the second component ) the product of two symmetric riemannian spaces of the compact type , thus this is a consequence of results in @xcite and @xcite .
however @xmath146 has some negative sectional curvatures ( see @xcite , cf .
@xcite , sect.3 ) , thus by theorem 2.4 in @xcite the adapted complex structure is not defined on all of @xmath147 .
let @xmath0 be a connected , non - compact , semisimple lie group and consider the one - parameter family of left - invariant metrics ( pseudo - riemannian for @xmath90 , degenerate for @xmath61 ) introduced in section 2 and defined by @xmath148 then ( cf . prop.@xmath16[semisimple ] ) the action of @xmath149 by left and right multiplication is by isometries , the isotropy in @xmath113 is @xmath150 and the quotient @xmath81 is a natural reductive realization of @xmath151 .
the riemannian exponential map @xmath152 in @xmath113 is given by @xmath153 @xmath154 for every @xmath155 .
note that the @xmath43-action on @xmath0 induces an action on the tangent space @xmath156 just by differentiation .
if one identifies @xmath156 with @xmath157 as usual , this action reads as @xmath158 the group @xmath66 also acts on the universal complexification @xmath44 of @xmath0 by left and right multiplication .
next we show the existence of a maximal complexification in @xmath156 adapted to the connection associated to @xmath27 for all real @xmath33 .
it can be characterized as follows .
[ polarmap ] let @xmath0 be a connected , non - compact , semisimple lie group endowed with a metric @xmath27 of the above family .
then there exists a maximal complexification @xmath41 adapted to the connection associated to @xmath42 .
let @xmath159 be the @xmath43-equivariant map defined by @xmath160 then @xmath41 is given by the connected component of @xmath161 containing @xmath162 .
the polar map @xmath163 is locally biholomorphic .
note that the universal complexifications of @xmath66 and @xmath71 are @xmath164 and @xmath165 , respectively .
let @xmath66 act on @xmath166 by left multiplication and consider the @xmath43-equivariant map @xmath167 defined , for @xmath168 and @xmath169 , by @xmath170 identify @xmath171 with the zero section in @xmath172 .
then an analogous argument as in @xcite,@[email protected] , applies to show that the connected component @xmath173 of @xmath174 which contains @xmath0 is the maximal adapted complexification and the restriction @xmath175 is locally biholomorphic .
since @xmath176 , the statement is a consequence of the following real - analytic , @xmath43-equivariant identification @xmath177 in order to describe @xmath41 it is convenient to determine a slice for the @xmath43-action .
let @xmath178 be the closure of a weyl chamber in a maximal abelian subalgebra @xmath50 of @xmath91 and define @xmath179 since @xmath180 is a fundamental domain for the @xmath181 action on @xmath51 , every @xmath43-orbit of @xmath182 meets @xmath183
. then one has [ slice ] denote by @xmath184 the connected component of @xmath185 in the subset @xmath186 of @xmath183
. then @xmath187 .
note that the @xmath43-equivariance of @xmath56 induces that of @xmath188 , i.e. @xmath189 for all @xmath168 and @xmath190 .
in particular @xmath191 has maximal rank if and only if so does @xmath192 , implying the statement .
let @xmath37 . here we choose @xmath193 and give an explicit description of @xmath184 in terms of fixed basis of @xmath194 . by proposition [ slice ] this determines the maximal adapted complexification @xmath41 associated to the connection of @xmath42 .
identify @xmath195 with the set of zero trace matrices and let @xmath196 then @xmath197 is a basis of @xmath25 , while @xmath198 gives a basis of @xmath51 . consider the polar map @xmath199 introduced in section 3 and choose @xmath200 by fixing @xmath201 .
given @xmath202 consider the natural identification @xmath203 and note that for @xmath204 one has @xmath205 where @xmath206 and @xmath207 denote left and right multiplication by @xmath208 , respectively , and @xmath209 is the exponential map of @xmath210 .
recall that by identifying @xmath211 with @xmath212 via left multiplication one has @xmath213 for all @xmath214 ( see , e.g. @xcite ) .
fix @xmath215 in @xmath216 . by using the above formulae
, one shows that the basis @xmath217 of @xmath218 is mapped by the differential @xmath219 into the image via @xmath220 of the following six vectors @xmath221 @xmath222 @xmath223 now note that @xmath224 and @xmath225 are even holomorphic , thus the functions @xmath226 and @xmath227 are well defined and holomorphic .
moreover the cofficients of their taylor series around the origin are real .
then , by restriction one obtains two real - analytic functions which will be denoted in the sequel by @xmath228 and @xmath229 , respectively . letting @xmath230 ,
a further computation ( see appendix ) shows that the above six vectors can be written as @xmath231 @xmath232 @xmath233 @xmath234 @xmath235 @xmath236 let us point out without proof some properties of the functions @xmath237 , @xmath238 , and @xmath239 , @xmath240 , which are used in the sequel . [ functions ] @xmath16for @xmath241 the real - analytic functions @xmath242 are strictly positive and @xmath243 are strictly increasing .
moreover @xmath244 we can now determine the maximal adapted complexification @xmath41 by computing the slice @xmath184 introduced in proposition [ slice ] .
[ domain ] the slice @xmath184 for the maximal adapted complexification consists of the elements @xmath245 in @xmath200 such that * @xmath246 * @xmath247 where the function @xmath248 is defined by @xmath249 .
let @xmath250 and @xmath251 . by the above computations @xmath252 is non - singular if and only if the two determinants @xmath253 and @xmath254 do not vanish .
a straightforward computation yields the two conditions above . note that in the pseudo - riemannian symmetric case @xmath255 , conditions @xmath256 and @xmath257 coincide and yield the set @xmath258 .
@xmath104 for @xmath259 let @xmath260 be the subdomain of @xmath200 defined by condition @xmath256 in proposition [ domain ] , i.e. @xmath261 [ thesame ] ( see picture 1 ) @xmath104for @xmath262 one has @xmath263 .
indeed by lemma [ functions ] one has @xmath264 , for @xmath265 . consequently for any @xmath266 @xmath267 thus condition @xmath257 is automatically fulfilled .
for @xmath268 the closure of @xmath184 in @xmath200 is contained in @xmath269 . indeed , for @xmath270
, one has @xmath271 therefore condition @xmath257 is not fulfilled .
in particular the boundary of @xmath272 in @xmath200 is defined by @xmath273 picture 1 @xmath104 @xmath274 = 0.34 = 0.225 = 0.3 @xmath275 boundary of @xmath276 - - - @xmath16boundary of @xmath277 @xmath104 for @xmath278 one checks that the vector fields tangential to @xmath279 , i.e. @xmath280 and @xmath281 , remain linearly independent on the boundary of @xmath282 .
thus the boundary of the maximal adapted complexification can be characterized by saying that @xmath43-orbits ( in fact @xmath283-orbits ) become tangential to @xmath284 .
on the other hand for @xmath285 it is the dimension of @xmath43-orbits at the boundary of @xmath286 to drop from 4 to 3 ( see @xmath287 of prop .
[ properties ] below ) , giving again a characterization of the maximal adapted complexification . before studying the restriction of @xmath56 to @xmath184 it will be useful to obtain a concrete realization of quotients of @xmath44 with respect to those actions which are involved .
here @xmath0 and @xmath35 denote @xmath289 and @xmath290 , respectively .
let @xmath291 act on @xmath44 by left and right multiplication .
note that since @xmath35 is compact , this action is proper .
the main goal of this section is to present models for the quotients @xmath292 , @xmath293 @xmath294 and for the relative quotient maps .
first consider the map @xmath295 where @xmath296 , @xmath297 , is the conjugation in @xmath210 , i.e. the unique antiholomorphic involutive automorphism of @xmath44 whose fixed point set is @xmath22 .
let @xmath0 act on @xmath44 by left multiplication and note that every fiber of this map consist of a single @xmath22-orbit .
thus @xmath298 is set theoretically equivalent to @xmath299 and @xmath300 is a realization of the quotient map .
moreover , a simple computation shows that ( cf .
, e.g. @xcite ) @xmath301 it is convenient to consider the automorphism @xmath302 transforming @xmath303 onto @xmath304 .
this is induced by the unique complex lie algebra morphism of @xmath212 mapping the basis @xmath305 ( cf . beginning of sect .
4 ) into @xmath306 . recall that the involution of @xmath44 defining @xmath307 is given by @xmath308 , where @xmath309 since the elements of the above basis are fixed by the lie algebra automorphisms induced by @xmath310 and @xmath311 respectively , it follows that @xmath312 note that @xmath313 can be identified with @xmath314 .
then @xmath315 @xmath316 gives a model of the quotient @xmath292 .
let us describe how the right @xmath317-action on @xmath44 is transformed after applying @xmath318 .
for @xmath319 and @xmath320 one has @xmath321 @xmath322 and by ( [ equi ] ) this gives @xmath323 @xmath324 thus @xmath325 is @xmath317-equivariant , if one let @xmath326 act on @xmath44 by right multiplication and on @xmath327 by @xmath328 @xmath329 for every @xmath330 .
in particular , after applying @xmath331 , the right @xmath93-action on @xmath332 reads as rotations on @xmath333 .
let @xmath334 and define @xmath335 by @xmath336 for every @xmath337 the inverse image @xmath338 consists of a single @xmath93-orbit given by @xmath339 in fact @xmath340 is a realization of the quotient @xmath341 . recall that @xmath342 act on @xmath44 by left and right multiplication .
let the one - parameter subgroup of @xmath343 defined by @xmath344 act on @xmath340 by @xmath345 , for all @xmath346 and @xmath347 .
one has [ quotient1 ] let @xmath326 act on @xmath327 by @xmath348 for @xmath349 .
then * a model for the quotient @xmath350 is @xmath327 with @xmath317-equivariant quotient map @xmath325 .
the fixed point set of the @xmath93-action on @xmath327 is given by @xmath351 , * a model for @xmath352 is given by @xmath340 with quotient map @xmath335 defined by @xmath353 , * a model for the quotient @xmath354 is @xmath340 with @xmath355-equivariant quotient map @xmath356 .
in particular @xmath357 .
@xmath104 [ simmetries ] let @xmath358 denote the normalizer of @xmath326 in @xmath210 .
then @xmath113 and @xmath359 represent the two elements of the quotient @xmath360 .
a simple computation shows that @xmath361 is equivariant with respect to right ( or left ) multiplication by @xmath362 in @xmath44 and reflection with respect to the origin in @xmath363 .
furthermore @xmath361 is equivariant with respect to conjugation by @xmath362 in @xmath44 and reflection with respect to the line of equation @xmath364 in @xmath363 .
picture 2 = 0.47 @xmath365 - @xmath16- @xmath16- @xmath366 @xmath367 r - orbits @xmath104 [ sliceforp]@xmath104 the action of @xmath368 on @xmath340 is not proper .
set theoretically @xmath369 can be identified with the slice for the @xmath355-action on @xmath340 defined by the union @xmath370 , where @xmath371 , @xmath372 , @xmath373 and @xmath374 correspond to the non - closed @xmath355-orbits in @xmath363 .
the closure of these orbits is obtained by adding the unique fixed point @xmath375 .
however for our purposes it is more convenient to consider the following slice ( cf .
picture 2 ) @xmath376 here and in the sequel a slice is assumed to intersect every orbit in a single point .
let @xmath377 be the slice in @xmath378 introduced in section 3 .
consider the subdomain @xmath379 and denote by @xmath380 its closure in @xmath183 .
one has [ properties ] @xmath16let @xmath381 act on @xmath332 by left and right multiplication . * a slice for the @xmath43-action on @xmath44
is given by @xmath382 + all @xmath43-orbits are closed ( the action is proper ) , the union of all @xmath39-dimensional @xmath43-orbits is given by @xmath383 .
all other orbits are @xmath384-dimensional with discrete isotropy given by the ineffectivity @xmath385 of @xmath43 . * a slice for the @xmath386-action
is given by @xmath387 @xmath388 + where @xmath389 , @xmath390 and @xmath391 .
the only non - closed @xmath386-orbits are those through @xmath392 , with @xmath393 and @xmath394 .
their closure is obtained by adding the orbit through @xmath395 .
the only @xmath384-dimensional orbits are those through @xmath396 and @xmath395 , all other orbits have maximal dimension . for all @xmath245 in @xmath380
one has @xmath397
@xmath398 where @xmath399 .
fix @xmath400 and note that the set @xmath401 is given by @xmath402 then , from the above formula it follows that @xmath403 consists of the intersection of @xmath340 with the line of equation @xmath404 . as a consequence @xmath405 is mapped bijectively onto @xmath406 .
this and remark [ simmetries ] imply that @xmath361 maps @xmath407 bijectively onto @xmath363 . thus @xmath407 is a slice for the @xmath43-action on @xmath210 . since @xmath408 , by @xmath287 of lemma [ quotient1 ]
the only 3-dimensional @xmath43-orbits are those through @xmath409 and all others are 4-dimensional . moreover the isotropy of the @xmath93-action on @xmath410 is given by the ineffectivity @xmath411 , implying that last claim in @xmath412 . for @xmath413 note that @xmath414 ( cf . the above formula ) .
this and remark [ simmetries ] apply to show that the restriction @xmath415 is bijective , where @xmath416 is the @xmath355-slice in @xmath340 introduced in remark [ sliceforp ] .
one has a commutative diagram of canonical quotients @xmath417 where , by @xmath418 of lemma [ quotient1 ] , the map @xmath361 is @xmath355-equivariant .
this gives a one to one correspondence between the @xmath386-orbit space of @xmath44 and the @xmath355-orbit space of @xmath363 . as
a consequence @xmath419 is a slice for the @xmath386-action on @xmath210 .
moreover the @xmath386-orbits through @xmath420 correspond to the five @xmath355-orbits in @xmath421 , implying the topological claim .
finally , the dimension of every @xmath422 orbit can be obtained by adding to the dimension of the corresponding @xmath355-orbit the dimension of the @xmath43-orbit given in @xmath412 .
since every @xmath355-orbit different from the fixed point @xmath423 is one - dimensional , this concludes the proof .
regard @xmath424 as a riemannian symmetric space of rank one .
its maximal adapted complexification can be realized in @xmath425 and it is usually described as @xmath426 ( see @xcite , @xcite ) .
its boundary is given by @xmath427 and we refer to @xmath428 as its singular boundary . for later use
we note the following direct consequence of @xcite , theorem 6.1 and example 6.3 .
[ envelope ] for @xmath37 , let @xmath6 be a @xmath22-invariant domain of @xmath425 which contains @xmath429 and its non - singular boundary , i.e. @xmath6 contains the invariant subset @xmath430 .
then @xmath431 is holomorphically convex if and only if it coincides with @xmath432 .
let @xmath1 be endowed with one of the metrics @xmath42 . by proposition [ slice ]
the associated maximal adapted complexification is given by @xmath187 , where @xmath184 consists of the elements of @xmath41 which are in the slice @xmath200 for the @xmath43-action on @xmath433 .
note that every @xmath43-orbit in @xmath157 intersects @xmath200 in a single element , thus @xmath434 .
one has a commutative diagram @xmath435 where @xmath436 is the realization of the quotient map given in lemma [ quotient1 ] and @xmath437 .
here we show that the polar map @xmath163 is injective if and only if so is the restriction @xmath438 .
we also point out certain properties of the map @xmath439 which are used in the remaining sections in order to discuss injectivity of @xmath440 .
it will turn out that @xmath41 is biholomorphic to an @xmath43-invariant domain of @xmath44 if @xmath441 , while it is a non - holomorphically separable riemann domain over @xmath44 if @xmath89 . in both cases @xmath41 is not holomorphically convex and its envelope of holomorphy is biholomorphic to @xmath210 .
we first compute the two components of @xmath439 with respect to the basis @xmath442 .
one has @xmath443 @xmath444 @xmath445 @xmath446 where @xmath447 .
this gives @xmath448 [ rank]@xmath104 the map @xmath439 has maximal rank on @xmath449 .
indeed it is easy to check that @xmath361 has maximal rank on @xmath450 , therefore so does @xmath439 on the set @xmath451 . from formula ( [ reducedmap ] ) it follows that this set coincides with @xmath452 .
recall that for all real @xmath38 the slice @xmath184 is contained in the domain @xmath453 ( cf .
[ thesame ] ) . for such bigger domain one has [ diffeo ] the restriction of @xmath56 to any @xmath43-orbit of @xmath454 is a diffeomorphism onto an @xmath43-orbit of @xmath210 .
note that the isotropy of @xmath66 at a point @xmath455 of @xmath456 is given by @xmath457 if @xmath458 , it consists of the ineffectivity @xmath385 otherwise .
moreover since @xmath35 and @xmath459 commute , the identity @xmath460 holds true if and only if @xmath461 now , a similar computation as in formula ( [ reducedmap ] ) yields @xmath462 which belongs to @xmath463 if and only if @xmath464 .
then , as a consequence of @xmath287 of proposition [ properties ] , the isotropy at @xmath465 is given by @xmath457 if @xmath466 or @xmath467 otherwise , which proves the statement .
@xmath104 for @xmath468 one has @xmath469 .
thus , by formula ( [ reducedmap ] ) @xmath470 then @xmath287 of proposition [ properties ] implies that the dimension of the @xmath43-orbit through @xmath471 is only three .
in particular an analogous statement as in the above lemma does not hold on domains larger than @xmath269 .
since @xmath472 and @xmath473 is contained in @xmath269 , from the above lemma it follows that @xmath163 is injective if and only if different @xmath43-orbits in @xmath41 are mapped by @xmath56 to different @xmath43-orbits in @xmath210 . recalling that for @xmath1 every orbit intersects @xmath184 in a single point and that @xmath474 is a realization of the quotient map with respect to the @xmath43-action on @xmath210 , one has [ iniini ] the polar map @xmath163 is injective if and only if @xmath475 is injective .
we conclude this section with a technical result which will be repeatedly used in the sequel .
consider the two involutions @xmath476 and @xmath477 and denote by @xmath478 and @xmath479 the associated fixed point sets .
note that @xmath480 is equivariant with respect to these involutions . as a consequence @xmath481
is contained in @xmath482 .
also consider the @xmath483-invariant map @xmath484 defined by @xmath485 with @xmath486 .
[ fixpoints]@xmath104 let @xmath487 and @xmath488 be the two involutions defined above .
* for @xmath489 one has @xmath490 * for @xmath491 one has ( cf .
picture 3 ) + @xmath492 + where @xmath493 is the real - analytic map implicitly defined by @xmath494 .
picture 3 @xmath495 = 0.47 = 0.47 @xmath496 - @xmath497 - @xmath498 - @xmath16- @xmath16- @xmath499 @xmath367 @xmath500 for @xmath501 @xmath104 formula ( [ reducedmap ] ) implies that the set @xmath502 is given by @xmath503 @xmath504 then the cases @xmath40 and @xmath28 are straightforward . for @xmath505
this set can be written as @xmath506 , with @xmath507 as in the statement .
for @xmath508 let @xmath509 and note that @xmath510 recalling that @xmath511 is strictly increasing this yields @xmath512 @xmath513 for any @xmath514 , implying that @xmath494 has no solution for @xmath515 .
along with the @xmath483-invariance of @xmath516 , this implies @xmath412 . for @xmath517 , @xmath518 and @xmath509
fixed , an analogous argument shows that @xmath519 since @xmath520 is strictly increasing for @xmath241 and @xmath521 for @xmath522 , it follows that there exists a unique @xmath523 with @xmath524 such that @xmath525 .
this and @xmath483-invariance of @xmath516 yield @xmath526 .
let @xmath1 be endowed with one of the metrics @xmath27 for some @xmath527 . in this case we show that the polar map @xmath56 is a biholomorphism from the maximal adapted complexification @xmath41 onto an @xmath43-invariant domain of @xmath210 .
this domain is given by removing from @xmath44 a single @xmath384-dimensional @xmath386-orbit .
its envelope of holomorphy turns out to be biholomorphic to @xmath210 .
consider the one parameter subgroup @xmath528 of @xmath529 .
the @xmath530-action on @xmath41 induces a local @xmath355-action whose infinitesimal generator is given , for all @xmath531 in @xmath532 , by @xmath533 here @xmath534 denotes the adapted complex structure of @xmath41 since @xmath56 is holomorphic and @xmath535-equivariant , it is also locally equivariant with respect to such local @xmath355-action on @xmath41 and the global @xmath355-action on @xmath210 .
furthermore both ( local ) actions commute with the @xmath43-actions on @xmath41 and on @xmath210 , thus they push down to ( local ) @xmath355-actions on @xmath536 and on @xmath537 , respectively .
since one has the commutative diagram @xmath538 the restriction @xmath539 is locally @xmath355-equivariant .
recall that by @xmath418 of lemma [ quotient1 ] , the @xmath355-action on @xmath340 is explicitly given by @xmath540 .
in particular the function @xmath541 , defined by @xmath542 , is @xmath355-invariant .
this is used to determine local @xmath355-orbits in @xmath184 as follows . for
@xmath543 let @xmath544 and denote by @xmath545 the set @xmath546 .
one has [ levelset ] let @xmath547 .
then * a local @xmath355-orbit of @xmath545 coincides with a connected component of @xmath548 , * the intersection @xmath549 has two connected components if @xmath550 , it is connected if @xmath551 , * @xmath439 maps different local @xmath355-orbits of @xmath545 to different @xmath355-orbits of @xmath363 , * @xmath439 is injective on @xmath552 , * @xmath439 maps different local @xmath355-orbits of @xmath184 to different @xmath355-orbits of @xmath363 , * local @xmath355-orbits closed to @xmath553 are mapped bijectively by @xmath439 to @xmath355-orbits of @xmath363 , i.e. @xmath368 acts globally in a @xmath355-invariant neighborhood of @xmath554 .
@xmath104 note that @xmath423 is the unique fixed point for the @xmath355-action on @xmath555 and that @xmath556 does not belong to @xmath557 .
furthermore , the restriction @xmath558 is locally diffeomorphic by remark [ rank ] and locally @xmath355-equivariant by construction .
this implies that every local @xmath355-orbit of @xmath545 is one - dimensional .
one checks that @xmath559 , where @xmath560 .
moreover @xmath561 on @xmath562 , therefore @xmath563 is a one - dimensional manifold . by construction
, @xmath564 is locally @xmath355-invariant , hence local @xmath355-orbits of @xmath545 are open and closed in @xmath565 , implying @xmath412 . for @xmath526 , recall that @xmath566 by remark [ thesame ] . as a consequence @xmath567 vanishes on the boundary of @xmath552 .
also note that @xmath568 and that for @xmath569 fixed and @xmath570 such that @xmath571 , the map @xmath572 , @xmath573 is strictly increasing ( cf .
lemma [ functions ] ) .
this implies that @xmath563 is the ( connected ) graph of a function defined on @xmath574 for @xmath575 , while , for @xmath576 , it consists of two connected components ( which are contained in @xmath577 ) . for @xmath418
first consider the case of @xmath578 .
one needs to show that the two components of @xmath579 are mapped to different components of the hyperbola in @xmath580 defined by @xmath581 . for this
it is enough to note that this holds for the two limit points given by @xmath582 .
the only other non - connected case is @xmath583 , when the two components of @xmath584 have the same limit point @xmath585 by remark [ rank ] the map @xmath439 is a diffeomorphism in a neighborhood ( in @xmath552 ) of this point , thus these two components are mapped to different components of @xmath586 , as claimed .
the case @xmath587 is straightforward .
since all @xmath355-orbits in @xmath340 have connected isotropy and @xmath558 is a local diffeomorphism , @xmath439 is necessarily injective on every local @xmath355-orbit of @xmath552
. therefore @xmath418 implies that @xmath558 is injective .
moreover , from @xmath287 of lemma [ fixpoints ] it follows that @xmath545 is mapped to one of the two connected components of @xmath588 . by @xmath483-@xmath589-equivariance of @xmath439
this implies that @xmath590 is injectively mapped to the other connected component .
finally it is easy to check that @xmath591 is injectively mapped into @xmath592 , implying @xmath593 .
now note that a local @xmath355-orbit of @xmath184 either meets @xmath594 in a unique point or is contained in @xmath595 . as noticed , @xmath439 maps different elements of @xmath594 into different @xmath355-orbits of @xmath596 .
then @xmath418 and @xmath483-@xmath589-equivariance of @xmath597 , imply @xmath598 . for @xmath599
, one can move towards infinity ( topologically ) along local @xmath355-orbits of @xmath184 which are closed to @xmath554 , apply @xmath439 and check that one is moving towards infinity along @xmath355-orbits in @xmath596 . recalling that @xmath600 is the second component of @xmath597 , this follows by showing that for @xmath601 small enough and @xmath602 , with @xmath603 , such that @xmath604 , one has @xmath605 the details of this computation are omitted .
[ domain<-1 ] let @xmath1 endowed with a metric @xmath42 , with @xmath547 . then the polar map @xmath606 is injective and consequently @xmath173 is @xmath43-equivariantly biholomorphic to @xmath607 .
this domain is not holomorphically convex and its envelope of holomorphy is biholomorphic to @xmath210 .
@xmath16injectivity follows from @xmath608 of lemma [ levelset ] and proposition [ iniini ] .
assume by contradiction that @xmath173 is holomorphically convex , i.e. that the domain @xmath609 is stein .
then the categorical quotient @xmath610 with respect to the @xmath93-action is stein ( see @xcite , sect .
note that all local @xmath317-orbits are closed in @xmath609 and by @xmath611 of lemma [ levelset ] the domain @xmath609 is @xmath93-orbit - convex in @xmath210 .
it follows that @xmath612 is biholomorphic to @xmath613 , where @xmath614 is the canonical projection .
in particular @xmath615 is stein .
denote by @xmath616 the image of @xmath617 in @xmath618 via the canonical projection .
consider the commutative diagram @xmath619 where @xmath620 is our usual quotient map .
one checks ( cf . the proof of lemma [ levelset ] ) that @xmath617 intersects all local @xmath355-orbits of @xmath621 but one , namely @xmath622 .
moreover , for @xmath623 one has @xmath624 . as a consequence @xmath625 which , by lemma [ envelope ] , is not holomorphically convex .
this gives a contradiction , showing that @xmath41 is not holomorphically convex .
for the last statement , identify @xmath609 with @xmath41 and note that its envelope of holomorphy @xmath626 is a stein , @xmath43-equivariant , riemann domain over @xmath210 .
one has an induced stein , @xmath22-equivariant , riemann domain @xmath627 , where @xmath628 denotes the @xmath93-categorical quotient of @xmath626 ( see @xcite , sect.@xmath163 ) .
in fact , this can be regarded as a @xmath629-equivariant , riemann domain over @xmath630 , since the subgroup @xmath631 of @xmath632 acts trivially on @xmath532 .
then , by [email protected] in @xcite the map @xmath633 is injective and [email protected] in @xcite implies that @xmath626 is univalent .
that is , @xmath626 is a stein , @xmath43-invariant domain of @xmath210 .
as a consequence @xmath628 is biholomorphic to @xmath634 and , by lemma [ envelope ] , it necessarily coincides with @xmath630 .
choose , on a neighborhood @xmath131 of @xmath635 in @xmath636 , a stein local trivialization @xmath637 of the principal @xmath638-bundle @xmath639 .
then one has latexmath:[$\,\hat \omega_m\cap ( u \times \c^*)= \{\,(z,\lambda ) \in u \times \c^ * \ : \
a(z ) < @xmath131 with values in the extended real line . since @xmath641 is stein , the functions @xmath642 and @xmath643 are plurisubharmonic .
moreover , @xmath644 of lemma [ levelset ] implies that @xmath645 for @xmath646 close to @xmath647 .
thus @xmath648 and @xmath649 on @xmath135 .
finally a connectedness argument using local trivializations covering all @xmath44 shows that this holds for any trivialization .
hence @xmath650 as claimed .
let @xmath1 be endowed with one of the metrics @xmath42 . for @xmath651 , similarly to the cases considered in section @xmath652 , we show that the polar map @xmath56 is a biholomorphism from the maximal adapted complexification @xmath41 onto an @xmath43-invariant domain of @xmath210 which is not holomorphically convex .
however , note that here the projection @xmath653 of @xmath41 to @xmath425 misses more than one @xmath22-orbit .
finally , in the riemannian cases @xmath36 the polar map @xmath56 is not injective and its fibers consists of at most two elements .
the maximal complexification @xmath41 is neither holomorphically separable , nor holomorphically convex . for all @xmath33 , its envelope of holomorphy
is shown to be biholomorphic to @xmath210 .
[ levelset2](cf .
picture 3 ) let @xmath517 .
then * different local @xmath355-orbits of @xmath654 are mapped by @xmath655 to different @xmath355-orbits of @xmath656 , * local @xmath355-orbits closed to @xmath553 are mapped bijectively by @xmath439 to @xmath355-orbits of @xmath363 , i.e. @xmath368 acts globally in an @xmath355-invariant neighborhood of @xmath554 , * for @xmath657 one has @xmath658 and different local @xmath355-orbits of @xmath184 are mapped by @xmath655 to different @xmath355-orbits of @xmath656 , * for @xmath657 the polar map @xmath439 is injective on @xmath552 , * for @xmath36 one has @xmath659 .
more precisely @xmath184 contains the graph of @xmath660 , with @xmath10 defined as in lemma [ fixpoints ] .
@xmath104 recall that by lemma [ thesame ] the slice @xmath184 is a proper subdomain of @xmath661 for @xmath662 let @xmath663 .
a similar argument as in lemma [ levelset ] shows that different components of @xmath664 are mapped by @xmath439 to different @xmath355-orbits of @xmath596 .
one also checks that local @xmath355-orbits of @xmath665 are connected components of @xmath666 .
then in order to prove @xmath287 it is enough to show that for every @xmath667 and every connected component @xmath668 of @xmath664 the locally @xmath669-invariant set @xmath670 is connected . for this recall that ( cf.@xmath16prop.@xmath16[domain ] and rem.@xmath16[thesame ] ) the boundary of @xmath545 in the set @xmath671 is given by @xmath672 and
that @xmath673 is a connected component of @xmath674 , where @xmath675 since @xmath676 , in the coordinates @xmath677 such equations read as @xmath678 and @xmath679 .
thus it is enough to note that the function @xmath680 can be rewritten as @xmath681 hence it is strictly decreasing by lemma [ functions ] .
this proves @xmath412 .
the analogous proof as in @xmath644 of [ levelset ] implies @xmath526 . for @xmath28 one checks directly that @xmath682 ( cf .
lemma [ fixpoints ] ) .
thus @xmath418 holds for @xmath61 .
now let @xmath683 and let @xmath684 be an element of @xmath685 .
that is , @xmath686 with @xmath687 and @xmath688 where @xmath689 .
in order to show that @xmath684 does not belong to @xmath184 we need to see that @xmath690 , i.e that ( cf . lemma@xmath16[functions ] ) @xmath691 since @xmath692 , equation ( [ fixed ] ) above implies that @xmath693 .
then this inequality can be written as @xmath694 which becomes @xmath695 note that @xmath696 also implies @xmath697 .
then , by using equation ( [ fixed ] ) , one obtains @xmath698 which is easily checked to hold , since @xmath699 and @xmath700 . finally ,
a similar argument as in lemma [ levelset ] implies the last statement in @xmath701 .
part @xmath608 follows from @xmath418 by using the same argument as in lemma [ levelset ] for @xmath611 first note that @xmath702 , with @xmath703 uniquely defined by @xmath704 and @xmath705 .
rewrite condition @xmath257 of proposition [ domain ] as @xmath706 .
that is @xmath707 recall that @xmath708 is strictly increasing for @xmath241 ( lemma [ functions ] ) .
then for @xmath709 and @xmath710 one has @xmath711 implying that @xmath712 . since one sees that @xmath713 , in order to prove that the graph of @xmath714 is contained in @xmath272 it is enough to check that @xmath10 is strictly decreasing ( increasing ) for @xmath509 ( @xmath715 )
this can be done by a direct computation showing that @xmath716 [ inipart]@xmath104 let @xmath517 .
then the restrictions @xmath717 and @xmath718 are injective .
first consider the case @xmath719 .
since @xmath439 is necessarily injective on local @xmath355-orbits , @xmath287 of lemma [ levelset2 ] implies that @xmath439 is injective on @xmath720 .
moreover , a straightforward computation shows that @xmath439 is also injective on @xmath721 .
then , for @xmath722 , the map @xmath439 is injective on @xmath723 by @xmath418 of lemma [ levelset2 ] .
now recall that every @xmath43-orbit of @xmath41 meets @xmath184 in a single element and @xmath56 is injective on such an orbit by lemma@xmath16[diffeo ] .
then the statement follows in the case @xmath722 . for @xmath36 note that by @xmath611 of lemma [ levelset2 ] the set @xmath724 consists of a one - dimensional manifold with two connected components .
then in these cases one can essentially argue as follows .
assume by contradiction that @xmath725 for some @xmath726 in @xmath727 and @xmath728 in @xmath729 .
since @xmath730 , @xmath731 are both in @xmath552 , where @xmath439 is a local diffeomorphism , this would imply that @xmath558 is not injective , which gives a contradiction .
finally the case @xmath732 follows from @xmath483-@xmath589-equivariance of @xmath597 .
let @xmath6 be a complex @xmath733-manifold .
we need the following remark on limits of local orbits for the induced local @xmath638-action . for details
on induced local actions in this particular situation we refer to @xcite .
[ oloconv ] let @xmath6 be a complex @xmath733-manifold and consider the induced local @xmath638-action .
assume that there exist a sequence @xmath734 of @xmath6 and an element @xmath735 such that @xmath736 acts on every @xmath737 and the sequences @xmath738 , @xmath739 converge to different local @xmath638-orbits .
then @xmath6 admits no continuous plurisubharmonic exhaustion .
in particular it is not holomorphically convex .
assume by contradiction that a continuous , plurisubharmonic exhaustion @xmath740 of @xmath6 exists .
after integration over @xmath733 , this function can be assumed to be @xmath733-invariant .
let @xmath741 and @xmath742 be the limit points of @xmath737 and of @xmath743 .
denote by @xmath744 and @xmath745 the local @xmath638-orbits through @xmath741 and @xmath746 , respectively , and choose @xmath747 such that @xmath748 and @xmath749 . by assumption @xmath750 , therefore @xmath744 is given by @xmath751 , with @xmath752 and @xmath753 .
then @xmath754 as @xmath755 , in the sense of leaving all compact subsets of @xmath756
. thus there exists a real @xmath757 with @xmath758 such that @xmath759 .
furthermore @xmath760 , thus for @xmath761 large enough @xmath762 , while @xmath763 and @xmath764 .
however @xmath733-invariance and plurisubharmonicity of @xmath765 imply that the function @xmath766 is convex .
this gives a contradiction and concludes the proof .
in fact a similar argument yields the analogous result for actions of compact lie groups on holomorphically separable complex manifolds .
the remaining cases with @xmath268 are discussed in the following theorem .
[ domain>-1 ] let @xmath1 endowed with a left invariant metric @xmath42 , with @xmath767 .
the polar map @xmath768 is * injective for @xmath769 and consequently @xmath41 is @xmath43-equivariantly biholomorphic to @xmath607 .
such domain is not holomorphically convex , * not injective for @xmath36 and its fibers have at most two elements .
the maximal complexification @xmath41 is a riemann domain over @xmath44 which is neither holomorphically separable , nor holomorphically convex . in both cases
the envelope of holomorphy of @xmath41 is biholomorphic to @xmath210 .
@xmath104from @xmath608 of lemma [ levelset2 ] and proposition [ iniini ] it follows that @xmath163 is injective for @xmath770 .
assume by contradiction that @xmath609 is stein .
then the analogous argument as in theorem [ domain<-1 ] shows that @xmath653 is stein .
however one checks that @xmath771h \in \sigma_m$ ] ( cf
. proof of lemma [ levelset2 ] ) .
then , by @xmath22-invariance of @xmath41 and @xmath22-equivariance of @xmath56 , one has @xmath772h ) k^\c \,\subset \,\pi(p_m(\omega_m))$ ] . as a consequence
the closure of the akhiezer - gindikin domain @xmath773 is contained in @xmath615 . since one sees that @xmath774 ( cf .
proof of lemma [ levelset2 ] ) , lemma [ envelope ] implies that @xmath653 is not stein .
this gives a contradiction , implying @xmath412 .
for @xmath775 note that by @xmath611 of lemma [ levelset2 ] there exists an element @xmath776 in @xmath777 with @xmath778 . then by @xmath779-@xmath780-equivariance of @xmath439 one has @xmath781 , showing that @xmath475 is not injective
then , by proposition [ iniini ] the polar map @xmath163 is not injective as well . since @xmath782
, from corollary [ inipart ] and proposition [ iniini ] it follows that the fibers of @xmath56 consist either of one point or two points @xmath783 and @xmath784 . in the second case , necessarily @xmath785 and @xmath786 , where @xmath787
. now let @xmath788 and note that local @xmath638-orbits through @xmath789 accumulate , for @xmath790 , to different local @xmath638-orbits .
indeed this holds true for their images in the quotient @xmath791 .
such images are given by the @xmath355-orbits through @xmath789 for which , chosen @xmath792 small enough , @xmath793 and @xmath794 accumulate to different local @xmath355-orbits . namely the two connected components of @xmath795 , with @xmath796 such that @xmath797 ( cf .
the proof of lemma [ levelset2 ] ) .
then proposition [ oloconv ] implies that @xmath41 is not holomorphically convex .
assume by contradiction that @xmath41 is holomorphically separable .
then it embeds in its envelope of holomorphy @xmath798 , which is a non - univalent , stein , @xmath43-equivariant , riemann domain over @xmath210 .
by [email protected] in @xcite , the induced stein , @xmath22-equivariant , riemann domain @xmath799 is also non - univalent .
however the subgroup @xmath800 of @xmath66 acts trivially on @xmath798 , thus @xmath628 can be regarded as a @xmath801-equivariant , riemann domain .
then [email protected] in @xcite implies that @xmath633 is injective , giving a contradiction .
finally , in view of @xmath413 of lemma [ levelset2 ] , the last statement follows from the same argument as in theorem [ domain<-1 ] .
here we carry out the computation used in section 4 . for this note that @xmath802\,=\,2w\,$ ] , @xmath803\,=\,-2h\,$ ] and @xmath804\,=\,-2u$ ] .
thus @xmath805\,=\,2aiw , \\ \ad^2(i(umu - ah))\,u\,=&\,[i(umu - ah),2aiw ]
\,=\\ & \,-4a(au - umh),\\ \ad^3(i(umu - ah))\,u\,=&\,[i(umu - ah),-4a(au - umh ) ] \,= \\ & \ , 8a(u^2m^2-a^2)iw,\\ \ad^4(i(umu - ah))\,u\,=&\,[i(umu - ah),8a(u^2m^2-a^2)iw ] \,=\\ & \ , -16a(u^2m^2-a^2)(au - umh).\end{aligned}\ ] ] where @xmath810 .
similarly , for the second vector one has @xmath811\,=\,2umiw,\\ \ad^2(i(umu - ah))h\,=&\,[i(umu - ah),2umiw ] \,=\\ & \,-4um(au - umh),\\ \ad^3(i(umu - ah))h\,=&\,[i(umu - ah),-4um(au - umh)]\,= \\ & \ , 8um(u^2m^2-a^2)iw,\\ \ad^4(i(umu - ah))h\ , = & \,[i(umu - ah),8um(u^2m^2-a^2)iw]\,=\\ & \,-16um(u^2m^2-a^2)(au - umh).\end{aligned}\ ] ] for the third vector one has @xmath817\,=\,2i(au - umh ) , \\
\ad^2(i(umu - ah))\,w\,= & \,[i(umu - ah),2i(au - umh)]\,=\\ & \,4(u^2m^2-a^2)\,w , \\ \ad^3(i(umu - ah))\,w\,= & \,[i(umu - ah),4(u^2m^2-a^2)w ] \,= \\ & \,8(u^2m^2-a^2)i(au - umh ) , \\ \ad^4(i(umu - ah))\,w\,= & \,[i(umu - ah),8(u^2m^2-a^2)i(au - umh)]\,= \\ & \,16(u^2m^2-a^2)^2w.\end{aligned}\ ] ] for the fourth vector one has @xmath823\,=\,-2amw,\\ \ad^2(-i(umu - ah))(-imu)\,=&\,[-i(umu - ah),-2amw ] \,=\\ & \,4ami(au - umh),\\ \ad^3(-i(umu - ah))(-imu)\,=&\,[-i(umu - ah),4ami(au - umh ) ] \,= \\ & \,-8am(u^2m^2-a^2)w,\\ \ad^4(-i(umu - ah))(-imu)\,=&\,[-i(umu - ah),-8am(u^2m^2-a^2)w]\,= \\ & \,16am(u^2m^2-a^2)i(au - umh).\end{aligned}\ ] ] for the fifth vector one has @xmath829\,=\,2umw , \\
\ad^2(-i(umu - ah))\,ih\,=&\,[-i(umu - ah),2umw]\,=\\ & \,-4umi(au - umh),\\ \ad^3(-i(umu - ah))\,ih\,=&\,[-i(umu - ah),-4umi(au - umh)]\,= \\
& \,8um(u^2m^2-a^2)w,\\ \ad^4(-i(umu - ah))\,ih\,=&\,[-i(umu - ah),8um(u^2m^2-a^2)w]\,=\\ & \,-16um(u^2m^2-a^2)i(au - umh).\end{aligned}\ ] ] therefore @xmath830 @xmath831 @xmath832 @xmath833 @xmath834 for the sixth vector one has @xmath835\,=\,2(au - umh),\\ \ad^2(-i(umu - ah))\,iw\,= & \,[-i(umu - ah ) , 2(au - umh)]\,=\\ & \,4(u^2m^2-a^2)iw,\end{aligned}\ ] ] @xmath836\,= \\ & \,8(u^2m^2-a^2)(au - umh),\\ \ad^4(-i(umu - ah))\,iw\,=&\,[-i(umu - ah),8(u^2m^2-a^2)(au - umh)]\,=\\ & \,16(u^2m^2-a^2)^2iw.\end{aligned}\ ] ] therefore @xmath837 @xmath838 @xmath839 @xmath840 @xmath841
|
let @xmath0 be a non - compact , real semisimple lie group .
we consider maximal complexifications of @xmath0 which are adapted to a distinguished one - parameter family of naturally reductive , left - invariant metrics . in the case of @xmath1
their realization as equivariant riemann domains over @xmath2 is carried out and their complex - geometric properties are investigated .
one obtains new examples of non - univalent , non - stein , maximal adapted complexifications .
|
we wish to consider the problem of simplifying immersed planar curves , in a sense which will later be made precise .
intuitively , a generic immersion @xmath1 can be considered as a knot diagram without the crossing data , and for such immersions we can apply planar versions of the reidemeister moves for knot diagrams . by applying all three reidemeister moves to such a diagram ,
one is able to obtain a standardly embedded circle with no double points .
one way to see this is to add crossing data so as to give a knot diagram of the unknot ; applying the standard three reidemeister moves to this knot diagram gives the standardly embedded circle @xcite . in @xcite , stlund observed that reidemeister 1 is the only move that changes the degree of the gauss map , and showed that reidemeister move 3 is the only move that can change the signed number of instances of certain subdiagrams of the gauss diagram for an embedding .
these properties were used to show that any knot @xmath2 admits a pair of diagrams such that every sequence of reidemeister moves connecting them contains instances of reidemeister moves 1 and 3 .
planar versions of the same arguments give immersions of the circle in which every connecting sequence contains instances of reidemeister moves 1 and 3 .
stlund conjectured that any sequence of reidemeister moves could be replaced with a sequence consisting of only moves 1 and 3 .
a counterexample to this conjecture for the case of knots appears in @xcite .
independently , the first auther showed in @xcite that every knot type admits pairs of diagrams such that every connecting sequence contains every reidemeister move type . these two arguments , however , rely heavily on information about the crossings , and do not generalize to the case of planar reidemeister moves .
the purpose of the present paper is to disprove the planar version of the conjecture .
since every sequence of planar reidemeister moves corresponds to a ( non - unique ) sequence of knot reidemeister moves , the counterexample also provides an alternate disproof of the conjecture for knots .
an _ immersed curve _ is the image of a map @xmath3 where @xmath4 is some surface , such that any point in the pair @xmath5 has a neighborhood homeomorphic to a neighborhood in the picture below .
the pair @xmath6 shall denote the immersed curve .
+ figure 1 in this paper , @xmath4 will usually be @xmath7 or @xmath8 . [ rmoves]given an immersed curve , the reidemeister moves are given , as numbered below ( 1a , 1b , 2a , 2b , or 3 ) , by identifying a disk in @xmath9 homeomorphic to the disk on the left side of the numbered picture and replacing it with the homeomorphic preimage of the disk on the right .
+ figure 2 by convention , planar isotopies are always allowed as moves , even when not explicitly mentioned .
any two homotopic immersed curves are connected by a sequence of reidemeister moves and planar isotopies .
the argument in @xcite works for an arbitrary surface when there is no crossing data .
an immersed curve @xmath10 is _
@xmath11-simplifiable _ if for some @xmath12 there exists a sequence of immersed curves @xmath13 such that @xmath14 is obtained from @xmath15 by applying one of reidemeister moves 1a , 1b , or 3 , and @xmath16 where @xmath17 is an embedding .
the sequence @xmath13 is called a _ simplifying sequence _ for the curve @xmath18 if @xmath4 is a surface of genus at least @xmath19 and @xmath10 is not nullhomotopic , then @xmath10 is not @xmath11-simplifiable .
this is because the reidemeister moves applied to curves preserve homotopy type .
stlund s conjecture , stated in our language , is that every immersed planar curve is @xmath11-simplifiable .
since any curve which is @xmath11 -simplifiable in @xmath20 is @xmath11-simplifiable in its one point compactification @xmath8 , the next theorem suffices to disprove the conjecture : [ main theorem][main]the following curve is not @xmath11-simplifiable in @xmath8 : + figure 3 the proof of this theorem does not rely on heavy machinery .
it should be noted that there are immersed curves without an obvious simplifying sequence , which are nonetheless @xmath11-simplifiable .
for example , consider the following : + figure 4 the easiest way to show that this curve is @xmath11-simplifiable is to apply this theorem : [ double bigon]let @xmath21 be a @xmath11-simplifiable curve .
suppose that in @xmath21 we replace some instances of the local picture with the local picture relative boundary ( i.e. double bigons replace double points ) to obtain curve @xmath22 .
if @xmath21 is @xmath11-simplifiable , then @xmath22 is @xmath11-simplifiable .
it should be noted that theorem [ double bigon ] does not say that moves 1 and 3 may be used to replace a double point with a double bigon in an arbitrary diagram .
nonetheless , applying theorem [ double bigon ] repeatedly to figure 5 gives the following @xmath11-simplifiable immersed curve : one could generalize stlund s conjecture and ask whether two homotopic curves on a surface are related by only the first and third reidemeister moves . this generalized conjecture is much easier to falsify .
it is in fact a generalization because all generic curves on @xmath7 or @xmath8 are homotopically trivial .
[ torus]the following two curves on @xmath23 are homotopic , but are not related by a sequence of reidemeister moves consisting of only the first and third moves .
this section proves theorem [ main ] .
consider the following shaded regions in the curve from figure 4 , interpreted as a diagram on @xmath8 : + figure 5 reinterpret the diagram as a collection of eight blue boxes containing immersed tangles , connected by lines with no double points .
each box has a left and right side , as labeled below ; the left side of a given box is connected to the right side of its neighbor .
two polygons in the diagram deserve special attention and are marked with a star .
+ figure 6 the diagram satisfies the following properties : 1 . [ strands ends ] each blue box contains a tangle with three strands .
one of the strands , denoted strand 1 , begins and ends at the left side .
strand 2 begins and ends at the right side .
strand 3 has one endpoint on each side of the box .
[ strands connect ] in each box , the left side of strand 3 connects to strand 2 in the adjacent box to the left .
the right side of strand 3 connects to strand 1 in the adjacent box on the right .
[ strands two double ] strands 1 and 2 intersect in exactly two double points . 4 .
[ strands four sides]the polygons marked with a star have at least four sides .
we will show that any application of moves 1a , 1b , or 3 to any copy of figure 6 with immersed tangles satisfying the above properties results in a diagram which may be interpreted as a copy of figure 6 with immersed tangles still satisfying those properties .
since property [ strands two double ] implies that any blue box has at least two double points , every sequence of such moves results in a diagram with at least sixteen double points .
this proves that the curve is not @xmath11-simplifiable .
first , note that a move of type 1a , 1b , or 3 occuring entirely within one of the blue boxes gives a diagram ( with the same boxes ) satisfying all of the above properties .
such a move fixes the endpoints of the strands , so properties [ strands ends ] and [ strands connect ] remain satisfied .
none of these moves change the number of times one strand intersects another within a box , so property [ strands two double ] holds after a move .
property [ strands four sides ] actually follows from the arrangement of the boxes and the other three properties .
fix a starred polygon and consider the portion of its boundary lying within a single blue box .
if the end points of that boundary portion belong to different strands within the box , that box contains at least one vertex for the starred polygon .
otherwise , property [ strands ends ] implies that both ends belong to strand 3 .
then property [ strands connect ] implies that the end points for the portion of the starred polygon s boundary lying in each of the adjacent blue boxes belong to different strands within that box .
thus each of the adjacent boxes contains a vertex for the starred polygon .
therefore , allowed moves can not reduce the number of vertices ( or edges ) for a starred polygon below four .
it remains to show that it suffices to consider only moves lying within a single blue box .
first , consider reidemeister move 1b . performing this move
requires a disk in our immersed curve that is homeomorphic to the disk on the left side of picture 1b in figure 2 in definition [ rmoves ] .
suppose that the segment on the left side of picture 1b in definition [ rmoves ] is not contained completely inside one of the blue boxes as specified above .
then one can redefine the blue box before performing the move so that it occurs entirely within a single blue box .
for example , suppose that the disk for move 1b is the following : one can then isotop the blue boxes , while leaving crossings fixed , as follows : move 1a , on the other hand , removes a one - sided polygon .
this polygon must lie entirely within a single blue box , for the following reason : clearly , a one - sided polygon can not separate the two starred regions .
if a closed smooth subcurve of @xmath24 does not lie in a single blue box , and does not separate the two starred regions , then it must enter and exit one of the blue boxes on the same side .
such a curve contains a segment of strand type 1 or 2 , and by property [ strands two double ] , any such curve will have at least two crossings
. finally , move 3 always occurs on a neighborhood of a triangle ( which can never be marked with a star ) .
if that triangle lies entirely in one blue box , that box may be isotoped as above to include the entire disk on which the move occurs . otherwise , the triangle intersects the white region ( an example of such a potential triangle is marked with a red dot in figure 6 ) .
one can verify that this implies that one of the blue boxes intersects the triangle only in a single corner , as shown in figure 7 , for example .
+ figure 7 assume without loss of generality that the triangle extends to the right of the blue box containing just the corner , as in figure 7 .
there are two possibilities for the strand ends on the right side of the leftmost box shown in figure 7 .
either exactly one of the ends belongs to strand 3 , or both ends belong to strand 2 . in the box to the right ,
either both of the pictured left ends belong to strand 1 , or exactly one belongs to strand 3 , respectively . in either case
, isotoping the blue boxes in figure 7 to give the blue boxes in figure 8 preserves the required properties and reduces the number of white regions in the triangle .
after at most two such box adjustments , all three vertices of the triangle must lie in the same blue box .
then , since there are no isolated blue corners , the entire triangle must lie within a single blue box .
+ figure 8 one could also prove this theorem using gauss diagrams .
we give the main outline , leaving the proof to the reader .
the gauss diagram for the immersed curve in figure 4 is as follows : + figure 9 move 1b adds a chord to the diagram , which by convention shall be colored gray .
the following properties are preserved by reidemeister moves 1 and 3 . 1 .
if all the gray lines are erased , the resulting diagram is exactly as shown in figure 9 above except that some of the adjacent and parallel pairs of black lines may be replaced with crossed pairs .
2 . both endpoints of every gray chord lie in one of the eight regions indicated in figure 9 .
[ double bigon lemma]the following pictures are connected by a sequence of reidemeister moves 1 and 3 : [ proof of theorem [ double bigon]]consider two immersed curves @xmath25 and @xmath26 , equal except inside of a box .
the contents of the box for the curves @xmath25 and @xmath26 are given respectively by the following pictures on the left and right : suppose @xmath25 is @xmath11-simplifiable
. then reidemeister moves performed on @xmath25 that are supported away from the box have analagous reidemeister moves on @xmath26 .
however , the simplifying sequence for @xmath25 may contain moves 1a and 3 which involve the box .
the following sequences of moves on @xmath26 are analogous to moves 1a and 3 on @xmath25 which involve the box . in these sequences
it may be necessary to first apply lemma [ double bigon lemma ] to obtain the leftmost picture .
such that the two strands of the tangle inside the disk intersect , one can never obtain an embedded curve ( i.e. a curve without double points ) . observe that in the picture above , there is exactly one region not contained in the disk , and this region has genus .
up to isotopy every r1a , r1b and r3 move is supported within the disk . however , as noted in the proof of the main theorem , r1 and r3 moves on tangles do not change the number of intersections between strands .
|
in 2001 , stlund conjectured that reidemeister moves 1 and 3 are sufficient to describe a homotopy from any generic immersion @xmath0 to the standard embedding of the circle .
we show that this conjecture is false .
|
the detection and the study of ionized absorbers is more difficult than that of cold neutral absorbers , but can yield much more detailed information about the nature of the absorbers and the state and geometry of the nuclear regions of agn .
if photoionization applies then the electron density of the gas and its distance from the ionizing source can be estimated .
absorption features from ionized gas , are common in the x - ray spectra of seyfert galaxies and some quasars.178 , halpern ( 1984 ) , pan , stewart and pounds ( 1990 ) ; 3c 351 , fiore et al .
( 1993 ) , nicastro et al . , 1998 ;
see also reynolds ( 1997 ) ] deep oxygen vii and viii absorption edges at 0.74 kev and 0.87 kev ( rest ) have been detected by the rosat pspc and the asca sis in quite a large number of seyfert 1 galaxies ( reynolds , 1997 ) .
we do not know yet what is the origin of the gas ionization .
models to date have assumed the simplest equilibrium photoionization case ( e.g. fiore et al .
, 1993 ; guainazzi et al . , 1996 ; reynolds et al . , 1995 ) . in this case , if the gas is confined in a single cloud of constant density , and if the recombination time is smaller than the typical variability timescale , then the ionization parameter , and hence the ionization state of the gas , should follow closely the intensity of the ionizing continuum .
this is not always observed . in two asca observations of mgc-6 - 30 - 15 ( fabian et al .
1994 , reynolds et al .
1995 ) the best fit ionization parameter is higher when the ionizing flux is lower , in contrast with the expectations of the simplest equilibrium photoionization model ; the asca observations of mr2251 - 178 show a roughly constant ionization parameter despite large variation in the 2 - 10 kev flux ( reynolds & fabian , 1995 ) ; finally , in a rosat observation of ngc 4051 ( mchardy et al . , 1995 )
, the ionization parameter does not linearly track the luminosity , but shows changes that are smoothed and delayed with respect to the luminosity changes .
it seems clear that , at least in the above three cases ( which are also the best studied ) , the simplest photoionization equilibrium model is inadequate .
we clearly need more complete and consistent models to interpret the available data .
the gas could well be distributed in an irregular region with varying density .
different ionization states would then apply to different region of this gas .
other authors have adopted such multi - zone models ( otani et al .
1996 , kriss et al .
1996 ) . in those models
the authors assume that absorption features from different ions of the same element are imprinted on the spectrum by the transmission of the ionizing radiation through multiple distinct clouds of gas with different geometrical configurations , ionization states and densities .
this is of course a possibility , but to us it seems rather ` ad hoc ' . in this paper instead we discuss the additional physics of collisional ionization , and of time - evolving photoionization in a single zone model .
time - evolving photoionization models have been applied in the past to the extended narrow line emitting region of radio galaxies , to predict the evolution of the line ratios in low - density ( @xmath10 @xmath4 ) , low ionization ( logu@xmath11 -3 ) , multi - zone photoionized gas after the switch - off of the ionizing source ( binette et al . , 1987 ) .
we present here an iterative solution of the time - dependent ionization balance equations ( accounting for all ions of the most abundant elements in gas with solar composition ) , and apply our models to a drastically different physical and geometrical configuration : a relatively high - density ( @xmath12 @xmath4 ) , medium - high gas ionization ( logu @xmath11 -1 , 1 ) , geometrically thin ( @xmath13 ) , single zone absorber along the line of sight , undergoing rapid and persistent variations of the ionizing flux .
as an example , we apply our models to the rosat data of the seyfert 1 galaxy ngc 4051 . a detailed analyses of the time behaviour of ionized gas in seyfert 1 galaxies , using the asca - sis relatively high resolution data ,
is deferred to a future pubblication .
the innermost regions of an agn are likely to be ` active ' , in the sense that the gas there confined is expected to be involved in significant bulk motion , as in strong outflows ( arav et al . , 1995 ) or inflows .
in particular the evidence for ionized outflows from the central regions of agn is strong ( mathur et al .
1994 , 1995 , 1997 , 1998 ) .
hence sources of mechanical heating of the gas ( for example adiabatic compression by shock waves ) may well be at work in the high density clouds , making collisional ionization the dominant ionization mechanism . in the low density clouds
the gas could be far from photoionization equilibrium , because the recombination timescales have become greater than the x - ray continuum variability timescales .
hereinafter , by _ photoionization time _ ( @xmath14 ) and _ recombination time _ ( @xmath15 ) we mean the time necessary for the gas to reach photoionization equilibrium with the ionizing luminosity during increasing and decreasing phases respectively .
we discuss in turn the possible importance of collisional ionization and time - evolving photoionization , in determining the final transmitted spectrum that we observe in agn . to study collisional ionization we constructed a series of pure collisional ionization models and compared them with the equilibrium photoionization models .
the models were created using cloudy ( version 90.01 , ferland , 1996 ) , fixing the ionization parameter ( i.e. the dimensionless ratio between the number of hydrogen ionizing photons and the electron density of the gas ) u to @xmath16 and calculating the spectra transmitted through clouds with a total column density , @xmath17 @xmath1 , constant density and temperature ( the calculation of the physical condition of clouds in coronal or collisional equilibrium is one of the options of cloudy ; ferland , 1996 ) .
the distribution of fractional abundances for the most important ions is essentially independent of the electron density value : for @xmath18 @xmath4 ; the variation is smaller than 13% for ions with fractional abundances greater than 0.01 . for temperatures @xmath19 in the range @xmath20 the main edges imprinted on the spectra
are roughly the same as those in the spectra from photoionized gas with @xmath21 . however , the distribution of the different ionic species is very different in the two cases , and so the relative optical depth of the edges differ markedly .
high quality x - ray spectra , where more than one edge is visible , could then discriminate between the two cases .
the different ionic abundances in the collisional and photoionized cases can be seen in figs .
1a , b and 2a , b . here
we plot the fractional abundances of neviii - nexi and ovi - oix ( fig
. 1 ) and fexiv - fexxv ( fig .
2 ) , computed in the case of photoionization and collisional ionization respectively as a function of u and @xmath19 . as fiore
et al ( 1993 ) and mathur et al . (
1994 ) have shown , the photo - ionization state of the gas depends strongly on the spectral energy distribution ( sed ) of the ionizing continuum , from radio to hard x - rays . to simplify the comparison between photoionization models and the data , we therefore used an ionizing continuum similar to the sed of ngc 4051 ( done et al 1990 ) .
firstly we note that the fractional abundances in photoionized gas are more smoothly distributed than those of collisionally ionized gas .
let us consider the regime in which the ovii and oviii fractions are higher than @xmath22 , so that both ovii and oviii edges are present in the emerging spectrum , as sometimes found in seyfert galaxies spectra ( e.g. otani et al .
1996 , guainazzi et al .
1996 ) . in the photoionization case
the range of u where this occurs is 1.23.6 , a factor 3 .
this is twice as wide as the corresponding range of @xmath19 : @xmath23 k ( a factor of 1.6 : dotted lines in panels _ a _ and _ b _ of fig .
moreover , neix is abundant and carbon is almost fully stripped in the collisional case , but not in the photoionized case .
the difference in the resulting spectra is shown in fig . 3a . here
we show the ratio between two power law spectra emerging respectively from clouds of collisionallly ionized ( @xmath24 k ) gas , and photoionized ( u=2 ) gas with solar abundances ( the choices of @xmath19 and u were made by looking for similar ovii - oviii abundances : see fig .
below 1 kev there is a large feature due to the deep cvi k absorption edge in the photoionization case .
the most important feature above 1 kev is the neix absorption edge in the collisional case , but this feature is smoothed by the presence , in both cases , of deep ovii and oviii absorption edges .
reynolds ( 1997 ) , fitted both a simple ovii and oviii k - edges model and a physical photoionization model to asca sis spectra of a sample of seyfert 1 known to host a warm absorber . comparing the results he found that in some cases , the best fit 2-edges model continue to show residuals at the energy of cvi k edge , while residuals to physical photoionization model do not .
this may point to photoionization as ionizing mechanism .
spectra with higher resolution from 0.4 kev to 3 kev should therefore allow one to diagnose absorption by collisionally ionized or photoionized gas that has significant ovii and/or oviii abundances ( e.g. otani et al .
1996 , guainazzi et al .
1996 ) .
the other important feature of collisional models is the inertia of the heavy elements ( from ne to fe ) against becoming highly ionized , even when the oxygen is almost fully stripped . in fig .
1a and 1b we have marked the range of u for which oix is by far the dominant ion ( @xmath25 ) and oviii is the only oxygen edge in the emergent spectrum ( @xmath26 , for an @xmath27 of @xmath28 @xmath1 and solar abundances ) .
these same intervals of u and @xmath19 are also shown in fig 2a and 2b . in the photoionization case
neon is almost fully stripped ( nexi ) , and the dominant ions of the iron are fexviii to fexxiii .
instead in the collisional case there is a range of temperature for which neon and iron are much less ionized , with the dominant ions being neix - nex , and fexvi to fexx .
in particular we note the large abundance of fexvii ( fexvii and fexxv are respectively ne - like and he - like , and so very stable ) .
the different ionization level of neon and iron implies a very different emergent spectrum between 1 and 3 kev .
this is illustrated in fig .
3b which shows the ratio between spectra transmitted from collisionally ionized gas with @xmath29 k , and photoionized gas with u=10 ( both with @xmath17 @xmath1 ) . with these values of @xmath19 and u , @xmath30 in both collisional ionization and photoionization models ( see fig .
the ratio does not show any significant feature at e@xmath31 kev ( implying similar ionization states of carbon , nitrogen and oxygen ) , but between 1 and 3 kev , the spectrum from collisionally ionized gas shows a large and complicated absorption structure due mainly to iron l xvii - xix and neon k ix - x absorption ( fig .
3b ) . this absorption structure may be mis - interpreted in moderate quality spectra as a flattening above @xmath2 kev .
again , higher spectral resolution data with good @xmath32 can distinguish between photoionization or collisional equilibrium . that the neix and fexvii edges are of similar depth is due to a coincidence of cross - sections and abundances .
the photoelectric cross section of the iron l shell atoms ranges from @xmath33 @xmath34 for the lp levels of fexvi - fexix ( @xmath35 kev ) , to @xmath36 @xmath34 for the lp level of fexxii ( e=1.78 kev ) ( kallman & krolik , 1995 ) .
furthermore in the collisional case the relative abundance of fexvii is in the range 0.20.7 , compared to @xmath37 in the analogous photoionization case .
the relative abundance of the nex , in the collisional case , is greater than 0.25 , but nex @xmath38 in the analogous photoionization case , while @xmath39 ( neix)@xmath40 @xmath41 and ( neix)@xmath42 @xmath43 ( fig .
1 a , b ) .
the k - edge energy of nex is 1.36 kev the same of that of the fexviii lp - edge .
while the photoelectric cross - section of neon is about a factor 4 lower than iron , the solar abundance of the neon is about a factor 4 greater than that of iron ( grevesse & anders 1989 ) .
furthermore the cross - section of neix is about a factor 3 higher than nex . as a result ,
spectra from a cloud of collisional ionized gas at @xmath44 k , will show a very deep ( @xmath45 ) edge at the fexvii lp - edge energy ( 1.26 kev ) , a deep ( @xmath46 ) edge at 1.36 kev due to the comparable contributions of the nex k - edge and the fexviii lp - edge , and a similarly deep absorption edge at the neix k - edge energy ( @xmath47 kev , @xmath48 ) . in the corresponding photoionization case ,
the relative abundances of both neix - x and fexvii - xviii are too low ( fig .
1a , b , 2a , b ) to imprint similar features on the spectrum . when collisions are the dominant ionization process , ( @xmath49 ) we can estimate the minimum distance , r , between the ionized gas cloud and the central source . for an ionizing luminosity @xmath50
, we find : @xmath51 ( with @xmath52 being the electron density in units of @xmath53 @xmath4 , and @xmath54 the rate of photons ionizing hydrogen in units of @xmath55 ph s@xmath56 : @xmath57 ) . for typical agn ionizing continuum shapes and luminosities ( @xmath58 , @xmath59 erg s@xmath56 ) , and assuming @xmath60 , we find @xmath61 cm , @xmath62 cm , respectively , similar to the size of the blr for such agn ( peterson et al . , 1993 ) . the above discussion concerns gas in pure collisional equilibrium
if the gas density is low enough and/or if the gas is close enough to the central x - ray source then photoionization can be important too .
we examined a number of models with a varying mixture of collisional ionization and photoionization .
for @xmath63 we find that for values of u @xmath64 the main features of the fractional abundance distribution are still those of pure collisional ionized gas .
increasing u to 0.3 causes a sharp change in the ionization structure . for @xmath65 u @xmath66 ( the exact value of the upper limit depending on the equilibrium temperature determined by collisional mechanisms ) the ionization structure
is determined by _ both _ processes , and the transmitted spectra resemble those from pure photoionized gas of _ much higher _ ionization parameter .
this is shown in fig .
4 where we plot in the lower panel three spectra from purely photoionized gas ( column density log@xmath27=22.5 and log(u)=0.7 , 1.1 and 1.4 ) , and in the upper panel 5 spectra from gas in which both mechanisms are at work ( log@xmath27=22.5 , @xmath63 and log(u)=-2 , -1 , 0.3 , 0 , 0.4 ) .
the pure photoionization spectra with log(u)=0.7 , 1 ( lower panel ) are very similar to the mixed collisional and photoionization spectra with a factor 4 - 10 times lower u ( log(u)=-0.3 , 0.4 ; upper panel ) . fitting the mixed spectra with pure photoionation models would give a much more compressed range of u than the real one , as found by mchardy et al .
in ngc 4051 . the above spectra in the two panels of fig .
5 are practically undistinguishable even at high resolution , because the shape of the fractional abundance distributions when both processes are working , closely resemble those of a purely photoionized gas ( see upper panels of fig . 1 and 2 ) .
fortunately the two models predict quite different delay properties ( 2.3 ) . if photoionization is the dominant process ( the gas being purely photoionized , or the density enough low to give u@xmath67 ) the main features of the fractional abundance distribution are those of photoionized gas .
however , if the x - ray source is variable , photoionization equilibrium will apply only if the density is high enough to make the ion recombination timescales shorter than the variability timescales .
there are regimes where photoionization equilibrium does not apply at all . here
we investigate the low density case in some detail , beginning with a discussion of the relevant physics .
let us suppose that a geometrically thin single cloud of optically thin , gas is illuminated by an intense flux of ionizing radiation emitted by a variable source located at a distance r from the cloud .
we can calculate the equilibrium distibution of the ionic species in the cloud , by setting the photoionization rate equal to the radiative recombination rate ( see , e.g. , netzer , 1990 ) .
@xmath68 and adding the condition for charge conservation @xmath69 . here
@xmath70 is the radiative recombination coefficient ( in @xmath71 s@xmath56 ) , which includes recombination to all levels , and @xmath72 is the photoionization rate of the ion @xmath73 , which , for optically thin gas clouds , can be written ( see , eg . , netzer , 1990 ) : @xmath74 the lower limit of the integral is the threshold ionization frequency of the ion @xmath73 .
@xmath75 is the photoelectric absorption differential cross section of the ion @xmath73 . in this simple scheme
the ionization state of the gas is completely determined , at equilibrium , by the value of the ionization parameter @xmath76.[multiblock footnote omitted ] with these approximations the ionization parameter u can be written as a function of the ratio between any two consecutive ionic species of the generic element x : @xmath77 if the gas is not in equilibrium eq .
5 is of course meaningless .
however , it is still possible to use it formally by introducing an ionization parameter @xmath78 which depends both on time _ and _ on the specific ionic species under consideration .
the time evolution of the relative density of the ion @xmath79 of the element @xmath80 , considering only radiative recombination and photoionization , is then given by ( krolik and kriss , 1995 ) : @xmath81n_{x^i } + f_{x^{i-1}}n_{x^{i-1 } } + \alpha_{rec}(x^i , t_e ) n_en_{x^{i+1}}. \leqno ( 6)\ ] ] the first term on the right of eq .
6 , is the _ destruction rate _ of the ion @xmath73 , both by photoionization @xmath82 and radiative recombination @xmath83 , while the other two terms indicate the _ formation rate _ of the ion @xmath73 by photoionization of the ion @xmath84 and radiative recombination of the ion @xmath85 respectively . the solution of eq .
6 is a system of n coupled integral equations in the n unknowns @xmath86 , which is analytically solvable only for n=2 , with the addition of the charge conservation condition .
these solutions define the time scale @xmath87 , that measures the time necessary for the gas to reach photoionization equilibrium with the ionizing continuum . this time is given at any point of the light curve of the ionizing continuum by the inverse of the destruction rate of the ion @xmath73 ( following krolik & kriss 1995 ) .
a useful analytical approximation for @xmath87 is : @xmath88 } { \left[{1 \over { \left({{\alpha_{rec}(x^{i-1},t_e ) } \over { \alpha_{rec}(x^i , t_e)}}\right)_{eq } + \left({n_{x^{i+1 } } \over { n_{x^i } } } \right)_{eq}}}\right]_{t+dt } } , \leqno ( 7)\ ] ] where _ eq _ indicates the equilibrium quantities . during increasing ionization flux phases we call @xmath89 _ photoionization time _
, @xmath14 ; during decreasing ionization flux phases we call @xmath89 _ recombination time _ , @xmath15 .
these times are generally different from @xmath90 and @xmath91 in eq . 3 and 4 .
equation 7 shows that the time @xmath92 necessary for the gas to reach equilibrium depends explicitly on the electron density @xmath93 in the cloud , and on the equilibrium ratio between two consecutive ionic species calculated at the time t+dt .
this is the key result of this work , which has major consequences : 1 .
the time scale on which the gas reaches equilibrium with the ionizing continuum depends on the electron density , _ even when the continuum increases _ ( fig .
5 ) ; 2 .
different ions reach their equilibrium relative abundances at different points of the light curve ( fig .
6 ) ; 3 .
if changes of intensity are not instantaneous ( dl / dt@xmath94 ) , the time behavior of the relative abundance of a given ion can be opposite to that of the ionizing source ( fig .
all of these effetcs have been seen in ionized absorbers in agn . using these effects non - equilibrium photoionization models
can strongly constrain the physical state of the absorber .
we created a program to solve the first order differential equation system of eq . 6 for all the ions of the elements h , he , li , c , n , o. the program uses an iterative method ( see gallavotti , 1983 ) , that permits the solution of any system of n first order differential equations in the n unknowns @xmath95 , of the form @xmath96 , ( @xmath97 ) , with the only conditions being that @xmath98 and a limited ensemble @xmath99 exists , such that @xmath100 .
we consider the photoionization from the k - shell of each element , and radiative recombination to all levels for each ion .
we use the recombination rates tabulated by shull and van steenberg ( 1982 ) for the metals .
we take the values of the recombination rate of hydrogen from ferland ( 1996 ) ; we get the total recombination rate by summing over levels n=1,20 .
we calculate the photoionization rate from the k - shell of each ion , carrying out the integrals in eq .
2 , using for the spectral shape of the ionizing continuum , from the lyman limit to @xmath101-rays , a simple power law with @xmath102 ( similar to the observed sed of ngc 4051 ) .
we use the photoelectric k - shell cross section tabulated by kallman and krolik ( 1995 ) .
we carry out the calculation of the time - evolving heating - cooling balance , as described below .
we calculate with cloudy ( ferland , 1996 ) a grid of models for 300 values of u ( from 0.01 to 100 ) , and build the curve u = u(@xmath19 ) , using the technique described in kallman and krolik ( 1995 ) .
we then interpolate on these , to obtain the initial self - consistent equilibrium electron temperature of the gas .
the time evolution of the temperature in the cloud is carried out performing an iterative calculation of the time dependent photoionization equations , using the definition of @xmath78 given above .
the inputs the program needs are : ( a ) the initial equilibrium value of the ionization parameter u ; ( b ) the spectral shape of the ionizing continuum ; ( c ) the light curve of the ionizing continuum ; ( d ) the electron density @xmath93 ; ( e ) the ratio , p , between the source intensity and the intensity the source should have in order to produce the observed degree of ionization , assuming equilibrium at the beginning of the light curve .
the output of the program is a list of the relative ionic abundances of the chosen element , and the source flux at the illuminated face of the cloud as a function of time .
we neglect photoionization from the l - shell .
this affects the computation of the abundance balance when the gas is allowed to recombine to medium - to - low ionization states , with the li - like ions ( or lower ) being highly populated .
however this is not likely to be the case for warm absorbers in agn , for which the mean degree of ionization is usually very high , and the most abundant element ( c to ne ) are almost equally distributed between he - like and fully stripped ions ( see fig .
. our code can provide accurate results only at medium - high ionization regimes , those of interest for the astrophysical problem discussed in this paper .
we have also neglected iron ions .
this will likely cause an overestimation of the ovii - ix and neix - xi k - shell photoionization , since all these ions compete for the same photons . by comparing the ovii - ix
, neix - xi equilibrium abundances obtained with our method with those obtained with cloudy for u in the range 0.1 - 50 we estimate that the percentage differences are smaller than 7 % the iterative technique presented in the previous section is only an approximation to the correct self - consistent time - evolving heating - cooling calculation accounting for all the physical and dynamical heating - cooling mechanisms .
we verified that for high gas electron densities ( and therefore in a situation close to equilibrium ) this method is rather accurate . by comparing the equilibrium relative abundances of the main ions obtained with our method with that obtained using cloudy we estimate that for u in the range 0.1 - 50 the method works with a precision better than 10% . in the opposite case ,
when the density is sufficiently low , the gas remains basically in the same state and does not respond to variation of the ionizing continuum .
of course in this case our approximation is also very good , albeit uninteresting . in the intermediate cases , when the heating timescale may be longer than the typical photoionization timescale , the recombination rate coefficients may be greater than those estimated by our assumption .
we have tried to estimate the magnitude of the uncertainty on the ion abundance distribution induced by this effect by performing the time - evolving calculations fixing the temperature at a constant value equal to the initial value , and letting the ionizing intensity be free to vary by a factor of 10 . for an average gas ionization state typical of agn warm absorbers
( u = 0.1 - 50 ) , the percentage difference between the relative abundances of the main ions calculated with the iterative solution of the time dependent photoionization equations , and those calculated for a gas with a constant temperature are @xmath103 .
the difference is so small because of the rather small dymamic range spanned by the temperature in a photoionized gas with u=0.1 - 50 ( @xmath104 ) and of the weak dependence of the recombination rates on @xmath19 ( about the square root of @xmath19 , see shull and van steenberg , 1982 ) .
given the configuration of the gas we considered ( a thin , plan - parallel slab obscuring the line of sight ) , and the range of electron and column densities , the light - crossing time is smaller than both the typical source intensity variability timescales and the associated photoionization recombination timescales .
we can then neglect the effects of the light - crossing time through the cloud of photoionized gas ( see binette , 1988 ) .
the simplest case is that of a two state light curve .
let us suppose that the ionizing continuum intensity goes _ instantaneously _ from a `` low '' to a `` high '' state and comes back to the `` low '' state after a time @xmath105 ( fig .
the time behaviour of the ionization state of the gas irradiated by this continuum depends on the value of the ratio @xmath106 , and the amplitude of the flux variation ( here we adopt a factor 10 change in flux ) .
we considered the case of an optically thin cloud of gas with an initial ionization such that the most relevant ionic species of the oxygen are oviii and oix ( corresponding to equilibrium value of u@xmath107 with the adopted sed ) .
the time behavior of the relative density of @xmath108 is shown in fig .
5b . in both panels
different lines identify different values of the electron density , @xmath109 @xmath4 . in the upper panel
different @xmath93 imply four values of the distance of the cloud from the ionizing source .
the photoionization time @xmath110 ( eq .
7 ) of the gas becomes progressively longer as the density decreases : from about @xmath111 s for @xmath112 @xmath4 , to @xmath113 s for @xmath114 @xmath4 to even longer timescales for @xmath115 @xmath4 ( for @xmath116 @xmath4 the changes of @xmath108 during the firt @xmath117 s are @xmath118 % ) . note that this is not true for @xmath119 in eq .
3 , which is the definition of photoionization time usually found in litterature . formally
the dependence of @xmath14 on @xmath93 is introduced by the choice of a particular set of boundary conditions when solving eq .
6 , and hence defining @xmath92 .
physically , fixing the boundary conditions of the system means to choose a particular initial ionic distribution in the gas , and hence its initial ionization state .
different values of @xmath93 , given the initial ionization state of the gas ( and hence the ratio between the flux of ionizing photons at the illuminated face of the cloud and the electron density ) , imply different distances of the gas from the x - ray source .
this is clear in the upper panels of fig .
5 where the ionizing flux at the illuminated face of the cloud is plotted for different values of @xmath93 .
recombination times are generally longer ( @xmath120 ) , and can be order of magnitude longer .
at the highest density tested ( @xmath121 @xmath4 ) the ionization state of the gas is able to relax to the initial equilibrium state in less than @xmath117 s after the source switch off .
recombination time scales for @xmath122 @xmath4 are long , the order of many times @xmath117 s. since the source switch off is instantaneous the relative density of oix never increases after the switch off .
this case illustrates clearly how photoionization recombination timescales can have a strong effect on the changes observed in ionized absorbers and why the photoionization timescale depends on electron density . in the following section we present a more realistic light - curve , and discuss in detail the main features of our models .
we now consider a more realistic light curve . in this case
the source intensity goes from a low state to a high state in a finite time ( 4,000 s ) , and after 2,000 s comes back to the initial low state with the same absolute gradient ( fig .
the entire up & down cycle lasts 10,000 s. as in the previous case the change in flux is a factor 10 .
the corresponding light curves of the relative abundances of the fully stripped ions of three different elements , cvii , nviii and oix , are shown in fig .
6b . in both panels
different lines correspond to different values of the electron density , @xmath123 @xmath4 .
the gradual changes of the ionizing continuum produce time delays between the source light curve and the relative ion abundance light curves . in the lower density case ( @xmath112 @xmath4 solid lines )
, @xmath108 reaches its maximum value ( well below the equilibrium value @xmath2 ) around the minimum of the luminosity intensity ( at @xmath124 s ) . during
the whole decreasing luminosity phase @xmath108 is slightly increasing or constant .
x - ray spectra taken during the high luminosity phase and the decreasing luminosity phase would show an oviii absorption edge correlated with the intensity of the ionizing continuum , as seen in mgc-6 - 30 - 15 ( fabian et al . ,
1994 , reynolds et al . , 1995 ) .
at this density cvii is able to reach its maximum equilibrium value ( fig .
6b ) , but even this ion does not relax to its initial equilibrium value for many times up and down cycle time .
6b also shows that the ions reach their maximum values at different times .
@xmath125 reach its maximum value about 4,000 s before @xmath108 . at higher densities ( @xmath121 @xmath4 , dashed lines ) , the fractional abundance of each of the three ions reaches its maximum equilibrium value during the first 6,000 s , and relaxes to its minimum equilibrium value during the following @xmath126 s , but with different times - scales : oix reaches its minimum equilibrium value after a time corresponding to two cycles , about one cycle after cvii
this could help to explain why models in which all ions of all elements are in photoionization equilibrium , so often fail to describe agn spectral evolution .
spectra accumulated immediately after a very steep decreasing intensity phase could contain no significant absorption features at the energies of ovii oviii k - edge ( the oxygen being completely ionized ) , but still show a deep absorption edge at e@xmath127 kev , due to the presence of a large amount of recombined cvi - cvii in the absorbing gas .
high quality spectra would allow powerful tests of non - equilibrium photoionization models .
a general result is that the observation of any delays in the response of the absorber to flux changes on time scales of @xmath128 s , immediately implies photoionization with a density in a reasonably restricted range , @xmath129 @xmath4 ( depending on the average ionization state ) .
as an example of applicability of our models we present here the case of a rosat pspc observation of the low luminosity , rapidly variable seyfert 1 galaxy ngc4051 . both the soft and hard x - ray flux of ngc4051 vary by large factors ( up to 20 ) on a time scale of hours ( e.g. lawrence et al . , 1985 , guainazzi et al . , 1996 ) in a roughly correlated way .
the presence of an ionized absorber in this source was first proposed on the basis of variations of the ginga softness ratio correlated with the flux ( fiore et al .
rosat pspc observations , and subsequently observations with the higher resolution ccds on board of the asca satellite ( pounds et al .
1994 , mc hardy et al .
1995 , mihara et al .
1994 , and guainazzi et al .
1996 ) , also suggested the presence of a high column density , highly ionized absorber through the detection of a deep absorption edge at 0.8 - 0.9 kev .
mchardy et al .
( 1995 ) , find that a simple single zone equilibrium photoionization model can provide a reasonably good representation of spectra accumulated in 1000 - 3000 seconds , but the best fit ionization parameter does not track the source intensity , as required by the model .
this behaviour can not be explained in a single zone ionization equilibrium model and therefore ngc4051 is a good target to test our time - evolving photoionization and collisional ionization models .
we decided to compare our models to the rosat pspc data aquired on 1991 november 16 , and reported by mchardy et al .
( 1995 ) , when the source count rate showed large and rapid variations ( up to a factor of 6 in a few thousand sec .
, see figure 5 of mchardy et al . , 1995 ) .
analysis of the asca observation is deferred to a future pubblication .
the data reduction and the timing analysis were performed using the pros package in iraf .
the observation spanned 77 ksec and contained 28.7 ksec of exposure time .
ngc 4051 gave a mean count rate of 1.6 s@xmath56 .
we accumulated eight spectra ( @xmath130 ) , using a 3 radius extraction region , accumulating contiguous data with similar count rates .
background counts and spectra were accumulated from an annulus of internal and external radius of 3.5 and 6 respectively .
independent of any spectral fit the behavior of the main physical properties of the absorber can be seen in a color - color diagram . in fig .
7 we plot the hardness ratios hr = h / m against the softness ratio sr = s / m from the count rates in three bands ( s=0.1 - 0.6 kev , m=0.9 - 1.5 kev , and h=1.7 - 2.5 kev ) for theoretical curves ( for log(@xmath27)=22 , 22.5 and 23 ) obtained by folding the equilibrium photoionization models ( for log(u ) in the range @xmath1310.3 to + 1.5 , and galactic @xmath27 : 1.31@xmath132 @xmath1 , elvis et al .
1989 ) with the response matrix of the pspc .
we also plot the colors of the source in the eight spectra ( @xmath133 @xmath134 ) .
all the data points lie in a region of this diagram corresponding to the high u ends of the photoionization theoretical curves where both sr and hr decrease , as u increases , until all the ions in the gas are full stripped and the gas is completely transparent to radiation of any energy .
the colors of ngc 4051 are all consistent with the oviii - oix ions being dominant . filled circles on the log@xmath27=22.5 curve mark values of u in the range 4.07 .
all the data points are between the two extreme values of u , a factor @xmath135 change , while the source intensity varies up to a factor @xmath136 in the eight spectra .
the gas is clearly not responding to the continuum variations , a conclusion equivalent to that obtained by mchardy et al .
( 1995 ) using spectral fits with equilibrium photoionization models ( see their figure 5 ) . in the framework of the models discussed in this paper
this behavior suggests three different possibilities ( we do not take in account pure collisional ionization in this case because the observed spectral variations should be attributed to ` ad - hoc ' changes of @xmath19 or @xmath27 on time scales as short as 2000 - 4000 s ) : 1 .
the gas is far from ionization equilibrium ; 2 .
the gas has a distribution of densities ; 3 . both collisional and photoionization
processes are comparably important in the same physical region .
we investigated these possibilities in turn using detailed spectral fits . from fig .
1 we see that a given ionization state can be roughly determined by the measure of at least two consecutive ion abundances , e.g. ovii and oviii .
the measure of a single edge in fact would not distinguish between ` low ' and ` high ' ionization solutions .
the same measured feature could be produced by a lower @xmath27 , lower mean ionization gas , or by an higher @xmath27 , higher mean ionization state gas .
the best derived quantities to compare observed spectra with our several physical models are atomic edge strengths . here
we are mainly interested in the ovii and oviii because they are the strongest and therefore the easiest to detect and measure . however , with the pspc ovii and oviii edges are not individually discernable , and we must resort to model fits with multiple components . our choice is to use the components that can ensure an estimate of the ovii and oviii @xmath137 as unbiased as possible .
we fitted the eight spectra with a model consisting of a power law reduced at low energy by the galactic column density , the ovii and oviii edges and another edge at 1.36 kev to account for possible spectral complexity in the 1 - 2 kev region ( in particular the neix - x k and fe xv - xx lp absorption discussed in 2.1 ) .
five parameters were at first allowed to be free to vary : the spectral index @xmath138 , the ovii and oviii edges @xmath137 , the 1.36 kev edge @xmath137 and the model normalization . the results are presented in table 3 .
the fits with the 3 edges model are acceptable in all cases .
we stress that @xmath137(1.36 kev ) in table 3 should not be regarded as a true measure of the optical depth of ne k and fe l ions .
this feature provides only one of the possible parameterizations of the spectrum in the 1 - 2 kev range , a band in which a change in the continuum spectral index could also be present ( the 2 - 10 kev ginga and asca spectra of this source are typically flatter by @xmath139 than the pspc 0.1 - 2 kev spectra ) . as discussed in 2.1
is difficult to discriminate between ne absorption and a real spectral flattening above 1 kev with instruments of moderate spectral resolution like the pspc . we have performed a series of fit using a broken power law with break energy in the 1 - 2 kev band and the two oxygen edges , obtaining @xmath137 similar to those reported in table 3 .
we are therefore confident that the estimation of the oxygen edge @xmath137 is robust , within the rather large uncertainties given in table 3 . in principle , a way to reduce the uncertainties is to fix the continuum spectral index to a common value .
the results of this series of fits are again given in table 3 .
the uncertainies on @xmath137(ovii ) and @xmath137(1.36 kev ) are indeed smaller than in the previous case but this is not true for the @xmath137(oviii ) uncertainties .
the reason is that there is a strong anti - correlation between @xmath138 and @xmath140 .
a similar anti - correlation is present between @xmath138 and @xmath141 .
in contrast no correlation is present between @xmath138 and @xmath137(oviii ) .
this is illustrated in fig .
8 where we show the @xmath142 contours of these parameters for spectrum _
g_. since the @xmath137(ovii ) values could be biased in the fit with fixed @xmath138 by the @xmath137(ovii)-@xmath138 correlation , we prefer to compare our time - evolving models to the ovii and oviii optical depths obtained leaving @xmath138 free to vary . in all but one case ( spectrum _
d _ ) the depth of the oviii edge is higher or comparable to that of ovii , suggesting a ` high ' ionization solution ( u@xmath1434 , see fig .
1 ) , consistent with the hardness ratio analysis of fig . 7 . .*ngc
4051 : 3-edge model fits * [ cols="^,^,^,^,^,^",options="header " , ] we converted the best fit @xmath137 into ovii and oviii relative abundances assuming a solar oxygen abundance and a total hydrogen column density @xmath27 .
an indication of @xmath27 comes from the color - color diagram of fig .
although calculated using a photoionization equilibrium model , the theoretical curves in this diagram suggest a value for log@xmath27 between 22 and 23 , and so we adopt log@xmath27=22.5 .
the three panels of fig .
9 show the light curves of the source count rate ( upper panel ) and of the oviii and ovii abundances ( middle and lower panels respectively ) .
the time evolution of the ionization structure of a cloud of gas photoionized by a variable source is complex and its behaviour sometime counter - intuitive .
we then discuss first the simplest case : high electron density , for which each ion is close to its equilibrium state .
we examine next the case of lower densities and hence non - equilibrium solutions . the dotted curves on the middle and lower panels of fig .
9 show the @xmath144 @xmath4 , p=1 ( the ratio of the incident flux to that needed to produce the initial ionization distribution assuming photoionization equilibrium ) curves , when the gas is close to equilibrium .
while the equilibrium ovii curve tracks the count rate variations ( it is strictly anticorrelated with the count rate light curve ) , the oviii curve does not .
the different oviii behaviour is due to the different balance in the destruction rates of oix and oviii .
when the ionizing flux is at its maximum most of the oxygen is oix .
when the flux decreases from the maximum ( from point @xmath145 to @xmath146 ) oix recombines to oviii increasing the oviii abundance .
when the flux decreases from point @xmath147 to point @xmath148 at first oviii increases again because of the high destruction rate of oix , but after a certain point the amount of oviii recombining to ovii start to be higher than the amount of oix recombining to oviii , and so the total oviii abundance start to decrease . instead , the amount of oviii recombining to ovii is always higher than the amount of ovii recombining to ovi .
this gives rise to a different behaviour of the oviii and ovii curves in response to the same ionizing flux variations .
it is interesting to note that in this case while the dynamical range of variation of the ovii curve is larger than that of the ionizing flux , the oviii equilibrium curve shows a more compressed range of variations ( when the ionizing flux varies by a factor of 6 the oviii and ovii abundances vary by a factor of 3 and 30 respectively , see fig .
. therefore variations of oviii would be much more difficult to detect than variations of ovii , at least at these regimes of ionization .
we re - emphasize : _ the behaviour of a single edge does not provide a unique interpretation of the data . _
as explained in 2.3.3 , a grid of theoretical ovii and oviii light curves was generated using our time - evolving photoionization code for 28 values of @xmath93 from @xmath149 @xmath4 to @xmath150 @xmath4 and 15 values of p from 0.5 to 2 .
we compared these curves with the measured relative abundances of ovii and oviii and found the best fit model using a @xmath142 technique .
the thick solid lines in the lower and middle panels of fig .
9 represent the best fit non - equilibrium ovii and oviii curves .
the agreement between the best fit model and the observed ovii and oviii abundances is good : @xmath151 for 14 dof .
the best fitting values for @xmath93 and p are tightly constrained : @xmath152 @xmath4 , p@xmath153 . from the best fit @xmath93
we can estimate the distance of the absorbing cloud from the central source .
we obtain r@xmath154 cm ( 3 light days ) .
the dashed lines represent the solutions obtained using the 1 @xmath155 confidence interval on @xmath93 and p. the best fit curves ( and the 1 @xmath155 confidence intervals curves ) show a compressed dynamical range of oviii and ovii abundances variations and a delay between the source maxima and the ion abundance minima of 3000 - 6000 sec .
the compressed dynamical range is due to a mean over - ionization of the gas . while the best fit p shows that the gas in the initial point @xmath133 is near to equilibrium , it departs strongly from equilibrium during the later low intensity states ( spectra @xmath146 to @xmath134 ) .
so , despite the fact that the source spends more time in low states than in high states , the gas density is sufficiently low that the gas does not have time to fully recombine after the few events when it suffers high illumination and becomes highly ionized .
the above results were obtained assuming a total hydrogen column density of log(@xmath27)=22.5 .
assuming an higher ( lower ) column would imply a mean lower ( higher ) ovii and oviii relative ion abundance .
therefore , in principle the accurate measure of both edges would constrain also the total warm column density . the uncertainties on the pspc determinations
however preclude this possibility , and better resolution measurements are therefore needed .
the energy resolution of the asca sis , for example , is just sufficient to separate the ovii and oviii absorption edges .
a quantitative test of non - equilibrium photoionization model using the asca data and a comparison between rosat pspc and asca data of ngc 4051 is beyond the scope of this paper and will be presented in a future pubblication .
a much better separation , and therefore characterization , of the absorption features will be possible with the high resolution ( factor of 10 - 30 better than asca sis ) gratings on axaf and xmm . despite the success of time - evolving photoionization models
, alternatives do exist .
we discuss two of them in the following .
if the gas is not confined to a single cloud with constant density but rather is distributed in a region with , say , an increasing density , then different ionization equilibria could apply to different regions in the cloud .
two extreme regions may exist : in the region with higher density , lower u , collisional ionization will be the dominant ionization mechanism and the spectra transmitted by this region would show always the same features , irrespective of the source intensity ; the other region , with lower density and higher @xmath156 , is completely ionized ( for carbon , oxygen , neon and iron ions up to fexxii ) when the source is in the high state , but when the source is in a low state the abundances of oviii are sizeble and imprint the edges in the spectrum seen in low intensity spectra .
here we are assuming that the density of the photoionized region is high enough for the gas to be in instantaneous equilibrium with the ionizing intensity .
if not , the average ionization degree of the gas would be very high during the whole observation , and the gas would be always almost transparent at the energies of the relevant absorption edges .
we tested this hypothesis by fitting the highest intensity spectrum ( @xmath145 ) with a simple power law model plus a collisionally ionized absorber , using the method of fiore et al .
the best fit temperature and @xmath27 are @xmath157 k and @xmath158@xmath1 respectively .
we then used the same model ( with fixed continuum parameters , fixed temperature and @xmath27 but variable normalization ) with the inclusion of an additional oviii edge to mimic a variable ionization state , equilibrium photoionization absorber .
the fits are all acceptable .
the worst fit is that of spectrum ( @xmath146 ) ( @xmath159 , 21 dof , probability of 10.4 % ) . in fig .
10 we plot @xmath160 as a function of the time .
the dynamic range of variations of @xmath137 is here larger than that on @xmath161 in fig .
9 but is still more compressed than that predicted by equilibrium photoionization model ( solid line ) .
we can therefore exclude the possibility that a major part of this absorber could be in pure photoionization equilibrium with the ionizing intensity .
the other possibility is that both collisional and photoionization processes are important in the same physical region . in this case
the transmitted spectra are very similar to those transmitted from purely photoionized clouds of gas which much higher ionization parameter ( @xmath162 , fig .
the electron temperature of the gas is mainly determined by collisions , and is higher than that expected in pure photoionization equilibrium .
the ionization parameter u is no longer linearly correlated with the ionizing intensity , and then its dynamical range of variations is compressed by a factor @xmath163 compared to the pure photoionization case .
this could in principle explain the compression observed in the measured dynamical range of variation of @xmath161 , but could not account for the delays observed on the response of the ionization state of the gas to source intensity variation .
unfortunately there is no way to distinguish between ` hot ' photoionized absorber models and simple pure photoionization models on the only basis of the spectral analysis .
nevertheless the delay observed in the response of the ionization degree of the gas between spectrum ( b ) and ( c ) suggests that a non - equilibrium photoionization component is mainly required by data .
we have investigated ionization models for agn in different regimes of gas volume densities and photoionization states .
in particular we focussed on high gas density , low photoionization parameter gas clouds , where collisional ionization is likely to play a significant role in the gas ionization , and on low gas densities , where the photoionization may be far from equilibrium .
we presented detailed model calculations in both regimes . while the time - evolving photoionization models in 2.3 are far from being complete or exhaustive they are nevertheless instructive , and reveal the main features of these kinds of models . 1 . in many important astrophysical conditions ( ovii , oviii regime )
the fractional abundances of the most important ions of o and ne in photoionized gas are more broadly distributed in @xmath156 than those of collisionally ionized gas are in @xmath164 .
2 . in the collisional ionization case
the heavy elements show a strong inertia against becoming highly ionized , even when lighter elements , like oxygen , are almost fully stripped . in this case
the transmitted spectrum shows a large and complicated absorption structure between 1 and 3 kev , mainly due to iron l xvii - xix and neon k ix - x absorption , which is much less visible in spectra emerging from photoionized gas with similar ovii and oviii abundances .
this absorption structure may be mis - interpreted as a flattening of the spectra above @xmath2 kev , when fitting low energy resolution data with a photoionization equilibrium model .
higher spectral resolution and good @xmath32 observations are therefore needed to distinguish between collisional ionization and photoionization .
3 . in non - equilibrium photoionization models
the response of the ionization state of the gas to sudden changes of the ionizing continuum is delayed even during increasing luminosity phases . the delays increase for decreasing electron densities , as changes of @xmath93 , require changes of the intensity of the ionizing flux , i.e. changes of the distance of the gas from the x - ray source ( taking as fixed the initial ionization state of the gas ) .
the recombination timescale is generally much longer ( up to orders of magnitudes ) than the photoionization timescale , because of the dependence of @xmath92 on the equilibrium ratio @xmath165 , evaluated at the time @xmath166 .
this means that a photoionized absorber undergoing frequent , quick , and consistent changes of ionizing luminosity is likely to be overionized with respect to the equilibrium ionization , a state that would be reached only after a sufficiently long low intensity phase .
if the changes of the ionizing luminosity are not instantaneous , and the electron density of the cloud is low enough , the ionization state of the gas could continue to increase during decreasing source luminosity phases .
this means that we may measure a maximum in ionization state of a given element , when the ionizing flux is at a minimum ( opposite to what is expected in equilibrium models ) .
different ions of different elements reach their equilibrium abundance on different timescales .
this is again because of the dependence of @xmath167 on the ratio @xmath165 .
therefore in the same cloud of gas carbon could be in equilibrium while oxygen could be very far from equilibrium .
this may help in explaining why models where _ all ions of all elements _ are in photoionization equilibrium so often fails to provide a reasonable description of agn spectra and spectra evolution we have tested the above models in the case of the seyfert 1 galaxy ngc4051 .
the rosat observations of ngc4051 are not consistent with a simple equilibrium model , but can be explained straightforwardly by our time - evolving photoionization models .
the two main features in the non - equilibrium best fit models are : ( a ) the compressed range of variability of the measured ovii and oviii relative abundances with respect to the amplitude of the source variations , and to the amplitude of the variations of the abundances of these ions expected in equilibrium photoionization models ; ( b ) the 3000 - 6000 sec delay between the maximum intensity state of the source ( spectrum @xmath145 ) and the minimum of the best fit oviii abundance curve , i.e. the maximum ionization state of the gas ( spectrum @xmath168 ) . as result
we were able to estimate the gas electron density , @xmath169 @xmath4 ( assuming log@xmath27=22.5 ) and hence the distance of the ionized gas cloud from the x - ray source in r@xmath8 cm ( 3 light days ) .
we explored alternative models and we also explored alternatives which we find to be less likely ; we discuss ways to distinguish between them conclusively .
we conclude that non - equilibrium photoionization and collisional models apply to wide zones of gas density and ionization , zones which are expected in agn .
these effects must be considered in understanding ionized absorbers , and seem likely to explain otherwise puzzling behavior , without resorting to ad hoc distributions of gas .
several clear diagnostics of these models exist so that decisive tests will soon be possible .
we thank giorgio matt for useful discussions .
we also thank an anonymous referee whose comments contributed to improve the manuscript . this work was supported in part by nasa grant nag5 - 3066 ( adp ) .
f.f . acknowledges support from nasa grant nag5 - 2476 .
this work made use of the iras / pros package and the rosat archive maintained at the heasarc .
|
in this paper we explore collisional ionization and time - evolving photoionization in the , x - ray discovered , ionized absorbers in seyfert galaxies .
these absorbers show temporal changes inconsistent with simple equilibrium models .
we develop a simple code to follow the temporal evolution of non - equilibrium photoionized gas . as a result
several effects appear that are easily observable ; and which , in fact , may explain otherwise paradoxical behavior .
specifically we find that : 1 . in many important astrophysical conditions ( ovii , oviii dominant ; and high ( @xmath0@xmath1 ) column density ) pure collisional and photoionization equilibria
can be distinguished with moderate spectral resolution observations , due to a strong absorption structure between 1 and 3 kev .
this feature is due mainly to iron l xvii - xix and neon k ix - x absorption , which is much stronger in collisional models .
this absorption structure may be mis - interpreted as a flattening of the intrinsic emission spectrum above @xmath2 kev in low resolution data .
2 . in time - evolving non - equilibrium photoionization models
the response of the ionization state of the gas to sudden changes of the ionizing continuum is smoothed and delayed at low gas densities ( usually up to @xmath3 @xmath4 ) , even when the luminosity increases .
the recombination time can be much longer ( up to orders of magnitude ) than the photoionization timescale .
hence a photoionized absorber subject to frequent , quick , and consistent changes of ionizing luminosity is likely to be overionized with respect to the equilibrium ionization state .
3 .
if the changes of the ionizing luminosity are not instantaneous , and the electron density is low enough ( the limit depends on the average ionization state of the gas , but is usually @xmath5 @xmath4 to @xmath6 @xmath4 ) , the ionization state of the gas can continue to increase while the source luminosity decreases , so a maximum in the ionization state of a given element may occur during a minimum of the ionizing intensity ( the opposite of the prediction of equilibrium models ) .
4 . different ions of different elements reach their equilibrium configuration on different time - scales , so models in which all ions of all elements are in photoionization equilibrium so often fail to describe agn spectral evolution .
these properties are similar to those seen in several ionized absorbers in agn , properties which had hitherto been puzzling .
we applied these models to a high s / n rosat pspc observation of the seyfert 1 galaxy ngc 4051 .
the compressed dynamical range of variation of the ionization parameter u and the ionization delays seen in the rosat observations of ngc 4051 may be simply explained by non - equilibrium photoionization model , giving well constrained parameters : @xmath7 @xmath4 , and r@xmath8 cm ( @xmath9 light days ) .
|
the thallium nuclei , with proton number z = 81 , are situated in a transition region between the deformed prolate rare earth nuclei and the spherical lead nuclei at z = 82 .
the proton fermi level in tl lies below the z = 82 shell closure and near the 2s@xmath6 orbtal .
the ground state spin - parity of the heavier odd - a thalium isotopes , are accordingly , 1/2@xmath7 @xcite .
however , the intruder 9/2@xmath8[505 ] and 1/2@xmath8[541 ] nilsson states , originating from the @xmath9h@xmath10 orbital above z = 82 , are also available near the fermi surface for oblate and prolate deformations , respectively .
the experimental evidence comes from the observation of a low lying 9/2@xmath8 isomeric state and strongly coupled rotational bands built on this in the odd - a thalium isotopes @xcite .
the neutron fermi surface , for the tl and hg isotopes in the @xmath11 region lie near the top of the @xmath12 orbital and decoupled bands with this configuration have been identified in odd - a hg ( z = 80 ) isotopes in this region @xcite . therefore , in even - mass tl isotopes in this region , collective rotational bands based on the @xmath5 configuration are expected .
such bands have been experimentally observed in a few odd - odd tl isotopes @xcite .
but in most of the cases , there are ambiguities on level energies , spins and parities .
recently , experimental studies on the structures of @xmath13tl @xcite and @xmath14tl @xcite have been reported with definite spin - parity assignment of the @xmath5 band . in @xmath13tl , this band , which shows a low spin signature inversion ,
was reported to have an oblate structure .
whereas in @xmath14tl , the possible chiral structure associated with this band was interpreted with a triaxial deformation . therefore , in order to understand the possible transition from oblate to triaxial shape induced by the change in the neutron fermi surface of tl nuclei , it is important to chracterize the @xmath5 band systematically in an isotopic chain .
but the high spin data on the other odd - odd tl nuclei are scarce . moreover , both the protons and the neutrons occupy high - j orbitals so , different kinds of collective and single particle excitations , like magnetic rotation , are expected .
although several magnetic rotational bands have been observed in pb and hg nuclei in this region @xcite , no such bands are reported for tl isotopes .
the knowledge of high spin states in @xmath0tl was very limited prior to the present work although several superdeformed bands have been reported for this nucleus by f. azaiez et al . , but these bands were not connected with the normal deformed bands @xcite .
the normal deformed high spin states in @xmath0tl were studied by kreiner _
_ @xcite and a level scheme was obtained using two ge(li ) detectors .
although indication of rotatonal bands based on the @xmath15 configuration was reported in that work , no definite spin - parity assignments could be done .
possible observation of chiral partner bands in @xmath0tl has been claimed recently @xcite but no such level scheme has been reported till date . in this work , we have studied the @xmath4-ray spectroscopy of @xmath0tl with the aim , in particular , to compare the band structures in odd - odd tl isotopes and to search for a possible magnetic rotational band .
the @xmath4-ray spectroscopy of @xmath0tl has been studied at 14-ud barc - tifr pelletron at mumbai , india using the indian national gamma array ( inga ) .
the inga consisted of 15 clover hpge detectors with bgo anti - compton shields at the time of the experiment .
the excited states of @xmath0tl were populated by fusion evaporation reactions @xmath1re(@xmath2c , xn)@xmath0tl at the beam energy of 75 mev .
the target was a thick ( 18.5 mg/@xmath16 ) natural rhenium target .
the recoils were stopped inside the target .
the isotopic ratio of @xmath17re and @xmath18re in the natural rhenium is 37:63 .
according to the pace - iv calculations , the 4n and 5n channels are the dominant ones at the beam energies encompassed inside the target in this experiment .
the beam energy at the exit of the target was calculated to be below the coulomb barrier .
the clover detectors were arranged in six angles with 2 clovers each at @xmath1940@xmath20 and @xmath1965@xmath20 while four clovers were at 90@xmath20 and three were at -23@xmath20 angles .
the average count rate in each crystal was limited to @xmath21 4k / sec .
the clover detectors were calibrated for @xmath4 ray energies and efficiencies by using @xmath22ba and @xmath23eu radioactive sources .
recently , a digital data acquisition ( ddaq ) system , based on pixie-16 modules developed by xia llc @xcite , has been adopted for the inga .
this system has provision for the digitization of 96 channels of 24 clover detectors ( maximum number of clover detectors that can be put in inga ) with 100 mhz sampling rate .
this ddaq system was used in the present experiment for the data collection .
time stamped data were collected when at least two clovers ( compton suppressed ) were fired in coincidence .
a time window of 150 ns was set for this coincidence between the fast triggers of individual channels and the coincidence trigger was kept open for 1.5 @xmath24s .
the bgo signals from the anti - compton shields of the respective clovers were used for vetoing the individual channels .
the detailed description of the ddaq has been given in the refs .
@xcite . the data sorting routine multi parameter time - stamped based coincidence search program ( marcos ) " developed at tifr , sorts the time stamped data to generate a @xmath25-@xmath25 matrix and a @xmath25-@xmath25-@xmath25 cube in a radware compatible format for further analysis .
to construct these , a coincidence time window of 400 ns was selected .
the radware software @xcite was used for the analysis of the matrix and the cube .
the @xmath25-@xmath25 matrix contained about 2.4 @xmath26 10@xmath27 coincidence events . to construct the level scheme of @xmath0tl , the coincidence and
the intensity relations of the @xmath4 rays were used .
gates were put on the known @xmath4-ray transitions for determining the coincidence relations .
single and double gated @xmath4-ray spectra from the @xmath25-@xmath25 matrix and @xmath25-@xmath25-@xmath25 cube are shown in figs . 1 and 2 , respectively .
the relevance of these spectra in the level scheme will be discussed in the next section .
the intensity of the @xmath4 rays were obtained from the @xmath25-@xmath25 matrix using a single gated spectrum .
the multipolarities of the @xmath4-ray transitions have been determined from the angular correlation analysis using the method of directional correlation from the oriented states ( dco ) ratio , following the prescriptions of krmer - flecken et al.@xcite . for the dco ratio analysis ,
the coincidence events were sorted into an asymmetry matrix with data from the 90@xmath20 detectors ( @xmath28 ) on one axis and -23@xmath20 detectors ( @xmath29 ) on the other axis .
the dco ratios ( for the @xmath4 ray @xmath30 , gated by a @xmath4 ray @xmath31 of known multipolarity ) are obtained from the intensities of the @xmath4 rays ( i@xmath32 ) at two angles @xmath28 and @xmath29 , as @xmath33 by putting gates on the transitions with known multipolarity along the two axes of the above matrix , the dco ratios are obtained for each @xmath4 ray . for stretched transitions ,
the value of r@xmath34 would be close to unity for the same multipolarity of @xmath30 and @xmath31 . for different multipolarities and mixed transition ,
the value of r@xmath34 depends on the detector angles ( @xmath28 and @xmath29 ) and the mixing ratio ( @xmath35 ) .
the validity of the r@xmath34 measurements was checked with the known transitions in @xmath0tl and with the calculated values . in the present geometry ,
the calculated value of r@xmath34 @xcite for a pure dipole transition gated by a stretched quadrupole transition is 1.65 while for a quadrupole transition gated by a pure dipole , the calculated value is 0.61 .
these compare well with the experimental values of 1.69(3 ) and 0.58(2 ) , respectively , for the known pure dipole ( 293-kev , @xmath36 ) and stretched quadrupole ( 687-kev , @xmath37 ) transitions .
the use of clover hpge detectors allowed us to assign definite parities to the excited states from the measurement of the integrated polarization asymmetry ( ipdco ) ratio , as described in @xcite , from the parallel and perpendicular scattering of a @xmath4-ray photon inside the detector medium .
the ipdco ratio measurement gives a qualitative idea about the type of the transitions ( @xmath38 ) .
the ipdco asymmetry parameters have been deduced using the relation , @xmath39 where @xmath40 and @xmath41 are the counts for the actual compton scattered @xmath4 rays in the planes parallel and perpendicular to the reaction plane .
the correction due to the asymmetry in the array and response of the clover segments , defined by @xmath42 ) = @xmath43 , was checked using @xmath23eu and @xmath22ba sources and was found to be 1.004(1 ) ; which is close to unity , as expected .
the data of @xmath42 ) and the fitting are shown in fig .
3 . by using the fitted parameter @xmath42 ) , the @xmath44 of the @xmath4 rays in @xmath0tl have been determined .
a positive value of @xmath44 indicates an electric type transition where as a negative value favors a magnetic type transition .
the @xmath44 could not be measured for the low energy and weaker transitions .
the low energy cut off for the polarization measurement was about 200 kev , in this work .
the validity of the method of the ipdco measurements was confirmed from the known transitions in @xmath0tl .
the parallel ( @xmath40 ) and perpendicular ( @xmath42)*@xmath41 ) count for two @xmath4 rays in @xmath0tl are shown in fig .
it shows that the magnetic ( 404-kev ) and the electric ( 687-kev ) transitions can easily be identified .
the level scheme of @xmath0tl , as obtained in the present work , is shown in fig .
5 . a total of 19 new @xmath4 rays have been found and placed in the level scheme .
they are marked as * " in the level scheme of fig .
the deduced excitation energy , spin and parity of the excited levels and the multipolarity of the @xmath4 rays , together with the other relevent information concerning their placement in the proposed level scheme of @xmath0tl , are summarized in table i. [ cols="^,^,^,^,^,^,^ " , ] the 2@xmath8 ground state ( not shown in fig .
5 ) and the 7@xmath7 isomeric state , ( [email protected] min ) were known in @xmath0tl from the beta decay studies @xcite .
no @xmath4-ray transition was known to decay from the isomeric state .
the excitation energy of this state was also not known .
both the ground and the isomeric states decay by electron capture decay .
the excitation energies of the states , in the level scheme presented in fig . 5 , have been given with respect to the 7@xmath7 isomeric state as was presented in refs .
@xcite for @xmath46tl . in the present work
, the level scheme of @xmath0tl has been extended up to an excitation energy of @xmath47 mev and ( 21@xmath8)@xmath48 spin and is a much improved one compared to the previouly known level scheme reported in ref .
the proposed level scheme is based on the following arguments . a 293-kev @xmath4 ray from the 8@xmath8 to the 7@xmath7 isomeric state in @xmath0tl was known , from the systematics of the tl isotopes , to be a hindered e1 transition . the dco ratio and the positive value of the ipdco ratio , obtained in the present work ,
confirm this assignment .
a 137-kev @xmath4 ray has been observed in the present work which is in coincidence with the known 293-kev transition from the 8@xmath8 to the 7@xmath7 state as can be seen from fig .
1 ( top panel ) . a spectrum gated by this 137-kev @xmath4 line , shown in the bottom panel of fig .
1 , shows all the @xmath4 rays in @xmath0tl present in the 293-kev gated spectrum , except for the lines at 96 kev and 374 kev .
this @xmath4 ray is also observed in the double gated spectra shown in fig .
this clearly indicates that there is a 137 kev @xmath4-ray transition above the 293-kev transition and in parallel to the 96-kev transition .
it established the 334-kev level as the previously unknown energy upon which the rest of the band is built .
the dco ratio for the 137-kev @xmath4 ray has been measured in this work gated by the 687-kev @xmath4 ray which was known to be a quadrupole ( @xmath37 ) transition @xcite .
the value of this ratio comes out to be close to unity ( see table -
i ) indicating the stretched quadrupole nature of the 137-kev transition .
although , the energy of this @xmath4 ray is too low for the ipdco measurement but the m2 assignment for this low - energy transition may be completely ruled out from the lifetime consideration .
therefore , considering the 137-kev @xmath4 ray as an @xmath37 transition , spin and parity of the 430-kev level has been assigned , in this work , as 10@xmath8 .
the 374-kev @xmath4 ray was reported as a tentative one in ref.@xcite .
the double gated spectrum of 293- and 245-kev shown in fig .
2(a ) has a peak at 374 kev which confirms its placement .
the @xmath37 nature of this transition is also evident from the measured r@xmath34 value ( close to unity ) gated by a known @xmath37 transition .
the placement of the above @xmath4 rays indicates that there should be a 41-kev transition from the 9@xmath8 to the 8@xmath8 state .
this low energy highly converted ( @xmath49 ) transition has not been observed in this work as our experimental set up was not suited to detect such transitions .
it can be seen from fig .
2(a ) that all the @xmath4 rays reported in ref . @xcite
have been observed in the present work up to the 1640-kev state .
however , the 748.6- , 741.9- , 458.6- , 289.4- kev @xmath4 rays , reported in ref .
@xcite and placed above the 1640-kev state , were not observed .
instead , we have observed a cascade of 468- , 292- , 478- and 249-kev @xmath50
@xmath4 rays above the 1640-kev state along with the cross - over @xmath37 transitions .
these @xmath4 rays have been placed above the 687-kev @xmath4 ray in band b1 as they satisfy the corresponding coincidence relations .
it may be pointed out that a more efficient experimental set up was used in the present work than the earlier work . with the observation of these transitions ,
the negative parity ground band b1 has been extended up to an excitation energy of 3.13 mev and a spin of 18@xmath48 .
we have also observed two other band like structures b2 and b3 , for the first time in @xmath0tl .
they are placed above the 16@xmath8 and the 14@xmath8 states of the band b1 , respectively .
the @xmath4 lines belonging to these bands have been seen in the single and double gated spectra shown in fig . 1 and fig . 2(a ) . the spectrum in fig .
2(b ) ( double gate on 760- and 322-kev ) shows the @xmath4 rays belonging to the band b2 .
the spectrum also shows the 761-kev @xmath4 ray , indicating its double placement .
the 687-kev line is observed in this spectrum but not the 726-kev line .
the 216- and the 381-kev @xmath4 rays , belonging to this band , are also observed in this spectrum .
the double gated spectrum of 687- and 726-kev [ fig .
2(c ) ] shows the 760-kev line , slightly lower in energy than the one in fig .
therefore , the higher energy transition ( i.e 761-kev ) of the 760-kev doublet has been assigned as the linking transition between the bands b2 and b1 .
therefore , the lower energy transition ( i.e 760-kev ) of the doublet corresponds to the member of the band b1 .
the 216- , 322- and 381-kev @xmath4 rays are not observed in the spectrum of fig .
2(c ) and hence , they form the members of the band b2 .
to get an unambiguous value for the dco ratios for the 760-and 761-kev transitions , the gating transitions were selected carefully .
the dco ratio values of the doublet can not be distinguised for any choice of a pure @xmath37 gating transition below the 14@xmath8 level .
however , in the 468-kev gate , the 760-kev transition ( in the band b1 ) would be absent and an unambiguous value of dco ratio can be obtained for the 761-kev linking transition . the 468-kev transition has been found to be an @xmath51 transition with very little or no @xmath37 admixture from the negative value of its ipdco ratio .
so , the 468-kev dipole transition was chosen to obtain the dco ratio of the 761-kev linking transition and similarly , the 726-kev quadrupole transition has been chosen as the gating transition to obtain the dco ratio value of the 760-kev in - band transition .
the dco ratios of the 760- and 761-kev transitions , obtained in this way , suggest that both of them are quadrupole transitions .
the ipdco ratio , measured for the in - band 760-kev transition , clearly suggests that it is an @xmath37 transition .
the ipdco ratio could not be obtained for the 761-kev transition and an @xmath37 character for this linking transition has been assumed to assign the parity of the band b2 , tentatively , as negative .
the @xmath52 character for the 761-kev transition could not , however , be ruled out and , hence the parity of the band b2 could be positive as well .
a lifetime measurement of the 18@xmath53 state is necessary to overcome this ambiguity .
the weisskopf estimate of the halflives corresponding to @xmath37 and @xmath52 character of this transition would be @xmath54 pico - sec and @xmath55 nsec , respectively . in the present work
, we could not distinguish these two halflives .
the 876-kev @xmath4 ray has been observed in the double gated spectrum in fig .
2(a ) but not in the double gated spectrum in fig .
2(b ) or 2(c ) .
this @xmath4 ray was observed with any combination of double gates below the 1640-kev level but was not observed with any @xmath4 ray above this level .
this indicates that the 876-kev @xmath4 ray must decay to the 14@xmath8 state of the band b1 .
the cascade of @xmath4 rays in the band b3 has been observed in fig .
2(a ) . a sum of double gated spectra , constructed from the combination of the 687- , 876- , 163- , 207 & 377-kev @xmath4 rays , is shown in fig .
this spectrum clearly shows the 163- , 207- , 328- , 377- & 429-kev @xmath4 rays belonging to this band .
the @xmath37 nature of the 876-kev @xmath4 ray is evident from its dco and ipdco ratios .
therefore , the bandhead of the band b3 has been assigned as 16@xmath8 .
the ordering of the @xmath4s in this band are based on their total intensities . however , since the cross over @xmath37 transitions in this band were not observed , the ordering of the levels may be considered as somewhat tentative . the spins and parities of the states in this band
are assigned from the measured dco and the ipdco ratios of the @xmath4-ray transitions , wherever possible .
the r@xmath34 values of these transitions are very close to the expected values for pure dipole transitions .
the negative values obtained for the ipdco ratios for the 328- and the 377-kev transitions , together with their r@xmath34 values , give clear evidence that they are predominantly m1 in nature . since the other @xmath4 rays are in - band with the 328- and the 377-kev m1 transitions , so , we have assigned m1 nature for all the @xmath4 rays in this band .
the high spin band structure in @xmath0tl , obtained in the present work , can be discussed in the light of the neighboring isotopes .
ground state band structures in the odd - odd tl isotopes in @xmath11 region are similar to band b1 in @xmath0tl . @xmath13tl and @xmath14tl are the other two isotopes for which definite excitation energies , spins and parities are known for this band .
however , it is interesting that the structure of this band in these two nuclei has been interpreted differently . in @xmath13tl ,
this band was interpreted with oblate deformation @xcite while in @xmath14tl , a possible chiral structure has been reported with triaxial deformation @xcite .
we have compared different properties of the ground state band in the odd - odd tl isotopes and have shown that they look very similar .
the configurations of the first few excited states in @xmath0tl are attributed by the following arguments .
the ground state of the odd - a tl isotopes are 1/2@xmath7 corresponding to the 3@xmath56 orbital near the proton fermi level . in the isotone @xmath57hg ,
the experimentally observed 5/2@xmath8 and 13/2@xmath7 states lie very close ( within 40- and 141-kev , respectively ) to the 3/2@xmath8 ground state @xcite .
these states correspond to the @xmath58 , @xmath59 and @xmath60 orbitals , respectively , near the neutron fermi level .
correspondingly , the configurations of the ground state ( 2@xmath8 ) and the 7@xmath7 isomeric state in the odd - odd @xmath0tl nucleus have been attributed to @xmath61 and @xmath62 configurations , respectively .
the 8@xmath8 state , at the excitation energy of 293-kev , in this nucleus , has been interpreted as the band head of the @xmath5 configuration @xcite .
these orbitals lie near the proton and the neutron fermi levels of @xmath0tl for an oblate deformation .
the excitation energies of the 9@xmath8 , 10@xmath8 and 11@xmath8 states in this band for @xmath63tl nuclei are shown in fig .
6 , with respect to the excitation energy of the 7@xmath7 isomeric state .
it can be seen that the energies of these states increase smoothly with mass number a , as expected . the alignment ( @xmath64 ) for the band b1 in @xmath0tl is shown in fig .
7 as a function of the rotational frequency @xmath65 .
the alignments for the neighboring odd - a nuclei @xmath57tl and @xmath57hg are also shown in fig .
7 which have been deduced from the corresponding level schemes reported by reviol et al .
@xcite and hubel et al .
@xcite , respectively . it can be seen that the band in @xmath0tl has an initial alignment of about 5@xmath48 .
this value is in good agreement with the value obtained from the initial alignments of the neighboring odd - a nuclei using the additivity rule @xcite .
there is a gain in alignment of about 5@xmath48 at rotational frequency of @xmath66 mev which is similar to @xmath13tl and @xmath14tl .
the similarity between these odd - odd tl isotopes are also reflected in their kinetic moments of inertia ( j@xmath67 ) values as shown in fig
this indicates similar deformation in these nuclei .
the particle alignments in these isotopes are also taking place at about the same frequency . in the present work
, the band b1 in @xmath0tl could be extended just beyond the band crossing at @xmath68 mev .
the band crossing phenomena in the odd - odd tl isotopes can be understood from the band crossings in the even - even core of hg isotopes and the band crossings in the odd - a tl isotopes . in @xmath69hg @xcite , the first two
observed band crossings , at @xmath70 mev and at @xmath71 mev , were interpreted as due to the alignments of the @xmath59 neutrons ( g @xmath72 ab and ab @xmath72 abcd crossing , following the nomenclature of hubel _ et al . _
the alignment of the protons takes place at even higher frequencies .
the band crossings in the odd - z @xmath57tl nucleus , in which the first proton crossing is blocked , were also interpreted as due to the alignments of the same @xmath59 neutrons .
the crossing frequency of @xmath68 mev in @xmath0tl agrees well with the @xmath73 value in @xmath69hg .
therefore , the observed band crossing in doubly - odd @xmath0tl may be attributed to the alignment of a pair of neutrons in the @xmath59 orbital .
the similarity and the systematic trend of the observed ab @xmath72 abcd crossing frequencies for @xmath13hg ( @xmath74 mev ) , @xmath69hg ( @xmath71 mev ) and @xmath0hg ( @xmath75 mev ) @xcite are also in good agreement with the crossing frequencies of @xmath69tl ( @xmath76 mev ) , @xmath0tl ( @xmath68 mev ) and @xmath77tl ( @xmath78 mev ) .
a low spin signature inversion has been observed in the rotational bands in odd - odd nuclei in various mass regions , involving high - j configurations . in particular , the signature inversion has been reported for the prolate deformed odd - odd nuclei in the rare earth region with a @xmath5 configuration @xcite . in the a @xmath21 190 region ,
the low - spin signature inversion for a band associated with an oblate @xmath5 configuration has been reported for the first time in @xmath13tl @xcite . in the present work
, a similar signature inversion has also been observed in @xmath0tl .
the signature inversion in a band can be identified experimentally in the plot of energy staggering , defined by @xmath79/2i$ ] , where @xmath80 is the energy of the state with spin @xmath81 , as a function of the spin .
the same has been plotted in fig . 9 for the negative parity yrast band of @xmath0tl along with those for @xmath46tl .
the staggering plots of these three nuclei show remarkably similar behavior with low - spin signature inversion at the same spin value of 11@xmath48 .
a j - dependent residual @xmath82 interaction was used to interpret the signature inversion in @xmath13tl @xcite .
the similarity in the staggering plots suggest that the residual @xmath82 interaction remains almost unchanged for the heavier odd - odd isotopes of @xmath83tl .
the similarities in the band crossing frequencies , the moments of inertia , the signature inversion and the staggering plots for the above three nuclei suggest that the high spin properties of the ground state band in all the odd - odd tl isotopes in @xmath11 region are very similar and therefore , they are expected to have similar structure . in the present work ,
we have not observed any indication of a chiral side band in @xmath0tl unlike in @xmath14tl .
the band b3 in @xmath0tl is a negative parity band with band head spin of 16@xmath48 at an excitation energy of 2.5 mev .
this band has been extended up to 21@xmath48 with predominnatly @xmath51 transitions .
no @xmath37 cross over transitions could be observed in this band .
the band b3 lies at an energy which is somewhat higher than the crossing of the 2- and 4-quasiparticle configurations of the band b1 .
therefore , the band b3 seems to be a 6-quasiparticle band with two more protons in the @xmath9h@xmath10 and @xmath9s@xmath6 orbitals .
the configuration of the band b3 could , therefore , be assigned as @xmath84 .
the respective proton and neutron orbitals constitute the 2-quasiparticle states in this nucleus at the lower excitation energies ( 2@xmath8 state , 7@xmath7 state and the 8@xmath8 band head of the band b1 ) .
the experimental initial aligned angular momentum of 7@xmath48 , deduced for this band , is consistent with the above configuration .
the particle - hole configuration of this band with proton particles in high - j , high-@xmath85 and neutron holes in high - j , low-@xmath85 orbitals ( for an oblate deformation ) is favorable for magnetic rotation ( mr ) and therefore , the excited states , in this band , may be generated by the shear mechanism .
this band , with 3 particle - hole pairs , seems to follow the same general features of mr bands @xcite . for the mr bands , the level energies ( @xmath86 ) and the spin ( @xmath81 ) in the band follow the pattern @xmath87 where , e@xmath88 and i@xmath88 are the energy and the spin of the band head , respectively . the plot of @xmath89 vs. @xmath90 for this band is shown in fig .
the solid line is the fit of the data using the above relation .
the good agreement of the data with the fitted curve , in this plot , clearly indicates that the band b3 closely follows the above relation .
considering a higher limit of the intensity of the unobserved crossover @xmath37 transitions as the level of the background in our data , the lower limit of the @xmath91 ratio has been estimated to be @xmath9228 @xmath93 .
this compares well with the typical value of @xmath9420 @xmath93 for a mr band .
the dynamic moment of inertia j@xmath95
@xmath96 mev@xmath97 obtained for this band is also within the typical value of j@xmath98 @xmath96 mev@xmath97 for an mr band .
all these indicate that the band b3 is , most likely , an mr band .
the band like structure b2 lies at an excitation energy of 3.16 mev with the band head spin of 18@xmath48 .
the parity of this band remains tentative and this band is also not well developed .
therefore , characterization of this band at this moment would be rather prematured .
however , from the systematics of the odd - a tl nuclei , it appears that this band may be generated with an additional proton pair in the i@xmath99 orbital with the 4-quasiparticle configuration of the band b1 above its band crossing . in order to get an idea about the shape of the tl nuclei for different configurations , the total routhian surface ( trs )
calculations have been performed , in the @xmath100 deformation mesh points , where @xmath101 and @xmath4 are the quadrupole deformation parameters , with minimization in the hexadcapole deformation @xmath102 , for the bands b1 and b3 in @xmath0tl .
the hartee - fock - bogoliubov code of nazarewicz _ et al . _
@xcite was used for the calculations .
a deformed woods - saxon potential and pairing interaction were used with the strutinsky shell corrections method .
the procedure has been outlined in refs .
the contour plots of the trss are shown in fig .
11 for the 2-quasiparticle @xmath5 configuration corresponding to the band b1 ( top panel ) and for the 6-quasiparticle @xmath84 configuration corresponding to band b3 ( bottom panel ) in @xmath0tl .
these surfaces were calculated at the rotational frequencies of @xmath103 mev and 0.16 mev , respectively .
the plots clearly show minima in the trs at an oblate deformation ( @xmath104 is prolate and @xmath105 is oblate ) with @xmath106 and @xmath107 for the band b1 and a near spherical shape with @xmath108 and @xmath109 for the band b3 .
the surfaces calculated for the same 2-quasiparticle configuration as in band b1 are shown in fig .
12 for @xmath77tl and in @xmath14tl .
the minima for these nuclei are also found to be at the oblate deformation with very similar @xmath101 values .
the similarity in the calculated shapes for these nuclei corroborates well with the observed similarities in the properties of the @xmath110 bands in odd - odd tl nuclei .
the oblate shapes in all these three nuclei are predicted to persist over a rotational frequency range up to the band crossing .
after the band crossing , the shape of @xmath0tl remains almost the same with deformation @xmath111 and @xmath112 .
the lack of quadrupole deformation calculated for the band b3 in @xmath0tl is consistent with the non - observation of @xmath37 transitions and the conjecture of mr nature of this band .
the calculated quasiparticle routhians for the protons ( z = 81 ) and the neutrons ( n = 113 ) in @xmath0tl are shown in left and right panels , respectively , of fig . 13 for the deformation parameters @xmath113 and @xmath114 corresponding to the minimum in the trs in fig . 11 .
these calculations predict the first neutron pair alignment ( ab crossing ) at around @xmath115 mev and a second one ( cd crossing ) around @xmath116 mev .
these are in good agreement with the values obtained using similar calculations for @xmath57tl by reviol et al . @xcite . the same nomenclature as that of ref .
@xcite has been used in our calculations to label the quasiparticle levels .
the observed band crossings at @xmath117 mev and at @xmath118 mev in @xmath57tl were identified as due to the first and second neutron pair alignments by reviol et al .
the calculated crossing frequencies , however , are found to underpredict both the experimental crossing frequencies by the same amount and this discrepancy was attributed to the fact that the core deformation is not stiff as assumed in these calculations .
in addition to the first proton pair alignment , as in @xmath57tl , the first neutron pair alignment is also blocked in the odd - odd nucleus @xmath0tl . therefore , the experimental band crossing in @xmath0tl , observed at a rotational frequency of @xmath119 mev , is due to the second neutron pair alignment .
this value of the crossing frequency is in fairly good agreement with the observed second crossing in @xmath57tl but at a slightly lower frequency .
the experimental crossing frequency in @xmath0tl is lowered by 0.03 mev compared to the second crossing in @xmath57tl .
this difference of the experimental crossing frequencies , in these two nuclei , is reproduced well in the calculated second neutron pair alignment frequencies for @xmath0tl ( @xmath116 mev ) obtained in our calculation , and that for the @xmath57tl ( @xmath120 mev ) as obtained by reviol et al .
@xcite .
to investigate the shears mechanism of the band b3 , it has been studied by the semiclassical approach of macchiavelli _ et al .
this is based on a schematic model of the coupling of two long j vectors , ( @xmath121 and @xmath122 ) , corresponding to the proton and the neutron parts of the angular momenta .
this aims at extracting the information on the effective interaction between the nucleons which are involved in the shears mechanism . in this model ,
the shears angle ( @xmath123 ) , between the two j vectors , is an important variable which can be derived using the equation @xmath124 where , @xmath81 is the total angular momentum . considering the @xmath125 and @xmath126 values of 8.5@xmath48 and 13.5@xmath48 , respectively , for the proposed @xmath127 configuration of the band b3
, the band head spin is calculated to be 16 @xmath48 , assuming perpendicular coupling .
this is in excellent agreement with the observed spin of the band head of this band .
the maximum spin for this configuration has been calculated as @xmath128 corresponding to @xmath129 , which is again consistent with the highest spin observed for this band in the present work .
therefore , the angular momentum along the entire range of the band is , most likely , generated through the shears mechanism .
the good agreement of the initial and the final spin values of this band with the calculated ones imply that the spins are generated solely by the shears mechanism with very little or no contribution from the rotation of the core .
this fact is , again , corroborated well by the small quadrupole deformation obtained for this band in the trs calculations .
according to the prescription of macchiavelli _ et al . _
@xcite , the excitation energies of the states in shears bands correspond to the change in the potential energy because of the recoupling of the angular momenta of the shears .
the excitation energies of the states in the band , with respect to the band head energy , can be written as : @xmath130 where , @xmath131 is the energy of the level with angular momentum @xmath81 , @xmath132 is the corresponding shears angle as given in eqn .
3 , @xmath133 is the band head energy and @xmath134 is the strength of the interaction between the blades of the shears .
therefore , @xmath134 can be calculated by using the experimentally observed energy levels of the shears band . in fig .
14 , the @xmath135 is plotted as a function of @xmath123 .
@xmath134 has been extracted from this plot by a fit of eqn .
the fitted curve is shown as the solid line in fig .
the extracted value of @xmath134 comes out to be 1661 kev , which corresponds to an effective interaction of 332 kev per particle - hole pair for the suggested configuration of the band b3 .
this value of the effective interaction is in good agreement with the typical value of @xmath136 kev observed for the mr bands in pb nuclei in this region @xcite .
the @xmath4-ray spectroscopy of the high spin states in the odd - odd nucleus @xmath0tl has been studied in the fusion evaporation reaction using @xmath1re target with @xmath2c beam at 75 mev .
a new and improved level scheme of @xmath0tl is presented in this work which includes 19 new @xmath4-ray transitions . the dco ratio and the polarization asymmetry ratio measurements
have been performed to assign the spins and parities of the levels . the @xmath5 band ( b1 ) in this nucleus has been extended just beyond the band crossing and up to the 18@xmath8 @xmath48 of spin .
the uncertainties in the excitation energies and the spins in this band have been removed .
the moment of inertia , alignment , the energy staggering and signature invesrion of this band have been found to be very similar to its neighoring odd - odd isotopes .
the trs calculations predict similar structure for these nuclei and support the observed similarity .
we have not found any indication of a chiral doublet band structure in @xmath0tl , unlike that which was reported for @xmath14tl .
the observed band crossing in @xmath0tl could also be understood from the band crossing in the neighboring odd - a nuclei with oblate deformation .
therefore , it needs further experimental and theoretical investigation to understand the possible change in the structure of this 2-quaiparticle band in @xmath14tl , if any .
two new side bands ( b2 and b3 ) have been observed in this work , for the _ first time _ , in @xmath0tl .
6-quasiparticle configurations for these bands have been suggested .
a near spherical shape is predicted by the trs calculations for the band b3 and is consistent with the suggested mr nature of this band .
this band has been discussed in the frame work of the semiclassical approach .
the observed band head spin and the range of the spin values of this band are in agreement with such calculations .
however , to get a microscopic understanding of this band , tilted axis cranking calculations are needed .
the untiring effort of the operators and support staff at barc - tifr pelletron are acknowledged for providing a good beam of @xmath2c .
the authors gratefully acknowledge the efforts of profs .
r.g . pillay and v. nanal for the smooth running of the experiment and illuminating discussion .
the help of the inga community is gratefully acknowledged to set up the array and its associated electronics .
99 r.m . diamond and f.s .
stephens , nucl .
phys . * a 45 * , 632 ( 1963 ) .
v.t . gritsyna and h.h .
foster , nucl .
phys . * a 61 * , 129 ( 1965 ) .
newton , s.d .
cirilov , f.s .
stephens and r.m .
diamond nucl . phys . * a 148 * , 593 ( 1970 ) .
newton , f.s .
stephens and r.m .
. phys . * a 236 * , 225 ( 1974 ) .
lieder et al . , nucl . phys . * a 299 * , 255 ( 1978 ) .
kreiner et al . , phys .
* c 38 * , 2674 ( 1988 ) .
porquet et al . , phys . rev .
* c 44 * , 2445 ( 1991 ) .
w. reviol et al . , phys .
* t 56 * , 167 ( 1995 ) .
bearden et al .
phys . * a 576 * , 441 ( 1994 ) .
h. hbel et al . , nucl . phys . * a 453 * , 316 ( 1986 )
xie et al . , phys . rev .
* c 72 * , 044302 ( 2005 ) .
kreiner et al . , phys . rev .
* c 21 * , 933 ( 1980 ) .
kreiner et al . , phys . rev .
* c 20 * , 2205 ( 1979 ) .
kreiner et al . , nucl . phys . * a 308 * , 147 ( 1978 ) .
kreiner et al . , nucl . phys . * a 282 * , 243 ( 1977 ) .
kreiner et al . , phys . rev .
* c 23 * , 748 ( 1981 ) .
lawrie et al . , phys . rev .
* c 78 * , 021305(r ) ( 2008 ) .
h. hbel , prog .
. phys . * 453 * , 1 ( 2005 ) .
clark and a.o .
macchiavelli , annu .
nucl . part .
sci . * 50 * , 1 ( 2000 ) .
f. azaiez et al . , z. phys . * a 336 * , 243 ( 1990 ) .
f. azaiez et al . , phys .
lett . * 66 * , 1030 ( 1991 ) .
masiteng et al . , acta phys .
b 40 * , 657 ( 2009 ) . h. tan , et al . , nuclear science symposium conference record , nss 08 , ieee , p 3196 ( 2008 ) .
r. palit et al .
inst . meth .
. res . * a 680 * 90 ( 2012 ) r. palit , aip conf .
* 1336 * , 573 ( 2011 ) .
d. c. radford , nucl .
methods phys .
res . * a 361 * , 297 ( 1995 ) .
a. krmer - flecken et al . , nucl .
methods phys .
res . * a 275 * , 333 ( 1989 ) .
macias et al . ,
computer phys .
comm * 11 * , 75 ( 1976 ) .
k. starosta et al . , nucl . instrum .
. res . * a 423 * , 16 ( 1999 ) .
droste et al . , nucl .
instrum . meth .
. res . * a 378 * , 518 ( 1996 ) .
balraj singh , nucl .
data sheets * 107 * , 1531 ( 2006 ) .
e. achterberg , et al . ,
data sheets * 107 * , 1 ( 2006 ) w. reviol et al . , nucl
. phys . * a 548 * , 331 ( 1992 ) .
r. bengtsson and s. frauendorf , nucl . phys . * a 327 * , 139 ( 1979 ) .
r. bengtsson and s. frauendorf , nucl
. phys . * a 314 * , 27 ( 1979 ) .
i. hamamoto , phys .
* b 235 * , 221 ( 1990 ) .
jain et al .
, pramana * 75 * , 51 ( 2010 ) . w. nazarewicz , et al . ,
phys . * a 512 * 61 ( 1990 ) .
w. nazarewicz , et al . ,
phys . * a 435 * 397 ( 1985 ) .
g. mukherjee , et al .
* c 64 * 034316 ( 2001 ) .
g. mukherjee , et al .
, nucl . phys . * a 829 * 137 ( 2009 ) .
a.o . macchiavelli _
et al . _ ,
c * 57 * , r1073 ( 1998 ) .
a.o . macchiavelli _
et al . _ , phys . rev .
c * 58 * , r621 ( 1998 ) .
clark and a.o .
macchiavelli , nucl . phys . *
a682 * , 415c ( 2001 ) .
|
the high - spin states in odd - odd @xmath0tl nucleus have been studied by populating them using the @xmath1re(@xmath2c , xn ) reactions at 75 mev of beam energy .
a @xmath3 coincidence measurement has been performed using the inga array with a digital data acquisition system to record the time stamped data .
definite spin - parity assignment of the levels was made from the dco ratio and the ipdco ratio measurements .
the level scheme of @xmath0tl has been extended up to 4.1 mev in excitation energy including 19 new @xmath4-ray transitions .
the @xmath5 band , in the neighboring odd - odd tl isotopes show very similar properties in both experimental observables and calculated shapes .
two new band structures , with 6-quasiparticle configuration , have been observed for the first time in @xmath0tl .
one of these bands has the characteristics of a magnetic rotational band . cranked shell model calculations , using a deformed woods - saxon potential ,
have been performed to obtain the total routhian surfaces in order to study the shapes of the bands and the band crossing in @xmath0tl .
the semiclassical formalism has been used to describe the magnetic rotational band .
|
first , a little theoretical background .
the question of flavor - changing neutral currents ( fcncs ) mediated by higgs bosons was first addressed two decades ago by glashow and weinberg @xcite .
it had always been obvious that the higgs boson of the _ minimal _ standard model could not have flavor - violating couplings since the couplings of fermions to the higgs field _ defines _ the fermion mass eigenstates and thus also defines our notion of flavor .
but in models with two ( or more ) higgs fields , the fermion mass eigenbasis can be different from the higgs interaction eigenbasis and thus higgs - mediated fcncs can occur . in order to avoid large fcncs inconsistent with experiment , the authors of ref .
@xcite proposed several solutions .
one of those solutions is the `` type ii '' two - higgs - doublet model : one higgs field ( @xmath8 ) couples only to up - type quarks , while the other ( @xmath9 ) couples only to the down - type : @xmath10 .
this guarantees the alignment of the higgs boson interaction eigenbasis with the fermion mass eigenbasis .
such a structure can be protected from quantum corrections by any number of discrete symmetries under which the two higgs fields transform differently .
the mssm possesses the structure of a type - ii model classically .
but it possesses no discrete symmetry which can protect this structure ; all such symmetries are broken by the @xmath11-term in the superpotential , which must be present to avoid disagreement with experiment . and though the type - ii structure is also protected by holomorphy of the superpotential , this too is ineffective after susy is broken .
thus new terms are generated in the low - energy lagrangian of the mssm of the form @xmath12 .
whether such terms will lead immediately to fcncs depends on the structure of the new @xmath13 yukawa matrices @xcite . in ref .
@xcite it was shown that there are contributions to @xmath14 which _ are _ flavor - violating and can have important effects at large @xmath1 .
it is actually quite easy to see why .
consider the diagram in fig .
1(a ) in which charged higgsinos propagate inside the loop .
if we work in a basis in which the down quarks couple diagonally to @xmath15 , then the up quarks have off - diagonal couplings to @xmath16 proportional to ckm elements .
in particular , the coupling of @xmath15 to @xmath17 is just the bottom yukawa coupling , @xmath18 .
but the coupling of @xmath16 to @xmath19 is then @xmath20 where @xmath21 is the top - quark yukawa coupling .
thus the diagram generates a new interaction @xmath22 with coefficient proportional to @xmath23 .
we can rewrite this coefficient as simply @xmath24 , where @xmath25 includes not only @xmath20 but also the loop kinematic and suppression factors .
( we ignore the phases induced by the sparticles in the loops ; work including them is under way @xcite . )
= 1.25 in when @xmath8 gets a vev ( @xmath26 ) , it contributes an off - diagonal piece to the fermion mass matrix : = ( cc |s_r & |b_r ) ( cc m_s & 0 + y_b v_u & m_b ) ( c s_l + b_l ) .
this lagrangian is written in the higgs interaction eigenbasis of the fermions , which at tree - level is also the mass eigenbasis ; we drop the first generation for simplicity .
the mass matrix can be diagonalized by a biunitary transformation which mixes the @xmath27 and @xmath28 by an angle @xmath29 .
but since @xmath30 , we have @xmath31 .
so although @xmath25 is one - loop suppressed , the factor of @xmath1 can allow @xmath32 to be @xmath33 . at tree - level ,
a @xmath34higgs transition is possible , with the higgs decaying to leptons , as in fig .
2 . this occurs by exchange of @xmath9 .
however the coupling of the @xmath35-quark sector to @xmath8 has shifted the @xmath35-quark interaction eigenstates away from their mass eigenstates . in order to replace the interaction eigenstates on the external legs with mass eigenstates , we must replace b_l
b_l= b_l + s_lwhich induces a @xmath36 transition through an @xmath9 higgs , fig .
then the flavor - changing amplitude @xmath37 is related to the flavor conserving amplitude @xmath38 by ^|bs ^|bb .
there are several noteworthy properties of the higgs - mediated fcncs .
first , the amplitude for @xmath2 scales as @xmath39 at large @xmath1 .
one factor of @xmath1 come from @xmath32 , the other two come from the @xmath40- and @xmath11-yukawa couplings which scale as @xmath41 .
thus the branching ratio scales as @xmath6 and provides an incredibly powerful tool for constraining @xmath1 .
second , at large @xmath1 , the @xmath9 higgs doublet is essentially decoupled from the electroweak symmetry breaking .
it contains the physical states @xmath42 , @xmath43 and @xmath44 which all have roughly equal masses . but because higgs flavor changing must disappear in a model with only one higgs doublet , the fcnc branching ratios must decouple as @xmath45 .
but this does _ not _ mean that the effects decouple as the susy mass scale increases .
rather , in the limit that all supersymmetric masses are taken heavy , while @xmath43 ( or @xmath9 ) remains light , the rate for higgs fcncs approachs a finite constant . what diagrams contribute to the off - diagonal higgs couplings ?
first , even if all flavor violation stems from the ckm matrix alone , there is the one - loop higgsino diagram of fig .
1(a ) already considered .
models with only this ckm - induced flavor violation are known as `` minimal flavor violation '' ( mfv ) models .
note that this kind of flavor violation has nothing to do with the `` susy flavor problem . ''
it is always present in susy because it is generated by the flavor violation in the ckm matrix and can not be eliminated simply by making superpartners heavy or by aligning quark / squark mass matrices .
= 5.25truein more generally there are also gluino loop diagrams contributing to the same interactions ( figure 1(b ) ) , if flavor - violating @xmath46 or @xmath47 squark mass insertions are non - zero .
such an insertion can be generated in minimal models by renormalization group running of the squark mass matrices , or may appear in the underlying theory .
we make no assumptions in our general analysis about the underlying source of this flavor - changing in the squark sector , and we refer to this as general flavor violation .
because this new source of fcncs emerges from the higgs sector , it preferentially generates processes involving heavy fermions .
thus the channels in which higgs - mediated fcncs can most easily be observed are @xmath48 and @xmath49 .
the channels with final state muons are suppressed relative to those with final state taus by @xmath50 but are much cleaner experimentally .
the channels with initial state @xmath51 are suppressed relative to @xmath52 by @xmath53 .
thus a machine which produces an abundance of @xmath52 and can cleanly tag and measure muons is ideal for studying this physics .
this machine is precisely the tevatron , either in cdf or d0 .
the current limits on @xmath54 are provided by cdf from run i at @xmath55 and @xmath56 respectively . given enough luminosity , the @xmath57-factories may be able to corroborate any tevatron discovery using @xmath58 or , if a suitable technique is found , @xmath59 .
if cdf and d0 each record even @xmath60 of data during run ii of the tevatron , now underway , they can probe the @xmath2 branching fraction below the @xmath61 level .
( see , , ref .
@xcite for a more detailed study of the cdf capabilities . ) we will for illustration assume that the tevatron can discover a signal in @xmath7 if its branching ratio is greater than @xmath61 . because the branching ratios scale as high powers of the input parameters ( @xmath45 and @xmath6 ) , small changes in the tevatron capabilities will generate infinitessimal changes in our results . and
once a signal is discovered , the numerical analysis can be redone to obtain more precise limits .
we demonstrate below that there is a minimum value of @xmath1 consistent with a signal at the @xmath61 level , and calculate it .
we also consider the interplay between the pseudoscalar mass and @xmath1 , a result which could be important in searches for the additional higgs bosons at future colliders . both of these analyses will be done in full generality within the mssm .
we will not constrain ourselves initially to any particular class of mssm models , such as supergravity- or anomaly - mediated susy - breaking .
this can be done because there are only a limited number of parameters on which the branching ratio will depend and these few parameters can be studied without further simplifications or assumptions . however in particular classes of models the constraints on @xmath1 and @xmath62 are stronger and therefore observation of a signal provides even more information .
we will consider specific susy models after doing the general analysis and we will find bounds on @xmath1 much stronger than in the general case .
regardless of any details of susy - breaking , the higgsino contribution of fig .
1(a ) to higgs - mediated flavor - changing processes must be present .
first we will consider the case where _ only _ this contribution is present and then generalize in the next section .
thus we begin by considering the mfv models .
the flavor - changing contribution of the higgsino loop diagram is encoded in a new dimensionless parameter @xmath63 such that ( b_s ) _
qcd^2 m_b_s^3 f_b_s^2 _
b_s m_b^2 m_^2 ( ) ( ) .
[ eq : br ] here @xmath64 is the qcd correction due to running between the susy and @xmath52 scales , and we take @xmath65 .
note that the fraction @xmath66 approaches @xmath6 at large @xmath1 .
the parameter @xmath63 is calculated to be @xcite _
= -a_t f(^2 , m_t_l^2 , m_t_r^2 ) where @xmath67 and @xmath68 are the left- and right - handed top squark masses , @xmath11 is the superpotential higgs mass parameter , @xmath69 is the top - squark trilinear term , and the function @xmath70 is defined in ref .
the important thing to know about @xmath70 is that it is positive definite , symmetric in its inputs and @xmath71 up to a constant of @xmath33 .
in particular , @xmath72 and @xmath73 .
in order to maximize the function @xmath63 ( and therefore maximize the branching fraction one can obtain for a given @xmath74 ) we need only consider the four parameters @xmath75 , @xmath76 , @xmath11 and @xmath69 .
this is easiest to do by considering the limits in which @xmath63 could become large .
first consider the limit in which @xmath11 is much larger than the squark masses .
then @xmath77 since @xmath11 appears both in @xmath70 and as a prefactor .
thus it appears that @xmath63 can grow unabated as @xmath69 becomes large .
such a runaway behavior would generate arbitrarily large branching ratios and invalidate our claim to a bound on @xmath1 .
however , there is a `` cosmological '' limit to @xmath69 : as @xmath69 increases far beyond the squark masses , the usual electroweak vacuum become unstable with the true vacuum breaking qed and qcd @xcite .
we apply this constraint by applying the famous condition a_t^2 < 3(m^2_t_l+m^2_t_r+m_2 ^ 2 ) .
( the last term , @xmath78 , is the mass parameter for @xmath8 appearing in the higgs potential .
since @xmath79 is necessary to break the electroweak symmetry , we can maximize @xmath69 by setting @xmath80 . )
because the squark masses are much smaller than @xmath11 in this limit , then @xmath81 can not become much larger than unity at best .
a similar argument holds for large @xmath69 but with @xmath11 smaller than the squark masses , except now @xmath63 is more highly suppressed , @xmath82 .
the actual limit is obtained when @xmath11 becomes large along with one of the squark masses ( with the other as small as possible )
. then @xmath83 . but
the qed / qcd - breaking constraint still limits @xmath69 , though now it becomes @xmath84 so that @xmath85 which implies that _ < . plugging into eq .
( [ eq : br ] ) immediately gives ( b_s ) < .
thus for @xmath86 ( approximately its current lower limit ) and a branching ratio greater than @xmath61 , we deduce in an mfv scenario that @xmath87 .
as @xmath62 increases , the bound on @xmath1 also increases rapidly ; for example , if @xmath88 , one must have @xmath89 . in general , > 11()^23 ^16 [ eq : mfv ] where we have taken @xmath90 for large @xmath1 . in figure [ mfvfig ]
, we plot contours representing the maximal value of the @xmath2 branching ratio consistent with a given choice of @xmath62 and @xmath1 . given a measured value of the branching ratio , only the region above the line is consistent in an mfv scenario .
in the case of mfv , only the higgsino loop has any significant effect in producing flavor violation .
however once one moves away from truly minimal flavor violation , a new source arises . in a truly general scenario ,
the @xmath91-@xmath92-@xmath93 couplings can be flavor changing . in the mass insertion approximation ,
these violations are moved to the squark propagators where they appear as bilinear interactions which mix flavors and even chiralities of the squarks .
the size of the insertions is parametrized by a dimensionless quantity , @xmath94 where @xmath95 are either @xmath96 or @xmath97 , and @xmath98 label the squark generation .
we are only interested in the down - sector @xmath99 and @xmath100 insertions for this work . at present
there are no strong experimental bounds on these insertions and so it is possible that the @xmath101 and @xmath102 ( or @xmath103 and @xmath104 ) are maximimally mixed , or @xmath105 .
of course , in particular models a large @xmath106 could lead to large @xmath107 or @xmath108 , but constraints from these processes are sensitive to @xmath109 while @xmath7 is sensitive to @xmath110 alone .
since we allow very large flavor - mixing , we must take care to note that the mass insertion approximation breaks down as @xmath111 approaches one ; we will discuss this issue shortly .
the new gluino - induced source of flavor changing in general models generates new diagrams in @xmath2 with @xmath112 and @xmath113 squarks and gluinos in the loop ; see fig .
1(b ) . for our purposes
here we can ignore the higgsino contributions since their maximal size is smaller than that of the gluino loop if @xmath114 .
we will also assume that only one of the two diagrams in fig .
1(b ) dominates the branching ratio ; when there is data , a fully combined analysis should be done .
specifically we will choose the @xmath99 diagram to dominate , though our results are identical if the @xmath100 diagram dominates instead .
we will discuss more realistic scenarios at the end of the section .
we parametrize the gluino contribution as _ = _
ll^23 m_3 f_2(m^2__l , m^2__l , m^2_,m_3 ^ 2 ) [ eq : gluino ] where the function @xmath115 is defined in ref . @xcite . the branching ratio is calculated from @xmath116 just as for @xmath63 in eq .
( [ eq : br ] ) .
notice in particular the lack of a @xmath117 suppression in @xmath116 ; this will allow much smaller values of @xmath1 to be consistent with an experimental signal .
there are two limits in which we can simplify the above expression . for small @xmath118@xmath119 mixing
, we can usually take @xmath120 in eq .
( [ eq : gluino ] ) and then work in the mass insertion approximation .
however , if there is a large hierarchy in between @xmath119 and @xmath118 , or if there is large mixing , that approximation breaks down . instead
, it is far easier to work directly in the basis of the mass eigenstates . defining @xmath121 @xmath122 as @xmath123 and its orthogonal combination ,
we must replace the product @xmath124 in eq .
( [ eq : gluino ] ) with f(m__l1 ^ 2 , ) -f(m__l2 ^ 2 , ) 4 ( ) ^23 ^16 ||^13 .
[ eq : nmfvmax ] contours of the maximum allowed branching ratio are plotted in fig .
[ nmfvfig ] as a function of @xmath62 and @xmath1 for the case of maximal mixing . for small
mixing we can return to the mass insertion approximation in which case > 7 ( ) ^13 ( ) ^23 ^16 ||^13 . [ eq : nmfvmin ] = 2.5 in the bounds presented in this section should be considered the absolute limits on @xmath1 given a signal at the tevatron at the @xmath61 level .
nowhere did we inject theoretical constraints stemming from some particular model of susy .
on the other hand , if we had a model , or other information , we could devise much stronger bounds .
for example , many observables in susy depend on @xmath11 , so that once susy is found , we will very quickly have constraints on @xmath11 , which would allow more precise bounds . and
if we have some theoretical or experimental reason to believe in a particular model of susy - breaking , then we know even more . as we will discuss in section [ models ]
, models produce bounds which are much stronger than the general case . for most purposes ,
the bounds in the previous section are too general .
susy models which are truly generic and general have an embarassing wealth of flavor - changing mass insertions like @xmath99 .
most of these insertions must be miniscule in order to avoid dangerously large fcncs in well - studied processes , such as @xmath125-@xmath126 mixing .
thus one usually imposes on susy some form of organizing principle which prevents most of these insertions from becoming large , whether that be mass degeneracies or mass matrix alignment . however even in models which naturally solve the susy flavor problem
, one usually still expects some flavor violation to reappear by way of renormalization effects .
the most common source of these is the presence of the up - quarks yukawa matrix in the renormalization group equations for the left - handed squarks .
because all yukawa matrices can not be simultaneously diagonalized , non - zero @xmath127 will be generated .
( the @xmath47 mass insertions will not be generated at leading order .
) however there is a significant difference between these mass insertions and those considered in the general case above : here @xmath127 will be proportional to the corresponding element of the ckm matrix . for example , @xmath99 will be naturally suppressed by @xmath117 .
thus one typically finds @xmath128 which in turn would force @xmath129 for @xmath130 .
however this bound is not a true limit the model can be tuned in order to obtain the same branching ratio but with much smaller values of @xmath1 .
thus , while one expects from theoretical arguments that @xmath1 must be greater than 9 in order to see a signal at the tevatron , the true mathematical limit remains at @xmath131 as found in the previous section , until the origins of susy flavor physics are better understood .
the bound on @xmath1 in the general case depends strongly on the @xmath46 ( or @xmath47 ) mixing present in the strange bottom squark sector .
but there are several other key observables which also depend on this mixing and which should be measured or constrained over the next few years .
key among them are @xmath132@xmath52 mixing , parametrized by @xmath133 , and the rate and cp asymmetry in @xmath134 . on the subject of @xmath134 , we will not say very much , but refer the reader to ref .
@xcite for a full discussion .
the situation for @xmath133 is more interesting . in the presence of a single ,
dominating @xmath46 ( or @xmath47 ) mass insertion , a calculation of @xmath133 yields the approximate formula : m_s |(m_s)_sm + ( ) ^2 ( _ ll^23)^2| where @xmath135 is roughly 15 to @xmath136 .
the same formula also holds with the substitution @xmath137 .
( to obtain this simple result , we have followed the calculations of ref .
@xcite for @xmath51@xmath138 mixing , appropriately modified , and assumed all squark and gluino masses are essentially degenerate at @xmath139 . ) since @xmath99 can be complex , the susy contributions can either add to or subtract from the sm prediction for @xmath133 .
a measurement of @xmath133 near its sm value would indicate either large @xmath140 or small @xmath99 , either one of which would suppress @xmath2 and thus tighten our bound on @xmath1 . once susy partners are discovered , the value of @xmath139 will be known and a very stringent bound on @xmath99 can be extracted from the @xmath133 data , strengthening our bound on @xmath1 further .
in the previous sections , we performed very general analyses in determining what could be learned about the susy parameter space by measuring @xmath141 . now we ask the same question , but in the context of specific models of susy - breaking , namely the three most commonly studied models : minimal supergravity models ( msugra , also called the constrained mssm ) , minimal gauge - mediated ( gmsb ) models , and minimal anomaly - mediated ( amsb ) models .
realistic models will necessarily yield much stronger bounds on @xmath62 and @xmath1 for two reasons : first , there is no reason for their contributions to fcncs to be maximal , and second there are often correlations with other processes which suppress the flavor - changing contributions .
several analyses of @xmath2 in the context of susy models have been completed @xcite and a more detailed study of this will appear shortly @xcite ; here we summarize some of the important results of this last study .
namely we will find bounds on @xmath1 in specific models for an observed @xmath141 and discuss their implications . for each of the three models to be discussed , we will vary all the free parameters of the models and at each `` model point '' calculate the @xmath2 branching fraction @xcite
. however we will also apply several phenomenological constraints to avoid points which are unphysical .
the most important constraints come from the muon magnetic moment @xcite and the @xmath142 transition @xcite . for the former
, we will demand that the susy constributions to @xmath143 fall between @xmath144 and @xmath145 . for the latter
, we will demand that calculated rate for @xmath142 fall between @xmath146 and @xmath147 . in both cases
these are rather wide windows . for the muon magnetic moment ,
a wide window is required by the inconsistency of the @xmath148 and @xmath149 data sets which are used to calculate the standard model hadronic contributions to @xmath150 .
for @xmath142 we use a wide window because of the uncertainties in the next - to - leading order calculation of the susy contributions .
minimal supergravity ( msugra ) actually contributes fairly efficiently to higgs - mediated fcncs .
it has three key ingredients that help to generate large branching ratios .
first , msugra models often have large @xmath69 trilinear terms , necessary for the higgsino contribution .
second , the extended running from the gut scale down to the weak scale generates considerable mixing between the second and third generation squarks , allowing for a sizable gluino contribution .
finally , the pseudoscalar higgs does not come out particularly heavy . in fig .
[ sugrafig](a ) , we show a scatter plot of points in the msugra model space , each one representing a distinct choice of @xmath1 , @xmath151 , @xmath152 , and @xmath153 ( we take @xmath154 only ) , consistent with all experimental and theoretical constraints .
there are several important lessons to draw from the figure .
one should immediately note that msugra models exist , consistent with all other constraints , that have rates for @xmath2 right up to the cdf bound ; in fact , models were found that passed all other constraints , except the @xmath2 bound , indicating that @xmath2 is _ already _ a non - trivial constraint on msugra models .
we also notice that msugra models with a branching ratio greater than @xmath61 require @xmath155 .
thus observation of @xmath2 at run ii would certainly indicate very large @xmath1 if msugra is the correct model .
one also notices another interesting feature in the figure , namely a bifurcation of the points at the largest branching fractions .
the points in the upper arm all have @xmath156 while all the points with @xmath157 lie in the main body of points .
this behavior is rooted in the @xmath142 constraint . for @xmath157 ,
models with very large @xmath2 also have chargino contributions to @xmath142 which tend to cancel the sm .
thus the resulting @xmath158 is too small and we throw the model point out . as the susy mass scales increase , the cancellation goes away , but the rate for @xmath2 is also suppressed .
however , the cases with @xmath156 exhibit a very different behavior . for light spectra , these models can have susy contributions to @xmath142 much larger than those of the sm , but with opposite sign , so that after adding the two pieces , the resulting rate is still roughly consistent with the sm and with experiment
. then as the susy mass scale is increased , the two pieces start to cancel to zero , ruling out the model points . as the susy scale
is further increased , the susy contribution decouples and the sm once again dominates . thus a gap is generated where the models predict @xmath159 .
thus models with @xmath157 can only be observed in run ii if @xmath160 while models with @xmath156 can have @xmath1 down to 40 .
the upshot of this is that a measurement of @xmath2 and an independent measurement of @xmath1 could indicate that @xmath156 .
however , these measurements alone can not be used to show that @xmath157 .
the simplest models with pure anomaly mediation have only one free parameter , an overall mass scale , but are not phenomenologically viable .
but one can define a `` minimal '' anomaly - mediated model which has two free parameters : an overall mass scale for the anomaly mediation , and a separate mass scale just for the scalars .
we considered this minimal model , by varying both of these parameters , as well as @xmath1 , and studying the resulting spectra . in principle , amsb shares many of the same features that make msugra a nice candidate for discovery of @xmath2 . in practice
, this is only partially true . in amsb
the @xmath142 constraint is much stronger and rules out many model points where a large higgs - mediated fcnc signal would have been predicted .
we do not find any models with observable branching fractions of @xmath52 into muons @xmath161 , consistent with the result of baek @xcite .
however , we note that this result depends very strongly on the @xmath142 calculation ; changes in the nlo calculation could easily allow for much larger ( or smaller ) @xmath2 branching fractions . a scatter plot of the allowed parameter space is shown in fig . [ amsbfig](a ) .
this figure again shows a bifurcation into two separate regions , though this time much more distinct than in the msugra case .
here there is no fundamental difference between the two regions , but the basic explanation is similar to the case of the last section . for amsb models with large @xmath1 and intermediate masses , the susy contribution to @xmath142 tends to cancel the sm , predicting a rate which is too small . for large masses , the susy contribution decouples , but then so does @xmath2 .
however , there is a region of very light masses where the susy contribution is roughly twice the sm contribution , but with opposite sign , so that their sum squared is consistent with data .
this last region is the upper - right region of the figure .
there we see that one can have @xmath141 are large as @xmath61 only if @xmath162 .
unlike the previous two cases , one would not expect a large higgs - mediated fcnc signal in models with gauge mediation . whereas msugra has a large @xmath69 and lots of running to generate squark mixing , gauge - mediated models have neither . at the mediation scale , @xmath163 ( so no higgsino contribution ) and
there is little running unless the mediation scale is large ( so no gluino contribution ) .
indeed , we find that it is very difficult to generate observable signals at the tevatron for gmsb models , even without invoking constraints from @xmath142 .
this is best summarized by fig .
[ gmsbfig ] , where no models are found with @xmath141 above @xmath164 , making this signal difficult if not impossible to observe in run ii . to obtain this figure , we varied all the parameters of the gmsb model , including the number of messengers , which we allowed to take values of @xmath165 , 3 , and 5 .
we also allowed the messenger scale , @xmath166 , to go as high as @xmath167 . at such large @xmath166
, one would expect to maximize your signal since both @xmath168-terms and squark mixing benefit from the running between @xmath166 and @xmath169 . and
indeed the maximum signals are occurring for large messenger scales .
however even with such a large scale , there is little signal in this channel compared to msugra or amsb models .
thus a signal at the tevatron tells us something significant that has nothing to do with @xmath1 it would be evidence _ against _ gauge mediation itself @xcite !
over the next several years , the tevatron s physics program has a strong opportunity to discover new physics in the rare decay @xmath2 . and
if either cdf or d do observe this signature , they will have simultaneously provided very strong evidence for a supersymmetric world just beyond our current reach . and they will have provided an important insight into susy - breaking and its mediation mechanism . and they will have placed a lower bound on the parameter @xmath1 , vitally important to future studies of susy . it is the existence of this lower bound , and its value , that we have derived in this work .
the susy parameter @xmath1 is one of the most important quantities which needs to be measured in a supersymmetric world . in a top - down approach
, the calculation of @xmath1 from first principles would be an important test of some underlying theory . in
a bottom - up approach , @xmath1 is a necessary input to the yukawa sector and thus relevant to everything from yukawa unification to radiative electroweak symmetry breaking .
almost all predictions that can test the form of the underlying theory depend on @xmath1 in some way , so it must be measured as a first step in studying and testing models of susy .
yet @xmath1 is infamously difficult to measure , particularly at hadron colliders or in rare decays all methods proposed to date are model and/or parameter dependent .
the decay @xmath2 may improve the situation dramatically because it is unusually sensitive to @xmath1 , providing an opportunity to bound @xmath1 if a signal is seen . in this paper
we showed that some very general theoretical assumptions that do not depend on specific models allow us to put significant lower bounds on @xmath1 given a signal . in the general case with large flavor changing effects allowed ( @xmath170 ) ,
the bound is given by eq .
( [ eq : nmfvmax ] ) , leading to @xmath171 .
this bound increases quickly with decreasing flavor violation : for @xmath172 the bound is pushed to @xmath173 ( eq . ( [ eq : nmfvmin ] ) ) . in more typical models in which @xmath174 or @xmath175 , then @xmath129 if a signal is seen at the tevatron .
however , the decay @xmath2 does not require new sources of flavor changing to be present ; it can be induced simply from the flavor changing already present in the standard model ckm matrix .
in such a case a stronger bound @xmath176 is obtained ( eq . ( [ eq : mfv ] ) ) .
these bounds are for a @xmath177 ; we have also shown how to scale the bounds as a function of the branching fraction . finally , we find that the bounds are significantly stronger in often studied models such as minimal supergravity . for many purposes ,
a lower limit is as good as a measurement since many observables ( @xmath178 , @xmath179 , dark matter signals , etc ) saturate as @xmath1 increases .
there is also an upper limit on @xmath1 of order 60 from the requirement that the theory remain perturbative up to high ( unificiation ) scales .
thus a lower limit often is tantamount to a measurement .
we are grateful to k. hidaka for helpful comments on the manuscript .
this work was supported in part by the national science foundation under grant phy00 - 98791 and by the u.s .
department of energy under grant de - fg02 - 95er40896 . c. s. huang , w. liao and q. s. yan , phys .
d * 59 * , 011701 ( 1999 ) ; + c. hamzaoui , m. pospelov and m. toharia , phys .
d * 59 * , 095005 ( 1999 ) ; + s. r. choudhury and n. gaur , phys .
b * 451 * , 86 ( 1999 ) ; + c. s. huang , w. liao , q. s. yan and s. h. zhu , phys .
d * 63 * , 114021 ( 2001 ) . c. s. huang and w. liao , phys .
b * 525 * , 107 ( 2002 ) and phys .
b * 538 * , 301 ( 2002 ) ; + c. bobeth , t. ewerth , f. kruger and j. urban , phys .
d * 64 * , 074014 ( 2001 ) and phys .
d * 66 * , 074021 ( 2002 ) ; + a. j. buras , p. h. chankowski , j. rosiek and l. slawianowska , nucl .
b * 619 * , 434 ( 2001 ) , phys .
b * 546 * , 96 ( 2002 ) and nucl .
b * 659 * , 3 ( 2003 ) ; + g. isidori and a. retico , jhep * 0111 * , 001 ( 2001 ) and jhep * 0209 * , 063 ( 2002 ) .
a. dedes , arxiv : hep - ph/0309233 .
a. dedes , h. k. dreiner and u. nierste , phys .
lett . * 87 * , 251804 ( 2001 ) ; + r. arnowitt , b. dutta , t. kamon and m. tanaka , phys .
b * 538 * , 121 ( 2002 ) ; + s. baek , p. ko and w. y. song , phys .
lett . * 89 * , 271801 ( 2002 ) and jhep * 0303 * , 054 ( 2003 ) ; + h. baer _ et al .
_ , jhep * 0207 * , 050 ( 2002 ) ; + a. dedes , h. k. dreiner , u. nierste and p. richardson , arxiv : hep - ph/0207026 ; + j. k. mizukoshi , x. tata and y. wang , phys . rev .
d * 66 * , 115003 ( 2002 ) ; + t. blazek , s. f. king and j. k. parry , arxiv : hep - ph/0308068 .
model points were generated using the package softsusy : + b. c. allanach , comput .
commun . * 143 * , 305 ( 2002 ) [ arxiv : hep - ph/0104145 ] .
g. l. kane , j. lykken , b. d. nelson and l. t. wang , phys .
b * 551 * , 146 ( 2003 ) .
g. l. kane , p. ko , h. b. wang , c. kolda , j. h. park and l. t. wang , arxiv : hep - ph/0212092 and phys .
lett . * 90 * , 141803 ( 2003 ) . c. kolda and j. lennon ,
_ in preparation_. d. becirevic _ et al .
_ , nucl .
b * 634 * , 105 ( 2002 ) ; + p. ko , j. park and g. kramer , eur .
j. c * 25 * , 615 ( 2002 ) .
the experimental constraints on @xmath143 are based on : + g. w. bennett _ et al . _ [ muon g-2 collaboration ] , phys .
lett . * 89 * , 101804 ( 2002 ) . + for theoretical constraints in the context of susy see , _
e.g. _ : + m. byrne , c. kolda and j. e. lennon , phys . rev .
d * 67 * , 075004 ( 2003 ) .
the experimental constraints on @xmath180 are based on : + k. abe _ et al . _
[ belle collaboration ] , phys .
b * 511 * , 151 ( 2001 ) ; + s. chen _ et al .
_ [ cleo collaboration ] , phys . rev .
lett . * 87 * , 251807 ( 2001 ) . + for the calculation within susy see : + m. ciuchini , g. degrassi , p. gambino and g. f. giudice , nucl .
b * 534 * , 3 ( 1998 ) ; + g. degrassi , p. gambino and g. f. giudice , jhep * 0012 * , 009 ( 2000 ) ; + m. carena , d. garcia , u. nierste and c. e. wagner , phys . lett .
b * 499 * , 141 ( 2001 ) .
|
recently it has been understood that flavor - changing processes mediated by higgs bosons could be a new and powerful tool for discovering supersymmetry . in this paper
we show that they may also provide an important method for constraining the parameters of the minimal supersymmetric standard model ( mssm ) .
specifically , we show that observation of @xmath0 at the tevatron implies a significant , model - independent lower bound on @xmath1 in the mssm .
this is very important because @xmath1 enters crucially in predictions and interpretations of the mssm , though it is difficult to measure . within specific models , or with other data ,
the bound becomes significantly stronger .
@=11 caption#1[#2]#3 @=12 .7ex .7ex gev mev ev kev tev m_z m_pl ^-1 ^-1 * d * * d^ * * u * * u^ * * y_d * * y_d^ * * y_u * * y_u^ * * v * * v^ * * v^0 * * v^0 * * x * b^0-|b^0 # 1#1 | # 1| # 1 _ max # 1#2#3nucl
. phys .
* b#1 * ( 19#2 ) # 3 # 1#2#3phys . lett . * b#1 * ( 19#2 ) # 3 # 1#2#3phys . lett . *
# 1b * ( 19#2 ) # 3 # 1#2#3phys . rev .
* d#1 * ( 19#2 ) # 3 # 1#2#3phys .
. lett . *
# 1 * ( 19#2 ) # 3 # 1#2#3phys . rep . *
# 1 * ( 19#2 ) # 3 # 1#2#3ann .
rev .
. astrophys . *
# 1 * ( 19#2 ) # 3 # 1#2#3ann . rev .
nucl .
part .
sci . * # 1 * ( 19#2 ) # 3 # 1#2#3mod .
phys .
lett . *
# 1 * ( 19#2 ) # 3 # 1#2#3zeit .
fr physik * c#1 * ( 19#2 ) # 3 # 1#2#3ap .
j. * # 1 * ( 19#2 ) # 3 # 1#2#3ann . phys . * # 1 * ( 19#2 ) # 3 # 1#2#3rev .
mod .
phys . *
# 1 * ( 19#2 ) # 3 # 1#2#3comm .
math .
phys . *
# 1 * ( 19#2 ) # 3 _ i.e. _ _ et al . _ _ e.g. _ _ etc .
_ _ c.f .
_ /#1/ |#1 # 1#1 mctp-03 - 42 +
hep - ph/0310042 +
september 2003 + * @xmath2 as a probe of @xmath1 at the tevatron * + 0.5 cm _
@xmath3michigan center for theoretical physics , university of michigan , ann arbor , mi 48109 , usa + @xmath4department of physics , university of notre dame , notre dame , in 46556 , usa + _ .5 cm over the next several years , the tevatron at fermilab will be searching for signs of physics beyond the standard model ( sm ) , by directly producing and observing new particles , and by observing rare processes at rates inconsistent with sm predictions . in the latter category
falls the search for the rare flavor - changing neutral current ( fcnc ) decay @xmath5 .
observation of this decay at the tevatron would necessarily imply new physics since the predicted rate for this process in the sm is far below the search capabilities of the machine .
however , supersymmetry ( susy ) naturally predicts large enhancements in the decay rate mediated by neutral higgs bosons @xcite , and in some cases yields branching ratios three orders of magnitude above the sm .
so observation of the decay would be strong , albeit indirect , evidence in favor of susy .
but observation of @xmath2 actually tells us more
. we will show that it is possible to deduce bounds on the fundamental susy parameter @xmath1 , the ratio of the two higgs expectation values , from a signal .
this is particularly important because @xmath1 is very difficult to measure there is no general technique for measuring it at hadron colliders yet almost all susy observables depend on it . and a lower limit may be almost as useful as an actual measurement because many susy observables quickly saturate as @xmath1 increases .
but we can obtain a limit precisely because the rate for @xmath2 does not saturate .
instead it is strongly dependent on @xmath1 , rising ( and falling ) as @xmath6 . as a secondary feature
, we will also be able to obtain some information on the mass scale of the new susy higgs bosons , since the branching ratio scales as the fourth power of their masses .
we will begin by considering two very general scenarios in which all flavor mixing does / does not come from the ckm matrix and in each case we will arrive at a very strong bound .
but we will also consider some specific models of susy breaking and show that even more stringent bounds can be obtained for these .
our focus will remain on the tevatron because it has especially good sensitivity to the signal even if it does not reach its full luminosity potential .
however the b - factories and lhc can also use this , and related processes , to further probe @xmath1 and the higgs sector . unfortunately ,
if a signal is not seen , non - observation can not be used to draw absolute conclusions about the parameter space of the mssm other than to rule out specific model points .
it is always possible to choose susy parameters in such a way to push the signal down to the level of the standard model where it would not be observed .
|
optimal experimental design ( approximate theory ) is a well - developed area , and we refer to kiefer ( @xcite ) , silvey ( @xcite ) , pzman ( @xcite ) and pukelsheim ( @xcite ) for a general introduction and basic results .
we consider computational aspects of optimal designs , focusing on a finite design space @xmath0 .
suppose the probability density or mass function of the response is specified as @xmath1 where @xmath2 is the parameter of interest .
let @xmath3 denote the @xmath4 expected fisher information matrix from a unit assigned to @xmath5 with the @xmath6 entry [ the expectation is with respect to @xmath7 @xmath8.\ ] ] the moment matrix , as a function of the design measure @xmath9 , is defined as @xmath10 which is proportional to the fisher information for @xmath11 when the number of units assigned to @xmath5 is proportional to @xmath12 . here
@xmath13 , and @xmath14 denotes the closure of @xmath15 . throughout
we assume that @xmath3 are well defined and hence nonnegative definite .
the set @xmath16 is assumed nonempty .
our approach may conceivably extend to the case where @xmath17 is allowed to be singular , by using generalized inverses , although we do not pursue this here .
given an optimality criterion @xmath18 defined on positive definite matrices , the goal is to maximize @xmath19 with respect to @xmath20 .
typical optimality criteria include : the d - criterion @xmath21 , the a - criterion @xmath22 , more generally , the @xmath23th mean criterion @xmath24 and the c - criterion @xmath25 , where @xmath26 is a nonzero constant vector . often only a linear combination @xmath27 , for example , a subvector of @xmath11 , is of interest . the fisher information for @xmath27 is naturally defined as @xmath28 , assuming invertibility [ pukelsheim ( @xcite ) ]
. we may therefore consider the d- and a - criteria for @xmath27 defined , respectively , as @xmath29\\[-8pt ] \phi_{-1 , k}(m ) & = & -\operatorname{tr}(k^\top m^{-1 } k).\nonumber\end{aligned}\ ] ] the c - criterion is a special case of @xmath30 .
motivations for such optimality criteria are well known . in a linear problem ,
the a - criterion seeks to minimize the sum of variances of the best linear unbiased estimators ( blues ) for all coordinates of @xmath11 while the c - criterion seeks to minimize the variance of the blue for @xmath31 .
similar interpretations ( with asymptotic arguments ) apply to nonlinear problems
. in general @xmath17 also depends on the unknown parameter @xmath11 which complicates the definition of an optimality criterion .
a simple solution is to maximize @xmath19 with @xmath11 fixed at a prior guess @xmath32 ; this leads to _ local optimality _ [ chernoff ( @xcite ) ] .
local optimality may be criticized for ignoring uncertainty in @xmath11 . however , in a situation where real prior information is available , or where the dependence of @xmath33 on @xmath11 is weak , it is nevertheless a viable approach and has been adopted routinely [ see , e.g. , li and majumdar ( @xcite ) ] .
henceforth we assume a fixed @xmath32 and suppress the dependence of @xmath33 on @xmath11 .
possible extensions are mentioned in section [ sec5 ] .
optimal designs do not usually come in closed form . as early as wynn ( @xcite ) , fedorov ( @xcite ) , atwood ( @xcite ) and wu and wynn ( @xcite ) , and as late as torsney ( @xcite ) , harman and pronzato ( @xcite ) and dette , pepelyshev and zhigljavsky ( @xcite ) ,
various procedures have been studied for numerical computation .
we shall focus on the following multiplicative algorithm [ titterington ( @xcite , @xcite ) , silvey , titterington and torsney ( @xcite ) ] which is specified through a power parameter @xmath34 $ ] .
[ algorithmi ] set @xmath35 $ ] and @xmath36 . for
@xmath37 compute @xmath38 where @xmath39 iterate until convergence . for a heuristic explanation ,
observe that ( [ alg1 ] ) is equivalent to @xmath40 the value of @xmath41 indicates the amount of gain in information , as measured by @xmath18 , by a slight increase in @xmath12 , the weight on the @xmath42th design point .
so ( [ heur ] ) can be seen as adjusting @xmath43 so that relatively more weight is placed on design points whose increased weight may result in a larger gain in @xmath18 . if @xmath18 is increasing and concave , then a convenient convergence criterion , based on the general equivalence theorem [ kiefer and wolfowitz ( @xcite ) , whittle ( @xcite ) ] , is @xmath44 where @xmath45 and @xmath46 is a small positive constant .
algorithm i is remarkable in its generality .
for example , little restriction is placed on the underlying model @xmath1 .
part of the reason , of course , is that we focus on fisher information and local optimality , which essentially reduces the problem to a linear one .
there exists a large literature on algorithm [ algorithmi ] and its relatives [ see , e.g. , titterington ( @xcite , @xcite ) , silvey , titterington and torsney ( @xcite ) , pzman ( @xcite ) , fellman ( @xcite ) , pukelsheim and torsney ( @xcite ) , torsney and mandal ( @xcite ) , harman and pronzato ( @xcite ) , dette , pepelyshev and zhigljavsky ( @xcite ) and torsney and martn - martn ( @xcite ) ] . one feature that has attracted much attention is that algorithm [ algorithmi ] appears to be monotonic , that is , @xmath47 increases in @xmath48 , at least in some special cases .
for example , when @xmath49 ( for d - optimality ) and @xmath50 , titterington ( @xcite ) and pzman ( @xcite ) have shown monotonicity using clever probabilistic and analytic inequalities [ see also dette , pepelyshev and zhigljavsky ( @xcite ) and harman and trnovsk ( @xcite ) ] .
algorithm [ algorithmi ] is also known to be monotonic for @xmath51 as in ( [ phik ] ) , assuming @xmath52 and @xmath3 are rank - one [ fellman ( @xcite ) , torsney ( @xcite ) ] .
monotonicity is important because convergence then holds under mild assumptions ( see section [ sec4 ] ) .
results in these special cases suggest a monotonic convergence theory for a broad class of @xmath18 which is also supported by numerical evidence presented in some of the references above .
we aim to state general conditions on @xmath18 that ensure that algorithm [ algorithmi ] converges monotonically . as a consequence certain known theoretical results are unified and generalized , and one particular conjecture [ titterington ( @xcite ) ]
is confirmed .
define @xmath53 the functions @xmath18 and @xmath54 are assumed to be differentiable on invertible matrices .
our conditions are conveniently stated in terms of @xmath54 . as usual , for two symmetric matrices , @xmath55 means @xmath56 is nonnegative ( positive ) definite .
* @xmath57 is increasing : @xmath58 or , equivalently , @xmath59 is nonnegative definite for positive definite @xmath33 .
* @xmath57 is concave : @xmath60 for @xmath61 , m_1 , m_2>0 $ ] .
equivalently , @xmath62 condition ( [ increase ] ) is usually satisfied by any reasonable information criterion .
also note that , if ( [ increase ] ) fails , then @xmath41 on the right - hand side of ( [ heur ] ) is not even guaranteed to be nonnegative .
the real restriction is the concavity condition ( [ concave0 ] ) .
for example , ( [ concave0 ] ) is not satisfied by @xmath63 ( the @xmath23th mean criterion ) when @xmath64 .
[ it is usually assumed that @xmath65 , rather than @xmath57 , is concave . ] nevertheless , ( [ concave0 ] ) _ is _ satisfied by a wide range of criteria , including the commonly used d- , a- or c - criteria [ see cases ( i ) and ( ii ) in the illustration of the main result below ] .
our main result is as follows .
[ main ] assume ( [ increase ] ) and ( [ concave0 ] ) .
assume that in iteration ( [ alg1 ] ) , with @xmath66 , we have @xmath67 then @xmath68 in other words , under mild conditions which ensure that ( [ alg1 ] ) is well defined [ specifically , the denominator in ( [ alg1 ] ) is nonzero ] , ( [ increase ] ) and ( [ concave0 ] ) imply that ( [ alg1 ] ) never decreases the criterion @xmath18 .
let us illustrate theorem [ main ] with some examples . for simplicity , in ( i)(iv )
we display formulae for @xmath69 only , although monotonicity holds for all @xmath35 $ ] .
take @xmath70 then @xmath71 satisfies ( [ increase ] ) and ( [ concave0 ] ) . by theorem [ main ] ,
algorithm [ algorithmi ] is monotonic for @xmath72 $ ] .
this generalizes the previously known cases @xmath73 and @xmath74 ( with particular values of @xmath75 ) .
the iteration ( [ alg1 ] ) reads @xmath76 more generally , given a full rank @xmath77 matrix @xmath78 ( @xmath79 ) , consider @xmath80 then @xmath81 satisfies ( [ increase ] ) and ( [ concave0 ] ) . by theorem [ main ] ,
algorithm [ algorithmi ] is monotonic for @xmath82 $ ] .
the iteration ( [ alg1 ] ) reads @xmath83 in particular , taking @xmath84 ( an @xmath85 vector ) and @xmath74 in case ( ii ) , we obtain that algorithm [ algorithmi ] is monotonic for the c - criterion @xmath86 .
the iteration ( [ pk ] ) reduces to @xmath87 as noted by a referee , with @xmath74 , the choice @xmath69 may lead to an oscillating behavior in the sense that @xmath88 alternates between two points at which @xmath89 takes the same value . while this does not contradict theorem [ main ]
, it suggests that other values of @xmath75 are more desirable for fast convergence .
following fellman ( @xcite ) and torsney ( @xcite ) , a practical recommendation is @xmath52 in the @xmath74 case .
consider another example of case ( ii ) , with @xmath90 and @xmath91 .
henceforth @xmath92 denotes the @xmath93 vector of zeros , and @xmath94 denotes the @xmath95 identity matrix .
assume @xmath96 and @xmath97 is @xmath98 .
this corresponds to a d - optimal design problem for @xmath99 under the linear model , @xmath100 where the parameter is @xmath101 .
that is , interest centers on all coefficients other than the intercept .
nevertheless , as far as the design measure @xmath43 is concerned , the optimality criterion , @xmath102 , coincides with @xmath103 , that is , @xmath104 after some algebra , ( [ pk ] ) reduces to @xmath105 where @xmath106 thus ( [ alg2 ] ) satisfies @xmath107 .
monotonicity of ( [ alg2 ] ) has been conjectured since titterington ( @xcite ) , and considerable numerical evidence has accumulated over the years . recently , extending the arguments of pzman ( @xcite ) , dette , pepelyshev and zhigljavsky ( @xcite ) have obtained results which come very close to resolving titterington s conjecture .
nevertheless , we have been unable to extend their arguments further .
instead we prove the general theorem [ main ] using a different approach , and settle this conjecture as a consequence .
the proof of theorem [ main ] is achieved by using a method of _ auxiliary variables_. when a function @xmath108 [ e.g. , @xmath109 to be minimized is complicated , we introduce a new variable @xmath110 and a function @xmath111 such that @xmath112 for all @xmath43 , thus transforming the problem into minimizing @xmath111 over @xmath43 and @xmath110 jointly .
then we may use an iterative conditional minimization strategy on @xmath111 .
this is inspired by the em algorithm [ dempster , laird and rubin ( @xcite ) , meng and van dyk ( @xcite ) ; in particular , see csiszr and tusnady s ( @xcite ) interpretation ; see yu ( @xcite ) for a related interpretation of the data augmentation algorithm ] . in section [ sec3 ]
we analyze algorithm [ algorithmi ] using this strategy .
although attention is paid to the mathematics , our focus is on intuitively appealing interpretations which may lead to further extensions of algorithm [ algorithmi ] with the same desirable monotonicity properties .
if the algorithm is monotonic , then convergence can be established under mild conditions ( section [ sec4 ] ) .
section [ sec5 ] contains an illustration with optimal designs for a simple logistic regression model .
a key observation is that the problem of maximizing @xmath19 , or , equivalently , minimizing @xmath113 can be formulated as a joint minimization over both the design and the estimator .
specifically , let us compare the original problem [ problemp1 ] with its companion problem [ problemp2 ]
. throughout @xmath114 denotes the symmetric nonnegative definite ( snnd ) square root of an snnd matrix @xmath115 .
[ problemp1 ] minimize @xmath116 over @xmath117 .
[ problemp2 ] minimize @xmath118 over @xmath117 and @xmath110 [ an @xmath119 matrix ] , subject to @xmath120 , where @xmath121 though not immediately obvious , problems [ problemp1 ] and [ problemp2 ] are equivalent , and this may be explained in statistical terms as follows . in ( [ gwq ] ) , @xmath122 is simply the variance matrix of a linear unbiased estimator , @xmath123 , of the @xmath85 parameter @xmath11 in the model @xmath124 where @xmath125 is the @xmath126 vector of observations .
the constraint @xmath120 ensures unbiasedness .
[ note that @xmath127 is full - rank since @xmath17 is nonsingular by assumption .
] of course , the weighted least squares ( wls ) estimator is the best linear unbiased estimator , having the smallest variance matrix ( in the sense of positive definite ordering ) and , by ( [ increase ] ) , the smallest @xmath54 for that matrix .
it follows that , for fixed @xmath43 , @xmath111 is minimized by choosing @xmath123 as the wls estimator , @xmath128 however , from ( [ gwq ] ) and ( [ qwls ] ) we get @xmath129 that is , problem [ problemp2 ] reduces to problem [ problemp1 ] upon minimizing over @xmath110 . since problem [ problemp2 ] is not immediately solvable , it is natural to consider the subproblems : ( i ) minimizing @xmath111 over @xmath110 for fixed @xmath43 and ( ii ) minimizing @xmath111 over @xmath43 for fixed @xmath110 . part ( ii )
is again formulated as a joint minimization problem . for a fixed @xmath119 matrix @xmath110 such that @xmath130 , let us consider problems [ problemp3 ] and [ problemp4 ] .
[ problemp3 ] minimize @xmath111 as in ( [ gwq ] ) over @xmath117 .
[ problemp4 ] minimize the function @xmath131 over @xmath117 and the @xmath4 positive - definite matrix @xmath132 .
the concavity assumption ( [ concave ] ) implies that @xmath133 with equality when @xmath134 , that is , problem [ problemp4 ] reduces to problem [ problemp3 ] upon minimizing over @xmath132 . since problem [ problemp4 ] is not immediately solvable , it is natural to consider the subproblems : ( i ) minimizing @xmath135 over @xmath132 for fixed @xmath43 and @xmath110 and ( ii ) minimizing @xmath135 over @xmath43 for fixed @xmath132 and @xmath110 . part ( ii ) , which amounts to minimizing @xmath136
admits a closed - form solution : if we write @xmath137 where each @xmath138 is @xmath4 , then @xmath139 should be proportional to @xmath140 . but algorithm [ algorithmi ] may not perform an exact minimization here [ see ( [ s3 ] ) ] . based on the above discussion
, we can express algorithm [ algorithmi ] as an iterative conditional minimization algorithm involving @xmath141 and @xmath132 . at iteration @xmath48 , define @xmath142 then we have @xmath143}\nonumber\\ & = & h\bigl(\sigma^{(t ) } , w^{(t ) } , q^{(t)}\bigr ) \qquad\mbox{[by ( \ref{hmin2})]}\nonumber\\ \label{s3 } & \geq & h\bigl(\sigma^{(t ) } , w^{(t+1 ) } , q^{(t)}\bigr ) \qquad\mbox{(see below)}\\ \label{s4 } & \geq & g\bigl(w^{(t+1 ) } , q^{(t)}\bigr ) \qquad\mbox{[by ( \ref{hmin1 } ) , ( \ref{gwq})]}\\ \label{s5 } & \geq & \psi\bigl(m^{-1}\bigl(w^{(t+1)}\bigr)\bigr ) \qquad\mbox{[by ( \ref{gmin1 } ) , ( \ref{gmin2})]}.\end{aligned}\ ] ] the choice of @xmath144 leads to ( [ s3 ] ) as follows .
after simple algebra , the iteration ( [ alg1 ] ) becomes @xmath145 where @xmath146 since @xmath147 , jensen s inequality yields @xmath148 that is , @xmath149 hence @xmath150 which produces ( [ s3 ] ) .
choosing @xmath52 , that is , @xmath151 , leads to exact minimization in ( [ s3 ] ) ; choosing @xmath69 yields equality in ( [ s3 ] ) . but
any choice of @xmath144 that decreases @xmath152 at ( [ s3 ] ) would have resulted in the desired inequality , @xmath153 we may allow @xmath75 to change from iteration to iteration , and monotonicity still holds , as long as @xmath35 $ ] .
see silvey , titterington and torsney ( @xcite ) and fellman ( @xcite ) for investigations concerning the choice of @xmath75 . also note that we assume @xmath154 for all @xmath42 .
this is not essential , however , because ( i ) the possibility of @xmath155 can be handled by restricting our analysis to all design points @xmath42 such that @xmath156 , and ( ii ) the possibility of @xmath157 can be handled by a standard limiting argument .
monotonicity holds as long as @xmath158 and @xmath159 are both positive definite , as noted in the statement of theorem [ main ] .
monotonicity ( theorem [ main ] ) plays an important role in the following convergence theorem .
[ thm2 ] denote the mapping ( [ alg1 ] ) by @xmath160 .
assume @xmath161 assume ( [ alg1 ] ) is strictly monotonic , that is , @xmath162 assume @xmath18 is strictly concave and @xmath163 is continuous on positive definite matrices .
assume that , if @xmath33 ( a positive definite matrix ) tends to @xmath164 such that @xmath65 increases monotonically , then @xmath164 is nonsingular .
let @xmath88 be generated by ( [ alg1 ] ) with @xmath165 for all @xmath42 .
then : all limit points of @xmath88 are global maxima of @xmath19 on @xmath166 , and as @xmath167 , @xmath47 increases monotonically to @xmath168 .
the proof of theorem [ thm2 ] is somewhat subtle .
standard arguments show that all limit points of @xmath88 are fixed points of the mapping @xmath160 .
this alone does not imply convergence to a global maximum , however , because there often exist sub - optimal fixed points on the boundary of @xmath169 .
( global maxima occur routinely on the boundary also . )
our goal is therefore to rule out possible convergence to such sub - optimal points ; details of the proof are presented in yu ( @xcite ) , an extended version of this paper .
we shall comment on conditions ( a)(d ) .
condition ( a ) ensures that starting with @xmath170 , all iterations are well defined .
moreover , if @xmath165 for all @xmath42 , then @xmath171 for all @xmath48 and @xmath42 .
this highlights the basic idea that , in order to converge to a global maximum @xmath172 , the starting value @xmath173 must assign positive weight to every support point of @xmath172 .
such a requirement is not necessary for monotonicity . on the other hand ,
assigning weight to nonsupporting points of @xmath172 tends to slow the algorithm down .
hence methods that quickly eliminate nonoptimal support points are valuable [ harman and pronzato ( @xcite ) ] . condition ( b ) simply says that unless @xmath43 is a fixed point , the mapping @xmath160 should produce a better solution .
let us assume ( [ increase ] ) , ( [ concave ] ) and condition ( a ) so that theorem [ main ] applies .
then , by checking the equality condition in ( [ s3 ] ) , it is easy to see that condition ( b ) is satisfied if @xmath174 . [
the argument leading to ( [ iff ] ) technically assumes that all coordinates of @xmath43 are nonzero , but we can apply it to the appropriate subvector of @xmath43 . ] if @xmath69 , then ( [ s3 ] ) reduces to an equality .
however , by checking the equality conditions in ( [ s4 ] ) and ( [ s5 ] ) , we can show that condition ( b ) is satisfied if @xmath54 is strictly increasing and strictly concave : @xmath175\\[-8pt ] & & \qquad m_1\neq m_2 \quad \longrightarrow\quad \psi(m_1)<\psi ( m_2);\nonumber\\ \label{sconcave } & & m_1 , m_2>0,\nonumber\\[-8pt]\\[-8pt ] & & \qquad m_1\neq m_2 \quad\longrightarrow\quad\psi(m_2)<\psi ( m_1)+\operatorname{tr}\bigl(\psi'(m_1)(m_2-m_1)\bigr).\nonumber\end{aligned}\ ] ] conditions ( c ) and ( d ) are technical requirements that concern @xmath18 alone . condition ( c ) ensures uniqueness of the optimal moment matrix which simplifies the analysis .
condition ( d ) ensures that positive definiteness of @xmath17 is maintained in the limit .
conditions ( c ) and ( d ) are satisfied by @xmath176 with @xmath177 , for example .
let us mention a typical example of theorem [ thm2 ] .
[ coro1 ] assume @xmath178 , and @xmath179 .
then the conclusion of theorem [ thm2 ] holds for algorithm [ algorithmi ] with @xmath49 .
conditions ( a ) , ( c ) and ( d ) are readily verified .
condition ( b ) is satisfied by ( [ sincrease ] ) and ( [ sconcave ] ) .
the claim follows from theorem [ thm2 ] .
when ( [ sincrease ] ) or ( [ sconcave ] ) fails , and @xmath69 , it is often difficult to appeal to theorem [ thm2 ] because strict monotonicity [ condition ( b ) ] may not hold .
we illustrate this with an example where the monotonicity is not strict , and the algorithm does not converge [ see pronzato , wynn and zhigljavsky ( @xcite ) , chapter 7 ; also the remark in case ( iii ) following theorem [ main ] ] .
consider iteration ( [ alg2 ] ) with @xmath180 and design space @xmath181 .
it is easy to show that , for any @xmath182 , iteration ( [ alg2 ] ) maps @xmath88 to @xmath183 . thus , unless @xmath184 to begin with , the algorithm alternates between two distinct points .
this appears to be a rare example , as ( [ alg2 ] ) usually converges in practical situations .
one can think of several reasons for the wide interest in algorithm [ algorithmi ] and its relatives .
similar to the em algorithm , algorithm [ algorithmi ] is simple , easy to implement and monotonically convergent for a large class of optimality criteria ( although this was not proved in the present generality ) .
algorithm [ algorithmi ] is known to be slow sometimes .
but it serves as a foundation upon which more effective variants can be built [ see , e.g. , harman and pronzato ( @xcite ) and dette , pepelyshev and zhigljavsky ( @xcite ) ] . while solving the conjectured monotonicity of ( [ alg2 ] ) holds mathematical interest , our main contribution is a way of interpreting such algorithms as optimization on augmented spaces .
this opens up new possibilities in constructing algorithms with the same desirable monotonic convergence properties . as a numerical example , consider the logistic regression model @xmath185 the expected fisher information for @xmath11 from a unit assigned to @xmath5 is @xmath186 we compute locally optimal designs with prior guess @xmath187 , and design spaces , @xmath188 the design criteria considered are @xmath189 ( for d - optimality ) and @xmath190 .
we use algorithm [ algorithmi ] with @xmath69 , starting with equally weighted designs . for @xmath189 , corollary [ coro1 ]
guarantees monotonic convergence .
this is illustrated by figure [ fig1 ] , the first row , where @xmath191 is plotted against iteration @xmath48 . using the convergence criterion ( [ conv ] ) with @xmath192 , the number of iterations until convergence is 93 for @xmath193 and 2121 for @xmath194 .
the actual locally d - optimal designs are @xmath195 for @xmath193 and @xmath196 for @xmath197 , as can be verified using the general equivalence theorem .
this simple example serves to illustrate both the monotonicity of algorithm [ algorithmi ] ( when theorem [ main ] applies ) and its potential slow convergence . and
@xmath198 for algorithm [ algorithmi ] with design spaces @xmath193 and @xmath194 . ] for @xmath199 , although algorithm [ algorithmi ] can be implemented just as easily , theorem [ main ] does not apply because the concavity condition ( [ concave ] ) no longer holds . indeed ,
algorithm [ algorithmi ] ( with @xmath69 ) is not monotonic , as is evident from figure [ fig1 ] , in the second row , where @xmath200 is plotted against iteration @xmath48 .
this shows the potential danger of using algorithm [ algorithmi ] when monotonicity is not guaranteed .
although theorem [ main ] does not cover the @xmath201 criterion for @xmath64 , it is still possible that monotonicity holds for a smaller range of @xmath75 .
calculations in special cases lead to the conjecture [ silvey , titterington and torsney ( @xcite ) ] that algorithm [ algorithmi ] is monotonic if @xmath202 .
theorem [ main ] provides further evidence for this conjecture , but new insights are needed to resolve it . we have focused on local optimality .
an alternative , _ bayesian optimality _ [ chaloner and larntz ( @xcite ) , chaloner and verdinelli ( @xcite ) ] , seeks to maximize the expected value of @xmath203 over a prior distribution @xmath204 .
the notation @xmath205 emphasizes the dependence of the moment matrix on the parameter @xmath206 .
it would be worthwhile to extend our strategy in section [ sec3 ] to bayesian optimality , and we plan to report both theoretical and empirical evaluations of such extensions in future works .
the author would like to thank don rubin , xiao - li meng and david van dyk for introducing him to the field of statistical computing .
he is also grateful to mike titterington , ben torsney and the referees for their valuable comments .
torsney , b. ( 1983 ) . a moment inequality and monotonicity of an algorithm . in _ proceedings of the international symposium on semi - infinite programming and appl . _ ( k. o. kortanek and a. v. fiacco , eds . ) . _ lecture notes in economics and mathematical systems _ * 215 * 249260 .
springer , berlin .
|
monotonic convergence is established for a general class of multiplicative algorithms introduced by silvey , titterington and torsney [ _ comm .
statist . theory methods _ * 14 * ( 1978 ) 13791389 ] for computing optimal designs . a conjecture of titterington [ _ appl .
stat . _
* 27 * ( 1978 ) 227234 ] is confirmed as a consequence .
optimal designs for logistic regression are used as an illustration . . .
|
Subsets and Splits
Data Mining Abstracts
This query searches for and retrieves 2000 documents containing the phrase 'data mining' in the abstract, providing limited but specific document samples.