article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
recently sub - diffusion processes have attracted increasing interest since the introduction of continuous time random walks ( ctrws ) in @xcite and a large number of contributions have been given to them ( @xcite , @xcite ) . since ctrw is a random walk subordinated to a simple renewal process , by @xcite , it can be regarded as a generalized physical diffusion process ( including the sub - diffusion process and the super - diffusion process ) and there exists a closed connection between the time fractional diffusion system and the sub - diffusion process .
moreover , it is confirmed in @xcite and @xcite that the time fractional diffusion systems can be used to well characterize those sub - diffusion processes , which offer better performance not achievable before using conventional diffusion systems and surely raise many potential research opportunities at the same time . in the case of diffusion system
, it is well known that in general , not all the states can be reached in the whole domain of interest .
so here , we first introduce some notations on the regional controllability of time fractional diffusion systems when the system under consideration is only exactly ( or approximately ) controllability on a subset of the whole space , which can be regarded as an extensions of the research work in ( @xcite , @xcite ) . besides
, focusing on regional controllability would allow for a reduction in the number of physical actuators , offer the potential to reduce computational requirements in some cases , and also possible to discuss those systems which are not controllable on the whole domain , etc .
furthermore , in @xcite and @xcite , the authors have shown that the measurements and actions in practical systems can be better described by using the notion of actuators and sensors ( including the location , number and spatial distribution of actuators and sensors @xcite ) .
then the contribution of this present work is on the regional controllability of the sub - diffusion processes described by riemann - liouville time fractional diffusion systems of order @xmath2 by using the notion of actuators and sensors .
as cited in @xcite , their applications are rich in many real life .
for example , the flow through porous media ( @xcite ) , or the swarm of robots moving through dense forest ( @xcite ) .
we hope that the results here could provide some insights into the qualitative analysis of the design and configuration of fractional controller .
the rest of the paper is organized as follows .
the mathematical concept of regional controllability problem is presented in the next section .
section 3 is focused on the characterizations of strategic actuators in the case of regional controllability . in section 4 ,
our main results on the regional controllability analysis of time time fractional diffusion systems are presented and the determination of the optimal control which achieves the regional controllability is obtained .
two applications are worked out in the last section .
let @xmath3 be an open bounded subset of @xmath4 with smooth boundary @xmath5 and we consider the following abstract riemann - liouville time fractional differential system : @xmath6,~0<\alpha<1 , \\
\lim\limits_{t\to 0^+ } { } _ 0d^{\alpha}_t z(t)=z_0 , \end{array}\right.\ ] ] where @xmath7 generates a strongly continuous semigroup @xmath8 on the hilbert space @xmath9 , @xmath10 is a uniformly elliptic operator ( @xcite , @xcite ) , @xmath11 and the initial vector @xmath12 . here
@xmath13 and @xmath14 denote the riemann - liouville fractional order derivative and integral , respectively , given by @xcite @xmath15 and @xmath16 in addition , @xmath17 is a control operator depends on the number and the structure of actuators . the control @xmath18 where @xmath19 is a hilbert space . in particular ,
if the system is excited by @xmath20 actuators , one has @xmath21 and @xmath22 we first recall some necessary lemmas to be used afterwards .
[ lem0 ] for any given @xmath23 @xmath24 a function @xmath25 is said to be a mild solution of the following system @xmath26 , \\
\lim\limits_{t\to 0^+}{}_0d_t^{\alpha-1}v(t)=v_0\in z , \end{array}\right.\ ] ] if it satisfies @xmath27 where @xmath28 here @xmath8 is the strongly continuous semigroup generated by operator @xmath7 , @xmath29 and @xmath30 is a probability density function defined by @xmath31 @xmath32 such that ( @xcite ) @xmath33 * proof . *
it follows from the laplace transforms @xmath34 that the system @xmath35 is equivalent to ( @xcite ) @xmath36 then @xmath37}ds .
\end{array}\ ] ] consider the stable probability density function @xmath38 . by the arguments in @xcite
, we see that @xmath39@xmath40 satisfies the following property @xmath41 let @xmath42 we obtain that @xmath43}d\tau\\&= & \alpha\int_0^\infty{\int_0^\infty{e^{-\lambda \tau\theta}\psi_\alpha(\theta)\phi(\tau^\alpha)\tau^{\alpha-1}[v_0+\tilde{f}(\lambda)]}d\theta}d\tau\\&= & \sigma_1(v_0)+\sigma_2(f),\end{aligned}\ ] ] where @xmath44 and @xmath45 suppose that @xmath46 . then we have @xmath47 and @xmath48 let @xmath29 and + @xmath49 then we get @xmath50 and the proof is complete .
@xcite[lem2 ] let @xmath51 be an open set and @xmath52 be the class of infinitely differentiable functions on @xmath3 with compact support in @xmath3 and @xmath53 be such that @xmath54 then @xmath55 almost everywhere in @xmath56 @xcite[lem1 ] let the reflection operator @xmath57 on interval @xmath58 $ ] be as follows : @xmath59 then the following equations hold : @xmath60 and @xmath61 let @xmath62 be a given region of positive lebesgue measure and @xmath63@xmath64the target function@xmath65 be a given element .
by lemma [ lem0 ] , the unique mild solution @xmath66 of @xmath35 can be given by @xmath67 taking into account that @xmath35 is a line system , by the proposition 3.1 in @xcite , it suffices to suppose that @xmath68 in the following discussion .
let @xmath69 be @xmath70 in order to state the main results , the following two assumptions are supposed to hold all over the article : @xmath71 @xmath17 is a densely defined operator and @xmath72 exists .
@xmath73 @xmath74 exists and @xmath75 . in particular , when @xmath76 is a bounded linear operator from @xmath77 to @xmath78 , it is easy to see that @xmath71 and @xmath73 hold .
suppose that @xmath79 , generated by the adjoint operator of @xmath7 , is also a strongly continuous semigroup in the space @xmath78 .
for any @xmath80 by @xmath81 , we have @xmath82 where @xmath83 is the duality pairing of the space @xmath78 , @xmath72 is the adjoint operator of @xmath17 and @xmath84 consider now the restriction map @xmath85 defined by @xmath86 , is the projection operator on @xmath87 .
then the adjoint operator of @xmath88 can be given by @xmath89 and we are ready to state the following definition .
[ def2 ] @xmath90 the system @xmath35 is said to be regionally exactly controllable on @xmath87 if for any @xmath63 , there exists a control @xmath91 such that @xmath92 @xmath93 the system @xmath35 is said to be regionally approximately controllable on @xmath87 at time @xmath94 if for any @xmath63 , given @xmath95 there exists a control @xmath91 such that @xmath96 [ prop1 ] let @xmath97 be defined as @xmath98
. then following properties are equivalent : @xmath99 the system @xmath35 is regionally exactly controllable on @xmath87 at time @xmath94 ; + @xmath100 @xmath101 ; + @xmath102 @xmath103 + @xmath104 for @xmath105 there exists a @xmath106 such that @xmath107 * proof .
* obviously , @xmath108
@xmath109 for any @xmath105 let @xmath110 be the extension of @xmath111 to @xmath112 since @xmath101 , there exists @xmath21 , @xmath113 such that @xmath114 @xmath115 for any @xmath116 , from @xmath102 , @xmath117 where @xmath113 and @xmath118 .
then there exists a @xmath21 such that @xmath119 .
hence , it follows from the definition of @xmath88 that @xmath101 .
@xmath120 here , we note that the equivalence between @xmath99 and @xmath104 can be deduced based on the following general result in @xcite : let @xmath121 be reflexive hilbert spaces and @xmath122 @xmath123 then the following two properties are equivalent @xmath99 @xmath124 ; + @xmath100 @xmath125 such that @xmath126
. + by choosing @xmath127 @xmath128 @xmath129 and @xmath130 , we then obtain the results and completes the proof . there is an equivalence among the following properties : @xmath131 the system @xmath35 is regionally approximately controllable on @xmath87 at time @xmath94 ; + @xmath132 @xmath133 + @xmath134 @xmath135 + @xmath136 the operator @xmath137 is positive definite . * proof .
* similar to the argument in proposition [ prop1 ] , we obtain that @xmath138 . finally , we show that @xmath139 .
in fact , it is well known that @xmath140 let @xmath141 .
then we see that @xmath142 i.e. , the operator @xmath137 is positive definite and the proof is complete .
[ remark7 ] @xmath99 the definition [ def2 ] can be applied to the case where @xmath143 .
note that there exists a system , which is not controllable on the whole domain but regionally controllable ( see example 5.1 below ) .
@xmath100 a system which is exactly ( respectively approximately ) controllable on @xmath87 is exactly ( respectively approximately ) controllable on @xmath144 for every @xmath145
in this section , we will explore the characteristic of actuators when the system @xmath35 is regionally approximately controllable . as pointed out in @xcite ,
an actuator is a couple @xmath146 where @xmath147 is the support of the actuator and @xmath148 is its spatial distribution . to state our main results , it is supposed that the system under consideration is excited by @xmath20 actuators @xmath149 and let @xmath150 , where @xmath151 , @xmath152 , @xmath153 and @xmath154 .
then the system @xmath35 can be rewritten as follows : @xmath155 , \\\lim\limits_{t\to 0^+ } z(t , x)=z_0(x ) \mbox { in } \omega . \end{array}\right.\ ] ] moreover , suppose that @xmath10 is a uniformly elliptic operator .
by @xcite , we get that there exists a sequence @xmath156 , @xmath157 such that @xmath99 for each @xmath158 , @xmath159 is the eigenvalue of the operator @xmath10 with multiplicities @xmath160 and @xmath161 @xmath100 for each @xmath158 , @xmath162 is the orthonormal eigenfunction corresponding to @xmath159 , i.e. , @xmath163 where @xmath164 @xmath165 and @xmath166 is the inner product of space @xmath167 .
then we see that the strongly continuous semigroup @xmath8 on @xmath78 generated by @xmath7 is @xmath168 and the sequence @xmath169 is an orthonormal basis in @xmath167 , then for any @xmath170 , it can be expressed as @xmath171 an actuators ( or a suite of actuators ) is said to be @xmath172strategic if the system under consideration is regionally approximately controllable on @xmath87 . before showing our main result
in this part , from eq.@xmath173 and eq.@xmath174 , for any @xmath175 we have @xmath176 where @xmath177 , @xmath178 , @xmath179 , @xmath180 is the generalized mittag - leffler function in three parameters and here , @xmath181 is the pochhammer symbol defined by ( see @xcite , section 2.1.1 ) @xmath182 moreover , if @xmath183 such that @xmath180 , @xmath184 , then ( see section 2.3.4 in @xcite , or section 5.1.1 in @xcite ) @xmath185 it follows that @xmath186 and @xmath187 where @xmath188 @xmath189 is known as the generalized mittag - leffler function in two parameters and @xmath190 , @xmath157 , @xmath191 , @xmath192
. then we have the following result . for @xmath193 ,
arbitrary given @xmath194 , define @xmath195 matrices @xmath196 as @xmath197_{p\times r_j},\ ] ] where @xmath190 , @xmath157 , @xmath191 , @xmath192 .
then the suite of actuators @xmath149 is said to be @xmath172strategic if and only if @xmath198 * proof . * for any given @xmath194 and all @xmath91 , suppose that @xmath199 satisfies @xmath200 where @xmath201 , @xmath193 , @xmath191 . moreover , since @xmath202 in @xmath203 is arbitrary , lemma [ lem2 ] leads us to @xmath204 then we see that the suite of actuators @xmath149 is @xmath172strategic if and only if for any @xmath199 , one has @xmath205\end{aligned}\ ] ] @xmath206 where @xmath207 , @xmath208 is a vector in @xmath209 and @xmath193 . finally , since @xmath210 for all @xmath211 , @xmath212 we then show our proof by using the reductio and absurdum .
@xmath213 if the actuators @xmath149 are not @xmath172strategic , i.e. , the system @xmath35 is not regionally approximately controllable on @xmath87 .
there exists a @xmath214 satisfying @xmath215 then if @xmath216 , we see that @xmath217 @xmath218 on the contrary , if @xmath216 and @xmath219 there exists a nonzero element @xmath220 with @xmath221 such that @xmath222 then there exists a nonzero element @xmath223 satisfying @xmath224 this implies that @xmath225 and the suite of actuators @xmath149 is not @xmath172strategic .
the proof is complete .
@xmath226 the system @xmath35 with @xmath227 , @xmath228 and @xmath229 being h@xmath230lder continuous on the compact domain of @xmath4 is discussed in @xcite , which can be considered as a particular case of our results .
@xmath231 if the multiplicity of the eigenvalues @xmath159 of the operator @xmath10 is infinite for some @xmath193 and if the system @xmath35 is regionally approximately controllable , then the number of the control functions should not be finite .
the purpose of this section is to present an approach on how to achieve the regional approximate controllability on @xmath87 with the minimum control energy to steer the system @xmath35 from the initial vector @xmath232 to a target function @xmath233 in the region @xmath87 .
the method used here is the hilbert uniqueness methods ( hums)@xcite .
let @xmath234 be the closed convex set defined by @xmath235 consider the following minimization problem @xmath236 next , we show a direct approach to the solution of the regional controllability problem with minimum control energy by utilizing the hums .
let @xmath237 and @xmath238 be the sets given by @xmath239 and @xmath240 then for @xmath241 , we have @xmath242 moreover , for any @xmath243 , consider the system @xmath244 , \\
\lim\limits_{t\to 0^+ } q{}_td^{\alpha-1}_b\varphi(t)=p_\omega^*g \end{array}\right.\ ] ] and the semi - norm on @xmath237 @xmath245 where the reflective operator @xmath57 is defined in @xmath246 .
[ lem4 ] @xmath247 defines a norm on @xmath237 if the system @xmath35 is regionally approximately controllable on @xmath87
. * proof .
* for any @xmath243 , by lemma [ lem1 ] , we see that system @xmath248 can be rewritten as @xmath249 , \\
\lim\limits_{t\to 0^+ } { } _ 0d^{\alpha-1}_tq\varphi(t)=p_\omega^*g \end{array}\right.\ ] ] and its unique mild solution is @xmath250 moreover , if the system @xmath35 is regionally approximately controllable on @xmath87 , we have @xmath251 i.e. , @xmath252 hence , for any @xmath243 , it follows from @xmath253 that @xmath254 is a norm of space @xmath237 and the proof is complete .
in addition , consider the following system @xmath255 , \\
\lim\limits_{t\to 0^+}{}_0d^{\alpha-1}_{t}\psi(t)=0 , \end{array}\right.\ ] ] which is controlled by the solution of the system @xmath248 .
let @xmath256 be @xmath257 suppose that @xmath258 satisfies @xmath259 for all @xmath260 we see that @xmath261 $ ] and the regional controllability problem is equivalent to solving the equation @xmath262 then we can obtain the following theorem .
[ th2 ] if the system @xmath35 is regionally approximately controllable on @xmath87 , then for any @xmath260 @xmath263 has a unique solution @xmath243 and the control @xmath264 steers the system @xmath35 to @xmath233 at time @xmath94 in @xmath87 .
moreover , @xmath265 solves the minimum problem @xmath266 .
* by lemma @xmath267 , we see that if the system @xmath35 is regionally approximately controllable on @xmath87 , then @xmath254 is a norm of space @xmath237 .
let the completion of @xmath237 with respect to the norm @xmath254 again by @xmath237 .
then we will show that @xmath263 has a unique solution in @xmath237 . for any @xmath243
, it follows from the definition of operator @xmath268 in @xmath269 that @xmath270 hence , @xmath271 is one to one .
it follows from theorem 2.1 in @xcite that @xmath263 admits a unique solution in @xmath237 .
further , let @xmath272 in problem @xmath35 , one has @xmath273 then for any @xmath274 with @xmath275 , we obtain that @xmath276=0.$ ] moreover , for any @xmath277 we have @xmath278\right>=0 $ ] and @xmath279 by the theorem 1.3 in @xcite , it then follows from @xmath280 that @xmath265 solves the minimum energy problem @xmath266 and the proof is complete .
this section aims to present two examples to show the effectiveness of our obtained results .
* example 5.1 .
* let us consider the following one dimensional time fractional order differential equations of order @xmath2 with a zone actuator to show @xmath99 of remark @xmath281 .
@xmath282}u(t)\mbox{in } [ 0,1]\times [ 0,b ] , \\
\lim\limits_{t\to 0^+}z(x , t)=z_0(x)\mbox { in } [ 0,1 ] , \\
z(0 , t)= z(1 , t)=0
\mbox { in } [ 0,b ] , \end{array}\right.\end{aligned}\ ] ] where @xmath283}u$ ] and @xmath284 . moreover , we see that @xmath285 with @xmath286 , @xmath287 and @xmath288 since @xmath289 is a self - adjoint operator , we have @xmath290(t)\\ = b^*(b - t)^{\alpha-1}\sum\limits_{i=1}^{\infty}e_{\alpha,\alpha}(-\lambda_i(b - t)^{\alpha})(z,\xi_i)\xi_i(x)\\ = ( b - t)^{\alpha-1}\sum\limits_{i=1}^{\infty}e_{\alpha,\alpha}(-\lambda_i(b - t)^{\alpha})(z,\xi_i)\int_{a_1}^{a_2}{\xi_i(x)}dx . \end{array}\end{aligned}\ ] ] by @xmath291 , we get that @xmath292 @xmath293 when @xmath294 .
then the system @xmath295 is not controllable on @xmath296 $ ] .
next , we show that there exists a sub - region @xmath297 such that the system @xmath295 is possible regional controllability in @xmath87 at time @xmath94 . without loss of generality ,
let @xmath298 , @xmath299 . based on the argument above
, @xmath300 is not reachable on @xmath301.$ ] however , since @xmath302 @xmath303 let @xmath304 $ ] , we see that @xmath305 \\\neq 0 . \end{array}\end{aligned}\ ] ] then @xmath300 is possible regional controllability in @xmath304 $ ] at time @xmath94 .
* example 5.2 .
* consider the following time fractional differential equations with a pointwise actuator @xmath306\times [ 0,b ] , \\
\lim\limits_{t\to 0^+}z(x,0)=z_0\mbox { in } [ 0,1 ] , \\z(0 , t)= z(1 , t)=0\mbox { in } [ 0,b ] , \end{array}\right.\end{aligned}\ ] ] which is excited by a pointwise control located at @xmath307.$ ] here @xmath308 generates a strongly continuous semigroup .
in addition , for any @xmath243 , by lemma @xmath267 , we see that @xmath309 defines a norm on @xmath237 , where @xmath310 is the unique mild solution of the following problem @xmath311 , \\
\lim\limits_{t\to 0^+ } q{}_td^{\alpha-1}_b\varphi(t)=p_\omega^*g .
\end{array}\right.\ ] ] now if we consider the following system @xmath312 , \\
\lim\limits_{t\to 0^+}{}_0d^{\alpha-1}_{t}\psi(t)=0 . \end{array}\right.\ ] ] let @xmath313 $ ] be a subinterval and let @xmath314 be @xmath315 then the regional controllability of the example @xmath316 is equivalent to solving the equation @xmath317 where @xmath258 is the solution of the following system @xmath318 , \\ \lim\limits_{t\to 0^+}\widetilde{\psi}(t)=z_0 .
\end{array}\right.\end{aligned}\ ] ] thus , by theorem @xmath319 , we can conclude that if the example @xmath316 is regionally approximately controllable on some subregion of @xmath296 $ ] , for any @xmath260 @xmath320 admits a unique solution @xmath243 .
moreover , the control @xmath321 steers @xmath316 to @xmath233 at time @xmath94 and @xmath265 solves the minimum control energy problem @xmath266 .
the purpose of this paper is to investigate the regional controllability of the riemann - liouville time fractional diffusion equations of order @xmath322 .
the characterizations of strategic actuators when the control inputs appear in the differential equations as distributed inputs and an approach on the regional controllability with minimum energy of the problems @xmath35 are solved . since @xmath323 , @xmath211 , together with @xmath324 we get that our results can be regarded as the extension of the results in @xcite and @xcite .
moreover , the results presented here can also be extended to complex fractional order distributed parameter dynamic systems .
for instance , the problem of constrained regional control of fractional order diffusion systems with more complicated regional sensing and actuation configurations are of great interest . for more information on the potential topics related to fractional distributed parameter systems
, we refer the readers to @xcite and the references therein . | this paper is concerned with the concepts of regional controllability for the riemann - liouville time fractional diffusion systems of order @xmath0 .
the characterizations of strategic actuators to achieve regional controllability are investigated when the control inputs emerge in the differential equations as distributed inputs . in the end ,
an approach to guarantee the regional controllability of the problems under consideration in the considered subregion with minimum energy control is described and successfully tested through two applications . ,
@xmath1 , regional controllability ; time fractional diffusion systems ; strategic actuators ; minimum energy control . |
a partition of a positive integer @xmath1 is a sequence of positive integers @xmath2 with @xmath3 and @xmath4 .
the _ ( ordinary ) ferrers diagram _ of shape @xmath5 is an array of cells with @xmath6 left - justified rows and @xmath7 cells in row @xmath8 . figure [ ferrer].a shows the ferrers diagram corresponding to @xmath9 .
( 14.5,4 ) ( 1,1 ) ( 8,1 ) ( 1,0.5)a .
the ferrers diagram ( 1,0)corresponding to @xmath9 ( 8,0.5)b . the shifted ferrers diagram ( 8,0)corresponding to @xmath10 if @xmath5 is a partition with distinct components ( strict partition ) then the _ shifted ferrers diagram _ of shape @xmath5 is an array of cells with @xmath6 rows , each row indented by one cell to the right with respect to the previous row and @xmath7 cells in row @xmath8 .
figure [ ferrer].b shows the shifted ferrers diagram corresponding to @xmath10 . given a partition @xmath5 of @xmath1 , respectively strict partition of @xmath1 , a _
standard tableau _ , respectively a _ shifted standard tableau _ of shape @xmath5 , is a filling of the cells of the ordinary ferrers diagram , respectively shifted ferrers diagram , of shape @xmath5 with @xmath0 , such that the entries along rows and columns are increasing .
figure [ standard].a displays an example of a standard tableau of shape @xmath9 and figure [ standard].b displays an example of a shifted standard tableau of shape @xmath10 , ( 14.5,4 ) ( 1,1 ) ( 8,1 ) ( 1,0.5)a . a standard tableau ( 8,0.5)b . a shifted standard ( 8,0)tableau once we have accepted these definitions it is a natural question to ask for the number of standard tableaux , respectively shifted standard tableaux , of a given shape @xmath5 .
surprisingly there exists a simple product formula for these numbers .
it involves objects called _ hooks _ , which are defined in the following paragraph .
we label the cell in the @xmath8-th row and @xmath11-th column of the ordinary , respectively shifted , ferrers diagram of shape @xmath5 by the pair @xmath12 .
the _ hook _ of a cell @xmath12 in an _ ordinary _ ferrers diagram is the set of cells that are either in the same row as @xmath12 and to the right of @xmath12 , or in the same column as @xmath12 and below @xmath12 , @xmath12 included .
the dots in figure [ hook].a indicate the hook of the cell @xmath13 .
the _ hook _ of a cell @xmath12 in a _ shifted _ ferrers diagram again includes all cells that are either in the same row as @xmath12 and to the right of @xmath12 , or in the same column as @xmath12 and below @xmath12 , @xmath12 included , but if this set contains the cell @xmath14 on the main diagonal , then also the cells of the @xmath15-st row belong to the hook of @xmath12 .
the dots in figure [ hook].b indicate the hook of cell @xmath16 .
the _ hook - length _
@xmath17 of the cell @xmath12 is the number of cells in the hook of @xmath12 .
( 14.5,4 ) ( 1,1 ) ( 8,1 ) ( 1,0.5)a . the hook of @xmath13 ( 8,0.5)b . the shifted hook of @xmath16 now we are in the position to state the hook - length formula .
the number of standard tableaux , respectively shifted standard tableaux , of shape @xmath5 is @xmath18 where the product in the denominator is taken over all cells in the ferrers diagram , respectively shifted ferrers diagram , of shape @xmath5 .
thus the number of standard tableaux of shape @xmath9 is @xmath19 and the number of shifted standard tableaux of shape @xmath10 is @xmath20 . for the history of the various proofs of the hook - length formula for ordinary standard tableaux
see @xcite .
a recipe for an inductive proof of the hook - length formula for shifted standard tableaux can be found in @xcite . in @xcite
it is shown that the nice probabilistic proof of the hook - length formula for ordinary standard tableaux in @xcite has an analog for shifted standard tableaux .
a majority among the combinatorialists considers a bijective proof as the most aesthetic type of proof for an enumeration result . in @xcite a bijective proof of the hook - length formula for ordinary standard tableaux based on a modified jeu de taquin is given .
there exists a bijective proof of the hook - length formula for shifted standard tableaux as well @xcite , however , it makes use of the involution principle by garsia and milne .
the aim of this paper is to present an involution principle - free bijective proof of the hook - length formula for shifted standard tableaux , which is in the spirit of the beautiful bijective proof in @xcite .
if we discover two sets @xmath21 of combinatorial objects with the same cardinality , we often believe that this fact is a projection of a canonical bijection between the two sets .
such a bijection is called a bijective proof of the equality @xmath22 .
now suppose we are in the following more general situation : there exists an integer @xmath23 such that @xmath24 .
then this fact could be a projection of a canonical @xmath23 to @xmath25 surjection from @xmath26 onto @xmath27 , i.e. a map from @xmath26 onto @xmath27 where every element in @xmath27 is assigned to exactly @xmath23 elements in @xmath26 . in the following @xmath28
denotes the set of shifted standard tableaux of shape @xmath5 .
a _ shifted tabloid _ of shape @xmath5 is an ( arbitrary ) filling of the cells of the shifted ferrers diagram of shape @xmath5 with @xmath0 .
we denote the set of shifted tabloids of shape @xmath5 by @xmath29 and observe that its cardinality is @xmath30 . in our main theorem ( theorem [ main ] ) we present a @xmath31 to @xmath25 surjection from @xmath32 onto @xmath28 , which is clearly a proof of the hook - length formula for shifted standard tableaux .
we prove theorem [ main ] by introducing the set of hook tabloids @xmath33 with @xmath34 and extending the surjection to a bijection from @xmath32 to @xmath35 .
the corresponding surjection from the set of ordinary tabloids onto the set of standard tableaux is similar to the surjection in the shifted case , see @xcite , the extension in the shifted case is however far more complicated compared to the ordinary case .
in this section we describe the ordering procedure which assigns to every ` scrambled ' shifted tabloid @xmath36 an ` ordered ' shifted standard tableau @xmath37 .
this map has the property that the number of shifted tabloids which are mapped to a fixed shifted standard tableau is @xmath31 .
the ordering procedure is based on a modified jeu de taquin , which we have to describe first
. * notation .
* let @xmath38 be a shifted tabloid of shape @xmath5 .
then @xmath39 denotes the entry of cell @xmath12 .
if the cell @xmath12 does not exist in the shifted ferrers diagram of shape @xmath5 , let @xmath40 .
for every @xmath41 , @xmath42 , there exists a unique cell @xmath12 in the shifted ferrers diagram of shape @xmath5 with @xmath43 .
we define @xmath44 .
* jeu de taquin in @xmath38 with entry @xmath41 and with respect to a set @xmath45 .
* let @xmath42 and @xmath45 be a set of cells in the shifted ferrers diagram of shape @xmath5 .
we define the routine _ jeu de taquin in @xmath38 with entry @xmath41 and with respect to @xmath45 _ inductively .
the output is another shifted tabloid @xmath46 of shape @xmath5 .
let @xmath44 .
set @xmath47 and stop if either @xmath48 or @xmath49 .
( in the latter case @xmath41 is said to be _
stable_. see figure [ stable ] . )
otherwise let @xmath50 be such that @xmath51 and let @xmath52 denote the tabloid we obtain by exchanging the entries @xmath41 and @xmath53 in @xmath38 .
( see figure [ unstable ] . )
next perform jeu de taquin in @xmath52 with entry @xmath41 and with respect to @xmath45 in order to obtain the final tabloid @xmath46 , i.e. we repeat this exchanging procedure with @xmath41 and either its current neighbour to the right or below until @xmath41 is either stable or in a cell of @xmath45 .
( for an example see figure [ jt ] . )
if @xmath45 is the empty set we omit ` with respect to @xmath45 ' .
the output tabloid @xmath46 is denoted by @xmath54 and the cell of @xmath41 in @xmath46 is denoted by @xmath55 . if @xmath45 is the @xmath56-th row of the shifted ferrers diagram we write @xmath57 and @xmath58 , respectively . * the total order . *
we define a total order on the cells of a shifted ferrers diagram of shape @xmath5
. this will be the order in which we perform the jeu de taquin just defined with the entries of a shifted tabloid .
a cell @xmath59 comes before cell @xmath60 if @xmath59 is in a lower row than @xmath60 or if both are in the same row but @xmath59 is to the right of @xmath60 .
phrased differently , to obtain the total order one starts with the rightmost cell in the last row and reads each row from right to left , beginning with the bottom row and continuing up to the first row .
figure [ ord ] displays this total order for the shifted ferrers diagram of shape @xmath10 . *
the @xmath31 to @xmath25 map from @xmath29 onto @xmath61 .
* let @xmath38 denote an arbitrary shifted tabloid of shape @xmath5 .
in order to construct the corresponding shifted standard tableau @xmath62 we perform step by step jeu de taquin with the entries of @xmath38 subject to the total order we have just defined and starting with the entry in the smallest cell . to be more accurate : consider the shifted tabloid @xmath63 of shape @xmath10 .
we construct the corresponding shifted standard tableau . according to the algorithm above we start with performing jeu de taquin with @xmath64 , but @xmath64 is stable .
the same is true for the entry @xmath65 . performing jeu de taquin with @xmath66 results in @xmath67 since @xmath68 and @xmath69 are stable the next change happens to be when performing jeu de taquin with entry @xmath70 .
@xmath71 we perform jeu de taquin with @xmath72 @xmath73 next with @xmath74 @xmath75 then with @xmath76 @xmath77 and with @xmath78 @xmath79 and finally with @xmath80 @xmath81 observe that the output tabloid @xmath62 is a shifted standard tableau by construction .
we denote it by @xmath82 .
we are in the position to state our main theorem .
note that the bijective proof in @xcite shows that a similar theorem is true for ordinary standard tableaux .
[ main ] the map @xmath83 is a @xmath31 to @xmath25 map from the set of shifted tabloids @xmath29 onto the set of shifted standard tableaux @xmath61 . as an interesting side - result we obtain a random algorithm which produces every shifted standard tableau of a given shape with the same probability .
the following algorithm produces every shifted standard tableau of a given shape @xmath5 with the same probability . 1 .
generate a permutation @xmath84 of @xmath85 subject to uniform distribution .
2 . construct the corresponding shifted tabloid @xmath86 of shape @xmath5 by filling the elements from @xmath84 into the shifted ferrers diagram of shape @xmath5 rowwise from top to bottom and in each row from left to right .
3 . apply @xmath87 in order to obtain the shifted standard tableau .
note that novelli , pak and stoyanovskii @xcite use another total order of the cells in the ferrers diagram , they perform jeu de taquin columnwise from right to left and within a column from bottom to top . by transposing
their algorithm it is clear that the order we defined in the shifted case would also induce a @xmath88 to @xmath25 map in the ordinary case .
however , computer experiments with the strict partition @xmath89 have shown that the order defined by novelli , pak and stoyanovskii is not admissible in the shifted case .
moreover it seems that the total order we have defined is the only admissible in the shifted case , whereas in the ordinary case there exist many total orders with the property that they induce a @xmath88 to @xmath25 map .
we plan to discuss this phenomenon in a forthcoming paper .
we prove theorem [ main ] by giving a bijective proof of the hook - length formula .
for that purpose we rewrite the hook - length formula as @xmath90 we define combinatorial objects that correspond to @xmath31 in this formula in our bijective proof .
a _ shifted hook tabloid _ of shape @xmath5 is a filling of the cells of the shifted ferrers diagram of shape @xmath5 with pairs of integers , such that the entry in a cell @xmath91 are the coordinates of a cell in the hook of @xmath91 .
( see figure [ shbsp ] .
this definition of a shifted hook tabloid has a natural analog for ordinary ferrers diagram , which is equivalent to the definition of a hook function in @xcite . )
we denote the set of shifted hook tabloids by @xmath33 . since @xmath92 it suffices to find a bijection between the set of shifted tabloids @xmath29 and the cartesian product of the set of shifted standard tableaux @xmath61 and the set of shifted hook tabloids @xmath33 to prove the hook - length formula for shifted standard tableaux .
@xmath93 in order to prove theorem [ main ] we construct such a bijection , where the shifted standard tableau is obtained from the shifted tabloid by the modified jeu de taquin described above .
unfortunately the second component of the bijection , i.e. the construction of the shifted hook tableau , is complicated .
thus it would be nice to find an easier - to - describe bijection .
also a shorter proof of theorem [ main ] would be of interest .
maybe the present paper serves as an inspiration in this task .
we give some further definitions we need for the rest of the paper .
* reverse jeu de taquin in @xmath46 with entry @xmath41 and with respect to a set @xmath45 .
* we define an inverse to jeu de taquin , which we need to construct the inverse of the map which assigns a pair of a shifted standard tableau and a shifted hook tabloid to a given shifted tabloid . in all our algorithms where we use reverse jeu de taquin we fix a row @xmath8
. reverse jeu de taquin depends on this row .
the output of reverse jeu de taquin in @xmath46 with entry @xmath41 and with respect to @xmath45 is again another shifted tabloid @xmath38 of shape @xmath5 .
let @xmath94 .
set @xmath95 and stop if either @xmath96 or @xmath97 .
otherwise let @xmath98 be such that @xmath99 if @xmath100 , @xmath101 if @xmath102 and @xmath103 if @xmath104 and let @xmath105 denote the tabloid we obtain by exchanging the entries @xmath41 and @xmath106 in @xmath46 . next perform jeu de taquin in @xmath105 with entry @xmath41 and with respect to @xmath45 in order to obtain the final tabloid @xmath38 .
( for an example read figure [ jt ] from right to left .
the left shifted tabloid can be obtained from the right by performing reverse jeu de taquin with @xmath107 and with respect to @xmath108 . )
if @xmath45 is the empty set we omit ` with respect to @xmath45 ' . in the following the output tabloid @xmath38
is denoted by @xmath109 and the cell of @xmath41 in @xmath38 is denoted by @xmath110 . if @xmath45 is the @xmath56-th row we write @xmath111 and @xmath112 , respectively .
let @xmath38 be a shifted tabloid of shape @xmath5 and @xmath42 .
the _ forward path of @xmath41 in @xmath38 with respect to @xmath45 _ is the set of cells @xmath41 comes across when performing jeu de taquin in @xmath38 with @xmath41 and with respect to @xmath45 .
for example the forward path of @xmath107 with respect to the cells on the main diagonal in the left tabloid of figure [ jt ] is @xmath113 . similarly the _
backward path in @xmath38 of @xmath41 with respect to @xmath45 _ is the set of cells @xmath41 comes across when performing reverse jeu de taquin in @xmath38 with @xmath41 and with respect to @xmath45 .
clearly the backward path of @xmath107 with respect to @xmath108 in the right shifted tabloid in figure [ jt ] coincides with the forward path in the left shifted tabloid .
we need one more definition before we are able to relate jeu de taquin and reverse jeu de taquin .
let @xmath38 be a shifted tabloid of shape @xmath5 and @xmath91 a cell in the shifted ferrers diagram of shape @xmath5 .
the shifted tabloid @xmath38 is said to be _ ordered up to cell @xmath91 _ if rows and columns are increasing in the subtabloid consisting of the cells that are smaller or equal to @xmath91 with respect to the total order .
( note that in figure [ jt ] the left tabloid is ordered up to @xmath114 . )
observe that the routines _ jeu de taquin _ and _ reverse jeu de taquin _ are inverse to each other in the following sense : let @xmath38 be a shifted tabloid and @xmath41 an entry in @xmath38 with @xmath44 .
suppose @xmath38 is ordered up to the predecessor of @xmath12 in the total order .
perform jeu de taquin in @xmath38 with @xmath41 and with respect to @xmath45 and obtain the tabloid @xmath46 . if we perform reverse jeu de taquin in @xmath46 with @xmath41 and with respect to @xmath12 ( @xmath8 being the fixed row ) , we reobtain @xmath38 . in symbols
@xmath115 furthermore observe that @xmath46 is ordered up to @xmath12 if @xmath116 .
conversely : let @xmath41 be an entry weakly below the @xmath8-th row of a shifted tabloid @xmath38 that is ordered up to @xmath12 and assume that the backward path of @xmath41 in @xmath38 contains @xmath12 ( @xmath8 being the fixed row ) .
perform reverse jeu de taquin in @xmath38 with @xmath41 and respect to @xmath12 and denote the output tabloid by @xmath46 . if we perform jeu de taquin in @xmath46 with @xmath41 and with respect to @xmath117 , we reobtain @xmath38 . in symbols
@xmath118 furthermore observe that @xmath46 is ordered up to the predecessor of @xmath12 .
therefore : _ in order to be able to reconstruct @xmath38 from @xmath82 we have to store the endcells of the forward paths we obtain in the application of modified jeu de taquin to @xmath38 in the shifted hook tabloid . since we want to obtain a bijection this storage has to be organized most efficiently .
the fact that the endcells are not independent from each other ( see lemma [ 1 ] ) makes this storage non - trivial .
_ * backward paths order .
* we define the _ backward paths order _ on the entries of a shifted tabloid @xmath38 .
let @xmath119 , @xmath120 be two entries in @xmath38 and @xmath121 , @xmath122 their backward paths ( @xmath25 being the fixed row ) in @xmath38 .
furthermore let @xmath12 denote the smallest cell in @xmath123 with respect to the total order .
if either @xmath124 or @xmath125 we define @xmath126 and say that @xmath120 is greater than @xmath119 with respect to the backward paths order in @xmath38 . for an example see figure [ bpo ] .
we introduce a manner - of - speaking : let @xmath38 be a shifted tabloid and @xmath127 an accompanying partial shifted hook tabloid , a partial shifted hook tabloid being a filling of some cells of the shifted ferrer diagram with entries satisfying the requirement for the entries in a shifted hook tabloid .
let @xmath91 be an entry of a cell @xmath128 in the fixed @xmath8-th row of @xmath127 , i.e. @xmath129 .
we say that @xmath53 is a _ horizontal candidate _ in @xmath38 with respect to @xmath127 , if @xmath91 is in the same column as @xmath128 in the shifted ferrers diagram .
for example consider the shifted hook tabloid in figure [ shbsp ] and an arbitrary shifted tabloid @xmath38 of shape @xmath10 . if @xmath130 then @xmath131 , @xmath132 and @xmath133 are the horizontal candidates .
if @xmath91 is neither in the same row nor in the same column as @xmath128 , @xmath53 is a _ vertical candidate _ with respect to @xmath127 . in our example in figure [ shbsp ]
@xmath134 is the only vertical candidate if @xmath130 .
the rest of the paper is organized as follows . in section [ alg ]
we describe the algorithm split that converts a shifted tabloid of shape @xmath5 into a pair of a shifted standard tableau of shape @xmath5 and a shifted hook tabloid of shape @xmath5 . in section [ examples ]
we give some examples of the application of split . in section [ inv ]
we describe the algorithm merge that ` merges ' a pair of a shifted standard tableau and a shifted hook tabloid to a shifted tabloid . in section [ proof ]
we prove that the algorithm merge is the inverse of the algorithm split .
in this section we describe the algorithm split that transforms a shifted tabloid into a pair of a shifted standard tableau and a shifted hook tabloid .
the construction of the shifted hook tabloid is more involved compared to the construction of the shifted standard tabloid , which we have already described .
it depends on the endcells of the forward paths we obtain in the course of performing jeu de taquin in the shifted tabloid and on the intersection of two such paths near the main diagonal .
we have to modify the order of some steps in the construction of the shifted standard tableau so that the shifted hook tabloid can be built up simultaneously .
but since these steps in question commute the modified algorithm for building the shifted standard tableau is equivalent to the origin algorithm .
before we are in the position to describe the algorithm , we introduce two routines on a partial shifted hook tabloid @xmath127 of shape @xmath5 . * a shift from cell @xmath12 to cell @xmath135 in a partial shifted hook tabloid .
* let @xmath12 , @xmath135 be two cells in @xmath127 , @xmath136 .
we define the term _ shift from @xmath12 to @xmath135 in @xmath127_. the output of this operation is another partial shifted hook tabloid @xmath137 which coincides with @xmath127 except for the cells @xmath138 , @xmath139 .
let @xmath140 . if @xmath141 , @xmath142 , or @xmath143 then set @xmath144 or @xmath145 , respectively .
if @xmath146 then set @xmath147 .
the cell @xmath135 in @xmath137 remains empty .
we denote @xmath137 by @xmath148 . for an example
see figure [ shift ] . whenever we perform jeu de taquin with an entry @xmath41 and with respect to a set @xmath45 in the shifted tabloid @xmath38 in the algorithms below
, we store the end of the forward path in the accompanying partial shifted hook tabloid @xmath127 .
if we are in the course of performing jeu de taquin with the entries in the @xmath8-th row and @xmath149 is the end of the forward path , this is either done by setting @xmath150 or by setting @xmath151 .
the latter possibility is used if @xmath41 was previously a vertical candidate . by reverse
jeu de taquin the knowledge of the end of the forward path is enough to undo the performance of jeu de taquin .
however , when @xmath152 , respectively @xmath153 , is occupied we have to perform a shift from the cell in @xmath127 that has previously pointed to @xmath41 to @xmath135 , respectively @xmath154 , in order to empty either @xmath135 or @xmath154 . more accurate : if @xmath41 is either a horizontal candidate or no candidate in @xmath155 we denote by @xmath156 the pair of a shifted tabloid @xmath52 and a shifted hook tabloid @xmath137 , which is obtained in the following way : @xmath157 , @xmath158 and @xmath159 , where @xmath160 is the column of @xmath41 in @xmath38 and @xmath161 . if @xmath41 is a vertical candidate we denote by @xmath156 the pair of a shifted tabloid @xmath52 and a shifted hook tabloid @xmath137 , which is obtained in the following way : @xmath157 , @xmath162 and @xmath163 , where @xmath164 is the row of @xmath41 in @xmath38 and @xmath161 .
if @xmath45 is empty we write @xmath165 and if @xmath45 is the @xmath56-th row we write @xmath166 . * a transfer from cell @xmath12 to cell @xmath138 in a partial shifted hook tabloid . *
let @xmath167 and @xmath168 with either @xmath102 or @xmath169 .
we define the term _ transfer from cell @xmath12 to cell @xmath138 _ in @xmath127 .
the output of this operation is again another partial shifted hook tabloid @xmath137 which coincides with @xmath127 except for the cells @xmath170 , @xmath171 . if @xmath102 and @xmath172 let @xmath173 , otherwise let @xmath174 . for @xmath175 let @xmath176 .
we denote @xmath137 by @xmath177 . for an example
see figure [ trans ] .
whenever we apply a transfer from @xmath12 to @xmath138 in @xmath127 we have @xmath178 for @xmath179 as in the example .
this operation we need in the algorithms for converting horizontal candidates into vertical candidates .
the algorithm split is divided into @xmath6 steps , where in the @xmath8-th step we perform jeu de taquin with the entries in the @xmath180-st row . within a row @xmath8 , split is divided into 3 steps , split 1 , split 2 and split 3 .
assume we just start performing jeu de taquin with the entries in the @xmath8-th row .
let @xmath38 denote the shifted tabloid we have constructed so far and @xmath127 the partial shifted hook tabloid .
at this point @xmath38 is ordered up to @xmath181 ( as it is in the origin algorithm for constructing the shifted standard tableau in section [ jeu ] ) , the first @xmath8 rows of @xmath127 are empty and the last @xmath182 rows of @xmath127 form a shifted hook tabloid . in split
1 we perform jeu de taquin with the entries in the @xmath8-th row from right to left and with respect to the set @xmath183 of cells on the main diagonal . after split 1 for every entry @xmath41 whose forward path terminates in split 1 in a cell @xmath184 on the main diagonal , we have @xmath185 and @xmath186 and therefore all unstable entries in @xmath38 are horizontal candidates with respect to @xmath127 after the application of split 1 .
this follows from the fact that whenever a forward path of an entry @xmath187 ends in a cell @xmath184 on the main diagonal , then the forward paths of the following entries @xmath188 , @xmath189 , end strictly left of the @xmath56-th column .
( see lemma [ 6 ] . )
observe the following ( see also figure [ split2 ] . ) : consider the backward paths with respect to the @xmath190-th row of the horizontal candidates strictly below the @xmath190-th row in the output tabloid of split 2 .
then either one of these paths ends weakly to the left of the vertical candidate in the @xmath190-th row or there exists no vertical candidate if and only if we were in case 2 of split 2 .
( if we were in case 1 of split 2 and the backward path of a horizontal candidate strictly below the @xmath190-th row would end weakly to the left of the vertical candidate in the @xmath190-th row after the application of case 1 split 2 then @xmath191 by the cases distinction in split 2 .
thus the backward path of the same horizontal candidate includes the cell @xmath192 on the main diagonal before the application of case 1 in split 2 ( by the argument in the proof of lemma [ 1 ] ) which implies @xmath193 ( by property after_split 1 in the proof of claim 1 in section [ proof ] ) and this is not possible by the choice of @xmath190 for our assumptions imply that @xmath194 is unstable before the application of case 1 in split 2 .
furthermore there always exists a vertical candidate after the performance of case 1 of split 2 . ) in other words : we were in case 2 of split 2 if and only if either the smallest horizontal candidate strictly below the row of the smallest vertical candidate is smaller than this smallest vertical candidate or there exists no vertical candidate in the output pair . besides the pair @xmath155 we need two sets of entries @xmath195 and @xmath196 as an input for split 3 .
the set @xmath195 contains the vertical candidates together with the smallest candidate with respect to the backward paths order if it is unstable .
the set @xmath196 contains the horizontal candidates on the main diagonal of @xmath38 which are not in @xmath195 .
observe that the entries in @xmath195 are strictly above of the entries in @xmath196 , i.e. the maximal row of an entry in @xmath195 is smaller than the minimal row of an entry in @xmath196 .
this will also be true in the course of performing split 3 .
similarly all vertical candidates are strictly above of the entries in @xmath196 before and while performing split 3 . in figure [ split3 ]
an example of case 2 in split 3 is displayed .
let @xmath38 be a shifted tabloid ordered up to @xmath181 and @xmath127 a partial shifted hook tabloid .
let @xmath197 denote the pair we obtain after the application of split 1 , split 2 and split 3 to the @xmath8-th row of @xmath155 .
we denote @xmath198 and @xmath199 .
we are finally in the position to formulate the algorithm split .
the input is a shifted tabloid @xmath38 of shape @xmath5 and an empty shifted tabloid @xmath137 of shape @xmath5 .
the output @xmath52 and @xmath137 of the algorithm is denoted by @xmath200 and @xmath201 .
in this section we give five examples for the application of split .
all examples are of shape @xmath202 and the input shifted tabloid is ordered up to @xmath203 .
we perform jeu de taquin with the entries in the first row of the shifted tabloid and simultaneously built up the first row of the shifted hook tabloid .
our examples cover all possible cases in split 2 : in example [ ex1 ] we are in case 2 of split 2 for @xmath23 does not exist , in example [ best ] we are in case 2 of split 2 for @xmath204 , in example [ ex3 ] we are in case 1 of split 2 for @xmath205 , in example [ ex4 ] we are in case 1 of split 2 with @xmath206 and in example [ ex5 ] we are in case 2 of split 2 with @xmath206 .
[ ex1 ] we consider the shifted tabloid @xmath207 first we have to apply split 1 to the first row .
if we perform jeu de taquin with @xmath208 and with respect to the cells on the main diagonal , the entry ends in cell @xmath209 .
@xmath210 thus we set @xmath211 and therefore the first row of the shifted hook tabloid is @xmath212 the entry @xmath213 is stable , thus the shifted tabloid does not change in the next step and we set @xmath214 , the cell of @xmath213 .
@xmath215 next we perform jeu de taquin with @xmath216 .
@xmath217 the forward path ends in cell @xmath218 and therefore we want to set @xmath219 .
however , @xmath220 is occupied , thus we perform a shift from @xmath221 to @xmath222 .
@xmath223 if we perform jeu de taquin with @xmath224 and with respect to the cells on the main diagonal , the entry gets stuck in @xmath225 and remains unstable there .
@xmath226 again we want to set @xmath227 , but @xmath228 is occupied and therefore we have to perform a shift .
@xmath229 next we perform jeu de taquin with @xmath230 and this entry again gets stuck in a cell on the main diagonal .
then we perform jeu de taquin with @xmath231 , @xmath232 , @xmath233 , @xmath234 , @xmath107 and finally with @xmath235 . these entries either end in a stable position or on the main diagonal .
simultaneously we built up the first row of the shifted hook tabloid .
there we put down the endcells of the paths in the appropriate columns and whenever we want to use a cell of the shifted hook tabloid which is already occupied we perform the appropriate shift .
this procedure results in @xmath236 and @xmath237 we apply split 2 to this pair . observe that @xmath238 since @xmath239 for @xmath240 and @xmath235 , @xmath107 , @xmath234 and @xmath233 are unstable .
we perform jeu de taquin with @xmath232 , @xmath233 , @xmath234 , @xmath107 and @xmath235 in that order and with repsect to the relative next row .
@xmath241 we observe that every entry changes row and therefore @xmath23 does not exist .
consequently we are in case 2 of split 2 .
we set @xmath95 .
since @xmath232 is in cell @xmath242 , we set @xmath243 after performing a shift , @xmath233 is in cell @xmath244 and therefore @xmath245 , @xmath234 is in cell @xmath246 and therefore @xmath247 , @xmath107 is in cell @xmath248 and therefore @xmath249 and @xmath235 is in cell @xmath250 and therefore @xmath251 .
thus @xmath252 finally we apply split 3 . observe that @xmath253 and @xmath254 . in the first step of split 3 @xmath255 and @xmath256
. the forward path of @xmath41 ends left of the column of @xmath257 and therefore we are in case 1 .
we obtain @xmath258 and @xmath259 we delete @xmath232 from @xmath195 : @xmath260 and @xmath254 . in the next step of split 3
we have @xmath261 and @xmath256 .
since the forward path of @xmath233 ends weakly right of the column of @xmath230 but does not contain @xmath262 , we are in case 3 . here
we do not change @xmath38 or @xmath127 , we only delete @xmath230 from @xmath196 and put it into @xmath195 : @xmath263 and @xmath264 .
now @xmath265 and @xmath266 and we are in case 2 for the forward path of @xmath230 contains @xmath267 and @xmath268 . therefore we want to change @xmath230 into a vertical candidate by setting @xmath269 , but since @xmath270 is occupied we perform a transfer from @xmath271 to @xmath272 . @xmath273 the shifted tabloid @xmath38 does not change .
furthermore @xmath274 , @xmath275 and we are always in case 1 for the rest of the application of split 3 .
next @xmath276 and its forward path ends in @xmath277 . therefore we want to set @xmath278 and since @xmath220 is occupied we perform a shift from the previous pointer @xmath221 of @xmath224 to @xmath222 .
we obtain @xmath279 @xmath280 and @xmath263 .
the forward path of @xmath265 ends in @xmath281 .
therefore we want to set @xmath282 and since @xmath283 is occupied we perform a shift from @xmath271 to @xmath272 .
this yields @xmath284 moreover @xmath285 and @xmath260 .
in the next step @xmath261 etc .
@xmath286 finally we obtain the following shifted standard tableau @xmath287 and the following first row of @xmath127 @xmath288 [ best ] we consider @xmath289 after the application of split 1 we obtain the pair @xmath290 @xmath291 since @xmath102 we are in case 2 of split 2
. we move @xmath107 to the second row @xmath292 and set @xmath293 after performing the appropriate shift @xmath294 we set @xmath295 and @xmath296 . in the first step of split 3
we have @xmath297 and @xmath298 .
since the forward path of @xmath107 includes @xmath203 and @xmath299 we are in case 2 of split 3 . the shifted tabloid @xmath38 does not change , but we want to set @xmath300 and therefore perform a transfer from @xmath301 to @xmath16 in @xmath127 .
@xmath302 furthermore @xmath303 and @xmath304 .
next @xmath305 and @xmath306 .
again we are in case 2 .
@xmath307 we set @xmath308 after performing an appropriate shift and a transfer from @xmath108 to @xmath114 .
@xmath309 we update @xmath310 and @xmath311 .
next @xmath312 and @xmath313 and since @xmath314 is stable we are in case 1 .
thus @xmath38 and @xmath127 do not change , but @xmath303 and @xmath311 . therefore @xmath305 and @xmath313 and we are in case 1 .
@xmath315 @xmath316 moreover @xmath295 and @xmath311 . in the next step @xmath297 and @xmath313 .
because of the run of the forward path of @xmath107 we are in case 3 .
thus @xmath38 and @xmath127 do not change , only @xmath317 and @xmath318 . therefore @xmath319 and @xmath320 .
we are in case 2 and @xmath38 does not change .
we want to set @xmath321 .
since @xmath322 is occupied by a pointer which does not point to @xmath323 , we perform a transfer from @xmath324 to @xmath271 .
@xmath325 moreover @xmath326 and @xmath327 .
next @xmath328 and @xmath329 .
we are in case 1 for @xmath330 is stable and thus @xmath317 and @xmath327 . therefore @xmath319 and @xmath329 .
again we are in case 1 .
@xmath331 @xmath332 moreover @xmath295 and @xmath327 .
now we have @xmath297 and @xmath333 .
we are in case 2 .
@xmath334 since @xmath107 was already a vertical candidate we want to set @xmath335 . for that purpose
we perform a transfer from @xmath221 to @xmath336 .
@xmath337 we update @xmath338 and @xmath275 .
now @xmath235 is stable and @xmath295 after the next step .
we terminate with the pair @xmath339
@xmath340 [ ex3 ] let @xmath341 after the application of split 1 we obtain @xmath342 and the first row of the shifted hook tabloid is @xmath343 we apply split 2 to the pair .
observe that @xmath344 since @xmath239 for @xmath345 and the entries @xmath346 , @xmath213 are unstable . if we perform jeu de taquin with @xmath347 , @xmath213 and @xmath346 and with respect to the relative next row we observe that @xmath346 does not change row and thus @xmath205 .
therefore we are in case 1 of split 2 .
we perform jeu de taquin with @xmath347 , @xmath213 and @xmath346 in that order and with respect to the last cell in the current row of the forward path in the course of constructing @xmath46 .
this results in @xmath348 since @xmath347 moved to cell @xmath349 we set @xmath350 , since @xmath213 moved to @xmath351 we set @xmath352 and since @xmath346 moved to @xmath221 we set @xmath353 .
@xmath354 after the application of split 3 we obtain @xmath355 and @xmath356 [ ex4 ] let @xmath357 after the application of split 1 we obtain the pair @xmath358 and @xmath359 next we apply split 2 .
observe that @xmath360 ( since @xmath64 is stable ) and that @xmath361 thus @xmath362 and no backward paths with respect to the @xmath78-th row of a horizontal candidate strictly below the @xmath78-th row ends weakly left of @xmath107 .
therefore we are in case 1 of split 2 .
we set @xmath47 .
since @xmath64 original ended in @xmath363 we set @xmath364 , since @xmath107 is in @xmath363 we set @xmath365 , since @xmath76 is in @xmath349 we set @xmath350 and since @xmath231 is in @xmath366 we set @xmath367 .
@xmath368 after the performance of split 3 we obtain the pair @xmath369 and @xmath370 [ ex5 ] we consider @xmath371 after the application of split 1 we obtain @xmath372 and @xmath373 next we apply split 2 . observe that @xmath374 and that @xmath375 thus @xmath376 and the backward path with respect to the @xmath69-nd row of the horizontal candidate @xmath377 ends weakly left of @xmath378 .
we are in case 2 of split 2 and set @xmath95 and @xmath379 after the application of split 3 we finally obtain @xmath380 and @xmath381
now we describe the algorithm merge that merges a pair of a shifted standard tableau and a shifted hook tabloid to a shifted tabloid .
this will turn out to be the inverse of the algorithm split .
again we have to introduce two routines on a partial shifted hook tabloid @xmath127 of shape @xmath5 . * a reshift from cell @xmath135 to cell @xmath12 in a partial shifted hook tabloid .
* let @xmath12 , @xmath135 be two cells in @xmath127 , @xmath136 .
we define the term _ reshift from @xmath135 to @xmath12 in @xmath127_. the output of this operation is another shifted tabloid @xmath137 which coincides with @xmath127 except for the cells @xmath138 , @xmath139 .
let @xmath382 . if @xmath383 or @xmath384 then set @xmath385 or @xmath386 , respectively . if @xmath387 , @xmath388 , then set @xmath147 .
the cell @xmath12 in @xmath137 remains empty .
we denote @xmath137 by @xmath389 .
for an example for a reshift from @xmath324 to @xmath16 read figure [ shift ] from right to left . observe that @xmath390 and @xmath391 are inverse to each other in the following sense : let @xmath392 and @xmath127 be a partial shifted hook tabloid in which only cell @xmath12 is empty .
then @xmath393 if @xmath137 is a partial shifted hook tabloid where only cell @xmath135 is empty then @xmath394 in general the output of a reshift in a partial shifted hook tabloid need not to be a partial shifted hook tabloid for @xmath395 , respectively @xmath396 , need not to be a cell in the shifted ferrers diagram of shape @xmath5 if @xmath397 , respectively @xmath398 , is a cell in this diagram .
however , in the following two situations the reshift is indeed a partial shifted hook tabloid and we apply the reshift only if one of these situations encounters . 1 .
let @xmath38 be a shifted tabloid and @xmath127 an accompanying partial shifted hook tabloid .
let @xmath41 be a vertical candidate in @xmath38 with respect to @xmath127 and the @xmath8-th row ( meaning that the pointer to @xmath41 is in the @xmath8-th row of @xmath127)and perform reverse jeu de taquin ( @xmath8 being the fixed row ) with @xmath41 in @xmath38 and with respect to a set @xmath45 .
let @xmath399 denote the cell of @xmath41 in the output tabloid and @xmath400 denote the cell of @xmath41 in @xmath38 .
suppose that @xmath41 is the maximal vertical candidate in @xmath38 under the vertical candidates in the rows @xmath23 , @xmath401 , with respect to the backward paths order of @xmath38 .
then the reshift from cell @xmath402 to cell @xmath403 in @xmath127 produces another partial shifted hook tabloid . 2 .
again let @xmath38 be a shifted tabloid and @xmath127 an accompanying partial shifted hook tabloid .
furthermore we assume in this situation that there exists no vertical candidate in @xmath38 with respect to @xmath127 .
let @xmath41 be a horizontal candidate in @xmath38 with respect to @xmath127 and the @xmath8-th row and perform reverse jeu de taquin ( @xmath8 being the fixed row ) with @xmath41 in @xmath38 and with respect to a set @xmath45 .
let @xmath399 denote the cell of @xmath41 in the output tabloid and @xmath400 denote the cell of @xmath41 in @xmath38 .
suppose that @xmath41 is the smallest horizontal candidate in @xmath38 under the horizontal candidates in the columns @xmath56 , @xmath139 , with respect to the backward paths order of @xmath38 .
then the reshift from cell @xmath135 to cell @xmath12 in @xmath127 produces another partial shifted hook tabloid .
( in order to see that use lemma [ 0 ] . ) as jeu de taquin in split is mostly followed by a shift in the accompanying shifted hook tabloid , reverse jeu de taquin in merge is mostly followed by a reshift .
let @xmath38 be a shifted tabloid , @xmath127 a shifted hook tabloid , @xmath41 a candidate and @xmath45 a set of cells .
if @xmath41 is a horizontal candidate then @xmath404 is obtained as follows : @xmath405 , @xmath406 and @xmath159 , where @xmath160 is the column of @xmath41 in @xmath38 and @xmath161 .
if @xmath41 is a vertical candidate then @xmath404 is obtained as follows : @xmath405 , @xmath407 and @xmath163 , where @xmath164 is the row of @xmath41 in @xmath38 and @xmath161 .
observe that @xmath408 and @xmath409 are each other s respective inverses . * a retransfer from cell @xmath138 to cell @xmath12 in a partial shifted hook tabloid .
* let @xmath167 and @xmath410 with either @xmath102 or @xmath411 .
we define the term _ retransfer from cell @xmath138 to cell @xmath12 in @xmath127 . _
the output of this operation is again another partial shifted hook tabloid @xmath137 which coincides with @xmath127 except for the cells @xmath170 , @xmath171 . if @xmath411 and @xmath412 let @xmath413 , otherwise @xmath414 . for @xmath179
set @xmath415 .
we denote @xmath137 by @xmath416 . for an example of a retransfer from @xmath324 to @xmath16 read figure [ trans ] from right to left .
again , whenever we will apply a retransfer from @xmath138 to @xmath12 in @xmath127 , we have @xmath417 for @xmath175 as in the example .
observe that @xmath418 and @xmath419 are inverse to each other in the following sense .
let @xmath420 .
if @xmath127 is a shifted hook tabloid such that @xmath421 is in the same row or column as @xmath12 and with @xmath178 for @xmath179 then @xmath422 if @xmath137 is a shifted hook tabloid such that @xmath423 is in the same row or column as @xmath138 and with @xmath176 for @xmath175 then @xmath424 now we are in the position to state the algorithm merge .
it is divided into @xmath6 steps , where in the @xmath8-th step we perform reverse jeu de taquin with the entries in the @xmath8-th row . immediately
before we perform the @xmath8-th step of merge to a pair @xmath155 , @xmath38 is ordered up to @xmath425 , the first @xmath426 rows of @xmath127 are empty and the last @xmath427 rows form a shifted hook tabloid . within a row @xmath8 ,
merge is divided into @xmath68 steps .
the @xmath68 parts of merge in a fixed row @xmath8 are numbered in reverse order to emphasize the connection between split @xmath56 and merge @xmath56 , @xmath428 .
all the horizontal and vertical candidates below are meant to be with respect to the @xmath8-th row , i.e. their pointers are in the @xmath8-th row of the current shifted hook tabloid .
the marker (@xmath429) in merge 3 is needed in the proof that split 3 and merge 3 are inverse to each other in subsection [ proof].6 .
observe that in case the if - condition ( if @xmath430 : [ @xmath286 ] ) at the beginning of a step of merge 3 is true , there exists no vertical candidate strictly below the @xmath23-th row by the maximality of @xmath41 . let @xmath38 be a shifted tabloid which is ordered up to @xmath425 and @xmath127 a partial shifted hook tabloid , where no entry in the @xmath8-th row is empty .
let @xmath197 denote the pair we obtain after the application of merge 3 , merge 2 and merge 1 to the @xmath8-th row of @xmath155 .
we denote @xmath432 and note that @xmath137 is equal to @xmath127 with the entries in the @xmath8-th row deleted .
we invite the reader to apply merge 3 , merge 2 and merge 1 to the first rows of the output pairs in the examples in section [ examples ] in order to find out that these applications result in the input pairs .
we are finally in the position to formulate the algorithm merge .
the input is a shifted standard tableaux @xmath38 of shape @xmath5 and a shifted hook tabloid @xmath127 of shape @xmath5 .
the output tabloid @xmath52 of the algorithm is denoted by @xmath433 .
note that the output tabloid @xmath137 is empty . in the remaining section we aim to show the following theorem .
[ key ] let @xmath434 be a partition with distinct components and @xmath435 . 1 .
let @xmath38 be a shifted tabloid of shape @xmath5 which is ordered up to @xmath181 and @xmath127 a partial shifted hook tabloid of shape @xmath5 , where the cells of the @xmath8-th row are empty .
let @xmath436 and @xmath199 .
then @xmath437 .
2 . let @xmath38 be a shifted tabloid of shape @xmath5 which is ordered up to @xmath425 and @xmath127 a partial shifted hook tabloid of shape @xmath5 , where no cell in the @xmath8-th row is empty .
let @xmath432 and @xmath137 denote the partial shifted hook tabloid @xmath127 with the entries in the @xmath8-th row deleted
. then @xmath438 and @xmath439 .
this theorem , once it is proved , shows that split is the desired algorithm , i.e. a bijection from @xmath32 onto @xmath35 , for merge is its inverse . in the following we fix a strict partition @xmath434 and a row @xmath8 , @xmath435 .
6.1 . variantes of split 1 and split 2 we modify split 1 and split 2 such that after this modification merge @xmath56 is the inverse of split @xmath56 , @xmath428 .
the original version of split and the modification of split , where we replace split @xmath56 , @xmath440 , by their modifications described below , are equivalent , since we only change the order of some commuting steps and swap the beginning of split 3 to the end of split 2 .
again the following holds for @xmath155 after the application of post - split 2 : we were in case 2 of post - split 2 if and only if either the smallest horizontal candidate strictly below the @xmath441-th row is smaller than every vertical candidate or no vertical candidate exists .
* touching from above .
* let @xmath155 denote a pair to which we apply split 2 , post - split 2 and split 3 .
assume that we are in case 2 in a certain step of split 3 .
then we say _
@xmath41 touches @xmath257 from above _ and for @xmath442 , @xmath443 touches @xmath444 from above .
if @xmath41 is a horizontal candidate before the application of case 2 we say that @xmath445 is the _ place of change for @xmath41 _ and @xmath14 is the place of change for @xmath444 , @xmath446 .
if @xmath41 is already a vertical candidate before the application of case 2 we say that @xmath14 is the place of change for @xmath444 , @xmath447 . furthermore
: if we are in case 1 of split 2 let @xmath448 , @xmath449 , be the exceptional entries , where @xmath38 denotes the shifted tabloid at the beginning of split 2 .
if we are in case 2 of split 2 let @xmath448 , @xmath449 and @xmath450 , be the exceptional entries , @xmath38 again being the shifted tabloid at the beginning of split 2 . if @xmath451 and we are in case 2 of split 2 then the vertical candidate in row @xmath452 after the application of split 2 is defined to be an entry which touches from above with place of change @xmath192 .
we replace split 1 by the following variante . again
the input is a shifted tabloid @xmath38 , which is ordered up to @xmath181 and a partial shifted hook tabloid @xmath127 , where the @xmath8-th row is empty . the change of split 1 forces us to replace split 2 by the following slight modification .
this is because for a pair @xmath155 which falls into case 2 in the original version of split 2 , the entry @xmath194 either moves to a stable position or to its place of change if we apply the variante of split 1 .
check that if we replace split 1 and split 2 in split by the variantes above , this variante of split is equivalent to the original version . in the following split
denotes this variante and split @xmath56 , @xmath428 , its parts .
furthermore split and split @xmath56 , @xmath428 , denote the versions that were valid before this paragraph .
note that after the application of split 2 the following holds : if we were in case 1 and @xmath453 is a horizontal candidate then @xmath454 is a stable vertical candidate , if we were in case 2 and @xmath453 and @xmath455 are horizontal candidates then @xmath453 is stable . in section [ alg ] , before the description of split 3
, we define a set @xmath196 , which we need as an input for split 3 .
note that we can omit the candidates in @xmath196 that neither touch nor are touched from above in the application of split 3 and consequently these entries are omitted in the following .
the lemmas in this subsection are needed in order to prove that split 1 and merge 1 are inverse to each other .
they have analogs that are used for showing that split 3 and merge 3 are each other s respective inverses as we see in subsection 6.4 .
we introduce some notation concerning the relative position of an entry and a path .
let @xmath41 and @xmath257 be two entries in a shifted tabloid @xmath38 in the @xmath8-th row and @xmath456 , @xmath457 . if @xmath458 , resp .
@xmath459 , we say that @xmath257 is right , resp .
weakly right , of @xmath41 .
if @xmath460 is a path in the shifted tabloid @xmath38 then @xmath257 is said to be right , respectively weakly right , of @xmath460 , if @xmath460 includes an entry @xmath41 such that @xmath257 is right , respectively weakly right , of @xmath41 .
similar definitions apply to left , weakly left , above , weakly above , below and weakly below . if the formulations of our lemmas and corollaries include phrases in brackets [ @xmath286 ] , the assertions are true with and without these phrases . 1 .
let @xmath462 and @xmath463 denote the backward path of @xmath464 .
then @xmath464 is the smallest entry in @xmath461 with respect to the backward paths order , if and only if every @xmath465 is either [ weakly right and ] above of @xmath463 or the backward path of @xmath466 enters the column of @xmath464 weakly above of @xmath464 .
2 . let @xmath467 and @xmath468 denote the backward path of @xmath469 .
then @xmath469 is the greatest entry in @xmath461 with respect to the backward paths order , if and only if every @xmath465 is either left [ and weakly below ] of @xmath468 or the backward path of @xmath466 enters the row of @xmath469 weakly left of @xmath469 .
we extend the definition of ` entry @xmath41 touches entry @xmath257 from above ' : let @xmath38 be a shifted tabloid , @xmath41 , @xmath257 two entries in @xmath38 and @xmath45 a set of cells in the associated shifted ferrers diagram .
let @xmath460 be the forward path of @xmath41 with respect to @xmath45 and @xmath470 the backward path of @xmath257 in @xmath38 .
if there exists a @xmath56 such that @xmath471 and @xmath472 , we say that _ @xmath41 touches
@xmath257 from above with respect to @xmath45_. if we want to refer to the restricted definition for touching from above which was valid before this paragraph we add the phrase ` in the application of split ' .
note that there exist pairs @xmath155 in which an entry @xmath41 touches another entry @xmath257 from above in the sense we just defined , but there is no step in the application of split 3 to @xmath155 where @xmath41 touches @xmath257 from above in the original sense . [ 1 ] let @xmath473 and @xmath12 be two cells in a shifted tabloid @xmath38 , which is ordered up to @xmath12 .
let @xmath474 denote the forward path of @xmath475 in @xmath38 with respect to the a set @xmath45 and @xmath257 the entry of the endcell of @xmath474 in @xmath38 .
let @xmath41 be an entry in @xmath38 whose backward path ( @xmath8 being the fixed row ) contains @xmath12 .
if @xmath475 does not touch @xmath41 from above with respect to @xmath45 , then @xmath41 is right [ and weakly above ] of @xmath474 or the backward path of @xmath41 enters the column of @xmath257 weakly above of @xmath257 .
( see figure [ el2 ] . )
we show the following : let @xmath476 be below of @xmath474 .
if @xmath475 does not touch @xmath476 from above with respect to @xmath45 then every cell in the backward path @xmath460 of @xmath476 ( @xmath8 being the fixed row ) in a column greater than @xmath477 is below of @xmath474 .
if the statment were false there would exist integers @xmath469,@xmath478 such that @xmath479 and @xmath480 .
since @xmath475 does not touch @xmath476 from above with respect to @xmath45 the cell @xmath481 is not a cell in the main diagonal and thus @xmath482 is a cell in the shifted ferrers diagram .
from @xmath479 it follows that @xmath483 , which is a contradiction to @xmath484 , which follows from @xmath480 .
( later we will often refer to this important argument as the argument in the proof of lemma [ 1]. ) if the statement in the lemma were false , @xmath41 is below of @xmath474 or the backward path of @xmath41 enters the column of @xmath257 below of @xmath257 . by the assertion we have just proved every entry in the backward path @xmath460 of @xmath41 which is in a column greater than @xmath477 and
weakly left of the column of @xmath257 must be below of @xmath474 .
but , since @xmath12 is contained in the backward path of @xmath41 and @xmath12 is not below of @xmath474 , this is a contradiction if @xmath474 does not consist solely of south steps .
note that if we are in the situation of lemma [ 1 ] and @xmath41 enters the column of @xmath257 weakly above of @xmath257 then @xmath41 enters the column of @xmath257 weakly above uppermost cell in @xmath474 and the column of @xmath257 by the proof of the lemma .
[ 5 ] mostly we apply the lemma in the following situation .
let @xmath473 and @xmath12 be two cells in a shifted tabloid @xmath38 which is ordered up to the predecessor of @xmath12 in the total order .
furthermore let @xmath485 denote the forward path of @xmath187 in @xmath38 with respect to a set @xmath45 and let @xmath474 denote the forward path of @xmath475 in @xmath52 with respect to a set @xmath486 , where @xmath52 denotes the shifted tabloid we obtain after performing jeu de taquin in @xmath38 with @xmath187 and with respect to @xmath45 .
let @xmath257 denote the entry in the last cell of @xmath474 in @xmath52 .
observe that @xmath485 coincides with the backward path of @xmath187 in @xmath52 ( @xmath8 being the fixed row ) except for the part of the @xmath8-th row left of @xmath12 .
thus , if @xmath475 does not touch @xmath187 from above with respect to @xmath486 in @xmath52 , then @xmath187 is right [ and weakly above ] of @xmath474 in @xmath52 or the backward path of @xmath187 in @xmath52 enters the column of @xmath257 weakly above of @xmath257 by lemma [ 1 ] . [ 3 ] let @xmath473 , @xmath12 , @xmath38 and @xmath52 be as in remark [ 5 ] .
furthermore let @xmath488 be a set of cells in the last @xmath427 rows of @xmath52 , such that @xmath187 is the smallest entry under the entries in the cells of @xmath488 with respect to the backward paths order in @xmath52 .
let @xmath489 denote the shifted tabloid we obtain after performing jeu de taquin with @xmath475 in @xmath52 and with respect to @xmath486 and let @xmath490 denote the column of @xmath475 in @xmath489 .
let @xmath491 denote the set of cells we obtain from @xmath488 by first replacing every cell @xmath492 with @xmath172 by @xmath493 and then deleting the replacing cells in the @xmath426-st row .
if @xmath475 touches no entry in a cell in @xmath488 from above with respect to @xmath486 in @xmath52 then @xmath475 is smaller than every entry in a cell in @xmath491 with respect to the backward paths order in @xmath489 .
_ by lemma [ 0 ] ( 1 ) , lemma [ 1 ] and remark [ 5 ] it is clear that either the entry in the cell of @xmath491 that came from @xmath187 is greater than @xmath475 with respect to the backward paths order in @xmath489 or the cell that came from @xmath187 was deleted in the course of constructing @xmath491 .
( this is due to the following observation : let @xmath91 be a cell which is right and weakly above of a path @xmath460 then @xmath494 is weakly right and above of @xmath460 or @xmath460 has no cell in the row of @xmath494 . ) now suppose that @xmath41 is an entry in @xmath52 weakly below the @xmath8-th row that it is greater than @xmath187 with respect to the backward paths order in @xmath52 .
furthermore suppose that @xmath475 does not touch @xmath41 from above with respect to @xmath486 in @xmath52 .
thus , by the relative position of @xmath41 and @xmath187 by lemma [ 0 ] , and since the backward path ( @xmath8 being the fixed row ) of @xmath187 in @xmath52 includes @xmath12 , the backward path of @xmath41 in @xmath52 includes @xmath12 as well .
( observe that the backward path of @xmath41 coincides with the backward path of @xmath187 after they intersect . ) by lemma [ 1 ] this implies that @xmath41 is either right [ and weakly above ] of @xmath474 in @xmath52 or the backward path of @xmath41 enters the column of the end @xmath257 of @xmath474 in @xmath52 weakly above of @xmath257 . from this
the assertion follows by lemma [ 0 ] ( 1 ) . in section [ alg ] after split 1 we claim that for an entry @xmath41 whose forward path terminates in a cell @xmath184 in the course of split 1 , we have @xmath185 and @xmath186 after the application of split 1 .
this is also true for split 1 and proved in the subsequent lemma .
[ 6 ] let @xmath38 be a shifted tabloid which is ordered up to @xmath181 and @xmath127 a partial shifted hook tabloid where the @xmath8-th row is empty .
let @xmath495 denote the pair we obtain in the course of applying split 1 to @xmath155 in the @xmath8-th row after performing jeu de taquin with @xmath187 .
assume that the forward path of @xmath187 in @xmath496 contains cells on the main diagonal and let @xmath184 be the first of these in the run of the forward path .
let @xmath497 be such that the forward path @xmath460 of @xmath498 in @xmath499 ends weakly right of the @xmath56-th column .
furthermore suppose that the forward path of @xmath500 in @xmath501 does not contain a cell on the main diagonal for @xmath502 .
then @xmath498 touches @xmath187 from above in @xmath499 , to be more accurate @xmath471 and @xmath184 is the cell of @xmath187 in @xmath499 .
_ we show that the backward path of @xmath187 in @xmath499 does not contain a cell @xmath503 on the main diagonal with @xmath504 .
furthermore this backward path contains @xmath505 and @xmath184 .
the assertion is then a consequence of lemma [ 1 ] : let @xmath38 in lemma [ 1 ] equal @xmath499 , @xmath11 equal @xmath506 and let @xmath507 , where @xmath508 is the first cell in the @xmath56-th column in the forward path @xmath460 of @xmath498 in @xmath499 . by lemma [ 1 ]
@xmath187 is either right of @xmath460 or the backward path of @xmath187 enters column @xmath56 weakly above of @xmath508 or @xmath498 touches @xmath187 from above in @xmath499 . since @xmath509 and @xmath187 is either strictly left of the @xmath56-th column in @xmath499 or in @xmath184 the first two options are impossible and thus @xmath498 touches @xmath187 from above .
the only possible place for touching is @xmath510 , @xmath511 , @xmath184 for the backward path of @xmath187 in @xmath499 does not contain a cell on the main diagonal strictly between the @xmath8-th and the @xmath56-th column .
observe that the backward path ( @xmath8 being the fixed row ) of @xmath187 in @xmath512 contains no cell @xmath503 with @xmath504 and it contains @xmath12 and therefore @xmath505 , for it coincides with the forward path of @xmath187 in @xmath496 except for the part strictly left of @xmath12 . if @xmath513 there is nothing to prove .
otherwise the forward path of @xmath475 in @xmath512 contains no cell on the main diagonal and thus the backward path of @xmath187 in @xmath514 does not contain a cell on the main diagonal strictly between column @xmath8 and @xmath56 and it contains @xmath473 , for this backward path coincides with the backward path of @xmath187 in @xmath512 until it meets the backward path of @xmath475 in @xmath514 and after that it coincides with the backward path of @xmath475 in @xmath514 .
the claim at the beginning of the proof now follows by induction with respect to @xmath515 .
now we are ready to show the assertion before the formulation of lemma [ 6 ] ( we use the notation of lemma [ 6 ] . ) : assume that the forward path of @xmath516 in @xmath517 contains a cell @xmath91 on the main diagonal and let @xmath518 be such that the forward path of @xmath498 in @xmath499 ends weakly right of the column of @xmath91 .
let @xmath45 be the appropriate set of cells in split 1 such that @xmath519 , then it suffices to show that the forward path of @xmath498 in @xmath499 with respect to @xmath45 ends strictly left of the column of @xmath91 .
we show the assertion by induction with respect to @xmath520 . to this end
let @xmath11 , @xmath521 , be minimal such that the forward path of @xmath187 in @xmath496 with respect to the appropriate set in split 1 contains a cell @xmath522 on the main diagonal .
by lemma [ 6 ] the forward path of @xmath498 in @xmath499 with respect to @xmath45 ends in a column strictly left of the column of @xmath522 .
if @xmath523 the assertion follows immediately . otherwise the column of @xmath522 is smaller than the column of @xmath91 by induction since @xmath524 and the assertion follows too .
actually the corollary also implies the following inversion of the statement above : if @xmath186 after the application of split 1 , then @xmath525 is an entry with which we have performed jeu de taquin in the course of applying split 1 and @xmath184 is the end of its forward path . in fact this is a consequence of the following observation which can be shown by lemma [ 1 ] and lemma [ 6 ] : assume that the backward path of a horizontal candidate in @xmath495 contains @xmath184 .
then there exists an @xmath478 , @xmath526 , such that the forward path of @xmath500 in @xmath501 with respect to the appropriate @xmath45 in split 1 contains @xmath184 and @xmath500 is the smallest horizontal candidate in @xmath495 , whose backward path contains @xmath184 .
we define property after_split 1 for a pair @xmath155 and will observe that the application of split 1 yields a pair with that property . a pair @xmath155 of a shifted tabloid @xmath38 and a partial shifted hook tabloid @xmath127 has property after_split 1 if 1 .
the last @xmath427 rows of @xmath127 form a shifted hook tabloid , 2 .
there exists no vertical candidate ( with respect to the @xmath8-th row ) in @xmath155 , 3 .
the subtabloid of @xmath38 consisting of the last @xmath427 rows is standard , except for @xmath527 there might hold @xmath528 , where @xmath196 is the set of entries @xmath525 with @xmath186 .
4 . for @xmath527
such that @xmath525 either proves to be exceptional or touched from above in the application of split and every diagonal cell @xmath529 in the backward path of @xmath525 , there exists a horizontal candidate @xmath41 in @xmath38 whose backward path includes @xmath14 and @xmath530 , and 5 .
@xmath527 is unstable and neither exceptional nor touched from above in the application of split , then @xmath525 touches another entry from above in the application of split and is in its place of change . 1 .
let @xmath38 be a shifted tabloid which is ordered up to @xmath181 and @xmath127 a partial shifted hook tabloid such that the @xmath8-th row is empty .
let @xmath197 denote the pair we obtain after the application of split 1
. then @xmath197 has property after_split 1 and the application of merge 1 to @xmath197 yields @xmath155 .
2 . let @xmath197 be a pair with property after_split 1 and @xmath155 denote the pair we obtain after the application of merge 1 .
then @xmath38 is ordered up to @xmath181 and @xmath127 is a partial shifted hook tabloid such that the @xmath8-th row is empty and the application of split 1 to @xmath155 yields @xmath197 . _ re 1 .
_ let @xmath495 denote the pair we obtain after the performance of jeu de taquin with @xmath187 and the corresponding shift in the partial shifted hook tabloid in split 1 and set @xmath531 . by construction
the current cell of @xmath532 in @xmath496 is in @xmath533 .
furthermore by induction we assume that @xmath532 is the smallest horizontal candidate in a cell in @xmath534 with respect to the backward paths order in @xmath496 .
let @xmath45 be the set such that @xmath512 is obtained from @xmath496 in split 1 by performing jeu de taquin with @xmath187 and with respect to @xmath45 . by construction
no horizontal candidate in a cell in @xmath534 is touched from above by @xmath187 with respect to @xmath45 in @xmath496 : let @xmath56 be minimal such that the backward path in @xmath496 of a horizontal candidate with respect to @xmath535 contains @xmath184 .
then the forward path of @xmath187 with respect to @xmath45 ends strictly left of the column of @xmath56 by lemma [ 6 ] .
( see also the last paragraph in the previous subsection . ) thus , by corollary [ 3 ] , @xmath187 is smaller or equal than every horizontal candidate in a cell in @xmath536 with respect to the backward paths order in @xmath512 .
therefore if we apply the @xmath537-st step of merge 1 to the pair @xmath495 we reobtain @xmath538 .
observe that the application of pre - merge 1 and merge 1 to @xmath197 yields the same result as the application of merge 1 to @xmath197 for the smallest horizontal candidate whose backward path contains @xmath14 is the smallest horizontal candidate with respect to the backward paths order weakly right of the @xmath11-th column .
let @xmath539 denote the pair we obtain after the application of pre - merge 1 to @xmath197 .
then @xmath539 has the properties 1 3 from after_split and for every cell @xmath184 on the main diagonal such that the backward path of a horizontal candidate contains @xmath184 , @xmath540 is itself a horizontal candidate .
we denote the property of @xmath539 by after_split 1 .
( if we apply split 1 to a pair @xmath155 , @xmath38 ordered up to @xmath181 and the @xmath8-th row of @xmath127 is empty , then the output pair has property after_split 1 . )
let @xmath541 denote the pair we obtain in the course of applying merge 1 to @xmath539 after performing reverse jeu de taquin with respect to @xmath542 and the corresponding reshift in the partial shifted hook tabloid , and set @xmath543 . by induction
we assume that the backward path of every horizontal candidate in a cell in @xmath544 in @xmath545 contains @xmath12 and it contains no cell on the main diagonal except for possibly @xmath546 and @xmath12 .
let @xmath41 be the smallest horizontal candidate in a cell in @xmath547 with respect to the backward paths order in @xmath545 .
furthermore let @xmath460 be its backward paths in @xmath545 with respect to @xmath12 . if @xmath490 is the column of @xmath41 in @xmath545 , the subtabloid of @xmath545 consisting of the first @xmath490 columns is ordered up to @xmath12 by property after_split 1 and the minimality of @xmath41 .
thus @xmath548 and performing jeu de taquin with @xmath41 in @xmath549 and with respect to the cells on the main diagonal together with the corresponding shift in @xmath550 results in the pair @xmath541 . by the minimality of @xmath41 , lemma [ 0 ] ( 1 ) and by the construction of the reshift in @xmath551 the entries in the cells in @xmath552 are right [ and weakly above ] of @xmath460 or their backward paths in @xmath549
enters column @xmath490 weakly above of @xmath553 .
thus , by the argument from lemma [ 1 ] , the backward paths of the entries in these cells in @xmath549 contain @xmath554 .
likewise it is easy to see that no backward path of these entries contains a cell on the main diagonal except for possibly its origin and terminus .
now the assertion follows since pre - merge 1 and post - split 1 are inverse to each other ( in order to show that use lemma [ 6 ] ) and since the application of split 1 and post - split 1 is equivalent to the application of split 1 . 1 .
let @xmath155 be a pair with the property after_split 1 and let @xmath197 denote the pair we obtain after the application of split 2 and split 3 to @xmath155 .
we have to show that if we apply merge 3 and merge 2 to @xmath197 we reobtain @xmath155 .
2 . let @xmath155 be such that @xmath38 is ordered up to @xmath425 and the subtabloid of @xmath127 consisting of the last @xmath427 rows is a shifted hook tabloid .
let @xmath197 denote the pair we obtain after the application of merge 3 and merge 2 .
we have to show that @xmath197 has property after_split 1 and that the application of split 2 and split 3 to @xmath197 yields @xmath155 .
[ 2 ] let @xmath555 and @xmath556 , @xmath557 , be two cells in a shifted tabloid @xmath38 such that the subtabloid of @xmath38 consisting of the last @xmath558 rows and without the entries in @xmath555 and @xmath556 has increasing rows and columns .
let @xmath559 denote the forward path of @xmath560 in @xmath52 with respect to a set , where @xmath52 denotes the shifted tabloid we obtain after performing jeu de taquin in @xmath38 with @xmath561 and with respect to @xmath45 . furthermore let @xmath41 denote the entry in the last cell of @xmath559 in @xmath52 .
then @xmath561 is [ weakly left and ] below of @xmath562 in @xmath52 or the backward path of @xmath561 with respect to @xmath555 in @xmath52 enters the row of @xmath41 weakly left of @xmath41 . _ proof . _
similar to the proof of lemma [ 1 ] .
observe that if the backward path of @xmath561 in @xmath52 enters the row of @xmath41 weakly left of @xmath41 then it enters the row of @xmath41 weakly left of the leftmost cell of @xmath559 in the row of @xmath41 .
[ 4 ] let @xmath556 , @xmath555 , @xmath38 , @xmath52 be as in lemma [ 2 ] . furthermore let @xmath488 be a set of cells weakly below the @xmath8-th row , which includes the cell of @xmath561 in @xmath52 , such that @xmath561 is the greatest entry in a cell of @xmath488 in @xmath52 with respect to the backward paths order .
let @xmath489 denote the shifted tabloid we obtain after performing jeu de taquin with @xmath560 in @xmath52 and with respect to a set and let @xmath190 denote the row of @xmath560 in @xmath489 .
let @xmath491 denote the set of cells we obtain from @xmath488 by replacing every cell @xmath563 in @xmath488 with @xmath564 by @xmath493 . then @xmath560 is greater than every entry in a cell of @xmath491 in @xmath489 with respect to the backward paths order .
we define property after_split 2 for a pair @xmath155 and will observe that a pair to which we have applied split 1 and split 2 has that property . a pair @xmath155 of a shifted tabloid @xmath38 and a partial shifted hook tabloid @xmath127 has property after_split 2 if 1 .
the last @xmath427 rows of @xmath127 form a shifted hook tabloid , 2 .
there exists an @xmath441 such that for every @xmath565 there is a vertical candidate in row @xmath469 and no other vertical candidates exist . for @xmath566 the vertical candidate in row
@xmath469 is greater than the vertical candidate in row @xmath567 with respect to the backward paths order ( @xmath8 being the fixed row ) .
if we perform jeu de taquin with the vertical candidates in row @xmath568 in that order , then every candidate is either stable or its forward path starts with a step into the next row .
moreover : @xmath454 is a stable vertical candidate or @xmath569 is a stable horizontal candidate or @xmath453 is not a horizontal candidate or @xmath455 is not a horizontal candidate , 3 .
the subtabolid of @xmath38 consisting of the last @xmath427 rows is standard , except for @xmath570 there might hold @xmath571 , where @xmath195 and @xmath196 are the two sets we define before the application of split 3 to @xmath155 ( see section [ alg ] before the description of split 3 ) , 4 .
@xmath41 is a horizontal candidate which is touched from above in the application of split , then for every cell @xmath572 with @xmath573 in the backward path of @xmath41 , there exists a horizontal candidate @xmath257 with @xmath574 and the backward path of @xmath257 contains @xmath14 , and 5 .
@xmath41 is an unstable horizontal candidate which is not touched from above in the application of split then @xmath41 touches from above in the application of split and is in its place of change . a remark on property ( 2 ) of after_split 2 : note that the vertical candidate in row @xmath469 is greater than the vertical candidate in row @xmath567 , @xmath566 in a pair to which we have applied split 2 by lemma [ 2 ] .
let @xmath155 be a pair with property after_split 1 and let @xmath197 denote the pair we obtain after the application of split 2
. then @xmath197 has property after_split 2 and the application of merge 2 to @xmath197 results in @xmath155 .
2 . let @xmath197 be a pair with property after_split 2 and @xmath155 denote the pair we obtain after the application of merge 2 .
then @xmath155 has property after_split 1 and the application of split 2 to @xmath155 results in @xmath197 .
_ left to the reader .
observe that we are in case 2 of split 2 , respectively case 2 of merge 2 , if and only if either the smallest horizontal candidate with respect to the backward paths order in @xmath197 in a row strictly below the lowest vertical candidate is smaller than the smallest vertical candidate or there exists no vertical candidate .
the rest follows from the fact that jeu de taquin and reverse jeu de taquin are inverse to each other and from the fact that ( with one exception ) the endcells of the forward paths are not distorted by a shift in the accompanying shifted hook tabloid .
the exception is the exceptional vertical candidate in the @xmath23-th row in case 1 of split 2 , whose endcell is stored in the @xmath8-th row . there
we need corollary [ 4 ] . *
touching from below . *
we define an analog to ` @xmath41 touches @xmath257 from above ' for the algorithm merge .
suppose we are at the beginning of a step of merge 3 and the if - condition ( if @xmath430 : [ @xmath286 ] ) is true .
in the following commands in the description of merge 3 we distinguish between two cases .
if we are in the first case we say that @xmath257 touches @xmath41 from below and for @xmath575 , @xmath444 touches @xmath443 from below .
furthermore for @xmath576 , let @xmath14 be the place of change for @xmath444 .
in the second case we say that for @xmath577 , @xmath444 touches @xmath443 from below and @xmath14 is the place of change for @xmath444 .
we define the exceptional entries with respect to merge 2 .
suppose we have applied merge 2 , case 1 .
define @xmath448 to be exceptional for @xmath449 .
now suppose we have applied merge 2 , case 2 . for @xmath449 and @xmath578
we define @xmath448 to be exceptional and @xmath579 is said to be touched from below by the smallest candidate strictly below the @xmath580-st row if @xmath451 .
a _ step of split 3 _ includes the choice of a protagonist @xmath581 and the performance of either case 1 , case 2 or case 3 .
a _ step of merge 3 _ is defined analogously .
the maximal vertical candidate @xmath41 at the beginning of a step of merge 3 is said to be the first protagonist of the step and the candidate @xmath41 after the performance of the if - condition ( if @xmath582 $ ] ) is said to be the second protagonist .
clearly the second protagonist can be equal to the first protagonist in a step of merge 3 . * irreducible pair with respect to split .
* let @xmath155 be a pair with property after_split 2 .
suppose that in the course of applying split 3 to @xmath155 every protagonist which touches an entry from above was either in @xmath195 at the beginning of split 3 or is itself touched by an entry from above in the application of split .
observe that this is equivalent to the fact that we are never in case 3 in split 3 and that @xmath583 at the beginning of every step of split 3 .
then @xmath155 is said to be irreducible with respect to split . in figure [ irred ]
the application of split 3 to an irreducible pair is illustrated . * irreducible pair with respect to merge .
* let @xmath38 be a shifted tabloid which is ordered up to @xmath425 and @xmath127 a partial shifted hook tabloid , such that the subtabloid consisting of the last @xmath427 rows is a shifted hook tabloid .
if every protagonist @xmath41 ( except for possibly the last ) in merge 3 either touches an entry from below or terminates in row @xmath584 in a certain step of merge 3 , then @xmath155 is said to be irreducible with respect to merge .
note that this is equivalent to the following : suppose the second protagonist in a step of merge 3 is a horizontal candidate .
then this is either the last step in the application of merge 3 to @xmath155 or this second protagonist is equal to @xmath257 in the if - condition at the beginning of the next step of merge 3
. * claim 3 . *
let @xmath155 be irreducible with respect to split and let @xmath197 denote the pair we obtain after the application of split 3 .
then @xmath52 is ordered up to @xmath425 , the last @xmath427 rows of @xmath137 form a shifted hook tabloid and the application of merge 3 to @xmath197 results in @xmath155 .
_ let @xmath585 denote the set @xmath195 at the beginning of the application of split 3 to @xmath155 .
suppose we are in the course of applying split 3 : we choose in a certain step entry @xmath119 to be the protagonist and obtain the pair @xmath586 , in the next step we choose @xmath120 to be the protagonist and obtain @xmath587 .
let @xmath588 be such that there exists a vertical candidate in row @xmath588 of @xmath586 and in @xmath585 , which is not the protagonist in and before the step of @xmath119 and if there exists a vertical candidate in the @xmath589-st row of @xmath586 and in @xmath585 , this entry was the protagonist in or before the step of @xmath119 .
if @xmath588 does not exist , set @xmath590 .
let @xmath591 be the corresponding quantity for the step of @xmath120 and observe that @xmath592 iff @xmath120 is the vertical candidate of @xmath586 in the @xmath588-th row
. otherwise @xmath593 .
if we are in case 2 in the step of @xmath594 let @xmath595 and @xmath596 denote @xmath23 and @xmath597 in the description of case 2 of split 3 , @xmath440 . 1 .
suppose we are in case 1 in the step of @xmath119 .
then @xmath119 is greater or equal than every vertical candidate strictly below the @xmath588-th row of @xmath586 with respect to the backward paths order and if @xmath119 is horizontal candidate then @xmath120 is the greatest vertical candidate strictly below the @xmath591-th row of @xmath586 .
suppose we are in case 2 in the step of @xmath119 .
then @xmath598 is greater or equal than every vertical candidate strictly below the @xmath588-th row of @xmath586 and if @xmath120 is a horizontal candidate in @xmath586 then @xmath599 is the greatest vertical candidate strictly below the @xmath588-th row of @xmath586 . 1 . @xmath586 or 2 .
the pair we obtain before the performance of @xmath418 in the step of @xmath601 ( in which we are in case 2 of split 3 and @xmath601 is a horizontal candidate at the beginning of the step of @xmath601 ) or 3 . a pair which results in @xmath586 after the application of a second step of merge 3 until the marker @xmath602 ( see the description of merge 3 ) . 1 .
suppose we are in case 1 in the step of @xmath120 .
then @xmath120 is greater or equal than every vertical candidate in @xmath587 strictly below the @xmath591-th row and if @xmath120 is a horizontal candidate then @xmath603 for a @xmath56 and @xmath604 is the greatest vertical candidate of @xmath587 strictly below the @xmath591-th row .
suppose we are in case 2 in the step of @xmath120 . in this case
@xmath605 is greater or equal than every vertical candidate of @xmath587 strictly below the @xmath591-th row of @xmath587 and if @xmath605 is a horizontal candidate then @xmath606 is the greatest vertical candidate of @xmath587 strictly below the @xmath591-th row .
1 . @xmath607 , @xmath120 is in the @xmath588-th row of @xmath46 and @xmath119 was either the vertical candidate of @xmath585 in the @xmath589-st row or the unique horizontal candidate of @xmath585 in @xmath155 or 2 .
@xmath608 for a @xmath56 , @xmath120 is unstable in @xmath46 and the backward path of @xmath119 in @xmath46 contains @xmath609 . _
_ clearly @xmath120 is the greatest vertical candidate in @xmath586 weakly below the @xmath588-th row for @xmath119 is greater or equal than every vertical candidate in @xmath586 strictly below the @xmath588-th row by induction and @xmath120 is greater than @xmath119 by property after_split 2 .
if @xmath120 changes row in the step of @xmath120 , the first cell in row @xmath610 in its forward path is weakly right of the first cell in row @xmath610 in @xmath586 in the backward path of @xmath601 by property after_split 2 .
thus @xmath120 is also the greatest vertical candidate in @xmath587 strictly below the @xmath591-th row by corollary [ 4 ] , if we are not in case 2 in the step of @xmath120 and in this case it is obvious that @xmath120 is the first and the second protagonist in the application of a step of merge 3 to @xmath587 with @xmath600 ( check that the if - condition at the beginning of the step of merge 3 is either not fulfilled or its application leaves @xmath587 unchanged in this step ) .
otherwise @xmath605 is the greatest vertical candidate in @xmath587 strictly below the @xmath591-th row for firstly @xmath611 is the greatest vertical candidate strictly below the @xmath591-th row and strictly above the @xmath612-th row of @xmath587 by corollary [ 4 ] , secondly there exists no vertical candidate strictly below the @xmath613-th row and thirdly @xmath605 is greater than @xmath614 for @xmath615 .
thus @xmath605 is the first protagonist in the application of a step of merge 3 to @xmath587 .
consequently @xmath120 is the second protagonist in the application of a step of merge 3 to @xmath587 if we are in case 2 in the step of @xmath120 .
it remains to show that the loop in the application of a step of merge 3 with @xmath600 to @xmath587 terminates with ` 1 .
@xmath41 is in row @xmath584 ' : observe that the forward path of @xmath120 in @xmath46 does not contain a cell @xmath184 such that there exists a vertical candidate strictly below the @xmath56-th row for such a vertical candidate would be greater than @xmath120 in @xmath46 by the arguments from lemma [ 1 ] , which is a contradiction to the maximality of @xmath120 weakly below the @xmath588-th row in @xmath586 .
thus the loop does not stop with ` 2 . ' .
furthermore the forward path of @xmath120 in @xmath586 with respect to the appropriate set does not contain a cell @xmath184 ( except for possibly the endcell of the forward path of @xmath616 with respect of the emptyset ) with @xmath617 such that the backward path in @xmath46 of a horizontal candidate contains @xmath609 , for otherwise @xmath120 touches the minimal horizontal candidate whose backward paths contains @xmath609 from above in the application of split by lemma [ 6 ] , which implies @xmath618 and this is a contradiction to the choice of @xmath612 .
therefore the loop does not stop with ` 3 . ' . finally , if @xmath56 is the row of @xmath120 in @xmath105 , then the subtabloid of @xmath46 consisting of the cells strictly below of the @xmath588-th row and weakly above of the @xmath56-th row is standard and consequently the loop does not stop with ` 4 . ' .
therefore the application of one step of merge 3 to @xmath587 with @xmath600 yields @xmath586 in this case .
let @xmath119 be a vertical candidate , then @xmath119 is the greatest vertical candidate strictly below the @xmath588-th row of @xmath586 by the induction hypothesis .
if we are not in case 2 in the step of @xmath120 then @xmath120 is the greatest vertical candidate strictly below the @xmath591-th row of @xmath587 by corollary [ 4 ] ( since @xmath120 is unstable in @xmath586 ) and thus the first and the second protagonist in the application of a step of merge 3 to @xmath587 with @xmath600 .
if we are in case 2 in the step of @xmath120 then again @xmath605 is the greatest vertical candidate strictly below the @xmath591-th row of @xmath587 and with this the first protagonist in the application of a step of merge 3 to @xmath586 .
thus @xmath120 is the second protagonist in the application of a step of merge 3 to @xmath587 with @xmath600 in this case .
it remains to show that the loop at the end of the application of a step of merge 3 to @xmath587 stops with ` 2 . ' and at the cell of @xmath120 in @xmath46 .
this is left to the reader , for it is similar to the previous case .
now let @xmath119 be a horizontal candidate , then @xmath119 is greater than every vertical candidate strictly below the @xmath588-th row of @xmath586 and @xmath120 is the greatest vertical candidate strictly below the @xmath588-th row of @xmath586 by the induction hypothesis
. thus either @xmath120 ( if we are in case 1 in the step of @xmath616 ) or @xmath605 ( if we are in case 2 in the step of @xmath616 ) is the greatest vertical candidate strictly below the @xmath591-th row of @xmath587 and therefore the application of a step of merge 3 to @xmath587 with @xmath600 yields @xmath586 .
note that the loop in merge 3 terminates with ` 3 . ' and at the cell of @xmath120 in @xmath46 , @xmath601 being the horizontal candidate in 3.. _ case 2 .
_ clearly @xmath598 and @xmath120 is greater or equal than every vertical candidate strictly below the @xmath619-th row of @xmath586 with respect to the backward paths order by the induction hypothesis .
thus either @xmath120 ( if we are in case 1 in the step of @xmath616 ) or @xmath605 ( if we are in case 2 in the step of @xmath616 ) is greater or equal than every vertical candidate strictly below the @xmath591-th row in @xmath587 .
suppose @xmath120 is a vertical candidate in @xmath586 .
then either @xmath120 or @xmath605 is the first protagonist in the application of a step of merge 3 with @xmath600 to @xmath587 and therefore @xmath120 is the second protagonist .
thus the application of a step of merge 3 yields @xmath586 .
observe that the loop in merge 3 terminates with ` 4 . ' and at the cell of @xmath120 in @xmath46 .
suppose @xmath120 is a horizontal candidate in @xmath586 .
first we suppose that we are in case 1 in the step of @xmath616 .
then @xmath620 is the greatest vertical candidate strictly below the @xmath591-th row of @xmath587 and consequently the first protagonist in the application of a step of merge 3 to @xmath586 .
thus @xmath621 in the if - condition at the beginning of this step of merge 3 and @xmath601 is the second ( horizontal ) protagonist .
we obtain the pair which we had immediately before the performance of @xmath418 in the step of @xmath601 .
now suppose we are in case 2 in the step of @xmath616 .
clearly @xmath606 is the greatest vertical candidate strictly below the @xmath591-th row of @xmath587 and consequently the first protagonist in the application of a step of merge 3 to @xmath587 with @xmath600 .
the second protagonist is the horizontal candidate @xmath120 and since @xmath620 is the first protagonist in the next step of merge 3 we have @xmath621 in the if - condition at the beginning of this next step .
thus the application of another step of merge 3 until the marker @xmath602 yields @xmath586 .
* claim 4 . *
let @xmath155 be irreducible with respect to merge and @xmath197 denote the pair we obtain after the application of merge 3 to @xmath155
. then @xmath197 has property after_split 2 and the application of split 3 to @xmath197 yields @xmath155 .
_ in the course of applying merge 3 to @xmath155 we construct two sets @xmath195 and @xmath196 .
we start with @xmath622 . if the if - condition ( if @xmath430 : [ @xmath286 ] ) at the beginning of a step of merge 3 is true and we have just applied the appropriate commands , we set @xmath623 and @xmath624 if we are in the first case in these commands , and we set @xmath625 and @xmath626 if we are in the second case in these commands . after the if - condition and its commands we let @xmath627 , whether or not the if - condition was fulfilled .
suppose that after a certain step in the application of merge 3 to @xmath155 we obtain the pair @xmath587 , the sets @xmath195 and @xmath196 and a row @xmath466 .
first we show that the second protagonist @xmath41 of this step is also the protagonist in the application of one step of split 3 to @xmath587 with the sets @xmath195 and @xmath196 .
this is equivalent to the fact that @xmath41 is the candidate in @xmath195 whose row is maximal .
if the second protagonist is a horizontal candidate in a certain step of merge 3 , then it is the horizontal candidate @xmath257 in the if - condition at the beginning of the next step ( if we are not in the final step ) of merge 3 and thus leaves @xmath195 at that point , for @xmath155 is irreducible with respect to merge 3 .
consequently @xmath195 includes at most one horizontal candidate and if this is the case then all other candidates in @xmath195 are strictly above the row of the unique horizontal candidate .
( in order to see the second assertion in the previous sentence note that whenever @xmath628 at the beginning of a step of merge 3 then there exists no vertical candidate strictly below the @xmath23-th row by the maximality of @xmath41 . )
therefore it suffices to show the assertion for second protagonists @xmath41 that are vertical candidates . in a step of merge 3
the second protagonist either terminates in row @xmath584 or in a cell @xmath184 on the main diagonal , if the protagonist is a vertical candidate . in the latter case by the maximality of the second protagonist (
observe that after the if - condition in a step of merge 3 the second protagonist is greater or equal than every vertical candidate strictly below the @xmath466-th row ; this is because the backward path of the first protagonist always contains the cell of the second protagonist at the beginning of a step of merge 3 ) every backward path of a vertical candidate strictly below the @xmath56-th row in the tabloid we obtain after this step contains @xmath609 by the argument in the proof of lemma [ 1 ] and therefore these vertical candidates are greater than @xmath41 with respect to the backward paths order .
however , the vertical second protagonist is still greater than every vertical candidate strictly below the @xmath466-th row and strictly above the @xmath56-th row , since this part of the shifted hook tabloid remains unchanged in this step of merge 3 .
consequently in the course of merge 3 the second vertical protagonists ( and with this the elements of @xmath195 ) are arranged in increasing rows ( as long as the first protagonist is equal to the second protagonist ) until from time to time former second vertical protagonists leave @xmath195 after the performance of the commands in the if - condition .
we conclude that the second protagonist in a step of merge 3 either starts ( by its maximality and the arguments in this paragraph ) and ends ( by the stopping criteria ` 4 .
@xmath629 and @xmath630 is an unstable vertical candidate ' in merge 3 and the arguments in this paragraph ) below the lowest vertical candidate in @xmath195 or is equal to the lowest vertical candidate in @xmath195 , whose backward path in this step ends below the second lowest vertical candidate in @xmath195 ( if it exists ) by the fourth stopping criteria .
consequently the row of the vertical second protagonist in a step of merge 3 is always below the other vertical candidates in @xmath195 and by the observation about the horizontal candidates in @xmath195 in the previous paragraph this proves our claim in the second paragraph of the proof .
let @xmath586 denote the pair before the step of merge 3 , in which @xmath41 is the second protagonist and let @xmath631 and @xmath632 denote the accompanying sets .
we have to show that if we apply one or two steps of split 3 to @xmath587 with the sets @xmath195 and @xmath196 , we reobtain @xmath586 together with @xmath631 and @xmath632 or a pair of a shifted tabloid and a shifted hook tabloid together with the two sets which we have obtained in the application of merge 3 to @xmath155 before the pair @xmath586 .
proving this claim is now a matter of considering every circumstance in the step of @xmath41 in merge 3 .
( case distinction : first suppose that we are in the first case in the if - condition at the beginning of the step of merge 3 ( this leaves no choice open for the rest of the step ) and second combine the second case in the if - condition and the case that the if - condition is false with each of the four stopping criterions of the loop in merge 3 . )
thus we have shown that for irreducible pairs split 3 and merge 3 are inverse to each other .
now we show the assertion for reducible pairs by induction with respect to the number of protagonists not contained in @xmath195 at the beginning of the application of split 3 which touch from above but are not touched from above in the application of split , respectively with respect to the number of second protagonists unequal to the last second protagonist which neither touch an entry from below nor terminate in row @xmath584 in the application of merge 3 .
observe that the pair in example [ best ] in section [ examples ] is reducible . in our proof of claim 5
we sometimes refer to that example . moreover in figure [ red ] the application of split 3 to a reducible pair is illustrated . for the proof of claim 5
we need the following observation : suppose we are in case 3 of split 3 and let @xmath586 denote the current pair before the application of case 3 .
then the protagonist @xmath41 in the description of case 3 of split 3 is a vertical candidate . in order to show that suppose that @xmath41 was the first horizontal candidate in the course of split 3 which contradicts this assertion .
then we are either in the first step of split 3 or @xmath41 joined @xmath195 in the previous step of split 3 in which we were in case 2 and @xmath41 equals @xmath633 in the description of case 2 .
we only consider the latter case here , for the first case is similar . by the relative position of @xmath41 and @xmath257 in the current step of split 3 and the argument from lemma [ 1 ] the backward path of @xmath257 in @xmath586 contains @xmath634 . by lemma [ 6 ] @xmath41 touches the smallest horizontal candidate @xmath476 in @xmath586 whose backward path in @xmath46 contains @xmath634 from above in the application of split 3 and this is a contradiction , since @xmath476 is weakly above of the row of @xmath257 by its minimality and by the fact that @xmath257 is on the main diagonal
. * claim 5 . *
let @xmath155 be a pair with property after_split 2 and @xmath197 the pair we obtain after the application of split 3 .
then @xmath52 is ordered up to @xmath425 , the subtabloid of @xmath137 consisting of the last @xmath427 rows is a shifted hook tabloid and the application of merge 3 to @xmath197 yields @xmath155 .
_ let @xmath585 and @xmath635 denote @xmath195 and @xmath196 at the beginning of the application of split 3 to @xmath155 .
we assume that every entry in @xmath635 is either touched from above or touches an entry from above in the application of split 3 . by claim 3
we furthermore assume that @xmath155 is reducible with respect to split .
let @xmath41 be the candidate in @xmath635 with maximal row @xmath56 , which is not touched from above in the application of split ( @xmath319 in example [ best ] ) .
furthermore , if it exists , let @xmath636 be the entry with minimal row strictly below the @xmath56-th row such that @xmath257 is touched from above in a certain step in the application of split 3 by a protagonist which is at the beginning of the step in a row strictly above of the @xmath56-th row ( @xmath329 in the example ) .
let @xmath637 denote the entries in @xmath635 which are strictly above the @xmath56-th row and weakly below the row of @xmath257 and @xmath638 denote the entries in @xmath635 which are strictly below the @xmath56-th row and strictly above the row of @xmath257 in @xmath38 .
let split 3.1 denote the application of split 3 with @xmath639 and @xmath640 to @xmath155 and let split 3.2 .
denote the application of split 3 with @xmath641 and @xmath642 to @xmath155 .
observe that the application of split 3 with @xmath639 and @xmath643 to @xmath155 is equivalent to the simultaneous application of split 3.1 and split 3.2 , where after every step of split 3.@xmath8 , @xmath644 , we can decide to make either the next step of split 3.1 or the next step of split 3.2 , but whenever a protagonist in split 3.1 moves weakly below the @xmath56-th row the application of split 3.2 has already terminated .
this is because @xmath56 is maximal and there exists a step of split 3 in which we are either in case 3 with @xmath41 being the @xmath257 in the description of case 3 of split 3 or @xmath645 at the beginning of the step and @xmath41 is the candidate with minimal row in @xmath196 at that point .
observe that after the termination of split 3.2 and before a protagonist of split 3.1 moves weakly below the @xmath56-th row , @xmath41 is the greatest vertical candidate under the vertical candidates that arose in the application of split 3.2 by the proof of claim 3 , for @xmath41 is the final protagonist in split 3.2 for @xmath56 is maximal .
the first protagonist of split 3.1 that moves weakly below of the @xmath56-th row is a vertical candidate ( since whenever we are in case 3 in split 3 the protagonist is a vertical candidate ; this is proved at the beginning of this subsection ) and greater than @xmath41 with respect to the backward paths order by corollary [ 6 ] .
moreover every subsequent protagonist of split 3.1 , which moves weakly below the @xmath56-th row , is greater than the previous .
thus the vertical candidates from split 3.1 dominate the vertical candidates from split 3.2 once they come weakly below @xmath56-th row .
let @xmath646 denote the vertical pointers that came from split 3.1 and let @xmath647 denote the vertical pointers that came from split 3.2 after the application of split . by the induction hypothesis we reobtain @xmath155 if we apply merge 3 to the pointers in @xmath646 ( this procedure is denoted by merge 3.1 ) and , seperately , to the pointers in @xmath647 ( this procedure is denoted by merge 3.2 ) such that when we start applying it to the pointers originated in @xmath647 the lowest vertical pointer that came from @xmath646 is strictly above the @xmath56-th row .
if we apply merge 3
not seperately to the vertical pointers in @xmath648 and @xmath649 , but to all vertical pointers in @xmath197 at once , then the first steps of this application are equal to the first steps of merge 3.1 until the vertical pointers of merge 3.1 have moved strictly above the @xmath56-th row , for the candidates originated in @xmath648 dominate the candidates originated in @xmath649 with respect to the backward paths order until they are strictly above the @xmath56-th row as we saw in the previous paragraph
. * claim 6 . *
let @xmath155 be such that @xmath38 is ordered up to @xmath425 , the subtabloid of @xmath127 consisting of the last @xmath427 rows is a shifted hook tabloid and let @xmath197 denote the pair we obtain after the application of merge 3
. then @xmath197 has property after_split 2 and the application of split 3 to @xmath197 yields @xmath155 .
_ by claim 4 we assume that @xmath155 is reducible with respect to merge 3 .
let @xmath41 be the protagonist in the application of merge 3 to @xmath155 which neither touches an entry from below nor terminates in row @xmath584 , whose row @xmath56 of the place of change is maximal .
we bipartite the vertical pointers of @xmath155 into two sets @xmath646 and @xmath647 .
the set @xmath647 includes the vertical pointer in @xmath155 that results in @xmath41 and if @xmath650 and the vertical pointer @xmath651 results in a protagonist that touches the protagonist that corresponds to @xmath652 from below then @xmath653 .
all other vertical pointers are in @xmath646 .
observe that by the algorithm a candidate associated with @xmath649 can be a protagonist only after the candidates associated with @xmath648 have moved strictly above the @xmath56-th row or changed to horizontal candidates : by the proof of claim 3 and 4 a candidate @xmath257 associated with @xmath649 and unequal to @xmath41 can only be a protagonist if @xmath41 was already a protagonist and has moved to a row above of @xmath257 .
therefore we only have to show that @xmath41 is a protagonist only after the pointers associated with @xmath648 have moved strictly above the @xmath56-th row or changes to horizontal candidates .
but this is obvious since if @xmath41 is a protagonist in a step of merge 3 then every vertical candidate weakly below the @xmath56-th row belongs to @xmath649 .
( after @xmath41 is protagonist for the first time and until it changes into a horizontal candidate , @xmath41 is greater than every vertical candidate strictly above of the current row of @xmath41 and strictly below the @xmath466-th row .
moreover if there exist vertical candidates in this phase , which are strictly below the current row of @xmath41 then they are greater than @xmath41 with respect to the backward paths order and their places of change are strictly below the current row of @xmath41 .
see the proof of claim 3 and 4 . )
we denote the application of merge 3 to @xmath155 restricted to the vertical pointers in @xmath646 by merge 3.1 and the application of merge 3 to @xmath155 restricted to the vertical pointers in @xmath647 by merge 3.2 .
then the application of merge 3 is equivalent to the simultaneous application of merge 3.1 and merge 3.2 , where merge 3.2 starts after the vertical pointers that came from @xmath646 are strictly above of the @xmath56-th row or have changed to horizontal candidates . by induction hypothesis we reobtain @xmath155 from @xmath197 if we apply split 3 simultaneously to the protagonists of merge 3.1 together with the minimal horizontal candidates @xmath257 from the if - condition ( we denote this restricted application of split 3 by split 3.1 ) and to the protagonists of merge 3.2 together with these horizontal candidates ( we denote this restricted application of split 3 by split 3.2 ) such that when a candidate in split 3.1 moves weakly below the @xmath56-th row the application of split 3.2 has already finished .
note that @xmath41 is not touched from above in the application of split 3 to @xmath155 ( if we apply it to all candidates at once ) , for @xmath41 does not touch from below in the application of merge 3 to @xmath155 .
consequently if we apply split 3 unrestricted to @xmath155 , then this application starts with the application of split 3.1 , until we are either in case 3 of split 3 with the @xmath257 in the description of case 3 in split 3 being equal to @xmath41 or split 3.1 has terminated . in the first case
we perform split 3.2 and after the termination of split 3.2 we continue with the application of split 3.1 . in the second case
the set @xmath195 is empty immediately after the termination of split 3.1 and @xmath41 is the candidate in @xmath196 with minimal row .
thus we terminate with the application of split 3.2 in this case . | we present a bijective proof of the hook - length formula for shifted standard tableaux of a fixed shape based on a modified jeu de taquin and the ideas of the bijective proof of the hook - length formula for ordinary standard tableaux by novelli , pak and stoyanovskii @xcite . in their proof
novelli , pak and stoyanovskii define a bijection between arbitrary fillings of the ferrers diagram with the integers @xmath0 and pairs of standard tableaux and hook tabloids . in our shifted version of their algorithm the map from the set of arbitrary fillings of the shifted ferrers diagram onto the set of shifted standard tableaux is analog to the construction of novelli ,
pak and stoyanovskii , however , unlike to their algorithm , we are forced to use the rowwise total order of the cells in the shifted ferrers diagram rather than the columnwise total order as the underlying order in the algorithm .
unfortunately the construction of the shifted hook tabloid is more complicated in the shifted case . as a side - result we obtain a simple random algorithm for generating shifted standard tableaux of a given shape , which produces every such tableau equally likely . |
the peripheral fragmentation of light relativistic nuclei can serve as a source of information about their exitations above particle decay thresholds including many - body final states .
the interactions of this type are provoked either in electromagnetic and diffraction processes , or in nucleon collisions at small overlapping of the colliding nucleus densities .
a fragmenting nucleus gains an excitation spectrum near the cluster dissociation thresholds . in the kinetic region of fragmentation of a relativistic nucleus there
are produced nuclear fragment systems the total charge of it is close to the parent - nucleus charge .
a relative intensity of formation of fragments of various configurations makes it possible to estimate the importance of different cluster modes .
the opening angle of the relativistic fragmentation cone is determined by the fermi - momenta of the nucleon clusters in a nucleus .
being normalized to the mass numbers they are concentrated with a few percent dispersion near the normalized momentum of the primary nucleus .
when selecting events with dissociation of a projectile into a narrow fragmentation cone we see that target - nucleus non - relativistic fragments either are absent ( whitestars in ref.@xcite ) , or their number is insignificant . the target fragments are easily separated from the fragments of a relativistic projectile since their fraction in the angular relativistic fragmentation cone is small and they possess non - relativistic momentum values . in the peripheral fragmentation of a relativistic nucleus with charge z the ionization induced by the fragments can decrease down to a factor z , while the ionization per one track down to z@xmath8 .
therefore experiment should provide an adequate detection range . in order to reconstruct an event , a complete kinematic information about the particles in the relativistic fragmentation cone is needed which , e.g. , allows one to calculate the invariant mass of the system . the accuracy of its estimation decisively depends on the exactness of the track angular resolution . to ensure the best angular resolution ,
it is necessary that the detection of relativistic fragments should be performed with the best spacial resolution .
an event of the type of whitestar from the fragmentation of a relativistic @xmath0be nucleus into two he fragments in emulsion .
the photograph was obtained on the pavikom(fian ) complex.,width=480 ] the nuclear emulsion technique , which underlies the becquerel project at the jinr nuclotron @xcite , well satisfies the above - mentioned requirements .
it is aimed at a systematic search for peripheral fragmentation modes with statistical provision at a level of dozens events , their classification and angular metrology .
emulsions provide the best spacial resolution ( about 0.5 @xmath9 m ) which allows one to separate the charged particle tracks in the three - dimensional image of an event within one - layer thickness ( 600 @xmath9 m ) and ensure a high accuracy of angle measurements .
the tracks of relativistic h and he nuclei are separated by sight . as a rule , in the peripheral fragmentation of a light nucleus its charge can be determined by the sum of the charges of relativistic fragments .
multiple - particle scattering measurements on the light fragment tracks enable one to separate the h and he isotopes .
the analysis of the products of the relativistic fragmentation of neutron - deficient isotopes has some additional advantages owing to a larger fraction of observable nucleons and minimal coulomb distortions .
irradiation details and a special analysis of interactions in the br-2 emulsion are presented in ref .
@xcite . in
what follows , we give the first results of the study of the @xmath0be,@xmath3b , @xmath7be @xmath4n nuclei fragmentation with a few a gev energy which are obtained with the use of a part of the material analyzed .
the @xmath0be nucleus is a loosely bound n+@xmath1+@xmath1 system .
the energy threshold of the @xmath0be@xmath5n+@xmath1+@xmath1 dissociation channel is 1.57 mev .
the study of the @xmath0be fragmentation at relativistic energies gives the possibility of observing the reaction fragments , which are the decay products of unbound @xmath3be and @xmath10he nuclei .
the method of nuclear emulsions used in the present paper allows one to observe the charged component of the relativistic @xmath0be@xmath52he+n fragmentation channel . owing to a good angular resolution of this method it is possible to separate the @xmath0be fragmentation events , which accompanied by the production of an unstable @xmath3be nucleus with its subsequent breakup to two a particles . in this case , the absence of a combinatorial background ( of three and more @xmath1 particles ) for @xmath0be , which is typical for heavier n@xmath1 nuclei @xmath11c and @xmath12o makes it possible to observe distinctly this picture .
nuclear emulsions were exposed to relativistic @xmath0be nuclei at the jinr nuclotron .
a beam of relativistic @xmath0be nuclei was obtained in the @xmath13b@xmath14be fragmentation reaction using a polyethylene target .
the @xmath0be nuclei constituted about 80% of the beam , the remaining 20% fell on li and he nuclei.@xcite events were sought by microscope scanning over the emulsion plates . in total 362 events of the @xmath0be fragmentation involving the two he fragment production in the forward fragmentation cone within a polar angle of 6@xmath15(0.1 rad ) were found .
the requirement of conservation of the fragment charge in the fragmentation cone was fulfilled for the detected events . in event
selection 5 - 7 tracks of various types were allowed in a wide ( larger than 6@xmath15 ) cone to increase statistics .
an example of the @xmath0be@xmath52he fragmentation event in emulsion is given in fig .
[ fig:1 ] @xcite .
this event belongs to the class of whitestars as far as it contains neither target nucleus fragments , nor produced mesons .
this event sample includes 144 white stars .
the angles of the tracks in emulsion for the detected events were obtained using a fine measuring microscope .
angular measurements for the 362 events were carried out with an accuracy not worse then 4.5@xmath1610@xmath17 rad . in analyzing the data both he fragments observed in the @xmath0be@xmath52he+n channel were supposed to be a particles .
this assumption is motivated by the fact that at small angles the @xmath0be@xmath18he+n fragmentation channel with an energy threshold of 1.57 mev must dominate the @xmath0be@xmath19he+@xmath20he+n channel whose energy threshold is 22.15 mev .
the @xmath21he fraction will not exceed a few percent in this energy range @xcite and all the he fragments in the detected events may be thought of as @xmath1 particles . the p@xmath22 transverse momentum distribution of @xmath1 particles in the laboratory system ( a ) , and the p@xmath23 momentum distribution in the c.m.s .
of an @xmath1 particle pair ( b ) .
the outer contour corresponds to all events .
the inner histogram is obtained for events , which accompanied by protons recoil of emulsion target ( dashed area).,width=480 ] in fig .
[ fig:2]a the p@xmath22 transverse momentum distribution of @xmath1 particles in the laboratory frame of reference is calculated without the account of particle energy losses in emulsion by the equation @xmath24 where p@xmath25 , a and @xmath26 are the momentum per nucleon , the fragment mass and the polar emission angle , respectively .
the outer contour corresponds to all events .
the inner histogram is obtained for events accompanied by protons recoil of emulsion target ( dashed area ) .
the mean value of the transverse momentum for the total event sample in the laboratory system is equal to @xmath27103 mev / c with fwhm @xmath2872 mev / c .
this may be an indication of the fact that the experimental data are not of the same kind which can be pronounced when going over to the c.m.s . of two @xmath1 particles . the p@xmath23 transverse momentum distribution of @xmath1 particles in the c.m.s
of two @xmath1 particles described by the equation @xmath29 where p@xmath30 is the transverse momentum of an i - th @xmath1 particle in the laboratory system n@xmath31=2 is given in fig .
[ fig:2]b .
there is observed a grouping of events around two peaks with the values @xmath3224 mev / c and @xmath32101 mev / c . in ref
@xcite the appropriate mean values of the @xmath1 fragment transverse momenta are @xmath32121 mev / c for @xmath12o@xmath54@xmath1,@xmath32141 mev / c @xcite for @xmath11c@xmath53@xmath1 and @xmath32200 mev/ for @xmath33ne@xmath55@xmath1 ( processing of the available data ) . there by we clearly see a tendency toward an increase of the mean @xmath1 particle momentum with increasing their multiplicity .
this implies a growth of the total coulomb interaction of alpha clusters arising in nuclei . the opening @xmath34 angle distribution of @xmath1 particles in the @xmath0be@xmath52@xmath1 fragmentation reaction at 1.2 a gev energy .
the outer contour corresponds to all events .
the inner histogram is obtained for events , which accompanied by protons recoil of emulsion target ( dashed area).,width=480 ] in the opening angle @xmath34 distribution ( fig .
[ fig:3 ] ) one can also see two peaks with mean values 4.6@xmath35rad . and 26.8@xmath35rad .
the ratio of the numbers of the events in the peaks is close to unity .
the @xmath34 distribution entails the invariant energy q@xmath36 distribution , which is calculated as a difference between the effective invariant mass m@xmath36 of an @xmath1 fragment pair and the doubled @xmath1 particle mass by the equations @xmath37 where p@xmath38 is the @xmath1 particle 4-momentum . the invariant energy q@xmath36 distribution of @xmath1 particle pairs in the @xmath0be@xmath52@xmath1 fragmentation reaction at 1.2 a gev energy .
on the intersection : the q@xmath36 range from 0 to 1 mev .
arrows mark the @xmath3be nucleus levels ( 0@xmath2 and 2@xmath2 ) .
the outer contour corresponds to all events .
the inner histogram is obtained for events , which accompanied by protons recoil of emulsion target ( dashed area).,width=480 ] in the invariant energy q@xmath36 distribution ( fig .
[ fig:4 ] ) there are two peaks in the ranges 0 to 1 mev and 2 to 4 mev .
the shape of the distribution does not contradict the suggestion about the @xmath0be fragmentation involving the production of an unstable @xmath3be nucleus which decays in the 0@xmath2 and 2@xmath2 states .
the values of the peaks of the invariant energy q@xmath36 and the transverse momenta p@xmath39 in the c.m.s . relate to each other . to the q@xmath36 range from 0 to 1 mev with a peak at 100 kev
there corresponds a peak p@xmath23 with @xmath3224 mev / c , and to the q@xmath36 range from 2 to 4 mev there corresponds a peak with @xmath32101 mev / c .
a stack of layers of br-2 emulsion was exposed to a beam of @xmath4n nuclei accelerated @xcite to a momentum of 2.86 a gev / c at the nuclotron of the laboratory of high energy physics ( jinr ) .
already been found amoung 950 inelastic events in which the total fragment charge was equal to the z@xmath25=7 fragment charge and there were no produced particles .
events were sought by viewing over the track length which provided the accumulation of statistics without selection .
the selected events are divided in two classes .
the events of the type of whitestar and the interactions involving the production of one or a few target - nucleus fragments belong to the first class .
table [ aba : tab1 ] shows the charge multi - fragmentation topology which was studied for the events satisfying the above - mentioned conditions .
the upper line is the z@xmath402 fragment charge , the second line is the number of single - charged fragments , the third one the number of two - charged fragments , and the fourth and fifth lines are the number of the detected events with a given topology for whitestars and events with target - nucleus excitation for each channel , respectively .
the two last lines present the total number of interactions calculated in absolute values and in percent .
l@|@c@|@c@| @c@|@c@| @c@| @c@| @c@|@c@ z@xmath41 & 6 & 5 & 5 & 4 & 3 & 3 & &
+ n@xmath42 & 1 & & 2 & 1 & 4 & 2 & 3 & 1 + n@xmath43 & & 1 & & 1 & & 1 & 2 & 3 + n@xmath44 & 13 & 4 & 3 & 1 & 1 & 1 & 6 & 17 + n@xmath45 & 15 & 1 & 3 & 3 & & 2 & 5 & 32 + n@xmath46 & 28 & 5 & 6 & 4 & 1 & 3 & 11 & 49 + n@xmath47 & 26 & 5 & 5 & 4 & 1 & 3 & 10 & 46 + the analysis of the data of table [ aba : tab1 ] shows that the number of channels involving z@xmath403 fragments for the whitestars is larger by about a factor of 1.5 than that for the events accompanied by a target breakup .
on the contrary , for the 2 + 2 + 2 + 1 charge configuration channel this number is smaller by about a factor of 1.5 .
thus , in the events with target breakup , the projectile fragments more strongly than in the whitestars .
the data of table [ aba : tab1 ] points to the predominance of the channel with the 2 + 2 + 2 + 1 charge configuration ( 49 events ) which has been studied in more detail .
the obtained results show that the @xmath4n nucleus constitutes a very effective source for the production of 3@xmath1 system . in order to estimate the energy scale of production of 3@xmath1 particle systems in the @xmath4n@xmath53@xmath1+x channel , we present the invariant excitation energy q@xmath48 distribution with respect to the @xmath11c ground state : @xmath49 where m(@xmath50 ) is the mass of the ground state corresponding to the charge and the weight of the system being analyzed , m@xmath51 the invariant mass of the system of fragments .
statistics was increased to 132 events @xmath4n@xmath53@xmath1+x including 50 white stars by scanning over the emulsion plates . the invariant excitation energy q@xmath48 distribution of three @xmath1 particles with respect to the @xmath11c ground state for the process @xmath4n@xmath53@xmath1+x .
the following notation is used@xmath52 1 ) all the events of the given dissociation , 2 ) whitestars.,width=480 ] the main part of the events is concentrated in the q@xmath48 area from 10 to 14 mev , covering the known @xmath11c levels ( fig . [ fig:5 ] ) .
softening of the conditions of the 3he + h selection , for which the target fragment production is allowed , does not result in a shift of the 3@xmath1 excitation peak .
this fact suggests the universality of the 3@xmath1 state population mechanism .
the invariant excitation energy q@xmath36 distribution of @xmath1 particle pairs for the process @xmath4n@xmath53@xmath1+x . in the inset : a fraction of the distribution at 0 - 500 kev .
, width=480 ] to estimate the fraction of the events involving the production of an intermediate @xmath3be nucleus in the reactions @xmath4n@xmath53be+x@xmath53@xmath1+x we present the invariant excitation energy distribution for an @xmath1 particle pair with respect to the @xmath3be ground state ( fig .
[ fig:6 ] ) .
the first distribution peak relates to the value to be expected for the decay products of an unstable @xmath3be nucleus in the ground state 0@xmath2 .
the distribution centre is seen to coincide well with the decay energy of the @xmath3be ground state .
the fraction of the @xmath1 particles originating from the @xmath3be decay is 25 - 30% .
the results of investigations dealing with the charge topology of the fragments produced in peripheral dissociation of relativistic @xmath3b , @xmath7be nuclei in emulsion are presented in ref @xcite .
@c|c|c|c|c|c|c|c|c|c@ channel&2he&2he&he+2h&he+2h&4h&4h&li+h&li+h&sum + & n@xmath540&n@xmath550&n@xmath540&n@xmath550&n@xmath540&n@xmath550&n@xmath540&n@xmath550 & + @xmath21he+@xmath20he&30&11 & & & & & & & 41 + @xmath21he+@xmath21he&11 & 7 & & & & & & & 18 + @xmath20he+2p & & & 13 & 9 & & & & & 22 + @xmath20he+d+p & & & 10 & 5 & & & & & 15 + @xmath21he+2p & & & 9 & 9 & & & & & 18 + @xmath21he+d+p & & & 8&10 & & & & & 18 + @xmath21he+2d & & & 1 & & & & & & 1 + @xmath21he+t+p & & & 1 & & & & & & 1 + 3p+d & & & & & 2 & & & & 2 + 2p+2d & & & & & 1 & & & & 1 + @xmath56li+p & & & & & & & 9 & 3 & 12 + sum & 41&18&42&33 & 2 & 1 & 9 & 3&149 + table [ aba : tab2 ] presents the numbers of the events detected in various channels of the @xmath7be fragmentation . of them
, the @xmath21he+@xmath20he channel noticeably dominates , the channels @xmath20he+d+p and @xmath56li+p constitute 10% each .
two events involving no emission of neutrons in the three - body channels @xmath21he+t+p and @xmath21he+d+d were registered . the reaction of charge - exchange of @xmath7be nuclei to @xmath7li nuclei
was not detected among the events not accompanied by other secondary charged particles.the events involving no target fragments ( n@xmath57=0 ) are separated from the events involving one or a few fragments ( n@xmath580 ) . for the first time , nuclear emulsions were exposed to a beam of relativistic @xmath3b nuclei .
we have obtained data on the probabilities of the @xmath3b fragmentation channels in peripheral interactions .
55 events of the peripheral @xmath3b dissociation which do not involve the production of the target - nucleus fragments and mesons ( white stars ) were selected .
a leading contribution of the @xmath3b@xmath59be+p mode having the lowest energy threshold was revealed on the basis of these events .
information about a relative probability of dissociation modes with larger multiplicity have been obtained . among the found events there are 320 stars in which the total charge of the relativistic fragments in a 8@xmath15 fragmentation cone @xmath60z@xmath41 satisfies the condition @xmath60z@xmath613 .
these stars were attributed to the number of peripheral dissociation events n@xmath62 .
the n@xmath62 relativistic fragment distribution of over charges n@xmath63 is given in table [ aba : tab3 ] .
there are given the data for 256 events containing the target - nucleus fragments - n@xmath64 , as well as for 64 events which contain no target - nucleus fragments ( white stars ) n@xmath62 .
the role of the channels with multiple relativistic fragments n@xmath652 is revealed to be dominant for the nwhite stars .
of peripheral events , the white stars n@xmath66 ( table [ aba : tab3 ] ) are of very particular interest .
they are not accompanied by the target - nucleus fragment tracks and makes it possible to clarify the role of different cluster degrees of freedom at a minimal excitation of the nuclear structure .
@c| c | c |c |c |c |c |c@ z@xmath41 & n@xmath67 & n@xmath68 & n@xmath69 & n@xmath70 & n@xmath71 & n@xmath64 & n@xmath66 + 7 & - & - & - & 1 & 5 & 1 & - + 6 & - & - & - & 2 & 2 & 8 & 2 + 6 & - & - & - & 1 & 4 & 6 & 4 + 6 & - & - & - & - & 6 & 1 & - + 5 & - & - & - & 1 & 3 & 61 & 14 + 5 & - & - & - & 2 & 1 & 44 & 12 + 5 & - & - & 1 & - & 2 & 8 & - + 5 & - & - & 1 & 1 & - & 1 & - + 5 & - & 1 & - & - & 1 & 17 & 24 + 5 & 1 & - & - & - & - & 17 & 1 + 5 & - & - & - & - & 5 & 21 & 4 + 4 & - & - & - & - & 4 & 5 & 1 + 4 & - & - & - & 2 & - & 24 & 1 + 4 & - & - & - & 1 & 2 & 42 & - + @c |c| c| c| c@ @xmath60z@xmath41=4 & @xmath7be & % & @xmath3b ( + h ) & % + 2he & 41 & 43 & 12 & 40 + he+2h & 42 & 45 & 14 & 47 + 4h & 2 & 2 & 4 & 13 + table [ aba : tab4 ] gives the relativistic fragment charge distribution in the white stars for @xmath72be and @xmath73b nuclei .
the @xmath73b events are presented without one single - charged relativistic fragment , that is a supposed proton halo .
the identical fraction of the two main 2he and he+2h dissociation channels is observed for @xmath72be and @xmath73b nuclei which points out that the @xmath73be core excitation is independent of the presence of an additional loosely bound proton in the @xmath73b nucleus .
the degree of the dissociation of the relativistic nuclei in peripheral interactions can reach a total destruction into nucleons and singly and doubly charged fragments .
the emulsion technique allows one to observe these systems to the smallest details and gives the possibility of studying them experimentally .
new experimental observations are reported from the emulsion exposures to @xmath4n , @xmath0be , @xmath3b , @xmath7be nuclei with energy above 1 a gev .
the main features of @xmath0be@xmath52he relativistic fragmentation are presented . for the particular case of the relativistic @xmath0be nucleus dissociation
it is shown that precise angular measurements play a crucial role in the restoration of the excitation spectrum of the alpha particle fragments .
this nucleus is dissociated practically totally through the 0@xmath2 and 2@xmath2 states of the @xmath3be nucleus .
the data obtained from @xmath0be angular measurements can be employed for the estimation of the role of @xmath3be in more complicated n@xmath1 systems .
the results of the study of the dissociation of @xmath4n nuclei of a primary momentum of 2.86 a gev / c in their interactions with the emulsion nuclei are also presented .
the present investigation indicates the leading role of the 2 + 2 + 2 + 1 charge configuration channel .
the energy scale of the 3@xmath1 system production has been estimated . according to the available statistics 80% of interactions
are concentrated at 10 - 14 mev .
the fraction of the @xmath4n@xmath53be+x@xmath53@xmath1+x channel involving the production of an intermediate @xmath3be nucleus is about 25% .
advantages of emulsion technique are exploited most completely in the study of peripheral fragmentation of light stable and neutron deficient nuclei .
the results of investigations dealing with the charge topology of the fragments produced in peripheral dissociation of relativistic @xmath7be , @xmath3b nuclei in emulsion are presented .
information on the relative probability of dissociation modes with a larger multiplicity was obtained .
the dissociation of a @xmath7be core in @xmath3b indicates an analogy with that of the free @xmath7be nucleus . | recent studies of clustering in light nuclei with an initial energy above 1 a gev in nuclear treack emulsion are overviewed .
the results of investigations of the relativistic @xmath0be nuclei fragmentation in emulsion , which entails the production of he fragments , are presented .
it is shown that most precise angular measurements provided by this technique play a crucial role in the restoration of the excitation spectrum of the @xmath1 particle sysytem . in peripheral interactions @xmath0be
nuclei are dissociated practically totally through the 0@xmath2 and 2@xmath2 states of the @xmath3be nucleus .
the results of investigations of the dissociation of a @xmath4n nucleus of momentum 2.86 a gev / c in emulsion are presented as example of more complicated system .
the momentum and correlation characteristics of @xmath1 particles for the @xmath4n@xmath53@xmath6 channel in the laboratory system and the rest systems of 3@xmath1 particles were considered in detail .
topology of charged fragments produced in peripheral relativistic dissociation of radioactive @xmath3b , @xmath7be nuclei in emulsion is studied . |
a typical wireless sensor network consists of sensor nodes with limited energy reserves .
many sensor network applications expect the sensor nodes to be active for months and may be years . however , in most of the situations , once the sensor nodes run out of their energy reserves , then replacing their batteries is not possible either due to the inaccessibility of sensor nodes or because such an endeavor may not be economically viable .
so , there is a great demand and scope of the strategies which attempt to reduce the energy consumption , hence increase the lifetime of the sensor nodes .
sensor nodes spend energy in receiving and transmitting data , sensing / actuating , and computation . in this paper , we concern ourselves with reducing the energy cost of transmission .
the energy cost of receiving data can be easily incorporated in the model that we propose .
the energy spent in sensing / actuating represents a fixed cost that can be ignored .
we assume computation costs are negligible compared to radio communication costs . though this is a debatable assumption in dense networks , but incorporating computation costs is not straightforward and is left for future work .
the transmission energy depends on three factors : the number of bits to be transmitted , the path - loss factor between the sensor nodes and the base - station , and the time available to transmit the given number of bits .
the path - loss factor describes the wireless channel between a sensor node and the base - station and captures various channel effects , such as distance induced attenuation , shadowing , and multipath fading . for simplicity
, we assume the path - loss factor to be constant .
this is reasonable for static networks and also the scenarios where the path - loss factor varies slowly and can be accurately tracked .
the idea of varying the transmission time to reduce the energy consumption was proposed in @xcite and explored in @xcite in the context of sensor networks , where its @xmath1-hardness is discussed . in this paper
, we attempt to reduce the transmission energy by reducing the number of bits transmitted by the sensor nodes to the base - station . in a data gathering sensor network ,
the spatio - temporal correlations in sensor data induce data - redundancy . in @xcite , slepian and wolf
show that it is possible to compress a set of correlated sources down to their joint entropy , without explicit communication between the sources .
this surprising existential result shows that it is enough for the sources to know the probability distribution of data generated .
recent advances @xcite in distributed source coding allow us to take advantage of data correlation to reduce the number of bits that need to be transmitted , with concomitant savings in energy .
however , finding the optimal rate allocation lying in slepian - wolf achievable rate region defined by @xmath2 constraints for @xmath3 nodes and designing efficient distributed source codes is a challenging problem .
we simplify this problem by allowing the interaction between the base - station and the sensor nodes , and introducing the notion of _ instantaneous decoding _ @xcite .
this reduces the rate allocation problem to the problem of finding the optimal scheduling order , albeit at some loss of optimality .
two results in the theory of interactive communication provided further motivation for our work .
first result states that even if the feedback does not help in increasing the capacity of a communication channel , it does help in reducing the complexity of communication @xcite .
second result states that for the worst - case interactive communication , for almost all _ sparse pairs _ the recipient , who has nothing to say , must transmit almost all the bits in an optimal interactive communication protocol and informant transmits _ almost _ nothing @xcite . based on these results ,
this paper proposes a formalism that attempts to reduce the number of bits sent by a sensor node .
the proposed formalism casts the problem of many sensor nodes communicating with a base - station as a problem where multiple informants with correlated information communicate with single recipient . here
we identify the base - station as the recipient of the information and the sensor nodes as the sources of information .
however , @xcite and subsequent papers @xcite consider only the interactive communication between a single informant - recipient pair , while in the sensor networks , we have as many informant - recipient pairs , as the number of sensor nodes . if the sensor data is assumed to be uncorrelated , then the results of @xcite can be trivially extended to the present scenario .
however , in a data - gathering sensor network , the sensor data is supposed to be correlated .
so , extending the results in @xcite to the scenarios where multiple informants with correlated information communicate with single recipient and then applying those results to the sensor networks , is not straightforward .
formally , the problem is the following .
there are @xmath3 sources of correlated information and there is one recipient of information @xmath4 , which needs to collect the information from these @xmath3 sources .
assume that @xmath4 can interact with any of those @xmath3 sources , but among themselves these @xmath3 sources can not interact directly .
at the end of communication , each of these @xmath3 sources need not know what other @xmath5 sources know .
we are looking for the most efficient communication schemes ( ones which minimize the communication complexity ) for this problem . in the context of sensor networks , if the base - station knows the joint distribution of the sensor data or the correlations therein , then it can play some distributed version of the ` game of twenty questions ' with @xmath3 sensors to retrieve their information .
if there is a single sensor node or if the sensor data are uncorrelated , then the base - station can ` play ' with individual sensor nodes and any node @xmath6 needs to send at least @xmath7 bits , where @xmath7 is the entropy of the information source at node @xmath6 .
so , assuming the sensor data at all the nodes are _ identically distributed _
, finally the base - station retrieves @xmath8 bits from the @xmath3 sensors .
however , in a more realistic situation , the sensors have correlated data and the base - station needs to retrieve only @xmath9 bits . note that we can talk in terms of entropies only when we are concerned with average number of bits communicated .
however , if we are interested in the worst - case number of bits , then it depends solely on the cardinality of the _ support set _ of the data of individual nodes than on the corresponding probabilities . in this work , we make two simplifying assumptions : the sensor nodes communicate with the base - station in a single - hop and only the total number of bits exchanged between the sensor nodes and the base - station are considered and this is what we refer to as ` communication complexity ' .
this paper shows that the communication complexity depends on the model of the spatial - correlation of the sensor data as well as on the order in which the sensor nodes communicate with the base - station . under the assumption of an omniscient base - station
, the base - station can compute the optimal number of bits , which any node in any schedule needs to send to it for any spatial correlation model of the sensor data .
however , a sensor node , even if it knows in how many bits it needs to send its information to the base - station , may not actually be able to compress its information to that many number of bits , without explicit knowledge of how its data is correlated with the data of the other nodes . in general
, this is neither possible given the limited knowledge of the network that a sensor node is supposed to have , nor desirable given the limited energy and computational capabilities of the sensor nodes .
an omnipotent base - station can take up the most of the burden of computation and communication , allowing the the sensor nodes to perform minimal computation and transmit minimal number of bits , hence conserving their precious energy reserves and possibly increasing their operational lifetime .
in the section [ example ] , we give an example of the situation where multiple informants with correlated information communicate with single recipient .
we use this example to illustrate the complexity of communication in such scenarios and various other fundamental issues and our results .
section [ worst_case_theory ] develops the formalism to compute the worst - case communication complexity in the scenarios where multiple correlated informants communicate with a single recipient .
section [ sensornet_appl ] illustrates some of the ideas proposed and developed in this paper in the context of a sensor network with one particular model of spatial correlation of sensor data .
finally , the section [ conclusions ] lists the contributions of our work and concludes the paper .
there are @xmath3 groups @xmath10 , and each group has @xmath11 teams . in every match , two teams from two different groups play against each other . the result of each match is announced over radio
. matches always result in a clear winner .
the format of the radio announcements is : `` today the teams from groups @xmath12 and @xmath13 played against each other .
the match between teams @xmath6 and @xmath14 was won by team @xmath15 . ''
three persons @xmath16 , @xmath17 , and @xmath18 are involved .
@xmath16 listens to the first part of the announcement `` today the teams from groups @xmath12 and @xmath13 played against each other . ''
and then the radio is snatched by person @xmath17 , who listens to the portion `` the match between teams @xmath6 and @xmath14 was won by '' before the radio is snatched from him by @xmath18 , who listens to the portion `` team @xmath15 '' .
now all three persons agree that @xmath16 must know which two teams played and who the winner was ( @xmath17 and @xmath18 need not learn which two teams from which two groups actually played and which one finally won ) .
they want to find the most efficient way to do this .
@xmath16 wants to know which two teams from the groups @xmath12 and @xmath13 played and which one actually won .
suppose only @xmath17 communicates with @xmath16 , then @xmath16 will only know which two teams from the groups @xmath12 and @xmath13 actually played the match , but not the winner .
on the other hand , if only @xmath18 communicates with @xmath16 , then @xmath16 knows the teams from which two groups played and who was the winner , but he may not know who the other team was . in this problem
, it is essential that the identity of the group to which a team belongs to is included in the name of the team , that is , the names of the teams have to be globally unique .
suppose on the contrary , that the names of the teams are only unique within a group , then two or more groups might have the the teams with the same name .
so , in the event of a match between two teams from two different groups with the same name playing against each other , @xmath16 will not be able to make the winner out of the information sent to him by @xmath18 , even if @xmath17 has already informed @xmath16 of which two teams have played .
so , we are demanding that once @xmath17 and @xmath18 have communicated all their information to @xmath16 , then @xmath16 must be able to unambiguously infer which two teams had played as well as who the winner was .
the previous argument proves that this demand is satisfied if and only if the team names are globally unique .
it should be noted that the total number of bits exchanged to complete the entire process of communication depends very much on the format of the announcement .
for example , in the original problem in @xcite , if the format of the announcement is : `` the match between teams @xmath6 and @xmath14 is won by first / second team '' with the protocol that the ` first ' ( ` second ' ) corresponds to the first ( second ) team mentioned in the announcement . in such a situation , only one bit needs to be sent to @xmath17 , with or without interaction between @xmath16 and @xmath17 .
@xmath17 sends @xmath19 bits to identify one of the groups and the team from it . after this message
, it does not need to identify the second group as it is obvious to @xmath16 .
so , @xmath17 needs to send another @xmath20 bits to enable @xmath16 know which is the team from the other group ( message 1 , 2 ) .
@xmath18 sends @xmath21 bits to @xmath16 to help him know who the winner was ( message 3 ) .
@xmath22 there are two scenarios here .
one where @xmath17 communicates its information to @xmath16 before @xmath18 communicates with @xmath16 .
the other scenario is the one where @xmath18 sends his information to @xmath16 before @xmath17 sends .
we have to compute the number of bits exchanged for both the scenarios .
* when @xmath17 communicates with @xmath16 before @xmath18 : * @xmath16 knows the two groups from which two teams played against each other .
@xmath16 encodes in @xmath23 bits the names of the groups .
so , @xmath16 sends in @xmath24 bits the first bit location at which the encodings of the two groups differ to @xmath17 ( message 1 ) .
@xmath17 on its turn , sends the value of the first bit at which the encodings of two groups differ along with the @xmath20 bits to identify the team within one of the groups . to help identify the team from the other group it just needs to send @xmath20 bits to @xmath16 as the identity of the group to which this team belongs to
is already known to @xmath16 by now ( message 2 , 3 , 4 ) . at the end of this step ,
@xmath16 knows which two teams had played .
so , now in @xmath25 bits it sends the identity of the first bit location at which the encodings of the two teams differ to @xmath18 ( message 5 ) and @xmath18 responds by sending the value of that bit ( message 6 ) . with this @xmath16
can determine who the winner was .
so under this scenario , the number of bits exchanged are : @xmath26 * when @xmath18 communicates with @xmath16 before @xmath17 : * @xmath16 sends in @xmath24 bits the location of the first bit at which the encodings of the two groups differ to @xmath18 ( message 1 ) .
@xmath18 in its turn , sends the value of the bit at that location along with the encoding of the winning team in @xmath20 number of bits ( message 2 , 3 ) . at the end of this step
, @xmath16 knows which was the winning team and to which group it belonged to too .
so , all that @xmath16 does not know now is that which team from the other group also played in the match .
it sends to @xmath17 the location and value of the first bit at which the encodings of two two teams differ in @xmath27 bits ( message 4 , 5 ) .
@xmath17 then responds by sending @xmath20 number of bits to identify the team from the given group ( message 6 ) .
so under this scenario , the number of bits exchanged are : @xmath28 note , , and show that if we consider , total number of bits exchanged in the entire communication or the total number of bits transmitted by persons @xmath17 and @xmath18 together , interaction helps in reducing the number of bits compared to when no interaction is allowed .
further , when @xmath18 communicates with @xmath16 before @xmath17 , the number of exchanged bits are less than when @xmath17 communicates with @xmath16 before @xmath18 . if we adopt the convention that all the messages sent by a source , until some other source sends the messages , form one message , then in all the above situations at most four messages are exchanged .
so , in a communication protocol following this convention , the source concatenates all the messages that it sends before some other source begins to send the messages , and receiver knows how to parse the concatenated message into its individual messages . as @xcite and the example above
prove , the interaction reduces the number of bits exchanged between the informants and the recipient .
however , from the above example , it also becomes clear that even for the given number of messages exchanged between the informants and the recipient , in general , the number of bits exchanged depends on the order in which the informants communicate with the recipient .
so in the above example , with four messages allowed , the number of bits exchanged between @xmath29 , and @xmath18 depend on whether @xmath17 communicates with @xmath16 before @xmath18 or after .
we conjecture that this is so due to `` somewhat '' asymmetric nature of the distribution of the information at the nodes @xmath17 and @xmath18 of this particular example and loosely , we can say that the messages from @xmath18 contain more information than those from @xmath17 .
in this section we attempt to develop a theory of worst - case interactive communication . to keep the discussion simple and clear , we consider a general scenario involving two informants ( @xmath17 and @xmath18 ) and one recipient ( @xmath16 ) .
however , the same formalism can be extended for the scenarios involving more than two informants . for the sake of completeness and to facilitate the discussion that follows , let us reintroduce the notion of _ ambiguity set _ and some related concepts , defined originally in @xcite .
let @xmath30 be a random pair , with _ support set _ @xmath31 .
the _ support set _ of @xmath32 is the set @xmath33 of possible @xmath32 values .
@xmath34 , the _ support set _ of @xmath35 , is similarly defined .
s _ ambiguity set _ when his random variable @xmath35 takes the value @xmath36 is @xmath37 the set of possible @xmath32 values when @xmath38 .
ambiguity _ in that case is @xmath39 the number of possible @xmath32 values when @xmath40 .
the _ maximum ambiguity _ of @xmath30 is @xmath41 the maximum number of @xmath32 values possible with any given @xmath35 value .
assume that a total of @xmath42 messages are allowed , with @xmath0 messages per informant allowed to be exchanged between the informant and the recipient .
there are two possible schedules in which the informants can communicate with the recipient : either @xmath17 communicates first with @xmath16 or @xmath18 communicates first . in the spirit of @xcite ,
let us introduce some more definitions : @xmath43 : @xmath42-message worst - case complexity of transmitting @xmath44 to a person who already knows @xmath32 , when @xmath35 communicates first .
@xmath45 : @xmath42-message worst - case complexity of transmitting @xmath44 to a person who already knows @xmath32 , when @xmath46 communicates first . using these definitions
, we have : @xmath47 in general , @xmath48 .
we omit the detailed proof for the sake of brevity . however , the league example of the previous section provides an example supporting the statement of the theorem .
_ corollary : _ the unbounded interaction complexities satisfy @xmath49 .
it is easy to prove the following trivial , but quite useful lower bounds on the unbounded interaction complexities . for all ( x , y , z ) tuples
, @xmath50 similar bounds exists for @xmath51 , @xmath52 , and @xmath53 .
since empty messages are allowed , it is obvious that @xmath54 is a decreasing function of @xmath0 .
this holds true for other complexities , such as @xmath55 etc too .
this fact together with previous lemma implies that @xmath56 with similar bounds for @xmath55 , @xmath57 , and @xmath45 .
we can use above results to find the communication complexity of the version of the league problem discussed in the previous section and arrive at the results of and .
if we identify the sets @xmath58 , and @xmath46 appropriately , we can directly use the results from @xcite .
for example , the relevant support sets @xmath59 , @xmath60 , @xmath61 , and @xmath62 , with @xmath63 are : @xmath64
in this section , we apply the formalism developed in the previous sections to illustrate the computation of worst - case and average - case interactive communication complexities . we assume following spatial correlation model for the sensor data .
let @xmath65 be the random variable representing the sampled sensor reading at node @xmath66 and @xmath67 denote the number of bits that the node @xmath6 has to send to the base - station .
let us assume that each node @xmath6 has at most @xmath68 number of bits to send to the base - station , so @xmath69 .
however , due to the spatial correlation among sensor readings , each sensor may send less than @xmath68 number of bits .
let us define a data - correlation model as follows .
let @xmath70 denote the distance between nodes @xmath6 and @xmath14 .
let us define @xmath71 , the number of bits that the node @xmath6 has to send when the node @xmath14 has already sent its bits to the base - station , as follows : @xmath72 figure [ fig1 ] illustrates this for @xmath73 .
it should be noted that when @xmath74 , the data of the nodes @xmath6 and @xmath14 differ in at most @xmath71 least significant bits .
so , the node @xmath6 has to send , at most , @xmath75 least significant bits of its @xmath68 bit data
. from the definition above in , follows the symmetry of the conditional number of bits : @xmath76 however , the definition of the correlation model is not complete yet and we must give the expression for the number of bits transmitted by a node conditioned on more than one node already having transmitted their bits to the base - station .
there are several ways in which this quantity can be defined , we choose the following definition : @xmath77 so , with this definition , the number of bits transmitted by any node in a schedule depends only on the node nearest to it among all the nodes already polled in the schedule . for a given schedule @xmath78
, first node @xmath79 transmits its @xmath68 bits to the base - station . the omniscient base - station based on its knowledge of the correlation model in as well as all the internode distances
, knows that its _ maximum ambiguity _ about second node s data is @xmath80 , as @xmath80 bit - patterns are possible for the @xmath81 least - significant bits of @xmath82 s data .
so , even if we allow unbounded interaction between the base - station and the node @xmath82 , it follows from the results of the previous section that at least @xmath81 bits are exchanged , even by the optimal communication protocol . however , here we are more concerned with demonstrating the reduction in the communication complexity due to the interaction between the sensor nodes and the base - station rather than with proposing the schemes which achieve the optimal lower bounds of the communication complexity .
so , here we propose an _ almost optimal _
protocol for the communication between the base - station and a sensor node .
base - station informs the node @xmath82 in @xmath83 bits ( if @xmath84 , then the base - station sends one bit to the corresponding node ) to transmit @xmath81 least significant bits of its information and the sensor node @xmath82 responds by sending corresponding bits .
note that if no interaction is allowed between the base - station and the sensor node , then the sensor node has to send @xmath68 bits to the base - station . continuing this process , the base - station queries all the @xmath3 nodes
so , the worst - case communication complexity @xmath85 of schedule @xmath78 is the sum of the total number of bits sent by the base - station and the number of bits sent by the sensor - nodes , in the worst - case .
it is given by : @xmath86 let @xmath87 be the set of all possible schedules .
then optimally minimum value of @xmath85 is achieved for that schedule @xmath88 that solves the following optimization problem : @xmath89 given the definition of the correlation model in and , it is easy to see that the optimum schedule is generated by a greedy scheme that chooses the next node in the schedule ( from the set of the nodes not already scheduled ) to be that node that is nearest to the set of already scheduled nodes . as noted above in the discussion of the correlation model of , the data of two correlated nodes @xmath6 and @xmath14 can differ in at most @xmath71 number of least significant bits . however , it is not necessary that it _ actually _ differs in those many bits .
so , there can be the situations where even if the base - station has estimated that the node @xmath6 has to send @xmath71 number of its least - significant bits and communicated this to the node @xmath6 , the node @xmath6 s data differs from the the data of node @xmath14 in less than @xmath71 least - significant bits .
in such situations , it is sufficient for the base - station to reconstruct the data of the node @xmath6 if the node @xmath6 sends only those bits where its data actually differs from that of node @xmath14 . given the @xmath71 number of its least - significant bits that the node @xmath6 has to send to the base - station , there are @xmath90 possible bit - patterns , one out of which the node @xmath6 has to communicate to the base - station . assuming that the each of these @xmath90 bit - patterns are uniformly distributed with probability @xmath91 , then following the typical huffman - coding argument , the node @xmath6 can communicate its data to the base - station in at most @xmath71 bits on average .
given a communication schedule @xmath78 , the first node in the schedule @xmath79 sends its @xmath68 bits to the base - station .
based on this and the knowledge of the correlation model given in , the base - station informs the second node @xmath82 in @xmath83 bits to send its information in @xmath81 bits . assuming the uniform - distribution
, the node @xmath82 sends the requested information in @xmath81 bits , in average . continuing this process , the base - station queries all the @xmath3 nodes .
so , the average - case communication complexity @xmath92 of schedule @xmath78 is the sum of the total number of bits sent by the base - station and the number of bits sent , on average , by the sensor - nodes .
it is given by : @xmath93 the optimal value of @xmath92 is achieved for that schedule @xmath88 that solves the following optimization problem : @xmath94 it is easy to see that , once again , the optimum schedule is generated by a greedy scheme that chooses the next node in the schedule to be that node that is nearest to the set of the nodes which are already scheduled . _
remark : _ comparing and , it may appear that the average - case performance of a schedule is no better than its worst - case performance , but it should be noted that the average - case analysis is done for the uniform distribution of the bit - patterns , that gives the maximum entropy . for any other distribution of the bit - patterns , the average communication complexity will be lesser than @xmath92 given by .
this paper proposes a new framework , based on exploiting the redundancy induced by the spatio - temporal correlations in the sensor data and the reduction in communication complexity due to interaction , to reduce the total number of bits sent by the sensor nodes in a single - hop data - gathering sensor network .
the proposed formalism views the problem of many sensor nodes communicating with the base - station as the problem of many informants with the correlated information communicating with single recipient .
we extend various existing results on single informant - recipient pair communication to the present case .
we show that such extensions lead to various non - trivial new results .
finally , we apply this new framework to compute the worst - case and average - case communication complexities of a typical sensor network scenario and demonstrate the significance of our contribution . 1 .
we show that interaction helps in reducing the communication complexity also in the scenarios where more than one informant are involved .
2 . we show that when the _ ambiguity _ is more than two , then no fixed number of messages are optimum .
we show that for the scenarios involving more than one informant , the order in which the informants communicate with the recipient may determine the communication complexity .
we conjecture that if the nodes which have ` more ' information communicate first , then this brings down the overall communication complexity .
essentially , we need a metric to quantify the amount of information , but not in ` shannon - sense ' .
we show that when multiple informants communicate with a recipient , the m - message complexity of communication between informant @xmath35 and recipient @xmath32 , can be computed by directly modifying the hypergraph @xmath95 based on the information provided by the previous informants in the communication schedule . | in this work , we are concerned with maximizing the lifetime of a cluster of sensors engaged in single - hop communication with a base - station . in a data - gathering network ,
the spatio - temporal correlation in sensor data induces data - redundancy .
also , the interaction between two communicating parties is well - known to reduce the communication complexity .
this paper proposes a formalism that exploits these two opportunities to reduce the number of bits transmitted by a sensor node in a cluster , hence enhancing its lifetime .
we argue that our approach has several inherent advantages in scenarios where the sensor nodes are acutely energy and computing - power constrained , but the base - station is not so .
this provides us an opportunity to develop communication protocols , where most of the computing and communication is done by the base - station .
the proposed framework casts the sensor nodes and base - station communication problem as the problem of multiple informants with correlated information communicating with a recipient and attempts to extend extant work on interactive communication between an informant - recipient pair to such scenarios .
our work makes four major contributions .
firstly , we explicitly show that in such scenarios interaction can help in reducing the communication complexity .
secondly , we show that the order in which the informants communicate with the recipient may determine the communication complexity .
thirdly , we provide the framework to compute the @xmath0-message communication complexity in such scenarios .
lastly , we prove that in a typical sensor network scenario , the proposed formalism significantly reduces the communication and computational complexities . |
recent surprising progress in microfabrication techniques has enabled us to make tiny josephson junctions with artificial geometry .
this opens a door to explore various mesoscopic effects in the transport properties of small superconducting junctions . in this paper
we review the dc josephson effect in a superconducting quantum point contact@xcite , a josephson junction consisting of two superconductors coupled through a few quantum channels ( fig . 1 ) .
such a superconducting quantum point contact can be created in a superconductor2deg superconductor junction ( 2deg : two - dimensional electron gas ) @xcite as well as in a mechanically controllable break junction @xcite .
main focus will be on current - carrying bound states that appear near a quantum point contact .
our discussion is based on the bogoliubov - de gennes equation , which is a two - component schrdinger equation describing electron - like and hole - like quasiparticle excitations in superconducting systems .
we show that the bogoliubov - de gennes equation has solutions describing bound states that are localized around the junction and decay exponentially into the superconductors .
these bound states play a central role in the dc josephson effect . the physical picture that emerges out of the solution
is that the supercurrent flows as a result of multiple andreev reflections @xcite that transform an electron into a hole or vice versa .
the bound states represent standing waves of electron and hole waves going back and forth between two superconductors .
this view might look slightly different from a traditional picture of the dc josephson effect explained in textbooks , according to which the supercurrent , calculated in lowest order in the tunneling hamiltonian , flows as a result of tunneling of cooper pairs .
in fact , these views are simply two sides of a same coin .
= 1.5 in the importance of the bound states caused by the multiple andreev reflections was first recognized in the early 1970s in the studies of the dc josephson effect in long superconductor normal metal
superconductor junctions @xcite . until the 1990s
, however , it was not fully appreciated that the presence of current - carrying bound states is a general feature in any symmetric josephson junctions including weak links , superconducting quantum point contacts , and conventional tunnel junctions .
the bound states in a short junction or a point contact , whose energy depends on the phase difference of the superconductors , were explicitly constructed by solving the bogoliubov - de gennes equation in @xcite .
it is a curious fact that almost no attention was paid to the bound states in earlier studies of short superconducting weak links @xcite , in which the bound states show up silently as poles of the green s function on the real axis .
in fact , the andreev bound states play a central role for various situations in mesoscopic superconducting systems .
in this paper we will concentrate on the dc josephson effect of @xmath0-wave superconductors . when a finite voltage is applied across the junction ,
the loop of the multiple andreev scattering process is no longer closed , and the solutions of the bogoliubov - de gennes equation contain infinitely many terms each representing andreev - reflected electron wave or hole wave @xcite . for
sufficiently small applied voltage , however , the current - voltage characteristics can be described in terms of the quasi - stationary bound states whose energy is changing slowly with time @xcite .
we will not discuss further the ac josephson effect and the current - voltage characteristics of superconducting quantum point contacts in this paper .
this paper is organized as follows . in sec .
ii we introduce our basic equation , the bogoliubov - de gennes equation , and a continuity equation for electric charge .
we solve a simplified model of a josephson junction in sec .
iii to illustrate the basic idea of the supercurrent carried by andreev bound states .
we point out that a midgap surface state in a @xmath1-wave superconductor is essentially the same as the bound state in a short josephson junction .
we then explain a scattering approach to calculate the dc josephson current in sec . iv .
we begin with the bogoliubov - de gennes equation@xcite @xmath2 where @xmath3 represents both confining potential and scattering potential at interfaces of a superconductor and a normal metal , and @xmath4 is the chemical potential .
this is the basic equation for mesoscopic superconducting systems .
the upper component @xmath5 corresponds to electrons and the lower one @xmath6 represents holes . in principle
the pair potential @xmath7 is determined from the selfconsistency condition or the gap equation , @xmath8 , where @xmath9 is a coupling constant of electron - electron attraction , and @xmath10 is an average in the equilibrium state .
the field operator @xmath11 of electrons with spin @xmath12 can be expanded with the solutions of the bogoliubov - de gennes equation : @xmath13 , \label{eq : expansion}\ ] ] where the creation and annihilation operators @xmath14 and @xmath15 satisfy @xmath16 and @xmath17 .
note that @xmath18 is an eigenfunction with energy @xmath19 if @xmath20 satisfies eq .
( [ eq : bdg ] ) .
that is why we sum over the positive energy states only and introduce the second term in eq .
( [ eq : expansion ] ) . following ref .
@xcite , we introduce the equation of continuity of electric charge , which are satisfied by the solutions of the bogoliubov - de gennes equation . we define @xmath21 , \label{eq : p_e}\\ \mbox{\boldmath$j$}_e&= & -\frac{e\hbar}{m}{\rm im}\left [ \psi^\dagger_\uparrow(\mbox{\boldmath$r$},t ) \frac{\partial}{\partial\mbox{\boldmath$r$ } } \psi_\uparrow(\mbox{\boldmath$r$},t ) + \psi^\dagger_\downarrow(\mbox{\boldmath$r$},t ) \frac{\partial}{\partial\mbox{\boldmath$r$ } } \psi_\downarrow(\mbox{\boldmath$r$},t ) \right ] .
\label{eq : j_e}\end{aligned}\ ] ] it is clear that @xmath22 and @xmath23 are electric charge density and electric current , respectively . the continuity equation for the electric charge reads @xmath24 , where the source term @xmath25 is given by @xmath26 $ ] .
when the gap equation is satisfied , @xmath27 vanishes , and we recover the ordinary continuity equation of the charge for the averaged quantities @xmath28 and @xmath29 , where @xmath10 is an average in equilibrium . when we calculate the josephson current in a model where the pair potential satisfies the gap equation only approximately , we have to take into account the contribution from @xmath25 @xcite .
the source term can be ignored when the current is calculated inside the normal region where @xmath30 .
we shall consider a simplest model of a josephson junction that shows the essential features of the josephson effect .
it is a one - dimensional model where the left and right superconductors have the pair potentials of the same magnitude but with different phases : @xmath31 we assume the normal region to be thin and consider an extreme case where its width is infinitesimal . in general we expect some scattering process to be present either inside the normal region or at the super normal interfaces .
we therefore introduce a scattering potential @xmath32 to describe such a scattering .
the model with ( [ eq : delta(x ) ] ) and ( [ eq : u(x ) ] ) is the simplest non - trivial model for the josephson effect .
the strength of the @xmath33-function potential is characterized by a dimensionless parameter @xmath34 , where @xmath35 is the fermi wave vector . when @xmath36 the model describes a short superconductor normal metal superconductor junction . in the opposite limit @xmath37
, the model can be regarded as a toy model of a superconductor insulator superconductor tunnel junction .
we are interested in a state in the energy gap , @xmath38 . in this energy
range a wave function has to decay exponentially for @xmath39 .
we can write the wave function as @xmath40 where @xmath41 the coefficients @xmath42 , @xmath43 , @xmath44 , and @xmath45 are determined from the matching condition at @xmath46 : @xmath47 the eigenenergy is determined from the condition that eq .
( [ eq : matching ] ) has a nontrivial solution . within the approximation @xmath48 we find@xcite @xmath49 the solution of eq .
( [ eq : matching ] ) with this energy is @xmath50,\quad b_b= i\sigma nz(1-iz),\cr c_b&= & n(1-iz)\left[\sqrt{\cos^2(\varphi/2)+z^2 } -\sigma\cos(\varphi/2)\right],\quad d_b= inz\sqrt{1+z^2},\end{aligned}\ ] ] where @xmath51 , @xmath52 is the normalization constant @xmath53},\ ] ] and @xmath54 for @xmath55 .
the wave function is normalized such that @xmath56dx=1 .
$ ] once the wave function of the positive energy state is known , the eigenstate with energy @xmath57 is obtained as @xmath58 the field operator is then expanded as @xmath59 let us assume for a while that the states above or below the energy gap ( @xmath60 ) do not contribute to the supercurrent .
we will see later that this is indeed the case for the model we are dealing with .
we can calculate the supercurrent as the average of @xmath23 at @xmath61 in the equilibrium state at temperature @xmath62 : @xmath63\cr & & \cr & = & -\frac{e\delta_0\sin\varphi}{2\hbar(1+z^2 ) } \sqrt{\frac{1+z^2}{\cos^2(\varphi/2)+z^2 } } \tanh\!\left(\frac{e_b}{2k_bt}\right ) , \label{eq : i_b}\end{aligned}\ ] ] where @xmath64^{-1}$ ] . since the transmission probability through the @xmath33-function barrier eq .
( [ eq : u(x ) ] ) is @xmath65 , the normal - state conductance of this model is @xmath66 $ ] .
we can thus rewrite eqs .
( [ eq : e_b ] ) and ( [ eq : i_b ] ) as @xmath67 when @xmath68 eq .
( [ eq : i_b-2 ] ) reduces to @xmath69 which is the well - known ambegaokar - baratoff formula @xcite . in the opposite limit @xmath70 , eq .
( [ eq : i_b-2 ] ) becomes @xmath71 which coincides with the result obtained by kulik and omelyanchuk@xcite for a point contact josephson junction . in this limit
the zero - temperature josephson current becomes maximum at @xmath72 , not at @xmath73 like in the tunnel junctions .
the bound state energy is simply given by @xmath74 , which equals the fermi energy @xmath75 at @xmath72 .
equation ( [ eq : i_b-2 ] ) thus interpolates between eq .
( [ eq : ab ] ) of tunnel junctions with very small transmission probability and eq . ( [ eq : ko ] ) of point contacts with perfect transmission
. the same expression of the supercurrent interpolating the two limits has been found by many people @xcite .
the phase - dependent bound state energy ( [ eq : e_b ] ) and its implication to the josephson effect in short junctions were emphasized in @xcite .
the results were then generalized to the multichannnel contacts in @xcite . equation ( [ eq : e_b ] ) shows that there is always a bound state ( andreev level ) with @xmath76 unless the phase difference @xmath77 is zero .
this is true even for the tunnel junctions in which @xmath68 and @xmath78 . in the above model two bound states ( @xmath79 )
appear in the energy gap @xmath80 but disappear when @xmath81 .
this is a general property of short @xmath0-wave junctions where @xmath82 with @xmath83 being the length of the normal region ( @xmath84 in our model ) . in the opposite regime
where @xmath85 , many ( @xmath86 ) andreev levels exist in the gap for any @xmath77 @xcite .
these andreev levels are easily understood as discrete levels of normal electrons confined in a quantum well of @xmath87 pair potential .
they are similar to the discrete energy levels in a vortex in an @xmath0-wave type ii superconductor @xcite .
a nontrivial observation we made in the above calculation is that , when @xmath88 , two andreev levels survive even in the limit @xmath89 , in sharp contrast with a case where a particle is confined in a quantum well of diagonal potential barrier , in which no bound state exists in the limit @xmath89 .
= 1.5 in the bound state @xmath90 is formed as a result of resonant multiple andreev reflection process .
equation ( [ eq : i_b ] ) shows that this state carries nonvanishing supercurrent .
this can be physically understood as follows ( fig .
[ fig : andreev ] ) .
each time an electron is andreev reflected into a hole , a cooper pair is effectively generated .
therefore the bound state , which represents an infinite loop of andreev reflections ( electron@xmath91hole@xmath91electron@xmath92 ) , serves as a pump that transfers cooper pairs from one superconductor to the other .
when @xmath93 the bound state energy has a strong dependence on the phase difference @xmath77 , which can be controlled externally by the current source .
it would be very interesting if these current - carrying andreev levels are observed experimentally using , for example , the scanning tunneling spectroscopy .
finally we comment on midgap states that appear near the surface of a @xmath1-wave superconductor when the surface is not parallel to a crystal axis ( fig .
[ fig : dwave ] ) @xcite .
in fact the mechanism of the formation of the zero - energy midgap state is the same as the andreev bound states we discussed above . as shown in fig .
[ fig : dwave ] , an electron traveling towards a surface is reflected back into the @xmath1-wave superconductor and is subsequently andreev - reflected into a hole by the _ positive _ pair potential . in the next step the hole follows the same path backwards , reflected at the surface , and finally andreev - reflected into another electron by the _ negative _ pair potential .
the analogy to the josephson junction is now obvious .
the surface of the @xmath1-wave superconductor plays a role of the point contact with @xmath70 , and the sign change in the pair potential corresponds to the phase difference @xmath72 .
equation ( [ eq : e_b-2 ] ) tells that there must be bound states with @xmath75 localized near the surface , which are the midgap states found in ref .
they give rise to a peak in differential conductance of a normal insulator@xmath1-wave tunnel junction @xcite .
the andreev levels also play an important role in a josephson junction made of an @xmath0-wave superconductor and a @xmath1-wave superconductor @xcite .
it is known that in such a junction there is a chance that andreev levels appear even in the lowest - energy state , thereby leading to a ground state with broken time - reversal symmetry @xcite .
we shall introduce a general expression of the dc josephson current under the assumption that we can ignore the spatial variation of the pair potential . here
we assume @xmath94 in the normal region ( @xmath95 ) the order parameter vanishes .
we also assume that the potential @xmath3 is independent of @xmath96 in the superconductors .
we emphasize that this step - function pair potential ( [ eq : delta(r ) ] ) is a good approximation for mesoscopic josephson junctions of our concern , where bulk superconductors are coupled via a few conduction channels .
the use of the step - function model is justified by the following reasons .
first , the presence of a point contact does not seriously affect the order parameter of bulk superconductors , because the width of the constrictions is much smaller than the superconducting coherence length @xmath97 , which is the characteristic length scale for the variation of the order parameter .
second , if the normal region is made of a material different from superconducting electrodes , the contact resistance of the sn interfaces is not small due to the fermi - velocity difference and/or due to the presence of a schottky barrier . in this case also the pair potential is not reduced significantly near the point contact . more generally , when the resistance of the junction is much larger than those of superconductors , we may safely use the step - function model @xcite .
since the dc josephson current is an equilibrium current , it is given by the thermal average of @xmath98 .
we calculate it at @xmath46 to avoid complication due to the source term @xmath25 : @xmath99 \biggr|_{x=0 } , \label{eq : i = g_omega_n}\ ] ] where @xmath100 is the temperature green s function at frequency @xmath101 in the nambu space ( @xmath102 matrix ) @xcite .
our assumption that the order parameter @xmath7 and the potential @xmath3 have no dependence on @xmath96 in the superconductors , is strong enough that we can immediately write down @xmath100 at @xmath103 in terms of scattering amplitudes .
its explicit expression can be found in ref . @xcite . after substituting it into eq .
( [ eq : i = g_omega_n ] ) , we obtain @xmath104 , \label{eq : i = a - a}\ ] ] where @xmath105 and we have used the andreev approximation where the velocities of quasiparticles are set equal to the fermi velocity . the quantity @xmath106 is a scattering amplitude for the process in which an electron - like quasiparticle of @xmath107th channel traveling from the left to the junction is reflected back as a hole - like quasiparticle of @xmath108th channel @xcite .
equation ( [ eq : i = a - a ] ) expresses the supercurrent in terms of the scattering data @xmath109 , just like the landauer formula does the normal - state conductance .
an important difference is that the supercurrent is proportional to the scattering amplitude , not to the probability .
equation ( [ eq : i = a - a ] ) is clearly an odd function of @xmath77 , and can be expanded formally as @xmath110 .
the term proportional to @xmath111 corresponds to @xmath112th order term in the tunneling hamiltonian approach that is a contribution from the simultaneous tunneling of @xmath112 cooper pairs .
equation ( [ eq : i = a - a ] ) is thus a useful , nonperturbative formula for the dc josephson current @xcite .
let us apply it to the simple one - dimensional model we analyzed in the last section . solving the bogoliubov - de gennes equation for @xmath113 with potentials given by ( [ eq : delta(x ) ] ) and ( [ eq : u(x ) ] ) gives @xmath114 } { e^2(1+z^2)-\delta^2_0[\cos^2(\varphi/2)+z^2]}. \label{eq : a(varphi , e)}\ ] ] note that the scattering amplitude has poles at @xmath79 . substituting ( [ eq : a(varphi , e ) ] ) with @xmath115 into ( [ eq : i = a - a ] ) , we then find @xmath116 } , \label{eq : i = sum_omegan}\ ] ] which agrees with eq .
( [ eq : i_b ] ) after the frequency sum is evaluated by reducing into a contour integral . in this calculation
all the contributions to the integral come from the residues at @xmath79 .
this proves that the andreev bound states carry the whole supercurrent in short symmetric josephson junctions .
we note that this is not true for long weak links where @xmath85 @xcite and for asymmetric junctions where @xmath117 .
let us examine the latter case in more detail .
we consider a one - dimensional model where the pair potential is given by eq .
( [ eq : delta(r ) ] ) with @xmath84 and the diagonal potential is ( [ eq : u(x ) ] ) .
a solution of the bogoliubov - de gennes equation for this model gives a scattering amplitude for @xmath118 , @xmath119 where @xmath120 and @xmath121 . from eq .
( [ eq : i = a - a ] ) we obtain @xmath122 where @xmath123 .
this result agrees with the ambegaokar - baratoff formula @xcite for a tunnel junction in the limit @xmath124 and with the kulik - omelyanchuk formula @xcite for a point contact with perfect transmission in the limit @xmath125 .
the summation can be reduced to an integral over @xmath126 plus contributions from residues at energies of bound states at @xmath127 .
the integral over @xmath128 corresponds to the process shown in fig .
[ fig : asymmetry ] and carry most of the supercurrent .
in a superconducting quantum point contact , the critical current or the maximum value of the dc josephson current can take a quantized value in units of @xmath129 .
this was first predicted theoretically @xcite and confirmed by experiments @xcite .
the effect can be understood easily by the following argument .
the critical current of a josephson junction is generally proportional to its normal - state conductance , as seen in eq .
( [ eq : i_b-2 ] ) .
it is well known that the conductance of a quantum point contact fabricated in a 2deg is quantized in units of @xmath130 @xcite : the conductance is simply @xmath130 times the number of open channels at the constriction .
we can thus naively expect that the critical current should be quantized as well . for a more quantitative discussion we can use the adiabatic approximation where
scattering between channels is neglected .
the transmission probability through the quantum point contact is then @xmath131 for open channels and @xmath132 for closed channels . for a short symmetric superconducting quantum point contact where @xmath82
, we obtain , from eqs .
( [ eq : i_b-2 ] ) and ( [ eq : i = a - a ] ) , @xmath133 where @xmath52 is the number of open channels at the point contact .
we thus find that the critical current is a multiple of @xmath129 at zero temperature .
the number of open channels is determined by the width of the constriction , which can be controlled externally , for example , by changing a gate voltage for a 2deg .
this was done in an experiment by takayanagi _
_ @xcite . equation ( [ eq : nko ] ) is valid only for short junctions . in the opposite case where @xmath85 , the unit of the quantization is different .
the supercurrent for a long superconducting quantum point contact is given by @xmath134 for @xmath135 at zero temperature , where @xmath136 is the longitudinal component of the fermi velocity in @xmath107th channel .
in this case the critical current is not exactly quantized in contrast with the short - junction limit .
this is nothing but a quantum limit of long sns junctions obtained in early 1970s @xcite .
in this paper we have reviewed the dc josephson effect in mesoscopic josephson junctions .
our discussion is based on a simple analysis of the bogoliubov - de gennes equation . in short junctions including superconducting quantum point contacts the whole supercurrent is carried by andreev bound states in the energy gap that appear as a result of multiple andreev reflections .
the wave function of a bound state is constructed explicitly for a simple @xmath33-function model by solving the bogoliubov - de gennes equation .
the bound state energies are sensitive to the difference of the phases of superconductors , and this sensitivity is the origin of the josephson effect .
our simple physical view of the josephson effect in terms of the current - carrying bound states is very general and applicable to many situations .
permanent address .
furusaki , h. takayanagi and m. tsukada , phys .
lett . * 67 * , 132 ( 1991 ) ; phys .
b * 45 * , 10563 ( 1992 ) .
beenakker and h. van houten , phys .
lett . * 66 * , 3056 ( 1991 ) .
h. takayanagi , j.b .
hansen and j. nitta , phys .
lett . * 74 * , 166 ( 1995 ) ; h. takayanagi , t. akazaki and j. nitta , phys .
* 75 * , 3533 ( 1995 ) .
muller , j.m .
van ruitenbeek , and l.j .
de jongh , phys .
rev.lett . * 69 * , 140 ( 1992 ) .
andreev , sov .
jetp * 19 * , 1228 ( 1964 ) .
kulik , sov .
jetp * 30 * , 944 ( 1970 ) . c. ishii , prog .
phys . * 44 * , 1525 ( 1970 ) ; * 47 * , 1464 ( 1972 ) .
j. bardeen and j.l .
johnson , phys .
b * 5 * , 72 ( 1972 ) .
svidzinsky , t.n .
antsygina and e.n .
bratus , j. low temp .
phys . , 131 ( 1973 ) .
a. furusaki and m. tsukada , physica b*165&166 * , 967 ( 1990 ) ; a. furusaki and m. tsukada , phys . rev .
b * 43 * , 10164 ( 1991 ) .
i.o . kulik and a.n .
omelyanchuk , sov .
j. low temp .
phys . * 4 * , 142 ( 1978 ) .
zaitsev , sov .
jetp * 59 * , 1015 ( 1984 ) .
arnold , j. low temp .
phys . * 59 * , 143 ( 1985 ) .
klapwijk , g.e .
blonder , and m. tinkham , physica b&c * 110 * , 1657 ( 1982 ) .
artemenko , a.f .
volkov and a.v .
zaitsev , sov .
jetp * 49 * , 924 ( 1979 ) .
k. flensberg , j.b . hansen , and m. octavio , phys .
b * 38 * , 8707 ( 1988 ) .
r. kmmel , u. gunsenheimer and p. nicolsky , phys . rev . * b * 42 , 3992 ( 1990 ) . u. gunsenheimer and a.d .
zaikin , phys .
b * 50 * , 6317 ( 1995 ) .
bratus , v.s .
shumeiko and g. wendin , phys . rev .
lett . * 74 * , 2110 ( 1995 ) .
d. averin and a. bardas , phys .
75 * , 1831 ( 1995 ) ; phys .
b * 53 * , r1705 ( 1996 ) ; d. averin and h.t .
imam , phys .
* 76 * , 3814 ( 1996 ) .
m. hurd , s. datta , and p.f .
bagwell , phys .
b * 54 * , 6557 ( 1996 ) .
de gennes , _ superconductivity of metals and alloys _
( benjamin , new york , 1966 ) . g.e .
blonder , m. tinkham and t.m .
klapwijk , phys .
b * 25 * , 4515 ( 1982 ) .
a. furusaki and m. tsukada , solid state commun .
* 78 * , 299 ( 1991 ) .
v. ambegaokar and a. baratoff , phys .
* 10 * , 486 ( 1963 ) ; * 11 * , 104(e ) ( 1964 ) . w. haberkorn , h. knauer , and j. richter , phys .
( a ) * 47 * , k161 ( 1978 ) .
glazman and k.a .
matveev , jetp lett . * 49 * , 659 ( 1989 ) .
beenakker , phys .
lett . * 67 * , 3836 ( 1991 ) ; * 68 * , 1442(e ) ( 1992 ) . c. caroli , p.g .
de gennes , and j. matricon , phys .
* 6 * , 307 ( 1964 ) .
hu , phys .
lett . * 72 * , 1526 ( 1994 ) .
y. tanaka and s. kashiwaya , phys .
, 3451 ( 1995 ) .
s. yip , j. low temp .
* 91 * , 203 ( 1993 ) ; y. tanaka , phys .
lett . * 72 * , 3871 ( 1994 ) ; a. m. zagoskin , j. phys . : condens .
matter * 9 * , l419 ( 1997 ) .
a. m. zagoskin , _ quantum theory of many - body systems : techniques and applications _
( springer - verlag , new york , 1998 ) .
see , for example , m. sigrist , prog .
theor . phys .
* 99 * , 899 ( 1998 ) and references therein .
likharev , rev .
phys . * 51 * , 101 ( 1979 ) .
abrikosov , l.p .
gorkov and i.e. dzyaloshinski , _ methods of quantum field theory in statistical physics _
( dover , new york , 1975 ) . a complementary method to compute the dc josephson current in lattice models is found in a. furusaki , physica b*203 * , 214 ( 1994 ) .
there is a small but important typo in this paper : the left hand side of eq .
( 10 ) should read @xmath137 .
see also y. takane , j. phys .
jpn . * 63 * , 2668 ( 1994 ) ; y. koyama , y. takane and h. ebisawa , j. phys .
* 66 * , 430 ( 1997 ) ; * 66 * , 3314(e ) ( 1997 ) .
van wees , h. van houten , c.w.j .
beenakker , j.g .
williamson , l.p .
kouwenhoven , d. van der marel and c.t .
foxon , phys .
60 * , 848 ( 1988 ) .
wharam , t.j .
thornton , r. newbury , m. pepper , h. ahmed , j.e.f .
frost , d.g .
hasko , d.c . peacock , d.a . ritchie and g.a.c .
jones , j. phys .
c*21 * , l209 ( 1988 ) . | the dc josephson effect in a superconducting quantum point contact , where supercurrent flows through a small number of channels , is reviewed . the central role of andreev levels is emphasized which carry the whole supercurrent in short symmetric josephson junctions including tunnel junctions .
a simple intuitive view of the dc josephson effect in a quantum point contact is given in terms of multiple andreev reflections .
the quantization of the critical current in superconducting quantum point contacts is briefly discussed . |
the brauer algebra @xmath1 was introduced in 1937 @xcite .
it is an enlargement of the symmetric group @xmath2 , designed to impose an orthosymplectic algebraic condition , depending on @xmath3 , on a schur - weyl dual algebraic group over the complex field
. however it may be defined over an arbitrary ring @xmath4 , for any @xmath5 and @xmath6 .
it has integral representations ( in the sense of @xcite ) that pass to simple modules over suitable splitting fields ( sometimes called cell modules in modern parlance ) , constructed by brown @xcite .
this raises the problem of determining simple decomposition matrices for these key modules ( and hence for indecomposable projective modules ) over other extensions @xmath7 ( @xmath7 a field equipped with the property of @xmath4-algebra ) .
this long - standing problem remains open .
the aim of the present work is twofold : to develop some tools to solve this problem , by constructing a formal ` weight space ' with a geometry and associated functors on the module categories , and to propose a possible combinatorial framework ( at least over @xmath8 ) in which the answer might be couched .
the problem can be addressed in two parts : first working over @xmath8 , and then over fields of prime characteristic .
( the latter can be anticipated to be significantly harder , as the representation theory of @xmath1 _ contains _ the representation theory of the symmetric group @xmath2 . ) a significant step towards an answer came with the determinantion in 2005 of the blocks of the algebra over @xmath8 ( and in 2006 of a geometric linkage principle in any characteristic different from @xmath9 ) @xcite .
these results were obtained by using functors that allow the algebras for all @xmath10 to be treated together ( as previously used in @xcite ) .
( an alternative approach to the characteristic @xmath0 result via characteristic @xmath11 has recently been developed by donkin and tange @xcite . ) the key observation that underpins the various geometric considerations in this paper is that the cell modules of @xmath1 may be indexed by certain orbits of lattice points in the euclidean space @xmath12 .
the orbits are those of a reflection group @xmath13 , where @xmath13 is the limit of the usual type-@xmath13 reflection group action on euclidean @xmath14-space @xmath15 .
the @xmath13-action is a parabolic in the limit type-@xmath16 reflection group action on @xmath12 ( as in finite rank ) , and the orbits of the @xmath16-action on coset space @xmath17 describe the blocks of @xmath1-mod over @xmath8 .
for this reason we will work over @xmath8 in this paper . using this parabolic / reflection group formulation we are able to : 1
. determine a translation principle ( morita equivalences between certain blocks ) ; 2 .
compute appropriate brauer analogues of the parabolic kazhdan - lusztig polynomials that determine decomposition matrices in lie ( quantum group ) theory ; 3 .
use these to encode the structure of the algebra in many special cases ( for example in low rank , with the obvious conjecture that this extends to all cases ) .
our methodology , and the structure of the paper , can be summarised as follows .
the well - established root system / weyl group analysis of high - weight theory reduces many questions in lie representation theory ( of algebraic and quantum groups ) to geometry and combinatorics @xcite , once the weyl group and affine weyl group action on weight space has been determined . of
course these weyl groups are reflection groups , one a parabolic in the other @xcite , facilitating , for example , an alcove geometric description of blocks .
note the obvious analogy with the role of reflection groups described above .
it was this which motivated our formulation of the results in @xcite ( guided by success with a similar approach to other ` diagram ' algebras @xcite ) .
in lie theory the euclidean space is finite and the reflection group is infinite by virtue of being affine ; here it is by virtue of unbounded rank .
nonetheless , all the geometric and combinatorial machinery goes through unchanged .
the development of this analogy in section [ weyl ] lies at the heart of our methodology .
arguably one of the most beautiful machines that exists for computing decomposition matrices in any setting is the method of ( parabolic ) kazhdan - lusztig polynomials in lie theory @xcite .
not all of the assumptions of this set - up hold for the brauer algebra , but in section [ klp ] we show how to bring the two theories close enough together that parabolic kazhdan - lusztig ( pkl ) polynomials suitable for the brauer algebra may be computed . in the group ( or quantum group ) case one has pkl polynomials associated to alcoves in the alcove geometry , determining ( at least for the @xmath18-group over @xmath8 , and its ringel dual hecke algebra quotient @xcite ) decomposition matrices in alcove blocks . in general
there is more than one block intersecting an alcove , but there is also a translation principle @xcite , which states that all these blocks are morita equivalent , and hence do indeed have the same decomposition matrices . in our case
the pkl method formally assigns the same decomposition matrix to every alcove block .
one is therefore led to seek a form of translation principle .
the brauer algebras @xmath1 as @xmath10 varies form a tower via an idempotent construction .
in @xcite we gave a general axiom scheme for studying such a tower as a _ tower of recollement_. the advantage of studying algebras in such a tower is the existence of four functors : induction , restriction , globalisation , and localisation , which relate the representation theories of the different algebras in a compatible manner . in section [ towermor ] we will show how towers of recollement , when combined with a suitable description of the blocks in the tower , give rise to analogues of translation functors and corresponding morita equivalences . these functors are defined using induction or restriction functors followed by projection onto a block , and are similar in spirit to @xmath19-induction and @xmath19-restriction functors for the symmetric group @xcite .
we apply this translation theory to the brauer algebras in section [ brauertran ] , with the aim of proving that two blocks corresponding to weights in the same facet have the same representation theory ( corollary [ sumup ] ) .
however , in order to do this we will need some additional functors , generalisations of induction and restriction , which are introduced in section [ genir ] .
we will also see that when @xmath20 there are translation equivalences _ between _ certain facets , which raises interesting questions as to the true geometric structure underlying the representation theory .
we can also consider an analogue of translation onto a wall in lie theory for towers of recollement . using this we show that the decomposition matrix for the brauer algebra is determined by the decomposition matrix for the block containing the weight @xmath0 ( theorem [ need0 ] ) .
in section [ manygraphs ] we consider various graphs associated to each block ( or each facet ) , and show that they are in fact all isomorphic
. for alcove graphs we can define associated kazhdan - lusztig polynomials ; using the graph isomorphisms these polynomials can more generally be associated to any block graph . in the final section
we show that when @xmath21 these polynomials correctly predict decomposition numbers and filtrations in the alcove case for standard modules in low rank examples .
in this section we will very briefly summarise the basic representation theory of the brauer algebras that will be needed in what follows .
details can be found in @xcite . in this paper
, we will restrict our attention to the case where the ground field is @xmath8 .
the brauer algebra @xmath1 is a finite dimensional algebra with parameter @xmath22 .
when @xmath23 this algebra is semisimple , so we will henceforth assume that @xmath3 is an integer .
_ we will also assume that @xmath24 .
_ it will be convenient to use the usual graphical presentation of brauer algebras .
an _ @xmath25 brauer algebra diagram _ will consists of a rectangular frame with @xmath10 marked points on the northern edge and @xmath26 on the southern edge called _ nodes_. each of these sets will be numbered from @xmath27 to @xmath10 ( respectively @xmath26 ) from left to right .
each node is joined to precisely one other by a line ; lines connecting the northern and southern edge will be called _ propagating lines _ and the remainder _ ( northern or southern ) arcs_. multiplication of two @xmath28 diagrams is by concatenation , where any diagram obtained with a closed loop is set equal to @xmath3 times the same diagram with the loop removed .
two diagrams are equivalent if they connect the same pairs of nodes .
the algebra obtained by taking linear combinations of @xmath28 diagrams is a realisation of @xmath29 .
note that @xmath30 is isomorphic to the subalgebra of @xmath31 spanned by diagram with only propagating lines .
moreover , @xmath29 is generated by this subalgebra together with the elements @xmath32 with @xmath33 consisting of @xmath34 propagating lines and arcs joining @xmath35 and @xmath36 on the northern ( respectively ) southern edges .
the brauer algebra can also be constructed via iterated inflations of the symmetric group @xcite , and thus is a cellular algebra . if @xmath24 , then it is even quasihereditary .
the standard modules @xmath37 are parameterised by partitions of @xmath38 ( where the final term depends on the parity of @xmath10 ) , and we will denote the set of such by @xmath39 . if @xmath24 then the same set parameterises the simple modules .
we have the following explicit construction of standard modules .
consider a brauer diagram with @xmath10 northern nodes and @xmath40 southern nodes , and with no southern arcs .
such a diagram must have exactly @xmath41 northern arcs .
we will denote this diagram by @xmath42 , where @xmath43 denotes the configuration of northern arcs , @xmath27 represents the fixed southern boundary , and @xmath44 is the permutation obtained by setting @xmath45 if the @xmath35th propagating northern node from the left is connected to the southern node labelled by @xmath36 .
the elements @xmath43 arising as above will be called _ partial one - row diagrams _ , and the set of such will be denoted by @xmath46 .
if a node @xmath35 in @xmath47 is not part of a northern arc we say that it is _
free_. the vector space spanned by the set of diagrams of the form @xmath48 where @xmath47 will be denoted @xmath49 .
note that @xmath50 acts on @xmath49 on the right by permuting the southern nodes .
given @xmath51 a partition of @xmath52 , let @xmath53 denote the specht module corresponding to @xmath51 for @xmath50
. then the standard module @xmath37 can be realised ( see ( * ? ? ?
* section 2 ) or ( * ? ? ?
* section 2 ) ) in the following manner . as a vector space
we have @xmath54 an element @xmath55 of @xmath29 acts on @xmath56 from the right by diagram multiplication .
if the resulting product has fewer than @xmath26 propagating lines then we define the action of @xmath55 on @xmath57 to be @xmath0 .
otherwise the product will result in a diagram with exactly @xmath26 propagating lines , but these may now be permuted .
we transfer this permutation ( thought of as an element of @xmath58 ) through the tensor product to act on @xmath53 . for @xmath24 ( or @xmath59
there is an idempotent @xmath60 such that @xmath61 , and so there are associated localisation and globalisation functors @xmath62 and @xmath63 . in this way we can regard @xmath64 as a full subcategory of @xmath65 , and hence @xmath66 .
we set @xmath67 , the set of all partitions .
we will abuse terminology and say that two labels are in the same block when the associated standard modules are in the same block . in order to describe the main results in @xcite
we will need some additional terminology .
recall that for a partition @xmath51 ( which we will identify with its young diagram ) , the _ content _ of the box in row @xmath35 and column @xmath36 of the diagram is defined to be @xmath68 .
a pair of partitions @xmath69 is said to be _
@xmath3-balanced _ if the following conditions are satisfied : 1 .
the boxes in the skew partition can be paired so that the contents of each pair sum to @xmath70 ; 2 .
if the skew partition contains boxes labelled by @xmath71 and there is only one such box in the bottom row then the number of pairs of such boxes is even .
we say that two general partitions @xmath51 and @xmath72 are _
@xmath3-balanced _ if the pairs @xmath73 and @xmath74 are both @xmath3-balanced .
the importance of the @xmath3-balanced condition is clear from the following result ( * ? ? ?
* corollary 6.7 ) : [ balver ] two partitions @xmath51 and @xmath72 are in the same block for @xmath1 if and only if they are @xmath3-balanced .
denote by @xmath75 the set of partitions @xmath72 such that @xmath72 and @xmath51 are @xmath3-balanced .
note that if @xmath76 then so too are @xmath77 and @xmath78 . thus @xmath75 forms a lattice under the inclusion relation .
we say that @xmath72 is a _ maximal balanced subpartition of _
@xmath51 if @xmath79 and there does not exist @xmath80 with @xmath81
. one of the main steps in the proof of theorem [ balver ] is ( * ? ? ?
* theorem 6.5 ) , which shows that if @xmath72 is a maximal balanced subpartition of @xmath51 then @xmath82 the standard module @xmath37 has simple head @xmath83 and all other composition factors are of the form @xmath84 where @xmath85 lies in the same block as @xmath51 ( * ? ? ?
* proposition 4.5 ) .
if @xmath51 and @xmath72 are such a pair with @xmath86 then @xmath87=\dim\operatorname{hom}(\delta_n(\lambda),\delta_n(\mu))=1\ ] ] by ( * ?
* theorem 3.4 and the remarks after theorem 3.1 ) ( see ( * ? ? ?
* theorem 4.4 ) ) . if @xmath88 are two weights in the same block and @xmath89 for some @xmath90 and @xmath55 with @xmath90 even then we also have ( * ? ? ?
* proposition 5.1 ) that @xmath91=1.\ ] ] in general , if @xmath92 and @xmath93 with @xmath94 the exactness of the localisation functor implies that @xmath95=[\delta_n(\mu):l_n(\lambda)]\ ] ] for all @xmath96 .
the algebra @xmath1 embeds inside @xmath97 , and so we may consider the associated induction and restriction functors @xmath98 and @xmath99 . if @xmath51 and @xmath72 are partitions we write @xmath100 , or @xmath101 if the young diagram for @xmath72 is obtained from that for @xmath51 by removing one box . then
* theorem 4.1 and corollary 6.4 ) we have short exact sequences @xmath102 and @xmath103 the restriction rule for simples is not so straightforward .
however we do have by ( * ? ? ?
* lemma 7.1 ) that if @xmath72 is a partition obtained from @xmath51 by removing one box then @xmath104\neq 0.\ ] ] the next two results are new , and show how the local data in ( [ indrule ] ) and ( [ resrule ] ) can be applied to explicit decomposition number calculations ( which illustrates one of the motivations for the tower of recollement formalism in @xcite ) .
[ projelim ] suppose that by removing @xmath26 boxes from @xmath92 it is possible to reach a partition @xmath105 such that @xmath106 is a projective @xmath107-module .
then the simple module @xmath83 does not appear as a composition factor in any @xmath108 with @xmath109 of degree less than @xmath110 .
suppose that @xmath83 does occur as a composition factor of @xmath111 . then by brauer - humphreys reciprocity (
* proposition a2.2(iv ) ) the projective cover @xmath112 of @xmath83 has a standard module filtration with @xmath111 as a factor . as @xmath113 is projective , so is @xmath114 . by repeated application of ( [ indrule ] )
we see that this contains @xmath83 in its head , and so must have as a summand @xmath112 .
we have that @xmath115 and @xmath116 therefore by repeated application of ( [ resrule ] ) to @xmath117 we see that @xmath112 can not have a standard module filtration with @xmath111 as a factor .
this gives the desired contradiction and so we are done .
note that any standard module which is alone in its block must be projective .
thus there are many circumstances where proposition [ projelim ] will be easy to apply .
indeed , this case will be sufficient for our purposes .
if @xmath69 are two partitions then their skew @xmath118 can be regarded as a series of disjoint partitions ; when considering such differences we will list the various partitions in order from top right to bottom left . thus a skew partition @xmath119 will consist of two disjoint partitions of the form @xmath120 .
[ largerhom ] if @xmath69 is a balanced pair with @xmath121 or @xmath122 then @xmath82 we may assume that @xmath92 by localisation . if @xmath121 let @xmath123 be @xmath51 less one of the two removable boxes in @xmath118 , and @xmath124 be the partition @xmath72 together with the addable box from the other component of the skew .
if @xmath122 then let @xmath123 be @xmath51 with any one of the boxes in @xmath118 removed , and @xmath124 be @xmath72 together with the unique box in @xmath118 making this a balanced pair . in each case
@xmath124 is a maximal balanced subpartition of @xmath123 and so by ( * ? ? ?
* theorem 6.5 ) we have that @xmath125 by ( * ? ? ?
* corollary 6.7 ) and ( [ resrule ] ) the only term in the block labelled by @xmath123 in the standard filtration of @xmath126 is @xmath127 .
therefore by frobenius reciprocity we have @xmath128 now @xmath51 is not the only weight in its block in the set of weights labelling term in the standard filtration of @xmath129 .
however , by ( * ? ? ?
* lemma 4.10 ) it follows from ( [ frex ] ) that @xmath130 as required .
in @xcite we identified partitions labelling brauer algebra modules with elements of @xmath131 ( for suitable @xmath14 ) only after transposition of the original partition to form its conjugate .
henceforth when we regard @xmath39 ( or @xmath132 ) as a subset of @xmath133 it will always be via this _ transpose map _ @xmath134 .
we wish to identify reflection groups associated to the brauer algebra which play the role of the weyl and affine weyl groups for reductive algebraic groups .
first let us recall the properties of weyl groups which we wish to replicate . in lie theory
a weyl group @xmath135 is a reflection group acting on a euclidean weight space with the following properties : 1 .
there is an integral set of weights on which @xmath135 acts via a dot action @xmath136 , 2 .
the reflection hyperplanes of @xmath135 under this action break space up into chambers ( and other facets ) , 3 .
a complete set of weights indexing simple ( or standard ) modules coincides with the weights in a single chamber under the dot action , namely that containing the zero weight .
such weights are said to be _
dominant_. thus the selection of an indexing set for the dominant weights is taken care of by the weyl group ( and its dot action ) . in positive characteristic @xmath11 or at a quantum @xmath137th root of unity
there is then a second stage , the introduction of an affine extension of @xmath135 ( with action depending on @xmath11 or @xmath137 ) , which has orbits whose intersection with the dominant weights determine the blocks .
this affine extension defines an additional set of reflecting hyperplanes , which break the set of weights up into a series of chambers ( now called alcoves ) and other facets .
we refer to this configuration of facets , together with the action of the affine extension , as the _ alcove geometry _ associated to the particular lie theory in question . the alcove geometry controls much of the representation theory of the corresponding reductive group . in particular
, we typically have a _ translation principle _ which says that there are morita equivalences between blocks which intersect a given facet , and so much of the representation theory does not depend on the weight itself but only on the facet in which it lies .
we will show how a version of the above programme can be implemented for the brauer algebra from scratch .
let @xmath138 be the @xmath139-vector space with basis @xmath140 , and @xmath141 .
we will define various reflections on @xmath138 corresponding to the standard action of the type @xmath16 weyl group .
let @xmath142 be the reflection in the hyperplane in @xmath138 through the origin which takes @xmath143 to @xmath144 and fixes all other unit vectors , and @xmath145 to be the reflection in the hyperplane perpendicular to @xmath146 which takes @xmath143 to @xmath147 .
we define @xmath148 which is the type @xmath16 weyl group .
note that it has a a subgroup @xmath149 which is just the type @xmath13 weyl group ( isomorphic to @xmath2 ) .
as explained in the previous section , the brauer algebras @xmath1 as @xmath10 varies form a tower .
thus it is natural to consider all such algebras simultaneously . in order to do this
we will work with the infinite rank case .
note also that orbits of the finite weyl group @xmath150 are not sufficient to define an indexing set for the simple @xmath1-modules ( one needs to consider @xmath151-orbits , but then this group is not a subgroup of @xmath152 ) , unlike the infinite rank case .
let @xmath153 be the @xmath139-vector space consisting of ( possibly infinite ) linear combinations of the elements @xmath154 .
we say that @xmath155 has _ finite support _ if only finitely many components of @xmath51 are non - zero , and write @xmath156 for the set of such elements .
( we define @xmath157 similarly . )
thus , with the obvious embedding of @xmath158 inside @xmath159 , we have that @xmath160 .
( note that the usual component - wise inner product on @xmath133 is finite on @xmath156 , so that @xmath156 is just the integral part of the usual @xmath161 hilbert space in @xmath162 . )
we say that an element @xmath163 of @xmath164 is _ dominant _ if @xmath165 for all @xmath35 .
( note that any such element must lie in @xmath166 . )
embed @xmath138 inside @xmath153 in the obvious way , and let @xmath167 , and @xmath135 be the corresponding limits of @xmath152 , and @xmath150 .
clearly the space @xmath156 is closed under the action of @xmath167 .
we will call elements in @xmath156 _ weights_. dominant weights are precisely those which label standard modules for the brauer algebras , and by analogy with lie theory we will denote the set of such weights by @xmath168 .
general elements of @xmath153 will be called _ vectors_. we will use greek letters for weights and roman letters for general vectors .
given a reflection group @xmath169 ( or the corresponding set of hyperplanes @xmath170 ) , we say that a vector is _ regular in @xmath169 ( or in @xmath170 ) _ if it lies in the interior of a chamber , i.e. in some connected component of @xmath171 .
otherwise we say the vector is _
singular_. in the case @xmath172 we shall call chambers _ alcoves _ to emphasise the distinction between this and the @xmath135 case .
for @xmath173 we define the _ degree of singularity _ @xmath174 ( which need not be finite in general ) .
note that a vector @xmath43 is regular in @xmath167 if and only if @xmath175 . the next lemma is clear .
[ basicalcoves ] ( i ) there is a chamber @xmath176 of the action of @xmath135 on @xmath153 consisting of all strictly decreasing sequences .
+ ( ii ) the boundary of @xmath176 consists of all non - strictly decreasing sequences .
recall that in lie theory we typically consider a shifted reflection group action with respect to some fixed element @xmath177 .
it will be convenient to consider a similar adjustment here .
let @xmath178 and @xmath179 . for @xmath180 define @xmath181 for @xmath182 and @xmath173 let @xmath183 where the right - hand side is given by the usual reflection action of @xmath167 on @xmath153 .
note that @xmath156 is closed under this action of @xmath167 .
we say that a weight @xmath51 is _
@xmath3-regular _ if @xmath184 is regular , and define the _ degree of @xmath3-singularity _ of @xmath51 to be @xmath185 .
[ wis ] let @xmath186 .
+ ( i ) for @xmath187 the weight @xmath188 does not depend on @xmath3 . moreover , if @xmath189 and @xmath190 then @xmath191 .
+ ( ii ) if @xmath190 then @xmath184 can only lie on a @xmath145-hyperplane .
+ ( iii ) we have @xmath190 if and only if @xmath192 .
\(i ) note that @xmath193 ( ii ) and ( iii ) are clear .
the description of the blocks of the brauer algebra in characteristic zero in theorem [ balver ] was given the following geometric reformulation in @xcite : [ geoblock ] two standard modules @xmath37 and @xmath194 for @xmath1 are in the same block if and only if @xmath195 and @xmath196 are in the same @xmath197-orbit .
in summary , we have shown that there is a space that plays a role analogous to a weight space ( in lie theory ) for the brauer algebra , together with an action of the type @xmath13 coxeter group which plays the role of the weyl group , while the corresponding type @xmath16 coxeter group plays the role of the affine weyl group .
we will now consider the geometry of facets induced by @xmath167 inside @xmath176 .
we will call reflection hyperplanes _ walls _ , and for any collection of hyperplanes we will call a connected component of the set of points lying on the intersection of these hyperplanes but on no other a _ facet_. ( then an alcove is a facet corresponding to the empty collection of hyperplanes . ) it will be convenient to have an explicit description of the set of vectors in a given facet . for vectors in @xmath176 ( which will be the only ones which concern us ) these facets are defined by the hyperplanes @xmath198 for some @xmath199 . for @xmath200 in @xmath176 , note that for all @xmath201 we have @xmath202 we will call @xmath203 a _ singleton _ if @xmath203 is the only coordinate with modulus @xmath204 , and the pair @xmath205 a _ doubleton _ if @xmath206 and @xmath207 . for a given facet @xmath208 with @xmath209 ,
a vector @xmath210 lies in @xmath208 if and only if @xmath211 whenever @xmath206 , and @xmath212 whenever @xmath213 .
therefore an alcove ( where every @xmath203 is a singleton ) is determined by a permutation @xmath214 from @xmath215 to @xmath215 where @xmath216 is the @xmath10th smallest modulus occurring in @xmath43 .
note that not every permutation corresponds to an alcove in this way .
further , if @xmath217 then @xmath218 , while if @xmath219 then @xmath220 . for more general facets we replace the permutation @xmath214 by a function @xmath221 such that @xmath222 is the coordinate , or pair of coordinates , where the @xmath10th smallest modulus in @xmath43 occurs . for example ,
if @xmath223 then the facet containing @xmath43 corresponds to a function whose first four values are @xmath224 , @xmath225 , @xmath226 , and @xmath227 .
we will denote by @xmath228 the alcove corresponding to the identity permutation .
thus @xmath228 consists of all @xmath229 such that @xmath230 and @xmath231 .
it is easy to see that , for any @xmath232 , the weight @xmath0 is @xmath3-regular , with the vector @xmath233 in @xmath228 ; in this case we will call the @xmath234-alcove the _
@xmath3-fundamental alcove_. [ poscase ] for @xmath235
the set of weights in the @xmath3-fundamental alcove is @xmath236 by our discussion above , the desired set of weights is precisely the set of dominant @xmath51 such that @xmath237 and @xmath238 . but
this means that @xmath239 which implies the result . although our alcove geometry is reminiscent of that arising in positive characteristic lie theory , there are also some striking differences .
consider for example the case when @xmath21 .
the alcove @xmath228 is non - empty and contains the two weights @xmath0 and @xmath240 .
the next lowest alcove contains @xmath241 and @xmath242 , and the third contains @xmath243 and @xmath244 .
however , the associated facets with singularity @xmath27 are not necessarily finite ( in @xmath153 ) ; for example there is a facet consisting of the weights @xmath245 and the @xmath10-tuple @xmath246 for all @xmath247 . in particular ,
not every weight on a wall is adjacent to a weight in an alcove .
recall that for @xmath248 the brauer algebra @xmath249 is in schur - weyl duality with o@xmath250 acting on the @xmath251th tensor product of the natural representation .
suppose that @xmath252 .
the elements of @xmath253 corresponding to weights in the @xmath3-fundamental alcove are in bijection with the set of partitions labelling the irreducible representations which arise in a decomposition of tensor powers of the natural representation of @xmath254 . for o@xmath254 tensor space components
are labelled by partitions whose first and second columns sum to at most @xmath10 ( see for example ( * ? ? ?
* theorem 10.2.5 ) ) .
the result now follows by comparing with lemma [ poscase ] via the transpose map on partitions .
the above result shows that the fundamental alcove arises naturally in the representation theory of o@xmath254 .
suppose that @xmath20 .
choose @xmath255 so that @xmath256 ( if @xmath3 is even ) or @xmath257 ( if @xmath3 is odd ) .
it is easy to see that @xmath0 is @xmath3-singular of degree @xmath26 .
indeed , any dominant weight @xmath51 is @xmath3-singular of degree at least @xmath26 .
thus there are no regular dominant weights for @xmath20 . instead of the @xmath3-fundamental alcove
, we can consider the _
@xmath3-fundamental facet _ containing @xmath0 , for which we have [ negcase ] for @xmath258 of the form @xmath259 or @xmath260
the set of weights in the @xmath3-fundamental facet is @xmath261 .
we consider the case @xmath256 ; the odd case is similar .
the element @xmath233 equals @xmath262 and hence our facet consists of all vectors of the form @xmath263 where the sequence @xmath264 is decreasing ( as any other weight would be non - dominant ) . but
this implies that @xmath265 .
this is very different from the case @xmath266 .
however we do have [ fakealc ] suppose that @xmath256 .
the set of elements of @xmath253 corresponding to weights which can be obtained from @xmath0 via a sequence of one box additions only involving intermediate weights of singularity @xmath26 is in bijection with the set of partitions labelling the irreducible representations which arise in a decomposition of tensor powers of the natural representation of @xmath267 .
first note that we can clearly add boxes in the first @xmath26 coordinate of @xmath0 without changing the degree of singularity . in order to change the @xmath268st
coordinate in our path , we will have to pass through some point of the form @xmath269 where @xmath270 .
but this implies that the first @xmath271 coordinates of the vector all pair up with the corresponding negative values later down the vector , and so this is a singular vector of degree @xmath271 .
the result now follows from the description of tensor space components ( see for example ( * ? ? ?
* theorem 10.2.5 ) ) .
we will see in section [ brauertran ] that there is a sense in which the set of weights occurring in theorem [ fakealc ] can be regarded as playing the role of an alcove in the @xmath20 case .
towers of recollement were introduced in @xcite as an axiom scheme for studying various families of algebras . the brauer algebra over @xmath8 was shown to satisfy these axioms in @xcite .
we will prove a general result about morita equivalences in such towers , and apply it in the following section to the brauer algebra . in this section
we will work over a general field @xmath7 .
suppose that we have a family of @xmath7-algebras @xmath272 forming a tower of recollement .
the precise properties of such a tower will not concern us here ; such properties as we need will be introduced in what follows , and details can be found in @xcite . in each @xmath273 with @xmath247 there exists an idempotent @xmath274 such that @xmath275 .
this determines a pair of functors : _ localisation _
@xmath276 from @xmath277 to @xmath278 , and _ globalisation _
@xmath279 from @xmath277 to @xmath280 given on objects by @xmath281 the functor @xmath276 is exact , @xmath279 is right exact , and @xmath279 is left adjoint to @xmath282
. we also have algebra inclusions @xmath283 for each @xmath284 , and associated functors @xmath98 from @xmath277 to @xmath285 and @xmath286 from @xmath277 to @xmath287 .
let @xmath39 be an indexing set for the simple @xmath273-modules .
globalisation induces an embedding of @xmath39 inside @xmath288 , and we take @xmath132 to be the disjoint union of @xmath289 and @xmath290 , whose elements we call _ weights_. the algebras @xmath273 are quasihereditary , and so there exists a standard @xmath273-module @xmath37 for each weight @xmath51 in @xmath39 , such that the associated simple @xmath83 arises as its head .
the restriction of a standard module has a filtration by standards and we denote by @xmath291 the multi - set of labels for standard modules occurring in such a filtration of @xmath292 .
the embedding of @xmath39 in @xmath288 induces an embedding of @xmath291 inside @xmath293 , which becomes an identification if @xmath294 .
we denote by @xmath295 the set @xmath291 with @xmath296 .
suppose that we have determined the blocks of such a family of algebras ( or at least a linkage principle ) ; we are thinking of the cases where we have an alcove geometry at hand , but will avoid stating the result in that form . let @xmath297 be the functor @xmath298 and @xmath299 be the functor @xmath300 where @xmath301 is projection onto the block containing @xmath51 for @xmath302 .
note that @xmath297 is exact and @xmath299 is right exact .
we will regard these functors as analogues of _ translation functors _ in lie theory .
let @xmath303 denote the set of weights in the block of @xmath273 which contains @xmath51 .
our embedding of @xmath39 into @xmath288 induces an embedding of @xmath303 into @xmath304 , and we denote by @xmath305 the corresponding limiting set .
we will say that two elements @xmath51 and @xmath123 are _ translation equivalent _ if they satisfy 1 .
the weight @xmath123 is the only element of @xmath306 .
the weight @xmath51 is the only element of @xmath307 .
3 . for all weights
@xmath308 there is a unique element @xmath309 , and @xmath72 is the unique element in @xmath310 . clearly conditions ( i ) and ( ii ) are special cases of ( iii ) ; we list them separately as in an alcove geometry ( i ) and ( ii ) will be enough for ( iii ) to hold .
when @xmath51 and @xmath123 are translation equivalent then we will denote by @xmath311 the bijection taking @xmath72 to @xmath124 .
we will see that translation equivalent weights belong to morita equivalent blocks
. we will put a very crude partial order on weights in @xmath305 by saying that @xmath312 if there exists @xmath10 such that @xmath313 and @xmath314 for some @xmath315 , but @xmath316 .
note that this is the opposite of the standard order arising from the quasi - hereditary structure ; we prefer to work with the natural order on the size of partitions . in the following proposition , by a unique element in a multi - set we mean one with multiplicity one .
[ transmain ] let @xmath13 be a tower of recollement .
suppose that @xmath317 and @xmath318 are translation equivalent , and that @xmath319 is such that the @xmath124 is in @xmath320 .
then we have that @xmath321 for all @xmath322 .
further if @xmath323 is such that @xmath324 is in @xmath320 then we have @xmath325=[\delta_{n-1}(\mu'):l_{n-1}(\tau')]\ ] ] and @xmath326 we begin with ( [ simres ] ) .
consider the exact sequence @xmath327 applying @xmath328 we obtain by our assumptions the exact sequence @xmath329 and hence @xmath330 has simple head @xmath331 , and possibly other composition factors @xmath332 with @xmath333 .
if @xmath334 is in the socle of @xmath330 then we have @xmath335 by our assumptions and frobenius reciprocity , which implies that @xmath336 and hence @xmath337 .
therefore @xmath338 as required .
next consider the exact sequence @xmath339 applying @xmath340 we obtain by our assumptions the exact sequence @xmath341 and hence @xmath342 has simple head @xmath343 , and possibly other composition factors @xmath344 with @xmath345 . now
apply @xmath328 to obtain the exact sequence @xmath346 then @xmath347 has simple head @xmath331 and possibly other composition factors @xmath332 with @xmath333 corresponding to those in @xmath342 .
we have @xmath348 and hence @xmath331 must appear in the socle of @xmath347 .
this forces @xmath349 and as we already have that @xmath328 takes simples to simples we deduce that @xmath350 which completes our proof of ( [ simres ] ) . now ( [ decres ] ) follows immediately as @xmath351 , while ( [ homsame ] ) follows from @xmath352 let @xmath112 denote the projective cover of @xmath83 .
as our algebras are quasihereditary we have that @xmath112 has a filtration by standard modules with well - defined filtration multiplicities ; we denote the multiplicity of @xmath353 in such a filtration by @xmath354 .
[ transproj ] suppose that @xmath317 and @xmath318 are translation equivalent
. then for all @xmath322 with @xmath355 we have @xmath356 if @xmath357
we have @xmath358 we begin with ( [ indproj ] ) .
the functor @xmath359 takes projectives to projectives , and hence so does @xmath340 .
we must show that inducing an indecomposable projective gives an indecomposable projective with the right weight .
suppose we have an exact sequence @xmath360 for some @xmath361 .
then we have @xmath362 by proposition [ transmain ] .
therefore we must have @xmath363 , and hence @xmath364 and @xmath365 this implies that @xmath366 has simple head @xmath84 and hence is isomorphic to @xmath367 .
next we consider ( [ resproj ] ) .
as @xmath368 is a direct summand of the left @xmath273-module @xmath273 , it is a projective @xmath273-module .
moreover , as @xmath275 we have that @xmath368 contains precisely those indecomposable projective @xmath273-modules labelled by weights in @xmath369 . in a tower of recollement
we have that @xmath370 as a left @xmath371-module .
this implies that for @xmath372 , the module @xmath373 ( and hence @xmath374 ) is projective .
as @xmath328 is an exact functor and @xmath367 has simple head @xmath84 we know from proposition [ transmain ] that @xmath375 for some projective @xmath371-module @xmath376 .
however , by brauer - humphreys reciprocity and proposition [ transmain ] we have @xmath377= [ \delta_{n-1}(\tau'):l_{n-1}(\mu ' ) ] = ( p_{n-1}(\mu'):\delta_{n-1}(\tau')).\ ] ] as @xmath328 is exact and takes @xmath378 to @xmath379 this implies that @xmath380 .
we would like to argue that two blocks labelled by translation equivalent weights are morita equivalent .
however , the fact that not every projective module restricts to a projective in ( [ resproj ] ) causes certain complications .
[ pstable ] if @xmath317 then @xmath381 by ( * ? ? ?
* chapter i , theorem 6.8 ) @xmath382 is an indecomposable projective .
we have an exact sequence @xmath383 and hence as @xmath279 is right exact and takes standards to standards we obtain @xmath384 this implies that @xmath385 .
[ phomstable ] if @xmath386 then @xmath387 and this extends to an algebra isomorphism @xmath388 where @xmath389 . see ( * ? ? ?
* chapter i , theorem 6.8 ) .
the algebra @xmath273 decomposes as a direct sum of indecomposable projective modules : @xmath390 for some integers
there is a corresponding decomposition of @xmath392 as a sum of ( not necessarily primitive ) orthogonal idempotents @xmath393 where @xmath394 . as @xmath39 decomposes as a union of blocks the algebra @xmath273 decomposes as a direct sum of ( block ) subalgebras @xmath395 where the sum runs over a set of block representatives and @xmath396 now let @xmath397 and consider the idempotent @xmath398 .
we define the algebra @xmath399 by @xmath400 by lemma [ phomstable ] we have that @xmath399 and @xmath401 are morita equivalent for all @xmath26 such that @xmath402 .
[ genmor ] suppose that @xmath51 and @xmath123 are translation equivalent with @xmath403 , and set @xmath404 then @xmath405 and @xmath406 are morita equivalent . in particular , if there exists an @xmath10 such that @xmath407 then @xmath405 and @xmath408 are morita equivalent
. we will show that the basic algebras corresponding to @xmath405 and @xmath406 are isomorphic ; i.e. that @xmath409 by lemma [ phomstable ] it is enough to show that @xmath410 suppose that @xmath411 . then by lemma [ phomstable ] and proposition [ transproj ] we have @xmath412 next we will show that these isomorphisms are also compatible with the multiplicative structure in each of our algebras .
let @xmath413 , @xmath376 , and @xmath414 be indecomposable projectives for @xmath415 labelled by elements from @xmath303 .
then there exist indecomposable projectives @xmath416 , @xmath417 and @xmath418 labelled by elements in @xmath419 such that @xmath420 an isomorphism @xmath19 giving a frobenius reciprocity of the form @xmath421 is given by the map taking @xmath422 to @xmath423 where @xmath424 for all @xmath425 and @xmath426 , and extending by linearity .
( recall that @xmath427 is just the function @xmath428. ) given @xmath429|{\times}\ar[d]_{\cong } & \operatorname{hom}_{n+2}(q , r)\ar[r]\ar[d]_{\cong}&\operatorname{hom}_{n+2}(p , r)\ar[d]_{\cong}\\ \!\!\!\!\!\!\!\!\!\operatorname{hom}_{n+2}(\operatorname{ind}_{n+1}^{\lambda}p',\operatorname{ind}_{n+1}^{\lambda}q')\ \
\ar@{}[r]|{\times } & \!\!\!\!\!\!\operatorname{hom}_{n+2}(\operatorname{ind}_{n+1}^{\lambda}q',r)\ar[r ] & \operatorname{hom}_{n+2}(\operatorname{ind}_{n+1}^{\lambda}p',r)\\ \!\!\!\!\!\!\!\!\ !
\operatorname{hom}_{n+1}(p',\operatorname{res}_{n+2}^{\lambda'}\operatorname{ind}_{n+1}^{\lambda}q'))\ \
\ar@{}[r]|{\times } \ar[u]^{\alpha } & \!\!\!\!\!\!\operatorname{hom}_{n+1}(q'\operatorname{res}_{n+2}^{\lambda'}r)\ar[r]\ar[u]^{\alpha } & \operatorname{hom}_{n+1}(p'\operatorname{res}_{n+2}^{\lambda'}r)\ar[u]^{\alpha}\\ \operatorname{hom}_{n+1}(p',q ' ) \ar@{}[r]|{\times}\ar@{}[d]|{\phi}\ar[u]^{\cong } & \operatorname{hom}_{n+1}(q',r')\ar@{}[d]|{\psi}\ar[u]^{\cong}\ar[r ] & \operatorname{hom}_{n+1}(p',r')\ar@{}[d]|{\psi\circ\phi}\ar[u]^{\cong}\\ & & } \ ] ] we need to check that @xmath430 .
we have @xmath431 where @xmath432 and @xmath433 . as @xmath434
we have @xmath435 where @xmath436 and @xmath437 .
now @xmath438 where the second equality follows from the action of @xmath415 on @xmath439 .
on the other hand @xmath440 and hence @xmath430 as required .
next we will consider how we can relate the cohomology of @xmath406 to that of @xmath408 and hence compare the cohomology of @xmath405 with that of @xmath408 .
we say that a subset @xmath441 is _ saturated _ if @xmath442 and @xmath443 with @xmath444 implies that @xmath445 .
a subset @xmath397 is _ cosaturated _ if @xmath446 is saturated .
[ cosat ] the set @xmath447 is cosaturated in @xmath419 .
we need to show that if @xmath448 and @xmath449 with @xmath450 then @xmath451 .
suppose for a contradiction that @xmath452
. then @xmath453 and @xmath454 . as @xmath455 we must have @xmath456 , and as @xmath457 we have @xmath458 now @xmath459 and so @xmath460 but this contradicts the assumption that @xmath461 .
if @xmath397 is cosaturated then by ( * ? ? ?
* a.3.11 ) the algebra @xmath399 is quasihereditary with standard modules given by @xmath462 moreover if @xmath463 is any @xmath273-module having a @xmath464-filtration with factors @xmath353 for @xmath442 , and @xmath465 is any @xmath273-module , then for all @xmath466 we have ( * ? ? ?
* a.3.13 ) @xmath467 combining the above remarks with theorem [ genmor ] and lemma [ cosat ] we obtain [ extsame ] if @xmath317 and @xmath468 are translation equivalent then for all @xmath466 and for all @xmath322 we have @xmath469 we will say that two weights @xmath51 and @xmath123 are in the same _ translation class _ if they are related by the equivalence relation generated by translation equivalence
. then analogues of ( [ decres ] ) , ( [ homsame ] ) , theorem [ genmor ] and corollary [ extsame ] also hold for weights in the same translation class
. in lie theory one can consider translation between two weights in the same facet ( corresponding to the case considered above ) , or from one facet to another .
we will now give an analogue of translation onto a wall for a tower of recollement .
we will say that @xmath123 _ separates _ @xmath470 and @xmath471 if 1 .
the weight @xmath123 is the only element of @xmath472 .
the weight @xmath123 is the only element of @xmath473 .
the weights @xmath471 and @xmath470 are the only elements of @xmath474 .
whenever we consider a pair of weights @xmath470 and @xmath471 separated by @xmath123 we shall always assume that @xmath475 .
[ ontowall ] ( i ) if @xmath318 separates @xmath470 and @xmath471 then @xmath476 ( ii ) if further we have @xmath477 then @xmath478 and @xmath479 is a nonsplit extension of @xmath480 by @xmath481 and has simple head @xmath482 .
arguing as in the proof of proposition [ transmain ] we see that @xmath483 is either @xmath0 or has simple head @xmath484 for @xmath485 . also , any other composition factors @xmath332 of @xmath483 must satisfy @xmath486 .
note that by assumption we have a short exact sequence @xmath487 and hence @xmath488 is non - zero when @xmath489 and @xmath490 , and is zero when @xmath486 by our assumptions .
this completes the proof of ( i ) .
now suppose that @xmath491 .
then we have that @xmath492\neq 0.\ ] ] by exactness and the first part of the theorem , the unique copy of @xmath484 in @xmath493 must come from @xmath494 , and hence @xmath495 can not have simple head @xmath484 . but
this implies by the first part of the proof that @xmath496 .
therefore the hom - space in ( [ twohom ] ) must be zero when @xmath490 and @xmath497 , which implies that ( [ sepses ] ) is a non - split extension whose central module has simple head @xmath482 as required .
suppose that @xmath123 and @xmath471 are weights with @xmath498 and @xmath499 such that for every weight @xmath500 either ( i ) there is a unique weight @xmath501 and @xmath324 is the unique weight in @xmath502 , or ( ii ) there exists @xmath503 such that @xmath324 separates @xmath504 and @xmath505 .
then we say that @xmath123 is in the _ lower closure _ of @xmath471 .
if further @xmath506 whenever @xmath507 separates @xmath504 and @xmath505 in @xmath508 then we shall say that @xmath508 has _ local homomorphisms _ with respect to @xmath509 .
[ alcok ] suppose that @xmath318 is in the lower closure of @xmath510 , and that @xmath508 has enough local homomorphisms with respect to @xmath509
. then @xmath511=[\delta_n(\lambda^+):l_n(\mu^+)].\ ] ] we have by our assumptions that @xmath512 as @xmath328 is an exact functor , it is enough to determine its effect on simples @xmath513 in @xmath481 .
if there exists @xmath124 separating @xmath514 from @xmath515 then the result follows from theorem [ ontowall ] , while if @xmath514 is the only element in @xmath516 then if follows as in the proof of proposition [ transmain ] .
thus , as long as there are enough local homomorphisms , the decomposition numbers for @xmath37 determine those for all weights in the lower closure of @xmath51 .
we can generalise the results of this section up to corollary [ extsame ] by replacing @xmath517 and @xmath518 by any pair of functor families @xmath519 and @xmath520 with the following properties .
let @xmath13 be a tower of recollement , with @xmath521 , and fix @xmath201 .
suppose that we have functors @xmath522 for @xmath523 and @xmath524 for @xmath284 satisfying 1 . the functor @xmath520 is left adjoint to @xmath525 for all @xmath10 .
the functor @xmath519 is exact and @xmath520 is right exact for all @xmath10 where they are defined .
3 . there is a bijection @xmath526 taking @xmath72 to @xmath124 such that for all @xmath523 , if @xmath322 and @xmath527 then @xmath528 and @xmath529 otherwise .
if @xmath530 for some @xmath26 then @xmath531 is cosaturated in @xmath532 .
there exists @xmath533 such that for all @xmath10 and for all @xmath534 the module @xmath535 is projective .
then we say that @xmath51 and @xmath123 are _
@xmath536-translation equivalent_. in this case the proofs of propositions [ transmain ] and [ transproj ] go through essentially unchanged , and we get [ ripair ] suppose that @xmath51 and @xmath123 are @xmath536-translation equivalent and @xmath523 .
then for all @xmath537 with @xmath538 we have @xmath539 and if @xmath534 then @xmath540 moreover , if the adjointness isomorphism @xmath541 is multiplicative ( i.e. makes the diagram ( [ comdiag ] ) commute ) then there is a morita equivalence between @xmath405 and @xmath542 and for all @xmath543 and @xmath544 we have @xmath545
we wish to show ( in section [ brauertran ] ) that two weights in the same facet for the brauer algebra give rise to morita equivalent blocks ( at least when we truncate the blocks to have the same number of simples )
. however , the usual induction and restriction functors are not sufficient to show this except in the alcove case . to remedy this , in this section we will consider a variation on the usual induction and restriction functors . as @xmath3 will be fixed throughout
, we will denote @xmath1 simply as @xmath29 .
first consider @xmath546 with @xmath24 .
it is easy to see that this is a semisimple algebra , with a decomposition @xmath547 of the identity into primitive orthogonal idempotents given by the elements in figure [ idemsare ] .
there are three standard modules for this algebra , which we will denote by @xmath548 for @xmath247 consider the subalgebra @xmath549 obtained by letting @xmath550 act on the leftmost @xmath34 lines and @xmath546 act on the rightmost pair of lines .
we will view elements of @xmath550 and @xmath546 as elements of @xmath29 via this embedding .
note that under this embedding the two algebras obviously commute with each other . in particular , for any @xmath29-module @xmath551 the vector spaces @xmath552 are @xmath550-modules .
thus we have a pair of functors @xmath553 from @xmath29-mod to @xmath550-mod given on objects by the map @xmath554 .
note that these functors can also be defined as @xmath555 we have @xmath556 and so the functors @xmath557 from @xmath550-mod to @xmath29-mod given by @xmath558 are left adjoint to @xmath553 .
[ altdef ] let @xmath14 be a @xmath550-module .
then we have @xmath559 as @xmath29-modules , where the action on the right - hand space is by left multiplication in @xmath29 .
define a map @xmath560 by @xmath561 we first show that this is well - defined .
let @xmath562 for some @xmath563 and @xmath564 .
then @xmath565 as required .
the map @xmath422 is clearly a @xmath29-homomorphism .
we also have a map @xmath566 given by @xmath567 it is easy to check that @xmath568 is well - defined and that @xmath569 and @xmath570 .
let @xmath571 be the idempotent in @xmath29 shown in figure [ bige ] .
[ alt ] as left @xmath29- and right @xmath550-modules we have @xmath572 consider the map from @xmath573 to @xmath574 given on diagrams as shown in figure [ e4iso ] .
the grey shaded regions show the actions of @xmath29 from above , of @xmath550 from below , and the dark shaded region the action of the element @xmath575 .
all lines in the diagrams except those indicated remain unchanged ; the two southern arcs in the left - hand diagram are removed , and the ends of the pair of lines acted on by @xmath575 are translated clockwise around the boundary from the northern to the southern side .
this gives an isomorphism of vector spaces , and clearly preserves the actions of @xmath29 and @xmath550 .
[ pmproj ] the module @xmath576 is projective for all @xmath577 .
first note that @xmath578 is a projective @xmath29-module .
moreover , as @xmath579 we have that @xmath578 contains precisely the indecomposable projectives labelled by elements of @xmath580 . by lemma [ alt ]
we have that @xmath581 as left @xmath550-modules , and hence @xmath576 is projective for all @xmath577 .
[ indrespm ] we have an isomorphism of functors @xmath582 by the definition of @xmath279 and @xmath583 we have @xmath584 where the final isomorphism follows from lemma [ alt ] .
but by lemma [ altdef ] this final module is isomorphic to @xmath585 .
corollary [ indrespm ] is an analogue of the relation between induction , restriction and globalisation in ( * ? ? ?
* lemma 2.6(ii ) ) , corresponding to axiom ( a4 ) for a tower of recollement . given two partitions @xmath51 and @xmath72 ,
we write @xmath586 , or @xmath587 , if @xmath72 can be obtained from @xmath51 by removing two boxes and @xmath118 is not the partition @xmath588 .
similarly we write @xmath589 , or @xmath590 if @xmath72 can be obtained from @xmath51 by removing two boxes and @xmath591
. we will write @xmath592 if @xmath72 is obtained from @xmath51 by removing a box and then adding a box .
the next theorem describes the structure of @xmath593 , and so is an analogue of the usual induction and restriction rules in ( * ? ? ?
* theorem 4.1 and corollary 6.4 ) ( and use the same strategy for the proof ) .
[ bigres ] suppose that @xmath51 is a partition of @xmath52 for some @xmath594 .
+ ( i ) there is a filtration of @xmath550-modules @xmath595 with @xmath596 where any @xmath597 which does not make sense is taken as @xmath0 .
+ ( ii ) there is a filtration of @xmath598-modules @xmath599 with @xmath600 part ( ii ) follows from part ( i ) by corollary [ indrespm ] . for the rest of the proof we will work with the concrete realisation of standard modules given in section [ basics ] . by definition
we have @xmath601 and we will represent an element @xmath602 in this space diagrammatically as shown in figure [ howrep ] .
we are now in a position to define the various spaces @xmath603 , @xmath604 , and @xmath605 .
choose a fixed basis @xmath606 for @xmath53 and set @xmath607 @xmath608 @xmath609 then for @xmath610 we set @xmath611 note that if @xmath612 and @xmath613 is obtained from @xmath614 by swapping nodes @xmath615 and @xmath10 , then @xmath616 moreover , if there is an arc linking nodes @xmath615 and @xmath10 in @xmath614 then @xmath617 thus we have that @xmath618 and @xmath603 and @xmath604 are submodules of @xmath619 .
we first show that @xmath620 where @xmath621 represents the symmetriser / antisymmetriser on the last two lines in @xmath58 and @xmath622 acts on the first @xmath623 lines .
note that @xmath624 where @xmath625 equals @xmath245 for @xmath626 and @xmath588 for @xmath627 .
as @xmath628 it will follow from ( [ concrete ] ) that @xmath629 as required .
note that for @xmath630 the lines from @xmath615 and @xmath10 are propagating in @xmath48 , and so we have @xmath631 for @xmath630 define @xmath632 by removing nodes @xmath615 and @xmath10 , and a map @xmath633 by @xmath634 it is clear that @xmath635 is an isomorphism of vector spaces , and commutes with the action of @xmath550 .
this proves ( [ firststep ] ) .
next we will show that @xmath636 note that @xmath637 and so it will follow from ( [ concrete ] ) that @xmath638 as required
. we will need an explicit description of @xmath639 .
the quotient @xmath640 has coset representatives @xmath641 where @xmath642 .
therefore @xmath639 has a basis @xmath643 and the action of @xmath644 is given by @xmath645 where @xmath646 for a unique @xmath647 and @xmath648 . for @xmath649 set @xmath650 and for @xmath651 define @xmath652 by removing the nodes @xmath615 and @xmath10 and removing the arc from @xmath615 ( which will thus introduce a new free node elsewhere in @xmath653 ) .
now we can define a map @xmath654 by @xmath655 if node @xmath615 is linked to node @xmath35 in @xmath614 .
this is illustrated graphically in figure [ phi1 ] .
note that for every @xmath656 there are exactly @xmath26 elements @xmath651 satisfying @xmath657 , as @xmath615 can be joined to any of the @xmath26 free vertices in @xmath43 .
note also that @xmath658 where @xmath659 , and so @xmath660 .
given @xmath656 , @xmath649 , and @xmath661 pick @xmath651 with @xmath657 and @xmath615 joined to the @xmath35th free node . then @xmath662 and so @xmath663 is surjective .
moreover @xmath664 and so @xmath663 is an isomorphism of vector spaces .
it remains to show that @xmath663 commutes with the action of @xmath550 .
first consider the action of @xmath665 .
the actions of @xmath663 and @xmath666 can be seen to commute by the schematic diagram in figure [ comm1 ] , noting that @xmath667 .
next consider the action of @xmath668 .
if @xmath669 then it is clear that @xmath670 commutes with @xmath663 . now consider the action of @xmath671 .
there are two cases : ( i ) @xmath36 is a free node in @xmath614 , and ( ii ) @xmath36 is linked to some node @xmath7 in @xmath614 . case ( i )
is illustrated schematically in figure [ comm2 ] .
the lower left diagram in figure [ comm2 ] represents @xmath0 as it lies in @xmath603 .
the lower right diagram represents @xmath0 as there is a decrease in the number of propagating lines .
therefore the dotted arrow is an equality and the diagram commutes . case ( ii ) is illustrated schematically in figure [ comm3 ] .
again we see that @xmath671 commutes with the action of @xmath663 , and so we have shown that @xmath663 is a @xmath550-isomorphism .
this completes the proof of ( [ secondstep ] ) .
finally , we shall show that @xmath672 as @xmath673 it will follow from ( [ concrete ] ) that @xmath674 which will complete the proof .
we will need an explicit description of @xmath675 .
the quotient @xmath676 has coset representatives @xmath677 where @xmath678 . therefore @xmath679 has a basis @xmath680 and the action of @xmath681 is given by @xmath682 where @xmath683 for a unique @xmath684 and @xmath685 .
for @xmath686 set @xmath687 and for @xmath688 define @xmath689 by removing the nodes @xmath615 and @xmath10 and removing the arcs from @xmath615 and @xmath10 ( which will thus introduce two new free nodes elsewhere in @xmath653 ) .
now we can define a map @xmath690 by @xmath691 if @xmath615 is linked to @xmath36 and @xmath10 is linked to @xmath35 in @xmath614 .
this is illustrated graphically in figure [ phi2 ] .
arguing as for @xmath663 we can show that @xmath692 is an isomorphism of vector spaces .
thus we will be done if we can show that @xmath692 commutes with the action of @xmath550 .
first consider the action of @xmath693 .
the actions of @xmath692 and @xmath666 are illustrated schematically in figure [ comm4 ] .
again we use the fact that @xmath694 , while in the bottom pair of diagrams we have used the action of @xmath575 on each side , which in each case gives a coefficient of @xmath695 .
we see that the actions of @xmath692 does commute with @xmath666 as required .
it remains to check that @xmath692 commutes with the action of @xmath696 .
if @xmath697 is disjoint from @xmath698 then it is clear that @xmath699 commutes with @xmath692 .
if @xmath700 and @xmath701 it is easy to verify that @xmath702 thus we just have to check what happens when @xmath700 and @xmath703 .
there are two cases : ( i ) @xmath137 is a free node in @xmath614 , and ( ii ) @xmath137 is linked to some node @xmath704 in @xmath614 . case ( i ) is illustrated schematically in figure [ comm5 ] .
the lower left diagram in figure [ comm5 ] represents @xmath0 as it lies in @xmath604 .
the lower right diagram represents @xmath0 as there is a decrease in the number of propagating lines .
therefore the dotted arrow is an equality and the diagram commutes .
case ( ii ) is illustrated schematically in figure [ comm6 ] .
again we see that @xmath705 commutes with the action of @xmath692 , and so we are done .
in section [ towermor ] we saw how translation equivalence of weights for a tower of recollement implies morita equivalences of the corresponding blocks ( when the blocks are truncated to contain the same number of simples ) .
we will now reinterpret this in the language of alcove geometry in the case of the brauer algebra .
given a partition @xmath51 , we saw in ( [ resrule ] ) that the set @xmath295 consists of those partitions obtained from @xmath51 by the addition or subtraction of a box from @xmath51 , all with multiplicity one .
we denote by @xmath706 the composition obtained by adding / subtracting a box from row @xmath35 of @xmath51 .
[ nosep ] if @xmath707 then there can not exist a reflection hyperplane separating @xmath51 from @xmath123 .
suppose that @xmath414 is a reflection hyperplane between @xmath51 and @xmath123 , and denote their respective reflections by @xmath708 and @xmath709 .
either @xmath51 is the reflection of @xmath123 or the line from @xmath51 to @xmath123 is not orthogonal to the hyperplane .
the former case is impossible as two weights differing by one box can not be in the same block . for the latter case , note
that the distance between @xmath51 and @xmath123 is one . therefore at least one of the distances from @xmath51 to @xmath708 and @xmath123 to @xmath709 is less than one . but this is impossible , as @xmath708 and @xmath709 are also elements of the lattice of weights . given a facet @xmath208 , we denote by @xmath710 the closure of @xmath208 in @xmath153 .
this will consist of a union of facets .
[ onwall ] if @xmath711 and @xmath712 for some facet @xmath208 then @xmath713 if and only if @xmath714 .
we first show that if @xmath715 then @xmath714 . by the interpretation of blocks in terms of contents of partitions ( theorem [ balver ] ) there is precisely one other weight @xmath716 in @xmath295 in the same block as @xmath123 .
also , one of these weights is obtained from @xmath51 by adding a box , and one by subtracting a box .
but this implies that @xmath123 is the reflection of @xmath716 about some hyperplane @xmath717 ; this reflection must fix the midpoint on the line from @xmath123 to @xmath716 , which is @xmath51 , and so @xmath718 . by lemma [ nosep ]
there is no hyperplane separating @xmath51 from @xmath123 .
however to complete the first part of the proof we still need to show that if @xmath719 for some hyperplane @xmath720 then @xmath721 too .
first note that by proposition [ wis](ii ) @xmath717 and @xmath720 must be @xmath722 and @xmath723 hyperplanes respectively for some quadruple @xmath724 .
it is easy to check that either @xmath722 fixes @xmath720 or @xmath725 is an @xmath726-hyperplane . but @xmath727 is dominant and so @xmath722 must fix @xmath720 . hence @xmath728 and as @xmath51 is the midpoint between @xmath123 and @xmath716 we must have @xmath721 .
for the reverse implication , suppose that @xmath729 .
then for all hyperplanes @xmath720 with @xmath719 we have @xmath721 and there is ( at least ) one hyperplane @xmath717 with @xmath718 and @xmath730 .
suppose that @xmath717 is an @xmath722-hyperplane , and consider @xmath731 . if @xmath732 then we are done .
otherwise by lemma [ nosep ] we have that @xmath716 must lie on the boundary of the dominant region , and hence in some @xmath733-hyperplane @xmath734 .
now @xmath735 and hence @xmath736 ( as @xmath737 ) .
therefore we must have @xmath721 .
but @xmath51 is fixed by @xmath722 and so @xmath738 .
this implies that @xmath739 which is a contradiction .
thus we have shown that @xmath732 and so @xmath740 .
[ alcovemor ] if @xmath51 is in an alcove then @xmath72 is in the same translation class as @xmath51 if and only if it is in the same alcove . by lemmas
[ nosep ] and [ onwall ] it is enough to show that if @xmath72 is in the same alcove as @xmath51 then @xmath72 can be obtained from @xmath51 by repeatedly adding or subtracting a box without ever leaving this alcove .
suppose that @xmath51 and @xmath72 are in the same alcove , and set @xmath237 and @xmath741 , the corresponding vectors in @xmath176 .
recall that there is a permutation @xmath214 defining the alcove @xmath13 introduced in section [ weyl ] .
we may assume that @xmath742 .
consider the sequence obtained by repeatedly adding ( or subtracting ) @xmath27 from @xmath743 until we obtain @xmath744 . at each stage
the vector @xmath43 obtained is of the form @xmath745 for some weight @xmath666 , and the sequence of weights thus obtained are such that each consecutive pair are translation equivalent .
now we repeat the process to convert @xmath746 into @xmath747 ( note that @xmath746 and @xmath747 have the same sign , and so the chain of weights constructed will always have @xmath748-coordinate satisfying the defining conditions for the alcove ) .
we continue in this manner until we have converted @xmath749 into @xmath750 .
this constructs a chain of translation equivalent weights connecting @xmath51 and @xmath72 and so we are done .
theorem [ alcovemor ] shows that the geometry on the weight space for @xmath1 comes naturally from the induction and restriction functors when the alcoves are non - empty ( i.e. for @xmath266 ) .
we would like to extend theorem [ alcovemor ] to the case of two weights in the same facet .
however , not all weights in the same facet are in the same translation class . to see this , note that a hyperplane is defined by the equation @xmath751 for some fixed pair @xmath35 and @xmath36 .
any modification of a weight in such a hyperplane by adding or subtracting a single box can not alter the value of the @xmath35th or @xmath36th coordinate without leaving the hyperplane .
however , we will see that if we also use the modified translation functors introduced in section [ genir ] then we do get the desired equivalences within facets .
let @xmath752 .
this set consists of those partitions obtained from @xmath51 by adding two boxes , removing two boxes , or adding a box and removing a box .
[ wallcase ] suppose that @xmath753 with @xmath754 .
if @xmath51 and @xmath755 are in the same facet then @xmath756 we take @xmath757 with @xmath758 and show that the above assumptions imply that @xmath759 .
there are six possible cases .
\(i ) suppose that @xmath760 and @xmath761 .
for these two weights to be in the same block the boxes @xmath762 and @xmath763 must pair up ( and so must @xmath764 and @xmath765 ) in the sense of condition ( 1 ) for a balanced partition .
this implies that there is a simple reflection @xmath766 taking @xmath767 to @xmath768 , which fixes @xmath51 . hence @xmath51 is in the @xmath766-hyperplane .
however , @xmath755 is not in this hyperplane , contradicting our assumption that they are in the same facet .
\(ii ) suppose that @xmath769 and @xmath770 .
for these two weights to be in the same block the elements @xmath771 must all be distinct and @xmath762 and @xmath764 must pair up ( and so must @xmath765 and @xmath763 ) . thus there is a reflection taking @xmath51 to @xmath755 , which contradicts our assumption .
\(iii ) suppose that @xmath772 and @xmath773 .
this is similar to ( ii ) .
\(iv ) suppose that @xmath760 and @xmath770 .
for these two weights to be in the same block we must have @xmath774 ( say ) .
but then @xmath775 is the reflection of @xmath776 through the @xmath777-hyperplane , and hence @xmath778 is in the @xmath777-hyperplane .
therefore @xmath51 is also in this hyperplane , but @xmath755 is not , which gives a contradiction .
\(v ) suppose that @xmath760 and @xmath773 .
this is similar to ( iv ) .
\(vi ) suppose that @xmath769 and @xmath773 .
first note that @xmath762 and @xmath764 can not pair up ( as this would imply that @xmath51 and @xmath755 are not in the same facet ) .
so for these two weights to be in the same block we must have @xmath762 pairing up with @xmath765 ( say ) and @xmath764 pairing up with @xmath763 .
but then @xmath779 is the reflection of @xmath780 through the @xmath777-hyperplane , which implies that @xmath51 is in this hyperplane but @xmath755 is not , which gives a contradiction .
[ samen ] suppose that @xmath781 with @xmath782 , and that @xmath51 lies on the @xmath145-hyperplane .
then @xmath51 and @xmath755 are in the same facet . moreover , if @xmath783 for some @xmath187 then @xmath784 satisfies @xmath785 for some @xmath786 .
the fact that @xmath51 and @xmath755 are in the same facet is clear .
now suppose that @xmath787 and @xmath784 .
then @xmath788 where @xmath789 or @xmath790 for some @xmath786 .
suppose for a contradiction that @xmath789 .
note that @xmath72 and @xmath791 are in the same facet , and so for any @xmath723-hyperplane on which @xmath72 and @xmath791 lie we must have that @xmath792 lies on the _ unshifted _ @xmath723-hyperplane .
this implies that @xmath793 and @xmath794 .
we have a sequence of dominant weights @xmath72 , @xmath795 and @xmath796 which are each at distance @xmath27 from their neighbours in the sequence .
we have already seen that they all lie on the same set of hyperplanes . moreover , by lemma [ nosep ] there can not exist a hyperplane separating @xmath72 from @xmath124 or @xmath124 from @xmath791 .
so @xmath72 , @xmath124 and @xmath791 all lie in the same facet .
now consider the image of these three weights under @xmath797 .
we get a corresponding sequence @xmath51 , @xmath123 and @xmath755 .
these weights must also lie in a common facet ( and hence @xmath123 is dominant ) and are distance @xmath27 from their neighbours .
this forces @xmath798 or @xmath799 .
however @xmath123 can not be in the same facet as @xmath51 as it does not lie on the @xmath722-hyperplane , which gives the desired contradiction .
let @xmath800 and @xmath801 .
we say that @xmath51 and @xmath72 are in the same _ @xmath802-translation class _ if there is a chain of dominant weights @xmath803 such that either @xmath804 with @xmath805 and @xmath806 translation equivalent or @xmath807 with @xmath808 ( for some @xmath809 and @xmath41 ) and @xmath805 and @xmath806 are @xmath810-translation equivalent .
suppose that @xmath811 with @xmath760 , and that @xmath51 lies in the @xmath722-hyperplane . by lemmas [ wallcase ] and [ samen ] we have a bijection @xmath812 which restricts to a bijection @xmath813 .
by corollary [ pmproj ] , theorem [ bigres ] , lemmas [ wallcase ] and [ samen ] , and standard properties of @xmath814 and @xmath815 it is clear that weights @xmath51 and @xmath755 are @xmath816 translation equivalent .
it is also easy to see that the adjointness isomorphism is multiplicative .
thus we can apply theorem [ ripair ] and get a morita equivalence between the two blocks @xmath817 and @xmath818 .
[ factran ] if @xmath51 and @xmath72 are in the same facet then they are in the same @xmath802-translation class . by lemmas
[ onwall ] and [ wallcase ] it is enough to show that if @xmath51 and @xmath72 are in the same facet then there is a chain of dominant weights @xmath803 in the same facet such that @xmath804 or @xmath807 for each @xmath35 .
let @xmath237 and @xmath741 , and recall that in section [ weyl ] we associated a function @xmath819 to each facet ( rather than just a permutation @xmath214 as for an alcove ) .
the proof now proceeds exactly as for the alcove case ( theorem [ alcovemor ] ) replacing @xmath214 by @xmath819 , until we reach some point where @xmath820 . in this case
we repeatedly add ( or subtract ) a box from @xmath821 and subtract ( or add ) a box to @xmath822 until we reach @xmath823 and @xmath824 . in each of these steps
we obtain some @xmath825 . applying the results on translation and @xmath536-translation equivalence from section [ towermor ] we deduce [ sumup ]
if @xmath51 and @xmath123 are in the same facet and @xmath72 and @xmath124 are such that @xmath826 , and @xmath827 , and @xmath124 is the unique weight in @xmath509 in the same facet as @xmath72 , then we have : + ( i ) @xmath828=[\delta_m(\lambda'):l_m(\mu')]\ ] ] ( ii ) @xmath829 ( iii ) @xmath830 for all @xmath831 . if further @xmath303 and @xmath832 contain the same number of simples then the corresponding blocks are morita equivalent . by (
* theorem 3.4 ) there are always enough local homomorphisms for the brauer algebra .
further , by lemma [ onwall ] any weight that is not adjacent to a weight in a less singular facet is translation equivalent to a weight of smaller total degree .
thus if @xmath266 then every weight can be reduced to a weight in some alcove by translation equivalence and repeated applications of proposition [ alcok ] .
this implies [ plus0 ] if @xmath266 then the decomposition numbers @xmath833 $ ] for arbitrary @xmath51 and @xmath72 are determined by those for @xmath51 and @xmath72 in an alcove .
note that the restriction on @xmath3 is necessary , as for @xmath20 there are no weights in an alcove .
in fact for @xmath256 or @xmath257 any dominant weight is @xmath3-singular of degree at least @xmath26 .
for the rest of this section we will see what more can be said in such cases .
we will denote the set of all partitions @xmath51 with at most @xmath26 non - zero parts by @xmath834 , and the set of those with at most @xmath271 non - zero parts with @xmath835 by @xmath836 .
( note that @xmath834 is precisely the set of weights considered in theorem [ fakealc ] . )
such weights lie in a union of facets , but we shall see that together they play a role analogous to that played by the fundamental alcove in the @xmath266 case .
we begin by noting [ fullrep ] ( i ) for @xmath256 , every @xmath3-singular weight of degree @xmath26 is in the same block as a unique element of @xmath834 .
+ ( ii ) for @xmath257 , every @xmath3-singular weight of degree @xmath26 is in the same block as a unique element of @xmath837 .
+ \(i ) suppose that @xmath256 and let @xmath51 be a @xmath3-singular weight of degree @xmath26 .
then @xmath184 is of the form @xmath838 where the only elements of equal modulus are those of the form @xmath839 , and the bracketed @xmath0 may or may not appear . note that as @xmath51 is a finite weight the tail of @xmath184 will equal the tail of @xmath840 , i.e. has value @xmath841 in position @xmath842 for all @xmath296 , and we assume that this holds for the @xmath10 in the expression above ( and similar expressions to follow ) .
first suppose that @xmath184 contains @xmath0 .
then @xmath184 is in the same @xmath167-orbit as @xmath843 thus the @xmath615 coordinates between the entries @xmath0 and @xmath841 must be strictly decreasing , which forces @xmath844 hence we deduce that @xmath845 as required .
next suppose that @xmath184 does not contain @xmath0 .
then there are two cases depending on the parity of the number of positive entries in @xmath184 .
the first case is when @xmath184 is in the same @xmath167-orbit as @xmath846 where @xmath749 is some positive integer and all entries after @xmath847 are negative . arguing as above we see that @xmath848 but this vector is @xmath3-singular of degree @xmath271 , which contradicts our assumptions on @xmath51 .
the second case is when @xmath184 is in the same @xmath167-orbit as @xmath849 where all entries after @xmath847 are negative . but
this implies that @xmath850 has @xmath10 strictly decreasing coordinates between the entries @xmath0 and @xmath841 , which is impossible .
the argument for ( ii ) is very similar .
we see that @xmath184 is in the same @xmath167-orbit as either @xmath851 or @xmath852 where in each case all entries after @xmath847 are negative . in the first case we deduce as above that @xmath853 and so @xmath854 as required . in the second case ,
as @xmath850 must be @xmath3-singular of degree @xmath26 , we deduce that @xmath855 where @xmath856 denotes the omission of the entry @xmath749 .
but this element is in the same @xmath167-orbit as @xmath857 ( by swapping @xmath749 and @xmath858 with a change of signs , and rearranging to get a decreasing sequence ) .
thus @xmath51 is in the same @xmath167-orbit as @xmath109 , and @xmath859 as required .
although the weights in @xmath834 ( respectively in @xmath837 ) lie in several different facets , the next result shows that all these facets have equivalent representation theories .
[ bigalc ] let @xmath256 ( respectively @xmath257 ) and @xmath860 ( respectively @xmath861 ) .
then @xmath51 and @xmath123 are translation equivalent if and only if @xmath862 ( respectively @xmath863 ) .
note that all weights in @xmath834 ( respectively in @xmath837 ) are @xmath3-singular of degree @xmath26 , and that any pair of such weights can be linked by a chain of weights in the same set differing at each stage only by the addition or subtraction of a single block . by lemma [ onwall ] we see that any such pair is translation equivalent . for the reverse implication , we first consider the @xmath256 case , with @xmath864 , and suppose that @xmath711 is not an element of @xmath834 .
then we must have @xmath865 .
now @xmath237 and @xmath866 differ only in the @xmath271st coordinate , which is @xmath0 or @xmath27 respectively . if @xmath819 is the function associated to the facet containing @xmath750 and @xmath867 is the corresponding function for @xmath868 then the only difference is that @xmath869 while @xmath870 .
thus @xmath871 , where @xmath208 is the facet containing @xmath51 , and so by lemma [ onwall ] this pair can not be translation equivalent .
the case @xmath257 is similar . arguing as above we have that @xmath872 or @xmath873 , and in each case it is easy to show that the pair @xmath51 and @xmath123 are not translation equivalent . combining propositions [ fullrep ] and [ bigalc ]
we deduce that for @xmath20 and @xmath874 ( respectively @xmath875 ) there is a bijection @xmath876 which as before we will denote by @xmath877 . applying the results from section [ towermor ] we obtain
let @xmath20 and @xmath878 ( respectively @xmath875 ) . if @xmath879 and @xmath880 then ( i iii ) of corollary [ sumup ] hold . if further @xmath881 and @xmath882 contain the same number of elements then the corresponding blocks are morita equivalent . as in corollary
[ plus0 ] , we obtain the following application of proposition [ alcok ] .
[ neg0 ] if @xmath20 then the decomposition numbers @xmath833 $ ] for arbitrary @xmath51 and @xmath72 are determined by those for @xmath51 and @xmath72 in a singular facet of degree @xmath26 .
combining corollaries [ plus0 ] and [ neg0 ] with our earlier remarks we obtain [ need0 ] for @xmath180 non - zero the decomposition numbers @xmath833 $ ] for arbitrary @xmath51 and @xmath72 are determined by those for @xmath51 and @xmath72 in @xmath883 .
thus ( at least at the level of decomposition numbers ) is it enough to restrict attention to a single block of the brauer algebra .
the decomposition numbers for the module @xmath884 are known by ( * ? ? ?
* proposition 5.1 and theorem 5.2 ) .
we would also like the representation theory to be independent of @xmath180 , in the sense that it should depend only on the geometry of facets . for weights in alcoves , this would in large part follow if we could show that decomposition numbers are given by some kind of parabolic kazhdan - lusztig polynomials . in the remaining sections we will consider some evidence for this .
recall from section [ basics ] the definition of a maximal balanced partition .
let @xmath885 be the directed graph with vertex set @xmath75 and edge @xmath886 if @xmath72 is a maximal @xmath3-balanced subpartition of @xmath666 .
the above graph appears to depend both on @xmath51 and @xmath3 , while the alcove geometry associated to @xmath167 does not .
let @xmath887 be the directed graph with vertex set the set of alcoves for @xmath167 in @xmath176 , and an edge @xmath888 if the closures of @xmath13 and @xmath889 meet in a hyperplane and this hyperplane separates @xmath228 and @xmath889 .
( note that the former condition corresponds to @xmath890 for some reflection @xmath145 . )
our goal in this section is to show that all the graphs @xmath885 are in fact isomorphic , and are isomorphic to the alcove graph @xmath887 .
this will be our first evidence that the representation theory depends only on the geometry of facets .
recall that for @xmath891 we have @xmath892 ( the set of strictly decreasing sequences in @xmath893 ) and the @xmath3-dot action of @xmath167 on @xmath51 corresponds to the usual action of @xmath167 on @xmath894 . for the rest of this section we will work with the usual action of @xmath167 on @xmath176 .
for @xmath229 we define @xmath895 we define a partial order on @xmath176 by setting @xmath896 if @xmath897 and all entries in @xmath898 are non - negative . for @xmath229
we define a directed graph @xmath899 with vertex set @xmath900 and arrows given as follows . if @xmath901 , we set @xmath902 if and only if @xmath903 and there is no @xmath904 with @xmath905 .
the reason for introducing this graph is clear from [ iso1 ] for @xmath891 we have @xmath906 . by theorem [ geoblock ] we have a bijection between @xmath907 and @xmath908
. moreover , for @xmath909 we have @xmath910 if and only if @xmath911 .
thus the two graph structures on these vertex sets are preserved under the correspondence .
recall the definition of singletons from section [ weyl ] . define @xmath912 to be the subsequence of @xmath43 consisting only if its singletons .
for example , if @xmath43 begins @xmath913 then @xmath912 begins @xmath914 note that if @xmath229 then @xmath915 and @xmath916 for all @xmath199 .
therefore @xmath912 is a regular element in @xmath893 ( as it does not lie on any reflecting hyperplane ) .
we define the _ regularisation map _ @xmath917 by setting @xmath918 the key result about the regularisation map is [ iso2 ] for all @xmath229 we have @xmath919 we first observe that the map @xmath920 gives rise to a bijection between @xmath900 and @xmath921 . for the set of doubletons
is an invariant of the elements in @xmath900 , and given this set there is a unique way of adding the doubletons into an element of @xmath921 keeping the sequence strictly decreasing .
now suppose that @xmath922 and @xmath923 are such that @xmath924 then it is easy to see that @xmath903 if and only if @xmath925 .
however , this implies that the set of edges coincide under the map @xmath920 , as required . for all @xmath926 we have @xmath927 hence for all
@xmath928 and @xmath929 we have @xmath930 note that any @xmath931 lies inside an alcove . for any vector @xmath932
the maximal weights below @xmath43 in the same orbit lie in the alcoves below and adjacent to the alcove containing @xmath43 .
thus for @xmath933 it is clear that we have @xmath934 .
now the result follows for general @xmath43 from proposition [ iso1 ] , and in its @xmath935 form from proposition [ iso2 ] . it will be convenient to give @xmath887 the structure of a graph with coloured edges .
an edge in @xmath887 corresponds to reflection from an alcove @xmath13 to an alcove @xmath889 through the facet separating them .
the action of @xmath167 on weights induces a corresponding action on facets , and we shall say that two edges have the same colour if and only if the corresponding facets lie in the same orbit .
we conclude this section with one final graph @xmath936 isomorphic to @xmath887 , whose structure can be described explicitly .
we fix the element @xmath937 . using the action of @xmath167
we can see that every @xmath938 corresponds uniquely to a strictly decreasing partition with an even number of parts , obtained by ignoring all parts of @xmath750 which are negative .
for example , the element @xmath939 corresponds to @xmath940 while @xmath43 corresponds to @xmath941 .
thus if we write @xmath942 for the set of strictly decreasing partitions with an even number of parts then we have a bijection @xmath943 consider the usual partial order @xmath944 on @xmath942 given by inclusion of partitions ( viewed as young diagrams ) .
it is clear that the partial order @xmath945 on @xmath900 corresponds to the partial order @xmath944 on @xmath942 under the bijection @xmath422 .
define a graph @xmath936 with vertex set @xmath942 and an arrow @xmath946 if and only if @xmath947 and there is no @xmath948 with @xmath949 .
it is easy to verify that the map @xmath422 induces a graph isomorphism between @xmath899 and @xmath936 .
the graph @xmath936 can easily be described explicitly as follows . for @xmath950
, there is an arrow @xmath951 if and only if either @xmath952 or @xmath953 to see this , first observe that in both cases there is no @xmath954 with @xmath949 . moreover , if @xmath955 with @xmath947 then @xmath72 can be obtained from @xmath51 by applying ( [ cond1 ] ) and ( [ cond2 ] ) repeatedly .
in this section we will introduce parabolic kazhdan - lusztig polynomials associated with the pair @xmath956 .
we briefly review the relevant theory ; details can be found in @xcite and @xcite .
recall that a coxeter system is a pair @xmath957 consisting of a group @xmath169 and a set @xmath958 of generators of @xmath169 such that all relations in @xmath169 are of the form @xmath959 where @xmath960 and @xmath961 otherwise ( including the possibility that @xmath962 denoting no relation between @xmath809 and @xmath963 ) .
note that the group @xmath169 does not need to be finite ( although this is often assumed ) . given a coxeter system ,
the associated coxeter graph is the graph with vertices the elements of @xmath958 , and @xmath964 edges between @xmath809 and @xmath963 ( or no edges when @xmath962 ) .
for example the @xmath965 coxeter graph is given by the graph shown in figure [ dcox ] .
coxeter system ] it is easy to verify that our group @xmath167 is generated by the elements @xmath966 and satisfies the relations given by the coxeter graph in figure [ dcox ] .
thus it must be a quotient of the coxeter group of type @xmath965 .
however , for each choice of @xmath10 , the subsystem generated by the first @xmath10 generators is precisely the type @xmath967 system ( see ( * ? ? ?
* planche iv ) ) , and so there can be no further relations , and our group is the type @xmath965 coxeter group associated to the given generators . given a coxeter system @xmath957 , any subgroup @xmath968 generated by a subset @xmath969 of @xmath958 defines a parabolic subsystem @xmath970 . in our case
the group @xmath135 clearly arises in this way from the generators of the form @xmath971 and so is a type @xmath972 parabolic subgroup of @xmath167 . when @xmath235 there is a bijection from @xmath167 to the set of alcoves , given by @xmath973 .
we will henceforth identify elements of @xmath167 with alcoves via this map . under this bijection
the standard length function on our coxeter system associated to @xmath167 ( given in terms of the number of terms occurring in a reduced expression for @xmath614 ) corresponds to the number of reflection hyperplanes between @xmath0 and @xmath974 .
we define @xmath975 to be the subset of @xmath167 corresponding to the alcoves in @xmath168 . by proposition [ wis](ii )
we then have a bijection @xmath976 we are thus in a position to define @xmath977 parabolic kazhdan - lusztig polynomials following @xcite ( although we use the notation of ( * ? ? ?
* section 3 ) ) .
their precise definition and general properties need not concern us , instead we will give a recursive construction corresponding to stepping away from the root of @xmath887 .
to do this we will first need to define a partial order on weights ( and alcoves ) .
two weights @xmath51 and @xmath72 such that @xmath978 for some reflection @xmath614 , lie in different components of the space formed by removing this hyperplane .
we say that @xmath979 if @xmath51 is in the component containing the fundamental alcove .
this extends to give a partial order on weights , which in turn induces a partial order on alcoves .
this agrees with the path - from - root order on @xmath887 .
two alcoves are said to be _ adjacent _ if there is precisely one reflecting hyperplane between them ( i.e. they are adjacent in @xmath887 ) .
suppose that @xmath109 and @xmath72 are dominant weights in adjacent alcoves with @xmath980 .
given a dominant weight @xmath981 we define @xmath982 to be the unique weight such that @xmath983 i.e. @xmath984 is an edge of the same colour as @xmath985 in @xmath887 .
we next define certain polynomials @xmath986 ( in an indeterminate @xmath43 ) for regular weights @xmath51 and @xmath72 in the following recursive manner .
let @xmath987 as @xmath51 runs over the regular weights be a set of formal symbols . 1 .
we set @xmath988 if @xmath989 or @xmath990 or either @xmath51 or @xmath109 is non - dominant
we set @xmath991 and @xmath992 .
3 . for each @xmath993
regular dominant , there exists some @xmath72 regular dominant below it such that @xmath994 and @xmath995 are in adjacent alcoves .
pick any such @xmath72 . then for any dominant @xmath51 with @xmath996 for some @xmath614 and @xmath997 we set @xmath998 where @xmath999 if @xmath1000 and @xmath1001 otherwise . note that for @xmath1000 we have @xmath1002 if @xmath1003 , respectively @xmath1004 if @xmath1005 .
let @xmath1006 be the sum @xmath1007 and @xmath1008 be the set of @xmath1009 such that @xmath1010 .
then @xmath1011 and @xmath986 is the coefficient of @xmath987 in @xmath1012 .
it is a consequence of ( deodhar s generalisation of ) kazhdan - lusztig theory that this process is well defined ( so does not depend on the choice of @xmath72 in step ( iii ) ) , and that each @xmath986 is a polynomial in @xmath43 with @xmath1013 only if @xmath1014 .
to illustrate the various constructions so far , we will consider the case when @xmath21 , and examine the regular block containing the weight @xmath0 .
first we calculate the associated parabolic kazhdan - lusztig polynomials , and then we compare these with the representation theoretic results
. we will also need to consider the block containing @xmath240 . as this is in the same alcove as @xmath0 these two blocks are translation equivalence .
however , in this simple case case we do not obtain any simplification to the calculations by applying the results from section 6 ; instead the results can be considered as a verification of the general theory in this special case . in figure [ blockex ]
we have listed all dominant weights of degree at most 16 that are in the same block as the weight @xmath0 .
we will abbreviate weights in the same manner as partitions ( and so write for example @xmath1015 instead of @xmath1016 ) .
an edge between two weights indicates that they are in adjacent alcoves , and the label @xmath145 corresponds to the reflection hyperplane between them .
( clearly only weights of the form @xmath145 can arise as such labels . )
given this data we can now compute the @xmath1017 .
the final results are shown in figure [ kltable ] .
we start with the weight @xmath0 having @xmath1018 . reflecting through @xmath1019
we obtain the weight @xmath120 , and we see that @xmath1020 .
( note that the term @xmath1021 is always @xmath27 . )
continuing we reflect @xmath120 through @xmath1022 to obtain @xmath1023 .
as @xmath1024 is not dominant we see that the only non - zero term apart from @xmath1025 is @xmath1026 .
identical arguments give all polynomials @xmath986 where @xmath109 is on the top row of figure [ blockex ] . for the second row
we obtain four terms as @xmath109 and @xmath72 both give dominant weights under the action of @xmath1027 for suitable @xmath35 ( as the parallelogram with @xmath109 as highest term has identically labelled parallel sides ) . for @xmath1028 we must observe that @xmath1029 and similar results give the remaining cases .
in all of these cases we have no constant terms arising at any stage ( apart from in @xmath1021 ) , and hence @xmath1030 for every weight considered .
+ @xmath1049 & @xmath1050 & & & & & & & & & & & & & & & & & + @xmath1032 & @xmath43&@xmath27 & & & & & & & & & & & & & & & & + @xmath1033 & .&@xmath43&@xmath27 & & & & & & & & & & & & & & & + @xmath1034 & .&.&@xmath43&@xmath27 & & & & & & & & & & & & & & + @xmath1035 & .&.&@xmath43&.&@xmath27 & & & & & & & & & & & & & + @xmath1036 & .&.&.&@xmath43&.&@xmath27 & & & & & & & & & & & & + @xmath1037 & .&.&@xmath1051&@xmath43&@xmath43&.&@xmath27 & & & & & & & & & & & + @xmath1038 & .&.&.&.&.&@xmath43&.&@xmath27 & & & & & & & & & & + @xmath1039 & .&.&.&@xmath1051&.&@xmath43&@xmath43&.&@xmath27 & & & & & & & & & + @xmath1040 & .&@xmath1051&@xmath43&.&.&.&@xmath43&.&.&@xmath27 & & & & & & & & + @xmath1041 & .&.&.&.&.&.&.&@xmath43&.&.&@xmath27 & & & & & & & + @xmath1042 & .&.&.&.&.&@xmath1051&.&@xmath43&@xmath43&.&.&@xmath27 & & & & & & + @xmath1043 & .&.&.&.&.&.&@xmath1051&.&@xmath43&@xmath43&.&.&@xmath27 & & & & & + @xmath1044 & .&.&.&.&.&.&.&.&.&.&@xmath43&.&.&@xmath27 & & & & + @xmath1045&.&.&.&.&.&.&.&@xmath1051&.&.&@xmath43&@xmath43&.&.&@xmath27 & & & + @xmath1046&.&.&.&.&.&.&.&.&@xmath1051&.&.&@xmath43&@xmath43&.&.&@xmath27 & & + @xmath1047&.&.&.&.&@xmath1051&.&@xmath43&.&.&.&.&.&@xmath43&.&.&.&@xmath27 & + @xmath1048&@xmath1051&@xmath43&.&.&.&.&.&.&.&@xmath43&.&.&.&.&.&.&.&@xmath27 + next we will determine the structure of certain low rank standard modules for @xmath1052 in the block containing @xmath0 .
these will then be compared with the kazhdan - lusztig polynomials calculated above
. we will proceed in stages , and will also need to consider the structure of modules in the block containing @xmath240 .
the submodule structure of modules will be illustrated diagrammatically , where a simple module @xmath463 is connected by a line to a simple module @xmath465 above it if there is a non - split extension of @xmath463 by @xmath465 .
note that in this section we follow the usual labelling of modules by partitions ( as in @xcite ) and _ not _ via the transpose map by weights . when @xmath1059 we have @xmath1060\\ l_6(321)}\ ] ] as in the case @xmath1057 . for the remaining module
@xmath1061 we know that @xmath1062=1 $ ] by localising to @xmath1057 . applying proposition [ projelim ] with @xmath1063 ( as this weight is minimal in its block )
we see that @xmath1064 can not occur in @xmath1061 .
hence we have that @xmath1065\\ l_6(22).}\ ] ] the odd @xmath10 cases are very similar .
arguing as above we see that @xmath1066\\ l_6(21)}\ ] ] and for @xmath1067 that @xmath1068\\ l_5(311)},\quad\quad \delta_5(1)= \xymatrix@=10pt { l_5(1 ) \ar@{-}[d]\\ l_5(21).}\ ] ] for the remaining two standard modules , all composition multiplicities are known ( by localising to the case @xmath1067 ) except for those for the new simple @xmath1071 .
however , this does not occur in @xmath1072 or in @xmath1073 by an application of proposition [ projelim ] with @xmath1074 ( as this weight is minimal in its block ) .
thus we have that @xmath1075\\ l_8(311)}\quad\quad{\mbox{\rm and } } \quad\quad \delta_7(1)= \xymatrix@=10pt { l_7(1 ) \ar@{-}[d]\\ l_7(21).}\ ] ] for the remaining two standard modules , all composition multiplicities are known ( by localising to the case @xmath1059 ) except for those involving @xmath1078 and @xmath1079
. however , neither of these occurs in @xmath1080 or in @xmath1081 by an application of proposition [ projelim ] with @xmath1082 , respectively @xmath1083 .
thus we have that @xmath1084\\ l_8(321)}\quad\quad{\mbox{\rm and } } \quad\quad \delta_8(0)= \xymatrix@=10pt { l_8(0 ) \ar@{-}[d]\\ l_8(22).}\ ] ] in this case we have @xmath1086 standard modules , labelled by @xmath240 , @xmath1087 , @xmath1088 , @xmath1089 , @xmath1090 and @xmath1091 . by ( [ twoboxstuff ] )
there is a homomorphism from each of these standards to the preceding one in the list , except in the case of @xmath1091 . for this weight
we instead use ( [ squarestuff ] ) which tells us that @xmath1092=1.\ ] ] as in earlier cases , we have that @xmath1093\\ l_9(3 ^ 3).}\ ] ] the module @xmath1094 can not occur in any other standards , by applying proposition [ projelim ] with @xmath1095 , and similarly @xmath1096 can only occur in @xmath1097 , by taking @xmath1098 . by the above observations and localisation to @xmath1069
we deduce that @xmath1099\ar@{-}[dr]&\\ l_9(41 ^ 3 ) & & l_9(3 ^ 3)}\quad\quad \delta_9(21)= \xymatrix@=10pt { l_9(21 ) \ar@{-}[d]\\
l_9(311)}\quad\quad \delta_9(1)= \xymatrix@=10pt { l_9(1 ) \ar@{-}[d]\\ l_9(21).}\ ] ] we will need to consider @xmath1100 . for this we need to understand the various standard modules arising in the short exact sequence ( [ resrule ] ) in this case . first note that @xmath1101 and @xmath1102 are the unique weights in their respective blocks when @xmath1085 . for the weights @xmath1103 , @xmath1104 and @xmath1105 there is exactly one larger weight in the same block in each case , respectively @xmath1106 , @xmath1107 , and @xmath1108 . from now on , we will summarise the results obtained for each value of @xmath10 in a single diagram , together with an explanation of how they were derived . in each such diagram
we shall illustrate the structure of individual modules as above , but label simple factors just by the corresponding partition .
we will indicate the existence of a homomorphism between two modules by an arrow .
( it will be clear which standard module is which by the label of the simple in the head . ) for @xmath1112 we claim that the structure of the block containing @xmath1113 is given by the data in figure [ case10 ] .
the structure of the modules @xmath1114 , @xmath1115 and @xmath1116 , follows exactly as in the preceding cases for partitions of @xmath10 and @xmath34 . for @xmath1117
we also need to note that @xmath1118 , and so @xmath1119 can not occur . to see that @xmath1119 can not occur anywhere else
it is enough to note ( by proposition [ projelim ] ) that @xmath1120 is projective .
similarly @xmath1121 can not occur in the standards @xmath1122 and @xmath1123 as @xmath1124 is projective .
the structure of @xmath1123 and @xmath1122 then follows by localisation to @xmath1076 .
the only remaining module is @xmath1125 .
it is clear that this must have at most the four factors shown .
the multiplicities of @xmath1126 and @xmath1127 must be @xmath27 by localisation to the case @xmath1076 .
it remains to show that the final factor has multiplicity @xmath27 , that there is a map to the module from @xmath1115 , and that the module structure is as shown .
consider @xmath1129 . by ( [ resrule ] ) , and the simplicity of standard modules @xmath37 when @xmath92 , this has simple factors @xmath1130 if we consider @xmath1131 and @xmath1132 , using the structure of @xmath1116 and @xmath1117 given above , it is easy to show that neither @xmath1133 nor @xmath1134 contain any of the factors in ( [ 10simfac ] ) . comparing with ( [ 9res ] ) , ( [ 9lower ] ) , and ( [ 9upper ] )
we see that @xmath1135\leq 1.\ ] ] further , either this simple does occur , or the simples in ( [ 10simfac ] ) all occur in @xmath1136 . by ( [ ressimple ] )
we have that @xmath1133 contains @xmath1137 and @xmath1134 contains @xmath1138 .
( in fact , all factors can be easily determined . ) comparing with ( [ 9upper ] ) we see that both @xmath1126 and @xmath1139 must occur above the simple whose restrictions contribute the terms in ( [ 10simfac ] ) .
thus @xmath1121 must occur and forms the socle of @xmath1125 .
this completes our verification of the various claims above , and so the block structure is as shown in figure [ case10 ] .
we claim that the structure of the block containing @xmath240 is given by the data in figure [ case11 ] .
the structure of the modules @xmath1141 , @xmath1142 , @xmath1143 and @xmath1144 follows as for @xmath1112 for partitions of @xmath10 and @xmath34 .
the modules @xmath1145 and @xmath1124 are projective .
therefore by proposition [ projelim ] the simple @xmath1146 can not occur in any of the remaining standards , and @xmath1147 can not occur in @xmath1148 or @xmath1149 .
the structure of these latter two modules now follows by localisation to the case @xmath1085 . to see that the structure of @xmath1150 is as illustrated follows from ( [ squarestuff ] ) and localisation . the only remaining module is @xmath1152 . by localising this contains @xmath1153 , @xmath1154 and @xmath1155 , all with multiplicity .
we have eliminated all other possible factors except @xmath1147 .
we proceed as for the module @xmath1125 above .
restriction of @xmath1152 contains @xmath1156 with multiplicity one , and this can only arise in the restriction of @xmath1157 . arguing as in the @xmath1112 case we also see that @xmath1157 must coincide with the socle of @xmath1152 , and so we are done .
we claim that the structure of the block containing @xmath1113 is given by the data in figure [ case12 ] . as usual ,
the structure of the modules labelled by partitions of @xmath10 and @xmath34 is straightforward .
the simple @xmath1159 can not occur anywhere else as @xmath1160 is projective .
the modules @xmath1161 and @xmath1162 are both projective
. therefore @xmath1163 can not occur in any standard labelled by a partition of @xmath1086 or smaller , while @xmath1164 can not occur in @xmath1165 .
the structure of @xmath1165 is then clear by localisation .
next consider @xmath1166 .
this can not contain @xmath1163 as @xmath1167 , so it is enough by localisation to verify that @xmath1164 can not occur .
but this is clear , as the restriction of this simple contains @xmath1168 , which is not in the same block as any of the standard modules in the restriction of @xmath1166 .
it remains to determine the structure of @xmath1169 and @xmath1170 . in each case we know the multiplicity of all composition factors by localisation and the remarks above , except for @xmath1164 .
using proposition [ largerhom ] we see that we have a non - zero homomorphism from @xmath1171 into each of the two standards , and so the multiplicity of @xmath1164 in each case is at least one .
the restriction of each of these standards contains precisely one copy of @xmath1172 and @xmath1168 , and so @xmath1164 must occur with multiplicity one in each standard . to determine the location of @xmath1164 in @xmath1169 note that it can not occur below any composition factor other than @xmath1173 , as this would contradict the existence of homomorphisms from @xmath1174 and @xmath1166 into @xmath1169 .
therefore the structure of this module must be as shown . for @xmath1170 , we have that the simple @xmath1175 in the restriction of @xmath1173 occurs above @xmath1176 , and hence the structure of @xmath1170 must be as shown .
suppose that @xmath135 is a weyl group , with associated affine weyl group @xmath1177 .
soergel has shown @xcite that ( provided @xmath11 is not too small ) the value of the parabolic kazhdan - lusztig polynomials @xmath1178 ( evaluated at @xmath1179 ) associated to @xmath1180 determine the multiplicity of the standard module @xmath1181 in the indecomposable @xmath1182 for the quantum group @xmath1183 associated to @xmath135 where @xmath18 is a @xmath11th root of unity . in the case of the quantum general linear group ,
ringel duality @xcite translates this into a result about decomposition numbers for the hecke algebra of type @xmath13 , where now @xmath72 labels a simple module and @xmath51 a specht module .
further , rouquier has conjectured that the coefficient of @xmath1184 occurring in @xmath1178 should correspond to the multiplicity of the simple @xmath1185 in the @xmath41th layer of the jantzen filtration of the specht module @xmath53 . in this spirit
, we can compare our results in this section with the polynomials in figure [ klcalc ] for @xmath1186 .
we see that in each case , the value of @xmath1187 from figure [ klcalc ] is exactly the multiplicity of @xmath84 in @xmath37 , and that there is a filtration of @xmath1188 corresponding to the powers of @xmath43 occurring in the polynomials for the @xmath84s .
this , together with the other lie - like phenomena we have observed leads us to ask \(i ) for the brauer algebra with kazhdan - lusztig polynomials as defined in section [ klpoly ] , is it true for weights in an alcove that @xmath828=n_{\lambda,\mu}(1)\ ? \ ] ] ( ii ) is there a ( jantzen ? )
filtration of @xmath37 such that the multiplicity of a simple @xmath84 in the @xmath41th layer is given by the coefficient of @xmath1184 in @xmath1017 ?
h. h. andersen , j. c. jantzen , and w. soergel , _ representations of quantum groups at a @xmath11-th root of unity and of semisimple groups in characteristic @xmath11 : independence of @xmath11 _ , astrisque * 220 * ( 1994 ) . | there are similarities between algebraic lie theory and a geometric description of the blocks of the brauer algebra in characteristic zero .
motivated by this , we study the alcove geometry of a certain reflection group action . we provide analogues of translation functors for a tower of recollement , and use these to construct morita equivalences between blocks containing weights in the same facet .
moreover , we show that the determination of decomposition numbers for the brauer algebra in characteristic zero can be reduced to a study of the block containing the weight @xmath0 .
we define parabolic kazhdan - lusztig polynomials for the brauer algebra and show in certain low rank examples that they determine standard module decomposition numbers and filtrations
. [ section ] [ thm]proposition [ thm]lemma [ thm]corollary [ thm]conjecture [ thm]question [ thm]claim [ thm]remark [ thm]assumption [ thm]definition [ thm]example |
let @xmath0 be a countable discrete group and let @xmath9 denote the hilbert space of square - summable functions @xmath10 .
let @xmath11 denote the algebra of bounded linear operators on @xmath9 .
note that elements of @xmath0 act on @xmath9 as bounded linear operators , so there is a natural map @xmath12 \to b(\ell^2(g))$ ] , where @xmath12 $ ] denotes the group algebra of @xmath0 .
we define the reduced group @xmath13-algebra , denoted @xmath14 , to be the operator norm closure of @xmath12 $ ] in @xmath11 .
the reduced group @xmath13 algebra plays an important role in noncommutative geometry , and these algebras are studied in detail in the book @xcite .
a trace on @xmath14 is a bounded linear map @xmath15 that satisfies @xmath16 for @xmath17 .
it is easy to see that such a map satisfies that for @xmath18 , we have @xmath19 , and also that if we had a bounded linear map @xmath15 @xmath19 , then it is a trace . in @xcite ,
weinberger and yu studied the degree of non - rigidity of manifolds using idempotents in @xmath13 algebras of groups associated to torsion elements in those groups , and exploring whether they are linearly independent in @xmath20-theory .
one approach to this is to study whether idempotents can be shown to be linearly independent using traces , as explored in @xcite .
the study of idempotents also leads to results in the study of the moduli space of positive scalar curvature metrics , as in @xcite and @xcite .
the question of studying the traces on reduced group @xmath13 algebras was introduced by powers in @xcite , who proved that up to scalar multiplication , there is only one trace on @xmath14 for non - abelian free groups @xmath0 .
it was also studied extensivesly in @xcite , @xcite , and @xcite , where it was shown that if @xmath0 is a torsion - free non - elementary hyperblic group , then @xmath14 has a only one trace , up to scalar .
similar questions were also studied in @xcite for relatively hyperbolic groups . in this paper , we introduce a new method of studying traces using property rd . recall that a length function on a group @xmath0 is a function @xmath21 , such that @xmath22 , @xmath23 , and @xmath24 , such that for @xmath25 finite @xmath26 is also finite . given such a length function on a group , there is a norm @xmath27 on @xmath28 given by @xmath29 let @xmath30 denote the completion of @xmath28 with respect to this norm .
then we say @xmath0 has rapid decay property , or property rd if for some @xmath31 , there is a length function on it and a constant @xmath32 such that @xmath33 this property was introduced by connes and studied extensively by jolissaint in @xcite and @xcite . for a countable discrete group @xmath0 with length function @xmath34 and a subset @xmath35 , we say @xmath36 has polynomial growth
if for @xmath37 and @xmath38 , we have @xmath39 for some polynomial @xmath40 . if such a @xmath40 does not exist , we say that @xmath36 grows super - polynomially .
we say @xmath36 grows at least exponentially if there are @xmath41 with @xmath42 such that @xmath43 . in @xcite
, i showed that for groups with property rd , if the conjugacy class of an element @xmath44 grows super - polynomially , then any trace @xmath15 vanishes on @xmath44 .
we will explain this result in more detail in section [ on traces ] .
we will use this method to classify all traces on all non - elementary hyperbolic groups , extending de la harpe s result on torsion - free non - elementary hyperbolic groups .
the author would like to thank prof .
guoliang yu for his helpful advice and guidance , and prof .
david kerr for his suggestion to look into questions regarding unique traces on reduced group @xmath13 algebras .
in this section , we show that if a conjugacy class in a hyperbolic group is infinite , then it grows exponentially . in the next section
, we will use this to find all the traces .
let @xmath0 be a finitely generate discrete group and @xmath45 a real number .
consider the metric space obtained by considering the cayley graph of @xmath0 , and identifying each edge with the unit interval .
then , we say @xmath0 is @xmath46-hyperbolic if for any @xmath47 ,
@xmath48 \subset b_{\delta}([y , z ] \cup [ z , x])\ ] ] where @xmath49 $ ] denotes the geodesic between @xmath50 and @xmath51 in the metric space , and for a subset @xmath35 , @xmath52 denotes points @xmath2 such that there is @xmath53 with @xmath54 .
we say that a group @xmath0 is elementary if @xmath0 contains @xmath55 as a finite index subgroup .
we study the problem of conjugacy class growth in non - elementary hyperbolic groups .
let @xmath0 be a non - elementary hyperbolic group and let @xmath1 be an element .
suppose that @xmath56 , the conjugacy class of @xmath2 in @xmath0 is infinite
. then @xmath56 grows at least exponentially .
[ exp conj class ] notation : for @xmath57 a non - negative real number , we let @xmath58 denote the ball in the cayley graph of @xmath0 around the identity of radius @xmath57 . to show this theorem , we want to consider some element in @xmath59 of length @xmath60 , where @xmath60 is large , and consider conjugating it to @xmath61 where @xmath62 .
the number of @xmath8 with @xmath62 grows exponentially in @xmath60 , and @xmath63 is then in the range @xmath64 $ ] so if all different @xmath8 in @xmath62 gave rise to different @xmath61 , we would be done , because we would have shown that the conjugacy class of @xmath2 has exponentially many elements in the range .
however , this is not quite true , because there could be @xmath8 and @xmath65 with @xmath66 , but so that @xmath67 .
however , we will show that this can not happen too much .
more precisely , we prove the following lemma : let @xmath0 be a non - elementary hyperbolic group and @xmath1 be a fixed element .
let @xmath50 be in @xmath56 , and let @xmath68 .
let @xmath69 have @xmath70 .
let @xmath71 .
let @xmath72 be an integer .
then the number of @xmath65 in @xmath0 such that 1 .
@xmath73 2 .
@xmath74 is @xmath75 , meaning it is bounded above by @xmath76 where @xmath32 and @xmath77 are constants that depend only on @xmath0 and @xmath2 , and independent of @xmath60 and @xmath78 .
thus , the number of elements @xmath79 such that 1 .
@xmath80 2 .
@xmath74 is @xmath81 , meaning it is bounded by @xmath82 . in triangle @xmath83 in a graph ,
let @xmath41 , denote the sides ; that is , @xmath84 , @xmath85 and @xmath86 .
we say that triangle @xmath83 is @xmath87-thin if for points @xmath88 , @xmath89 and @xmath90 with @xmath91 , @xmath92 and @xmath93 , we have @xmath94 . proposition 2.12 in @xcite
says that for a @xmath46-hyperbolic graph , there is @xmath87 such that all triangles are @xmath87-thin .
let us replace @xmath46 with the max of @xmath46 and @xmath87 so that now the graph is @xmath46-hyperbolic and all triangles are @xmath46-thin .
let us assume @xmath95 , which we may , because there are finitely many cases for @xmath96 , so we can shove those cases into the constant @xmath77 .
consider the geodesic picture shown in the diagram , where @xmath97 , @xmath98 , @xmath99 and @xmath100 , and @xmath101 , @xmath102 , @xmath103 , and @xmath104 are geodesics .
thus , @xmath102 is composed of the same sequence of @xmath105 edges as the geodesic for @xmath50 would be , and @xmath103 the same as the geodesic for @xmath65 .
( in my picture the cayley graph is formed by taking generators @xmath106 and joining @xmath8 to @xmath107 for @xmath108 . thus , my picture is invariant under the right action of @xmath0 . )
now consider triangle @xmath83 .
applying that it is @xmath46-thin , we have @xmath109 , @xmath110 and @xmath111 such that @xmath112 , and @xmath113 and @xmath114 .
now , let us apply hyperbolicity to triange @xmath115 .
we have @xmath111 so there is @xmath116 such that @xmath117 .
we claim that @xmath118 .
if not , @xmath119 , but then @xmath120 and @xmath121 , and @xmath122 , so @xmath123 , but @xmath124 , so @xmath125 , so @xmath126 , which is impossible because @xmath127 .
thus @xmath118 .
so we have @xmath118 with @xmath128 .
also , @xmath114 , so @xmath129 @xmath130
so now for fixed @xmath50 and @xmath51 , and for @xmath131 , @xmath132 , we have fixed @xmath110 , because @xmath133 has @xmath134 , and @xmath102 is just the @xmath135 right translate of the geodesic for @xmath50 , so @xmath136 is fixed , because it is just the product of the first @xmath137 edges of the geodesic of @xmath50 .
now , @xmath138 , as a group element , has distance at most @xmath46 from @xmath136 , so there are @xmath139 possibilities for @xmath138 . for @xmath140 fixed , as above , there are @xmath141 possibilities for @xmath142 , because @xmath143 is along @xmath104 , which is the geodesic of @xmath51 , and we showed above that @xmath144 .
so @xmath142 is the product of the first @xmath57 edges of the geodesic of @xmath51 .
for each of the @xmath145 choices of @xmath142 , there are at most @xmath146 choices of @xmath147 , because @xmath147 and @xmath142 are within @xmath145 of each other as group elements .
thus , in total , for fixed @xmath140 , there are @xmath148 possibilities for @xmath147 .
so we have , for fixed @xmath149 , there are @xmath150 choices of @xmath147 and @xmath148 choices of @xmath138 so there are @xmath151 choices of @xmath152 .
but @xmath153 , so there are at most @xmath75 choices of @xmath65 for a given @xmath78 and @xmath140 . there are @xmath78 choices of @xmath154 $ ] , so in total there are at most @xmath155 possible @xmath65 for a given @xmath78 , giving us the first statement in the lemma .
adding up the values for different @xmath156 is adding up @xmath157 summands , each of which is @xmath158 , so we get that there are @xmath81 possible @xmath65 with @xmath159 and @xmath74 .
now we complete the proof of the theorem .
we are assuming that the conjugacy class of @xmath2 is infinite , so there are @xmath59 of arbitrarily large length . for large @xmath60 , consider some @xmath59 with @xmath160 .
then for @xmath161 , we may consider @xmath162 .
then @xmath163 , so we can map @xmath164 to the elements of the conjugacy class of @xmath2 of length in the range @xmath64 $ ] .
@xmath165 by the above lemma , the pre - image of any element under @xmath166 has size at most @xmath82 , for constants @xmath167 depending only on @xmath2 and @xmath0 . but @xmath168 grows exponentially by a theorem in koubi ( also attached ) , that says that non - elementary hyperbolic groups have exponential growth .
so the number of elements in the image of @xmath166 , grows at least with @xmath169 , which is still exponential .
thus the conjugacy class of @xmath170 grows exponentially in @xmath60 , so @xmath56 has exponential growth , as desired . for @xmath0 a non - elementary hyperbolic group and @xmath1 a non - torsion element
, the conjugacy class of @xmath2 grows exponentially . by the theorem , it suffices to show that the conjugacy class is infinite .
let @xmath56 be the conjugacy class and @xmath171 be the centraliser .
then @xmath172 , so if the conjugacy class is finite , then @xmath173 has finite index .
but , by theorem 3.35 , @xmath171 is a finite extension of @xmath174 , so it is virtually cyclic . but
any group is quasi - isometric to any finite index subgroup , so that would mean that @xmath0 is also virtually cyclic , which is impossible .
in this section , we show the main result ; namely , we classify all traces on non - elementary word - hyperbolic groups .
let @xmath0 be a non - elementary word - hyperbolic group .
then the traces @xmath4 are exactly the linear combinations of @xmath5 given by @xmath6 where @xmath8 is an element of finite conjugacy class and @xmath7 is its conjugacy class .
[ all traces ] as was shown in @xcite [ theorem 3.27 ] , in a hyperbolic group , there are only finitely many conjugacy classes of torsion elements
. we will show that all infinite order elements have infinite conjugacy class .
suppose , for the sake of contradiction that there is @xmath44 an element of infinite order with finite conjugacy class @xmath7 .
then @xmath175 , where @xmath176 is the centraliser of @xmath0 . however , it was shown in @xcite [ theorem 3.35 ] that @xmath176 is a finite extension of the subgroup generated by @xmath8 , so that it is elementary .
but then if @xmath7 is finite , then @xmath0 would also be elementary , a contradiction .
so we see that there are only finitely many elements that have finite conjugacy class . for each such element @xmath8 , it is easy to see that the map @xmath6 defines a bounded linear functional @xmath177 .
any linear combination of these finitely many traces is also a bounded trace . to do this
, we will use the fact that word hyperbolic groups have property rd : given such a length function on a group , there is a norm @xmath27 on @xmath28 given by @xmath29 let @xmath30 denote the completion of @xmath28 with respect to this norm .
then we say @xmath0 has rapid decay property , or property rd if for some @xmath31 , there is a length function on it and a constant @xmath32 such that @xmath33 it was shown in @xcite that all hyperbolic groups have property rd .
now , we will apply lemma 2.8 in @xcite , which states the following : let @xmath0 be a group with property rd and @xmath44 be an element whose conjugacy class grows faster than polynomially .
then all traces @xmath178 vanish on @xmath8 .
in our case , word - hyperbolic groups have property rd , so applying theorem [ exp conj class ] , we have that all traces
@xmath178 vanish on all elements with infinite conjugacy class , and they must be linear combinations of the above traces , as desired . in particular , this gives another proof of de la harpe s result regarding traces on torsion - free non - elementary word - hyperbolic groups , which can be found in @xcite : let @xmath0 be a torsion - free non - elementary word - hyperbolic group .
then there is only one trace @xmath15 , up to scalar multiples .
theorem [ all traces ] also implies that every hyperbolic group can be written as an extension of a hyperbolic group with unique trace by a finite group , namely : let @xmath0 be a hyperbolic group .
then it has a finite normal subgroup @xmath179 such that @xmath180 is a hyperbolic group with unique trace , up to scalar multiples .
[ normal subgp ] let @xmath181 be the subset of elements with finite conjugacy class .
note that all elements in it have finite order , as we showed in the proof of theorem [ all traces ] .
we claim that @xmath179 is a subgroup .
it is easy to see that @xmath179 is closed under inversion , so we want that if @xmath182 then @xmath183 . for an element @xmath44 ,
let @xmath184 be the subgroup consisting of elements that commute with @xmath8 .
then , note that @xmath185 if and only if @xmath176 has finite index as a subgroup of @xmath0 .
thus , it suffices to show that if @xmath186 and @xmath187 have finite index , then @xmath188 does also .
note that @xmath189 , so it suffices to show that if @xmath190 are finite index subgroups , then @xmath191 is also .
there is a natural injection @xmath192 .
thus , @xmath193 . let @xmath194 and @xmath195 then @xmath196 , so @xmath191 has finite index , as desired . by the lemma
, we have that @xmath179 is a subgroup of @xmath0 . as we showed in the proof of theorem
[ all traces ] , @xmath179 is finite . also , it is easy to see that @xmath179 is closed under conjugation . thus , @xmath179 is a finite normal subgroup .
note that @xmath180 is hyperbolic because hyperbolicity is preserved under quasi - isometries , and it is easy to see that the map @xmath197 is a quasi isometry .
thus , @xmath180 is hyperbolic .
it remains to show that @xmath180 has unique trace @xmath198 .
such a trace @xmath199 induces a trace @xmath177 , because @xmath179 is a finite normal subgroup . by theorem [ all traces ] , the trace @xmath177 vanishes on all elements outside of @xmath179 , so the map @xmath200 vanishes on all elements except the identity , as desired .
in this section , we give some examples of hyperbolic groups with torsion that have unique trace by theorem [ all traces ] .
let @xmath201 be the free product of two non - trivial hyperbolic groups @xmath202 and @xmath203 .
suppose that @xmath204 .
this assumption does not lose any generality , because @xmath205 is elementary .
then @xmath0 is hyperbolic and all its conjugacy classes are infinite .
it is hyperbolic because the free product of two hyperbolic groups is hyperbolic . to see that the conjugacy classes have infinite order : note that every element @xmath206 can be uniquely expressed as @xmath207 where @xmath208 , @xmath209 and @xmath210 for @xmath211 and @xmath212 for @xmath213 . in the case @xmath214 and @xmath215 , then consider some element @xmath216 and @xmath217 , and consider @xmath218 , that is conjugation of @xmath50 by @xmath219 . as @xmath57 varies amongst positive integers
, @xmath220 goes through different elements of @xmath0 , all in the conjugacy class of @xmath50 . for the case
@xmath223 , and @xmath215 , conjugate by @xmath219 where @xmath224 has @xmath225 and @xmath226 .
then the same argument as the above applies .
similarly , for the case @xmath227 and @xmath228 , pick @xmath224 , with @xmath225 and @xmath229 .
this example shows that many hyperbolic groups with torsion elements still have unique traces , extending de la harpe s results in @xcite , @xcite , and @xcite that torsion - free non - elementary hyperbolic groups have unique traces .
now let us give an example of a non - elementary hyperbolic group that does not have unique trace ( up to scalar multiples ) .
for this , by theorem [ all traces ] , it suffices that there are non - identity elements with finite conjugacy class .
let @xmath0 by a non - elementary hyperbolic group and @xmath202 be a non - trivial finite group .
then in @xmath230 , the direct product of @xmath202 and @xmath0 , is hyperbolic , because it is quasi - isometric to @xmath0 . the elements in @xmath231 , where @xmath232 is the identity in @xmath0 , have finite conjugacy class .
let @xmath233 .
this group @xmath0 is hyperbolic , because it is quasi - isometric to the free group on two generators , @xmath234 . in @xmath0 , @xmath2 and @xmath235
clearly have finite conjugacy class , as their conjugacy class is @xmath236 .
thus , by theorem [ all traces ] , @xmath0 does not have unique trace .
however , it is not the direct product of a finite group and a hyperbolic group whose reduced @xmath13 algebra has unique trace up to scalar multiples , because the only elements of finite order in @xmath0 are @xmath2 and @xmath237 , so if @xmath0 were such a direct product , it would be @xmath238 for some hyperbolic group @xmath202 .
but in that case , @xmath2 would not be conjugate to @xmath237 .
z. xie and g. yu .
_ a relative higher index theorem , diffeomorphisms and positive scalar curvature _ , advances in mathematics , 250 ( 2014 ) 35-73 .
z. xie and g. yu .
_ higher rho invariants and the moduli space of positive scalar curvature metrics _ , arxiv:1310.1136 , 2012 | in this paper , we show that if @xmath0 by a non - elementary word hyperbolic group and @xmath1 an element , if the conjugacy class of @xmath2 is infinite , then all traces @xmath3 vanish on @xmath2 .
we show that all traces @xmath4 are linear combinations of traces @xmath5 given by @xmath6 where @xmath7 is the conjugacy class of @xmath8 . to do this ,
we introduce a new method to study traces , using sobolev norms and property rd . |
in their influential paper the geometry of graphs and its algorithmic applications " , linial et al . @xcite introduce a novel and powerful set of techniques to the algorithm designer s toolkit .
they show how to use the mathematics of metric embeddings to help solve difficult problems in combinatorial optimization .
the approach inspired a large body of further work on metric embeddings and their applications .
our objective here is to show how this extensive body of work might be generalized to the _ geometry of hypergraphs_. recall that a _ hypergraph _
@xmath1 consists of a set of vertices @xmath2 and a set of _ hyperedges _ @xmath3 , where each @xmath4 is a subset of @xmath2 .
the underlying geometric objects in this new context will not be metric spaces , but _ diversities _
, a generalization of metrics recently introduced by bryant and tupper @xcite .
diversities are a form of multi - way metric which have already given rise to a substantial , and novel , body of theory @xcite .
we hope to demonstrate that a switch to diversities opens up a whole new array of problems and potential applications , potentially richer than that for metrics .
the result of @xcite which is of particular significance to us is the use of metric embeddings to bound the difference between cuts and flows in a _ multi - commodity flow _ problem .
let @xmath5 be a graph with a non - negative edge capacity @xmath6 for every edge @xmath7 .
we are given a set of _ demands _
@xmath8 for @xmath9 .
the objective of the multi - commodity flow problem is to find the largest value of @xmath10 such that we can simultaneously flow at least @xmath11 units between @xmath12 and @xmath13 for all @xmath12 and @xmath13 . as usual
, the total amount of flow along an edge can not exceed its capacity .
multi - commodity flow is a linear programming problem ( lp ) and can be solved in polynomial time .
the _ dual _ of the lp is a relaxation of a min - cut problem which generalizes several np - hard graph partition problems .
given @xmath14 let @xmath15 be the sum of edge capacities of edges joining @xmath16 and @xmath17 and let @xmath18 denote the sum of the demands for pairs @xmath19 with @xmath20 and @xmath21 .
we then have @xmath22 for every @xmath14 .
when there is a single demand , the minimum of @xmath23 equals the maximum value of @xmath10 , a consequence of the max - flow min - cut theorem . in general , for more than one demand there will be a gap between the values of the minimum cut and the maximum flow .
linial et al @xcite , building on the work of @xcite , show that this gap can be bounded by the _ distortion _ required to embed a particular metric @xmath24 ( arising from the lp dual ) into @xmath0 space .
the metric @xmath24 is _ supported _ on the graph @xmath25 , meaning that it is the shortest path metric for some weighting of the edges @xmath3 . by applying the extensive literature on distortion bounds for metric embeddings
they obtain new approximation bounds for the min - cut problem . in this paper
we consider generalizations of the multi - commodity flow and corresponding minimum cut problems .
a natural generalization of the single - commodity maximum flow problem in a graph is _ fractional steiner tree packing _
@xcite . given a graph @xmath5 with weighted edges , and a subset @xmath14 ,
find the maximum total weight of trees in @xmath26 spanning @xmath16 such that the sum of the weights of trees containing an edge does not exceed the capacity of that edge .
whereas multi - commodity flows are typically used to model transport of physical substances ( or vehicles ) , the steiner tree packing problem arises from models of information , particularly the broadcasting of information ( see @xcite for references ) .
the fractional steiner tree packing problem generalizes further to incorporate multiple commodities , a formulation which occurs naturally in multicast and vlsi design applications ( see @xcite ) . for each @xmath14
we have a demand @xmath27 ( possibly zero ) and the set @xmath28 of trees in the graph spanning @xmath16 .
a _ generalized flow _ in this context is an assignment of non - negative weights @xmath29 to the trees in @xmath28 for all @xmath16 , with the constraint that for each edge , the total weight of trees including that edge does not exceed the edge s capacity .
the objective is to find the largest value of @xmath10 for which there is a flow with weights @xmath29 satisfying @xmath30 for all demand sets @xmath16 .
these problems translate directly to hypergraphs , permitting far more complex relationships between the different capacity constraints . as for graphs , we have demands @xmath27 defined for all @xmath14 .
each hyperedge @xmath4 has a non - negative capacity .
we let @xmath28 denote the set of all _ minimal connected sub - hypergraphs _ which include @xmath16 ( not necessarily trees ) .
a flow in this context is an assignment of non - negative weights @xmath29 to the trees in @xmath28 for all @xmath16 , with the constraint that for each hyperedge , the total weight of trees including that hyperedge does not exceed the hyperedge s capacity . as in the graph case ,
the aim is determine the largest value of @xmath10 for which there is a flow with weights @xmath29 satisfying the constraint for all demand sets @xmath16 .
all of these generalizations of the multi - commodity flow problem have a dual problem that is a relaxation of a corresponding min - cut problem . for convenience , we assume any missing edges or hyperedges are included with capacity zero . for
a subset @xmath31 let @xmath32 be the set of edges or hyperedges which have endpoints in both @xmath33 and @xmath34 .
the min - cut problem in the case of graphs is to find the cut @xmath33 minimizing @xmath35 where @xmath36 runs over all pairs of distinct vertices in @xmath2 , while in hypergraphs we find @xmath33 which minimizes @xmath37 where @xmath38 run over all subsets of @xmath2 . in both problems
the value of a min - cut is an upper bound for corresponding value of the maximum flow .
linial et al .
@xcite showed that the ratio between the min - cut and the max flow can be bounded using metric embeddings .
our main result is that this relationship generalizes to the fractional steiner problem with multiple demand sets , on both graphs and hypergraphs , once we consider diversities instead of metrics .
the following theorems depend on the notions of diversities being supported on hypergraphs and @xmath0-embeddings of diversities , which we will define in subsequent sections .
[ major1 ] let @xmath39 be a hypergraph .
let @xmath40 be a set of edge capacities and @xmath41 a set of demands .
there is a diversity @xmath42 supported on @xmath43 , such that the ratio of the min - cut to the maximum ( generalized ) flow for the hypergraph is bounded by the minimum distortion embedding of @xmath44 into @xmath0 .
gupta et al .
@xcite proved a converse of the result of linial et al . by showing that , given any graph @xmath26 and metric @xmath24 supported on it
, we could determine capacities and demands so that the bound given by the minimal distortion embedding of @xmath24 into @xmath0 was tight . we establish the analogous result for the generalized flow problem in hypergraphs .
[ major2 ] let @xmath39 be a hypergraph , and let @xmath44 be a diversity supported on it .
there is a set @xmath40 of edge capacities and a set @xmath41 of demands so that the ratio of the min - cut to the maximum ( generalized ) flow equals the distortion of the minimum distortion embedding of @xmath44 into @xmath0 .
a major benefit of the link between min - cut and metric embeddings was that linial et al . and
others could make use of an extensive body of work on metric geometry to establish improved approximation bounds . in our context , the embeddings of diversities is an area which is almost completely unexplored .
we prove a few preliminary bounds here , though much work remains .
+ the structure of this paper is as follows .
we begin in section 2 with a brief review of diversity theory , including a list of examples of diversities . in section 3
we focus on @xmath45 and @xmath0 diversities , which are the generalizations of @xmath45 and @xmath0 metrics .
these diversities arise in a variety of different contexts .
fundamental properties of @xmath45 diversities are established , many of which closely parallel results on metrics . in section 4
we show how the concepts of metric embedding and distortion are defined for diversities , and establish a range of preliminary bounds for distortion and dimension . finally ,
in section 5 , we prove the analogues of linial et al s @xcite and gupta et al s @xcite results on multi - commodity flows , as stated in theorems [ major1 ] and [ major2 ] above .
a _ diversity _ is a pair @xmath46 where @xmath47 is a set and @xmath44 is a function from the finite subsets of @xmath47 to @xmath48 satisfying _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ( d1 ) @xmath49 , and @xmath50 if and only if @xmath51 .
+ ( d2 ) if @xmath52 then @xmath53 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for all finite @xmath54 .
diversities are , in a sense , an extension of the metric concept .
indeed , every diversity has an _ induced metric _ , given by @xmath55 for all @xmath56 .
note also that @xmath44 is _ monotonic _ : @xmath57 implies @xmath58 . for convenience ,
in the remainder of the paper we will relax condition ( d1 ) and allow @xmath50 even when @xmath59 .
likewise , for metrics we allow @xmath60 even if @xmath61 .
we define embeddings and distortion for diversities in the same way as for metric spaces .
let @xmath62 and @xmath63 be two diversities and suppose @xmath64 . a map @xmath65 has _ distortion _
@xmath66 if there is @xmath67 such that @xmath68 and @xmath69 for all finite @xmath70 .
we say that @xmath71 is an _
isometric embedding _ if it has distortion @xmath72 and an _ approximate embedding _ otherwise .
bryant and tupper @xcite provide several examples of diversities .
we expand that list here . 1 .
_ diameter diversity .
_ let @xmath73 be a metric space .
for all finite @xmath74 let @xmath75 2 . _
@xmath0 diversity .
_ let @xmath76 denote the diversity @xmath77 , where @xmath78 for all finite @xmath79 .
@xmath45 diversity .
_ let @xmath80 be a measure space and let @xmath45 denote the set of all all measurable functions @xmath81 with @xmath82 .
an @xmath45 diversity is a pair @xmath83 where @xmath84 is given by @xmath85 for all finite @xmath86 . to see that @xmath83 satisfies ( d2 ) , consider the triangle inequality for the diameter diversity on a real line and integrate over @xmath87 .
phylogenetic diversity .
_ let @xmath88 be a phylogenetic tree with taxon set @xmath47 .
for each finite @xmath74 , the _ phylogenetic diversity _ of @xmath89 is the length @xmath90 of the smallest subtree of @xmath88 connecting taxa in @xmath89 @xcite .
steiner diversity .
_ let @xmath73 be a metric space .
for each finite @xmath74 let @xmath90 denote the minimum length of a steiner tree connecting elements in @xmath89 .
_ hypergraph steiner diversity .
_ let @xmath91 be a hypergraph and let @xmath92 be a non - negative weight function .
given @xmath74 let @xmath90 denote the minimum of @xmath93 over all subsets @xmath94 such that the sub - hypergraph induced by @xmath95 is connected and includes @xmath89 . then @xmath46 is a diversity .
_ measure diversity .
_ let @xmath96 be a measure space , where @xmath97 is a @xmath98-algebra of subsets of @xmath99 and @xmath100 $ ] .
let @xmath47 be the collection of all sets in @xmath97 with finite measure . for sets
@xmath101 we let @xmath102 8 .
_ smallest enclosing ball diversity_. let @xmath73 be a metric space . for each finite @xmath74
let @xmath90 be the diameter of the smallest closed ball containing @xmath89 .
note that if that every pair of points in @xmath73 are connected by a geodesic then @xmath73 will be the induced metric of @xmath46 , though this does not hold in general .
9 . _ travelling salesman diversity .
_ let @xmath73 be a metric space .
for every finite @xmath74 , let @xmath90 be the minimum of @xmath103 over all orderings @xmath104 of @xmath89 .
_ mean - width diversity_. we define the _ mean - width diversity _ for finite @xmath105 as the mean width of @xmath106 , the convex hull of @xmath89 , suitably scaled .
specifically , given a compact convex set @xmath107 and unit vector @xmath108 , the width of @xmath109 in direction @xmath12 is given by @xmath110 that is , @xmath111 is the minimum distance between two hyperplanes with normal @xmath12 which enclose @xmath109 .
mean width _ of @xmath109 is given by @xmath112 where @xmath113 denotes the surface measure on the unit sphere @xmath114 @xcite .
shephard @xcite observed that the mean width varies according to the space that @xmath109 is sitting in , whereas a scaled _ absolute mean width _
@xmath115 does not .
a simple calculation gives that for points @xmath116 in @xmath48 we have @xmath117 ) = \frac{1}{\pi } |a - b|$ ] .
hence we define the _ mean - width diversity _
@xmath118 so that the induced metric of @xmath119 is the euclidean metric . here
@xmath120 is the beta function .
note that @xmath121 , see @xcite .
_ @xmath16-diversity_. let @xmath47 be a collection of random variables taking values in the same state space .
for every finite @xmath122 let @xmath90 be the probability that @xmath123 do not all have the same state .
then @xmath46 is a diversity , termed the _ @xmath16-diversity _ since @xmath16 , the proportion of segregating ( non - constant ) sites , is a standard measure of genetic diversity in an alignment of genetic sequences ( see , e.g. @xcite ) .
below , we will show that @xmath0 diversities , phylogenetic diversities , measure diversities , mean - width diversities and @xmath16-diversities are all examples of @xmath45-embeddable diversities .
in metric geometry we say that one metric _ dominates _ another on the same set if distances under the first metric are all greater than , or equal to , distances under the second .
the relation forms a partial order on the cone of metrics for a set : given any two metric spaces @xmath124 and @xmath125 we write @xmath126 if @xmath127 for all @xmath128 .
the partial order @xmath129 provides a particularly useful characterization of the standard shortest - path graph metric @xmath130 .
let @xmath131 be a graph with edge weights @xmath132 .
the shortest path metric @xmath130 is then the unique , maximal metric ( under @xmath129 ) which satisfies @xmath133 for all @xmath7 . given that the _ geometry of graphs _ of @xcite is based on the shortest path metric , it is natural to explore what arises when we apply the same approach to diversities .
we say that a diversity @xmath134 dominates another diversity @xmath135 if @xmath136 for all finite @xmath74 , in which case we write @xmath137 .
applying these to graphs , and hypergraphs , we obtain the diversity analogue to the shortest - path metric .
[ thm : steinmax ] 1 .
let @xmath131 be a graph with non - negative weight function @xmath138 .
the steiner tree diversity is the unique maximal diversity @xmath44 such that @xmath139 for all @xmath7 .
2 . let @xmath39 be a hypergraph with non - negative weight function @xmath92 .
the hypergraph steiner diversity is the unique maximal diversity @xmath44 such that @xmath140 for all @xmath4 .
note that 1
. is a special case of 2 .
we prove 2 .
let @xmath141 denote the hypergraph steiner diversity for @xmath43 . for any edge @xmath89 , the edge itself forms a connected sub - hypergraph , so @xmath142 .
let @xmath44 be any other diversity which also satisfies @xmath140 for all @xmath4 .
for all @xmath143 there is @xmath94 such that the sub - hypergraph induced by @xmath95 is connected , contains @xmath144 , and has summed weight @xmath145 .
multiple applications of the triangle inequality ( d2 ) gives @xmath146 as a further consequence , we can show that the hypergraph steiner diversity dominates all diversities with a given induced metric .
[ divbounds ] let @xmath46 be a diversity with induced metric space @xmath73 .
let @xmath147 denote the diameter diversity on @xmath47 and let @xmath148 denote the steiner diversity on @xmath47 .
then for all finite @xmath74 , @xmath149 if @xmath150 then @xmath151
. suppose @xmath152 .
there is @xmath153 such that @xmath154 the last inequality following from the monotonicity of @xmath44 .
let @xmath26 be the complete graph with vertex set @xmath89 and edge weights @xmath155 .
then @xmath156 by theorem [ thm : steinmax ] . to obtain the final inequality ,
consider any ordering of the elements of @xmath89 : @xmath157 .
then , using the triangle inequality repeatedly gives @xmath158
@xmath45 diversities were defined in section [ sec : diversitylist ] .
we say that a diversity @xmath46 is @xmath45-embeddable if there exists an isometric embedding of @xmath46 into an @xmath45 diversity . a direct consequence of the definition of @xmath45 diversities ( and the direct sum of measure spaces )
is that if @xmath135 and @xmath134 are both @xmath45 diversities then so are @xmath159 and @xmath160 for @xmath161 . hence the @xmath45-embeddable diversities on
a given set form a cone .
deza and laurent @xcite make a systematic study of the identities and inequalities satisfied by the cone of @xmath45 _ metrics_. much of this work will no doubt have analogues in diversity theory .
for one thing , every identity for @xmath45 metrics is also an identity for the induced metrics of @xmath45 diversities .
however @xmath45 diversities will satisfy a far richer collection of identities .
one example is the following .
[ prop : circlel1 ] let @xmath46 be @xmath45-embeddable and let @xmath162 be finite subsets of @xmath47 with union @xmath89 .
then @xmath163 first suppose @xmath46 embeds isometrically in @xmath164 , the diameter diversity on @xmath48 .
let @xmath165 and @xmath166 be the minimum and maximum elements in @xmath89 . identify @xmath167 with @xmath168 and @xmath169 with @xmath170 .
there is @xmath171 such that @xmath172 , @xmath173 and , without loss of generality , @xmath174 . if @xmath175 then @xmath176 if @xmath177 then , without loss of generality , @xmath178 .
select @xmath179 such that @xmath180 , @xmath181 and @xmath182 for all @xmath183 . then , considering two different paths from @xmath184 to @xmath185 we obtain @xmath186 the case for general @xmath45-embeddable diversities can be obtained by integrating this inequality over the measure space .
espnola and piatek @xcite investigated when hyperconvexity for diversities implied hyperconvexity for their induced metrics , proving that this held whenever the induced metric @xmath73 of a diversity satisfies @xmath187 for all @xmath188 .
( see @xcite for definitions and results ) .
a consequence of proposition [ prop : circlel1 ] is that this property holds for all @xmath45-embeddable diversities .
if @xmath46 is @xmath45 embeddable then its induced metric @xmath73 satisfies for all finite @xmath74 .
suppose that @xmath189 .
there are @xmath190 cycles of length @xmath183 through @xmath89 , and each edge is contained in exactly @xmath191 such cycles . for each cycle
@xmath192 we have from proposition [ prop : circlel1 ] that @xmath193 hence @xmath194 we now examine three examples of diversities @xmath46 which are @xmath45-embeddable . in all three cases , the diversity need not be finite , nor even finite dimensional .
later , we examine @xmath45-embeddable diversities for finite sets .
measure diversities , @xmath16-diversities and mean - width diversities are all @xmath45-embeddable .
we treat each kind of diversity in turn .
+ _ measure diversities_. + in a measure diversity any element @xmath195 can be naturally identified with the function @xmath196 in @xmath197 .
observe now that @xmath198 _ mean - width diversities .
_ + let @xmath199 be the @xmath200-dimensional mean - width diversity . consider @xmath201 where @xmath114 is the unit sphere in @xmath202 , @xmath203 is the borel subsets of @xmath114 and @xmath204 is the measure given by @xmath205 for all @xmath206 where @xmath113 is the surface measure on @xmath114 .
let @xmath207 for @xmath208 be the function @xmath209 for @xmath210 . then @xmath211 thus @xmath212 is embedded in @xmath45 .
+ _ @xmath16-diversities_. + let @xmath46 be an @xmath16-diversity .
suppose that the random variables in @xmath47 have state space @xmath213 and that they are defined on the same probability space @xmath214 .
for each @xmath215 let @xmath216 be given by @xmath217 if @xmath218 and @xmath219 otherwise .
then @xmath220 in the case of measure diversities , we can also prove a converse result , in the sense that every @xmath45 diversity can be embedded in a measure diversity .
we first make some observations about @xmath48 .
consider the map @xmath221 given by @xmath222 & \mbox{if $ x \geq 0 $ } ; \\
\left [ x , 0 \right ) & \mbox{if $ x < 0$}. \end{array } \right.\ ] ] note that @xmath223 , where @xmath224 is lebesgue measure on @xmath48 .
furthermore , we have that @xmath225 to see that this is true , we consider three cases .
we let @xmath165 be the minimum of all @xmath226 and @xmath166 be the maximum . in case 1 ,
all the @xmath226 are non - negative
. then @xmath227 $ ] and @xmath228 $ ] .
this gives the result . in case 2 ,
all the @xmath226 are negative and the result follows similarly . in case 3 , some of the @xmath226 are positive and some of the @xmath226 are negative . in this case
@xmath229 $ ] and @xmath230 is empty .
any @xmath45-embeddable diversity can be embedded in a measure diversity . without loss of generality , consider the diversity @xmath135 where @xmath47 is a subset of @xmath231 .
we construct a new measure space @xmath232 , i.e. the product measure of @xmath233 with lebesgue measure on @xmath48 .
for @xmath234 we define @xmath235 by @xmath236 we then have that for all finite subsets @xmath237 of @xmath47 we have @xmath238 further results can be obtained for @xmath45-embeddable diversities
@xmath46 when @xmath47 is finite , say @xmath239 . in this case , the study of @xmath45 diversities reduces to the study of non - negative combinations of _ cut diversities _ , also called _ split diversities _
, that are directly analogous to _ cut metrics_. given @xmath240 define the diversity @xmath241 by @xmath242 in other words , @xmath243 when @xmath33 cuts @xmath89 into two parts .
the set of non - negative combinations of cut diversities for @xmath47 form a cone which equals the set of @xmath45-embeddable diversities on @xmath47 .
[ charact_l1_embed ] suppose that @xmath239 and @xmath46 is a diversity .
the following are equivalent . 1
. @xmath46 is @xmath45-embeddable .
@xmath46 is @xmath244-embeddable for some @xmath245 .
3 . @xmath46 is a _ split system diversity _ ( see @xcite ) .
that is , @xmath44 is a non - negative combination of cut diversities .
( i)@xmath246(iii ) + let @xmath247 be an embedding from @xmath47 to @xmath231 . for each @xmath240 and each @xmath248 let @xmath249 letting @xmath250 if this is negative .
define @xmath251 then for all @xmath87 and all @xmath74 we have @xmath252 and so @xmath253 ( iii ) @xmath246 ( ii ) .
+ fix @xmath254 .
we can write @xmath44 as @xmath255 for all @xmath74 where @xmath33 runs over all subsets of @xmath47 containing @xmath256 .
this collection of subsets of @xmath47 can be partitioned into @xmath257 disjoint chains by dilworth s theorem .
denote these chains by @xmath258 so that @xmath259 we will show that for every chain @xmath260 the diversity @xmath261 is @xmath262-embeddable
. the result follows . to this end , define @xmath263 by @xmath264 where @xmath265 is the minimal element of the chain @xmath266 that contains @xmath267
. then @xmath268 \(ii ) @xmath246 ( i ) .
+ follows from the fact that @xmath244 is itself an @xmath45 diversity .
diversities formed from combinations of split diversities were studied by @xcite and in literature on phylogenetic diversities @xcite .
proposition [ prop : finitel1 ] is a restatement of theorems 3 and 4 in @xcite .
[ prop : finitel1 ] let @xmath46 be a finite , @xmath45-embeddable diversity , where for all @xmath74 , @xmath269 where we assume @xmath270 . for all @xmath74 we have the identity @xmath271 and
if @xmath272 we have @xmath273 from these we obtain the following characterization of finite , @xmath45-embeddable metrics .
a finite diversity @xmath46 is @xmath45-embeddable if and only if it satisfies and @xmath274 for all @xmath74 , such that @xmath272 .
necessity follows from proposition [ prop : finitel1 ] . for sufficiency , observe that the map from a weight assignment @xmath224 to a diversity @xmath275 is linear and , by proposition [ prop : finitel1 ] , invertible for the space of weight functions @xmath224 satisfying @xmath276 for all @xmath33 .
the image of this map therefore has dimension @xmath277 . from we that the diversities @xmath90 for @xmath278 odd can be written in terms of diversities @xmath90 for @xmath278 even .
hence the space of diversities satisfying has dimension @xmath277 and lies in the image of the map .
condition [ eq : finitel1splits ] ensures that the diversity is given by a _
non - negative _ combination of cut diversities .
given two metric spaces @xmath279 and @xmath280 we can ask what is the minimal distortion embedding of @xmath281 into @xmath282 , where the minimum is taken over all maps @xmath283 .
naturally , we can ask the same question for diversities . whereas the question for metric spaces is well - studied ( though still containing many interesting open problems ) the situation for diversities is almost completely unexplored .
we state some preliminary bounds here , most of which leverage on metric results . we begin by proving bounds for several types of diversities defined on @xmath284 .
[ lem : divrk ] let @xmath285 and @xmath286 be the diameter diversities on @xmath284 , evaluated using @xmath0 and @xmath287 metrics respectively .
let @xmath288 and @xmath119 be the @xmath0 and mean - width diversities on @xmath284
. then for all finite @xmath289 @xmath290 all bounds are tight .
the inequalities @xmath291 and @xmath292 are due to theorem [ divbounds ] . to prove the @xmath0 bounds , note that for each dimension @xmath293 there are @xmath294 which maximize @xmath295 .
hence @xmath296 with equality given by subsets of @xmath297 . to prove the mean - width bound note
that , by jung s theorem @xcite , a set of points in @xmath284 with diameter @xmath298 is contained in some sphere with radius @xmath299 , where @xmath300 hence @xmath106 is contained in a set with mean width @xmath301 . from
we have @xmath302 where again @xmath120 denotes the beta function .
the bound holds in the limit for points distributed on the surface of a sphere .
we now investigate upper bounds for the distortion of diversities into @xmath45 space .
to begin , we consider only diversities which are themselves diameter diversities . in many senses ,
these diversities are similar to metrics , and it is perhaps no surprise that they can embedded with a similar upper bound as their metric counterparts .
let @xmath73 be a metric space , @xmath239 , and let @xmath303 be the corresponding diameter diversity . 1 .
there is an embedding of @xmath303 in @xmath304 with distortion @xmath305 and @xmath306 .
2 . there is an embedding of @xmath303 in @xmath307 with distortion @xmath308 and @xmath309 .
any metric on @xmath200 points can be embedded into the metric space @xmath310 with distortion @xmath311 , where @xmath312 @xcite .
let @xmath71 be an embedding for @xmath73 with @xmath313 for all @xmath128 , where @xmath109 is @xmath314 .
as above , we let @xmath315 denote the diameter diversity for the @xmath304 metric . for all
@xmath74 we have from lemma [ lem : divrk ] that @xmath316 the result now follows since @xmath183 is @xmath311 and @xmath109 is @xmath311 . \2 . as shown in @xcite
( see also @xcite ) , there is an embedding @xmath71 of @xmath73 into @xmath317 with @xmath318 for all @xmath128 , where @xmath183 and @xmath109 are @xmath319 . for all @xmath74 we have from lemma [ lem : divrk ] that @xmath320
the result follows .
we now consider the problem of embedding general diversities .
the bounds we obtain here can definitely be improved : we do little more than slightly extend the results for diameter diversities . [ thm : upperb ] let @xmath46 be a diversity with @xmath321 . 1 .
@xmath46 can be embedded in @xmath304 with @xmath309 with distortion @xmath322 .
@xmath46 can be embedded in @xmath317 with @xmath309 with distortion @xmath323 .
any diversity can be approximated by the diameter diversity of its induced metric with distortion @xmath200 , as shown in theorem [ divbounds ] .
this fact together with the previous theorem gives the required bounds . from upper bounds we switch to lower bounds .
any embedding of diversities with distortion @xmath109 induces an embedding of the underlying metric with distortion at most @xmath109 .
hence we can use the examples from metrics @xcite to establish that there are diversities which can not be embedded in @xmath0 with better than an @xmath324 distortion .
we have been able to obtain slightly tighter lower bounds for embeddings into @xmath304 where @xmath183 is bounded .
[ prop : nobourgain ] let @xmath46 be the @xmath200-point diversity with @xmath325 for all non - empty @xmath74 .
then the minimal distortion embedding of @xmath46 into @xmath304 has distortion at least @xmath326 . for any embedding @xmath71 of @xmath46 ,
lemma [ lem : divrk ] shows that @xmath327 for some @xmath56 , @xmath328 .
the distortion of @xmath71 is equal to @xmath329 taking @xmath330 and @xmath331 shows that the distortion is at least @xmath326 .
a consequence of proposition [ prop : nobourgain ] is that there will , in general , be no embedding of diversities in @xmath0 for which both the distortion and dimension is @xmath311 , or indeed polylog , ruling out a direct translation of the classical embedding results for finite metrics .
even so , we suspect that the upper bounds achieved in theorem [ thm : upperb ] can still be greatly improved .
having reviewed diversities , @xmath0 diversities , and the diversity embedding problems , we return to their application in combinatorial optimization .
we will here establish analogous results to those of @xcite and @xcite for hypergraphs and diversity embeddings into @xmath0 .
we first state the extensions of maximum multicommodity flows and minimum cuts a little more formally . given a hypergraph @xmath1 , non - negative weights @xmath332 for @xmath333 and @xmath14 ,
the goal is find the maximum weighted sum of minimal connected sub - hypergraphs covering @xmath16 without exceeding the capacity of any hyperedge .
let @xmath334 be set of all minimal connected sub - hypergraphs of @xmath43 that include @xmath16 .
for each sub - hypergraph @xmath335 assign weight @xmath336 .
we consider the following generalization of _ fractional steiner tree packing _ @xcite which we call _ maximum hypergraph steiner packing _ : identify @xmath336 satisfying the lp : @xmath337 as before , if we define @xmath332 for all subsets @xmath36 of @xmath2 , and let it be zero for @xmath338 , we can drop the dependence of the problem on @xmath3 .
the reference @xcite studies an oriented version of this problem . as with flows ,
maximum hypergraph steiner packing has a multicommodity version . for each subset @xmath16 of @xmath2
suppose we have non - negative demand @xmath27 .
we view @xmath339 and @xmath266 as non - negative vectors indexed by all subsets of @xmath2 .
suppose we want to simultaneously connect up all @xmath14 with minimal connected sub - hypergraphs carrying flow @xmath340 for all @xmath14 and we want to maximize @xmath10 .
the corresponding optimization problem is : @xmath341 note that we use @xmath29 rather than just @xmath336 because the same connected sub - hypergraph @xmath342 might cover more than one set @xmath16 in the hypergraph .
we call the optimal value of @xmath10 for this problem @xmath343 , for _ maximum multicommodity hypergraph steiner packing_. next we define the appropriate analogues of the min - cut problem , which we call _ minimum hypergraph cut_. as before , we let @xmath32 be the set of hyperedges which have endpoints in both @xmath33 and @xmath34 , and we make the simplifying assumption that every subset is a hyperedge , including any missing hyperedges with capacity zero .
we define @xmath344 below we will show that @xmath345 we define @xmath346 we say that a non - negative vector @xmath266 is _ supported on the hypergraph _
@xmath1 if @xmath347 for @xmath338
. then for any hypergraph @xmath43 we define @xmath348 to be the greatest value of @xmath349 over all nonnegative @xmath266 and @xmath339 such that @xmath266 is supported on @xmath43 .
we say that a diversity @xmath44 on @xmath2 is _ supported on _
@xmath1 if it is the hypergraph steiner diversity of @xmath43 for some set of non - negative weights @xmath332 for @xmath350 . for any diversity @xmath44 on @xmath2 we define @xmath351 to be the minimal distortion between @xmath44 and an @xmath0-embeddable diversity on @xmath2 . for any hypergraph @xmath43
we define @xmath352 to be the maximum of @xmath351 over all diversities @xmath44 supported on @xmath43 .
the major result for this section is that for all hypergraphs @xmath43 @xmath353 the fact that @xmath354 ( our theorem 1 ) is the analogue of results in section 4 of @xcite and the fact that equality holds ( our theorem 2 ) is the analogue of theorem 3.2 in @xcite . [
prop : characterizemaxhsp ] for all @xmath355 , @xmath356 where @xmath357 is the set of all diversities on @xmath2 .
in particular , the optimal @xmath358 is supported on the hypergraph @xmath1 where @xmath3 is the set of all @xmath36 such that @xmath359 .
we rewrite the linear program in standard form .
we break the equality constraint into @xmath360 and @xmath361 and note that we can omit the @xmath361 constraint , because it will never be active .
then we get @xmath362 let @xmath363 be the dual variables corresponding to the first set of inequality constraints , and let @xmath364 be the inequalities corresponding to the second set of inequality constraints .
then the dual problem is @xmath365 by strong duality , and have the same optimal values .
next we show that is equivalent to @xmath366 where the minimum is taken over all diversities . to see the equivalence of and , suppose that @xmath44 is a diversity solving .
let @xmath367 for all @xmath368 and @xmath369 for all @xmath14 .
then the objective function of is the same , the second line of still holds , the third line holds by the triangle inequality for diversities , and the fourth and fifth line hold by the non - negativity of diversities . to see the other direction ,
suppose @xmath363 and @xmath364 solve .
let @xmath44 be the steiner diversity on @xmath2 generated edge weights @xmath363 , @xmath368 . since @xmath370 for all @xmath371 , this can only decrease the objective function .
also , by the definition of @xmath44 , @xmath372 for all @xmath16 , so the inequality of is satisfied too .
thus the two lps have the same minima .
note we can assume that @xmath44 is the steiner diversity for a weighted hypergraph with hyperedges @xmath373 . if not , we can replace @xmath44 with the steiner diversity on the hypergraph whose hyperedges are the set @xmath373 and whose weights are the @xmath374 .
this steiner diversity will have the same value on the hyperedges as @xmath44 , so the objective function will not change , but the value can only increase on other subsets of @xmath2 , and so the constraint is still satisfied .
finally , is equivalent to @xmath375 this is because , any solution of will only have a smaller or equal value for the objective function of . and
any solution of can be rescaled without changing the objective function so that @xmath376 , giving a feasible solution to with the same objective function . this rescaling will not change the hypergraph that @xmath44 is supported on . for any cut @xmath380 of @xmath2 , let @xmath241 be the corresponding cut diversity .
then by definition we have that @xmath381 where we restrict @xmath33 to values where the denominator is non - zero .
we need to show that this value is not decreased by taking the minimum over all @xmath0-embeddable diversities instead .
let @xmath44 be @xmath0-embeddable diversity that minimizes the ratio . by proposition [ charact_l1_embed ]
, @xmath44 can be expressed as a finite linear combination of cut - diversities : @xmath382 for some non - negative @xmath383 and some subsets @xmath384 of @xmath2 .
let @xmath385 be the index @xmath293 that minimizes @xmath386 .
then we claim that @xmath387 to see this , observe that @xmath388 let @xmath389 for vectors @xmath390 with @xmath391 for all @xmath293 , where @xmath267 and @xmath392 are non - negative vectors of the same size .
we claim that @xmath26 attains its minimum on this domain at a value of @xmath390 consisting of a vector with a single non - zero entry . to show this
, we compute the gradient of @xmath26 @xmath393.\ ] ] if @xmath267 and @xmath392 are parallel then the result immediately follows so assume that they are not .
then @xmath394 is not zero anywhere in the domain , and so the maximum of @xmath26 must be taken on boundary of the domain .
so at least one @xmath383 must be zero . discard this term from the numerator and the denominator of @xmath26 . then repeat the argument for @xmath26 as a function of a vector of one fewer entries .
repeating gives a single non - zero value , which may be set to 1 .
for the second inequality , given @xmath355 and hypergraph @xmath1 supporting @xmath266 , let @xmath44 solve the maxhsp linear program . by proposition [ prop : characterizemaxhsp ]
we know that @xmath44 is supported on @xmath43 .
let @xmath397 be the minimal - distortion @xmath0 embeddable diversity of @xmath44 .
we may assume that @xmath398 .
then @xmath399 as required .
let @xmath402 be positive vectors .
define @xmath403 if @xmath404 is a closed set of positive vectors , define @xmath405 as @xmath406 . if @xmath407 is a closed convex cone , then @xmath408 where the maximum is taken over all non - negative vectors @xmath409 for which @xmath410 for any @xmath411 .
( of theorem [ thm : hyrax ] ) let @xmath44 be a diversity supported by the hypergraph @xmath43 that maximizes @xmath351 , and define @xmath412 .
we need to show that @xmath413 where the maximum is taken over all @xmath414 where @xmath266 is supported on @xmath43 .
let @xmath13 be given by @xmath44 , and let @xmath109 be the cone of all @xmath0-embeddable diversities on @xmath2 .
then @xmath415 .
we apply the lemma to show that @xmath416 where the maximum is taken over all non - negative vectors @xmath414 which satisfy the restriction @xmath417 for any @xmath0-embeddable diversity @xmath418 .
this tells us that there exists @xmath414 such that @xmath419 and @xmath417 for any @xmath0-embeddable diversity @xmath418 .
first we show that we may assume that @xmath266 is supported on @xmath43 .
suppose that for some set @xmath368 , @xmath420 we have @xmath421 .
since @xmath44 is supported on @xmath43 there are hyperedges @xmath422 that form a connected set covering @xmath371 with @xmath423 .
define a new vector @xmath424 by @xmath425 even with this new @xmath426 we still have @xmath427 and @xmath428 for any @xmath0-embeddable diversity @xmath418 . to see this
, first note that @xmath429 so @xmath430 secondly , since @xmath44 satisfies @xmath423 and these are the only sets on which @xmath266 is changed , it follows that @xmath431 .
we repeat this procedure until we have @xmath432 only if @xmath433 .
m. abramowitz and i. a. stegun .
_ handbook of mathematical functions with formulas , graphs , and mathematical tables _ , volume 55 of _ national bureau of standards applied mathematics series_. for sale by the superintendent of documents , u.s .
government printing office , washington , d.c . , 1964 .
d. bryant and p. f. tupper .
hyperconvexity and tight - span theory for diversities .
_ advances in mathematics _ , 2310 ( 6):0 3172 3198 , 2012 .
issn 0001 - 8708 .
doi : http://dx.doi.org/10.1016/j.aim.2012.08.008 .
k. jain , m. mahdian , and m. r. salavatipour .
packing steiner trees . in _ proceedings of the fourteenth annual acm - siam symposium on discrete algorithms ( baltimore , md , 2003 ) _ ,
pages 266274 , new york , 2003 .
t. kirly and l. c. lau .
approximate min - max theorems for steiner rooted - orientations of graphs and hypergraphs .
_ journal of combinatorial theory , series b _ , 980 ( 6):0 1233 1252 , 2008 .
issn 0095 - 8956 .
doi : http://dx.doi.org/10.1016/j.jctb.2008.01.006 .
t. leighton and s. rao .
an approximate max - flow min - cut theorem for uniform multicommodity flow problems with applications to approximation algorithms . in _ foundations of computer science , 1988 .
, 29th annual symposium on _ , pages 422431 , 1988 .
doi : 10.1109/sfcs.1988.21958 .
v. moulton , c. semple , and m. steel .
optimizing phylogenetic diversity under constraints .
_ journal of theoretical biology _ , 2460 ( 1):0 186 194 , 2007 .
issn 0022 - 5193 .
doi : http://dx.doi.org/10.1016/j.jtbi.2006.12.021 . | the embedding of finite metrics in @xmath0 has become a fundamental tool for both combinatorial optimization and large - scale data analysis .
one important application is to network flow problems as there is close relation between max - flow min - cut theorems and the minimal distortion embeddings of metrics into @xmath0 . here we show that this theory can be generalized to a larger set of combinatorial optimization problems on both graphs and hypergraphs .
this theory is not built on metrics and metric embeddings , but on _ diversities _ , a type of multi - way metric introduced recently by the authors .
we explore diversity embeddings , @xmath0 diversities , and their application to _ steiner tree packing _ and _ hypergraph cut _ problems . |
to our knowledge , the first studies dealing with interfacial phase transitions in coulomb fluids were conducted by freyland _
et al . _
their _ in situ _ stm study of the interface between the [ c@xmath0mim@xmath1[pf@xmath2 and the ( 111 ) face of gold reported the formation of moir - like patterns at potentials greater than -0.2 v with respect to a platinum reference electrode .
these were attributed to the formation of an ordered adlayer of pf@xmath3 . at negative potentials ,
the stm images were consistent with the formation a layer of anions with the ( @xmath4 ) structure , indicating a two - dimensional ordering transition at this interface .
it is worth noting that these observations closely follow a previous work performed on the adsorption of iodine from aqueous solutions on similar gold surfaces @xcite .
a further study by the same authors on the electrodeposition of cd on au(111 ) in a chloroaluminate ionic liquid has also revealed the formation of an ordered alcl@xmath5 adlayer @xcite . 7.5 nm ) of the bmp@xmath6 adlayer on au(111 ) at @xmath71.4 v and b ) proposed structural model , showing two domains of the ( @xmath8 ) bmp@xmath6 adlayer separated by a translational domain boundary ( indicated by dashed line ) . reproduced with permission from reference ; copyright : wiley , 2015.,scaledwidth=50.0% ] when changing both the nature of the anion ( from pf@xmath3 to bf@xmath5 ) and of the surface of gold in contact with the rtil ( from ( 111 ) to ( 100 ) ) , su _ et al .
_ have also evidenced the existence of potential - driven ordering transitions @xcite .
increasing the potential from @xmath70.3 v , an ordered layer of anions is formed between @xmath70.1 and 0.4 v. on the contrary , when scanning in the negative potentials direction , they first observed a loose film - like layer which was attributed to a disordered adsorption of c@xmath0mim@xmath6 cations .
then , for potentials lower than @xmath70.95 v , perpendicularly oriented double - row strips were observed .
these strips were assigned to the formation of micelle - like arrangements of aligned c@xmath0mim@xmath6 cations .
these structures also formed with pf@xmath3 and so@xmath9cf@xmath10 anions , but not on ( 111 ) surfaces of gold @xcite , which shows that in this case a structural commensurability of the adsorbed layer and the metal surface is necessary for the formation of ordered structures .
this conclusion was confirmed in a study using _ in situ _ video - stm to probe the ( 111 ) au interface with a rtil composed of a different cation , namely 1-butyl-1-methylpyrrolidinium ( bmp@xmath6 ) associated with the bis(trifluoromethylsulfonyl)imide anion ( tfsi@xmath11 ) @xcite .
stable images could only be obtained for negative potentials below @xmath71 v. for such potentials the images showed the formation of ordered structures .
several distinct arrangements of cations were proposed in order to interpret the observations at various potentials ; one of them is shown in [ fig : stm ] . in both structures proposed by wen _
_ , the cation rings are adsorbed on the surface ; in contrast , the alkyl chains lie flat on the surface only for the lower charge density ( hence lateral cation density in the adsorbed fluid ) and extend into the perpendicular direction for the higher density
. the lattice parameters for the adlayer superstructure decrease accordingly and may change symmetry , resulting for the densest packing in a square lattice which differs from the hexagonal substrate . finally , the video - stm furter allowed the first direct observation of the dynamical evolution of the adsorbed liquid .
in particular , it was found that the fluctuations occur mainly at the boundaries between ordered domains .
all the structural transitions observed with stm have so far involved au electrodes because this metal can be produced as single crystals with well - defined surfaces .
have recently used another technique , _ in situ _ amplitude - modulated afm , to study interfaces of highly ordered pyrolytic graphite ( hopg ) instead @xcite .
this substrate presents the advantage of having flat surfaces with high area , and to avoid the surface reconstructions or etching which can occur in the case of gold @xcite .
these authors studied the effect of applied potential on the adsorbed layer structure for a [ c@xmath0mim@xmath1[tfsi@xmath11 ] ionic liquid . at the open - circuit potential , well - defined rows are present on the surface . unlike previous works , in which the ordered structures were apparently formed of only one type of ions ,
the unit cell is composed of an anion - cation - cation - anion arrangement @xcite
. this structure changes markedly with surface potential or when relatively low concentrations of lithium or chloride ions are present in the rtil .
the variety of systems in which transitions are observed clearly show that it is a common feature of metal / ionic liquids interfaces .
however , the few works reported so far raise very interesting questions .
in particular , all the ordered structures proposed so far to interpret the stm data are composed of a single species only .
however , the relatively small applied potentials which are used ( @xmath131 v ) may not be sufficent to fully separate cations from anions . in the case of the afm study , the anion - cation - cation - anion rows result in an overall neutral layer but there is a strong charge imbalance on the nanometre scale .
a few hypotheses can thus be proposed to explain the observations : i ) there may be a specific absorption of the ions on gold with the formation of partially covalent bonds @xcite ii ) on top of the observed layer , there could be an oppositely charged layer of ions which is not observed by the experiment and iii ) it remains possible that the proposed structures , which are only based on the relative size of the ions ( keeping the possibility of some kind of conformational ordering , for example only the imidazolium rings of the cations would lie parallel to the surface ) , are not the correct ones .
using additional techniques such as sfg , which is sensitive to the orientation of the ions @xcite , could possibly shed a complementary light on this issue .
another open question is whether commensurability between the adsorbed layer and the metal substrate is necessary to observe a transition . here
also , the recent study performed with hopg electrodes @xcite suggests that it is not the case , and that the electric fluctuations at an homogeneous and flat metallic surface are sufficient to trigger ordering transitions in the interfacial layer of rtil .
in addition , the use of carbon electrodes in this study demonstrates that a perfect metallic behavior is not necessary to induce such transitions . in order to simulate electrochemical systems
, it is necessary to fix the potential of the electrode . in classical molecular dynamics or monte - carlo ,
this can be done by various methods @xcite .
our approach consists in treating the partial charges carried by the electrode atoms as additional degrees of freedom which fluctuate during the simulation .
their values are determined at each time step from a self - consistent calculation @xcite . in such simulations ,
the electrochemical cell consists in a wide slab of electrolyte held between two electrodes with different voltages . like in experiments ,
the potentials are not absolute .
the only fixed quantity is the potential difference between the two electrodes @xmath14 , although it is also possible to calculate the potential of each electrode with respect to the bulk liquid in the case of flat electrodes ; we will note this potential @xmath15 . in the following
, we will assimilate an ordering transition to an abrupt change in the structure observed when changing the potential .
however , it is worth noting that first - order transitions are associated with a discontinuity in an order parameter and a corresponding singularity in a partition function , which are not easy to prove in simulations @xcite .
this point will be further discussed in the next section . using this simulation approach , a first example of voltage - driven transition
was reported for a rather exotic system , formed with a high temperature molten salt ( licl ) and an aluminum electrode with its ( 100 ) surface in contact with the liquid @xcite .
an advantage of this system is that a polarizable force field could be built directly from accurate dft calculations using a generalized force - matching approach @xcite.for potential drops @xmath15 across the interface more negative than @xmath71.76 v , which corresponds to the point of zero charge ( pzc ) , the molten salt adopted a disordered structure at the interface , while for larger potentials an ordered structure was obtained @xcite .
this structure , which is shown on the top panel of [ fig : snapshots ] , was commensurate with the aluminum substrate , and a strong alignment of the dipole components of the chloride anion and the normal of the surface was observed for large potentials @xcite .
interestingly , no transition was observed when the plane of the metal was changed to ( 110 ) instead of ( 100 ) , but a different ordered structure was then obtained showing that an epitaxial mechanism is at play , whereby the molten salt adapts it structure to that of the electrode surface . in a recent work ,
studied interfaces between primitive models of ionic liquids and solid surfaces with various net charges ( _ i.e. _ the electric potential was not controlled ) @xcite . at certain charge densities ( @xmath16 @xmath17 @xmath18c @xmath19 ) the structure of the adsorbed layer of cations undergoes a structural transition to a surface - frozen monolayer of densely packed counter - ions with a moir - like structure . at lower surface charge densities ( _ i.e. _ lower than @xmath20
@xmath18c @xmath19 ) , they even observed the formation of an herring - bone structure arising from the superposition of two ordered monolayers of ions ( see the bottom panel of [ fig : snapshots ] ) .
these findings provide an interesting support for the stm studies discussed above , but it is worth noting that the charge densities employed are somewhat larger than the experimental ones they would correspond to potentials which are above the electrochemical window of typical rtils .
going towards more realistic models , an ordering transition was reported from molecular simulations for the interface between [ c@xmath0mim@xmath1[pf@xmath2 ( for which a coarse - grained force field was used ) and an electrified surface of graphite @xcite .
the presence of the ordered structure could be monitored by computing the in - plane structure factor in the first layer of the adsorbed liquid .
this structure factor was liquid - like on a wide range of potentials , but it showed some strong bragg - like peaks suggesting a 2-dimensional lattice - like organization for both the anions and cations , which is shown in the top - right panel of [ fig : snapshots ] .
this ordered structure contained on average as many anions as cations , and it was also observed by kislenko _
et al . _ in simulations of the same rtil ( with an all - atom model ) adsorbed on an uncharged surface of graphite @xcite . by using importance sampling techniques ,
it could be shown by merlet _
_ that this structure was the most stable one for small positive potentials ( @xmath21 v ) and metastable for small negative electrode potentials .
it is worth noting that similarly to the experimental work of elbourne _ et al . _
involving hopg electrodes , no commensurability with the electrode surface seems necessary to observe such ordered structures .
so far , no ordering transitions have been observed using more elaborate , all - atom models of rtils in contact with electrodes at fixed potential . in particular , the adsorption of [ c@xmath0mim@xmath1[pf@xmath2 and [ c@xmath0mim@xmath1[bf@xmath5 ] on electrified surfaces of gold was studied by hu _
et al . _ ,
@xcite but they did not report any ordering transition similar to the experimental observations by stm . these simulation results , while confirming the possibility of transitions in the adsorbed layer of the fluid , also open their share of questions .
future works will need to address the issue of finite - size effects , since there must be a commensurability between the formed ordered structure and the simulation cell .
timescales are important too , since metastable states may be much longer - lived than the typical simulation times , which are on the order of the nanosecond only due to the computational cost .
there is therefore a possibility that the reported transitions are artefacts of the simulation setups , but the similarities with experimental findings seem to weaken this hypothesis .
the question of specific interactions with surfaces such as gold will also have to be treated .
this requires in turn the development of accurate force fields for this purpose .
first steps in this direction have recently been made in the case of carbon materials @xcite .
we now turn to the consequences of structural transitions within the adsorbed fluid on the physico - chemical properties of the interface .
specifically , we discuss the impact of voltage - induced transitions on the electrochemical response of the electrode - rtil interface in terms of differential capacitance , cyclic voltammograms and electrochemical reaction , as well as on the mechanical response ( solid - liquid friction ) .
the differential capacitance @xmath23 measures the response of the average surface charge density @xmath24 to changes in the voltage @xmath25 : @xmath26 by definition , a capacitor corresponds to a voltage - independent differential capacitance .
however , the charge of the electrode reflects the composition and the charge distribution within the interfacial liquid . as a result
, one should expect a signature of abrupt structural changes at voltages corresponding to putative phase transitions in the electrode charge , hence peaks in the corresponding differential capacitance .
while experimentally such peaks have indeed been observed @xcite , their possible link with changes in the structure or the interface has been difficult to demonstrate until recently , due to the experimental challenges of _ in situ _ imaging and the occurence of other processes such as surface reconstruction of the electrode .
indeed , in their stm study of [ c@xmath0mim@xmath1[bf@xmath5 ] on a ( 100 ) gold electrode , su _ et al . _
have also measured the capacitance of the interface .
they observed a fivefold increase in this quantity in the potential region of transition from anion adsorption to cation adsorption . in the case of the _ in situ _
video - stm study of the [ bmp@xmath1[tfsa@xmath11 ] on a ( 111 ) gold electrode @xcite , thanks to the high temporal resolution or the video - stm technique , the authors were able to visualize the evolution of the interfacial fluid during cyclic voltammetry ( cv ) experiments .
the cyclic voltammogram displays two current peaks associated with two surface transitions which could also be linked to the formation of the ordered cationic structures , such as the one shown on [ fig : stm ] , upon increasingly negative surface charge density . in computer simulations , it is relatively straightforward to calculate the capacitance of the interface in constant potential simulations .
the generic method consists in simulating an electrochemical cell at various voltage and extracting the average surface charge .
then the @xmath27 plot is differentiated , which provides the differential capacitance through [ eq : capa1 ] .
however , close to a transition , a large peak in the capacitance is expected , so that many voltages should in principle be sampled in this region .
an alternative was recently proposed , which consists in using importance sampling methods @xcite . in short , by using the whole distribution of surface charges during the simulations , it is possible to sample the probability distributions of any variable as continuous functions of the applied potential .
there is in principle no need to acquire more data close to the transition , the only requisite is to have a good overlap between the histograms of surface charges from the various voltages .
the probability distribution of the charge density @xmath28 of graphite electrodes in contact with the [ c@xmath0mim@xmath1[pf@xmath2 ionic liquid @xcite obtained with this approach is shown on [ fig : capa](a ) .
the figure shows the probability distribution on a logarithmic scale .
it is clear that there are three branches along which the distribution of the surface charge distribution shifts almost linearly upon increasing the potential .
these branches are separated by more complex changes in the distribution around particular voltages .
we will focus on the one occuring at @xmath29 = 0.9 v since this is the potential for which the order - disorder transition discussed above occurs .
[ fig : capa](b ) shows the distribution @xmath30 at three applied voltages ( 0.8 v , 0.9 v and 1.0 v ) .
they are characteristic of a first - order phase transition .
away from phase coexistence ( at 0.8 v and 1.0 v ) , they show the presence of `` fat tails '' , which are due to the small probability of obtaining the metastable phase . at the transition
the distribution displays hints of bimodality , which is expected if the two phases are equiprobable
. however it would be necessary to simulate larger systems to fully conclude on this point .
the differential capacitance computed from these simulations is shown on [ fig : capa](c ) .
note that much better statistics could again be obtained compared to the usual method involving [ eq : capa1 ] by using the johnson - nyquist relation , @xmath31 where @xmath32 is the surface of the electrode and @xmath33 are the fluctuations in the electrode surface charge density .
a large peak in the differential capacitance is observed at the applied voltage where the transition occurs , which is consistent with the experimental findings of su _ et al . _
note that again , larger systems should lead to a singular charge - density transition in a macroscopic limit @xcite .
our simulations therefore confirm that the presence of large peaks in experimental measures of the capacitance of an interface can indeed be the signature that a potential - driven transition is occuring .
the above - mentioned domain boundaries between phases also have important implications by themselves , due to the entailed free energy cost . in three dimensions
, this would be a surface free energy . in the present case
the topology of the boundary between interfacial domains at the surface of the electrode remains to be clarified @xcite .
as a result , annealing these boundaries , either between grains of otherwise identical domains , or between different domains , requires overcoming the corresponding free energy barrier . in practice , the consequences of these barriers are observable as long time scales in the dynamics of the interface or as hysteresis in cyclic voltammetry .
uysal _ et al .
_ reported a potential - dependent hysteresis at an electrified graphene / rtil interface @xcite .
x - ray reflectivity measurements during cyclic voltammetry and potential step measurements are used to probe the electronic density in the direction perpendicular to an epitaxial graphene surface , within the adsorbed [ c@xmath34mim@xmath1[tfsi@xmath11 ] ionic liquid .
the resulting profiles were consistent with that obtained from md simulations , by assuming a combination of two limiting structures , with weights varying as a function of applied voltage .
the structure evolves very slowly after a potential step , with processes occuring over time scales exceeding 10 s. in addition , the cv scans exhibit significant ( scan rate dependent ) hysteresis . while in this work the authors safely indicated that the nature of the apparent barrier and the associated mechanism require further investigation , these observations clearly point to the crucial role of structural transitions and the associated domain boundaries in the observed hysteresis and slow dynamics .
another manifestation of slow processes occuring at the ionic liquid / electrode interface was reported by roling _
et al . _ , who have carefully analysed the capacitance spectra on a broad range of frequencies @xcite .
although no particular ordering transition was observed by stm , these authors concluded that the slower capacitive process could be related to structural reorganisations of the gold surface or to strong rearrangements in the first adsorbed layer of ions .
recently , limmer proposed a detailed study of these effects using a coarse - grained model capturing strong inter - ionic correlations @xcite .
its limited computational cost compared to molecular simulations allowed for a systematic finite - size scaling analysis , which demonstrated the first - order nature of the fluctuation - induced transition and spontaneous charge density ordering at the interface , in the presence of an otherwise disordered bulk solution , already observed in molecular simulations @xcite .
a crucial step in this demonstration is the extensive growth of the free energy barrier between phases analogous to the ones observed in reference , which indeed implies hysteresis and long time scales .
structural changes in the ionic liquid at the interface also have implications on the local environment of other species in the liquid , in particular electro - active species .
this in turn may result in changes in their reactivity .
a direct observation of this feature has recently been reported by garcia - rey and dlott , who studied co@xmath35 reduction on a polycrystalline ag electrode , with 1-ethyl-3-methylimidazolium tetrafluoroborate [ c@xmath35mim@xmath1[bf@xmath5 ] containing 0.3 mol% water as electrolyte @xcite .
such systems have been shown to reduced the overpotential for co@xmath35 reduction .
sfg and ir were used to probe the surface field experienced by the adsorbed co molecules produced by the electrochemical reduction of co@xmath35 . from the co stark shift ,
a sudden increase of the field at the electrode surface was observed at the threshold potential for co@xmath35 reduction , which could be traced back to a structural transition within the rtil even though no information could be obtained on the nature of these structural changes .
nevertheless , this study illustrates the potential benefit of exploiting the peculiar structure of ionic liquid interfaces and the voltage - driven changes thereof ( with potentially much greater diversity than in solvent - based electrolytes ) for electrochemical reactions .
[ fap@xmath11 ] .
each data point represents the average over a full scan frame of 14 nm side length .
the potential is switched between @xmath71.5 and 0.7 v for each value of the applied normal load .
reprinted with permission from reference ; copyright 2012 by the american physical society.,scaledwidth=50.0% ] finally , voltage - driven changes in the structure and composition of the interfacial fluid also have implications from the dynamical point of view .
_ e__t al . conducted nanotribology experiments to probe the lubrication properties of 1-butyl-1-methylpyrrolidinium tris(pentafluoroethyl ) trifluorophosphate ( [ bmp@xmath1[fap@xmath11 ] ) confined between silica colloid probes or sharp silica tips and a au(111 ) substrate , using afm @xcite .
as the composition of the adsorbed layers is tuned by the electrode potential , from cation - enriched to anion - enriched , the friction also evolves .
[ fig : friction ] illustrates that these variations are directly linked to the nature of the sliding plane , which may correspond to cation or anion layers , depending on the electrode potential and on the normal load exerted on the confined fluid .
while the voltage - driven structural changes on the microscopic interfacial structures remain to be investigated , such studies open the way to a new tuning of frictional forces at the molecular scale without changing the substrate .
more detailed information on the role of key microscopic and macroscopic factors can be obtained using molecular simulations , such as load , shear velocity , surface topology and length of alkyl side chains in the ionic liquid @xcite .
simulations with fixed surface charge density ( instead of potential ) have further evidenced two mechanisms underlying friction changes in such systems , namely charge effects on normal and in - plane ordering in the film , as well as swapping between anion and cation layers at the surface @xcite .
there is now a large body of experiments pointing towards the existence of potential - driven transitions at the interface between ionic liquids and metallic electrodes .
however , as discussed above , the exact structure and composition of the ordered phases remain open questions .
computer simulations bring some theoretical support on the question , but they are still scarce because of the technical difficulty associated with the use of constant applied potential ensemble .
they also suffer from sampling issues ( both in size and time ) which render the observation and the characterization of the transitions difficult .
the current works , in which the interactions are determined using classical force fields may also be limited if particular bonding occurs at the interface .
the recent inclusion of constant voltage methods in dft - based molecular dynamics packages @xcite may open new opportunities for tackling this difficult problem , as it was shown recently in the context of nanotribology .
@xcite in addition , the role of many parameters remain to be established . for example , how do the composition of rtil and the possible presence of impurities affect the occurence of ordering transitions ?
what is the impact of the temperature of the systems ?
also many applications of rtils use them in the presence of a solvent , which will also impact the whole structure of the electric double layer . finally , although it is clear that the nature of the substrate plays a strong role , it is not sure that there is always a commensurability between the ordered structure of the liquid and the metal .
topological defects at the surface of the metal may also play a predominant role , @xcite and it is likely that corrugation effects can modify the formation and/or the detection of ordered layers : recent simulations have shown that the heterogeneous nucleation of ice at a surface depended markedly on the morphology of the latter .
@xcite additional works with varying metal electrodes will allow to better understand these issues .
whether these transitions will have practical applications remains an open question , but they clearly impact a lot the physico - chemical properties : peaks in the differential capacitances , slow dynamics at the interface , varying reactivity and voltage - dependent friction properties have already been reported .
overall such transitions reinforce the view of rtils as solvents with multifaceted properties , with a composition that can be specifically tailored to a given task .
lin , i .- c . ; seitsonen , a. p. ; coutinho - neto , m. d. ; tavernelli , i. ; rothlisberger , u. importance of van der waals interactions in liquid water .
_ j. phys .
b _ * 2009 * , _ 113 _ , 11271131 baldelli , s. surface structure at the ionic liquid - electrified metal interface . _
res . _ * 2008 * , _ 41 _ , 421431 pealber , c. y. ; baldelli , s. observation of charge inversion of an ionic liquid at the solid salt - liquid interface by sum frequency generation spectroscopy .
_ j. phys .
lett . _ * 2012 * , _ 3 _ , 844847 baldelli , s. interfacial structure of room - temperature ionic liquids at the solid - liquid interface as probed by sum frequency generation spectroscopy .
_ j. phys .
lett . _ * 2013 * , _ 4 _ , 244252 atkin , r. ; warr , g. g. structure in confined room - temperature ionic liquids .
_ j. phys .
c _ * 2007 * , _ 111 _ , 51625168 hayes , r. ; warr , g. g. ; atkin , r. structure and nanostructure in ionic liquids . _ chem .
rev . _ * 2015 * , _ 115 _ , 63576426 mezger , m. ; schrder , h. ; reichert , h. ; schramm , s. ; okasinski , j. s. ; schrder , s. ; honkimki , v. ; deutsch , m. ; ocko , b. m. ; ralston , j. ; _ et al_. molecular layering of fluorinated ionic liquids at a charged sapphire ( 0001 ) surface .
_ science _ * 2008 * , _ 322 _ , 424428 mezger , m. ; roth , r. ; schrder , h. ; reichert , p. ; pontoni , d. ; reichert , h. solid - liquid interfaces of ionic liquid solutions interfacial layering and bulk correlations .
_ j. chem .
* 2015 * , _ 124 _ , 164707 vandevondele , j. ; krack , m. ; mohamed , f. ; parrinello , m. ; chassaing , t. ; hutter , j. quickstep : fast and accurate density functional calculations using a mixed gaussian and plane waves approach . _
* 2005 * , _ 167 _ , 103128 reed , s. k. ; lanning , o. j. ; madden , p. a. electrochemical interface between an ionic liquid and a model metallic electrode .
_ j. chem .
_ * 2007 * , _ 126 _ , 084704 carrasco , j. ; hodgson , a. ; michaelides , a. a molecular perspective of water at metal interfaces .
* 2012 * , _ 11 _ , 667674 limmer , d. t. ; willard , a. p. ; madden , p. ; chandler , d. hydration of metal surfaces can be dynamically heterogeneous and hydrophobic .
* 2013 * , _ 110 _ , 42004205 cheng , j. ; sprik , m. alignment of electronic energy levels at electrochemical interfaces .
phys . _ * 2012 * , _ 14 _ , 1124511267 nielsen , m. ; bjrketun , m. e. ; hansen , m. h. ; rossmeisl , j. towards first principles modeling of electrochemical electrode - electrolyte interfaces . _ surf .
sci . _ * 2015 * , _ 631 _ , 27 willard , a. p. ; limmer , d. t. ; madden , p. a. ; chandler , d. characterizing heterogeneous dynamics at hydrated electrode surfaces _
phys . _ * 2013 * , _ 138 _ , 184702 armand , m. ; endres , f. ; macfarlane , d. r. ; ohno , h. ; scrosati , b. ionic - liquid materials for the electrochemical challenges of the future .
mater . _ * 2009 * , _ 8 _ , 621629 kornyshev , a. a. double - layer in ionic liquids : paradigm change ? _ j. phys .
b _ * 2007 * , _ 111 _ , 55455557 perkin , s. ; salanne , m. ; madden , p. ; lynden - bell , r. is a stern and diffuse layer model appropriate to ionic liquids at surfaces ? _ proc .
* 2013 * , _ 110 _ , e4121 lee , a. a. ; vella , d. ; perkin , s. ; goriely , a. are room - temperature ionic liquids dilute electrolytes ?
_ j. phys .
lett . _ * 2015 * , _ 6 _ , 159163 fedorov , m. v. ; kornyshev , a. a. ionic liquids at electrified interfaces .
_ * 2014 * , _ 114 _ , 29783036 perkin , s. ionic liquids in confined geometries . _ phys .
phys . _ * 2012 * , _ 14 _ , 50525062 merlet , c. ; rotenberg , b. ; madden , p. a. ; salanne , m. computer simulations of ionic liquids at electrochemical interfaces .
chem . phys . _
* 2013 * , _ 15 _ , 1578115792 ivanistsev , v. ; oconnor , s. ; fedorov , m. v. poly(a)morphic portrait of the electrical double layer in ionic liquid _ electrochem .
commun . _ * 2014 * , _ 48 _ , 6164 kornyshev , a. a. ; qiao , r. three - dimensional double layers _ j. phys . chem .
c _ * 2014 * , _ 118 _ , 1828518290 liu , q. x. ; zein el abedin , s. ; endres , f. electroplating of mild steel by aluminium in a first generation ionic liquid : a green alternative to commercial al - plating in organic solvents _ surf .
_ * 2006 * , _ 201 _ , 13521356 jeon , y. ; vaknin , d. ; bu , w. ; sung , j. ; ouchi , y. ; sung , w. ; kim , d. surface nanocrystallization of an ionic liquid .
* 2012 * , _ 108 _ , 055502 freyland , w. interfacial phase transitions in conducting fluids .
phys . _ * 2008 * , _ 10 _ , 923936 tao , n. j. ; lindsay , s. m. in situ scanning tunneling microscopy study of iodine and bromine adsorption on gold(111 ) under potential control .
_ j. phys .
_ * 1992 * , _ 96 _ , 52135217 pan , g .- b . ;
freyland , w. in situ stm investigation of spinodal decomposition and surface alloying during underpotential deposition of cd on au(111 ) from an ionic liquid .
phys . _ * 2007 * , _ 9 _ , 32863290 wen , r. ; rahn , b. ; magnussen , o. m. potential - dependent adlayer structure and dynamics at the ionic liquid / au(111 ) interface : a molecular - scale in situ video - stm study .
* 2015 * , _ 54 _ , 60626026 su , y .- z . ;
fu , y .- c . ; yan , j .- w . ;
chen , z .- b . ; mao , b .- w .
double layer of au(100)/ionic liquid interface and its stability in imidazolium - based ionic liquids .
, int . ed .
_ * 2009 * , _ 48 _ , 51485151 elbourne , a. ; mcdonald , s. ; voitchovsky , k. ; endres , f. ; warr , g. g. ; atkin , r. nanostructure of the ionic liquid - graphite stern layer . _
acs nano _ * 2015 * , _ 9 _ , 76087620 anderson , e. ; grozovski , v. ; siinor , l. ; siimenson , c. ; lust , e. in situ stm studies of bi(111)|1-ethyl-3-methylimidazolium tetrafluoroborate + 1-ethyl-3-methylimidazolium iodide interface _ electrochem .
* 2014 * , _ 46 _ , 1821 tazi , s. ; salanne , m. ; simon , c. ; turq , p. ; pounds , m. ; madden , p. a. potential - induced ordering transition of the adsorbed layer at the ionic liquid / electrified metal interface . _ j. phys .
b _ * 2010 * , _ 114 _ , 84538459 merlet , c. ; limmer , d. t. ; salanne , m. ; van roij , r. ; madden , p. a. ; chandler , d. ; rotenberg , b. the electric double layer has a life of its own .
_ j. phys .
* 2014 * , _ 118 _ , 1829118298 kirchner , k. ; kirchner , t. ; ivanistsev , v. ; fedorov , m. v. electrical double layer in ionic liquids : structural transitions from multilayer to monolayer structure at the interface .
_ electrochim .
acta _ * 2013 * , _ 110 _ , 762771 siepmann , j. i. ; sprik , m. influence of surface - topology and electrostatic potential on water electrode systems .
_ j. chem .
phys . _ * 1995 * , _ 102 _ , 511524 merlet , c. ; pan , c. ; rotenberg , b. ; madden , p. a. ; simon , p. ; salanne , m. simulating supercapacitors : can we model electrodes as constant charge surfaces ?
_ j. phys .
lett . _ * 2013 * , _ 4 _ , 264268 chandler , d. _ introduction to modern statistical mechanics _ ; oxford university press , 1987 pounds , m. ; tazi , s. ; salanne , m. ; madden , p. a. ion adsorption at a metallic electrode : an ab initio based simulation study . _ j. phys . : condens . matter _
* 2009 * , _ 21 _ , 424109 kislenko , s. ; samoylov , i. ; amirov , r. molecular dynamics simulation of the electrochemical interface between a graphite surface and the ionic liquid @xmath36bmim@xmath37[$]pf@xmath38 $ ] .
phys . _ * 2009 * , _ 11 _ , 55845590 hu , z. ; vatamanu , j. ; borodin , o. ; bedrov , d. a molecular dynamics simulation study of the electric double layer and capacitance of @xmath36bmim@xmath37[$]pf@xmath38 $ ] and @xmath36bmim@xmath37[$]bf@xmath39 $ ] room temperature ionic liquids near charged surfaces _ phys .
phys . _ * 2013 * , _ 15 _ , 1423414247 pensado , a. s. ; malberg , f. ; costa gomes , m. f. ; pdua , a. a. h. ; fernndez , j. ; kirchner , b. interactions and structure of ionic liquids on graphene and carbon nanotubes surfaces .
_ rsc adv . _
* 2014 * , _ 4 _ , 1801718024 cannes , c. ; cachet , h. ; debiemme - chouvy , c. ; deslouis , c. ; de sanoit , j. ; le naour , c. ; zinovyeva , v. a. the double layer at [ bumeim][tf@xmath35n ] ionic liquid - pt or c materials interfaces .
_ j. phys .
c _ * 2013 * , _ 117 _ , 2291522925 costa , r. ; pereira , c. m. ; silva , a. f. structural ordering transitions in ionic liquids mixtures .
_ electrochem .
commun . _ * 2015 * , _ 57 _ , 1013 limmer , d. t. ; merlet , c. ; salanne , m. ; chandler , d. ; madden , p. a. ; van roij , r. ; rotenberg , b. charge fluctuations in nanoscale capacitors .
lett . _ * 2013 * , _ 111 _ , 106102 uysal , a. ; zhou , h. ; feng , g. ; lee , s. s. ; li , s. ; fenter , p. ; cummings , p. t. ; fulvio , p. f. ; dai , s. ; mcdonough , j. k. ; gogotsi , y. structural origins of potential dependent hysteresis at the electrified graphene / ionic liquid interface .
_ j. phys .
c _ * 2014 * , _ 118 _ , 569574 roling , b. ; drschler , m. ; huber , b. slow and fast capacitive process taking place at the ionic liquid / electrode interface .
_ faraday discuss . _
* 2012 * , _ 154 _ , 303311 druschler , m. ; borisenko , n. ; wallauer , j. ; winter , c. ; huber , b. ; endres , f. ; roling , b. new insights into the interface between a single - crystalline metal electrode and an extremely pure ionic liquid : slow interfacial processes and the influence of temperature on interfacial dynamics .
phys . _ * 2012 * , _ 14 _ , 50905099 limmer , d. interfacial ordering and accompanying unbounded capacitance at ionic liquid - metal interfaces .
_ arxiv _ * 2015 * , 1506.02667 , n. ; dlott , d. d. a structural transition in an ionic liquid controls co@xmath35 electrochemical reduction . _
c _ * 2015 * , _ 119 _ , 2089220899 sweeney , j. ; hausen , f. ; hayes , r. ; webber , g. b. ; endres , f. ; rutland , m. w. ; bennewitz , r. ; atkin , r. control of nanoscale friction on gold in an ionic liquid by a potential - dependent ionic lubricant layer . _ phys .
lett . _ * 2012 * , _ 109 _ , 155502 , a. c. f. ; pdua , a. a. h. ; malfreyt , p. non - equilibrium molecular simulations of new ionic lubricants at metallic surfaces : prediction of the friction .
_ j. chem .
theory comput . _ * 2013 * , _ 9 _ , 16001610 fajardo , o. y. ; bresme , f. ; kornyshev , a. a. ; urbakh , m. electrotunable lubricity with ionic liquid nanoscale films .
rep . _ * 2015 * , _ 5 _ , 7698 golze , d. ; iannuzzi , m. ; nguyen , m .-
t . ; passerone , d. ; hutter , j. simulation of adsorption processes at metallic interfaces : an image charge augmented qm / mm approach . _
theory comput . _ * 2013 * , _ 9 _ , 50865097 page , a. j. ; elbourne , a. ; stefanovic , r. ; addicoat , m. a. ; warr , g. g. ; votchovsky , k. ; atkin , t. 3-dimensional atomic scale structure of the ionic liquid - graphite interface elucidated by am - afm and quantum chemical simulations .
_ nanoscale _ * 2014 * , _ 6 _ , 81008106 li , h. ; atkin , r. ; page , a. j. combined friction force microscopy and quantum chemical investigation of the tribotronic response at the propylammonium nitrate - graphite interface .
* 2015 * , _ 17 _ , 1604716052 black , j. m. ; okatan , m. b. ; feng , g. ; cummings , p. t. ; kalinin , s. v. ; balke , n. topological defects in electric double layers of ionic liquids at carbon interfaces
. _ nano energy _ * 2015 * , _ 15 _ , 737745 fitzner , m. ; sosso , g. c. ; cox , s. j. ; michaelides , a. the many faces of heterogeneous ice nucleation : interplay between surface morphology and hydrophobicity .
soc . _ * 2015 * , _ doi:10.1021/jacs.5b08748 _ | recent advances in experimental and computational techniques have allowed for an accurate description of the adsorption of ionic liquids on metallic electrodes .
it is now well established that they adopt a multi - layered structure , and that the composition of the layers changes with the potential of the electrode . in some cases ,
potential - driven ordering transitions in the first adsorbed layer have been observed in experiments probing the interface on the molecular scale or by molecular simulations .
this perspective gives an overview of the current understanding of such transitions and of their potential impact on the physical and ( electro)chemical processes at the interface .
in particular , peaks in the differential capacitance , slow dynamics at the interface and changes in the reactivity have been reported in electrochemical studies .
interfaces between ionic liquids and metallic electrodes are also highly relevant for their friction properties , the voltage - dependence of which opens the way to exciting applications . solid - liquid interfaces play key role in many processes , such as catalysis or electrochemical reactions , to mention only chemistry and energy related applications . despite their importance , our understanding of the molecular - scale structure of such interfaces , where all the essential ( electro)chemical processes occur , has long remained limited compared to the case of the corresponding pure solid and liquid phases . probing directly
the interface in experiments is indeed particularly challenging and computer simulations are also more involved due to the symmetry breaking in the direction perpendicular to the interface , which hinders the efficient use of periodic boundary conditions . indeed , although simulating a few tens of water molecules may be sufficient to investigate the bulk properties of the liquid @xcite , a similar number of molecules results in finite - size effects different from that , physically relevant , induced by the presence of the interface .
the past 10 years have witnessed the development of many experimental techniques which are sensitive to molecular arrangements at the interface , such as scanning tunneling microscopy ( stm ) , sum - frequency generation ( sfg ) , atomic force microscopy ( afm ) , high - energy x - ray reflectivity ( xr ) or surface force apparatus ( sfa ) .
they probe the structure of the liquid via different observables ( vibrations , electron density , resistance to shear , etc ) , thus providing complementary views of the interface : for example the sfg signal is dominated by the innermost adsorbed layer @xcite , while afm or xr studies probe several layers of fluid .
@xcite in parallel , the access to high performance computers and the development of new algorithms @xcite also allowed to simulate more accurately solid - liquid interfaces , shedding a new light on interfacial processes such as adsorption .
for example , the combination of stm and density functional theory ( dft ) calculations demonstrated that water molecules adsorbed at metal surfaces exhibit a surprisingly rich variety of structures @xcite .
their arrangement depends on the interplay between the geometry and energetics of the water - metal interaction and of the hydrogen bonding between the water molecules , which varies strongly from one metal to another ( and even from one crystal plane of a given metal to another ) and with the water coverage of the surface .
this structuring impacts the dynamics at the interface @xcite and ultimately the kinetics of electrochemical processes . due to the range of length and time scales involved , from the electron transfer event to the local rearrangements of the interfacial fluid ,
a full understanding of the water - splitting mechanisms from computer simulations will therefore require bridging the gap between _ ab initio _
@xcite and classical @xcite approaches . here we will focus on a particular class of electrolytes , namely room - temperature ionic liquids ( rtils )
they are increasingly used in electrochemistry , with applications ranging from energy storage ( batteries , supercapacitors ) to electrodeposition @xcite .
since they are made of ions , their interfacial properties have long been interpreted following the gouy - chapman - stern theory . however , many of the underlying asumptions are not valid due the very high density of ions an extreme case considering the absence of solvent in these liquids @xcite .
a significant number of experiments and molecular simulations have thus been devoted to the study of the interfaces of ionic liquids with a solid @xcite .
the main conclusion arising from xr , afm , sfa and molecular dynamics ( md ) is that the structure perpendicular to the interface is characterized by a strong layering of the liquid @xcite , as expected for a molecular liquid , which extends up to a few nanometres .
the local composition of the layers mostly depends on the surface charge of the solid @xcite and displays strong local correlations due to charge - ordering .
many recent studies on interfaces of rtils reported intriguing results , highlighting the role of the molecular structure within the adsorbed layers . as pointed out in an editorial by kornyshev and qiao , it is indeed necessary to account for the three - dimensionality of the interface @xcite . in particular , the formation of an ordered layer of ions has been reported at the interfaces of 1-butyl-3-methylimidazolium - hexafluorophosphate ( [ c@xmath0mim@xmath1[pf@xmath2 ) with mica @xcite or with vapor @xcite . at electrochemical interfaces
, the contact between rtils with an electrified metal opens the way to voltage - induced ordering transitions within the adsorbed liquid .
the universality of such transitions is far from being established , in particular the extent of concerned rtils - substrate combinations should be clarified .
a first objective of this perspective article is therefore to summarize the studies , both experimental and theoretical , in which such transitions have been observed .
we then discuss the impact of this finding on the physico - chemical properties of the interface .
in particular , the following questions will be addressed : how can we detect structural transitions in experiments and in simulations ?
is there a templating action from the solid ?
what is the main electrochemical signature of these transitions ?
is there an impact on the friction properties of the interface ? some of these questions remain open and call for further studies . |
intermittency is a prominent phenomenon observed in a large variety of nonlinear dynamical systems .
the classical examples for intermittent behaviour , the so - called pomeau - manneville types i - iii @xcite , can be found in deterministic systems where upon a certain change of a control parameter a fixed point of the system ( corresponding to a periodic trajectory ) becomes unstable .
one of the characteristic features for the distinction of the different types of intermittency is the statistics of the duration of the quasi - periodic ( laminar ) phases which are irregularly interrupted by chaotic parts of the trajectory .
a fundamentally different type of intermittent behaviour has been observed in coupled chaotic oscillators @xcite .
this phenomenon can be found in dynamical systems at the stability threshold when a stochastic or chaotic process couples multiplicatively to the system variables .
the term on - off intermittency has been coined for this phenomenon . in systems which exhibit this type of intermittency , there is no sharp transition from an equilibrium quiescent state into an active state but intermittent behaviour occurs for a range of values of the control parameter , and the system has to be characterized by a statistical description .
it resides in a ground ( off ) state during quiescent or laminar periods , which are interrupted by bursts of large scale excursions of the system variables into the on - state . as well as the other types , on - off
intermittency is characterized by fundamental statistical properties of the intermittent process which have been extensively studied during recent years , experimentally as well as theoretically .
a statistical analysis reveals characteristic asymptotic laws that describe the universal behaviour of such systems .
it has been shown that the distribution of the durations @xmath2 of the laminar phases in on - off intermittency follow a characteristic power law with exponent -3/2 @xcite in the vicinity of the instability threshold .
other fundamental scaling laws have been predicted for the distribution of the amplitudes of the bursts @xcite and the power spectrum of the trajectories @xcite . experimental evidence for on - off intermittent
behaviour has been reported in a number of very different physical systems .
a simple experimental realization can be achieved in coupled oscillator circuits @xcite , other systems described in literature involve a gas discharge plasma @xcite and a ferromagnetic resonance spin wave experiment @xcite . while the fundamental validity of the asymptotic scaling laws is established theoretically , it is not easy to confirm this prediction in the experiments . in the spin wave system @xcite , a power law in the distribution of laminar phases has been reported over just slightly more than one order of magnitude in @xmath2 . in the gas discharge experiment @xcite ,
the power law behaviour is covered by an exponential function . among the experimental situations where on - off intermittent behaviour has been unambiguously detected
is electrohydrodynamic convection ( ehc ) in nematic liquid crystals driven by multiplicative noise @xcite .
this system turns out to be particularly well suited for an experimental characterization .
it represents the first spatially extended dissipative system where on - off intermittency has been identified . compared to other reported systems , the ehc experiment shows an additional spatial periodicity of the intermittent bursts where a wavelength selection process is involved .
the access to the control parameters and observation of trajectories is straightforward .
the physical mechanisms are well understood .
many involved material parameters are accessible in independent experiments .
the validity of the asymptotic scaling law for the duration @xmath2 of laminar phases has been confirmed experimentally @xcite . in this study ,
a modified optical setup is used in order to record pattern wavelengths and amplitudes with high sampling rates .
the trajectories of pattern amplitudes are extracted from the laser scattering profile produced by the nematic cell .
in addition , a simulation of the linearized dynamic equations is presented .
trajectories obtained in the simulations are compared to the experimental results to test the validity of a linearized treatment of the system dynamics @xcite near the instability threshold .
nematic ehc @xcite represents a standard system of dissipative pattern formation , its fundamental features are well understood today @xcite .
the instability is driven by interactions of an external electric field with electric charges , present as either impurities or dopants in the nematic material .
the experimental geometry is sketched in fig .
[ figehc ] . the ground state with the nematic director ( optic axis ) uniform along an easy axis in the cell plane
is achieved by proper surface treatment of the cell plates .
an electric field is applied normal to the cell plane .
nematic material with negative dielectric anisotropy @xmath3 is chosen to prevent the splay fredericksz instability . in the electric field , the dielectric torque stabilizes the ground state , but small thermal fluctuations of the director field in connection with an anisotropic conductivity of the nematic generate a periodic space charge modulation in the cell plane .
the interaction of these space charges with the electric field lead to convective flow which in turn generates a destabilizing viscous torque on the director field . above a threshold voltage
u@xmath4 , this torque exceeds the stabilizing dielectric and elastic torques and a periodic pattern of convection rolls and corresponding director deflections is formed . in the simplest case ,
the wave vector of the first instability is along the director ( normal rolls ) .
the roll structure appears in optical transmission as an array of parallel stripes in the cell plane ( fig .
[ figehc ] ) .
electroconvection is conventionally driven with a periodic ac voltage ( to avoid sample degradation in dc fields ) . for the understanding of the mechanism of pattern formation it is essential to note that director and charge fields respond on different time scales to the alternation of the electric field .
the symmetry of the dynamic equations requires that their time dependence with respect to a periodic driving field is antisymmetric .
this is reflected in the existence of two different types of patterns . in the low frequency conduction regime ,
charge relaxation is fast compared to the ac frequency .
the charge density alters its sign with the applied field , the sign of the director deflection is preserved .
the voltage for the onset of conduction rolls increases monotonously with frequency . at the cut - off frequency @xmath5
, the threshold voltage curve intersects that of the dielectric patterns . in the dielectric regime above @xmath6 , the director deflections alternate with the field while the charge density modulation retains its sign .
figure [ figthresholds ] shows the stability diagram for the system studied in this paper .
figure [ figthresholds](a ) gives the onset voltage for the first instability , towards normal rolls .
the corresponding wave numbers are shown in figure [ figthresholds](b ) . when a stochastic excitation scheme is used where the driving field has no deterministic component ( as e.g. the dichotomous markov process that is considered in this paper ) the system does not exhibit a sharp transition from the quiescent to the convective state upon variation of the control parameter but shows intermittent behaviour .
two different regimes are found which have many features in common with the corresponding conductive and dielectric regimes in the deterministic case . at onset of the instability ,
one of the system s characteristic times ( director or charge relaxation ) becomes comparable to the characteristic time of the noise @xmath7 .
a considerable qualitative change of onset and appearance of the convection patterns @xcite is observed .
a typical snapshot is shown in fig .
[ figortho ] . nevertheless , there was little interest until recently in the quantitative statistical interpretation of the structures at the instability threshold .
main focus of research has been directed to the study of superimposed deterministic and stochastic driving fields and the construction of pattern state diagrams .
the empirical concept of a threshold voltage has been applied in previous experimental investigations of noise driven ehc @xcite and statistical methods have been applied to test various stability criteria @xcite .
the largest lyapunov exponent @xmath8 of the trajectories of the system variables , which is analytically known @xcite , provides a quantitative measure for the instability threshold , but the crucial problem is the experimental determination of the threshold in a system with limited dynamical range and additive noise .
the statistical analysis of the intermittent trajectories provides an experimental tool to characterize the stability threshold , @xmath9 @xcite . the observation and quantitative characterization of the dissipative patterns under stochastic excitation
is achieved in this study by exploring the phase grating formed by the spatially periodic deformed director field .
the laser scattering profile resulting from the nematic convection patterns is recorded .
a two - dimensional ( 2d ) detector allows to observe the complete scattering profile of the transmitted laser light and to study the wave vector orientation and wavelength selection process with a sampling rate of 25 s@xmath10 , and alternatively a photo diode positioned at the scattering reflex enables us to record the trajectory of the dominant mode with faster sampling rate and higher intensity resolution .
we use the nematic mixture _ mischung 5 _ ( halle ) which consists of four disubstituted alkyl(oxy)phenyl - alkyloxybenzoates .
this material has been used in previous ehc experiments @xcite .
the mixture is nematic at room temperature , its clearing point is 70.5 @xmath11c .
the commercial sandwich cells ( _ linkam _ ) used in the experiments have cell gaps of 25.8 @xmath12 m .
the glass plates are covered with transparent indium - tin - oxide electrodes ( 5 mm@xmath135 mm ) , they are polyimide coated and antiparallelly rubbed for planar surface alignment .
the temperature of the samples is controlled by a linkam heating stage with an accuracy of 0.1 k. the sample temperature has been set to 32@xmath14c in all experiments .
all relevant material parameters for the simulation of the electrohydrodynamic equations except for some viscosities have been measured in independent experiments @xcite .
the conductivity of the nematic samples differs by about 20% between individual cells , values for a given cell are almost constant in time . in order to prevent long - term trends of the conductivity ,
we reheat the material into the liquid phase between subsequent runs of the experiment , similar procedures have been proposed in @xcite .
all measurements presented in the diagrams of this paper have been performed consecutively with the same cell , in order to obtain quantitatively comparable results for the different statistical investigations .
differences in conductivity between individual cells may lead to variations of the respective thresholds but do not influence the statistical characteristics . in order to complete the parameter set for the numerical simulations , we fitted the threshold curves and critical wave numbers for deterministic square wave excitation where the experiment yields sharp thresholds towards the first instability . with a fixed parameters set ( table i )
, good agreement between numerical results and experimental data is achieved across the whole frequency range investigated ( fig .
[ figthresholds ] ) .
the same parameter set is used thereafter for the calculation of the lyapunov exponents and stochastic trajectories .
the waveform of the driving electrical voltage is synthesized by a computer and subsequently amplified .
we generate the driving voltage curve with a sampling rate of 10 khz .
although the spectrum of this waveform is in principle discrete , we can consider it as quasi - continuous in the frequency range relevant in the experiments ( below @xmath15 hz ) . within this study , we have used a few special waveforms , which are detailed in the following , figure [ figexcitations ] visualizes their signal shapes and shows the respective power spectra .
_ periodic square wave _ excitation is used to construct the stability diagram for deterministic driving .
we exploit the sharp threshold towards the first instability for the adjustment of system parameters and test of the long term stability of the samples .
for example , the cut - off frequency is a sensitive measure to reveal even small changes of the sample conductivity .
the _ dichotomous markov process _
( dmp ) is the stochastic waveform used in all experiments presented below .
it is characterized by random jumps of the electric field between the two values @xmath16 and @xmath17 , with an average jump rate @xmath18 .
the time intervals between consecutively jumps are distributed as @xmath19 , where @xmath20 $ ] is a uniformly distributed random number . in analogy to a deterministic square wave excitation with frequency @xmath21 , we define the average frequency @xmath22 .
the dmp power spectrum is lorentzian with its maximum at frequenzy zero and a half width related to the jump rate by @xmath23 ( fig .
[ figexcitations](b ) ) .
an important feature that facilitates the numerical calculations is that all terms quadratic in the electric field are time independent in the dmp .
an analytical calculation of the sample stability threshold and the critical wave numbers have been performed @xcite . technically , identical realizations of the stochastic process can be reproduced with the computer .
this allows us to use identical noise sequences for experiment and simulations when details of the trajectories of pattern amplitudes are of interest .
_ other stochastic processes _ have been tested in additional experiments .
while the dmp randomizes the phases of a periodic square wave , another stochastic process can be synthesized that randomizes the amplitudes of the square wave but keeps the jumps equidistant .
this randomization of the amplitude can be combined with a random phase of the jumps .
such processes do not complicate the numerical simulations , the electric field is piecewise constant and the equations of motion can be integrated straightforwardly .
it turns out that the statistical properties of the trajectories and the derived scaling laws for these stochastic processes are in full accordance with those for the dmp .
therefore , we will discuss only the dmp results being representative for other types of stochastic excitation .
the sample cell is irradiated by a he - ne laser with wavelength @xmath24=632.8 nm at normal incidence .
the beam diameter is about 1 mm . the scattered light is monitored on a screen in a distance of maximum 1.3 m from the cell ( fig .
[ figsetup ] ) .
when the sample is in the ground state ( zero electric field ) , only a weak background scattering @xmath25 is observed around the primary beam .
since we are interested in scattering from spatially deformations of the director field , we correct the raw data with a time averaged intensity @xmath26 by @xmath27 .
this correction is marginal since the background signal is in general three orders of magnitude smaller than the amplitude of the scattering reflex at the position of the photo diode . from the two - dimensional scattering images , wavelengths and orientations of the patterns
can in principle be continuously extracted . however , because of bandwidth limitations in signal processing we use two different equipments to record the scattering data .
a commercial video camera is employed to take 2d scattering images with an acquisition rate of 25 frames per second and 8 bit intensity resolution .
this enables us to study the evolution of the mode spectrum and to access the full wave vector information , although time and intensity resolution are limited .
the 2d images show that at given frequency , the pattern is dominated by a single mode ( see fig .
[ figscatter ] ) with fixed wave vector but varying amplitude .
any low order scattering reflex of that mode is representative for the momentary pattern , and it is sufficient to record only the scattering intensity at a fixed position ( here , we use one of the two symmetric second order reflexes ) . for that purpose
we employ a photo diode with aperture 7 mm@xmath28 adjusted to the reflex of interest .
the photo diode signal is digitized and the trajectory is stored in a computer .
we use a fast ( 0.2 ms time resolution ) 12 bit ad concerter for measurements with high time resolution as for example the determination of laminar phases , and alternatively a slow ( 0.1 s ) 24 bit adc for accurate amplitude measurements over large dynamic ranges .
the electroconvection rolls in the nematic material produce spatial periodic director deflections @xmath29 which lead to a modulation of the effective refractive index @xmath30 for the transmitted extraordinary beam , @xmath31 where @xmath32 are material parameters and @xmath33 is associated with the angle between the light ray and an optical axis @xcite .
a phase and amplitude grating is formed in the cell .
there has been some discrepancy in literature about the usage of refractive or ray index in these calculations . for the small fluctuations considered here ,
both approximations lead to similar results . in order to establish a relation between the experimental observations and the results of the simulations of the director and charge fields ,
it is necessary to calculate the diffraction efficiency of a given periodic director field .
we use the eigenfunctions @xmath34 from the linear stability analysis . using fermat s principle
it is possible to determine the optical path , the resulting phase difference @xmath35 and the intensity profile of light passing the cell @xcite . at small deflection angles
, one can assume that individual light rays pass the cell without deflection , creating a phase grating @xmath36 \text{d}z .
\nonumber\end{gathered}\ ] ] analytical integration over @xmath37 along a straight path through the cell @xcite yields a periodic phase modulation @xmath38 with twice the wave number @xmath39 and a quadratic dependence on the director deflection @xmath40 in this approximation , the intensity modulation with @xmath39 due to focussing effects of the inhomogeneous refractive index profile is neglected and only even order reflexes appear in the scattering image .
this is in agreement with the experimental observations .
the second order reflex dominates the small amplitude patterns , and with increasing amplitude of the director deflections @xmath41 , higher order reflexes can be observed .
we note that the amplitude grating ( which produces the well known shadowgraph images in conventional orthoscopic microscopy ) is effective as well , it is most prominently reflected in the weak first order reflexes ( see fig .
[ figscatter ] ) .
the relation @xmath42 connects the scattering angle @xmath43 of the @xmath44th order reflex with the wave number @xmath45 .
the scattered light intensity @xmath46 of the phase grating at the second order reflex is related to the square of the bessel function of first kind , @xmath47 , with the amplitude of the phase grating @xmath48 in the argument @xcite . for small director deflections @xmath41 ,
the intensity at the second order reflex is proportional to the fourth power of @xmath49 @xmath50 by numerical integration of the nonlinear euler - lagrange equations we have calculated the actual propagation of light beams through the sample .
the numerical integration allows for the deflection of light and thus for intensity modulations by focussing effects .
these simulations confirm the relation @xmath51 even for large director deflections @xmath52 .
the numerical simulation of ehc patterns yields @xmath53 , ( see below ) which represents the director deflection for a given mode . for comparison of simulated @xmath49 and experimental intensities
we use the relation @xmath54 we emphasize that we compare the experimental and simulated pattern amplitudes only on a relative level , since no efforts have been made to relate calculated absolute scattering intensities to voltages measured by the photosensor . since the simulations use a linear model , absolute scaling of pattern amplitudes is not relevant for the fundamental statistical properties of the system .
the basic equations for the charge and director fields and the method for analytical derivation of the lyapunov exponents in the electrohydrodynamic instability have been described in detail in @xcite .
the important quantities describing the system dynamics are the charge density @xmath55 and the amplitude of the director deflection @xmath56 ( see fig . [ figehc ] ) .
a standard technique to describe the time evolution of small amplitude periodic patterns is the use of linearized differential equations and a two - dimensional mode ansatz for the two involved quantities , @xmath57 in the relevant parameter range , the pattern is spatially periodic along the director easy axis ( @xmath58 direction ) this result of the linear stability analysis is in agreement with the experimental observations .
therefore , we consider only the stability of modes with the wave vector parallel to the @xmath58 axis .
the ( stress free ) boundary conditions for the director field at the glass plates @xmath59 are satisfied by @xmath60 for convenience , the director deflection @xmath56 is substituted by the curvature @xmath61 .
a system of two ordinary differential equations is derived from the torque balance , navier - stokes and maxwell equations .
after linearization we obtain an ordinary differential equation system in time for the vector @xmath62 , @xmath63 where @xmath64 and @xmath65 are parameters related to the viscous , elastic and electric properties of the liquid crystal as well as to the wavelength @xmath39 of the modes @xcite .
the electric field amplitude @xmath66 @xmath67u@xmath68 corresponds to the excitation voltage u@xmath69 . in case of piecewise constant excitation ( as the deterministic square wave and dmp described above ) ,
@xmath66 assume only two values @xmath70 , and all elements of the matrix are constant between consecutive jumps . within a time interval @xmath71 of constant excitation field , integration of the differential equation
gives a solution in the form of a sum of two exponentials .
the solution @xmath72 from the @xmath73th integration step is taken as the starting value for the @xmath74st step .
the complete trajectory for given @xmath39 , @xmath75 and set of ( random ) times @xmath71 for jumps of the excitation voltage is calculated with an initial vector @xmath76 . at the discrete jump times @xmath77 ,
the solution is @xmath78 where @xmath79 is the 2@xmath802 time evolution matrix for the @xmath73th interval . for a detailed statistical analysis of @xmath81 in particular for the calculation of power spectral densities ( psd ) , the trajectory can be filled in the intervals between the jumps using the exact exponential solutions @xmath82 . in the particular system
studied here , the eigenvalues @xmath83 are real and positive . for periodic square wave driving , all intervals @xmath71 are equal and the product is @xmath84 . for the calculation of the lyapunov exponents
it is sufficient to consider @xmath85 .
this reproduces the well known results of classical theory using floquet methods @xcite . in the case of stochastic excitation
all the @xmath71 are different and the calculation of the lyapunov exponents leads to an infinite product of 2@xmath802 random matrices @xcite . this system yields two real lyapunov exponents @xmath86 which are related to the eigenvalues of the product of stochastic matrices @xmath87 .
in particular , the largest lyapunov exponent ( in the following denoted by @xmath8 ) is found from @xmath88 for dmp excitation , the lyapunov exponents can be obtained analytically @xcite .
figure [ figlambda ] shows the lyapunov exponent for the critical wave length calculated with the parameters specified above . in the linear model , @xmath89 is growing to infinity ( @xmath90 ) or shrinking to zero ( @xmath91 ) , depending on the value of the largest lyapunov exponent .
the selected wavelengths and threshold voltages for a given frequency and set of material parameters are calculated from the neutral curve .
the wave number is varied and the minimum @xmath92 of the neutral curve provides the critical wave number @xmath93 .
because of the symmetries of eq .
( [ basicqpsi ] ) , one of the system variables ( @xmath94 ) keeps its sign while the other variable has to change its sign with the polarity of the applied field . at periodic excitation ,
the system is synchronized with the applied field and the conduction ( @xmath95 alternating periodically ) and dielectric ( @xmath96 alternating periodically ) regimes are distinguished . in case of dmp excitation , the support of one of the variables still preserves the sign and the two regimes can still be distinguished @xcite . in the following , we will discuss only the low frequency conduction regime , fig .
[ figqpsi5e3 ] shows details of a simulated trajectory .
the slow variable is the director deflection which is directly related to the measured quantity , the scattering intensity . due to the coupling of ( @xmath94 ) ,
both the director deflection and the envelope of the charge density curves show the same long term ( @xmath97 ) characteristics . in the linear description ,
the trajectories tend to infinity or zero for all values of @xmath98 . however , a realistic description of the experiments has to consider boundaries for @xmath99 .
nonlinearities in the system limit the excursion of the system variables @xmath99 to large values while additive noise prevents their unlimited decay . in order to compare the results of simulations and experiments , we introduce limits for the director deflection ( curvature @xmath100 ) by clipping each time step @xmath101 because of the linearity of the ode system ( eq .
( [ basicqpsi ] ) ) , only the ratio of the upper and lower thresholds is important .
a constant factor in the amplitudes is irrelevant for the statistical properties and scaling laws . here
, we assume that the dynamic range is two orders of magnitude , @xmath102 .
this dynamical range reflects approximately the situation of a thermal background stimulation @xmath103 mrad @xcite and an upper limit of 0.5 rad . for negative lyapunov exponents ,
the lower boundary is necessary to prevent the unlimited decay of @xmath104 .
the choice of the value of the lower boundary has strong effect to the number of burst per unit time ( see fig .
[ figtrajbackground ] ) , but only small effects on the fundamental statistical laws ( see below ) .
the assumption of a well defined lower limit is of course artificial and can not describe the actual experimental behaviour for very small pattern amplitudes .
a more realistic assumption considers low amplitude additive noise in the variable @xmath105 .
we have studied this case by adding gaussian random numbers with a given variance @xmath106 @xcite at each time step . for zero electric field ,
( [ basicqpsi ] ) decouple and @xmath100 describes an ornstein - uhlenbeck process ( oup ) which is characterized only by @xmath106 and the exponentially decaying autocorrelation with the correlation time @xmath107 related to the lyapunov exponent at zero voltage ( see fig .
[ figlambda ] ) .
one consequence of such additive noise is that the simulated trajectories for the slow variable @xmath108 can change sign ( see fig .
[ figqpsi5e4 ] ) , in contrast to eq .
( [ boundarypsi ] ) .
this is not relevant for the comparison with the scattering experiment which is insensitive for the spatial phase of the pattern .
if the variance @xmath106 is chosen such that the mean square amplitude of @xmath100 in absence of electric fields gives @xmath109 , the statistical properties of @xmath110 are qualitatively identical with the simulations using conditions ( [ boundarypsi ] ) .
the conventional technique to observe pattern formation in ehc is the shadowgraph method @xcite in combination with a transmission microscope .
an instant picture of a pattern burst , recorded by means of orthoscopic microscopy , is shown in fig . [ figortho](a ) .
the dynamics of this pattern can be visualized best when the intensity in a cross section perpendicular to the rolls is scanned and a spatio - temporal plot as in fig .
[ figortho](b ) is analyzed .
some bursts , in particular those which appear in fast sequence , are correlated in their spatial phase .
after long laminar phases , however , there is in general no spatial correlation remaining between the bursts . for stationary rolls ,
this loss of correlation is a consequence of additive noise in the system , it triggers new modes whenever the pattern amplitude reaches thermal noise level . in such cases ,
the new fluctuation mode has an arbitrary spatial phase with respect to the previous convection pattern . on the other hand , there are sequences of bursts in fig .
[ figortho](b ) which preserve spatial correlation .
they can be attributed to amplifications of the same mode which has disappeared in the optical image but has not reached the level of additive noise during the intermediate laminar phase .
figure [ figscatter](a ) shows a snapshot of the scattering image originating from a burst in a dmp driven d=25.8 @xmath12 m thick nematic cell , taken with the 2d camera detector .
the image shows the primary beam at @xmath111 and diffuse scattering spots from the periodic spatial director modulation in the cell .
the scattering reflexes concentrate on the @xmath112 axis , i.e. normal electroconvection rolls are observed .
a wave number @xmath112 of the pattern producing this image of 0.2 @xmath12m@xmath10 is derived from the spot positions , the director period is @xmath113 @xmath12 m .
below the scattering image , in fig [ figscatter](b ) , the time dependence of the intensity profile taken along the horizontal symmetry axis of the profile ( @xmath114 ) is plotted . for better reproduction ,
the image is plotted with inverse gray scale , dark spots correspond to high amplitudes of scattered light and consequently to large director modulations ( bursts ) .
the bursts are characterized by a narrow wavelength band and the reflexes remain approximately at the same positions , i.e. in all bursts the patterns have nearly the same wavelengths .
the wavelength selection process itself is not observable in the images , it is obviously fast compared to the video rate of 25 hz and occurs below the level of optical sensitivity of our camera .
the information about the spatial phase of the patterns gets lost in the scattering image , therefore we can not determine from the scattering images whether the convection rolls of subsequent bursts appear spatially phase correlated or not . the essential information taken from the 2d images
is that the wavelength of the noise driven patterns is constant and the scattering image consists of reflexes at fixed scattering angles with varying amplitudes .
thus , we achieve a considerable data reduction by restricting to the strongest scattering reflex .
in particular , the detector is placed at the second order scattering reflex of the most critical wave number , indicated by the arrow in fig .
[ figscatter](b ) .
the intensity of scattered light at the position of the second order scattering reflex of the first unstable mode is shown in fig . [ figtrajectories ] .
the curves have been digitized by means of the 24 bit ad converter .
the random electric excitation field @xmath115 uses identical realizations of the stochastic process for all three trajectories , with different amplitudes @xmath75 . concerning the effects of the multiplicative noise in the system , details of the three curves can be directly compared .
the characteristic frequency @xmath116 was 80 hz , corresponding to an average of 160 jumps / s of the sign of @xmath115 .
[ figtrajectories](a ) shows the raw signal from the detector for a voltage below the critical u@xmath117 .
the contributions of ( additive ) background fluctuations to the detector signal are of the order of @xmath118 v around a constant offset @xmath119 v. bursts of the spatially periodic pattern that exceed the background noise level occur infrequently . from arguments discussed in the next sections we conclude that the excitation voltage is below the stability threshold ( defined by the lyapunov exponent @xmath9 ) . in fig .
[ figtrajectories](b ) , the voltage is approximately equal to the critical voltage ( @xmath9 ) .
the characteristic feature of the trajectory is its mirror symmetry of high and low amplitude excursions of the scattering intensity @xmath46 on logarithmic scale .
rise and decay sections of the graph are symmetric .
we note that in a linear presentation of the same plot @xmath46 , the high amplitudes will appear as prominent bursts out of the background level , and the typical intermittent behavior is acknowledged . at higher voltages ( fig .
[ figtrajectories](c ) ) , the trajectory will be predominantly in the saturation region ( on state ) , interrupted by short laminar phases .
these intrinsic symmetries reflect theoretical predictions for on - off intermittent behaviour @xcite .
amplitudes in figs .
[ figtrajectories](b , c ) have been corrected for the background intensity , @xmath120 .
this correction attempts to separate the constant ( stray light ) background signal and additive ( thermal ) noise in the trajectories .
since these contributions are comparably small , the correction effects only the low amplitude sections of the trajectories .
it faciliates the identification of laminar phases and enables us to visualize the symmetry of burst and quiescent periods in the logarithmic scale .
the statistics of experimental and simulated trajectories can be compared on a quantitative level . in the experiment ,
the distribution of laminar phases is calculated by introducing an arbitrary threshold intensity @xmath121 , and the durations of periods where the intensity curve stays completely below that threshold are determined .
a threshold @xmath122 v has been chosen here , it corresponds to the geometrical average of lower and upper bound of the photo sensor signal . as expected in on - off intermittency , the choice of the actual threshold value is not critical .
[ figlamphasen ] ( left ) depicts the distribution of the laminar phase durations extracted from trajectories for dmp excitation with three different voltages , the same values as in fig .
[ figtrajectories ] .
the time axis is scaled with the jump rate @xmath123 of the dmp .
these distributions were extracted from 4200 sec runs of the experiment , each trajectory has been recorded with a sampling rate of 1000 s@xmath10 ( with the fast ad converter ) , so that a range of 6 orders of magnitude in @xmath1 is resolved .
the dash - dot line indicates a @xmath0 dependence , which is predicted theoretically exactly at the sample stability threshold , @xmath9 , when no additive noise is present . in the short time range , for @xmath124 ,
the curve deviates from this predicted fundamental dependence , because one approaches the time scale of the driving dmp process and specific details of the driving process become important . in the long time
limit of the curve , the power law behaviour of the experimental trajectories breaks down mainly because of the lower boundary ( additive noise level ) for the system variables . for long periods
at least one of the variables ( @xmath125 ) reaches a ( thermal ) noise level which prevents excursions of the system variables to values much below @xmath121 , trajectories are practically reflected there .
this lead to faster injections of the next burst above @xmath121 , and thus to a lower probability of long duration periods . the flat shoulder indicated in fig .
[ figlamphasen ] is the outcome of this effect @xcite .
figure [ figlamphasen ] ( right ) shows the results of the corresponding simulations .
limits to the system parameters were set as given in eq .
( [ boundarypsi ] ) .
the voltages used in the simulation are u@xmath126=14.2 v ( corresponding to the critical value @xmath9 ) , and [email protected] v chosen in the vicinity of this threshold .
the critical voltage in the simulated curves is derived from the analytically calculated lyapunov exponent @xcite .
it is in perfect agreement with the numerical simulation of the trajectories in absence of upper and lower limits . in the experiment ,
it has been proposed earlier that a reasonable definition of the critical voltage can be found when the statistical distribution of the laminar phases duration is analyzed .
thus we assume that the experimental critical driving voltage u@xmath128 is reached when the distribution of laminar phases durations is most adapted to a @xmath0 dependence @xcite .
knowledge of the relation ( [ psitoa ] ) between director deflection amplitudes and scattering intensities is not necessary for the determination of the distribution of laminar phases . on the other hand ,
the laminar phase distribution is not the most sensitive criterion for the determination of the sample stability threshold as will be shown in the next section .
another fundamental prediction for the statistics of on - off intermittent processes is the distribution of amplitudes in the trajectory .
it has been shown @xcite that the distribution of the amplitudes of the intermittent variable @xmath129 should follow a power law @xmath130 in the vicinity of the threshold @xmath9 .
the parameter @xmath131 vanishes at the threshold of the instability .
a statistical analysis of the recorded trajectories requires knowledge of the quantitative relation ( [ psitoa ] ) between the measured scattering intensity @xmath132 and the amplitude of the pattern @xmath49 .
one can easily show that a similar power law as for the amplitude distributions of the intermittent variable @xmath49 holds also for quantities @xmath133 that depend on the @xmath44th power of @xmath49 .
when @xmath134 and @xmath135 then @xmath136 the hyperbolic dependence at the threshold @xmath9 is preserved .
our observable @xmath46 , the scattered intensity in the @xmath137nd order reflex , is related to the director deflection amplitude in first approximation by eq .
( [ psi4 ] ) , and this relation provides another opportunity to determine the critical voltage u@xmath4 .
[ figampl ] shows the distribution of scattering amplitudes @xmath46 for zero driving voltage and four representative voltages in the vicinity of the stochastic threshold u@xmath4 .
a power law can be fitted to the middle part of all these distributions . for low scattering amplitudes
, the curve deviates from the fit because of the superimposed background scattering . for large amplitudes ,
the power law breaks down because of the saturation of the system variables and because the optical characteristics , eq . ( [ psi4 ] ) , is not valid for large director deflections . in the numerically simulated trajectories , fig .
[ figamplsim ] , where hard boundaries , cf .
( [ boundarypsi ] ) have been used , these effects are condensed in the edges of @xmath138 .
@xmath139 is computed from @xmath49 by means of eq .
( [ psitoa ] ) . in fig .
[ figampl](b ) , the amplitude for u=12.9 v is closest to the exponent -1 in the power law , therefore we assign this voltage to the experimental threshold voltage u@xmath128 .
this value is in good coincidence with the critical voltage found from the best fit of the laminar phase distributions to a @xmath0 law .
both statistical definitions of the experimental threshold voltage agree consistently , and the numerical simulations confirm the equivalence of the thresholds determined from the distributions of amplitudes and laminar phase durations . at voltages near the threshold , the power law still holds and the exponents @xmath140 can extracted .
typical amplitude distributions are depicted in figs .
[ figampl](c , d ) .
the exponents extracted from experimental data as well as from simulated trajectories are shown in fig .
[ figeta ] . in accordance with the natural scales of the system ,
the axis was normalized with the critical voltage to obtain a control parameter @xmath141 .
the optical relation ( [ psitoa ] ) has been applied to retrieve pattern amplitudes @xmath142 which can be compared to the simulation from experimental scattering intensities @xmath132 .
the good agreement in the slopes of the @xmath143 curves at @xmath144 justifies the application of the optical relation eq .
( [ psitoa ] ) .
finally , we discuss the power spectrum of the intermittent process .
the main general theoretical prediction for the power spectral density ( psd ) for @xmath9 is a square root frequency dependence @xcite in a certain frequency range . if any quantity with a power law dependence on the intermittent variable is observed , this prediction is equally valid @xcite .
the psd predicted for a process driven with multiplicative noise is qualitatively different from a process where noise couples additively to the system variables . in the range of very small frequencies , a constant psd
is expected because any time correlation in the system variables is destroyed by additive noise and the limited dynamical range of ( @xmath145 ) . in the high frequency tail ,
a @xmath146 dependence is expected , similar to that of a simple stochastic process with exponentially decaying autocorrelation function .
the relation for the high frequency limit @xmath147 can be obtained analytically ( from eq .
( 50 ) in @xcite ) @xmath148 the crossover frequencies @xmath149 , @xmath150 where the asymptotic exponent changes depend upon the lyapunov exponent and specific properties of the additive and multiplicative noise .
[ figpsd ] displays the psd of experimentally recorded trajectories @xmath46 for three voltages ( same as in fig .
[ figlamphasen ] ) .
it was obtained from fourier transformation of 4.2@xmath151 data points of the optical trajectories in an interval of 4200 s. in the logarithmic plot , the curves have been smoothed by averaging the spectral energy over intervals @xmath152 proportional to @xmath153 . at the threshold voltage @xmath154 ( b ) , the @xmath155 and @xmath156 regions
are well separated , and the existence of a constant psd in the low frequency wing seems to be indicated . in the high frequency wing , influences of the mean frequency of the driving process , @xmath157
hz , are observable .
the psd obtained from the numerical simulation ( see fig . [ figpsdsim ] ) are not in agreement with the experiment , in particula the square root dependence is not reproduced .
only when the dynamical range is chosen unrealistically large , we obtain a psd @xmath158 .
the choice of a realistic lower bound ( eq . ( [ boundarypsi ] ) ) , i.e. additive noise with reasonable amplitude , destroys any long time correlations .
therefore , a simulation of the psd with similar parameters as in figs . [ figamplsim ] yields a pronounced constant low - frequency region .
on - off intermittency in stochastically driven electroconvection of nematic liquid crystals in the conductive regime has been investigated experimentally and by numerical simulations .
results have been presented for excitation with the dichotomous markov process , but the resulting fundamental statistical behaviour is qualitatively similar for many other types of stochastic excitations . laser scattering has been used to determine the wavelengths and time resolved pattern amplitudes . it has been shown experimentally , and confirmed in the simulation of the electro - hydrodynamic equations that under stochastic excitation the pattern selects its wavelength within a narrow band , therefore the intensity of scattered laser light at a fixed scattering angle can used to characterize the temporal behavior of the pattern amplitude .
the resulting trajectories have been analyzed quantitatively and their statistical properties have been extracted .
the statistical analysis confirms that the distribution density of laminar phase durations @xmath1 is in full agreement with theoretical predictions .
in particular , the @xmath0 power law describes the statistics of laminar phase durations at the stability threshold in the conduction regime over four decades in @xmath1 .
the distribution density of pattern amplitudes @xmath132 in the vicinity of the instability threshold is also in quantitative agreement with the predicted power law .
deviations are found in the experiment in the limits of low and high pattern amplitudes where additive noise ( small a ) and nonlinearities in the dynamic equations ( large a ) influence the dynamical behaviour of the system variables . with increasing voltage the exponent @xmath140 increases .
theory predicts a relation @xmath159 for on - off intermittence , which allows us to define a critical voltage @xmath154 from the slopes of the amplitude distribution functions .
the critical voltage obtained with this method agrees with the value derived from the analysis of the laminar phase durations , i.e. the two definitions characterize consistently the experimental system as well as the numerically simulations .
this provides , irrespective of the fact that the lyapunov exponent is not directly accessible in the experiment , a quantitative criterion for the stability threshold @xmath154 of stochastically driven ehc patterns . from a practical aspect ,
the distribution density of the pattern amplitudes is the more sensitive measure for the determination of @xmath154 . in the simulations of the corresponding model system , trajectories which are characterized by the @xmath0 and @xmath160 power laws
are obtained when the driving parameters are chosen such that the lyapunov exponent is zero ( sample stability threshold @xcite ) .
this result coincides with theoretical predictions of the universal behaviour of on - off intermittency @xcite . in the vicinity of the critical voltage
the linear dependence of the exponent @xmath140 on @xmath8 is also in quantitative agreement with the numerical simulation . in the experiment
, we can not relate the exponent of the amplitude distribution to a lyapunov exponent .
however , the functional dependence of @xmath161 on the reduced voltage @xmath141 is in satisfactory agreement with the calculated data ( see fig . [ figeta ] ) .
we note that although the statistical characterization of experimental and simulated data are in good quantitative agreement , the experimental and calculated threshold voltages and wave numbers can differ on an absolute scale by roughly 10% ( see fig .
[ figthresholds ] ) .
this is mainly the consequence of the simplified assumptions of director and flow modes in the model , it is not relevant for the description of on - off - intermittent behaviour . in the power spectrum of the experimental trajectories ,
a @xmath155 dependence is indicated in a small frequency range , in agreement with predictions of general theories @xcite , in the high frequency tail , the psd adopts an @xmath156 behavior ( fig .
[ figpsd ] ) . in the numerical simulations with boundaries to the system variables , ( fig .
[ figpsdsim ] ) we did not find the @xmath155 dependence , it is reproduced only when boundaries of the trajectory are disregarded and a simulation with quasi unlimited dynamical amplitude range is performed .
this discrepancy leads us to the conclusion that the psd is particularly sensitive to additive noise and the full nonlinear dynamical equations .
the appearance of the @xmath155 range in the experimental data in apparent agreement with general predictions should therefore not be overestimated .
we note that although the investigated experimental system represents a spatially extended dissipative system , it has been shown in this study that its fundamental statistical properties can be reproduced in simulations of a 2@xmath802 evolution matrix model with global pattern amplitude , thus neglecting spatial details of the pattern . in the case of a global driving parameter
, the spatial phase does not play a role as long as other noise sources are excluded .
additive noise , however , may introduce phase drifts in the system @xcite and is responsible for a complex spatio - temporal characteristics .
a detailed spatio - temporal description of the system represents an ongoing interesting task .
.material parameters ( cgs ) used in the simulations .
the parameter set for the simulations of stochastic trajectories is taken from the fit of threshold voltage and wave number characteristics for periodic ac driving , see also fig .
[ figthresholds ] . experimentally data for _ mischung 5 _ taken from @xcite [ cols= " < , < , > , < , > , < " , ] .
periodic excitation : experiment ( @xmath162 ) , theory ( * --- * ) , stochastic excitation : experiment ( @xmath163 ) , theory ( * ----- * ) .
the method for the experimental determination of stochastic thresholds is explained in the text .
the symbol @xmath164 indicates the frequency , where the stochastic measurements presented below have been performed . ]
s@xmath10 ( jumps per sec ) dmp excitation recorded with ccd camera .
( a ) 2d snapshot during an intensive burst at @xmath165 s , ( b ) scattering angle - time plot of intensity profiles taken at @xmath166 , presented in inverse grey scale .
the constant angles of the individual reflexes in subsequent bursts reflects the fast and stable wavelength selection mechanism .
an arrow marks the position of the photo diode , set to the most intense scattering reflex of the most instable wave number . ]
of the most instable mode , @xmath167 @xmath12m@xmath10 , for dmp excitation with @xmath157 s@xmath10 .
material parameters are taken from the fit of periodic excitation thresholds and wavelengths .
the value @xmath9 defines the critical voltage @xmath168 v. ] and @xmath95 at @xmath157 s@xmath10 , at @xmath9 ( conductive regime ) for a constant lower boundary @xmath169 , see eq .
( [ boundarypsi ] ) . in the low frequency regime
the slow variable @xmath104 keeps its sign , whereas @xmath170 oscillates synchronously with the applied field @xmath115 . ] for different constant background levels ( a ) @xmath171 and ( b ) @xmath172 .
the trajectories appear to be significantly different , whereas the statistical analysis produce the same fundamental power laws . ] but with background additive noise in @xmath104 . with additive noise , @xmath104 can change its sign occasionally . the pattern amplitude as well as
the optical scattering intensities are , however , insensitive to the sign of @xmath100 . ] .
( a ) at low voltage ( u=12.2 v ) the measured raw signal @xmath173 fluctuates around a background level @xmath26 , interrupted by infrequent bursts .
( b ) corrected intensity @xmath27 at u=12.9 v where we assume @xmath9 ( see statistical analysis ) .
the up - down symmetry of the curve is recognized .
the experimental dynamic range is limited by saturation for large amplitudes and by background noise for small amplitudes .
( c ) at u=13.6 v the on - state is dominant , intermitted infrequently by breakdowns to the quiescent off - state . ] of durations @xmath1 of laminar phases from experimental ( left ) and simulated ( right ) trajectories , voltages as in figs .
[ figtrajectories](a - c ) . here
the trajectories were recorded in 12 bit resolution with a 1 ms sampling rate .
the time axis was scaled with the jump rate @xmath174 .
the threshold @xmath121 is set to 0.05 v. the @xmath0 power law holds over several decades , best agreement is found at @xmath175 which we assign to @xmath9 . ]
measured over 4200 sec for same dmp voltages as in figs .
[ figtrajectories](a)-(c ) .
( d ) for u=11.8 v ( large negative @xmath8 and ( e ) background at zero excitation voltage . in their middle parts , the distributions fit to a power laws @xmath176 . from the curve with slope
-1 we find the experimental voltage corresponding to @xmath9 . ] of the simulated trajectories .
the power law holds in the complete dynamical range , the fixed boundaries ( eq . ( [ boundarypsi ] ) )
generate abrupt edges in the support of @xmath177 .
the critical voltage u@xmath178 v deviates somewhat from the experimental value .
the graphs represent u=14.2 v ( solid ) , 14.9 v ( dotted ) and 13.5 v ( dashed ) . ] ( @xmath162 experiment , * --- * simulation ) .
the empirical error bars illustrate the uncertainty of the fit and variations between individual runs of the measurements . for @xmath179 , burst appear infrequently and the statistic is rather poor . a scattering efficiency according to eq .
( ) has been used to relate measured optical data to the simulations ( @xmath180 ) and assumed the dash - dotted line the lyapunov exponent is zero and we correspond this voltage to @xmath141=0 .
bigger symbols indicate the applied voltages depict in fig .
[ figampl ] . ] .
ad - converter with 1 ms sample time and 12 bit resolution was used .
general predictions for the psd in on - off intermittency are an exponent zero for very low frequencies , ( -1/2 ) for medium frequencies and
( -2 ) for high frequencies , see eq .
( [ psdomega ] ) . near the critical voltage u@xmath128 ,
the experimental data indicate such a behaviour . the curve has been smoothed by averaging the density over intervals proportional to frequency . ] with different lower bounds ( @xmath181 ( @xmath163 ) , @xmath182 ( @xmath183 ) , @xmath184 ( @xmath162 ) ) , upper limit @xmath185 . the theoretical exponent -1/2 is found only for unrealistically low background . ] | noise driven electroconvection in sandwich cells of nematic liquid crystals exhibits on - off intermittent behaviour at the onset of the instability .
we study laser scattering of convection rolls to characterize the wavelengths and the trajectories of the stochastic amplitudes of the intermittent structures .
the pattern wavelengths and the statistics of these trajectories are in quantitative agreement with simulations of the linearized electrohydrodynamic equations .
the fundamental @xmath0 distribution law for the durations @xmath1 of laminar phases as well as the power law of the amplitude distribution of intermittent bursts are confirmed in the experiments .
power spectral densities of the experimental and numerically simulated trajectories are discussed . |
the idea of elaborating the foundation of space - time ( or foundation of relativity ) in a spirit analogous with the rather successful foundation of mathematics ( fom ) was initiated by several authors including , e.g. , david hilbert @xcite or leading contemporary logician harvey friedman @xcite .
foundation of mathematics has been carried through strictly within the framework of first - order logic ( fol ) , for certain reasons .
the same reasons motivate the effort of keeping the foundation of space - time also inside fol .
one of the reasons is that staying inside fol helps us to avoid tacit assumptions , another reason is that fol has a complete inference system while higher - order logic can not have one by gdel s incompleteness theorem , see e.g. , vnnen @xcite or ( * ? ? ?
* appendix ) . for more motivation for staying inside fol ( as opposed to higher - order logic ) , cf .
e.g. , ax @xcite , pambuccian @xcite , ( * ? ? ?
* appendix 1 : why exactly fol ) , @xcite , but the reasons in vnnen @xcite , ferreirs @xcite , or woleski @xcite also apply . following the above motivation , we begin at the beginning , namely first we recall a streamlined fol axiomatization of special relativity theory , from the literature .
is complete with respect to ( w.r.t .
) questions about inertial motion
. then we ask ourselves whether we can prove the usual relativistic properties of accelerated motion ( e.g. , clocks in acceleration ) in . as it turns out ,
this is practically equivalent to asking whether is strong enough to `` handle '' ( or treat ) accelerated observers .
we show that there is a mathematical principle called induction ( ) coming from real analysis which needs to be added to in order to handle situations involving relativistic acceleration .
we present an extended version of which is strong enough to handle accelerated clocks , in particular , accelerated observers .
we show that the so - called twin paradox becomes provable in .
it also becomes possible to introduce einstein s equivalence principle for treating gravity as acceleration and proving the gravitational time dilation , i.e. that gravity `` causes time to run slow '' .
what we are doing here is not unrelated to field s `` science without numbers '' programme and to `` reverse mathematics '' in the sense of harvey friedman and steven simpson .
namely , we systematically ask ourselves which mathematical principles or assumptions ( like , e.g. , ) are really needed for proving certain observational predictions of relativity .
( it was this striving for parsimony in axioms or assumptions which we alluded to when we mentioned , way above , that was `` streamlined '' . )
the interplay between logic and relativity theory goes back to around 1920 and has been playing a non - negligible role in works of researchers like reichenbach , carnap , suppes , ax , szekeres , malament , walker , and of many other contemporaries .
in section [ ax - s ] we recall the fol axiomatization complete w.r.t .
questions concerning inertial motion .
there we also introduce an extension of ( still inside fol ) capable for handling accelerated clocks and also accelerated observers . in section [ main - s ]
we formalize the twin paradox in the language of fol .
we formulate theorems [ thmtwp ] , [ thmeq ] stating that the twin paradox is provable from and the same for related questions for accelerated clocks .
theorems [ thmnoind ] , [ thmmo ] state that is not sufficient for this , more concretely that the induction axiom in is needed . in sections [ an - s ] , [ proofss ] we prove these theorems
. motivation for the research direction reported here is nicely summarized in ax @xcite , suppes @xcite ; cf .
also the introduction of @xcite .
harvey friedman s
@xcite present a rather convincing general perspective ( and motivation ) for the kind of work reported here .
in this paper we deal with the kinematics of relativity only , i.e. we deal with motion of _ bodies _ ( or _ test - particles _ ) .
the motivation for our choice of vocabulary ( for special relativity ) is summarized as follows
. we will represent motion as changing spatial location in time .
to do so , we will have reference - frames for coordinatizing events and , for simplicity , we will associate reference - frames with special bodies which we will call _ observers_. we visualize an observer - as - a - body as `` sitting '' in the origin of the space part of its reference - frame , or equivalently , `` living '' on the time - axis of the reference - frame .
we will distinguish _ inertial _ observers from non - inertial ( i.e. accelerated ) ones .
there will be another special kind of bodies which we will call _
photons_. for coordinatizing events we will use an arbitrary _ ordered field _ in place of the field of the real numbers . thus the elements of this field will be the `` _ quantities _ '' which we will use for marking time and space . allowing arbitrary ordered fields in place of the reals
increases flexibility of our theory and minimizes the amount of our mathematical presuppositions .
e.g. , ax @xcite for further motivation in this direction .
similar remarks apply to our flexibility oriented decisions below , e.g. , keeping the number @xmath0 of space - time dimensions a variable .
using coordinate systems ( or reference - frames ) instead of a single observer independent space - time structure is only a matter of didactical convenience and visualization , furthermore it also helps us in weeding out unnecessary axioms from our theories .
motivated by the above , we now turn to fixing the fol language of our axiom systems .
the first occurrences of concepts used in this work are set by boldface letters to make it easier to find them . throughout this work ,
if - and - only - if is abbreviated to * iff*. let us fix a natural number @xmath1 for the dimension of the space - time that we are going to axiomatize .
our first - order language contains the following non - logical symbols : * unary relation symbols @xmath2 ( for * bodies * ) , @xmath3 ( for * observers * ) , @xmath4 ( for * inertial observers * ) , @xmath5 ( for * photons * ) and @xmath6 ( for * quantities * which are going to be elements of a field ) , * binary function symbols @xmath7 ,
@xmath8 and a binary relation symbol @xmath9 ( for the field operations and the ordering on @xmath6 ) , and * a @xmath10-ary relation symbol @xmath11 ( for * world - view relation * ) . the bodies will play the role of the `` main characters '' of our space - time models and they will be `` observed '' ( coordinatized using the quantities ) by the observers .
this observation will be coded by the world - view relation @xmath11 .
our bodies and observers are basically the same as the `` test particles '' and the `` reference - frames '' , respectively , in some of the literature .
we read @xmath12 as `` @xmath13 is a body '' , `` @xmath13 is an observer '' , `` @xmath13 is an inertial observer '' , `` @xmath13 is a photon '' , `` @xmath13 is a field - element '' .
we use the world - view relation @xmath11 to talk about coordinatization , by reading @xmath14 as `` observer @xmath13 observes ( or sees ) body @xmath15 at coordinate point @xmath16 '' .
this kind of observation has no connection with seeing via photons , it simply means coordinatization .
@xmath17 are the so - called atomic formulas of our first - order language , where @xmath18 can be arbitrary variables or terms built up from variables by using the field - operations `` @xmath7 '' and `` @xmath8 '' .
the * formulas * of our first - order language are built up from these atomic formulas by using the logical connectives _ not _
( @xmath19 ) , _ and _ ( @xmath20 ) , _ or _ ( @xmath21 ) , _ implies _ ( @xmath22 ) , _ if - and - only - if _ ( @xmath22 ) and the quantifiers _ exists _
@xmath13 ( @xmath23 ) and _ for all @xmath13 _ ( @xmath24 ) for every variable @xmath13 .
usually we use the variables @xmath25 to denote observers , @xmath26 to denote bodies , @xmath27 to denote photons and @xmath28 to denote quantities ( i.e. field - elements ) .
we write @xmath29 and @xmath30 in place of @xmath31 and @xmath32 , e.g. , we write @xmath33 in place of @xmath34 , and we write @xmath35 in place of @xmath36 etc .
the * models * of this language are of the form @xmath37 where @xmath38 is a nonempty set , @xmath39 are unary relations on @xmath38 , etc .
a unary relation on @xmath38 is just a subset of @xmath38 .
thus we use @xmath40 etc . as sets as well , e.g. , we write @xmath41 in place of @xmath42 . having fixed our language , we now turn to formulating an axiom system for special relativity in this language .
we will make special efforts to keep all our axioms inside the above specified first - order logic language of @xmath43 . throughout this work , @xmath44 , @xmath45 and
@xmath46 denote positive integers .
@xmath47 ( @xmath46-times ) is the set of all @xmath46-tuples of elements of @xmath6 .
if @xmath48 , then we assume that @xmath49 , i.e.@xmath50 denotes the @xmath44-th component of the @xmath46-tuple @xmath51 .
the following axiom is always assumed and is part of every axiom system we propose .
: : @xmath52 , @xmath53 , @xmath54 , @xmath55 , @xmath7 and @xmath8 are binary operations on @xmath6 , @xmath56 is a binary relation on @xmath6 and @xmath57 is an * euclidean ordered field * , i.e. a linearly ordered field in which positive elements have square roots . in pure first - order logic
, the above axiom would look like @xmath58 $ ] etc . in the present section we will not write out the purely first - order logic translations of our axioms since they will be straightforward to obtain .
the first - order logic translations of our next three axioms , , can be found in the appendix .
let @xmath43 be a model in which is true .
let @xmath59 denote the * ordered field reduct * of @xmath43 . here
we list the definitions and notation that we are going to use in formulating our axioms .
let @xmath60 be the usual field operations which are definable from `` @xmath7 '' and `` @xmath8 '' .
we use the vector - space structure of @xmath61 , i.e. if @xmath62 and @xmath63 , then @xmath64 ; and @xmath65 denotes the * origin*. the * euclidean - length * of @xmath48 is defined as @xmath66 .
the set of positive elements of @xmath6 is denoted by @xmath67 .
let @xmath68 .
we use the notation @xmath69 for the * space component * of @xmath29 and @xmath70 for the * time component * of @xmath29 .
we define the * line * through @xmath29 and @xmath30 as @xmath71 .
the set of * lines * is defined as @xmath72 .
the * slope * of @xmath29 is defined as @xmath73 if @xmath74 and is undefined otherwise ; furthermore @xmath75 if @xmath76 and is undefined otherwise .
@xmath77 is called the * coordinate system * and its elements are referred to as * coordinate points*. the * event * ( the set of bodies ) observed by observer @xmath78 at coordinate point @xmath29 is : @xmath79 the mapping @xmath80 is called the * world - view * ( function ) of @xmath78 .
the * coordinate domain * of observer @xmath78 is the set of coordinate points where @xmath78 observes something : @xmath81 the * life - line * ( or * trace * ) of body @xmath26 as seen by observer @xmath78 is defined as the set of coordinate points where @xmath26 was observed by @xmath78 : @xmath82 the life - line @xmath83 of observer @xmath78 as seen by himself is called the * self - line * of @xmath78 .
the * time - axis * is defined as @xmath84 .
[ bl][bl]@xmath83 [ bl][bl]@xmath85 [ t][t]@xmath86 [ tl][tl]@xmath87 [ br][br]@xmath88 [ bl][bl]@xmath89 [ t][t]@xmath90 [ tl][tl]@xmath91 [ r][r]@xmath29 [ l][l]@xmath30 [ lb][lb]@xmath92 [ t][t]@xmath93 [ t][t]@xmath94 [ cb][cb]world - view of @xmath95 [ cb][cb]world - view of @xmath78 .
, title="fig:",scaledwidth=80.0% ] now we are ready to build our space - time theories by formulating our axioms .
we formulate each axiom on two levels .
first we give an intuitive formulation , then we give a precise formalization using our notation .
the following natural axiom goes back to galileo galilei and even to the norman - french oresme of around 1350 , cf .
e.g. , ( * ? ? ?
* , 5 ) .
it simply states that each observer thinks that he rests in the origin of the space part of his coordinate system .
: : the self - line of any observer is the time - axis restricted to his coordinate domain : @xmath96 a fol - formula expressing can be found in the appendix .
the next axiom is about the constancy of the speed of the photons , cf .
e.g. , @xcite . for convenience ,
we choose @xmath97 for their speed .
: : for every inertial observer , the lines of slope 1 are exactly the traces of the photons : @xmath98 a fol - formula expressing can be found in the appendix .
we will also assume the following axiom : : : all inertial observers observe the same events : @xmath99 a fol - formula expressing can be found in the appendix .
@xmath100 since , in some sense , is only an `` auxiliary '' ( or book - keeping ) axiom about the `` mathematical frame '' of our reasoning , the heart of consists of three very natural axioms , , , . these are really intuitively convincing , natural and simple assumptions . from these three axioms
one can already prove the most characteristic predictions of special relativity theory .
what the average layperson usually knows about relativity is that `` moving clocks slow down '' , `` moving spaceships shrink '' , and `` moving pairs of clocks get out of synchronism '' .
we call these the * paradigmatic effects * of special relativity .
all these can be proven from the above three axioms , in some form , cf .
theorem [ spec - thm ] .
e.g. , one can prove that `` if @xmath101 are any two observers not at rest relative to each other , then one of @xmath101 will `` see '' or `` think '' that the clock of the other runs slow '' .
however , does not imply yet the inertial approximation of the so - called twin paradox . below theorem [ thmmo ] . ] in order to prove the inertial approximation of the twin paradox also , and to prove all the paradigmatic effects in their strongest form , it is enough to add one more axiom to .
this is what we are going to do now .
we will find that studying the relationships between the world - views is more illuminating than studying the world - views in themselves .
therefore the following definition is fundamental .
the * world - view transformation * between the world - views of observers @xmath95 and @xmath78 is the set of pairs of coordinate points @xmath102 such that @xmath78 and @xmath95 observe the same nonempty event in @xmath29 and @xmath30 , respectively : @xmath103 cf .
figure 1 .
we note that although the world - view transformations are only binary relations , axiom turns them into functions , cf .
( iii ) of proposition [ prop0 ] way below .
[ fmkconv ] whenever we write `` @xmath104 '' , we mean that there is a unique @xmath105 such that @xmath106 , and @xmath107 denotes this unique @xmath30 .
i.e. if we talk about the value @xmath107 of @xmath108 at @xmath29 , we tacitly postulate that it exists and is unique ( by the present convention ) .
the following axiom is an instance ( or special case ) of the principle of special relativity , according to which the `` laws of nature '' are the same for all inertial observers , in particular , there is no experiment which would decide whether you are in absolute motion , cf .
e.g. , einstein @xcite or @xcite or @xcite .
to explain the following formula , let @xmath109 .
then @xmath110 is the time passed between the events @xmath111 and @xmath112 as seen by @xmath78 and @xmath113 is the time passed between the same two events as seen by @xmath95 .
hence @xmath114 is the rate with which @xmath95 s clock runs slow as seen by @xmath78 .
the same explanation applies when @xmath78 and @xmath95 are interchanged .
: : any two inertial observers see each other s clocks go wrong in the same way : @xmath115 all the axioms so far talked about inertial observers , and they in fact form an axiom system complete w.r.t . the inertial observers , cf .
theorem [ spec - thm ] below .
@xmath116 let @xmath68 .
then @xmath117 is the ( signed ) * minkowski - length * of @xmath29 and the * minkowski - distance * between @xmath29 and @xmath30 is defined as follows : @xmath118 a motivation for the `` otherwise '' part of the definition of @xmath119 is the following . @xmath119 codes two kinds of information , ( i ) the length of @xmath29 and ( ii ) whether @xmath29 is time - like ( i.e. @xmath120 ) or space - like . since the length is always non - negative , we can use the sign of @xmath119 to code ( ii ) .
let @xmath121 be a function .
@xmath122 is said to be a * poincar - transformation * if @xmath122 is a bijection and it preserves the minkowski - distance , i.e. @xmath123 for all @xmath68 .
@xmath122 is called a * dilation * if there is a positive @xmath124 such that @xmath125 for all @xmath126 and @xmath122 is called a * field - automorphism - induced * mapping if there is an automorphism @xmath127 of the field @xmath128 such that @xmath129 for all @xmath126 .
the following is proved in @xcite and in @xcite .
let @xmath130 be a set of formulas and @xmath43 be a model .
@xmath131 denotes that all formulas in @xmath130 are true in model @xmath43 . in this case
we say that @xmath43 is a * model of * @xmath130 .
[ spec - thm][thmpoi ] let @xmath132 , let @xmath43 be a model of our language and let @xmath101 be inertial observers in @xmath43 .
then @xmath133 is a poincar - transformation whenever @xmath134 .
more generally , @xmath133 is a poincar - transformation composed with a dilation and a field - automorphism - induced mapping whenever @xmath135 .
[ rem - specthm ] assume @xmath132 .
theorem [ spec - thm ] is best possible in the sense that , e.g. , for every poincar - transformation @xmath122 over an arbitrary euclidean ordered field there are a model @xmath134 and inertial observers @xmath101 in @xmath43 such that the world - view transformation @xmath133 between @xmath78 s and @xmath95 s world - views in @xmath43 is @xmath122 , see @xcite , 2.9.5(iii ) ) .
similarly for the second statement in theorem [ spec - thm ] .
hence , theorem [ spec - thm ] can be refined to a pair of completeness theorems , cf .
* , p.271 ) .
roughly , for every euclidean ordered field , its poincar - transformations ( can be expanded to ) form a model of .
similarly for the other case .
@xmath136 it follows from theorem [ spec - thm ] that the paradigmatic effects of relativity hold in in their strongest form , e.g. , if @xmath78 and @xmath95 are observers not at rest w.r.t
. each other , then both will `` think '' that the clock of the other runs slow
. also implies the `` inertial approximation '' of the twin paradox , see e.g. , @xcite , and @xcite .
it is necessary to add to in order to be able to prove the inertial approximation of the twin paradox by theorem [ spec - thm ] , cf .
e.g. , @xcite .
we now begin to formulate axioms about non - inertial observers .
the non - inertial observers are called * accelerated * observers . to connect
the world - views of the accelerated and the inertial observers , we are going to formulate the statement that at each moment of his life , each accelerated observer sees the nearby world for a short while as an inertial observer does . to formalize this ,
first we introduce the relation of being a co - moving observer . the ( open ) * ball * with center @xmath137 and radius @xmath138
is : @xmath139 @xmath78 is a * co - moving observer * of @xmath95 at @xmath105 , in symbols @xmath140 , if @xmath141 and the following holds : @xmath142 behind the definition of the co - moving observers is the following intuitive if @xmath140 .
the following axiom gives the promised connection between the world - views of the inertial and the accelerated observers : : : at any point on the self - line of any observer , there is a co - moving inertial observer : @xmath143 let be the collection of the axioms introduced so far : @xmath144 let @xmath145 denote the ordered field of the real numbers .
accelerated clocks behave as expected in models of when the ordered field reduct of the model is @xmath145 ( cf .
theorems [ thmtwp ] , [ thmeq ] , and in more detail prop.[proptr ] , rem.[trrem ] ) ; but not otherwise ( cf
. theorems [ thmnoind ] , [ thmmo ] ) .
thus to prove properties of accelerated clocks ( and observers ) , we need more properties of the field reducts than their being euclidean ordered fields . as it turns out , adding all fol - formulas valid in the field of the reals does not suffice ( cf.corollary [ cornoind ] ) .
the additional property of @xmath145 we need is that in @xmath145 every bounded non - empty set has a * supremum * , i.e. a least upper bound .
this is a second - order logic property ( because it concerns all subsets of @xmath145 ) which we can not use in a fol axiom system .
instead , we will use a kind of `` induction '' axiom schema .
it will state that every non - empty , bounded subset of the ordered field reduct which can be defined by a fol - formula using possibly the extra part of the model , e.g. , using the world - view relation , has a supremum .
we now begin to formulate our fol induction axiom schema .
to abbreviate formulas of fol we often omit parentheses according to the following convention .
quantifiers bind as long as they can , and @xmath20 binds stronger than @xmath22 .
e.g. , @xmath146 means @xmath147 . instead of curly brackets
we sometimes write square brackets in formulas , e.g. , we may write @xmath148\land\eta)$ ] . if @xmath149 is a formula and
@xmath13 is a variable , then we say that @xmath13 is a * free variable * [ free variable ] of @xmath149 iff @xmath13 does not occur under the scope of either @xmath23 or @xmath24 .
let @xmath149 be a formula ; and let @xmath150 be all the free variables of @xmath149 .
let @xmath151 be a model .
whether @xmath149 is true or false in @xmath43 depends on how we associate elements of @xmath38 to these free variables .
when we associate @xmath152 to @xmath150 , respectively , then @xmath153 denotes this truth - value , thus @xmath153 is either true or false in @xmath43 .
for example , if @xmath149 is @xmath154 , then @xmath155 is true in @xmath145 while @xmath156 is false in @xmath145 .
@xmath149 is said to be * true * in @xmath43 if @xmath149 is true in @xmath43 no matter how we associate elements to the free variables .
we say that a subset @xmath157 of @xmath6 is * definable by * @xmath149 iff there are @xmath158 such that @xmath159 .
: : every subset of @xmath6 definable by @xmath149 has a supremum if it is non - empty and * bounded * : @xmath160 \;\land\ ; [ \exists b\in { { \mathrm}{f}}\quad ( \forall x\in { { \mathrm}{f}}\quad \varphi \longrightarrow x \le b)]\;\\ \longrightarrow\ ; \big(\exists s \in { { \mathrm}{f}}\enskip \forall b \in { { \mathrm}{f}}\quad ( \forall x\in { { \mathrm}{f}}\quad \varphi \longrightarrow x \le b)\iff s\le b\big ) . \end{split}\ ] ] we say that a subset of @xmath6 is * definable * iff it is definable by a fol - formula .
our axiom scheme below says that every non - empty bounded and definable subset of @xmath6 has a supremum .
@xmath161 notice that is true in any model whose ordered field reduct is @xmath145 .
let us add to and call it : @xmath162 is a countable set of fol - formulas .
we note that there are non - trivial models of , cf .
e.g. , remark [ trrem ] way below .
furthermore , we note that the construction in misner - thorne - wheeler ( * ? ? ?
* chapter 6 entitled the local coordinate system of an accelerated observer , especially pp .
172 - 173 and chapter 13.6 entitled the proper reference frame of an accelerated observer on pp . 327 - 332 ) can be used for constructing models of .
models of are discussed to some detail in @xcite .
theorems [ thmtwp ] and [ thmeq ] ( and also prop.[proptr ] , rem.[trrem ] ) below show that already implies properties of accelerated clocks , e.g. , it implies the twin paradox
. implies all the fol - formulas true in @xmath145 , but is stronger .
let denote the set of elements of that talk only about the ordered field reduct , i.e. let @xmath163 now , together with the axioms of linearly ordered fields is a complete axiomatization of the fol - theory of @xmath145 , i.e. all fol - formulas valid in @xmath145 can be derived from them . herein or ( * ? ? ?
* proposition a.0.1 ) .
] however , is stronger than , since @xmath164 by corollary [ cornoind ] below , while @xmath165 by theorem [ thmtwp ] .
the strength of comes from the fact that the formulas in can `` talk '' about more `` things '' than just those in the language of @xmath145 ( namely they can talk about the world - view relation , too ) . for understanding how works
, it is important to notice that does not speak about the field @xmath166 itself , but instead , it speaks about connections between @xmath166 and the rest of the model @xmath167 . why do we call a kind of induction schema
the reason is the following . implies that if a formula @xmath149 becomes false sometime after @xmath168 while being true at @xmath168 , then there is a `` first '' time - point where , so to speak , @xmath149 becomes false .
this time - point is the supremum of the time - points until which @xmath149 remained true after @xmath168 .
now , @xmath149 may or may not be false at this supremum , but it is false arbitrarily `` close '' to it afterwards . if such a `` point of change '' for the truth of @xmath149 can not exist , implies that @xmath149 has to be true always after @xmath168 if it is true at @xmath168 .
( without , this may not be true . )
twin paradox ( tp ) concerns two twin siblings whom we shall call ann and ian .
( `` a '' and `` i '' stand for accelerated and for inertial , respectively ) .
ann travels in a spaceship to some distant star while ian remains at home .
tp states that when ann returns home she will be _ younger _ than her _
twin brother _ ian . we now formulate tp in our fol language .
the * segment * between @xmath169 and @xmath105 is defined as : @xmath170:=\{{\lambda}p+(1-{\lambda})q:\lambda\in { { \mathrm}{f}}\;\land\ ; 0\le{\lambda}\le1\}.\ ] ] we say that observer @xmath95 is in * twin - paradox relation * with observer
@xmath78 iff whenever @xmath95 leaves @xmath78 between two meetings , @xmath95 measures less time between the two meetings than @xmath78 : @xmath171 \subseteq & tr_k(k)\;\land\ ; [ p ' q']\not\subseteq tr_m(k)\\ & \longrightarrow\ ; \big|q_t - p_t\big|<\big|q'_t - p'_t\big| , \end{split}\ ] ] cf . figure 2 . in this case
we write @xmath172 .
we note that , if two observers do not leave each other or they meet less than twice , then they are in twin - paradox relation by this definition . thus two inertial observers are always in this relation .
[ l][l ] [ l][l ] [ b][b]@xmath108 [ rt][rt]@xmath29 [ rb][rb]@xmath30 [ l][l]@xmath85 [ lt][lt]@xmath173 [ lb][lb]@xmath174 [ c][c]@xmath175 [ c][c]@xmath176 [ b][b]@xmath83 [ b][b]@xmath88 [ r][r]@xmath177 [ r][r]@xmath178 [ c][c]same events : : every observer is in twin - paradox relation with every inertial observer : @xmath179 let @xmath149 be a formula and @xmath130 be a set of formulas .
@xmath180 denotes that @xmath149 is true in all models of @xmath130 .
gdel s completeness theorem for fol implies that whenever @xmath181 , there is a ( syntactic ) derivation of @xmath149 from @xmath130 via the commonly used derivation rules of fol .
hence the next theorem states that the formula formulating the twin paradox is provable from the axiom system .
[ thmtwp ] @xmath182 if @xmath132 .
the proof of the theorem is in section [ proofss ] .
now we turn to formulating a phenomenon which we call duration determining property of events . : : if each of two observers observes the very same ( non - empty ) events in a segment of their self - lines , they measure the same time between the end points of these two segments : @xmath183\}=\{ev_m(r ' & ) : r'\in [ p ' q']\}\ ; \longrightarrow\;\\ \big|&q_t - p_t\big|=\big|q'_t - p'_t\big| , \end{split}\ ] ] see the right hand side of figure 2 . the next theorem states that also can be proved from our fol axiom system .
[ thmeq ] @xmath184 if @xmath132 .
the proof of the theorem is in section [ proofss ] .
the assumption @xmath132 can not be omitted from theorem [ thmtwp ] .
however , theorems [ thmtwp ] and [ thmeq ] remain true if we omit the assumption @xmath132 and assume auxiliary axioms and below , i.e.@xmath185 holds for @xmath186 , too .
a proof for the latter statement can be obtained from the proofs of theorems [ thmtwp ] and [ thmeq ] by ( * ? ? ?
* items 4.3.1 , 4.2.4 , 4.2.5 ) and ( * ? ? ?
* theorem 1.4(ii ) ) .
: : in every inertial observer s coordinate system , every line of slope less than 1 is the life - line of an inertial observer : @xmath187 : : traces of inertial observers are lines as observed by inertial observers : @xmath188 @xmath136 can the assumption @xmath132 be omitted from theorem [ thmeq ] , i.e.does @xmath189 hold for @xmath186 ? the following theorem says that theorems [ thmtwp ] and [ thmeq ] do not remain true if we omit the axiom scheme from .
if a formula @xmath149 is not true in a model @xmath43 , we write @xmath190 .
[ thmnoind ] for every euclidean ordered field @xmath166 not isomorphic to @xmath145 , there is a model @xmath43 of @xmath191 such that @xmath192 , @xmath193 and the ordered field reduct of @xmath43 is @xmath166 .
the proof of the theorem is in section [ proofss ] . by theorems
[ thmtwp ] and [ thmeq ] , is not true in the model @xmath43 mentioned in theorem [ thmnoind ] .
this theorem has strong consequences , it implies that to prove the twin paradox , it does not suffice to add all the fol - formulas valid in @xmath145 ( to ) . let @xmath194 denote the set of all fol - formulas valid in @xmath145 .
[ cornoind ] @xmath195 and @xmath196 .
the proof of the corollary is in section [ proofss ] .
an ordered field is called * non - archimedean * if it has an element @xmath51 such that , for every positive integer @xmath46 , @xmath197 .
we call these elements * infinitesimally small*. the following theorem says that , for countable or non - archimedean euclidean ordered fields , there are quite sophisticated models of in which and are false . [ thmmo ] for every euclidean
ordered field @xmath166 which is countable or non - archimedean , there is a model @xmath43 of @xmath191 such that @xmath192 , @xmath193 , the ordered field reduct of @xmath43 is @xmath166 and ( i)(iv ) below also hold in @xmath43 .
* every observer uses the whole coordinate system for coordinate - domain : @xmath198 * at any point in @xmath77 , there is a co - moving inertial observer of any observer : @xmath199 * all observers observe the same set of events : @xmath200 * every observer observes every event only once : @xmath201 @xmath136 the proof of the theorem is in section [ proofss ] .
[ l][l]@xmath83 [ rt][rt]@xmath202 [ rb][rb]@xmath203 [ rt][rt]@xmath204 [ rb][rb]@xmath205 [ l][l]@xmath173 [ l][l]@xmath174 [ l][l]@xmath206 [ lb][lb]@xmath207 [ lb][lb]@xmath208 [ t][t]@xmath207 [ b][b]@xmath208 [ r][r]@xmath29 [ r][r]@xmath30 [ r][r]@xmath209 finally we formulate a question .
to this end we introduce the inertial version of the twin paradox and some auxiliary axioms . in the inertial version of the twin paradox
, we use the common trick of the literature to talk about the twin paradox without talking about accelerated observers .
we replace the accelerated twin with two inertial ones , a leaving and an approaching one .
we say that observers @xmath210 and @xmath211 are in * inertial twin - paradox relation * with observer @xmath78 if the following holds : @xmath212 \\ \longrightarrow & \ ; |q'_t - p'_t| > |q_t - r_t|+|r_t - p_t\big| , \end{split}\ ] ] cf . figure 3 . in this case
we write @xmath213 . : : every three inertial observers are in inertial twin - paradox relation : @xmath214 : : to every inertial observer @xmath78 and coordinate point @xmath29 there is an inertial observer @xmath95 such that the world - view transformation between @xmath78 and @xmath95 is the translation by vector @xmath29 : @xmath215 : : the world - view transformation between inertial observers @xmath78 and @xmath95 is a linear transformation if @xmath216 : @xmath217 [ qtwp ] does theorem [ thmtwp ] remain true if we replace in with the inertial version of the twin paradox and we assume the auxiliary axioms , and ?
question [ qconv ] .
we note that and are true in the models of in case @xmath132 , cf .
* theorem 1.2 ) , ( * ? ? ?
* theorem 2.8.28 ) and @xcite .
in this section we gather the statements ( and proofs from ) of the facts we will need from analysis .
the point is in formulating these statements in fol and for an arbitrary ordered field in place of using the second - order language of the ordered field @xmath145 of reals . in the present section
is assumed without any further mentioning .
let @xmath218 .
we say that @xmath26 is * between * @xmath51 and @xmath219 iff @xmath220 or @xmath221 .
we use the following notation : @xmath222:=\{x\in { { \mathrm}{f } } : a\le x \le b\}$ ] and @xmath223 . whenever we write @xmath222 $ ] , we assume that @xmath224 and @xmath225 .
we also use this convention for @xmath226 .
let @xmath227 .
then @xmath228 is said to be an * accumulation point * of @xmath157 if for all @xmath229 , @xmath230 has an element different from @xmath29 .
@xmath157 is called * open * if for all @xmath231 , there is an @xmath229 such that @xmath232 .
let @xmath233 and @xmath234 be binary relations .
the * composition * of @xmath235 and @xmath236 is defined as : @xmath237 . the * domain * and
the * range * of @xmath235 are denoted by @xmath238 and @xmath239 , respectively .
@xmath240 denotes the * inverse * of @xmath235 , i.e. @xmath241 .
we think of a * function * as a special binary relation .
notice that if @xmath242 are functions , then @xmath243 for all @xmath244 .
@xmath245 denotes that @xmath122 is a function from @xmath246 to @xmath247 , i.e. @xmath248 and @xmath249 .
notation @xmath250 denotes that @xmath122 is a * partial function * from @xmath246 to @xmath247 ; this means that @xmath122 is a function , @xmath251 and @xmath249 .
let @xmath252 .
we call @xmath122 * continuous * at @xmath13 if @xmath253 , @xmath13 is an accumulation point of @xmath254 and the usual formula of continuity holds for @xmath122 and @xmath13 , i.e. @xmath255 we call @xmath122 * differentiable * at @xmath13 if @xmath256 , @xmath13 is an accumulation point of @xmath254 and there is an @xmath257 such that @xmath258 this @xmath51 is unique .
we call this @xmath51 the * derivate * of @xmath122 at @xmath13 and we denote it by @xmath259 .
@xmath122 is said to be continuous ( differentiable ) on @xmath260 iff @xmath261 and @xmath122 is continuous ( differentiable ) at every @xmath262 .
we note that the basic properties of the differentiability remain true since their proofs use only the ordered field properties of @xmath145 , cf.propositions [ propdiff ] , [ propaff ] and [ propmax ] below .
let @xmath263 and @xmath264
. then @xmath265 and @xmath266 are defined as @xmath267 and @xmath268 .
let @xmath269 .
@xmath270 is said to be * increasing * on @xmath157 iff @xmath271 and for all @xmath272 , @xmath273 if @xmath274 , and @xmath270 is said to be * decreasing * on @xmath157 iff @xmath271 and for all @xmath272 , @xmath275 if @xmath274 .
[ propdiff ] let @xmath263 and @xmath269 . then ( i)(v ) below hold .
* if @xmath122 is differentiable at @xmath13 then it is also continuous at @xmath13 . *
let @xmath63 .
if @xmath122 is differentiable at @xmath13 , then @xmath276 is also differentiable at @xmath13 and @xmath277 . *
if @xmath122 and @xmath278 are differentiable at @xmath13 and @xmath13 is an accumulation point of @xmath279 , then @xmath280 is differentiable at @xmath13 and @xmath281 .
* if @xmath270 is differentiable at @xmath13 , @xmath278 is differentiable at @xmath282 and @xmath13 is an accumulation point of @xmath283 , then @xmath284 is differentiable at @xmath13 and @xmath285 . *
if @xmath270 is increasing ( or decreasing ) on @xmath226 , differentiable at @xmath286 and @xmath287 , then @xmath288 is differentiable at @xmath282 . since the proofs of the statements
are based on the same calculations and ideas as in real analysis , we omit the proof , cf .
* theorems 28.2 , 28.3 , 28.4 and 29.9 ) .
let @xmath289 .
@xmath290 denotes the @xmath44-th projection function , i.e. @xmath291 .
let @xmath252 .
we denote the @xmath44-th coordinate function of @xmath122 by @xmath292 , i.e. @xmath293 .
we also denote @xmath294 by @xmath295 .
a function @xmath296 is said to be an * affine map * if it is a linear map composed by a translation .
is an affine map if there are @xmath297 and @xmath298 such that @xmath299 , @xmath300 and @xmath301 for all @xmath62 and @xmath264 . ]
the following proposition says that the derivate of a function @xmath122 composed by an affine map @xmath246 at a point @xmath13 is the image of the derivate @xmath259 taken by the linear part of @xmath246 .
[ propaff ] let @xmath252 be differentiable at @xmath13 and let @xmath302 be an affine map .
then @xmath303 is differentiable at @xmath13 and @xmath304 .
in particular , @xmath305 , i.e. @xmath306 .
the statement is straightforward from the definitions .
@xmath307 is said to be * locally maximal * at @xmath13 iff @xmath256 and there is a @xmath308 such that @xmath309 for all @xmath310 .
the * local minimality * is defined analogously .
[ propmax ] if @xmath307 is differentiable on @xmath226 and locally maximal or minimal at @xmath286 , then its derivate is @xmath168 at @xmath13 , i.e. @xmath311 .
the proof is the same as in real analysis , cf.e.g .
, ( * ? ? ?
* theorem 5.8 ) .
let @xmath312 be a model .
an @xmath46-ary relation @xmath313 is said to be * definable * iff there is a formula @xmath149 with only free variables @xmath314 and there are @xmath315 such that @xmath316 recall that says that every non - empty , bounded and definable subset of @xmath6 has a supremum .
[ thmboltzano ] assume .
let @xmath317 be definable and continuous on @xmath222 $ ] .
if @xmath219 is between @xmath318 and @xmath319 , then there is an @xmath320 $ ] such that @xmath321 .
let @xmath219 be between @xmath318 and @xmath319 .
we can assume that @xmath322 .
let @xmath323 : f(x ) < c\}$ ] .
then @xmath157 is definable , bounded and non - empty .
thus , by , it has a supremum , say @xmath324 .
both @xmath325 and @xmath326 are non - empty open sets since @xmath122 is continuous on @xmath222 $ ]
. thus @xmath327 can not be less than @xmath219 since @xmath324 is an upper bound of @xmath157 and can not be greater than @xmath219 since @xmath324 is the smallest upper bound .
hence @xmath321 as desired .
[ thmsup ] assume .
let @xmath317 be definable and continuous on @xmath222 $ ] .
then the supremum @xmath324 of @xmath328\}$ ] exists and there is an @xmath329 $ ] such that @xmath330 . the set @xmath331 : \exists c\in { { \mathrm}{f}}\enskip\forall x\in [ a , y]\quad f(x)<c\}$ ] has a supremum by since @xmath157 is definable , non - empty and bounded .
this supremum has to be @xmath26 and @xmath332 since @xmath122 is continuous on @xmath222 $ ] .
thus @xmath333\}$ ] is bounded .
thus , by , it has a supremum , say @xmath324 , since it is definable and non - empty .
we can assume that @xmath334 .
let @xmath335 : \exists c \in { { \mathrm}{f}}\enskip \forall x\in [ a , y]\quad f(x)<c < s\}$ ] . by , @xmath246 has a supremum . at this supremum , @xmath122 can not be less than @xmath324 since @xmath122 is continuous on @xmath222 $ ] and @xmath324 is the supremum of @xmath336 . throughout this work
@xmath337 denotes the identity function , i.e. @xmath338 .
[ thmlagrange ] assume .
let @xmath307 be definable , differentiable on @xmath222 $ ] .
if @xmath339 , then there is an @xmath340 such that @xmath341 .
assume @xmath339 .
let @xmath342 .
then @xmath270 is differentiable on @xmath222 $ ] and @xmath343 for all @xmath344 $ ] by ( ii ) and ( iii ) of proposition [ propdiff ] since @xmath345 is differentiable on @xmath222 $ ] and its derivate is @xmath97 for all @xmath344 $ ] . if @xmath270 is constant on @xmath222 $ ] , such that @xmath346 for all @xmath347 $ ] , ] then @xmath348 for all @xmath349
. otherwise , by theorem [ thmsup ] , there is a maximum or minimum of @xmath270 different from @xmath350 at an @xmath340 .
hence @xmath348 by proposition [ propmax ] .
this completes the proof since @xmath339 and @xmath351 .
[ thmroll ] assume .
let @xmath307 be definable and differentiable on @xmath222 $ ] .
if @xmath352 and @xmath339 , then there is an @xmath340 such that @xmath353 .
[ propint ] assume .
let @xmath354 be definable and differentiable on @xmath226 .
if @xmath355 for all @xmath356 , then there is a @xmath357 such that @xmath358 for all @xmath356 .
assume that @xmath355 for all @xmath356 .
let @xmath359 .
then @xmath360 for all @xmath356 by ( ii ) and ( iii ) of proposition [ propdiff ] .
if there are @xmath361 such that @xmath362 and @xmath363 , then , by the mean value theorem , there is an @xmath13 between @xmath15 and @xmath364 such that @xmath365 and this contradicts @xmath366 .
thus @xmath367 for all @xmath361 .
hence there is a @xmath357 such that @xmath346 for all @xmath356 .
in the present section is assumed without any further mentioning .
let @xmath368 be the natural embedding defined as @xmath369 .
we define the * life - curve * of observer @xmath95 as seen by observer @xmath78 as @xmath370 . throughout this work
we denote @xmath371 by @xmath372 , for @xmath373 .
thus @xmath374 is the coordinate point where @xmath78 observes the event `` @xmath95 s wristwatch shows @xmath375 '' , i.e. @xmath376 iff @xmath377 . in the following proposition ,
we list several easy but useful consequences of some of our axioms .
[ prop0 ] let @xmath378 and @xmath379 . then ( i)(viii ) below hold .
then @xmath380 and for all distinct @xmath68 , @xmath381 . *
. then @xmath382 . *
. then @xmath383 and @xmath384 . *
assume and .
if @xmath385 , then @xmath386 is a bijection and @xmath387 is an injection . *
. then @xmath389 and @xmath390 . *
. then @xmath391 . *
. then @xmath392 and @xmath393 .
* assume , and .
then @xmath394 . to prove ( i ) , let @xmath68 be distinct points . then there is a line of slope @xmath97 that contains @xmath29 but does not contain @xmath30 . by , this line is the trace of a photon .
for such a photon @xmath27 , we have @xmath395 and @xmath396 .
hence @xmath381 and @xmath397 . thus ( i ) holds .
\(ii ) follows from ( i ) since @xmath398 by .
\(iii ) and ( iv ) follow from ( i ) by the definitions of the world - view transformation and the life - curve . to prove ( v ) , let @xmath399 . then @xmath400 .
since , by , @xmath270 and @xmath78 observe the same set of events , there is an @xmath401 such that @xmath402 .
but then @xmath403 and @xmath404 . hence @xmath405 .
thus @xmath406 .
the other inclusion follows from the definition of the world - view transformation .
thus @xmath389 and @xmath407 . to prove ( vi ) , let @xmath408 . by
, there is an @xmath388 such that @xmath270 is a co - moving observer of @xmath95 at @xmath30 .
for such an @xmath270 , we have @xmath409 and , by ( v ) , @xmath410 .
thus @xmath411 . to prove ( vii ) , let @xmath412 . then @xmath413 .
but then @xmath414 . thus @xmath415 . by ,
@xmath416 ; and this proves the first part of ( vii ) .
by @xmath416 , we have @xmath417 .
thus @xmath418 and this proves the second part of ( vii ) .
the `` @xmath419 part '' of ( viii ) follows from ( vii ) . to prove the other inclusion ,
let @xmath420 . then , by and ( vi ) , @xmath421 .
thus there are @xmath422 and @xmath105 such that @xmath423 and @xmath424 .
but then @xmath425 .
hence @xmath426 .
we say that @xmath122 is * well - parametrized * iff @xmath427 and the following holds : if @xmath256 is an accumulation point of @xmath254 , then @xmath122 is differentiable at @xmath13 and its derivate at @xmath13 is of minkowski - length @xmath97 , i.e. @xmath428 .
assume @xmath429 .
then the curve @xmath122 is well - parametrized iff @xmath122 is parametrized according to minkowski - length , i.e. for all @xmath430 , if @xmath431\subseteq dom(f)$ ] , the minkowski - length of @xmath122 restricted to @xmath431 $ ] is @xmath432 .
( by minkowski - length of a curve we mean length according to minkowski - metric , e.g. , in the sense of wald ( * ? ? ? * , ( 3.3.7 ) ) ) . *
proper time * or * wristwatch time * is defined as the minkowski - length of a time - like curve , cf .
e.g. , wald ( * ? ? ?
* , ( 3.3.8 ) ) , taylor - wheeler @xcite or dinverno ( * ? ? ?
* , ( 8.14 ) ) .
thus a curve defined on a subset of @xmath145 is well - parametrized iff it is parametrized according to proper time , or wristwatch - time .
e.g. , ( * ? ? ?
* , ( 8.16 ) ) . )
the next proposition states that life - curves of accelerated observers in models of are well - parametrized .
this implies that accelerated clocks behave as expected in models of .
remark [ trrem ] after the proposition will state a kind of `` completeness theorem '' for life - curves of accelerated observers , much in the spirit of remark [ rem - specthm ] .
[ proptr ] assume and @xmath132 .
let @xmath378 and @xmath379 .
then @xmath433 is well - parametrized and definable .
let @xmath378 , @xmath379 .
then @xmath433 is definable by its definition .
furthermore , @xmath383 and @xmath434 by ( iii ) of proposition [ prop0 ] .
let @xmath426 be an accumulation point of @xmath435 .
we would like to prove that @xmath433 is differentiable at @xmath13 and its derivate at @xmath13 is of minkowski - length @xmath97 .
@xmath416 by ( vii ) of proposition [ prop0 ] .
thus , by , there is a co - moving inertial observer of @xmath95 at @xmath372 . by proposition [ propaff ]
, we can assume that @xmath78 is a co - moving inertial observer of @xmath95 at @xmath372 , i.e. @xmath436 , because of the following three statements . by ( v ) of proposition [ prop0 ] , for every @xmath388 , either of @xmath433 and @xmath437 can be obtained from the other by composing the other by a world - view transformation between inertial observers . by theorem [ thmpoi ] ,
world - view transformations between inertial observers are poincar - transformations .
poincar - transformations are affine and preserve the minkowski - distance . now , assume that @xmath78 is a co - moving inertial observer of @xmath95 at @xmath372
. then @xmath438 , @xmath439 and @xmath440 for every @xmath441 .
therefore @xmath442 since @xmath443 and @xmath444 if @xmath445 , we have that for all @xmath446 , @xmath447 let @xmath138 be fixed .
since @xmath448 and @xmath449 , there is a @xmath450 such that @xmath451 let such a @xmath452 be fixed . by ( [ tp - e2 ] ) ,
( [ tp - e3 ] ) and the fact that @xmath453 , we have that @xmath454 by this and ( [ tp - e1 ] ) , we have @xmath455 thus @xmath456 .
this completes the proof since @xmath457 .
[ trrem ] well parametrized curves are exactly the life - curves of accelerated observers , in models of , as follows .
let @xmath166 be an euclidean ordered field and let @xmath427 be well - parametrized .
then there are a model @xmath43 of , @xmath378 and @xmath379 such that @xmath458 and the ordered field reduct of @xmath43 is @xmath166 . recall that
if @xmath429 , then this @xmath43 is a model of .
this is not difficult to prove by using the methods of the present paper .
we say that @xmath169 is * vertical * iff @xmath459 .
[ lemwp ] let @xmath427 be well - parametrized . then ( i ) and ( ii ) below hold .
* let @xmath256 be an accumulation point of @xmath254 .
then @xmath295 is differentiable at @xmath13 and @xmath460 .
furthermore , @xmath461 iff @xmath259 is vertical .
* assume and that @xmath122 is definable .
let @xmath222\subseteq dom(f)$ ] .
then @xmath295 is increasing or decreasing on @xmath222 $ ] .
if @xmath295 is increasing on @xmath222 $ ] and @xmath339 , then @xmath462 for all @xmath344 $ ] .
let @xmath122 be well - parametrized . to prove ( i ) , let @xmath256 be an accumulation point of @xmath254 .
then @xmath259 is of minkowski - length @xmath97 . by proposition [ propaff ]
, @xmath295 is differentiable at @xmath13 and @xmath463 .
now , ( i ) follows from the fact that the absolute value of the time component of a vector of minkowski - length 1 is always greater than 1 and it is 1 iff the vector is vertical . to prove ( ii ) , assume and that @xmath122 is definable .
let @xmath222\subseteq dom(f)$ ] . from ( i )
, we have @xmath464 for all @xmath465 $ ] .
thus , by rolle s theorem , @xmath295 is injective on @xmath222 $ ] .
thus , by bolzano s theorem , @xmath295 is increasing or decreasing on @xmath222 $ ] since @xmath295 is continuous and injective on @xmath222 $ ] .
assume that @xmath295 is increasing on @xmath222 $ ] and @xmath339 .
then @xmath466 for all @xmath344 $ ] by the definition of the derivate .
hence , by ( i ) , @xmath467 for all @xmath344 $ ] .
[ thmjtwp ] assume .
let @xmath427 be definable , well - parametrized and @xmath222\subseteq dom(f)$ ] .
then ( i ) and ( ii ) below hold .
* @xmath468 . * if @xmath469 for an @xmath465 $ ] , then @xmath470 .
let @xmath427 be definable , well - parametrized and @xmath222\subseteq dom(f)$ ] .
we can assume that @xmath339 .
for every @xmath471 , @xmath292 is definable and differentiable on @xmath222 $ ] by proposition [ propaff ] .
then , by the main value theorem , there is an @xmath340 such that @xmath472 . by ( i ) of lemma [ lemwp ] , we have @xmath473 .
but then , @xmath474 .
this completes the proof of ( i ) . to prove ( ii ) , let @xmath344 $ ] be such that @xmath469 .
let @xmath475 be such that @xmath476 .
then , by the main value theorem , there is an @xmath477 such that @xmath478 . thus @xmath479 is not vertical .
therefore , by ( i ) of lemma [ lemwp ] , we have @xmath480 .
thus , by the definition of the derivate , there is a @xmath481 such that @xmath482 .
hence we have @xmath483 let us note that @xmath484 . by applying ( i )
to @xmath485 $ ] and @xmath486 $ ] , respectively , we get @xmath487 @xmath295 is increasing or decreasing on @xmath222 $ ] by ( ii ) of lemma [ lemwp ] . thus @xmath488 or @xmath489 .
now , by adding up the last three inequalities , we get @xmath470 . let @xmath490 . for convenience ,
we introduce the following notation : @xmath491 if @xmath492 and @xmath493 if @xmath494 .
a set @xmath495 is called * twin - paradoxical * iff @xmath496 , @xmath497 , @xmath498 if @xmath231 , for all @xmath169 if @xmath498 , then there is a @xmath63 such that @xmath499 , and for all distinct @xmath500 and for all @xmath501 , @xmath502 implies that @xmath503 . a positive answer to the following question would also provide a positive answer to question [ qtwp ] , cf .
@xcite .
[ qconv ] assume .
let @xmath427 be definable such that @xmath122 is differentiable on @xmath222 $ ] and @xmath504 .
furthermore , let the set @xmath505 \}$ ] be a subset of a twin - paradoxical set .
are then ( i ) and ( ii ) below true ?
* @xmath474 . *
if @xmath469 for an @xmath465 $ ] , then @xmath470 .
@xmath136 [ thmjeq ] assume .
let @xmath506 be definable and well - parametrized .
let @xmath222\subseteq dom(f)$ ] and @xmath507\subseteq dom(g)$ ] be such that @xmath508\}=\{g(r'):r'\in [ a',b']\}$ ] . then @xmath509 . by ( ii ) of lemma [ lemwp ]
, @xmath295 is increasing or decreasing on @xmath222 $ ] and so is @xmath510 on @xmath507 $ ] .
we can assume that @xmath511 $ ] , @xmath512 $ ] and that @xmath295 and @xmath510 are increasing on @xmath222 $ ] and @xmath507 $ ] , respectively .
is increasing on @xmath222 $ ] because the assumptions of the theorem remain true when @xmath122 and @xmath222 $ ] are replaced by @xmath513 and @xmath514 $ ] , respectively , and @xmath295 is decreasing on @xmath222 $ ] iff @xmath515 is increasing on @xmath514 $ ] . ] then @xmath516 .
furthermore , @xmath122 and @xmath278 are injective since @xmath295 and @xmath510 are such . since @xmath516 and @xmath510 is injective , @xmath517 .
let @xmath518 .
since @xmath519 and @xmath295 and @xmath510 are increasing , @xmath270 is an increasing bijection between @xmath222 $ ] and @xmath507 $ ] .
hence @xmath520 and @xmath521 .
we are going to prove that @xmath509 by proving that there is a @xmath357 such that @xmath522 for all @xmath344 $ ] .
we can assume that @xmath339 and @xmath523 .
by lemma [ lemwp ] , @xmath295 and @xmath510 are differentiable on @xmath222 $ ] and @xmath507 $ ] , respectively , and @xmath524 for all @xmath465 $ ] and @xmath525 for all @xmath526 $ ] . by ( iv ) and ( v ) of proposition [ propdiff ] , @xmath527 is also differentiable on @xmath226 .
by @xmath528 , we have @xmath529 .
thus @xmath530 for all @xmath356 by ( iv ) of proposition [ propdiff ] .
since both @xmath259 and @xmath531 are of minkowski - length @xmath97 and their time - components are positive and @xmath532 for all @xmath356 , we conclude that @xmath533 for all @xmath356 . by proposition [ propint ]
, we get that there is a @xmath357 such that @xmath522 for all @xmath356 and thus for all @xmath465 $ ] since @xmath270 is an increasing bijection between @xmath222 $ ] and @xmath507 $ ] .
[ thmtwp - proof ] assume and @xmath132 .
let @xmath378 and @xmath379 .
let @xmath534 , @xmath535 be such that @xmath536 , @xmath537\subseteq tr_k(k)$ ] and @xmath538\not\subseteq tr_m(k)$ ] , cf . figure 2 .
let us abbreviate @xmath433 by @xmath539 .
we are going to prove that @xmath540 by applying theorem [ thmjtwp ] to @xmath539 and @xmath541 $ ] . by proposition [ proptr ] ,
@xmath542 by , @xmath543 .
by @xmath544 , by @xmath545 , by @xmath546 and by @xmath547 , @xmath548 by @xmath534 and @xmath549 , we have that @xmath550 .
thus , by @xmath538\not\subseteq tr_m(k)$ ] , we have that @xmath551 .
hence , by ( [ twp - e2 ] ) , @xmath76 .
we can assume that @xmath552 . by ( viii ) of proposition [ prop0 ] , @xmath553 .
since @xmath537\subseteq tr_k(k)$ ] , @xmath554\subseteq dom(tr).\ ] ] by ( i ) of lemma [ lemwp ] , ( [ twp - e1 ] ) and ( [ twp - e3 ] ) , we have that @xmath555 is differentiable on @xmath556 $ ] , thus it is continuous on @xmath556 $ ] .
let @xmath557\subseteq \bar{t}$ ] be such that @xmath558 . by bolzano s theorem and ( [ twp - e2 ] ) ,
there is an @xmath559 $ ] such that @xmath560 .
let such an @xmath13 be fixed .
@xmath561 since @xmath562 by ( vii ) of proposition [ prop0 ] .
but then @xmath563 .
hence @xmath564 .
thus @xmath565\quad\mbox{and}\quad tr(x)_s\neq tr(p_t)_s\ ] ] since @xmath566 .
now , by ( [ twp - e1])([twp - e4 ] ) above , we can apply ( ii ) of theorem [ thmjtwp ] to @xmath539 and @xmath556 $ ] , and we get that @xmath567 .
[ thmeq - proof ] assume and @xmath132 .
let @xmath95 and @xmath78 be observers .
let @xmath568 , @xmath535 be such that @xmath569\}=\{ev_m(r'):r'\in [ p ' q']\}$ ] , cf . the right hand side of figure 2 .
thus @xmath537\subseteq cd(k)$ ] and @xmath538\subseteq cd(m)$ ] . by ,
@xmath570 and @xmath571 . therefore @xmath537\subseteq tr_k(k)\subseteq\bar t$ ] and @xmath538\subseteq tr_m(m)\subseteq\bar t$ ] .
we can assume that @xmath572 and @xmath573 .
let @xmath388 .
we are going to prove that @xmath574 , by applying theorem [ thmjeq ] as follows : let @xmath222:=[p_t , q_t]$ ] , @xmath507:=[p'_t , q'_t]$ ] , @xmath575 and @xmath576 . by ( viii ) of proposition [ prop0 ] , by @xmath537\subseteq tr_k(k)$ ] and by @xmath538\subseteq tr_m(m)$ ] , we conclude that @xmath222\subseteq dom(f)$ ] and @xmath507\subseteq dom(g)$ ] . by proposition
[ proptr ] , @xmath122 and @xmath278 are well - parametrized and definable .
we have @xmath508\}=\{g(r'):r'\in [ a',b']\}$ ] since @xmath577\}=\{ev_m(r'):r'\in [ p ' q']\}$ ] .
thus , by theorem [ thmjeq ] , we conclude that @xmath509 .
thus @xmath574 and this is what we wanted to prove .
[ thmnoind - proof ] [ thmmo - proof ] we will construct three models .
let @xmath578 be an euclidean ordered field different from @xmath145 .
for every @xmath169 , let @xmath579 denote the translation by vector @xmath29 , i.e. @xmath580 .
@xmath121 is called * translation - like * iff for all @xmath581 , there is a @xmath308 such that for all @xmath582 , @xmath583 and for all @xmath68 , @xmath584 and @xmath585 imply that @xmath586 .
let @xmath587 be translation - like .
first we construct a model @xmath588 of and ( i ) and ( ii ) of theorem [ thmmo ] for @xmath166 and @xmath95 , which will be a model of ( iii ) and ( iv ) of theorem [ thmmo ] if @xmath95 is a bijection .
we will show that is false in @xmath588
. then we will choose @xmath166 and @xmath95 appropriately to get the desired models in which is false , too .
let the ordered field reduct of @xmath588 be @xmath166 .
let @xmath589 be a partition s are disjoint and @xmath590 .
] of @xmath6 such that every @xmath591 is open , @xmath592 and for all @xmath593 and @xmath594 , @xmath595 .
such a partition can easily be constructed .
be a non - empty bounded set that does not have a supremum .
let @xmath596 , @xmath597 , @xmath598 , @xmath599 and @xmath600 .
] let @xmath601 for every @xmath169 , cf . figure 4 .
[ r][r]@xmath173 [ r][r]@xmath174 [ r][r]@xmath30 [ r][r]@xmath29 [ b][b]world - view of @xmath78 [ b][b]world - view of @xmath602 [ l][l]@xmath603 [ l][l]@xmath604 [ l][l]@xmath605 [ l][l]@xmath606 [ l][l]@xmath607 [ r][r]@xmath608 [ r][r]@xmath609 [ b][b]@xmath602 [ t][t]@xmath610 [ t][t]@xmath611 [ t][t ] for the proofs of theorems [ thmnoind ] and [ thmmo].,title="fig:",scaledwidth=90.0% ] it is easy to see that @xmath602 is a translation - like bijection .
let @xmath612 , @xmath613 , @xmath614 and @xmath615 . recall that @xmath65 is the origin .
first we give the world - view of @xmath616 then we give the world - view of an arbitrary observer @xmath270 by giving the world - view transformation between @xmath270 and @xmath616 .
let @xmath617 and @xmath618 for all @xmath619 and @xmath620 . and let @xmath621 for all @xmath169 .
let @xmath622 for all @xmath620 . from these world - view transformations , we can obtain the world - view of each observer @xmath270 in the following way : @xmath623 for all @xmath624 . and from the world - views
, we can obtain the @xmath11 relation as follows : for all @xmath620 , @xmath625 and @xmath169 , let @xmath626 iff @xmath627 .
thus we are given the model @xmath588 .
we note that @xmath628 and @xmath629 for all @xmath630 and @xmath105 .
it is easy to check that the axioms of and ( i ) and ( ii ) of theorem [ thmmo ] are true in @xmath588 and that if @xmath95 is a bijection , then ( iii ) and ( iv ) of theorem [ thmmo ] are also true in @xmath588 .
let @xmath631 be such that @xmath632 , @xmath633 ; and let @xmath634 , @xmath635 and @xmath636 .
it is easy to check that is false in @xmath588 for @xmath602 , @xmath78 , @xmath29 , @xmath30 , @xmath173 and @xmath174 , i.e. @xmath637 , @xmath638 , @xmath639 , @xmath537 \subseteq tr_{k'}(k')$ ] , @xmath538\not\subseteq tr_m(k')$ ] and @xmath640 , cf .
figure 4 .
[ t][t]world - view of @xmath78 [ t][t]world - view of @xmath95 [ r][r]@xmath173 [ r][r]@xmath174 [ r][r]@xmath30 [ r][r]@xmath29 [ rt][rt]@xmath29 [ lb][lb]@xmath30 [ rt][rt]@xmath173 [ lb][lb]@xmath174 [ r][r]@xmath88 [ r][r]@xmath85 [ b][b]@xmath95 [ t][t]@xmath108 [ t][t]@xmath133 [ t][t ] [ t][t ] [ t][t ] [ l][l]@xmath83 [ r][r]@xmath83 [ r][r]@xmath641 [ r][r]@xmath642 [ r][r]@xmath51 [ l][l]@xmath643 [ l][l]@xmath93 [ l][l]@xmath603 [ l][l]@xmath604 for the proofs of theorems [ thmnoind ] and [ thmmo].,title="fig:",scaledwidth=85.0% ] to construct the first model , let @xmath166 be an arbitrary euclidean ordered field different from @xmath145 and let @xmath644 be a partition of @xmath6 such that for all @xmath645 and @xmath646 , @xmath274 .
let @xmath647 for every @xmath169 , cf .
figure 5 .
it is easy to see that @xmath95 is translation - like .
let @xmath631 be such that @xmath648 and @xmath649 ; and let @xmath650 , @xmath651 and @xmath636 .
it is also easy to check that is false in @xmath588 for @xmath95 , @xmath78 , @xmath29 , @xmath30 , @xmath173 and @xmath174 , i.e.@xmath568 , @xmath535 , @xmath569\}=\{ev_m(r'):r'\in [ p ' q']\}$ ] and @xmath652 , cf .
figure 5 .
this completes the proof of theorem [ thmnoind ] . to construct the second model ,
let @xmath166 be an arbitrary non - archimedean euclidean ordered field .
let @xmath653 if @xmath224 and @xmath654 is infinitesimally small .
it is easy to see that @xmath655 is an equivalence relation .
let us choose an element from every equivalence class of @xmath655 and let @xmath656 denote the chosen element equivalent with @xmath657 .
let @xmath658 for every @xmath169 , cf.figure 5 .
it is easy to see that @xmath95 is a translation - like bijection .
let @xmath659 , @xmath660 , @xmath661 , @xmath662 and @xmath636 .
it is also easy to check that is false in @xmath588 for @xmath95 , @xmath78 , @xmath29 , @xmath30 , @xmath173 and @xmath174 , cf .
figure 5 .
[ t][t]@xmath51 [ t][t]@xmath663 [ t][t]@xmath664 [ t][t]@xmath665 [ r][r]@xmath97 [ l][l]@xmath666 [ l][l]@xmath667 [ l][l]@xmath668 [ l][l]@xmath669 [ rb][rb]etc .
[ t][t]@xmath642 [ t][t]@xmath670 [ b][b]@xmath603 [ b][b]@xmath604 for the proofs of theorems [ thmnoind ] and [ thmmo].,title="fig:",scaledwidth=90.0% ] to construct the third model , let @xmath166 be an arbitrary countable archimedean euclidean ordered field and let @xmath671 for every @xmath169 where @xmath672 is constructed as follows , cf .
figures 5 , 6 .
we can assume that @xmath166 is a subfield of @xmath145 by ( * ? ? ?
* theorem 1 in viii ) .
let @xmath51 be a real number that is not an element of @xmath6 .
let us enumerate the elements of @xmath673\cap { { \mathrm}{f}}$ ] and denote the @xmath44-th element with @xmath674 .
first we cover @xmath673\cap { { \mathrm}{f}}$ ] with infinitely many disjoint subintervals of @xmath673 $ ] such that the sum of their length is @xmath97 , the length of each interval is in @xmath6 and the distance of the left endpoint of each interval from @xmath51 is also in @xmath6 .
we are going to construct this covering by recursion . in the @xmath44-th step
, we will use only finitely many new intervals such that the sum of their length is @xmath675 .
in the first step , we cover @xmath663 with an interval of length @xmath676 .
suppose that we have covered @xmath674 for each @xmath677 .
since we have used only finitely many intervals yet , we can cover @xmath678 with an interval that is not longer than @xmath679 .
since @xmath680 , it is easy to see that we can choose finitely many other subintervals of @xmath673 $ ] to be added to this interval such that the sum of their length is @xmath679 .
we are given the covering of @xmath673 $ ] .
let us enumerate these intervals .
let @xmath591 be the @xmath44-th interval , @xmath681 be the length of @xmath591 , @xmath682 and @xmath683 the distance of @xmath51 and the left endpoint of @xmath591 .
@xmath684 since @xmath685 .
let @xmath686 for all @xmath422 , cf .
figure 6 .
it is easy to see that @xmath95 is a translation - like bijection .
let @xmath68 be such that @xmath687 and @xmath688 ; and let @xmath650 , @xmath651 and @xmath636 .
it is also easy to check that is false in @xmath588 for @xmath95 , @xmath78 , @xmath29 , @xmath30 , @xmath173 and @xmath174 , cf .
figure 5 .
let @xmath166 be a field elementarily equivalent to @xmath145 , i.e. such that all fol - formulas valid in @xmath145 are valid in @xmath166 , too .
assume that @xmath166 is not isomorphic to @xmath145 .
e.g. the field of the real algebraic numbers is such .
let @xmath43 be a model of @xmath191 with field - reduct @xmath166 in which neither @xmath689 nor @xmath690 is true .
such an @xmath43 exists by theorem [ thmnoind ] .
since @xmath691 by assumption , this shows @xmath692 . in a subsequent paper , we will discuss how the present methods and in particular and can be used for introducing gravity via einstein s equivalence principle and for proving that gravity `` causes time run slow '' ( known as gravitational time dilation ) . in this connection we would like to point out that it is explained in misner et al .
@xcite that the theory of accelerated observers ( in flat space - time ! ) is a rather useful first step in building up general relativity by using the methods of that book .
a fol - formula expressing is : @xmath693 .
\end{split}\ ] ] a fol - formula expressing is : @xmath694 \wedge\\ \big[\forall ph\;\forall\lambda\quad { { \mathrm}{ph}}(ph)\wedge{{\mathrm}{f}}(\lambda)\wedge{{\mathrm}{w}}(m , ph , p)\wedge&{{\mathrm}{w}}(m , ph , q)\\ \longrightarrow\ ; { { \mathrm}{w}}\big(&m , ph , q+\lambda(p - q)\big)\big ] .
\end{split}\ ] ] a fol - formula expressing is : @xmath695\big ) .
\end{split}\ ] ]
we are grateful to victor pambuccian for careful reading the paper and for various useful suggestions .
we are also grateful to hajnal andrka , ramn horvth and bertalan pcsi for enjoyable discussions .
special thanks are due to hajnal andrka for extensive help and support in writing the paper , encouragement and suggestions .
h. andrka , j. x. madarsz and i. nmeti , `` logical axiomatizations of space - time , '' in _ non - euclidean geometries _ , e. molnr , ed .
( kluwer , dordrecht , 2005 ) http://www.math-inst.hu/pub/algebraic-logic/lstsamples.ps .
h. andrka , j. x. madarsz and i. nmeti , with contributions from a. andai , g. sgi , i. sain and cs .
t oke , `` on the logical structure of relativity theories , '' research report , alfrd rnyi institute of mathematics , budapest ( 2002 ) http://www.math-inst.hu/pub/algebraic-logic/contents.html . | we study the foundation of space - time theory in the framework of first - order logic ( fol ) . since the foundation of mathematics has been successfully carried through ( via set theory ) in fol , it is not entirely impossible to do the same for space - time theory ( or relativity ) .
first we recall a simple and streamlined fol - axiomatization of special relativity from the literature .
is complete with respect to questions about inertial motion .
then we ask ourselves whether we can prove the usual relativistic properties of accelerated motion ( e.g. , clocks in acceleration ) in . as it turns out , this is practically equivalent to asking whether is strong enough to `` handle '' ( or treat ) accelerated observers . we show that there is a mathematical principle called induction ( ) coming from real analysis which needs to be added to in order to handle situations involving relativistic acceleration .
we present an extended version of which is strong enough to handle accelerated motion , in particular , accelerated observers . among others ,
we show that the twin paradox becomes provable in , but it is not provable without .
key words : twin paradox , relativity theory , accelerated observers , first - order logic , axiomatization , foundation of relativity theory |
in nuclear , atomic , and chemical physics , it is often useful to introduce an effective interaction acting in a chosen model space ( p space ) of tractable dimension . in nuclear physics ,
much effort has been made both as regards formal theories and their applications @xcite .
recently the effective interaction method has been applied to new fields of many - body physics , such as quantum dots @xcite and many - boson systems @xcite . among many approaches , we here direct our attention to the krenciglowa - kuo ( kk ) @xcite and the lee - suzuki ( ls ) @xcite methods .
these two methods are constructed in terms of the so called @xmath2-box as a building block of formulation .
the kk approach has very simple structure and the effective interaction is obtained in an iterative way .
if the iteration converges , in almost all numerical calculations , eigenvalues are given for the states which have the largest overlaps with the chosen model space .
on the other hand , the ls method reproduces eigenvalues for the states which lie closest to the chosen unperturbed energy .
originally the ls method had been presented to resolve the difficulty of divergence in the perturbation expansion .
the ls method is rather complicated in structure and higher derivatives of the @xmath2-box with respect to starting energy are necessary if one wishes to obtain more accurate solutions .
both of the two theories yield only certain of the true eigenvalues of the original hamiltonian .
this restriction is not desirable . in a formal point of view
, the @xmath2-box itself contains information regarding all of the true eigenvalues . for a given model space of dimension @xmath4
, there would be a method of reproducing any @xmath4 eigenvalues among all the true eigenvalues .
in many cases of applying the effective interaction theories an iteration or a recursion method has often been employed . in general , the convergence in such a method depends strongly on the properties of the eigenstates of the hamiltonian @xmath3 . in actual calculations
the information on the true eigenstates is not given beforehand .
therefore , it is impossible to control the convergence in the iteration . in many cases we can not know whether the iteration is convergent or not before starting calculations . even if the iteration converges , we do not know which eigenvalues are reproduced in all of the eigenvalues .
another difficulty encountered in actual calculations is the pole problem .
the @xmath2-box itself has poles at the energies which are the eigenvalues of @xmath3 in the complement space ( q space ) .
the presence of poles causes often instability in numerical calculations . we shall show that it is indeed possible to resolve these difficulties by introducing a new vertex function , @xmath0 , in place of the @xmath2-box by kuo _
et al . _
preliminary version of the present work has been reported in ref .
very recently dong , kuo , and holt have followed the present approach and applied to the actual calculations of the shell - model effective interactions @xcite .
these works have shown that the present method has a possibility of providing a suitable framework with @xmath0 as well as the kk approach .
the organization of the present paper is as follows : in sec .
[ formulation ] we outline the @xmath2-box formalism and the kk method .
a new vertex function operator , @xmath0 , is introduced and some of its mathematical properties are clarified .
the algorithm of the calculation procedure is given by applying the secant and the newton - raphson methods . in sec .
[ test ] test calculations are made in order to assess the present approach .
we examine whether the theoretical predictions are verified or not in an exactly solvable model .
concluding remarks are given in sec .
[ concluding ] .
we consider a general quantum system which is described by a hamiltonian @xmath3 .
we write an eigenvalue equation with the eigenvalues @xmath5 and the eigenstates @xmath6 as @xmath7 the hamiltonian @xmath3 is supposed to be composed of the unperturbed hamiltonian @xmath8 and the perturbation @xmath9 , _
i.e. _ , @xmath10 .
we decompose the entire hilbert space into the model space ( p space ) and its complement ( q space ) with the projection operators @xmath11 and @xmath12 , respectively .
basic properties of the projection operators are @xmath13 and @xmath14 .
we here assume that @xmath8 is decoupled between the p and q spaces as @xmath15 in the present work we consider a case that the p - space states have a degenerate unperturbed energy @xmath16 , _
i.e. _ , @xmath17 then the p - space eigenvalue equation is written with the effective interaction @xmath18 and the eigenstate @xmath19 as @xmath20 where @xmath4 is the dimension of the p space . in the above @xmath4 eigenvalues
@xmath21 s should agree with @xmath4 of the true eigenvalues in eq .
( [ eq : true - eigen - eq ] ) .
various solutions for @xmath18 are possible , and many theoretical frameworks have been given for obtaining @xmath18 . probably the most widely applied effective interaction is given by imposing the condition that the model - space eigenstate @xmath19 in eq .
( [ eq : p - space - eigen ] ) should agree with the p - space component of the true eigenstate @xmath22 of @xmath3 , _
i.e. _ , @xmath23 . this restriction on @xmath19 leads to the standard non - hermitian form of @xmath18 @xcite . among many approaches to the effective interaction @xmath18
we here discuss the kk formalism @xcite .
originally the kk method is based on a diagrammatic representation of the effective interaction , which has been known as the @xmath2-box folded - diagram method originated by kuo _
et al . _
@xcite . for obtaining @xmath18 ,
one first calculates the vertex function called the @xmath2-box which is defined as the sum of all the linked and non - folded diagrams .
next one should add the folded diagrams , which can be carried out rather simply by applying the energy - derivative expression of the @xmath2-box @xcite .
originally the kk approach was proposed to derive an effective interaction acting among a few valence particles outside the closed - shell core . however , in the present work , we consider a general quantum system and wish to reproduce @xmath4 total energies of @xmath3 by introducing an effective interaction . for this case
the kk method can also be applied by defining the @xmath2-box in an operator form as @xmath24 which is a function of an energy variable @xmath1 .
equation ( [ eq : q - box ] ) is equivalent to the energy - dependent form of the effective interaction given by bloch and horowitz in the many - body perturbation theory @xcite . with the @xmath2-box
the effective interaction @xmath18 can be expanded into @xmath25 where @xmath26 and @xmath27 with @xmath28 here @xmath29 is the starting energy or the degenerate unperturbed energy as given in eq .
( [ eq : model - space - hamiltonian ] ) . if the series expansion in eq .
( [ eq : r - expansion ] ) converges , @xmath18 is given in a formal way by @xmath30 where @xmath31 and @xmath19 are given in eq .
( [ eq : p - space - eigen ] ) , and @xmath32 is the biorthogonal state defined through the orthogonality @xmath33 .
we here note that , as seen in eqs .
( [ eq : p - space - eigen ] ) and ( [ eq : rkk-1 ] ) , the derivation of the effective interaction @xmath18 is equivalent to determining @xmath4 eigenvalues \{@xmath21 } and the corresponding p - space eigenstates @xmath34 . using eqs .
( [ eq : p - space - eigen ] ) and ( [ eq : rkk-1 ] ) , the effective interaction @xmath18 is rewritten simply as @xmath35 and the model - space eigenvalue equation with the effective interaction @xmath18 is expressed as @xmath36|{\phi_k}\rangle = e_{k}|{\phi_k}\rangle .
\end{aligned}\ ] ] from eqs .
( [ eq : rkk-1 ] ) and ( [ eq : p - space - eigen - eq-2 ] ) we understand that @xmath18 can be given by calculating the @xmath2-box at the true eigenvalues @xmath37 of @xmath3 as starting energies which are determined self - consistently . we here show that there is another way of deriving the effective interaction @xmath18 .
we first note that the @xmath2-box is a function of @xmath1 and resultantly the eigenvalues of @xmath38 are also functions of @xmath1 .
we write the eigenvalue equations in the p space for an arbitrary energy @xmath1 as @xmath39 |\psi_{m}\rangle = g_{m}(e ) | \psi_{m}\rangle , m = 1 , 2 , \cdots , d.\end{aligned}\ ] ] since the p space is @xmath4 dimensional , we have @xmath4 eigenvalues denoted by @xmath40 , which we label in order of energy as @xmath41 .
it may be clear from eqs .
( [ eq : p - space - eigen - eq-2 ] ) and ( [ eq : q - box - eigen - g ] ) that the eigenvalues \{@xmath21 } in eq .
( [ eq : p - space - eigen - eq-2 ] ) can be given by solving the following equations @xmath42 various mathematical methods have been known to solve such equations .
we should note that a set of the equations ( [ eq : egk ] ) are independent of the properties of the eigenstates of @xmath3 , such as p - space overlaps or energy spacings .
therefore , in principle , it is possible to reproduce the true eigenvalues @xmath43 of @xmath3 more than in the usual kk method based on eq .
( [ eq : p - space - eigen - eq-2 ] ) . however , as seen in eq .
( [ eq : q - box ] ) , there appear poles in the @xmath2-box when @xmath1 approaches one of the eigenvalues of @xmath44 .
the poles in @xmath45 induce also the poles in @xmath46 in eq .
( [ eq : q - box - eigen - g ] ) .
such a situation causes instability in numerically solving eq .
( [ eq : egk ] ) for @xmath47 around the pole positions . in order to resolve the pole problem
we introduce a new vertex function of an energy variable @xmath1 , which is a p - space operator defined in terms of the @xmath2-box and its energy derivative as @xcite @xmath48 .
\end{aligned}\ ] ] hereafter we shall refer to @xmath49 as the @xmath50-box .
we note here that the @xmath50-box agrees , at @xmath51 , with the first - order recursive solution in the ls method @xcite . in a recent paper of dong , kuo , and holt @xcite the definition of @xmath0 is given in a more general case with a non - degenerate p - space unperturbed hamiltonian @xmath52 . in this case
one should replace @xmath53 by @xmath54 . as in eq .
( [ eq : q - box - eigen - g ] ) with the @xmath2-box , we consider an eigenvalue problem @xmath55 |\psi_{m}\rangle = f_{m}(e ) | \psi_{m}\rangle,\ \
m=1 , 2 , \cdots , d,\end{aligned}\ ] ] where @xmath56 are @xmath4 eigenvalues which are functions of @xmath1 .
we here label @xmath56 in order of energy as @xmath57 .
the @xmath50-box and the associated functions @xmath58 have the following properties : 1 .
using eqs .
( [ eq : q - box ] ) , ( [ eq : rkk-1 ] ) , ( [ eq : rkk-2 ] ) , and ( [ eq : z - operator ] ) , we have , for the p - space eigenstates @xmath59 in eq .
( [ eq : p - space - eigen - eq-2 ] ) , @xmath60 } | { \phi_k } \rangle\langle { \tilde\phi _ k } | \nonumber\\ & = & r . \end{aligned}\ ] ] the above fact means that , replacing @xmath61 by @xmath49 in eq .
( [ eq : rkk-1 ] ) , a new solution for the effective interaction @xmath18 can be derived as @xmath62 and equivalently eq . ( [ eq : p - space - eigen - eq-2 ] ) with the @xmath2-box
is replaced by @xmath63|\phi_{k}\rangle = e_{k}|\phi_{k}\rangle.\end{aligned}\ ] ] from the above relations between two approaches with @xmath45 and @xmath0 we may call @xmath64 in eq .
( [ eq : rekk ] ) the extended krenciglowa - kuo ( ekk ) solution . in the same way as in eq .
( [ eq : egk ] ) , the true eigenvalues @xmath65 can be given by solving the equations @xmath66 2 .
using eqs .
( [ eq : q - box ] ) , ( [ eq : q - e0-derivative ] ) , and ( [ eq : z - operator ] ) we can derive a formal expression for the energy derivative of @xmath0 as @xmath67 .
\end{aligned}\ ] ] in the paper of dong _ et al . _
@xcite , the above expression has also been given for a general case with the non - degenerate @xmath54 .
if @xmath1 is one of the true eigenvalues @xmath68 satisfying eq .
( [ eq : ze ] ) , we see that the energy derivative of @xmath49 becomes zero at @xmath69 , namely @xmath70 resultantly we have for the energy derivative of @xmath71 @xmath72 these results for the energy derivatives have been pointed out in the previous paper @xcite .
we discuss here some problems associated with the poles of @xmath61 .
first we consider the eigenvalue equation for the q - space hamiltonian @xmath44 written as @xmath73 where @xmath74 and @xmath75 are the eigenvalue and the eigenstate , respectively .
it may be clear from eq .
( [ eq : q - box ] ) that @xmath45 has a pole at @xmath76 .
we define a p - space operator @xmath77 with the q - space eigenstate @xmath78 in eq .
( [ eq : qhq - eigenvalue eq ] ) as @xmath79 we write the eigenvalue equation for @xmath77 with an eigenvalue @xmath80 as @xmath81 from the definition of @xmath82 in eq .
( [ eq : x - operator ] ) we see that @xmath82 is hermitian and positive semi - definite , that is , @xmath82 has positive or zero eigenvalue because of the inequality @xmath83 we can further prove that there is only one eigenstate , denoted by @xmath84 , with a positive eigenvalue and that the eigenvalues of all the other eigenstates are zero .
the only one positive eigenvalue of eq .
( [ eq : x - eigenvalue - eq ] ) is given by @xmath85 where @xmath86 are the basis - state vectors of the p space .
the proof is as follows : we condsider a matrix representation of @xmath77 with the matrix element @xmath87 .
then we obtain the trace of @xmath77 as @xmath88 we note that the state vector @xmath89 can be written explicitly as @xmath90 where the symbol t means transpose of @xmath91 matrix , and @xmath92 is the normalization factor with @xmath93 .
this state @xmath84 becomes an eigenstate of @xmath77 and has a positive eigenvalue as @xmath94 the eigenvalue @xmath95 coincides with the trace of @xmath77 . recalling a well - known theorem on the trace of a square matrix , we see that the trace of @xmath82 should be equal to the sum of the eigenvalues of @xmath77 .
it follows immediately that all the other eigenvalues except @xmath95 become zero .
next , we prove that eq .
( [ eq : ze ] ) or ( [ eq : fe ] ) has additional solutions other than the true eigenvalues @xmath5 .
we note , in the vicinity of a pole of @xmath45 where @xmath96 with a small deviation @xmath97 , that @xmath45 in eq .
( [ eq : q - box ] ) is expressed as @xmath98 here we have used the definition of @xmath77 in eq .
( [ eq : x - operator ] ) .
the operations of @xmath99 , and @xmath100 on the state @xmath89 yield in leading order , respectively @xmath101 using eq .
( [ eq : z - operator ] ) and the above relations , we obtain at @xmath96 @xmath102 |\mu_{0}\rangle } \nonumber \\ & = & \left [ e_{0}+(\frac{x_{\mu_0}}{\delta^2})^{-1 } \right .
\nonumber\\ & & \hspace{10 mm } \times \left .
\left [ \frac{x_{\mu_0}}{\delta}+\frac{x_{\mu_0}(\varepsilon_{q}+\delta - e_0 ) } { \delta^2 } \right ] \right ] |\mu_{0}\rangle \nonumber\\ & = & ( \varepsilon_{q}+2\delta ) |\mu_{0}\rangle.\end{aligned}\ ] ] then we have in the limit of @xmath103 @xmath104 |\mu_{0}\rangle&=&\varepsilon_{q}|\mu_{0}\rangle.\end{aligned}\ ] ] this fact means that the state @xmath89 and the energy @xmath74 can be an additional solution to eq .
( [ eq : ze ] ) .
this also means that the pole energy @xmath105 satisfies eq .
( [ eq : fe ] ) .
we further consider the energy derivatives of @xmath0 and @xmath71 at the pole positions .
subtraction of eq .
( [ eq : z - mu ] ) from eq .
( [ eq : z - mu - delta ] ) follows @xmath106 we can derive the above result , in another way , by using eqs .
( [ eq : z - diff ] ) and ( [ eq : z - mu ] ) @xmath107 then , using eqs .
( [ eq : q1-mu ] ) and ( [ eq : q2-mu ] ) and taking the limit of @xmath108 , we obtain @xmath109 as a consequence we have @xmath110 this property of the energy derivative of @xmath71 at a pole position is quite contrast to that for a true eigenvalue in eq .
( [ eq : fk - derivative ] ) .
these properties of the energy derivative of @xmath71 given in eqs .
( [ eq : fk - derivative ] ) and ( [ eq : f - derivative - at - pole ] ) can be used to distinguish the pole energy solutions from those of the true eigenvalues of @xmath3 .
it is also noted , from eqs .
( [ eq : z - mu ] ) and ( [ eq : f - derivative - at - pole ] ) , that the function @xmath71 is a well - behaved function of @xmath1 and has no singularity even at the pole energies .
iteration methods have been employed quite often for obtaining the effective interaction or equivalently for determining the true eigenvalues of @xmath3 . in the previous subsections
we have given four basic equations , namely eqs .
( [ eq : p - space - eigen - eq-2 ] ) and ( [ eq : egk ] ) for the kk and eqs .
( [ eq : ze ] ) and ( [ eq : fe ] ) for the ekk methods . in the kk approach eqs .
( [ eq : p - space - eigen - eq-2 ] ) and ( [ eq : egk ] ) lead , respectively , to the following iterative equations : @xmath111|{\phi_k^{(n+1)}}\rangle&=&e_{k}^{(n+1)}|{\phi_k^{(n+1)}}\rangle \end{aligned}\ ] ] and @xmath112 in the same way two iterative equations can be derived from eqs .
( [ eq : ze ] ) and ( [ eq : fe ] ) for the ekk method , respectively as @xmath113|{\phi_k^{(n+1)}}\rangle&=&e_{k}^{(n+1)}|{\phi_k^{(n+1)}}\rangle \end{aligned}\ ] ] and @xmath114 the iterative equation ( [ eq : p - space - eigen - eq-2-iteration ] ) has long been used as a standard method in the kk approach
. the convergence condition has been investigated in many of the theoretical and numerical studies @xcite . from these studies
it has been known that , if the iteration converges , the kk approach in eq .
( [ eq : p - space - eigen - eq-2-iteration ] ) derives @xmath4 eigenvalues of the eigenstates of @xmath3 with the largest p - space overlaps .
however , it has also been known that the iteration in eq .
( [ eq : p - space - eigen - eq-2-iteration ] ) does not always converge .
recently takayanagi @xcite has pointed out an exceptional case that the kk method does not reproduce states which have the largest p - space overlaps even if the iteration converges . although the kk iteration method has been applied widely and
has brought remarkable results in actual calculations , the rigorous condition of convergence for eq .
( [ eq : p - space - eigen - eq-2-iteration ] ) has not yet been made clear .
the convergence condition for the iteration in eq .
( [ eq : egk - iteration ] ) requires mathematically that the energy derivative of @xmath115 at the solution @xmath116 should satisfy @xmath117 . in general , this convergence condition is considered to be satisfied by certain of the true eigenvalues of @xmath3 .
it implies that this iteration method is restrictive for the purpose of reproducing the solutions of @xmath5 as many as possible .
the iteration in eq .
( [ eq : z - p - space - eigen - eq-2-iteration ] ) leads to a new scheme of the calculations of @xmath47 and @xmath118 in eq .
( [ eq : rekk ] ) . as in the case of the kk iteration in eq .
( [ eq : p - space - eigen - eq-2-iteration ] ) , the convergence condition in eq .
( [ eq : z - p - space - eigen - eq-2-iteration ] ) is quite complicated and has not yet been made clear .
however , the iteration in eq .
( [ eq : z - p - space - eigen - eq-2-iteration ] ) would be applicable in actual calculations . in the paper of dong _
et al . _
@xcite , they have employed this iteration scheme to a model problem and actual calculations of the shell - model effective interaction .
they have concluded that the two iterations in eq .
( [ eq : p - space - eigen - eq-2-iteration ] ) for the kk and eq .
( [ eq : z - p - space - eigen - eq-2-iteration ] ) for the ekk methods are both suitable and efficient .
graphical representation of the modified secant method .
the @xmath119 is the straight line passing through the points a and b. the @xmath120 and @xmath121 are the tangents to the function @xmath122 at the points a and b , respectively . see text for detailed explanation . ]
another iterative equations in the ekk method is given in eq .
( [ eq : efk - iteration ] ) .
the convergence condition for this iteration is given by @xmath123 .
it is clear from the property of the energy derivative of @xmath124 in eq .
( [ eq : fk - derivative ] ) that the convergence condition is satisfied for any of the true eigenvalues @xmath47 .
the iteration in eq .
( [ eq : efk - iteration ] ) converges surely to one of @xmath47 .
we also note that this iteration in eq .
( [ eq : efk - iteration ] ) never reaches the pole - energy solutions due to the property in eq .
( [ eq : f - derivative - at - pole ] ) of the energy derivative of @xmath124 .
we can show that the iteration in eq .
( [ eq : efk - iteration ] ) is equivalent to that in the newton - raphson method which is often used to solve non - linear equations .
the newton - raphson method has been known to derive quadratic convergence , that is , the number of correct digits is doubled at each step of iteration . on the contrary the usual methods , including the kk and the ls methods ,
derive single ( or linear ) convergence .
the iteration in eq .
( [ eq : efk - iteration ] ) guarantees surely faster convergence than the usual iterations given so far .
we present a new method for solving eq .
( [ eq : fe ] ) derived on the basis of the @xmath50-box and its associated functions @xmath125 .
the solutions can be obtained by finding the energies of the intersections of two graphs , @xmath126 and @xmath127 .
as we have already shown , the @xmath50-box has no singularities at the pole positions of the @xmath2-box .
thus , @xmath128 can be considered to be well - behaved functions of @xmath1 .
these characteristics of the functions @xmath128 enable us to employ some of the mathematically well - established methods for solving non - linear equations .
the procedure of solving eq .
( [ eq : fe ] ) in the present approach , which we call the graphical method , is given as follows : we first draw graphs of @xmath126 for @xmath129 and @xmath127 .
the energies at the intersections of these graphs become the solutions to @xmath130 . furthermore
, the energy derivatives of @xmath128 at the intersections distinguish the pole - energy solutions @xmath131 from those of the true eigenvalues @xmath5 .
as was proved in the previous subsections , the energy derivatives , denoted by @xmath132 , should be @xmath133 for @xmath134 and @xmath135 for @xmath136 .
with these simple regulations for @xmath132 we can easily specify the solutions of the true eigenvalues @xmath43 .
we next figure out , from the intersections of the graphs , roughly estimated solutions for @xmath137 . starting with these approximate solutions , we proceed to make further precise calculations for the solutions . for this purpose
we employ a combined method of the secant and the newton - raphson methods @xcite .
the algorithm of this method is as follows : let us define a function @xmath122 as @xmath138 , where @xmath139 is one of the functions @xmath128 .
the solutions to @xmath130 are obtained as the roots of @xmath140 .
suppose that @xmath141 and @xmath142 be two values which bracket one of the roots and satisfy @xmath143 and @xmath144 .
we suppose also that @xmath122 be a monotone function on the interval @xmath145 .
these suppositions on @xmath141 and @xmath142 follow that only one root exists on the interval @xmath145 . as shown in fig .
[ fig-1new ] , we further determine five points according to @xmath146 the value @xmath147 in eq .
( [ eq : e2 ] ) is the better approximate solution in the usual secant method .
the @xmath148 and @xmath149 are the approximate values in the newton - raphson method .
we can easily select two values , @xmath150 and @xmath151 , among @xmath152 such that they are neighboring on the @xmath1-axis and satisfy @xmath153 and @xmath154 .
.[table-1 ] convergence of the eigenvalues of the lowest - lying and second lowest - lying states obtained in the iterations @xmath155 and @xmath156 for the kk and for the ekk methods , respectively , for the model hamiltonian with the strength @xmath157 .
correct digits in the kk and the ekk methods are given for @xmath158 , namely the number of iterations .
the starting energies @xmath159 are taken to be @xmath160 .
the notation c indicates convergence to more than fifteen decimal places .
the exact eigenvalues here are @xmath161 and @xmath162 . [ cols="^,<,<,<",options="header " , ] as a whole the convergence rates are reasonable in the cases with the coupling strength @xmath163 and @xmath164 .
three steps of the calculations in the modified secant method are enough for yielding the results with accuracy to six decimal places . however , the convergence depends strongly on the choice of the starting energies @xmath165 .
when the spacing of two solutions among @xmath43 and @xmath131 is very narrow , we need to draw graphs accurately enough for finding approximate values of @xmath166 to bracket each of the solutions .
we have introduced a new vertex function called the @xmath50-box , which is an operaor defined in a model space and a function of an energy @xmath1 .
the @xmath50-box is constructed in terms of the @xmath2-box and its energy derivative originated by kuo _
et al_. we have proved that a new expression of the effective interaction can be derived by replacing the @xmath2-box by the @xmath50-box in the kk method . with the @xmath50-box
we have also introduced a set of scalar functions @xmath125 .
it has been shown that the @xmath50-box and the associated functions @xmath167 have the following properties : 1 .
the true eigenvalues @xmath47 of the original hamiltonian @xmath3 can be given by the roots of a set of equations @xmath168 .
these equations have the roots , as additional solutions , at the pole energies @xmath131 of the @xmath2-box .
the @xmath50-box and the functions @xmath125 have no singularities even at the poles of the @xmath2-box , and they are well - behaved functions of @xmath1 .
3 . the derivatives of @xmath167 at the solutions for @xmath1 take two values , _
i.e. _ , zero at the true eigenvalues @xmath43 of @xmath3 , and two at the pole energies @xmath131 .
on the basis of the properties ( i ) , ( ii ) , and ( iii ) , we have proposed a new iteration scheme written as @xmath156 .
this iteration always converges and reproduces the true eigenvalues @xmath43 of @xmath3 .
since the energy derivatives of @xmath125 at the solutions for @xmath43 are always zero , this method can be understood to be equivalent to the newton - raphson method used to solve non - linear equations .
the newton - raphson iteration derives quadratic convergence .
therefore , we can expect that this new iteration leads to fast convergence .
we have carried out a test calculation and confirmed the quadratic convergence . as another method for solving a set of equations @xmath168 ,
we have proposed a new non - iterative method which we call the graphical method .
the solutions of these equations can be obtained as the energies at the intersections of the graphs @xmath169 and @xmath127 . using the property ( iii ) we can classify the intersections into two parts , _
i.e. _ , those for the solutions of the true eigenvalues @xmath43 and of the pole energies @xmath131 .
these graphs make us possible to estimate roughly the positions of the roots . with approximate solutions obtained from the graphs we proceed to make more accurate calculations of the solutions , where we have employed a modified secant method which is a combined method of the secant and the newton - raphson methods .
this method has been shown to provide a suitable scheme for obtaining accurate results if we start with approximate energies close to the solutions .
we have made test calculations to assess the graphical method .
we have confirmed numerically that the modified secant method reproduces successfully all the solutions , including the true eigenvalues @xmath43 of @xmath3 and the pole energies @xmath131 .
the theoretical predictions in the property ( iii ) on the energy derivatives of @xmath125 are also verified numerically .
we wish to note that the graphical method , implemented by the modified secant method , yields always convergent results , where we do not need any information on the eigenstates and/or the eigenvalues of @xmath3 such as p - space overlaps and/or energy spacings .
we may conclude that the present approach , the graphical method with the vertex function @xmath0 , would be promising in resolving some of the difficulties in the derivation of the effective interaction .
99 t. t. s. kuo and e. osnes , _ lecture notes in physics _ vol .
364 ( springer - verlag , new york , 1990 ) .
p. j. ellis and e. osnes , rev .
* 49 * , 777 ( 1977 ) .
m. hjorth - jensen , t. t. s. kuo , and e. osnes , phys .
rep . * 261 * , 125 ( 1995 ) .
p. navrtil , j. p. vary , and b. r. barrett , phys .
lett . * 84 * , 5728 ( 2000 ) .
p. navrtil , j. p. vary , and b. r. barrett , phys .
c * 62 * , 054311 ( 2000 ) .
k. suzuki , r. okamoto , and h. kumagai , phys .
c * 36 * , 804 ( 1987 ) . f. andreozzi , phys . rev .
c * 54 * , 684 ( 1996 ) .
r. okamoto , s. fujii , and k. suzuki , int . j. mod
e * 14 * , 21 ( 2005 ) .
k. varga , p. navrtil , j. usukura , and y. suzuki , phys .
b * 63 * , 205308 ( 2001 ) . m. pedersen lohne , g. hagen , m. hjorth - jensen , s. kvaal , and f. pederiva , [ cond-mat.mes-hall ] arxiv:1009.4833 .
j. christensen , c. forssn , s. berg , and s. m. reimann , phys .
a * 79 * , 012707 ( 2009 ) .
e. m. krenciglowa and t. t. s. kuo , nucl .
phys . a * 235 * , 171 ( 1974 )
. t. t. s. kuo , f. krmpoti , k. suzuki , and r. okamoto , nucl .
phys . a * 582 * , 205 ( 1995 ) . s. y. lee and k. suzuki , phys .
lett . b * 91 * , 173 ( 1980 ) .
k. suzuki and s. y. lee , prog .
. phys . * 64 * , 2091 ( 1980 ) .
r. okamoto , k. suzuki , h. kumagai , and s. fujii , to be published in j. phys .
g , [ nucl - th ] arxiv:1011.1994 .
huan dong , t. t. s. kuo , and j. w. holt , [ nucl - th ] arxiv:1011.1487 .
b. h. brandow , rev .
phys . * 39 * , 771 ( 1967 ) .
t. t. s. kuo , s. y. lee , and k. f. ratcliff , nucl .
phys . a * 176 * , 65 ( 1971 ) . c. bloch and j. horowitz , nucl .
* 8 * , 91 ( 1958 ) .
k. suzuki , r. okamoto , p. j. ellis , and t. t. s. kuo , nucl .
phys . a * 567 * , 570 ( 1994 ) . k. takayanagi , private communication . w. h. press , b. p. flannery , s. a. teukolsky , and w. t. vetterling , _ numerical recipes in fortran _ ( cambridge university press , 1992 ) .
t. schucan and h. a. weidenmller , ann . of phys .
* 73 * , 108 ( 1972 ) ; _ ibid . _
* 76 * , 483 ( 1973 ) .
h. m. hoffmann , s. y. lee , j. richert , and h. a. weidenmller , phys .
b * 45 * , 421 ( 1973 ) . | introducing a new vertex function , @xmath0 , of an energy variable @xmath1 , we derive a new equation for the effective interaction .
the equation is obtained by replacing the @xmath2-box in the krenciglowa - kuo ( kk ) method by @xmath0 .
this new approach can be viewed as an extension of the kk method .
we show that this equation can be solved both in iterative and non - iterative ways .
we observe that the iteration procedure with @xmath0 brings about fast convergence compared to the usual kk method .
it is shown that , as in the kk approach , the procedure of calculating the effective interaction can be reduced to determining the true eigenvalues of the original hamiltonian @xmath3 and they can be obtained as the positions of intersections of graphs generated from @xmath0 .
we find that this graphical method yields always precise results and reproduces any of the true eigenvalues of @xmath3 .
the calculation in the present approach can be made regardless of overlaps with the model space and energy differences between unperturbed energies and the eigenvalues of @xmath3 .
we find also that @xmath0 is a well - behaved function of @xmath1 and has no singularity .
these characteristics of the present approach ensure stability in actual calculations and would be helpful to resolve some difficulties due to the presence of poles in the @xmath2-box . performing test calculations ,
we verify numerically theoretical predictions made in the present approach . |
it well known that thin films of high - temperature superconductors exhibit larger critical currents than their single - crystal counterparts .
thin films of the high - temperature superconductor yba@xmath0cu@xmath1o@xmath2 ( ybco ) grown by pulsed laser deposition ( pld ) , which are actively being developed for wire technology , achieve critical currents that are a significant fraction of the maximum depairing current , for example@xcite .
evidence exists that lines of dislocations that run parallel to the crystalline @xmath3 axis in pld - ybco act as correlated pinning centers for vortex lines inside of the superconducting state@xcite , and thereby give rise to such high critical currents .
this is confirmed by the peak observed in the critical current of pld - ybco at orientations of the @xmath3 axis aligned parallel to an applied magnetic field , as well as by the dramatic enhancement of the former peak after more material defects in the form of nano - rods aligned parallel to the @xmath3 axis are added@xcite@xcite .
the microstructure described above for pld - ybco films immediately suggests that the vortex lattice that emerges from the superconducting state in applied magnetic field aligned parallel to the @xmath3 axis is some form of bose glass characterized by a divergent tilt modulus@xcite . in the limit of a rigid vortex lattice , to be expected at large magnetic fields , two - dimensional ( 2d ) collective
pinning of vortex lines by the material line defects then determines the critical current @xcite@xcite@xcite@xcite .
recent theoretical calculations that follow this line of reasoning find moderate quantitative agreement with the critical current measured in films of pld - ybco in @xmath3-axis magnetic fields of a few to many kg , at liquid nitrogen temperature@xcite .
a potential problem with the bose glass hypothesis , however , is that the material line defects in pld - ybco films can meander , or they can be of relatively short length@xcite@xcite .
that question is addressed in this paper , where we find that the bose glass _ breaks up _ into pieces along the direction of the correlated pinning centers when the effective length of the latter is less than the film thickness . in particular ,
the profile of the critical current versus film thickness is predicted to reflect a crossover from two - dimensional to three - dimensional ( 3d ) collective pinning of the vortex lines@xcite .
this cross - over can occur at a length scale that is notably much larger than that expected from point pins of comparable strength@xcite .
we also find that the unbroken bose glass state is robust to the addition of weak point pins .
good agreement is achieved between the dependence on thickness shown by the critical current in certain films of pld - ybco at applied magnetic field@xcite and that predicted for the true bose glass state@xcite .
last , the effective restoring force per vortex line due to a rigid translation of the bose glass about mechanical equilibrium , which is gauged by the labusch parameter@xcite , is found to depend only weakly on applied magnetic field .
this prediction agrees with recent measurements of the microwave surface resistance on pld - ybco films with nano - rod inclusions@xcite .
material line defects in thin enough films of pld - ybco can be considered to be perfectly parallel to the @xmath3-axis .
they notably arrange themselves in a manner that resembles a snapshot of a 2d liquid@xcite , as opposed to a gas@xcite@xcite .
in particular , such correlated pinning centers do not show clusters or voids . a defective vortex lattice that assumes a _ hexatic _
bose glass state will then occur for external magnetic fields aligned in parallel to such a microstructure , in the limit of weak correlated pinning .
it is characterized by parallel lines of edge dislocation defects that are injected into the pristine vortex lattice in order to relieve shear stress due to the correlated pins .
the former do _ not _ show any intrinsic tendency to arrange themselves into grain boundaries , however , due to the absence of clusters and voids in the `` liquid '' arrangement of linear pinning centers .
@xcite and @xcite .
) this results in a vortex lattice whose translational order is destroyed at long range by the isolated lines of edge dislocations , but which retains long - range orientational order .
collective pinning of the dislocation defects by the correlated pins then results both in an elastic response to shear and in a net superfluid density .
theoretical calculations@xcite and numerical monte carlo simulations ( see fig . [ hex - glass ] and caption ) of the corresponding 2d coulomb gas ensemble@xcite confirm this picture .
the correlation length @xmath4 for order along the magnetic field direction is infinite for a bose glass state . in the limit of weak correlated pins , the density of dislocation defects that thread the corresponding vortex lattice
can then be obtained by applying the theory of 2d collective pinning@xcite@xcite@xcite .
each line of unbound edge dislocations is in one - to - one correspondence with a well - ordered bundle of vortex lattice , or larkin domain , that has dimensions @xmath5 in the directions transverse to the magnetic field .
the injection of the dislocation lines into the pristine vortex lattice then results in plastic creep of each larkin domain by a burgers vector@xcite .
the transverse larkin scale is hence obtained by minimizing the sum of the elastic energy cost due to the edge dislocations with the gain in pinning energy due to the translation of a larkin domain by an elementary burgers vector of the triangular vortex lattice , @xmath6 .
this yields the estimate@xcite@xcite@xcite @xmath7 for the density of larkin domains , which coincides with the density of lines of unbound dislocations . here
@xmath8 denotes the density of pinned vortex lines , @xmath9 denotes the maximum pinning force along a material line defect per unit length , and @xmath10 denotes the elastic shear modulus of the pristine vortex lattice .
the prefactor above is of order@xcite @xmath11 , where @xmath12 is the core diameter of a dislocation in the vortex lattice .
consider now the limit of weak pinning centers that do not crowd together : @xmath13 and @xmath14 , respectively , where @xmath15 denotes the density of material line defects , and where @xmath16 denotes their range .
simple considerations of probability then yield the identity @xmath17 between the fraction of occupied pinning centers and the product of the density of vortex lines , @xmath18 , with the effective area of each pinning center . substituting it plus the estimate @xmath19 for the shear modulus of the pristine vortex lattice@xcite into eq .
( [ ratio ] ) then yields the result @xmath20 for the density of larkin domains , which depends only weakly on magnetic field . here , @xmath21 denotes the london penetration depth and @xmath22 is the maximum tension of a flux line in the superconductor .
all of the above is valid in the 2d collective pinning regime that exists at perpendicular magnetic fields beyond the threshold @xmath23 , in which case many vortex lines are pinned by material line defects within a larkin domain of transverse dimensions @xmath5 @xcite .
we will now exploit the boson analogy for vortex matter in order to compute the uniform tilt modulus of the hexatic bose glass in the absence of point pinning centers@xcite .
it amounts to a london model set by the free - energy density @xmath24 for vortex lines located at transverse positions @xmath25 , at a coordinate @xmath26 along the field direction . here
@xmath27 denotes the tension of an isolated vortex line , while the pair potential @xmath28 describes the interaction between vortex lines at the same longitudinal coordinate @xmath26 .
the energy landscape for the correlated pinning centers is described by the potential energy @xmath29 , which is independent of the coordinate @xmath26 along the field direction .
the thermodynamics of this system in the presence of an external tilt stress @xmath30 is then set by the partition function @xmath31 = ( \pi_i\int{\cal d } [ { \bf r}_i(z)]){\rm exp } [ -(k_b t)^{-1}\int_0^{l_z } dz [ g_{\rm b } ( \{{\bf r}\ } , z ) - \int d^2 r\ , { \bf j}_{\rm b}\cdot { \bf a } ] ] , \label{z_b}\ ] ] which under periodic boundary conditions , @xmath32 , is equivalent to a system of 2d bosons . above , @xmath33 ( d { \bf r}_i / d z)$ ] is the current density within the boson analogy .
observe now that the kernel @xmath34 of the uniform electromagnetic response for alternating current ( ac ) is connected to this partition function through the proportionality relationship @xmath35 \propto { \rm exp } \bigl[(k_{\rm b } t)^{-1 } v\sum_{i\omega_n } { 1\over 2 } { \bf a}\cdot{\bf\pi}\cdot{\bf a } \bigr ] \label{pi}\ ] ] in the limit that the corresponding uniform tilt stress vanishes , @xmath36 . above , the matsubara frequencies @xmath37 are given by the allowed wavenumbers along the magnetic field , @xmath38 , and @xmath39 is the volume of the system .
the fact that the tilt stress is given by @xmath30 then yields the identity @xmath40 between the uniform tilt modulus and the uniform ac electromagnetic response of the 2d bose glass . the subscript `` @xmath41 '' above is a tag for the pure shear component of the electromagnetic kernel @xmath42 ( cf .
@xcite ) .
the hexatic bose glass state is clearly a 2d dielectric insulator within the boson analogy .
its electromagnetic response can therefore be modeled by the kernel @xmath43 which is dielectric in the low - frequency limit , and which conserves charge by satisfying the oscillator f - sum rule . above
, @xmath44 is the natural frequency of the electric dipole degrees of freedom in the bose glass .
the latter correspond to the 2d larkin domains in reality , since they represent the smallest units of well - ordered vortex lattice that can respond independently to an applied force .
we then have that the above natural frequency is of order the resonant frequency for transverse sound inside a larkin domain of the 2d bose glass : @xmath45 , where @xmath46 is the effective mass anisotropy parameter equal to the transverse sound speed within the boson analogy . here
@xmath47 is the tilt modulus due to isolated flux lines .
after substitution into eqs .
( [ pi ] ) for the ac response , the identity ( [ identity ] ) then yields a divergent tilt modulus for the bose glass at long wavelength @xmath48 , \label{c_44}\ ] ] with a longitudinal scale @xmath49 that is related to the transverse larkin length by the anisotropic scale transformation @xmath50 .
expression ( [ c_44 ] ) for the uniform tilt modulus of a bose glass is the central result of the paper .
we shall first extract the labusch parameter@xcite from the singular behavior that it shows at long wavelength .
in particular , observe that the elastic energy density for a periodic tilt of the bose glass by a displacement @xmath51 at long wavelength , @xmath52 , acquires a contribution of the form @xmath53 from this divergence , with @xmath54 .
the latter is simply the spring constant per unit volume of the restoring force for a rigid translation of the bose glass state about mechanical equilibrium . using the estimate @xmath55 for the shear modulus of the vortex lattice@xcite yields an effective spring constant per vortex line due to 2d collective pinning limited by plastic creep , @xmath56 , that is given by @xmath57 .
notice that @xmath58 depends only weakly on magnetic field .
the labusch parameter extracted from the microwave surface resistance on pld - ybco films with nano - rod inclusions also exhibits only a weak dependence on external magnetic field aligned parallel to the nano - rods ( or c - axis)@xcite !
the above should be compared to the corresponding labusch parameter due to point pins@xcite , which depends strongly on magnetic field .
indeed , given the conjecture @xmath59 for the labusch parameter due to 3d collective - pinning implies that it decays with increasing magnetic field instead as @xmath60 in such case .
here we have assumed anisotropic scaling , and we have used@xcite @xmath61 .
we shall next use eq .
( [ c_44 ] ) for the uniform tilt modulus to test how robust the hexatic bose glass is to the addition of point pinning centers .
the hexatic bose glass shows long - range orientational order in _ all _ directions ( see fig .
[ hex - glass ] ) .
the addition of point pins will therefore break it up into larkin domains , of dimensions @xmath62 transverse and parallel to the magnetic induction , that tilt in the transverse direction by a distance of order the size of the vortex core , @xmath63 .
such a break - up then has an elastic energy cost per unit volume and a pinning energy gain per unit volume that sum to@xcite@xcite @xmath64 here @xmath65 denotes the shear modulus of the hexatic bose glass , which can be approximated by @xmath10 in the limit of weak correlated pinning , while the corresponding tilt modulus @xmath66 is given by expression ( [ c_44 ] ) evaluated at wavenumber @xmath67 .
also , @xmath68 denotes the density of point pins , while @xmath69 denotes the magnitude of their characteristic force . minimizing @xmath70 with respect to the dimensions of the larkin domains then yields standard results for these@xcite : @xmath71 $ ] and @xmath72^{1/2}$ ] , where @xmath73 and @xmath74 are respectively the longitudinal larkin scale and the transverse larkin scale in the absence of correlated pinning centers .
the first equation is quadratic in terms of the variable @xmath75 , and it has a formal solution @xmath76^{1/2}$ ] .
the hexatic bose glass is therefore robust to the addition of weak point pins .
in particular , the longitudinal larkin scale @xmath77 remains divergent for @xmath78 , which is equivalent to the inequality @xmath79 .
the critical current density of the above hexatic bose glass , which is robust to the addition of point pins , can be obtained by applying 2d collective pinning theory@xcite .
all vortex lines can be considered to be rigid rods .
balancing the lorentz force against the collective pinning force over a larkin domain@xcite@xcite yields the identity @xmath80^{1/2}$ ] for the product of the critical current density along the film , @xmath81 , with the perpendicular magnetic field , @xmath82 , aligned parallel to the material line defects@xcite . here
@xmath8 and @xmath83 denote the density of vortex lines pinned by material line defects and the density of interstitial vortex lines pinned by material point defects , while @xmath9 and @xmath84 are the maximum force exerted on the respective vortex lines per unit length .
again , it is important to observe that the critical current is limited by _
plastic creep _ of the vortex lattice due to slip of the quenched - in lines of edge dislocations along their respective glide planes@xcite .
minimization of the sum of the elastic and pinning energy densities then yields a higher density of larkin domains in the hexatic bose glass with material point defects added by comparison ( [ ratio ] ) : @xmath85 .
this reflects the injection of extra lines of edge dislocations that relieve shear stress caused by point pins .
substitution in turn yields a critical current density , @xmath86 , that has a component due to correlated pins set by the identity @xmath87 and that has a component due to point pins set by the ratio @xmath88 .
the critical current density notably varies as @xmath89 with magnetic field in the limit of weak pinning@xcite .
consider now a film geometry of thickness @xmath90 along the axis of the material line defects .
the forces due to point pins add up statistically along a rigid interstitial vortex line . the effective pinning force per unit length experienced by an interstitial vortex line
is then given by@xcite @xmath91 at film thicknesses @xmath90 that are much greater than the average separation @xmath92 between such pins along the field direction .
again , @xmath69 denotes the maximum force exerted by a point pin .
the relative contribution by point pins to the critical current density is then predicted to show an inverse dependence on film thickness , @xmath93 , that is set by the scale @xmath94 .
we thereby obtain the linear dependence on film thickness @xmath95 for the net critical current per unit width , @xmath96 .
last , comparison of eq .
( [ ratio ] ) with eq .
( [ balance2 ] ) yields the useful expression @xmath97^{1/2 } \label{l*}\ ] ] for the longitudinal scale characteristic of the bose glass as a function of the bulk critical current density , @xmath98 . here
@xmath99 is the average distance between vortex lines and @xmath100 is the depairing current density ( see ref .
@xcite ) . figure [ 2d-3d ] shows a fit to data for the critical current versus thickness obtained from a thin film of pld - ybco at liquid - nitrogen temperature in @xmath101 magnetic field aligned parallel to the @xmath3 axis@xcite .
a bulk critical current density @xmath102 is extracted from it . using a value of @xmath103 for the depairing current ( ref .
@xcite ) , of @xmath104 for the coherence length , of @xmath105 for the mass anisotropy parameter , and setting @xmath106 yields a longitudinal scale @xmath107 . finally , consider again an arrangement of material line defects that are aligned along the @xmath3 axis , that show no voids or clusters in the transverse directions , but that are broken up into relatively long rods of length @xmath108 .
it can be realized by meandering material line defects@xcite@xcite , where @xmath109 is the correlation length for alignment along the @xmath3-axis , or by artificial nano - rod defects@xcite@xcite .
expression ( [ c_44 ] ) for the divergent tilt modulus indicates proximity to the regime of 2d collective pinning , where both @xmath77 and @xmath109 are divergent .
it hence indicates that @xmath77 is also large compared to @xmath110 here .
second , observe that @xmath111 coincides with the longitudinal larkin scale if the rods are considered to be point defects .
the previous ultimately implies the chain of inequalities @xmath112 .
they are consistent with a _ broken _ bose glass state for the vortex lattice , which is threaded by isolated lines of edge dislocations of length @xmath77 along the @xmath3-axis that are connected together by lines of screw - dislocations@xcite along the transverse directions@xcite . as in the limiting case of the true bose glass state ( fig .
[ hex - glass ] ) , the lines of edge dislocations are injected into the pristine vortex lattice in order to relieve shear stress due to the correlated pins .
larkin domains hence are finite volumes@xcite@xcite .
in contrast to their transverse dimensions , however , larkin domains exhibit well defined boundaries along the direction parallel to the correlated pinning centers , across which the vortex lattice slips by a burger s vector due to the presence of the screw dislocations@xcite .
the critical current density expected from this peculiar example of 3d collective pinning therefore coincides with that of a bose glass of thickness @xmath77 : @xmath86 , with a _ bulk _ component due to interstitial vortex lines @xmath113 .
the critical current per unit width then varies with film thickness as @xmath114 , showing no offset .
figure [ 2d-3d ] depicts the predicted dependence on film thickness for the critical current of such a broken bose glass .
it notably exhibits 2d-3d cross - over at film thicknesses in the vicinity of @xmath77 @xcite .
the dependence of the critical current on thickness @xmath90 shown by certain films of pld - ybco is in fact consistent with dimensional cross - over at @xmath115 . a previous attempt by gurevich to account for such behavior by collective pinning of individual vortex lines at point defects yields a longitudinal larkin scale @xmath116 that is too small , however@xcite .
we find here , on the other hand , that pinning due to correlated material defects of length @xmath109 yields a larkin scale @xmath77 that is much longer than that [ @xmath107 , see eq .
( [ l * ] ) ] expected from point pins of comparable strength if @xmath109 is much longer than that scale as well .
indeed , we predict here that the film thickness at which the critical current crosses over from 2d to 3d behavior is of order the effective length of the correlated pinning centers when that length satisfies @xmath117 .
the author thanks leonardo civale , chandan das gupta and sang - il kim for discussions .
this work was supported in part by the us air force office of scientific research under grant no .
fa9550 - 06 - 1 - 0479 .
a. goyal , s. kang , k.j .
leonard , p.m. martin , a.a .
gapud , m. varela , m. paranthaman , a.o .
ijaduola , e.d .
specht , j.r .
thompson , d.k .
christen , s.j .
pennycook and f.a list , supercond .
technol . * 18 * , 1533 ( 2005 ) . | the tilt modulus of a defective abrikosov vortex lattice pinned by material line defects is computed using the boson analogy .
it tends to infinity at long wavelength , which yields a bose glass state that is robust to the addition of weak point - pinning centers , and which implies a restoring force per vortex line for rigid translations about mechanical equilibrium that is independent of magnetic field .
it also indicates that the bose glass state breaks into pieces along the direction of the correlated pinning centers if the latter have finite length . the critical current is predicted to crossover from two dimensional to three dimensional behavior as a function of sample thickness along the correlated pinning centers in such case .
that crossover notably can occur at a film thickness that is much larger than that expected from point pins of comparable strength .
the above is compared to the dependence on thickness shown by the critical current in certain films of high - temperature superconductors currently being developed for wire technology . |
according to the opinion expressed by the most part of the scientific community , the present - day situation of the superconductivity theory can hardly be called satisfactory . indeed ,
while the bcs theory @xcite provides a qualitative explanation of phenomena observed in superconductors , it is unable to produce reliable predictions of their quantitative parameters . at the same time , it is implicitly suggested that the bcs theory contains no contradictions in its description of the conventional superconductivity ; problems are only believed to appear in connection with the description of high - temperature superconductivity .
that is why recent studies mostly concern high - temperature superconductors alone @xcite .
all these studies include the bcs theory considered as an extreme case ; they only differ in mechanisms producing the attraction between electrons .
therefore , these models do not essentially change the microscopic picture of the phenomenon . according to the concept of the so called bcs - bec crossover @xcite ,
the phase transition in the bcs theory shares many features with the bose - einstein condensation ( bec ) phenomenon , which , eventually , allowed connecting it with the ginzburg - landau theory @xcite .
currently , attempts are being made at providing a more precise description of superconductors within the framework of this concept .
these models also incorporate the bcs model for an extreme case .
it should be noted that , in 1980 - 90s , superconductivity theories unrelated to the bcs theory have been put forward .
most of these models contradict experimental data even at the level of the qualitative description . by the present date
, the interest towards such models has seen a substantial decrease .
the superconductor model to be described in the present paper does not contain the bcs model as an extreme case either , and has nothing in common with the bec phenomenon .
nevertheless , the model under development can already claim , at least , a semi - quantitative agreement with the experiment .
its present - day status is described in sections 2 and 3 .
this approach has been discussed in detail in a number of publications @xcite ; therefore , here we restrict ourselves to a shortest possible description , only indicating its initial differences with currently universally acknowledged concepts .
the goal of the present study is a calculation of the superconducting gap in the framework of the proposed model in order to compare it with the analogous calculation results of the bcs theory .
the study is performed on simple metals belonging to pt .
fitting parameters obtained out of the quantitative calculation have the order of magnitude corresponding to that predicted by the theory .
this agreement makes the approach under consideration look more attractive than the bcs theory .
a procedure of renormalization for a system comprised of electrons and phonons correctly taking into account the singular nature of matrix elements of electron - phonon interaction ( epi ) in the long - wave limit has been performed previously @xcite .
the hamiltonian of a monatomic metal electron system , as obtained from the renormalization , has the following form : @xmath3 where @xmath4 @xmath5 @xmath6 @xmath7,\ ] ] @xmath8 and @xmath9 are creation and annihilation operators , respectively , of conductivity band electrons ; @xmath10 and
@xmath11 electron spin indexes ; @xmath12 is the number of conductivity electrons per atom ; @xmath13 is the electron mass ; @xmath14 ; @xmath15 is the ion mass ; @xmath16 is a parameter related to the observed sonic speed of the longitudinal polarization @xmath17 as @xmath18 ; @xmath19 is fermi vector ; @xmath20 ; @xmath21 is the crystal dimension , and @xmath0 is the statical long - wave limit of the metal dielectric function caused by interband transitions .
the above expression for the initial potential of eei substantially differs from other known results , first of all because @xmath22 at @xmath23 .
evidently , this result can not be obtained within the framework of frhlich model @xcite as this model lacks a consistent accounting for screening effects when calculating the epi hamiltonian terms .
this circumstance has long been known @xcite ; now it is worthwhile to note here that it is the frhlich model that makes a basis for eliashberg equations , which the majority of authors @xcite consider to represent the most integral superconductor description in bcs theory .
the expression ( 4 ) obtained for the initial eei excludes the formation of bound two - electron states in the system at this stage of problem examination .
indeed , the coulomb term is significantly larger than @xmath24 , since @xmath25 at @xmath26 .
as a formation of bound two - electron states of the hamiltonian ( 1 ) appears to be impossible , any bosonization of the system is completely out of question .
that is why the models involving bec to clarify the superconductivity phenomenon @xcite seem to be insufficiently substantiated .
let us assume by definition that the hamiltonian @xmath27,defined as : @xmath28 has the same two - electron excitation spectrum as the initial hamiltonian ( 1 ) accounting for multi - particle effects .
thus , in fact , we rigorously define the notion of the eei effective potential @xmath29 involved in the right - hand side of eq .
we start the procedure of definition of @xmath30 by considering the following transformation : @xmath31 where @xmath32 @xmath33 and @xmath34 are expressed in terms of wave functions ( wfs ) ) of stationary two - particle states @xmath35 and @xmath36 in the following way : @xmath37 it is easy to demonstrate that the transformation ( 7 ) reduces the hamiltonian @xmath27 to the diagonal form in the two - particle state space : @xmath38 taking into account multi - particle effects in electron gas is performed in the random phase approximation ( rpa ) as follows .
consider the inverse transformation of ( 7 ) : @xmath39 then express the initial hamiltonian ( 1 ) in terms of operators @xmath40 and @xmath41 using the transformations ( 11 ) . in this form , the hamiltonian
comprises terms of the structure @xmath42 , which , in the rpa , can be reduced to the form of @xmath43 .
these terms define additional potentials @xmath44 , which can be included in the equation for effective potentials @xmath45 : @xmath46 eqs .
( 9 ) and ( 12 ) form a self - consistent system for @xmath47 definition .
_ it is important to note that , in the framework of the proposed approach , it is the hamiltonian ( 10 ) which is related to the initial hamiltonian ( 1 ) by the unitary transformation ( 7 ) , and the introduction of the hamiltonian ( 6 ) into the theory is of an auxiliary nature and is used to calculate the parameters @xmath48 and @xmath49 . _ we should specify more explicitly in which sense we name the transformation ( 7 ) unitary , since the anti - commutators @xmath50 _ + $ ] and @xmath51_+$ ] differ from standard fermion commutation relations in the presence of terms @xmath52 and @xmath53 , respectively @xcite .
it is , however , easy to demonstrate that , in two - electron state space , a transition from @xmath54 states to @xmath55 does not alter the wfs scalar products ; therefore , the transformation ( 7 ) for this case should be considered to be unitary .
then , the right - hand sides of eqs .
( 6,7,11 ) can be considered as two first terms of the series with the number of terms equal to the number of particles in the system .
these series are composed while determining the multi - particle excitation spectrum using the same procedure as invoked above for the calculation of a two - electron excitation spectrum .
of course , the transformation ( 7 ) completed with the terms of the corresponding series is unitary for all states , in which the number of particles is less than or equal to the number of particles in the system . using the rpa for the examination of an excitation spectrum allows for an introduction the hartree - fock method for calculating of one - electron energies into the above calculation routine . from this point of view
, the proposed approach is essentially a generalization of the hartree - fock approximation for a multi - particle spectrum calculation .
previously @xcite , the above mentioned self - consistent system for the potential ( 4 ) have been formulated and solved using the perturbation theory for solution of eigenvalue problems ( 9 ) .
it should also be noted that considering the coulomb eei alone ( see appendix a ) within the framework of the proposed scheme is well consistent with the modern idea of the effective eei in a degenerate plasma .
this fact confirms the consistency of the proposed approach for treatment of multi - particle problems with interaction . combining ( a5 ) with previously obtained results @xcite yields the following equation : @xmath56 where @xmath57 . in this form ,
the equation for @xmath30 is valid in two limits : @xmath26 and @xmath23 .
it ensues from eq .
( 13 ) that an attraction between electrons is present in the long - wave limit .
now according to eliashberg equations , cooper pair formation is mainly due to short - wave phonons with a wave vector @xmath58 .
evidently , the authors of the theory were guided by the erroneous hypothesis of absence of electrons attraction in the long - wave limit , which had been suggested based on the plasma model analysis performed back in 1950 s @xcite .
( 9 ) do not allow the appearance of bound states , then @xmath59 , and the theory describes a normal metal .
appearance of bound states with the binding energy @xmath60 allows to simplify the hamiltonian ( 10 ) . neglecting energy corrections @xmath61 for unbound states of electron pairs
, the hamiltonian ( 10 ) for a superconductor can be written as @xcite : @xmath62 it is easily seen that the ground state @xmath63 of the hamiltonian ( 14 ) coincides with the well - known expression for a non - interacting fermi gas : @xmath64 therefore , from the point of view of the new theory , there is no ground for suggesting a possibility of formation of bec , although the theory does discuss wave functions of bound states of electron pairs .
hence , the ginsburg - landau theory @xcite also lacks microscopic grounds .
nevertheless , we take it as evident that main predictions of the ginsburg - landau theory are reproduced within the framework of the proposed theory ; therefore , the absence of bec in the system in question is not to be worried about . along with the ground state @xmath65 , also the one - electron and one - hole excitations
are the stationary states of hamiltonian ( 14 ) .
this allows to calculate the arising energy gap value @xmath66 in the one - electron spectrum of the ground state .
if there exists only one bound solution with a zero spin and a fixed pair momentum , the value of @xmath67 is calculated as : @xmath68
the expression ( 13 ) is evidently inapplicable for calculation of superconductive properties of metals due to the presence of singularities at @xmath69 and @xmath70 . on the other hand ,
we have not up to now taken into account the hamiltonian terms usually related to scattering processes , which can result in an uncertainty of electron momentum and , consequently , in a removal of singularities even in the initial eei potential @xmath71 .
let us begin with the simplest case of a metal that has been made on the basis of a semiconductor material by implanting electro - active defects into it .
evidently , in this case , the abovementioned momentum uncertainty is @xmath1 , where @xmath2 is the free path of electrons as determined by the scattering on lattice defects . in the case of pure pt metals ,
the following considerations should be taken into account .
the renormalization procedure for metals @xcite used only longitudinal phonons with a distinctly expressed singularity of the epi potential in the long - wave limit . however , due to various reasons , the epi terms remaining in the hamiltonian after this renormalization have a weak non - analyticity in the said limit for acoustic phonons . neglecting relativistic effects these terms
have the following form : @xmath72 where @xmath73 is the phonon frequency ; @xmath74 and @xmath75 are phonon creation and annihilation operators , respectively , including the case of transversal polarization , @xmath13 and @xmath76 are the phonon mode index and phonon quasi - momentum , respectively , and @xmath77 is the reciprocal lattice vector . as to @xmath78 , it is a piecewise regular function with respect to variations of the arguments @xmath79 and @xmath76 .
taking into account the term @xmath80 radically alters the wfs of stationary one - electron states due to the possibility of spontaneous creation of phonons from electrons , since fermi speed is much greater than the sonic speed .
therefore , from the point of view of calculation of the effective eei , the phonon system acts as a reservoir that electrons can exchange momentum with , which can be accounted for by a regularization of eei potential expressions .
let us try to determine the arising momentum uncertainty from the consideration of the temperature dependence of the specific resistance @xmath81 at temperatures @xmath82 , where @xmath83 is debye temperature . then , according to the fermi s golden rule , the following relation is valid for electron - phonon transition frequency @xmath84 of an electron of momentum @xmath79 : @xmath85 which takes into account that in this case , the transitions are induced by phonons with a distribution function approximately equal to @xmath86 . of course , what we are interested in is the spontaneous transition frequency @xmath87 .
it can be expressed as : @xmath88 comparison of eqs . ( 18 ) and ( 19 ) yields : @xmath89 where @xmath90 it is important to note that the following relationships are strictly fulfilled for @xmath91 : @xmath92 where @xmath93 is the maximum frequency of the phonon spectrum . on the other hand
, the hamiltonian ( 17 ) is in certain cases the main cause of energy dissipation in electron tunneling between superconductors .
that is why a correct numerical processing of the current - voltage characteristics ( cvc ) in the region where the tunnel junction ( tj ) behaves itself similarly to resistance yields a more precise evaluation of the parameter @xmath91 than the relationship ( 22 ) .
the @xmath94 value can be roughly evaluated from drude formula for specific resistance @xmath81 at high temperatures : @xmath95 where @xmath96 is plasma frequency determined in optical studies of metals in the infrared ( ir ) range , @xmath97 is the derivative of the specific electric resistance over temperature , for which , evidently , @xmath98 . considering that the free path of a single electron in the system is calculated as @xmath99 , where @xmath100 is an average electron velocity over the fermi surface , the definitive expression for the sought uncertainty @xmath101 of a stationary single - electron state is : @xmath102 in the further study , taking into account that for the metals considered in the present paper @xmath103
, we use the following regularized expression of the eei effective potential ( cf .
( 13 ) ) : @xmath104 this version of regularization allows to obtain an analytical expression for the eei potential in a coordinate representation , which simplifies the calculation procedure .
equation ( 25 ) is suitable for calculating the superconductivity gap . from the technical point of view
, the task can be reduced to determining the energy of bound states of pairs for a potential that is spherically symmetrical in the coordinate representation , whose fourier transform was defined in eq.(25 ) . considering the validity of the relationship @xmath105 , it can easily be concluded that in the coordinate representation the first term of the right - hand side of eq .
( 25 ) , which describes a screened coulomb repulsion of electrons , has a form of a delta - shaped function as compared with the second term .
therefore , its existence can be accounted for virtually without affecting the precision of the calculation by introducing the boundary condition of @xmath106 at @xmath107 into the appearing eigenvalue problem : @xmath108 where @xmath109 and @xmath110 is equal to the second term of the right - hand side of eq .
( 25 ) . having solved eq .
( 26 ) we evaluated the appearing superconductivity gap using eq .
( 16 ) or its generalized form @xmath111 , where @xmath112 is the number of the bound state if eq . (
26 ) had multiple bound solutions , as in the case of pb , where two bound states are present , and the ratio of their binding energies is approximately equal to 0.04 .
let us dwell on the problem of estimation of value @xmath113 , introduced in eq .
( 24 ) . to estimate it
, we use the method similar to that suggested by g. p. motulevich @xcite .
namely , consider the parameter @xmath114 , which relates the plasma frequency value @xmath115 , as observed in ir - range measurement to the corresponding value @xmath116 calculated based on the metal valence : @xmath117 assume that the approximate equation @xmath118 , where @xmath119 is the area of the fermi surface , is valid for the integral @xmath120 over the fermi surface , where @xmath121 is the electron velocity .
then , according to drude formula : @xmath122 where @xmath123 and @xmath124 are fermi velocity and fermi surface area as calculated based on the metal valence , similarly to @xmath116 . on the other hand , considering the ratio of density of states at fermi level over its value calculated for an ideal electron gas of the same concentration , we obtain : @xmath125 where the parameter @xmath126 can be determined based on the metal electronic specific heat data at low temperatures . comparing eqs .
( 29 - 30 ) yields : @xmath127 which provides a solution of the problem as stated .
the only ill - defined parameter of the calculation is the value of @xmath91 , which is only limited by the relationships ( 22 ) .
that is why the proposed version of the calculation used this value as fitting parameter .
in other terms , the parameter @xmath91 used in the definition of @xmath128 was chosen such as to make the resulting width of the superconductivity gap coincide with its experimentally known values . table [ tab : metals ] lists both initial parameters ( @xmath129,@xmath130 , @xmath126 , @xmath0 , @xmath67 ) and the calculated values of the parameter @xmath91 for al , zn , pb , and sn .
an analysis of the ratio @xmath131 , as provided in the same table indicates that fitting parameters @xmath91 have the same order of magnitude as the debye temperature @xmath83 in full accordance with eq.(21 ) . it should be noted here that the initial data of the calculations had a fairly wide variation range : @xmath132 , @xmath133 , @xmath134 , @xmath135 , @xmath136 . in view of this fact
, a physicist would normally consider accidental coincidences out of question , and acknowledge the proposed theory .
this circumstance suggests that a discussion on the approximations introduced within the framework of the approach under development , involving experts in both theoretical physics and mathematics , would be quite useful
. figs .
1 - 4 illustrate the calculation results for the considered metals .
[ cols="^,^,^,^,^,^,^,^,^,^,^",options="header " , ] calculated potential @xmath137 ( lower semiplane ) and wf @xmath138 ( upper semiplane ) for bound state in al .
values of @xmath139 are given in degrees kelvin , @xmath140 is in angstroms and @xmath141 in arbitrary units . ]
calculated potential @xmath137 ( lower semiplane ) and wf @xmath138 ( upper semiplane ) for bound state in zn .
values of @xmath139 are given in degrees kelvin , @xmath140 is in angstroms and @xmath141 in arbitrary units . ] calculated potential @xmath137 ( lower semiplane ) and wfs @xmath138 ( left - hand sides of the curves are in the upper semiplane ) for two bound states in pb .
values of @xmath139 are given in degrees kelvin , @xmath140 is in angstroms and @xmath141 in arbitrary units . ]
calculated potential @xmath137 ( lower semiplane ) and wf @xmath138 ( upper semiplane ) for bound state in sn .
values of @xmath139 are given in degrees kelvin , @xmath140 is in angstroms and @xmath141 in arbitrary units . ]
the following circumstance attracts attention .
the data on the superconductivity gap for the considered metals contradict the natural consequences of the bcs theory . indeed , according to this theory , the higher is the value of @xmath142 the larger should be @xmath67 ; now comparing , _ e.g. _ , the table 1 data for al and zn with those for pb and sn does not confirm this prediction .
in this situation , a study of beryllium , for which @xmath143 k , and the transition temperature @xmath144 k , would be of an incontestable interest . in the framework of the bcs theory ,
such deviations from its predictions are compensated by considering the value of the so - called coulomb pseudopotential @xmath145 , which is used as a fitting parameter @xcite .
this is due to impossibility of a satisfactorily precise evaluation of @xmath145 without a comparison with the data on transition temperature or superconductivity gap . on the other hand ,
as specified above , the value of @xmath91 can be estimated within the framework of the new theory after a correct processing of the cvc of tj ; therefore , a final version of superconductivity gap calculation for pt metals can be totally free of fitting parameters . from this point of view
, the proposed theory appears to be more advantageous than the bcs theory .
we may note as a conclusion that a calculation of the gap value for superconductors produced from semiconductors can be performed without determining the parameter @xmath91 at all .
indeed , assuming that the frequency of electron transitions induced by a scattering on an ionized impurity is substantially higher that the frequency of epi - caused spontaneous transitions , the value of @xmath146 can be estimated , _
e.g. _ , based on hall mobility data . in this case , considering the superconductors obtained from semiconductors could provide the most convincing evidence in favor of our theory .
the author expresses his gratitude to prof .
a. a. rukhadze for his unfailing attention towards the problem and a moral support .
self - consistency equations for the effective eei potential simplified based on the approximate evaluation of integrals within fermi sphere @xcite for this case can be expressed in the form of : @xmath147 where @xmath148 and the following relationships are satisfied : @xmath149 it follows immediately that : @xmath150 where @xmath151 . henceforth , we restrict ourselves to the case @xmath152 , _
i.e. _ that of electron interaction in the vicinity of fermi surface . then , considering that @xmath153 , the equation for @xmath154 can be simplified omitting a series of terms @xmath155 . as a result , it becomes as follows : @xmath156\tilde u_{\bf q}^{{\bf p},0}= \frac{{2\pi e^2 } } { { \varepsilon q^2 } } + \frac{{me^2 } } { { 2\varepsilon \pi ^2 q^3 } } i\left ( q \right)\tilde u_{\bf q}^{0,0 } .\end{aligned}\ ] ] now take into account that the value @xmath157 averaged over the angles between the vectors @xmath79 and @xmath76 is approximately equal to @xmath158 at @xmath23 .
then , simple calculation yields : @xmath159 which , on the whole , agrees with the present - day concept of the eei effective potential , as @xmath160 at @xmath161 .
schrieffer j r , _ theory of superconductivity _ , w.a benjamin , inc . , new york ( 1964 ) .
orenstein j , millis a j , _ science _ , * 288 * , 468 ( 2000 ) .
carlson e w et al . , in _ the physics of conventional and unconventional superconductors _
( eds k.h .
bennemann , j.b .
ketterson ) , berlin : springer - verlag , ( 2002 ) .
qijin chen et al . , _ physics reports _ , * 412 * , 1 ( 2005 ) .
ginzburg v.l . ,
landau l.d . , _ jetp ( rus ) _ , * 20 * , 1064 ( 1950 ) ; landau l.d . , _ collected papers _ , oxford : pergamon press . , 546 , ( 1965 ) .
i.m.yurin , v.b .
kalinin , _ prikl .
( rus ) _ , * 2 * , 12 ( 2003 ) ; i.m.yurin , v.b .
kalinin , _ e - print _ cond - mat/0102407 .
i.m.yurin , v.b .
kalinin , _ prikl . fiz .
( rus ) _ , * 5 * , 5 ( 2004 ) ; i.m.yurin , v.b .
kalinin , _ e - print _ cond - mat/0309516 .
h.frhlich , _
phys.rev._ , * 79 * , 845 ( 1950 ) .
e.g. brovman , yu.m .
kagan , _ usp.fiz.nauk ( rus ) _ , * 112 * , 369 ( 1974 ) . e.g. maksimov , d.yu .
savrasov , s. yu .
savrasov , _ usp.fiz.nauk ( rus ) _ , * 40 * , 337 ( 1997 ) . ogg r.a . (
jr ) , _ phys .
_ , * 69 * , 243 ( 1946 ) ; ogg r.a .
( jr ) , _ phys .
_ , * 70 * , 93 ( 1946 )
. r. schafroth , _ phys .
_ , * 96 * , 1442 ( 1954 ) .
d. bohm , d. pines , _ phys .
_ , * 82 * , 625 ( 1951 ) . g.p .
motulevich , _ usp.fiz.nauk ( rus ) _ , * 97 * , 211 ( 1969 ) .
golovashkin , _ trudy fian ( rus ) _ , * 39 * , 91 ( 1967 ) .
golovashkin , g.p .
motulevich , _ jetp ( rus ) _ , * 57 * , 1054 ( 1969 ) .
golovashkin , g.p .
motulevich , _ kratkie soobch .
po fiz.(rus ) _ , * 2 * , 69 ( 1970 ) . | a previously considered model interpreted a superconductor as an electron gas immersed in a medium with the dielectric constant @xmath0 and a certain elasticity , which could be determined by measured sonic speed in the metal .
the obtained expression of effective electron - electron interaction ( eei ) potential unambiguously implied that , contrary to the suggestions of bcs theory , it is its long - wave limit which is responsible for the emergence of bound two - electron states and , consequently , for gap formation in one - electron spectrum of the metal . however , the existence of singularities in the eei potential expression continued to pose a problem , which did not allow a calculation of the gap value for specific superconducting materials , first of all , for metals belonging to periodic table ( pt ) . in the present work ,
i suggest taking into account matrix elements traditionally attributed to electron scattering in eei effective potential calculations . for superconductors that has been made on the basis of a semiconductor material by implanting electro - active defects into it , this inclusion results in the appearance of an uncertainty of electron momentum @xmath1 , where @xmath2 is the electron free path . when considering pure pt metals , hamiltonian terms relating to creation and annihilaton of phonons should be taken into account , which also produces an uncertainty of electron momentum .
this uncertainty results in a regularization of eei potential expression and , therefore , in a possibility of examination of physical properties of specific superconductors .
results of calculation of superconductivity gap value for a range of simple metals ( al , zn , pb , sn ) confirm the consistency of the developed approach . |
recent imaging studies of the disks of s0 and spiral galaxies have demonstrated that not all stellar disks have simple exponential surface - brightness profiles ( e.g. , erwin , beckman , & pohlen 2005 ; @xcite ; @xcite ; erwin , pohlen , & beckman 2007 ) . instead , disk profiles appear to fall into three general categories : single - exponential @xcite type i ; freeman type ii , with a shallow inner slope and a steeper outer slope ( including so - called `` truncations '' ) ; and type iii ( `` antitruncations '' ) , with a steep inner slope and a shallow outer slope @xcite .
see pohlen et al .
( this volume ) for more background and illustrations of the three types .
we report here on analysis of a deep imaging study of lenticulars and early - type spirals ( hubble types s0sb ) , focused on tracing the outer disk structure using azimuthally averaged surface - brightness profiles .
the analysis of the barred - galaxy subsample ( 66 galaxies ) is complete @xcite ; analysis of the 45 _ unbarred _ galaxies is nearly complete ( see aladro et al . , this volume ) .
pohlen et al .
( this volume ) discuss the general trend of these disk types versus hubble type , including the complementary late - type sample of @xcite ; here , we highlight what we can learn about profiles in the earlier types and their connections to bars and galaxy environment .
for barred galaxies , we can use the size of the bar as a measuring rod .
type ii profiles ( downward - bending broken exponentials ) fall into two fairly distinct categories : those with the break inside the bar ( type ii.i ) and those with the break outside ( type ii.o ) .
the first category is rare ( 6% of the barred galaxies ) , but does match profiles of some @xmath0-body simulations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , so we suspect that this is a side - effect of the bar formation process .
type ii.o profiles are much more common ( 42% of barred galaxies ) , but more difficult to explain .
one possibility is `` truncations '' due to a threshold in star formation .
@xcite recently showed that such thresholds could produce broken - exponential type ii profiles .
however , the threshold ( gas ) surface densities predicted by such models ( e.g. , 310 @xmath1 pc@xmath2 in @xcite 2004 ) are quite low .
while this is plausibly consistent with some breaks in late - type galaxies ( e.g. , * ? ? ?
* ) , the surface brightnesses at the break radius which we observe correspond to stellar mass densities which an order of magnitude or more higher : 202000 @xmath1 pc@xmath2 ( median @xmath3 pc@xmath2 ) .
one potential clue is that we often find outer rings in type ii.o galaxies with the ring located at or very close to the break ( figure [ fig : type2olr - vs - ct ] ) .
since outer rings are well understood as an effect of a bar s outer lindblad resonance ( olr ; see , e.g. * ? ? ?
* ) , we suspect that these breaks are a related phenomenon . supporting evidence comes from recent @xmath0-body simulations by @xcite , who found that type ii.o profiles could form at the olr of a bar - driven spiral .
this leads us to suggest that the breaks in most type ii.o profiles are olr - related , and we refer to these as `` type ii.o - olr '' profiles .
-band image ; for ngc 2950 , we have subtracted a model of the outer disk from the sdss image in order to bring out the ( faint ) outer pseudo - ring . [
fig : type2olr - vs - ct ] ] figure [ fig : histogram ] demonstrates that the break radii in our type ii.o profiles follow the same size distribution ( break radius in terms of bar radius ) as the outer rings in our sample , as well as the outer rings in the sample of @xcite .
a handful of galaxies have breaks at larger radii ( @xmath4 ) corresponding to low stellar mass densities which suggests that the star - formation threshold mechanism might be at work in these galaxies , instead of an olr - related mechanism .
figure [ fig : type2olr - vs - ct ] shows one galaxy where this may be the case : the break is located well _ outside _ the bar s outer ring , and is thus _ not _ connected with the olr .
we label such profiles `` classical truncations '' ( type ii.o - ct ) . ]
note that the ct profiles are rare in the early type galaxies ( 5% of the barred galaxies , 27% of the unbarred galaxies ) ; however , @xcite found that they are more common in late type spirals ( see pohlen et al .
, this volume ) .
a striking contrast emerges when barred and unbarred galaxies are compared : unbarred galaxies are twice as likely to have type iii profiles ( 60% of unbarred galaxies , 32% of barred galaxies ) .
this carries over into the barred galaxies themselves : weakly barred ( sab ) galaxies are more likely to have type iii profiles than strongly barred ( sb ) galaxies .
similarly , the distribution of bar strengths ( deprojected maximum isophotal ellipticity ) is different for type iii and non - type iii profiles : bars in type iii galaxies have a median ( deprojected ) ellipticity of @xmath5 , versus @xmath6 for types i and ii .
type ii profiles ( of all kinds ) are clearly more common in barred galaxies ( 49% of barred galaxies versus 27% of unbarred galaxies ) .
however , they seem to be relatively unaffected by bar _ strength _ : for example , the type ii.o frequency is essentially identical for both sb and sab galaxies , and there is no difference in bar ellipticity between type i and ii.o profiles .
what _ does _ appear to affect the presence of the type ii.o profiles is the _ environment_. specifically , barred galaxies in the virgo cluster have a much lower frequency of type ii.o profiles ( 10% ) than do barred galaxies in the field ( 49% ) . most of the `` missing '' ii.o profiles in the virgo cluster are type i profiles instead .
although the numbers are relatively small , the result is significant at the 99.8% level .
_ why _ this should be so is hard to say , particularly since we do nt yet know why some field galaxies have type i profiles and others have type ii .
one speculative possibility is that type ii.o - olr profiles ( which are the vast majority of the field ii.o profiles ) require several gyr of interaction between the bar and gas in the outer disk .
if the gas is removed rapidly enough , as might be the case for ram - pressure stripping of spirals which fell into the virgo cluster several gyr ago , then the olr break formation process might be choked off before it can proceed to completion .
p.e . was supported by dfg priority program 1177 ; this work was also supported by grants no .
aya2004 - 08251-co2 - 01 from the spanish ministry of education and science and p3/86 of the instituto de astrofisica de canarias . | surface - brightness profiles for early - type ( s0sb ) disks exhibit three main classes ( type i , ii , and iii ) .
type ii profiles are more common in barred galaxies , and most of the time appear to be related to the bar s outer lindblad resonance . roughly half of barred galaxies in the field have type ii profiles , but almost none in the virgo cluster do . a strong _ _ anti__correlation is found between type iii profiles ( `` antitruncations '' ) and bars : type iii profiles are most common when there is no bar , and least common when there is a strong bar . |
the recent discovery of superconductivity with t@xmath2 of 18 k in potassium doped picene ( c@xmath3h@xmath4 ) @xcite has attracted great attention on the aromatic superconductor , a new class of organic superconductor , which was formed by intercalating alkali metal atoms into the interstitial space between aromatic molecules in the molecular crystal .
subsequently , alkali metal doped phenanthrene ( c@xmath5h@xmath6 ) @xcite and dibenzopentacene ( c@xmath7h@xmath8 ) @xcite with superconducting transition temperature of t@xmath2 = 5.6 k and 33 k were reported , respectively . in the three kinds of superconductors ,
the electronic charges are transferred from the doped metal atoms to aromatic molecules and delocalized throughout the molecular crystal .
apart from alkali metal , alkaline earth and rare earth metal are also adopted to synthesize metal - doped phenanthrene crystal to explore their superconductivity . in experiment ,
the high quality ba@xmath9phenanthrene @xcite and la@xmath0phenanthrene @xcite superconductors were obtained with superconducting shielding fractions @xmath10 and @xmath11 , and for la@xmath0phenanthrene , the superconducting transition temperature significantly increases with the pressure and sustains high t@xmath2 in a wide range of pressure @xcite .
the determination of crystal structure of metal doped picene , phenanthrene , and dibenzopentacene is an important prerequisite to understand the electronic behavior and superconducting mechanism in the aromatic superconductors .
however , the detailed crystal structures of doped materials have not yet been reported in experiment , especially the metal atomic position , due to the degradation of sample in air and the limit of measurement techniques @xcite .
several theoretical groups have devoted to the researches on the crystal and electronic properties @xcite . for exmaple , kosugi _ et al .
_ systematically investigated the crystal structure of k - doped picene with different k concentrations , and found that multiple structures exist at the same k concentration ; andres _ et al .
_ fully optimized the geometries of k@xmath12picene and acquired two distinct structures , one with a herringbone structure and the other with a laminar structure . for tripotassium - intercalated phenanthrene
, there also exist the great discrepancy between the optimized structure and the measured one @xcite .
other theoretical researches on the electronic correlation effect and magnetism of k@xmath12picene were also performed and the lattice parameters were fixed at the measured values @xcite in their first principle calculations .
we notice that van der waals interactions have not been taken into account in previous theoretical studies .
the organic compounds of picene , phenanthrene , and dibenzopentacene are the typical molecular crystals , which have low density and hardness because of the light element and relatively long intermolecular bonds . in these materials ,
van der waals interactions existing among molecules play a key role in determining the intermolecular distance and the stacking pattern .
therefore , it is quite necessary to include this factor in simulation of the structural and electronic properties of metal - doped picene , phenanthrene , and dibenzopentacene superconductors . in this paper
, we focus on the structural and electronic properties of rare earth metal la doped phenanthrene , which became a valuable research issue because of the following reasons .
firstly , potassium is an alkali metal element with one electron in the outermost shell . due to the high metal reactivity ,
potassium is usually used to intercalate into graphite , c@xmath13 , phenanthrene , picene , and dibenzopentacene to induce superconductivity .
lanthanum belongs to rare earth metal with an electron configuration [ xe]5d@xmath146s@xmath15 and exhibits two oxidation states , + 3 and + 2 in general .
the metal reactivity of lanthanum is much lower than potassium in the reactivity series of metal . because of such difference
, a large discrepancy relative to k - doped phenanthrene can be expected in la - doped phenanthrene . secondly , the recent report from x. j. chen group indicated that la - doped phenanthrene displays a significant positive pressure effect of t@xmath2 and a sustainable high critical temperature within a wide pressure range , which are unique superconducting features different from k - doped phenanthrene @xcite .
the paper is organized as follows . in sec .
[ method ] the computational method is briefly described . in sec .
[ result ] the results obtained are described and discussed . in sec .
[ summary ] , the conclusions are presented . .
2 @xmath16 1 supercell ( a ) and the interlayer arrangement of molecules is shown in a 1 @xmath16 1 @xmath16 2 supercell ( b ) for pristine phenanthrene .
there are two phenanthrene molecules in a unit cell.,title="fig:",width=321 ]
in our calculations the plane wave basis method was used @xcite .
we adopted the generalized gradient approximation ( gga ) with perdew - burke - ernzerhof ( pbe ) formula @xcite for the exchange - correlation potentials .
the projector augmented - wave method ( paw ) pseudopotentials were used to model the electron - ion interactions .
the local - density approximation ( lda ) is appropriate for largely homogeneous systems , for example , simple metals and semiconductors .
the semilocal - density approximation - gga works well for inhomogeneous systems , for example , transition metals , ionic crystals , compound metals , surfaces , interfaces , and some chemical systems .
however , for sparse systems , including soft matter , van der waals complexes , and biomolecules , they have inter - particle separations , for which nonlocal , long - ranged interactions , such as van der waals ( vdw ) forces , are influential .
the van der waals interaction is considered in the scheme developed by dion , thonhauser _ et al . _
@xcite and enclosed in the first principle calculation code _ quantum - espresso _ @xcite . in the functional the exchange - correlation energy takes the form of @xmath17 where the exchange energy @xmath18 uses the revpbe gga functional and @xmath19 is lda to the correlation energy .
@xmath20 is the nonlocal energy term which accounts approximately for the nonlocal electron correlation effects .
the vdw - df method greatly improves the interaction energies of dispersion bonded systems @xcite , and its application to hard solid can also obtain a systematic improvement in atomization energy and cohesive properties @xcite .
the gaussian broadening technique was employed and a mesh of @xmath21 k - points were sampled for the brillouin - zone integration .
after the full convergency test , the kinetic energy cut - off and the charge density cut - off of the plane wave basis were chosen to be 80 ry and 600 ry , respectively . in our calculations , the lattice parameters were firstly fixed at the experimental values and the internal atomic coordinates were optimized by the energy minimization .
the variable unit cell calculations were performed to compare with the results in the case of fixed lattice constants .
the structure relaxation is performed by broyden - fletcher - goldfarb - shanno ( bfgs ) quasi - newton algorithm @xcite and the maximum and minimum ionic displacement are set to 0.8 and 0.001 bohr .
the convergence thresholds of the total energy and force on atom are 10@xmath22 ry and 10@xmath23 ry / bohr respectively , and for variable cell calculation , the residual stress on cell is less than 0.5 kbar .
pristine phenanthrene crystallizes in the space group @xmath24 with the lattice parameters @xmath25 , @xmath26 , @xmath27 , and @xmath28 measured at room temperature @xcite .
the crystal geometry is shown in fig .
[ geometry - no - la ] , where a unit cell contains two molecules arranged in a herringbone structure . after the lattice parameters and the inner atomic coordinates
are all fully relaxed , the obtained parameters @xmath29 , @xmath30 , @xmath31 , and @xmath32 are 8.401 , 6.166 , 9.445 , and 97.65@xmath33 , which are perfectly consistent with the ones in experiment .
if van der waals interaction is not included in the calculation , the optimized parameters are 8.791 , 6.368 , 9.608 , and 98.66@xmath33 .
the differences between two sets of computational cell parameters demonstrate that van der waals force is an important aspect of physics in phenanthrene and related compounds , and it is necessary to take account of this interaction in the first principles simulation for these materials .
the rare earth metal la atoms are intercalated into phenanthrene solid with the ratio of la atom and molecule 1:1 to form la@xmath0phenanthrene .
la - doped phenanthrene crystallizes in p@xmath34 space group with the lattice parameters a = 8.482 , b = 6.187 , c = 9.512 , and @xmath35 = 97.95@xmath33 @xcite . in the relaxation calculations ,
we have checked most possible configurations to explore the reasonable geometry structure . for la@xmath0phenanthrene ,
the possible arrangements for two la atoms and two molecules in a unit cell are much reduced compared to the situation of k@xmath12phenanthrene with six k atoms and two molecules in a unit cell .
the following four cases were considered and believed to be sufficient for the arrangement of molecules and la atoms , as shown in fig .
[ la - initial - position ] : ( a ) two la atoms are located at two sides of a molecule in a unit cell ; ( b ) two la atoms are located at the same side of a molecule ; ( c ) one la atom sits at the end of molecule ( between two molecular layers ) and the other at the side of molecule ( in the molecular layer ) ; ( d ) two la atoms sit at the molecular ends .
we relaxed the inner atomic positions in the unit cell with cell parameters fixed at the measured ones .
the initial configurations ( a ) , ( c ) and ( d ) would evolve into the same structure , called the structure- , whereas the configuration ( b ) leads to the structure- .
[ geometry-1la ] shows one of the two stable structures , namely the structure- . in a molecular layer of la - doped phenanthrene ,
the aromatic molecules are arranged in a herringbone structure , with an angle about 90@xmath36 between nearest - neighbor molecular planes .
the arrangement of molecules forms a network structure , and la atoms enter into the net hole . for the structure- ,
la atoms are distributed more uniformly , i.e. one la atom sited in each hole . the positions of two la atoms in a unit cell are about @xmath37 and @xmath38 along the @xmath31 axis . as for the symmetry of unit cell
, the structure - i also keeps p@xmath34 group symmetry similar to pristine phenanthrene , namely all atomic positions obey the symmetry operation ( x , y , z ) @xmath39 ( -x , y+0.5 , -z ) .
phenanthrene for the structure- with a la atom inserted in each hole .
la atomic positions are viewed from two different angles in panels ( a ) and ( b ) .
hydrogen atoms are not shown .
, width=321 ] phenanthrene with experimental lattice parameters for the structure- .
( b ) the brillouin zone and selected @xmath40 point paths for energy band .
( c ) total dos per unit cell and orbital - resolved partial dos per c atom and per la atom .
the fermi level is set to zero .
, title="fig:",width=302 ] phenanthrene with experimental lattice parameters for the structure- .
( b ) the brillouin zone and selected @xmath40 point paths for energy band .
( c ) total dos per unit cell and orbital - resolved partial dos per c atom and per la atom .
the fermi level is set to zero .
, title="fig:",width=226 ] phenanthrene with experimental lattice parameters for the structure- .
( b ) the brillouin zone and selected @xmath40 point paths for energy band .
( c ) total dos per unit cell and orbital - resolved partial dos per c atom and per la atom .
the fermi level is set to zero .
, title="fig:",width=302 ] because of intercalation of la atom , the molecular plane of phenanthrene is distorted .
then the interactions among molecules , especially by the bridge of la atom , are enhanced and lead to a stronger band dispersion , as shown in fig .
[ 1la - band - dos ] .
[ 1la - band - dos](a ) clearly shows that there are two bands crossing the fermi level .
the dispersion of bands along @xmath41 and @xmath42 ( parallel to the axis @xmath43 in the brillouin zone ) is small , which is related to the weak interlayer coupling mediated by h - h overlap . in the direction of axis @xmath30 ,
the la atoms and molecules are stacked most closely because of the short axis @xmath30 , which is reflected in the stronger dispersion of bands along @xmath44 , @xmath45 and @xmath46 ( parallel to the axis @xmath47 ) .
compared with the band structure of k@xmath12phenanthrene in the lower panel of fig.2 in andres paper @xcite , the total width of two bands crossing the fermi level is about 0.8 ev for la@xmath0phenanthrene , which is larger than 0.5 ev in the k doped case .
in addition , the two bands below the fermi level locate around -1.0 ev for la@xmath0phenanthrene , lower than -0.5 ev in k@xmath12phenanthrene .
an important feature of the electronic structure of la - doped phenanthrene is revealed in fig .
[ 1la - band - dos](c ) that besides the c @xmath48 states , another important contribution to the dos around the fermi energy is from la @xmath1 electrons with the weight of about 27@xmath49 , which is estimated based on the total dos ( @xmath50 ) and projected dos on la @xmath1 orbitals ( @xmath51 ) .
in order to understand this significant feature , we performed l@xmath52wdin population calculations to discover the charge transfer from la atom to phenanthrene molecule and resulting electron occupations in atomic orbitals .
the calculations indicate that each of two c atoms close to la atom gets about [email protected] electron from the neighbored h and la atoms .
the @xmath54 orbital is empty and @xmath1 orbital has 1.8 electron for la atom . as can be seen from table .
[ lowdin ] , the charge transferred from each la atom to a molecule is about 1.0 electron .
this result has a great difference from the situation of k - doped phenanthrene .
three k atoms are required for each aromatic molecule in k - doped phenanthrene , picene , dibenzopentacene , coronene@xcite , as well as fullerene @xcite , where each molecule can get three electrons from metal atoms .
the k 4s states are far higher than the fermi energy and completely empty in the k - doped system @xcite , while there exists strong hybridization between la 5d states and c 2p states in la@xmath0phenanthrene , which can explain the difference between their band structures .
the contribution of la @xmath1 states to the dos at the fermi level implies that superconductivity in this compound is intimately related to these la @xmath1 electrons .
.[lowdin ] l@xmath52wdin population of la@xmath0 phenanthrene in the structure- .
the number of valence electron in la pseudopotential is 11 and the total valence charge of phenanthrene molecule is 66 .
the spiling charge that can not be projected to the atomic orbital is 0.32 . [
cols="^,^,^,^ " , ] the positive pressure effect on t@xmath2 is a common feature in the doped aromatic superconductors . especially for la - doped phenanthrene ,
application of pressure significantly increases t@xmath2 from 4.9 to 8.5 k as the pressure increases from ambient pressure to 3.2 gpa before the structural phase transition@xcite .
a similar increase of tc from 6.1 to 7.6 k with the pressure from ambient pressure to 1.0 gpa was also found by wang et al.@xcite . to understand the effect of pressure on the superconductivity in la - doped phenanthrene
, we keep the shape of crystal unit cell and decrease the volume to simulate the pressure effect . when the volume changes from @xmath55 to @xmath56 , @xmath57 , and @xmath58 , the corresponding pressure is increased from 2.27 to 3.61 , 5.52 and 8.20 gpa , respectively .
the total dos are displayed in fig .
[ 1la - perhole - press - dos ] as a function of the volume .
the value of dos at fermi level increases gradually from 3.2 to 3.6 , 3.9 , 4.7 states / ev with the volume decreasing . according to the bardeen - cooper - schrieffer ( bcs ) theory of electron - phonon induced superconductivity , the increase of dos with pressure will result in an increase of t@xmath2 , which is in consistent with the experimental measurements @xcite .
for la@xmath0phenanthrene , one unit cell includes two la atoms and two molecules .
when two la atoms enter the same hole in the network formed by phenanthrene molecules , we obtain the structure- , as shown in fig .
[ geometry-2la ] . because two la atoms do not obey the symmetry operation ( x , y , z ) @xmath39(-x , y+0.5 , -z )
, the structure - ii does not keep the p@xmath34 group symmetry and it belongs to p1 space group .
compared with the structure- , the energy is about 0.42 ev lower per unit cell composed of 50 atoms .
hence , the structure- is more stable from the energetic view .
our results show that the stress of the structure - ii with the experimental unit cell is 2.186 gpa , which is comparable to the stress of the structure -i and less than the pressure 3 gpa for tripotassium - intercalated phenanthrene @xcite .
phenanthrene for the structure- with two la atoms inserted in the same hole .
la atom positions are viewed from two different angles in panels ( a ) and ( b ) .
hydrogen atoms are not shown .
, width=321 ] phenanthrene with experimental lattice parameters for the structure- .
( b ) total dos per unit cell and orbital - resolved partial dos per c atom and per la atom .
the fermi level is set to zero .
, title="fig:",width=302 ] phenanthrene with experimental lattice parameters for the structure- .
( b ) total dos per unit cell and orbital - resolved partial dos per c atom and per la atom .
the fermi level is set to zero .
, title="fig:",width=302 ] fig .
[ 2la - dos ] shows the band structure and dos for the structure- .
the bands are more flat and less dispersive relative to the ones of the structure- .
an remarkable feature revealed by our calculations is that it displays a semiconducting state with an energy gap of 0.15 ev .
the lower panel in fig .
[ 2la - dos ] shows the partial dos of la @xmath1 states , which make a great contribution to the total dos around the fermi level , similar to the situation for the structure- .
l@xmath52wdin population calculations show that each of two c atoms close to la atom gets about [email protected] electron from the neighbored h and la atoms , and la @xmath54 orbital is empty and la @xmath1 states have about 2.0 electron for la atom .
the lost charge per la atom is about 1.0 electron . in experiment ,
the effect of charge transfer in la - doped phenanthrene was studied by raman spectroscopy .
based on the similar behavior of phonon softening between la - doped and k - doped phenanthrene , x. f. wang _
_ suggested that three electrons are transferred from a la atom to a phenanthrene molecule@xcite .
obviously , our l@xmath52wdin population analysis for the structure- and structure does not support their deduction .
more experiments are needed for clarifying the actual charge transfer in la - doped phenanthrene .
the semiconcuting structure - ii phase is not the superconducting phase , but it may be a parent compound of superconductor when the concentration of la has few percents deviation from 1 in la@xmath0phenanthrene , similar to the parent compounds of superconductor lacuo@xmath59 @xcite , kfe@xmath9se@xmath60 @xcite and k@xmath61fe@xmath62se@xmath63 @xcite .
the simulation of crystal structure of doped phenanthrene or doped picene is a great challenge , because the optimized crystal structures of doped phenanthrene@xcite or doped picen@xcite reported at present deviate largely from the measurements , even is completely different from the experiment .
the variable cell calculations were also performed for la - doped phenanthrene . for structure- with p2@xmath0 group symmetry ,
the optimized lattice parameters are @xmath648.632 , @xmath65 6.505 , @xmath66 9.661 , and @xmath67108.03@xmath33 . compared with the experimental ones @xmath648.482 , @xmath656.187 , @xmath669.512 , and @xmath6797.95@xmath33 , the corresponding fractional errors are + 1.7@xmath49 , + 5.1@xmath49 , + 1.5@xmath49 , and + 10.3@xmath49 , respectively . for the structure- with p1 group symmetry , the calculated lattice parameters are @xmath64 9.202 , @xmath656.005 , @xmath66 9.829 , @xmath6886.59@xmath33 , @xmath67103.66@xmath33 and @xmath6993.06@xmath33 , and the corresponding fractional errors are + 8.5@xmath49 , -2.9@xmath49 , + 3.3@xmath49 , -3.8@xmath49 , + 5.8@xmath49 and + 3.4@xmath49 , respectively ( @xmath70 = @xmath71 = 90@xmath33 in the experimental cell with p2@xmath0 symmetry ) .
the unit cell volumes for the structure - i and structure - ii are 515.8 @xmath72 and 525.6 @xmath72 , very close to the experimental value 494.3 @xmath72 . with the optimized lattice constants
, we have investigated the electronic structure of la@xmath0phenanthrene and found that their electronic properties are similar to ones with the experimental lattice parameters .
the structure- is a metal and the structure- lies in the semiconducting state with a larger energy gap of 0.30 ev .
still , the dos of la @xmath1 orbital have a substantial contribution to the total dos around the fermi level .
there is very small difference of lattice parameters in experiment between pristine and la - doped phenanthrene .
however , in our optimized structure- and structure- , the errors for some parameters compared to the measurement are larger than 5@xmath49 .
one reason may be that in the compound containing rare earth metal la whose atomic radius is much larger than the ones of c and h atom , it is difficult to deal with the non - bonding interaction between c atom and la atom with available pseudopotentials . in experiment , the metallic phase of la@xmath0phenathrene was observed , which corresponds to the structure- .
now the question arising is why the structure- with lower energy was not observed in measurement ?
this issue could be addressed in the relaxation calculations . as already mentioned before , the initial configurations ( a ) , ( c ) and ( d ) in fig .
[ la - initial - position ] evolve into the structure- , while only the particular configuration ( b ) results in the structure- .
this demonstrates that it is easier to form the metallic phase during the crystal growing process .
in summary , by the first principles calculations , we study the crystal structures and electronic properties of pristine and la - doped phenanthrene under considering van der waals interaction . for la@xmath0phenanthrene ,
two stable atomic geometries are obtained with different la atom distributions in the interstitial space among molecules .
the structure- is a metal and two energy bands cross the fermi level .
the structure- displays the semiconducting state with an energy gap of 0.15 ev . from the energetic viewpoint ,
the structure- is more stable because it has an energy gain of 0.42 ev per unit cell compared to the structure- .
the most striking feature of la@xmath0phenanthrene is that la @xmath1 electrons make a great contribution to the total dos around the fermi level , which is obviously distinct from potassium doped phenanthrene and picene .
the similar results can be obtained when we carried out the related calculations with the optimized lattice parameters , demonstrating that the electronic property of la@xmath0phenanthrene is strongly correlated with the specific positions of la atoms .
our findings provide an important information for understanding superconductivity in la@xmath0phenanthrene . whether the superconductivity in these metal doped aromatic crystals
is driven by electron - phonon or electron - electron mechanism is still in debate .
the value of @xmath73 is special for k@xmath74picene , k@xmath74phenanthrene and k@xmath74dibenzopentacene , indicating that three electrons doped into each molecule seem to be a key factor to induce superconductivity in k - doped aromatic materials . for la@xmath0phenanthrene
, la atom is expected to provide three electrons to each phenanthrene molecule .
however , only about one electron is transferred from la atom to aromatic molecule , as indicated by l@xmath52wdin population .
this is drastically different from usual lanthanide compounds and lanthanide complexes , where la takes the valence state of + 3 and each la atom loses three electrons .
moreover , la @xmath1 electrons have a great contribution to the dos around the fermi level , which might be intimately related to superconductivity , even unconventional superconductivity .
all these make la@xmath0phenanthrene special and rich physical properties can be expected in further research .
we acknowledge the valuable discussion with guo - hua zhong .
this work was supported by most 2011cb922200 , the natural science foundation of china under grants nos.91230203 , 91221103 , 11174072 and u1204108 . _ note added . _ while in the preparation of this manuscript , we learnt of a paper by s. shahab naghavi group ( phys .
b 88 , 115106 ( 2013 ) ) , which reports that the lowest energy state of la@xmath0phenanthrene is band insulating and the best metallic structure is slightly higher in energy but retains the @xmath24 symmetry .
the structure- and structure- in our results are accordant with the best metallic structure and the lowest energy structure shown in fig . 3 in phys .
b 88 , 115106 ( 2013 ) , respectively .
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 link:\doibase 10.1038/nature08859 [ * * , ( ) ] link:\doibase 10.1038/ncomms1513 [ * * , ( ) ] link:\doibase 10.1038/srep00389 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.214523 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/24/34/345701 [ * * , ( ) ] http://arxiv.org/abs/1303.5184 [ ( ) ] , link:\doibase 10.1103/physrevb.84.214506 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.245113 [ * * , ( ) ] link:\doibase 10.1039/c1cp20961b [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.155118 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.134508 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.165102 [ * * , ( ) ] link:\doibase 10.1038/srep00922 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/12/10/103036 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.144501 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.214510 [ * * , ( ) ] http://www.quantum-espresso.org [ * * , ( ) ] link:\doibase 10.1103/physrevlett.77.3865 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.246401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.125112 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.201105 [ * * , ( ) ] link:\doibase 10.1063/1.3193462 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/22/2/022201 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.195131 [ * * , ( ) ] @noop _ _
( , ) @noop _ _ ( , ) link:\doibase 10.1103/revmodphys.81.943 [ * * , ( ) ] link:\doibase 10.1088/0953 - 8984/15/13/202 [ * * , ( ) ] link:\doibase 10.1007/bf01303701 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.087005 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.233205 [ * * , ( ) ] | by the first principles calculations based on the van der waals density functional theory , we study the crystal structures and electronic properties of la - doped phenanthrene .
two stable atomic geometries of la@xmath0phenanthrene are obtained by relaxation of atomic positions from various initial structures .
the structure- is a metal with two energy bands crossing the fermi level , while the structure- displays a semiconducting state with an energy gap of 0.15 ev , which has an energy gain of 0.42 ev per unit cell compared to the structure- .
the most striking feature of la@xmath0phenanthrene is that la @xmath1 electrons make a significant contribution to the total density of state ( dos ) around the fermi level , which is distinct from potassium doped phenanthrene and picene .
our findings provide an important foundation for the understanding of superconductivity in la - doped phenanthrene . |
the discovery by belle @xcite of the very narrow narrow @xmath1 resonance has revived the interest in heavy quarkonium , both experimentally and theoretically . the observation of @xmath1 has been confirmed by cdf @xcite , d0 @xcite and babar collaborations @xcite .
other charmonium - like states , as for example @xmath2 @xcite , @xmath3 @xcite and @xmath4 @xcite have been uncovered in b - factory experiments .
a partial list is shown in table [ exotics ] .
several review papers , as for example @xcite , discuss the difficulty of interpreting these resonances as charmonium states .
seen as exotics , they can possibly be tetraquarks , meson - meson molecules , hybrids , glueballs , etc .
.properties of newly discovered charmonium - like resonances [ cols="^,^,^,^,^ " , ] whatever picture is adopted , it is important to properly determine the charge conjugation number @xmath5 of the resonance . from table
[ exotics ] one can see that @xmath5 is known in nearly all cases . here
we shall discuss the tetraquark option , for example @xcite , and the molecular option , for example @xcite or @xcite , where we noticed that some difficulty has been encountered in defining it .
we recall the definition of tetraquark basis states as given in refs . @xcite .
this basis can provide a direct connection with the diquark - antidiquark states @xcite and also with the molecular states @xcite .
[ fig1 ] in fig .
[ fig1 ] we suppose that the particle 1 is the charmed quark @xmath6 , the particle 2 a light quark @xmath7 , particle 3 the anticharmed quark @xmath8 and particle 4 a light antiquark @xmath9 .
the total wave function of a tetraquark is a linear combination of products of orbital , spin , flavour and colour parts .
we shall successively introduce the orbital , colour and spin parts of the wave function .
the flavor part is fixed .
there are at least three possible ways to define the relative coordinates .
the three relevant possibilities for our problem are shown in fig .
1 . in the cases @xmath10 and @xmath11 the internal coordinates are @xmath12 @xmath13 @xmath14 the first system of coordinates is convenient when the quarks or antiquarks are correlated to form diquarks , as in the diquark - antidiquark model . the coordinates ( 2 ) and ( 3 ) called direct and exchange meson - meson channels , are useful in describing strong decays or introduce a molecular picture @xcite , provided a quark structure is imposed at short separations .
one should use the system which is more appropriate for a given problem .
but in specific calculations one can pass from one coordinate system to the other by orthogonal transformations @xcite . in the colour space
one can construct a colour singlet tetraquark state using intermediate couplings associated to the three coordinate systems defined above . in this way one obtains three distinct bases
@xcite @xmath15 ) the states 3 and @xmath16 are antisymmetric and 6 and @xmath17 are symmetric under interchange of quarks or antiquarks .
this basis is convenient in a diquark - antidiquark picture . in ref .
@xcite only the @xmath18 state has been considered which restricts the spectrum to half of the allowed states @xcite .
the sets ( [ directcolour ] ) or ( [ exchangecolour ] ) , contain a singlet - singlet colour and an octet - octet colour state .
the amplitude of the latter vanishes asymptotically when a tetraquark decays into two mesons .
these are called _ hidden colour _ states , by analogy to states which appear in the nucleon - nucleon problem . when the amplitude of a _ hidden colour _ state is far the most dominant in the wave function , the corresponding open channel acquires , in exchange , a tiny value for its amplitude and in this way the authors of ref .
@xcite explained the small width of @xmath1 , when interpreted as a @xmath19 tetraquark .
as the quarks and antiquarks are spin 1/2 particles , the total spin of a tetraquark can be @xmath20 or 2 .
this can be obtained by first coupling two particles to each other and then couple the two subsystems together .
let us denote the intermediate coupling states between two quarks ( antiquarks ) by @xmath21 for spin 0 ( scalar ) and by @xmath22 for spin 1 ( axial ) .
for a quark - antiquark pair we denote the states by @xmath23 for spin 0 ( pseudoscalar ) and by @xmath24 for spin 1 ( vector ) .
we need their permutation symmetry properties under a given transposition @xmath25 . for quark ( antiquark ) pairs we have @xmath26 and @xmath27 for quark - antiquark pairs we have @xmath28 and @xmath29 for @xmath30 there are two independent basis states for each channel .
the bases associated to ( [ diquark ] ) , ( [ directcolour ] ) and ( [ exchangecolour ] ) are @xmath31 there are three independent states in each channel , to be identified by three distinct young tableaux .
as an example we give the basis for the direct meson - meson channel @xcite @xmath32 the lower index indicates the total spin 1 .
the case @xmath33 is trivial .
there is a single state @xmath34 which is symmetric under any permutation of particles .
we deal with the ground state , _
we have @xmath35 . making the identification 1 = @xmath6 , 2 = @xmath36 , 3 = @xmath37 and 4 = @xmath38 , introduced above , it is convenient to first couple 1 to 3 and 2 to 4 and then the subsystem 13 to 24 , as in fig .
the charge conjugation is equivalent to applying the permutation ( 13)(24 ) to the wave function .
from the properties ( [ transp13 ] ) and ( [ transp24 ] ) one can see that the spin states ( [ directcolour0 ] ) and ( [ spin2 ] ) remain unchanged under the permutation ( 13)(24 ) .
it follows that the states @xmath39 and @xmath40 have charge conjugation @xmath41 . for @xmath42
the situation is slightly more complicated .
there are six linearly independent basis vectors built as products of colour ( [ directcolour ] ) and spin ( [ directcolour1 ] ) states .
@xmath43 , @xmath44 , @xmath45 , @xmath46 change sign thus have charge conjugation @xmath47 . on the other hand @xmath48 and @xmath49
do not change sign , thus have charge conjugation @xmath50 .
also , by construction , they correspond to tetraquarks where the spin of the @xmath51 pair is @xmath52 . @xmath56 one can express @xmath57 in terms of @xmath53 or vice - versa .
in particular , one has @xmath58 @xmath59 @xmath60
@xmath61 @xmath62 @xmath63 from the properties of @xmath53 one can infer that @xmath64 , @xmath65 , @xmath66 and @xmath67 do not have a definite charge conjugation .
contrary , @xmath68 and @xmath69 have @xmath70 , because they are linear combinations of states with @xmath70 .
inverting the above relations one can obtain @xmath53 in terms of @xmath57 .
we have , for example @xmath71 section 3.1 implies that both @xmath72 and @xmath73 have @xmath41 . from the identification 1 = @xmath6 , 2 = @xmath36 , 3 = @xmath37 and 4 = @xmath38 and the definitions ( [ eq : betai ] ) it follows that @xmath74 thus one has @xmath75
@xmath76 where the second component is a _ hidden colour _ state with the same spin structure as @xmath77 .
it follows that a molecular structure with @xmath41 can be obtained only from the difference @xmath72 plus its _ hidden colour _ counterpart . except for this part ,
our result is in agreement with the discussion given in ref .
@xcite , but in disagreement with the molecular interpretation of ref .
the _ hidden colour _ component with @xmath41 can influence the energy of the system at short separations @xcite .
such contributions have been ignored in the simple molecular picture , because the molecules are point - like particles @xcite .
one can also find that @xmath78 which show that @xmath79 and @xmath80 have @xmath70 .
thus the ( open channel ) molecular state with @xmath47 reads @xmath81 again in agreement with ref .
interestingly , its hidden channel counterpart , @xmath80 is completely decoupled from the open channel ( [ open ] ) .
m. b. voloshin and l. b. okun , jetp lett .
* 23 * ( 1976 ) 333 ; a. de rujula , h. georgi and s. l. glashow , phys .
* 38 * ( 1977 ) 317 ; n. a. trnqvist , phys .
* 67 * ( 1992 ) 556 ; z. phys .
* c61 * ( 1994 ) 526 ; phys .
* b590 * , ( 2004 ) 209 and references therein ; t. e. o. ericson and g. karl , phys .
* b309 * ( 1993 ) 426 . | we discuss the charge conjugation quantum number for tetraquarks and meson - meson molecules , both seen as possible interpretations of the newly found @xmath0 charmonium - like resonances .
address = universit de lige , institut de physique b5 , sart tilman , b-4000 lige 1 , belgium + e - mail : [email protected] |
lots of research has been done to control energy utilization in the network layer of wsn .
clustering technique got great attention in resolving energy utilization issue . in clustering technique ,
data is first gathered and then forwarded to base station ( bs ) .
a uniform distribution of nodes and optimum number of chs in each round helps to control load distribution in clustering technique which , ultimately utilizes energy efficiently . as a second step controlling the number of clusters formed during network operation enhances the network stability and lifetime .
a network is said to be stable if difference between first node died time and last node died time is minimum @xcite .
optimum number of chs not only control load distribution but also uses energy efficiently .
one of the problem in clustering technique is the creation of energy holes . in random distribution of nodes
chs which are overloaded cause the creation of energy holes . in multi - hop data forwarding technique , nodes near the bs consume large energy .
these area of nodes are also called hotspots @xcite .
energy depletes quickly in the hotspot areas of network .
hienzelman , a.p .
chandrakasan and h. balakrishnan proposed leach @xcite ; one of the first clustering routing protocol for wsns .
according to leach algorithm , selection of ch for current round is probabilistic .
therefore , in this approach , of ch selection chs formed are not uniformly distributed , which may cause existence of disconnected nodes .
leach - centralized ( leach - c ) is an extension of leach which is proposed by balakrishnan , chandrakasan and heinzelman @xcite .
the plus point of this algorithm is that , bs makes sure that node with less energy does not become ch .
however in large scale network , nodes far away from bs are unable to send their status to bs .
multihop - leach protocol is proposed by nauman israr and irfan awan @xcite .
multihop - leach has mainly two modes of operations , i.e. , multihop inter - cluster operation and multihop - intra cluster operation .
in the leading operation nodes sense the environment and send their data to ch , this data is received by bs through a chain of chs , while the lagging operation is performed in time out period . however , in both modes of operations ch is selected randomly . this agreement does not guaranty full area coverage of the entire network which , it is monitoring .
et al_. @xcite proposed leach - selective cluster ( leach - sc ) .
ch selection in leach - sc is like leach .
the algorithm changes the theme of cluster formation in such a way that node finds ch closest to the mid point between itself and bs , then joins that cluster . however , number of chs fluctuate as rounds proceed .
localization problem is discussed in @xcite and @xcite .
authors divided the network area in to sub areas .
the localization technique helps to improve the coverage hole problem . in this research work
we introduce a new clustering technique of routing layer communication . in density
controlled divide - and - rule ( ddr ) , nodes are distributed uniformly in the network and randomly distributed in different segments of network ; to control the density . in this way coverage hole problem can be avoided . secondly in ddr ,
clusters formed are static and number of clusters remain fix during network operation .
the number of clusters formed are near to optimum number .
this helps in efficient energy utilization and uniform load distribution .
rest of the paper is organised as under .
in this section , we introduce ddr .
first we describe how the network area is logically divided into segments then we find out the energy consumption in these segments .
.notations used in mathematical model [ cols="<,<",options="header " , ] by increasing network area and number of nodes with a ratio of @xmath0 per node ; for 134 nodes we divide network area into four concentric squares , incase of 167 and 200 nodes we divide network area into five and six concentric squares respectively . in such situations communication distance is no more ; less than or equal to reference distance and overhead on chs near to bs increases as a result stability period and network lifetime decreases to an extent .
in this article we focused on energy efficient routing in wsns . our technique , ddr is based on static clustering and optimum number of ch selection in each round . in ddr
we divided the network field into logical segments .
the segmentation process helps to reduce communication distance between node and ch and between ch and bs .
multi - hop communication in inter - cluster further reduces communication distance . in ddr
we have tried to overcome the problem of coverage hole and energy hole through density controlled uniform distribution of nodes in different segments of network .
optimum number of chs in each round helps to achieve balanced load distribution . which enhances stable period and network life time .
00 a. liu , p. zhang , z. chen , theoretical analysis of the lifetime and energy hole in cluster based wireless sensor networks , journal of parallel and distributed computing 71 ( 10 ) ( 2011 ) 13271355 .
w. heinzelman , a. chandrakasan , h. balakrishnan , energy - efficient communication protocol for wireless microsensor networks , in : system sciences , 2000 .
proceedings of the 33rd annual hawaii international conference on , ieee , 2000 , pp . 10pp .
w. heinzelman , application - specific protocol architectures for wireless networks , ph.d .
thesis , massachusetts institute of technology ( 2000 ) .
n. israr , i. awan , multihop clustering algorithm for load balancing in wireless sensor networks , international journal of simulation , systems , science and technology 8 ( 1 ) ( 2007 ) 1325 .
w. jun , z. xin , x. junyuan , m. zhengkun , a distancebased clustering routing protocol in wireless sensor networks , in : communication technology ( icct ) , 2010 12th ieee international conference on , ieee , 2010 , pp .
r. sugihara , r. gupta , sensor localization with deterministic accuracy guarantee , in : infocom , 2011 proceedings ieee , ieee , 2011 , pp .
. m. jin , s. xia , h. wu , x. gu , scalable and fully distributed localization with mere connectivity , in : infocom , 2011 proceedings ieee , ieee , 2011 , pp .
. j. lian , l. chen , k. naik , t. otzu , g. agnew , modeling and enhancing the data capacity of wireless sensor networks , ieee monograph on sensor network operations | cluster based routing technique is most popular routing technique in wireless sensor networks ( wsns ) . due to varying need of wsn applications efficient energy utilization in routing protocols
is still a potential area of research . in this research work
we introduced a new energy efficient cluster based routing technique . in this technique we tried to overcome the problem of coverage hole and energy hole . in our technique
we controlled these problems by introducing density controlled uniform distribution of nodes and fixing optimum number of cluster heads ( chs ) in each round .
finally we verified our technique by experimental results of matlab simulations .
energy , efficient , routing , wsn , static , clustering , hole . |
preparing mechanical harmonic oscillators in their ground states is potentially important for future quantum technologies @xcite , and is presently relevant for experimental work in optomechanics @xcite and nano - electromechanics @xcite .
here we consider two simple and rather different methods for achieving this goal .
the first , called resolved - sideband cooling , is an example of coherent feedback control @xcite in which the mechanical oscillator is coupled linearly to an `` auxiliary '' microwave or optical mode @xcite . since the auxiliary oscillator has a much higher frequency than the mechanics , it is in its ground state at the ambient temperature . because of this the coupling between the two transfers both energy and entropy from the mechanics to the auxiliary , cooling the former .
sideband cooling has already allowed experimentalists to prepare mechanical oscillators in a state with less than one phonon .
the second method we investigate is that in which an explicit continuous measurement is made on the mechanical oscillator ( from now on just `` the oscillator '' ) , and the information from this measurement is used to apply a force to the oscillator to damp its motion in the manner of traditional feedback control @xcite .
our motivation for comparing the performance of resolved - sideband cooling and this measurement - based feedback cooling is to determine how differently the two forms of feedback behave , in the regime of good ground - state cooling , and to understand better the origin of this difference .
two previous works have examined , and to varying extents compared , the two cooling methods we consider here . to explain how our work extends and complements these previous analyses we now summarize them briefly .
the work by genes _
et al . _
@xcite was the first to obtain a complete analytical solution for resolved - sideband cooling .
in addition to presenting this solution they also analyzed a measurement - based feedback protocol for cooling in which the raw signal from a continuous measurement of position is processed by taking its derivative , and a force applied to the oscillator proportional to this processed signal .
nevertheless , genes _ et al .
_ were not able to compare quantitatively the effectiveness of the two methods because they did not have a means to quantitatively compare the resources used by each : sideband cooling employs a unitary coupling to the oscillator , whereas measurement - based feedback employs an irreversible coupling quantified by a damping rate .
hamerly and mabuchi , employing the theory developed in @xcite , made a direct quantitative comparison of the effectiveness of sideband cooling and measurement - based feedback by using the fact that both cooling methods can be realized by coupling the mechanical oscillator to a traveling - wave electromagnetic field ( also known as an _ output channel _
@xcite ) @xcite .
that is , a traveling - wave field can be used to mediate both the continuous measurement used in measurement - based feedback and the unitary coupling of sideband cooling ( coherent feedback ) . because both cooling methods can be implemented using the same coupling , one can ask which method is able to make the best use of the information obtained by the coupling for a given coupling rate . hamerly and
mabuchi ( hm ) also used the optimal estimates of the mean position and momentum in the measurement - based feedback protocol , as we do here , whereas genes _
_ did not . for a weakly - damped ( high - q ) mechanical oscillator , and for a fixed set of parameters ,
hm compared measurement - based cooling to coherent feedback as a function of the bath temperature .
they found that for weak damping , and for a given set of parameters , coherent feedback was able to cool better than the best linear measurement - based feedback .
the results of hamerly and mabuchi were purely numerical .
technologically the most interesting regime for cooling is that in which the mechanical oscillator has a high @xmath0 factor ( weak damping ) , and in which the coupling rate to the controller is strong enough that the control protocol can keep the mechanical oscillator close to its ground state .
here we show that it is possible to obtain simple analytic expressions for the optimal cooling achieved by both control methods to first order in the small parameters that define the regime of high @xmath0 and ground - state cooling . for sideband cooling this
is achieved merely by expanding the full expression for the performance to second order in these parameters . for measurement - based cooling the equations that determine the performance
are non - linear , and can only be solved exactly for zero damping ( @xmath1 ) .
we obtain analytic expressions for weak damping to first - order in the small parameters by using a perturbative method that expands about this exact solution . having analytic expressions for both cooling methods sheds light on the origins of the limits of each , the relationship between these limits , and reveals the dependance on the various key parameters .
the small parameters that define the regime of high @xmath0 and ground - state cooling are as follows .
we define the regime of ground - state cooling , which is also the regime of `` good control '' @xcite , as that in which the control method can maintain the average number of phonons in the oscillator , denoted by @xmath2 , at a value much less than unity ( @xmath3 ) .
if we define the steady - state probability that the system will be found outside the ground state by @xmath4 , then this is also the regime in which @xmath5 .
the rate at which energy flows into the oscillator from the environment is given by @xmath6 where @xmath7 is the damping rate of the oscillator and @xmath8 is the average number of phonons that the oscillator would have if it were at the ambient temperature @xmath9 .
the regime of ground - state cooling requires that the rate at which the control process extracts energy from the oscillator is much greater than @xmath6 . for coherent feedback
this means that the rate of the interaction with the auxiliary , @xmath10 , ( defined precisely below ) , and the damping rate of the auxiliary , @xmath11 , satisfy @xmath12 for measurement - based feedback the regime of ground - state cooling requires that the measurement rate , @xmath13 , ( a scaled version of the measurement strength , defined in section [ physimp ] ) , and the damping rate induced by the feedback force , @xmath14 , satisfy @xmath15 a further requirement for both methods to provide ground - state cooling is that the rate of the linear coupling , @xmath10 , between the oscillator and the auxiliary mode , or the measurement rate @xmath13 , is slower than the frequency @xmath16 of the oscillator .
this stems from the fact that a linear interaction is not the ideal interaction for cooling , and it only works well in the weak - coupling regime .
this requirement is not as strict as the above inequalities , however , since a value of @xmath17 as low as 5 can be sufficient to achieve optimal cooling @xcite . to obtain our simple expressions
we do assume that @xmath18 and expand to second order in the small parameters @xmath19 and @xmath17 . the fact that these parameters need not be very small is indicated by the fact that they do not affect the cooling to first order but only to second order .
further , as part of our analysis we derive results that are exact in @xmath19 and @xmath17 ; it is only in the small parameters @xmath20 and @xmath21 for which our results are necessarily perturbative . given the above time - scale separations , the small parameters that define our regime are @xmath22 for resolved - sideband cooling and @xmath23 for measurement - based cooling .
the size of the ratio @xmath24 is determined by further considerations that we discuss below .
the ratio @xmath21 is second order in the above parameters : @xmath25 we obtain analytic expressions for the steady - state average phonon number in the oscillator , @xmath2 , either to leading or next - to - leading order in the parameters @xmath26 .
resolved - sideband cooling is traditionally implemented by coupling the mechanical oscillator to the auxiliary directly via the linear interaction @xmath27 where @xmath28 , with @xmath29 the oscillator annihilation operator , and @xmath30 with @xmath31 the annihilation operator of the optical or superconducting mode .
the interaction rate @xmath10 is modulated at the frequency difference between the oscillator and the cavity mode , which is what allows them to exchange energy as if they were resonant .
what enables the direct comparison between the two cooling methods is that resolved - sideband cooling can be implemented by coupling the oscillator and auxiliary via a propagating electromagnetic field , and this is also how measurement - based feedback is implemented . in the latter the propagating field is measured using homodyne detection .
thus both cooling methods are able to use the same interface to the resonator , and thus extract information from the resonator in an identical way . for a given rate , @xmath13 ,
at which the propagating field couples to the oscillator , we can then ask which cooling method performs better , and is thus able to make better use of the information .
when the propagating field is measured , the coupling rate @xmath13 becomes the `` measurement strength '' ( defined below ) characterizing the rate at which the measurement extracts information .
when the field is used instead to create the linear coupling @xmath32 with the auxiliary oscillator , the resulting interaction rate @xmath10 realized by the field coupling rate @xmath13 is @xmath33 where @xmath11 is the damping rate of the auxiliary oscillator .
it is useful in our analysis below to define a variable @xmath34 that represents an expression that is first order , and only first order in all of the small parameters @xmath35 that appear in it .
this allows us in what follows to indicate that an expression @xmath36 is order @xmath37 in any ( or all ) of the @xmath35 with the notation @xmath38 . in the next section we present the physical implementation of both cooling methods via an irreversible output coupling . in section [ cohfb ]
we analyze resolved - sideband cooling and derive the expressions for the performance . to do this we use a slightly different approximate master equation to describe the thermal noise of the oscillator than that used by genes _
et al . _
their approximation was valid for all damping rates of the oscillator at high temperature , while ours is valid for weak damping at all temperatures . in section [ measfb ]
we derive the expressions for the performance of the ( optimal ) linear measurement - based feedback cooling . in section [ seccomp ]
we compare and discuss the performance of the two methods , and the origin of their respective limitations . in appendix
a we discuss how the measurement - based cooling scheme can be treated using the quantum noise equations of input - output theory . in appendix
b we show how the steady - state for resolved - sideband cooling is obtained by integrating the spectrum using a remarkable integral formula .
this formula can be used to integrate the spectra of the coordinates for any linear input - output network .
the interface via which both cooling methods interact with the oscillator is shown in fig .
it involves two optical ( or superconducting ) cavities , each with a single mode , where the right - hand end - mirror of each cavity is attached to the oscillator , and thus oscillates with it . to understand how the interface works ,
consider the effect of bouncing a beam of light off the oscillator .
this beam of light provides both an interface to extract information and to apply a force to the oscillator : i ) when the photons in the light beam are reflected from the surface of the oscillator they apply a force to it , and the size of the force is proportional to the beam intensity ; ii ) monitoring the phase of the reflected light provides a continuous measurement of the position of the oscillator .
each of the two optical cavities that are attached to the oscillator provides essentially the same interface as a single beam of light .
the photons in the single mode apply a force to the oscillator as they are reflected off it , and the number of photons in the mode can be adjusted by changing the intensity of the light incident on the cavity ( e.g. the light entering the top cavity via input 1 ) .
the phase of the light that exits each of the cavities provides information about the oscillator position .
the faster the light leaks out of the cavities , characterized by their respective damping rates , the more closely each cavity acts like a beam of light reflected from the oscillator .
while we could use a single optical cavity , with a single mode , to provide both a measurement and a feedback force , we choose to use two cavities because this configuration is required to implement resolved - sideband cooling .
for both cooling methods the top cavity will be used as a `` measurement interface '' to extract information , and the bottom cavity will be used as an `` actuation '' interface to apply a force to the oscillator . to compare the two control methods
, it is the measurement interface , the interface implemented by the top cavity , that we will demand is the same for both methods .
that is , both methods will extract information at the same rate using this interface .
as far as the physical implementation is concerned , this means that laser 1 has the same power , @xmath39 , and the top cavity the same damping rate , @xmath40 , for both methods . to use the top cavity to create an interface that provides continuous information about the position of the oscillator , @xmath41 , we set the cavity damping rate , @xmath40 , to be much larger than the opto - mechanical coupling rate between the cavity mode and the oscillator , and adiabatically eliminate the cavity .
this procedure is detailed in a number of places ( e.g. @xcite ) and we wo nt repeat it here . the resulting interface can be described by writing the electromagnetic field output from the cavity as @xcite @xmath42 where @xmath43 is the input to the cavity .
this is the input - output formalism of collett and gardiner @xcite .
the constant @xmath44 characterizes the rate at which the output channel provides information about the position , and is given by @xmath45 . here
@xmath46 is the steady - state number of photons in the cavity , @xmath47 is the single - photon optomechanical coupling rate @xcite , and @xmath48 is the ground - state position uncertainty of the mechanical oscillator , with @xmath16 the oscillator frequency and @xmath49 its mass . the fact that the interface provides information about position , rather than any other observable , and the fixed information rate @xmath44 are the only limits imposed on our two control protocols .
given this interface , we wish to know which protocol is able to provide the best ground - state cooling , and under what circumstances . here
we will use scaled position and momentum variables for the oscillator , @xmath50 , and @xmath51 , where @xmath52 .
the correspondingly scaled information rate constant is @xmath53 allowing us to write the output field as @xmath54 .
the field operators @xmath55 and @xmath56 are continuum versions of annihilation operators .
the output field @xmath56 has the same correlation functions as the input field , which are @xmath57 while the interface that provides the information will be the same for both control methods , the interface that provides the feedback force will be used differently in each case .
we now describe the two cases in turn .
the configuration that implements measurement - based feedback control is shown in fig .
[ fig2]b . in this case
the output field @xmath56 is measured by homodyne detection that monitors the phase of the output light @xcite , and the second interface ( cavity 2 ) is used merely to apply a classical force to the oscillator . to use cavity 2 to apply a classical force we make the damping rate of this cavity , @xmath58 , sufficiently large that the information rate provided by output 2 in fig .
[ fig1 ] goes to zero @xmath59 . to apply a force to the oscillator we shine a laser into input 2 ( see fig .
[ fig2]b ) and the resulting force on the oscillator in units of @xmath60 is @xmath61 where @xmath62 is the frequency of the optical mode in the cavity .
we therefore apply a time - dependent force by changing the laser power @xmath63 .
while it may appear that we can apply only a positive force , this is illusory .
the equilibrium position of the oscillator is determined by the force @xmath64 . thus applying a constant offset force @xmath65 ,
the force on the oscillator with respect to the resulting equilibrium position is @xmath66 . to peek ahead , the optimal feedback force for the oscillator under linear measurement - based feedback is @xmath67 $ ] , where @xmath14 is chosen to be as large as possible . the homodyne measurement on output 1 results in a stochastic master equation for the density matrix of the mechanical oscillator describing a measurement with strength @xmath13 . along with the thermal noise to which the oscillator is subjected , the full master equation describing the dynamics of the oscillator under the measurement - based feedback is @xcite @xmath68 dt - \tilde{k } \mathcal{k}(\tilde{x})\rho dt \nonumber \\ & & + \sqrt{2 \eta \tilde{k } } \left ( \tilde{x}\rho + \rho \tilde{x } - 2\langle \tilde{x}\rangle \rho \right ) dw , \nonumber \\ & & - \frac{\gamma}{2 } ( n_t + 1 ) \mathcal{k}(b)\rho dt - \frac{\gamma}{2 } n_t \mathcal{k}(b^\dagger)\rho dt , \label{ch8mbfbcx } \end{aligned}\ ] ] where @xmath69 for an arbitrary operator @xmath70 . here
@xmath71 is the efficiency of the measurement on output 1 , and @xmath8 is the average number of phonons in oscillator at the ambient temperature @xmath9 , given by @xmath72 - 1)^{-1}$ ] , where @xmath73 is boltzmann s constant .
the damping rate of the oscillator is @xmath7 . as noted above the feedback force
is a function of the state @xmath74 at time @xmath75 .
resolved - sideband cooling is traditionally implemented using a linear interaction between the mechanical resonator and an auxiliary optical or superconducting resonator , as discussed in the introduction .
we can use the two interfaces provided by cavities 1 and 2 in fig .
[ fig1 ] to reproduce this linear interaction .
this is done by choosing cavity 2 to have the same parameters as cavity 1 , and by applying a @xmath76 phase shift to the light in output 1 ( or alternatively input 2 ) , and by connecting the auxiliary optical resonator to output 1 and input 2 as shown in fig .
[ fig2]a . in this case
it is most convenient to use the quantum langevin equations of the input - output formalism to describe the dynamics of the auxiliary cavity mode and the oscillator .
for the mechanical oscillator these equations are given by @xmath77 with @xmath78 the input noise operators @xmath79 and @xmath80 describe the noise from the thermal bath , @xmath81 describes the field entering through input 1 , and @xmath82 that entering through input 2 .
the correlation functions of the thermal noise operators are given in appendix a. the langevin equations for the auxiliary cavity are @xmath83 where the operators @xmath30 and @xmath84 are the amplitude and phase quadratures of the cavity mode , with @xmath31 the annihilation operator .
the damping rate of the cavity is @xmath11 , and the input noise operators describe the single input .
these operators are @xmath85 and @xmath86 where @xmath87 is a continuum annihilation operator with the same correlation functions as @xmath55 . to connect the auxiliary input to output 1 of the oscillator we simply set @xmath88 where the minus sign accounts for the @xmath76 phase shift shown in fig .
similarly we connect the output of the auxiliary to input 2 by setting @xmath89 substituting eqs.([ch8links ] ) and ( [ ch8links2 ] ) into the equations of motion above for the oscillator and the cavity , the resulting coupled langevin equations for the two systems are @xmath90 with @xmath91 and we have defined @xmath92 .
the only noise driving the mechanical oscillator is now the thermal noise ; the noise coming into input 2 from the auxiliary has cancelled the noise coming in input 1 , which is a result of the @xmath76 phase shift applied to the auxiliary input .
the auxiliary is driven by the noise from input 1 and is damped via the corresponding output channel at rate @xmath11 .
it is this output that takes aways the entropy in the mechanical oscillator , since it is effectively damping to a thermal bath at zero temperature .
the coupling between the two oscillators is given by the last terms on the rhs in the equations for @xmath93 and @xmath94 .
these are the same as would be generated by an interaction hamiltonian @xmath95 .
since the oscillation of the cavity mode continually transforms @xmath96 into @xmath97 , we can replace @xmath96 with @xmath97 in this hamiltonian without affecting the steady - state cooling , and this gives us a hamiltonian equivalent to that in eq.([hint ] ) .
if we now modulate the coupling strength @xmath10 at the frequency difference between the two oscillators , then in the interaction picture the oscillators look as though they are resonant , and the result is the equations of motion for resolved sideband cooling .
these equations are the same as those in eq.([coup1 ] ) , but with @xmath62 replaced with @xmath16 .
the modulation of the coupling strength can be realized by modulating the strength of the effective linear interaction between the mechanics and the transduction oscillators .
alternatively it can be achieved by imprinting a modulation on the fields that couple the auxiliary to the other components .
here we use the average number of phonons in the steady - state , @xmath98 , to measure the degree of cooling . to calculate this quantity for resolved - sideband cooling we solve the quantum langevin equations ( eqs.([coup1 ] ) and ( [ cflang2 ] ) with @xmath62 replaced by @xmath16 ) in frequency space , and this gives us the spectrum of fluctuations of @xmath99 and @xmath100 . integrating these spectra over all frequencies gives us the steady - state expectation values of @xmath101 and @xmath102 , which in turn gives us the average phonon number via @xmath103 we derive the full expressions for @xmath104 and @xmath105 in appendix a. these expressions are rather complex , but simplify greatly if we expand them to second - order in the small parameters given in eq.([sbsmallp ] ) . performing this expansion we find that @xmath106 \label{bdbss } \\ & & - \frac{n_t}{2}\left ( \frac{\gamma}{\kappa } \right)^2 \left [ 1 + \frac{\kappa^2}{\lambda^2 } + \frac{\kappa^4}{\lambda^4 } \right ] + \frac{1}{16}\left ( \frac{\kappa}{\omega } \right)^2 + \frac{1}{8}\left ( \frac{\lambda}{\omega } \right)^2 .
\nonumber \end{aligned}\ ] ] here the first line gives the dominant term , since it is first order while all terms on the second line are second order .
note that @xmath107 is not necessarily a small parameter ; we will show it is of order @xmath108 for optimal cooling . note
also that @xmath21 is second order in the small parameters .
we note first that the dominant term in @xmath2 gives the cooling performance as the ratio between the rate at which energy flows into the oscillator , @xmath109 , to the maximum rate at which it can flow out of the auxiliary ( and thus out of the oscillator ) , being @xmath11 , and tells us that that this rate is achieved when @xmath110 .
this makes sense , and if the dominant term were the only term determining @xmath2 then we would get the best cooling by making @xmath10 as large as possible . but
this is not the case .
the last two terms in @xmath2 show that both @xmath11 and @xmath10 must be much smaller than @xmath16 to achieve ground - state cooling .
this is because the linear @xmath111 coupling between the oscillators only transfers energy efficiently between the two under the rotating - wave approximation , as is well - known @xcite .
the remaining term in @xmath2 merely provides a correction to the dominant term , since it is second order in @xmath112 , as long as @xmath10 is not too much smaller than @xmath11 .
curiously it improves the cooling a little . the parameters @xmath7 and @xmath16 are properties of the oscillator we want to cool , while @xmath10 and @xmath11 are parameters that we would ideally be able to choose as part of designing our controller .
it is therefore natural to ask what values of @xmath10 and @xmath11 will give us the best cooling .
it turns out that we can determine analytically the optimal value of @xmath10 for a given value of @xmath11 because this optimal value falls within the validity of our approximation . to do this we start by discarding the term proportional to @xmath113 , an action that we will justify shortly . differentiating the remaining terms with respect to @xmath10 , we find that the minimal value of @xmath2 is reached when @xmath114 the assumption that we used in our expansion in powers of our small parameters was that @xmath115 , but inspection shows that @xmath116 , and so is lower than first - order by a factor of @xmath117 .
this is not problematic unless higher - order terms in @xmath118 that we have previously dropped ( e.g. third and forth order terms ) now have an order that is sufficiently low as to be near to the order of the leading - order terms from any of the other small parameters , such as @xmath119 , which now has order @xmath120 . in that case
we would have to include these high - order terms to be consistent .
we will check the orders of the relevant terms below .
we note now that with this value for @xmath10 , the term @xmath121 , and so its total contribution to @xmath2 is @xmath122 .
this is why we were justified in discarding it before we performed the minimization to obtain @xmath123 . substituting @xmath123 into eq.([bdbss ] ) , and keeping terms only up to second order in @xmath34 , the minimum average phonon number is @xmath124 the first term is the dominant term , proportional to @xmath34 , the second term is proportional to @xmath125 , and the last two terms are proportional to @xmath126 . the lowest - order term that we have discarded is @xmath127 .
the final step is to minimize @xmath2 over @xmath11 to obtain @xmath128 . in doing this
we might assume that we can first discard the two second - order terms , because the two leading - order terms are sufficient to provide us with a minimum .
however , upon doing this and substituting in the resulting optimal value for @xmath11 , we find that the fourth term contributes to the same order as the first and second .
we therefore discard only the third term . minimizing the remaining terms we obtain @xmath129^{2/3 } , \label{bdbssx2}\ ] ] giving @xmath130 .
we now substitute this value for @xmath11 into the expression for @xmath2 above , and examine the order of the four terms .
the first , second and fourth terms now all have order @xmath120 , and are thus all leading order .
the third term can be discarded as it has order @xmath131 .
we note also that we do not need to include any higher - order terms in @xmath17 or @xmath132 that we have previously discarded : since @xmath2 is a symmetric function of @xmath10 and @xmath11 , the lowest - order terms that we discarded are @xmath133 and @xmath134 , and these are all significantly higher than the leading - order terms in @xmath2 .
the minimal value of @xmath2 is @xmath135 to leading order in our small parameters , with @xmath136 the maximal cooling factor , defined as the ratio of the cooled phonon number , @xmath2 , to the initial phonon number @xmath8 , is @xmath137 the best possible cooling is not the only thing we wish to know . in order to compare with measurement - based cooling we would also like to know the best cooling that can be achieved for a given value of the output coupling rate @xmath13 . to answer this question we substitute @xmath13 in for @xmath10 in eq.([bdbss ] ) , which gives @xmath138 - \frac{1}{2}\left ( \frac{\gamma}{\kappa } \right)^2 \left [ 1 + \frac{\kappa}{8 \tilde{k } } + \frac{\kappa^2}{(8 \tilde{k})^2 } \right ] \right ) \nonumber \\ & & + \frac{\kappa^2}{16\omega^2 } + \frac{\tilde{k } \kappa}{\omega^2 } .
\label{bdbssk}\end{aligned}\ ] ] minimizing this expression exactly with respect to @xmath11 gives a rather complex result , due to the need to solve a quartic equation .
we can nevertheless obtain a simple expression that provides an upper bound on this minimum by choosing @xmath11 so as to minimize the sum of the first and second - to - last terms only .
the resulting value of @xmath11 is @xmath139 so that @xmath140 and @xmath141 .
substituting this into eq.([bdbssk ] ) , and keeping terms up to order @xmath142 , we obtain @xmath143 with @xmath144
we now consider the best cooling that can be obtained by linear measurement - based feedback control , under which the dynamics of the system is described by eq.([ch8mbfbcx ] ) . because the operator being measured is linear in the position and momentum of the oscillator , the oscillator is also linear , and the state of the system is always gaussian , the dynamics is equivalent to that of a linear classical oscillator under a continuous measurement of position , driven by an additional white noise force that simulates exactly the quantum back - action of the measurement @xcite . because of this standard results from classical control theory can be applied to our system .
the classical theory of linear optimal control , referred to as `` linear quadratic gaussian '' ( lqg ) control tells us that if we wish to minimize a weighted sum of a quadratic function of the coordinates and a quadratic function of the control `` inputs '' ( these inputs are the terms in the equations of motion for the momentum that come from the feedback force ) , then we should choose the feedback force to be a linear combination of the means of @xmath41 and @xmath145 given the observer s state of knowledge .
however lqg theory does not apply directly to our problem , since we are interested in minimizing the energy without any particular reference to the control inputs . nevertheless , the coherent feedback protocol we analyzed above is restricted , by the fact that the interaction is linear , to generating only linear dynamics in the system .
it is reasonable therefore , in the interests of a fair comparison , that we also restrict the measurement - based feedback to generating linear dynamics .
this means that the feedback force must be a linear combination of the expectation values of @xmath99 and @xmath100 , and so can be written as @xmath146 for two rate constants @xmath147 and @xmath14 .
the control inputs are then @xmath148 and @xmath149 . in practice , and certainly in experiments today , the amount of force that can be applied induces motion on a timescale much slower than the oscillation of the resonator , and so we restrict ourselves to this regime here .
the only effect of @xmath147 is to modify the frequency of the oscillator as @xmath150 , and since @xmath151 in our regime , this has little effect on the dynamics .
we can therefore drop @xmath147 , leaving us with only one control parameter @xmath14 .
clearly the larger @xmath14 the smaller will be the resulting steady - state energy , @xmath2 . substituting the feedback force @xmath152 into the master equation , eq.([ch8mbfbcx ] ) , the equations of motion for the means and variances of the real variables @xmath41 and @xmath145 are @xmath153 and those for the variances are @xmath154 where @xmath155 is the symmetrized `` covariance '' of @xmath41 and @xmath145 .
it is important to remember that the means and variances in these equations are those of the state - of - knowledge of an observer who has access to the measurement results .
we refer to them as the _ conditional _ means and variances , which is why we denote the means with the subscript `` c '' .
the rate constant @xmath147 in the feedback force merely changes the effective frequency of the oscillator via @xmath156 , and so we have absorbed it into the definition of @xmath16 . in practical situations , certainly those for nano - mechanical resonators ,
the feedback rate constants @xmath147 and @xmath14 are much smaller than @xmath16 .
the result of this is that @xmath147 will in fact have little effect on the cooling , but @xmath14 is very important as we will see below .
since the conditional means of @xmath41 and @xmath145 will be randomly fluctuating , the total variances averaged over all possible trajectories that the system may take while it is being controlled are given by adding the variances of the conditional means of @xmath41 and @xmath145 to the conditional variances .
that is , if we denote the total variances by @xmath157 , @xmath158 , and @xmath159 , and the variances of the conditional means by @xmath160 , @xmath161 , and @xmath162 , then @xmath163 we can derive the equations of motion of the variances of the conditional means by first using ito calculus to derive the differential equations for @xmath164 , @xmath165 , and @xmath166 from eq .
( [ ch8mx1 ] ) . taking averages on both sides of the differential equations for
these square means gives us the differential equations for the second moments of the means . from these
we can obtain the equations of motion for the variances of the means , and these are @xmath167 here we have written the equations in terms of dimensionless ( scaled ) versions of the variances , defined by @xmath168 these scaled variances are those of the scaled variables @xmath99 and @xmath100 . using the scaled variances simplifies the equations , and exposes the important rate constants in the dynamics . from now on any variance with a tilde will indicate the dimensionless version of that variance ( e.g. @xmath169 ) .
the scaled versions of the thermal variances are @xmath170 the harmonic oscillator ground state has @xmath171 . to calculate the total variances in the steady - state we need to determine the steady - states of both the conditional variances and the variances of the means
this can be done by setting the left - hand - sides of the equations of motion to zero , and solving the resulting algebraic equations .
there is a big difference between the differential equations for the conditional variances and those for the variances of the means : we have written the equations for the latter in matrix form because they are linear , whereas the equations for the former are not , and as they stand do not have analytic solutions .
if the harmonic oscillator had no damping , so that @xmath7 were zero , there would be an analytic solution for the steady - states of the conditional variances , being @xmath172 where @xmath173 we see from these solutions that if we want to keep the oscillator close to the ground state , for which @xmath171 , then @xmath174 must be close to unity , and thus @xmath175 ( or equivalently @xmath176 , assuming that @xmath177 ) . while we can not obtain an analytic solution for the steady - states of the conditional variances for all values of @xmath7
, we can obtain an approximate solution valid when @xmath6 is much smaller than @xmath13 and @xmath16 . since @xmath13 is also much smaller than @xmath16
there is more than one way to do this expansion .
we do it by using the solution above for @xmath178 as our zeroth - order solution , and writing @xmath179 and @xmath180 , where @xmath181 is the small parameter , and @xmath182 is unrestricted . we then solve to obtain the steady - states to first - order in @xmath181 .
we are subsequently free to expand the zeroth - order solutions to second order in @xmath183 to obtain solutions to second order in @xmath34 .
the result of the first expansion is @xmath184 now expanding to second order in @xmath185 we obtain @xmath186\ ! \tilde{v}^t
-1 \!\right ) , \nonumber \\
\tilde{c}^{{\mbox{\scriptsize ss } } } & = & \tilde{c}^0 + \left(\frac{\gamma}{2\omega}\right ) \left(1 - \sqrt{\eta } r \right ) \left(\tilde{v}^t - \frac{[1- \eta r^2/8]}{\sqrt{\eta } } \right ) , \nonumber \\
\tilde{v}_p^{{\mbox{\scriptsize ss } } } & = & \tilde{v}_p^0 + ( \tilde{v}_x^{{\mbox{\scriptsize ss } } } -\tilde{v}_x^0 ) + \sqrt{\eta } r ( \tilde{c}^{{\mbox{\scriptsize ss } } } - \tilde{c}^0 ) .
\nonumber\end{aligned}\ ] ] for efficient detection these equations simplify considerably , and we see more clearly the effects of the measurement and thermal noise : @xmath187 , \nonumber \\
\tilde{c}^{{\mbox{\scriptsize ss } } } & = & \tilde{c}^0 + \left(\frac{\gamma}{\omega}\right ) \left[n_t \left(1 - \frac{8\tilde{k}}{\omega } \right ) + \left ( \frac{2\tilde{k}}{\omega } \right)^{\!\ !
2 } \right ] , \nonumber \\
\tilde{v}_p^{{\mbox{\scriptsize ss } } } & = & \tilde{v}_p^0 + ( \tilde{v}_x^{{\mbox{\scriptsize ss } } } -\tilde{v}_x^0 ) + r ( \tilde{c}^{{\mbox{\scriptsize ss } } } - \tilde{c}^0 ) .
\nonumber\end{aligned}\ ] ] since we are expanding to first - order in @xmath188 and second order in @xmath189 , we should drop terms proportional to @xmath190 as they contribute no more than the other third - order terms that have already been dropped .
we have kept these in the above equations merely to show how the second - order terms in @xmath189 affect the solution . calculating the steady - state variances of the means
is straightforward because the equations of motion are linear , although the resulting expressions are rather cumbersome .
we find that these variances will only be small , and thus the oscillator close to the ground state , when @xmath191 .
this makes sense because the noise from the measurement that causes the means to fluctuate is proportional to @xmath13 , and it is the job of the feedback damping at rate @xmath14 to counteract it .
we therefore expand the solutions for the variances of the means in the small parameter @xmath24 .
if we want to allow @xmath14 to be smaller than @xmath16 then we can assume that @xmath192 .
we keep only first - order terms in @xmath24 but we do not drop any terms in @xmath193 , so @xmath14 is not restricted to being small compared to @xmath16 .
the resulting expressions for the variances of the means are @xmath194 , \nonumber \\
\tilde{\mbox{\textbf{\textit{v}}}}_{\bar{p}}^{{\mbox{\scriptsize ss } } } & = & \frac{4\tilde{k}}{\gamma } \left [ ( \tilde{v}_x^{{\mbox{\scriptsize ss}}})^2 + ( \tilde{c}^{{\mbox{\scriptsize ss}}})^2 \right ] , \nonumber \\ \tilde{\mbox{\textbf{\textit{c}}}}^{{\mbox{\scriptsize ss } } } & = & -\frac{4\tilde{k}}{\omega } ( \tilde{v}_x^{{\mbox{\scriptsize ss}}})^2 . \nonumber \end{aligned}\ ] ] we can see from these expressions that we can not achieve good cooling if we make @xmath14 too large .
this is because our feedback force damps only the momentum , and so to confine the position as well as the momentum we need the oscillation of the oscillator to transform position into momentum ( and vice versa ) on a timescale at least as fast as the damping rate @xmath14 .
now we have the steady - state solutions for the conditional variances and the variances of the means , we can combine them to obtain the total variances as per eq.([cvhm2 ] ) .
since @xmath195 , eq.([bbrel ] ) tells us that @xmath196 , which we can use to obtain the mean steady - state phonon number . to second order in @xmath34
the result is @xmath197 the answer depends both on @xmath14 and @xmath13
. it would be nice to find the optimal value of @xmath14 and thus eliminate @xmath14 from the expression .
however this optimal value is @xmath198 , which is unrealistic for practical purposes , and would also imply that the feedback can significantly change the frequency @xmath16 . in this case
we would also be able to increase @xmath16 to reduce @xmath2 further .
so instead of minimizing with respect to @xmath14 , we assume @xmath199 , and thus @xmath200 . keeping only terms up to first - order in @xmath34 we then have @xmath201
finally , we minimize this expression with respect to the measurement strength to determine the best possible cooling .
the optimal measurement strength is @xmath202 and the best cooling is @xmath203 the cooling factor is @xmath204
we begin by examining the expressions for @xmath2 for sideband cooling and measurement - based feedback for a fixed output coupling rate @xmath13 , given respectively by eqs.([bdbssk ] ) and ( [ nbarm ] ) . in writing these equations now , we keep only the terms that make the most important contribution to limiting the cooling .
we have @xmath205 the origins of the various terms in these expressions can be clearly identified .
the terms proportional to @xmath7 give the value of @xmath2 that results from the balance between the rate at which energy is injected into the oscillator from the bath at the rate @xmath206 , and the rate at which it is extracted by the controller .
we can read off the energy extraction rates as @xmath207 the rate for sideband cooling is the series combination of two conductances , which makes intuitive sense . curiously the measurement has the advantage as far as the extraction rate is concerned , but this is because this rate takes into account only the desirable effect of the purification induced by the measurement .
this purification comes at the expense of projection noise , to which we will return below .
the second set of terms , those proportional to @xmath208 are remarkably similar for the two controllers .
the term in the expression for measurement - based feedback comes purely from the squeezing of the conditional momentum variance .
it is due to the fact that the reduction in the position variance due to the measurement of position causes an increase in the momentum variance , which is precisely the back - action noise of the measurement .
this term is not a fundamental restriction of measurement - base cooling , it is present only because we are restricted to a linear interaction with the resonator , and thus a measurement that is linear in the coordinates @xmath41 and @xmath145 . for sideband cooling
the term proportional to @xmath209 is due to the breakdown of the rotating - wave approximation ; when @xmath10 is small compared to the oscillator frequency the @xmath111 interaction acts purely to transfer energy between the two , but this is no longer true as @xmath10 is increased relative to @xmath16 , in which case it generates excitations in both systems .
this limitation on the coherent feedback cooling is not a fundamental one , but is due to the linearity of the interaction .
the term proportional to @xmath210 is due to the fact that the damping of the auxiliary interferes with the energy transfer process .
quantum mechanically this can be attributed to the quantum zeno effect , since the damping is a measurement process that inhibits the unitary dynamics .
since the joint system is linear , and is therefore equivalent to a noisy classical system , there must also be a classical interpretation .
one possibility is that in changing the transfer function of the auxiliary , the damping inhibits the energy transfer in a way that is similar to taking the auxiliary off - resonance with the oscillator .
the only way to avoid this limitation on the cooling appears to be to make the coherent control process time - dependent , rather than using an auxiliary with a constant hamiltonian .
we will return to this topic below .
so far the terms in the expressions for both cooling schemes parallel each other to a large extent .
while they may have somewhat different origins they have very similar forms , and would lead to similar cooling behavior . for example , the heating due to the back - action noise of the measurement is similar to the heating due to the correction to the rotating - wave approximation that appears in sideband cooling , and both are caused by the nature of the interaction .
the final term in the expression for measurement - based cooling is quite different , as it has no parallel in sideband cooling .
it is the heating due to the noise on the mean position and momentum that comes from the random nature of the measurement results , and is a necessary companion to the purification generated by the measurement .
this noise is sometimes referred to as the _ projection noise _ of the measurement .
this noise , being proportional to @xmath13 , is the projection noise of the position measurement when the oscillator is in its ground state .
it is the role of the feedback force to counteract this noise , which is why the resulting heating is proportional to @xmath24 .
this noise is not a fundamental limitation on measurement - based cooling , but is due to the fact that it is the position of the oscillator that is measured .
it is the heating due to the projection noise that makes measurement - based feedback significantly inferior to sideband cooling for cooling an oscillator via a linear interaction .
there are two reasons for this .
the first is that the heating terms in sideband cooling have @xmath16 on the bottom line , whereas the heating due to the projection noise is suppressed only by @xmath14 . in practical situations , and
certainly in current experiments , @xmath14 is considerably smaller than @xmath16 .
the second reason is that the heating term coming from the projection noise is first - order in @xmath13 .
the heating term in sideband cooling that is proportional to @xmath13 also has a factor of @xmath211 .
since @xmath11 and @xmath13 can be expected to be similar for optimal cooling , the heating for sideband cooling is effectively second - order in @xmath19 . because of this , even if we set @xmath212 , and thus replace @xmath213 in eq.([nmbf ] ) with @xmath0 , the maximal cooling for measurement - based feedback scales as @xmath214 , while that for sideband cooling scales as @xmath215 .
what would happen if we were able to apply a classical feedback `` force '' to damp the position as well as the momentum ? in this case the term @xmath216 would no - longer appear in eq.([nbarm ] ) and we would no - longer need @xmath217 . in the limit
in which @xmath218 the projection noise would be eliminated , and the performance of measurement - based cooling would be @xmath219 which is obtained by keeping only the first two terms in eq.([mbfx ] ) .
this is slightly better than that for sideband cooling , but requires quite different interactions and very large feedback forces .
we have found that the factors that place limits on both coherent and measurement - based feedback for cooling an oscillator with a linear interaction are not fundamental restrictions imposed by quantum mechanics , but are due to the linear nature of the interaction . both control methods could perform much better with a non - linear interaction . nevertheless , resolved - sideband cooling is able to make better use of the linear interaction and achieve much better cooling that measurement - based feedback .
we have found that it is not the back - action noise of the measurement which leads to this difference in performance , but the projection noise of the measurement .
finally , there is another important difference between coherent feedback and measurement - based feedback in this linear cooling scenario .
the performance of the coherent scheme can be greatly improved even without a nonlinear coupling , merely by making the interaction rate @xmath10 time - dependent @xcite .
this eliminates the need for the rotating - wave approximation , with the result that the energy in the oscillator can be swapped into the auxiliary within a single period of the oscillator .
the maximal cooling is then @xmath220 measurement - based feedback can not be improved in this way , and instead requires a non - linear interaction .
kj and fs were partially supported by the nsf under project nos .
phy-1005571 and phy-1212413 , and kj was partially supported by the aro muri grant w911nf-11 - 1 - 0268 .
hn and mj were supported by the australian research council , and mj was also supported by the air force office of scientific research ( afosr ) under grant afosr fa2386 - 09 - 1 - 4089 aoard 094089 .
to analyze measurement - based feedback in section [ measfb ] we used the stochastic master equation , while we used the quantum noise formalism of input - output theory to analyze coherent feedback . in the standard formalism of quantum mechanics used by physicists ,
the approach used to derive the former is very different from the analysis that leads to the latter . since the derivation of the quantum langevin equations by collett and gardiner ( cg ) involves approximations , it is not at all clear that they describe the same physical process as the stochastic master equation
. nevertheless , one can show explicitly that the auto - correlation functions of the output fields of the former agree exactly with those of the measurement records of the latter , and this is enough to show equivalence for most applications .
but doing so is not simple ( see for example @xcite ) .
there is another way to formulate measurement theory in quantum mechanics , which uses measure theory in the way that it is used in probability theory .
the resulting structure is called _ quantum probability _ @xcite .
this formulation of measurement theory can be used to construct both the quantum noise formalism and continuous measurement theory , and in this case it is clear by construction that the two descriptions refer to the same process .
this quantum probability formulation of input - output theory was first developed by hudson and parthasarathy ( hp ) @xcite , and exploited for continuous measurement by belavkin @xcite . because the hp formalism contains an explicit mapping between the quantum noise operators and the classical measurement record
the latter being a classical stochastic process it allows us to write a measurement - based feedback process using quantum langevin equations , something that is not possible in the input - output formalism as derived by cg using the standard formulation of quantum mechanics @xcite . to do this for the measurement - based feedback cooling scheme described in section [ physimp ] we first write down the quantum langevin equations for the oscillator , which are given by eq.([ddtmeans1 ] ) .
the output field that our controller measures is @xmath221 the hp formalism now goes beyond the cg formalism by telling us that the white noise quantum field @xmath222 can be interpreted immediately , without any further machinery , as a classical white noise process .
that is , we can fully describe the stream of measurement results from a homodyne detection performed on the field @xmath222 by @xmath222 _
itself_. the reason for this is that , in the quantum probability framework , @xmath222 is a classical noise process ; its quantum nature is captured by the fact that it does not commute with other noise processed that are also contained in the full probability space of events .
the quantum nature of the output field @xmath222 is important : it means that we can not treat _ both _ @xmath222 and @xmath223 as classical noise sources .
this is because when we measure the output field we can not choose to measure both in the @xmath41-basis and @xmath145-basis at the same time .
we could chose a measurement that gave us partial information about both @xmath41 and @xmath145 , but it would produce a stream of measurement results that was neither equal to @xmath222 or @xmath223 .
practically what this means is that we are free to send both @xmath222 and @xmath223 into a quantum system in order to process them , but we can only send one of them through a classical processing device .
the dynamics of all quantum systems will preserve the correct relationship between non - commuting operators , but classical processing will in general not do so because it is less restricted . interpreting @xmath222 now as the classical measurement record
, we can obtain our estimates @xmath224 and @xmath225 from @xmath222 in the usual way by using eqs.([ch8mx1 ] ) and setting @xmath226 to complete the feedback loop we include the feedback force @xmath227 in the quantum langevin equations for the mechanical oscillator , which are then @xmath228 while it may seem odd that the c - number @xmath225 now appears in a differential equation for operators , as usual any c - number merely acts as a multiple of the identity operator .
the measurement - based feedback process is now described by the coupled equations ( [ ch8mx1 ] ) , ( [ ddtcondvx ] ) ( [ ddtcondvp ] ) , and ( [ ddtmeans1x ] ) .
these equations can be compared more easily to the langevin equations describing the coherent feedback protocol than can the sme .
we can further eliminate the output field from eqs.([ch8mx1 ] ) and ( [ ddtmeans1x ] ) by using eq.([x1out ] ) .
the result is a set of langevin equations driven by the input noise operators . as with the langevin equations for sideband cooling
, we can use these to calculate power spectra and correlation functions , thus providing an alternative method for analyzing measurement - based feedback protocols .
examples of the use of the hp input - output theory to describe measurement feedback can be found in @xcite .
steady - states for linear open quantum systems can be obtained by solving the langevin equations in the frequency domain , and then integrating the spectrum over all frequencies .
this integration can be done with an integral formula that can be found in gradshteyn and ryzhik , and which we give below . to begin we recall that the langevin equations for the coupled oscillators , when the interaction is modulated at the frequency @xmath229 , is given by @xmath230 and for compactness we have defined @xmath231 and @xmath232 . to solve the equations of motion for @xmath233 in the frequency domain we take the fourier transform of both sides of the equation . denoting the frequency space variables with a caret , e.g. , @xmath234 the equations of motion become @xmath235 .
rearranging gives @xmath236 the dynamical variables are therefore given by a linear combination of the noise sources , where the coefficients are functions of @xmath237 and therefore filter the noise .
inverting the matrix @xmath238 , for which an algebraic software package is invaluable , we obtain the matrix @xmath239 with @xmath240 and @xmath241 \left [ f(\gamma)^2 + \omega^2 \right ] - \lambda^2 \omega^2 .
\end{aligned}\ ] ] two important properties of the matrix @xmath242 are i ) that each element is a ratio of polynomials , and ii ) that the imaginary unit @xmath243 and the frequency @xmath237 always appear together in eq.([eq132 ] ) .
this second property means that taking the complex conjugate of any element of @xmath242 is the same as replacing @xmath237 with @xmath244 .
the steady - state variance of a dynamical variable is given by integrating the spectrum for that variable over all @xmath237 .
the spectrum for @xmath99 ( for example ) is given by @xmath245 we can obtain the correlation functions for the dynamical variables @xmath246 directly from those of the noise sources : @xmath247 where @xmath248 is the correlation matrix for the noise sources , and is given by @xmath249 the spectrum for @xmath99 is @xmath250 and that for @xmath100 is @xmath251 + ( \lambda^2 \omega)^2 \right\ } } { d(\nu)d(-\nu ) } + \frac{k(\lambda \omega)^2 |f(\gamma)|^2 \left ( |f(k)|^2 + \omega^2 \right)}{d(\nu)d(-\nu ) } \;\;\;\end{aligned}\ ] ] the expressions for the spectra contain @xmath252-order polynomials in the denominator .
if these polynomials had no special structure , they would likely be impossible to integrate analytically .
the fact that this is possible is due to the following remarkable integral formula , which is a slightly simplified version of a formula in gradshteyn and ryzhik @xcite : @xmath253 where @xmath254 must satisfy @xmath255 , @xmath256 and @xmath257 note that @xmath258 and @xmath259 are determinants of matrices that differ only by their first row .
here we need the case @xmath260 , for which the integral is @xmath261 using the above integral formula , and the fact that the steady - state mean squares of @xmath99 and @xmath100 are @xmath262 we obtain @xmath263 here we have defined @xmath264 \\
c & = & ab - \lambda^2 \omega^2 , \end{aligned}\ ] ] with @xmath265 44ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1016/j.physrep.2011.12.004 [ * * , ( ) ] link:\doibase 10.1038/nature05244 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/nphys939 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/nphys1301 [ * * , ( ) ] link:\doibase 10.1038/nature08524 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.123601 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/nature10628 [ * * , ( ) ] link:\doibase 10.1038/nature09898 [ * * , ( ) ] link:\doibase 10.1038/nature10261 [ * * , ( ) ] \doibase doi:10.1038/nature11915 [ * * , ( ) ] @noop * * , ( ) \doibase http://dx.doi.org/10.1016/j.automatica.2009.04.018 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.173602 [ * * , ( ) ] link:\doibase 10.1103/physrevx.3.021013 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.107.177204 [ * * , ( ) ] @noop `` , '' ( ) link:\doibase 10.1103/physrevlett.108.153601 [ * * , ( ) ] @noop _ _ , vol .
( , ) link:\doibase 10.1103/physreva.77.033804 [ * * , ( ) ] link:\doibase 10.1103/physreva.87.013815 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _
( , , ) link:\doibase 10.1103/physreva.30.1386 [ * * , ( ) ] link:\doibase 10.1103/physreva.31.3761 [ * * , ( ) ] link:\doibase 10.1103/physreva.55.3042 [ * * , ( ) ] @noop * * , ( ) @noop _ _
( , , ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevx.4.041029 [ * * , ( ) ] @noop _ _ ( , , ) | we show that in the regime of ground - state cooling , simple expressions can be derived for the performance of resolved - sideband cooling an example of coherent feedback control and optimal linear measurement - based feedback cooling for a harmonic oscillator .
these results are valid to leading order in the small parameters that define this regime .
they provide insight into the origins of the limitations of coherent and measurement - based feedback for linear systems , and the relationship between them .
these limitations are not fundamental bounds imposed by quantum mechanics , but are due to the fact that both cooling methods are restricted to use only a linear interaction with the resonator .
we compare the performance of the two methods on an equal footing that is , for the same interaction strength and confirm that coherent feedback is able to make much better use of the linear interaction than measurement - based feedback .
we find that this performance gap is caused not by the back - action noise of the measurement but by the projection noise .
we also obtain simple expressions for the maximal cooling that can be obtained by both methods in this regime , optimized over the interaction strength . |
the way in which energy is transferred from the time dependent mean field to the individual nucleons is an important ingredient , for instance , in fission processes @xcite and in large amplitude collective motion at low energies . because of the pauli principle it is expected that one body effects , _
i.e. _ , loss of energy due to collision of independent individual nucleons with the mean field , should dominate the dissipation mechanism .
+ several descriptions of these processes involving different approximations are available in the literature @xcite .
these theories are perturbative in character and linear in the collective motion .
therefore they are not suited to address the issues related to nonlinear dynamics and the onset of chaos . on the other hand ,
the integrable or chaotic nature of the motion is of crucial importance to the dissipation mechanism @xcite,_i.e .
_ , the transition from order to chaos provides the possibility for a variety of nuclear responses ( from elastic to elastoplastic to dissipative ) .
therefore detailed studies of simplified models resembling a nucleus may be of interest .
+ planar billiards are perhaps the best systems to model the processes described above in which the nucleus can be imagined as a time - dependent container filled with a gas of non interacting point particles @xcite .
billiard systems have been thoroughly studied in the context of classical and quantum chaos @xcite .
in particular it has been shown that the quantum spectra of generic planar billiards have goe ( gaussian orthogonal ensemble ) characteristics that are observed in the excited spectra of nuclei @xcite .
+ in a seminal paper , hill and wheeler @xcite suggested the landau@xmath0zener ( l - z ) transitions as a mechanism for nuclear dissipation . the mechanism is based on the excitation of the individual nucleons via transitions at avoided level crossings near the fermi surface .
these excitations produce the damping of collective motion describing deformation of the nucleus . in an adiabatic evolution of the collective coordinates ,
the nucleus changes its shape relatively slowly , while the nucleonic levels move up and down in energy .
small deformations in the nuclear shape produce occasionally that two nucleonic levels almost cross each other and experience an avoided crossing . during the whole process many avoided crossings occur with more or less random transitions between nearest neigbours , in such a way that the system may end up in an arbitrary energy state . within this picture
one can imagine a stochastic dynamics in which by simply reversing the temporal evolution one does not recover the initial state .
therefore , the internal degrees of freedom are excited and the motion of the collective coordinates is thus damped . + more recently wilkinson , making good use of the properties of complex spectra , introduced a statistical treatment of dissipation in finite - sized quantum systems in terms of l - z transitions in the context of random matrix theory @xcite .
however , a l - z mechanism as the generator of dissipation in systems with complex spectra , has recently been seriously questioned @xcite . in their works bulgac and
collaborators suggest that the diffusive process in energy is dominated by memory effects and that the picture for dissipation through l - z transitions is likely to be incorrect ( see section [ 3niv ] for a detailed analysis of these arguments ) .
+ obviously the best way to elucidate this question is to solve the quantum mechanical evolution of a generic system , though this is difficult even for planar billiards with moving walls .
+ the goal of the present article is to present a new formulation to solve the quantum mechanical evolution of planar billiards with moving boundaries . using this formulation
we study the time evolution of a specific billiard , the bunimovich stadium with externally driven walls .
we restrict the analysis to slowly varying ( adiabatic ) evolutions @xcite .
the notion of slow motion will be quantified in section [ fr ] .
our aim is to understand whether a l - z mechanism generates the damping of the slow degree of freedom in a system with complex spectra .
+ the work outline is as follows . in the next section
we introduce a one - dimensional new formulation in order to solve the quantum evolution of planar billiards with moving walls .
section [ numres ] is devoted to the numerical results obtained for the bunimovich stadium billiard with goe spectrum characteristics . using these results ,
we evaluate relevant properties for the dissipation process . in section [ 3niv ]
we discuss in detail whether a l - z transition mechanism describes the dissipation of the slow degree of freedom or equivalently the difussion of the fast ones .
the last section is devoted to final remarks and conclusions .
+ before proceeding we want to stress that planar billiards systems externally driven can also be used to model other problems often encountered for example in mesoscopic systems , atomic clusters and of course deformable cavities @xcite .
therefore , the results presented in this work could also contribute to domains other than nuclear physics .
in a recent work vergini and saraceno developed a method to calculate directly _ all _ eigenvalues and eigenfunctions in a narrow energy range of quite general time independent @xmath1 billiards , by solving a generalized eigenvalue problem in terms of quantities defined over the boundary .
the method is based on the use of scaling that enables to write the boundary norm explicitly as a function of the energy @xcite .
+ the aim of this section is to extend the method of scaling to solve the schrdinger equation for @xmath2 billiards with time dependent boundary conditions . + let @xmath3 be a closed curve defining at time @xmath4 a two dimensional domain @xmath5 . we restrict to star shaped domains , this means that @xmath6 is the outgoing normal to @xmath7 .
consider a particle of mass @xmath8 inside the billiard , then the schrdinger equation reads , @xmath9 @xmath10 satisfies the time dependent boundary condition @xmath11 where @xmath12 is a point on @xmath3 , and we consider functions normalized to one on the domain .
a standard procedure is to expand the solution in terms of the adiabatic basis , @xmath13 @xmath14 is the contribution of the dynamical phase with @xmath15 . the adiabatic basis @xmath16 constitutes a complete set of real eigenfunctions of the billiard at each time ; that is
, @xmath17 satisfies the helmholtz equation @xmath18 with dirichlet boundary condition @xmath19 , and it is a continuous function of time . + we generate from the original domain defined by @xmath7 a family of systems that depends on a parameter @xmath20 .
these systems evolve with the curves @xmath21 that are obtained from @xmath7 through a scaling transformation on the plane @xmath22 ( if @xmath12 is a point on @xmath7 , then @xmath23 is the corresponding point on @xmath21 ) .
+ to each function @xmath24 we associate the scaling function @xmath25 .
this family of functions depending on the scaling parameter @xmath20 verifies the helmholtz equation with wave number @xmath26 and satisfies the dirichlet condition over the scaled boundary .
moreover , we require that the mass of the particle in the scaled systems changes as @xmath27 in such a way that @xmath28 is independent of @xmath20 .
the last statement implies that the time evolution is the same for _ all _ the scaled systems .
+ our approach to solve the schrdinger equation is to expand the wave function in terms of the adiabatic basis represented by the scaling functions . after replacing the expansion into the equation @xmath29 we obtain , @xmath30 differentiating this equation with respect to @xmath20 results in , @xmath31 the remainder of the calculus consists on commuting the order of the partial derivation in the rhs of @xmath32 , multiply the equation by @xmath33 and specialize the resulting equation in @xmath34 .
finally we integrate over the boundary of the billiard @xmath7 . after this straightforward calculation
, the final equation reads , @xmath35 where @xmath36 is the length element on the boundary . for the sake of simplicity
we have omitted the argument @xmath37 in the last equation . in @xcite
it was proved that the integral in the lhs of the last equation verifies a quasiorthogonality relation , this means @xmath38 employing this important relation in ( [ final ] ) , we derive the standard system of differential equations in the adiabatic basis @xmath39 with the coefficients @xmath40 computed approximately in terms of quantities defined over the boundary , @xmath41 the exact expression for the coefficients follows from @xmath42 , @xmath43 in the last equation each @xmath44 involves an integration on the domain @xmath5 .
it is also very easy to prove that they are antisymmetric .
+ to compute each @xmath45 via the equation @xmath46 the domain of integration has to be partitioned at least in @xmath47 cells , with @xmath48 equal to the maximum among the wave numbers of the functions @xmath49 in the region of integration , @xmath50 . if @xmath51 is the dimension of the adiabatic basis restricted to the energy region where the evolution will take place , one needs to know @xmath52 coefficients . as a consequence ,
the dimension of the problem of finding the coefficients @xmath45 from equation @xmath46 turns to be of the order of @xmath53 at each time . + as a way to check the goodness of the new formulation we have computed for the specific billiard studied in section [ numres ] , the coefficients @xmath54@xmath55 for a fixed @xmath56 , with @xmath57 @xmath58 calculated exactly ( equation ( [ 2cmunu ] ) ) and using the equation ( [ 1cmunu ] ) ( see fig .
[ coe ] ) .
the correspondance is extremely good over a great number of levels .
the departure between the two plots begins for @xmath59 , but in this region the values of the coefficients are very small . + with the present formulation the cpu time necessary to compute the coefficients is considerably reduced in comparison with the time needed in the standard approach ( equation ( [ 2cmunu ] ) ) . from the preceding remarks , and in order to study the interaction between neighbouring levels in the spectrum , we will calculate the coefficients @xmath40 employing the relation ( [ 1cmunu ] ) .
using the method presented in section [ sec : met ] , we will analize the dynamics of a particle of mass @xmath8 inside a bunimovich stadium billiard with moving boundaries . +
a point particle inside the static stadium billiard is a very well known example of a fully classical chaotic system @xcite .
the particle moves freely on the two dimensional domain and is perfectly reflected from its boundary .
the boundary is formed by two semi - circles of radius @xmath60 connected by two straight lines of length @xmath61 .
[ estadio ] shows a desymmetrized version of the system with area @xmath62 .
+ to study the dynamics , the parameter @xmath63 is changed with a finite velocity @xmath64 in such a way that the total area of the billiard remains unchanged .
we have fixed the area to avoid a drift in the energy spectrum ; this situation is characteristic of nuclear processes where the nucleonic density is approximately constant .
the drift term represents a reversible change in the energy of the system and can be neglected in the analysis of an irreversible dissipation process @xcite .
therefore , the dynamics of the boundary is introduced through the function @xmath65 .
[ spectra ] shows the spectrum of @xmath66 as a function of @xmath67 , @xmath68 .
we have selected the wave numbers @xmath69 between @xmath70 and @xmath71 because in this region a large number of energy levels exists in a narrow portion of the spectrum .
although the properties that we are going to evaluate are characteristic of this region of the spectrum ( @xmath72 ) , as we will show below , a proper scaling can be done in order to evaluate them in other energy regions .
+ the spectrum exhibits the typical behavior of avoided levels crossings that characterizes the energy levels as a function of a parameter for general systems without constants of motion @xcite .
also , we recognize that some avoided crossings are situated on two parallel lines labeled @xmath73 and @xmath74 ( see fig .
[ spectra ] ) .
these lines are associated to bouncing ball states with three and four low excitations respectively .
these states are highly localized in the momenta space @xcite , therefore their interaction with neighbouring states is smaller than the interaction between delocalized generic states .
+ let us analize the coefficients @xmath45 which determine the quantum mechanical evolution of the system ( see section [ sec : met ] ) .
they may be expressed in terms of the deformation parameter @xmath67 , as they satisfy @xmath75 .
[ cmunu ] shows the functions @xmath76 for several pairs of nearest neighbouring levels .
a well defined structure of peaks is observed .
the peaks appear each time two neighbouring energy levels experience an avoided crossing ( it is very easy to follow in the fig .
[ spectra ] a pair of energy levels as a function of the parameter @xmath67 in order to confirm this assertion ) .
the height of the peaks diminishes when the energy gap between levels at the avoided crossing increases . for this reason
the peaks corresponding to interaction with bouncing ball states are one order of magnitude greater than the generic ones .
we label with _ a , b , c .. , _ small peaks that correspond to not well defined avoided crossings or to situations where it is still difficult to decide whether an avoided crossing exists by simple inspection of the spectrum ( see also fig . [ spectra ] ) .
+ for second neighbouring levels , we also find some well defined peaks , they appear essentially when three levels come close to each other ( this situation is discussed in the next section ) .
the heights of the @xmath77 peaks are one order of magnitude smaller than those of the @xmath78 .
+ for coefficients with @xmath79 , we do not observe any simple structure , however the amplitude of these coefficients is indeed very small , lower than five in the scale of fig .
[ coe ] , in comparison with the amplitude observed for nearest neighbouring coefficients .
+ from the present analysis it is clear that the information contained in the coefficients enables a complete definition of the avoided crossings and that this information is not always available in the spectrum .
+ the peaks between first neighbouring levels are very well fitted by lorentzian functions , as it is expected for a l - z transition ( the first part of the next section is devoted to explain the expected lorentzian behavior of the coefficient ) , for almost all peaks with the exception of some small ones .
[ coef ] summarizes the preceding remarks .
it shows on the top the function @xmath80 where @xmath81 is a sum of lorenztian functions centered on the well defined peaks of the coefficient @xmath82 .
each lorentzian function is defined in the next section by the equation @xmath83 .
the widths and the position of the centers are @xmath84 and @xmath85 respectively .
+ the remainder of the fig .
[ coef ] shows the functions @xmath86 @xmath87 as a function of @xmath67 .
the figure reveals the lorentzian behavior of the first neighbouring levels coefficients , and the lack of a defined structure in the coefficients @xmath88 for @xmath79 .
+ the previous numerical study would be still more appealing if we knew how the spectrum scales to other energy regions .
the weyl s law @xcite tells us that the density of states associated to the vertical axis in fig .
[ spectra ] , scales as @xmath48 .
the problem appears with the horizontal axis because the scaling of the density of consecutive avoided crossings @xmath89 is unknown .
+ working in different energy regions and after an exhaustive numerical analysis , we have obtained that @xmath89 scales as @xmath90 with @xmath91 .
[ plot ] shows @xmath89 as a function of @xmath48 for gap sizes less than one quarter of the mean level spacing . in this calculation
we have not considered avoided crossings with bouncing ball states because their relative contribution to the density of states decreases as @xmath92 @xcite .
+ another important fact to stress is that each peak is very well defined ; its width ( given by @xmath84 ) is much smaller than the mean distance between consecutive peaks @xmath93 . only in few cases a small overlap between consecutive avoided crossings
is observed ( see for example the peaks labeled by _
o _ in fig .
[ cmunu ] ) . for generic avoided crossings
we have obtained that @xmath94 ; for avoided crossings with bouncing ball states this product is even smaller as expected .
as we have mentioned in the introduction , our aim is to understand whether l - z transitions govern the mechanism of one body dissipation in systems with complex spectra .
+ in the previous section we have analyzed the coefficients @xmath40 that describe the quantum evolution of a particular system with complex spectrum .
the analysis revealed that the coefficients have a simple structure of well defined lorentzian peaks as the dominant contribution , plus a very small component without any defined structure ( see fig .
[ coef ] ) .
these peaks are concentrated in the first neighbouring levels coefficients , which , as we will show below , is a characteristic of l - z transitions . in few cases
well defined peaks appear in coefficients between second neighboring levels , but these peaks are one order of magnitude smaller than the previous ones .
we will discuss this situation below using an idealized three level system .
+ we begin this section with a brief review on the theory of l - z transitions .
consider a two level system that depends on a parameter @xmath67 , in such a way that for @xmath95 the energy levels experience an avoided crossing .
let @xmath96 be the adiabatic eigenstates and @xmath97 the energy gap between the associated eigenvalues , with @xmath98 and @xmath99 constants .
the adiabatic theorem @xcite tells us that if the system is initially in the state @xmath100 and @xmath67 changes infinitely slowly from @xmath101 to @xmath102 the system will remain in the state @xmath100 .
however , if @xmath67 changes with a finite velocity the final state will be a linear combination of the basis states .
zener derived the probability of an adiabatic transition employing the diabatic basis @xcite for a constant velocity of the parameter @xmath103 .
if at time @xmath104 the system were in the state @xmath100 the transition probability at time @xmath105 is @xmath106 @xcite .
+ using the adiabatic basis and ( [ 2cmunu ] ) is straightforward to derive @xmath107 where @xmath108 is the width of the lorentzian function ; that is , the characteristic time for a l - z process .
+ it has been suggested in recent literature that it is very difficult to characterize the interaction between neighbouring levels in spectra like the present one in terms of l - z transitions @xcite .
the arguments could be summarized as follows : i ) it is not always possible from the spectrum to localize the position of the avoided crossings and to determine the parameters that define the l - z transitions .
this assertion is partially true ; it is not in the spectrum where all the information is contained .
we have solved this problem employing the adiabatic basis , in which the position of the avoided crossing and the interaction length @xmath84 are well defined in terms of the coefficients @xmath40 .
+ ii ) the l - z transition probability is exponentially small when the length of the transition process goes to infinity , but it could be strongly affected for lengths of the order of @xmath84 @xcite . this problem could emerge if the mean distance between avoided crossings @xmath89 is of the order or less than @xmath84 . in the preceding section
we have obtained @xmath109 for generic avoided crossings ( between delocalized eigenfunctions ) and this value is highly reduced for localized eigenfunctions . in a physical system
the eigenfunctions present some degree of localization because the associated classical phase space is not fully chaotic .
therefore , we do not expect correlations between consecutive avoided crossings . in terms of the coefficients @xmath110 ,
the last assertion means that each individual peak is very well defined .
+ iii ) one often encounters a situation where three levels come close to each other and by simple inspection one can think in a three level crossing ( see for example points @xmath111 and @xmath112 in the fig .
[ spectra ] ) . in order to understand this process
we will analyze a three level system which mimics such a circumstance . + consider a one parameter dependent hamiltonian defined in the diabatic basis by the following matrix , @xmath113 with @xmath99 the perturbation and @xmath84 the characteristic transition length .
this hamiltonian can be diagonalized analytically for each time .
the upper and lower eigenenergies are represented by hyperbolas @xmath114 , as in the l - z process , and the middle energy is @xmath115 for all times ( see fig .
[ 3nive ] ) .
obviously , for a diabatic evolution , if the system were in the upper state at @xmath116 , there is a high probability @xmath117 that the system decays to the state @xmath118 at @xmath119 . in other words ,
the presence of @xmath118 affects enormously the transition probability between @xmath120 and @xmath121 .
however , we are interested in an adiabatic evolution where the transition probability to @xmath118 turns to be small . in such a situation
we want to determine whether the l - z parameters for @xmath120 and @xmath121 , that is @xmath122 at @xmath123 ( we denote it @xmath124 ) and @xmath84 , adequately describe the transition probability between these two states .
this point is not obvious at all .
for example , the distance @xmath125 is largely affected by the presence of @xmath118 and @xmath126 , contrary to the case of a two level system .
+ we have computed numerically the dynamical evolution of this model obtaining the following result for the transition probability between @xmath120 and @xmath121 : @xmath127\ ] ] for @xmath128 .
that is , although the parameters need to be renormalized , the factor is very close to @xmath129 . as a conclusion , this three levels system may be thought as two independent avoided crossings with l - z interactions .
the results of this paper attempt to extend the present understanding of one body dissipation processes . to analyze the way in which energy is transferred form the time dependent mean field to the individual nucleons we have modeled the mean field by a slowly time dependent container . the same approach has been employed by many other authors in related nuclear models @xcite .
the container is represented by a planar billiard with externally driven moving walls . to solve the quantum mechanical evolution of these simplified systems ,
we have derived a one - dimensional formulation .
this approach gives the possibility to study the evolution of highly excited states and reduces the cpu time involved in the calculations .
+ we have devoted part of the work to answer a fundamental question ; whether a landau@xmath0zener excitation mechanism governs the irreversible transport of energy from the driven wall to the particles in an adiabatic evolution .
we have analyzed a parameter dependent billiard system with goe character spectrum , concluding that in an adiabatic evolution of the external parameter , the dissipation is dominated by l - z transitions at the avoided crossings .
adiabatic limit _ is attained in the limit of an infinitely slow evolution . on the other hand ,
adiabatic evolution refers to slowly varying evolutions @xcite .
of course , the notion of slow motion needs to be clarified .
for example , we have excluded in our analysis the structure showed by the function @xmath130 in fig .
[ coef ] because its height is very small ; although the area under it is comparable to the area under any peak observed in fig .
[ cmunu ] .
however , because @xmath130 is multiplied in the differential equation ( [ auto2 ] ) by an oscillatory function with period @xmath131 ( @xmath132 is the density of energy levels ) , its effective contribution is canceled if the time required by the collective motion to sweep the structure @xmath133 is larger than @xmath134 .
+ the above adiabaticity condition is satisfied by systems where quantum effects are very important , such as nuclei .
however , as the wave number increases , the collective velocity needs to be reduced drastically . taking the semiclassical limit @xmath135 , @xmath136 , with @xmath137
, it results that @xmath138 and @xmath139 .
therefore , for any finite value of @xmath64 , the evolution is always * diabatic * in the semiclassical limit . in another terms , a semiclassical theory of dissipation requieres a scaling of @xmath64 . to our knowledge , this important point has not been taken into account in previous works @xcite .
+ the description of the damping process in terms of l - z transitions has been already done by wilkinson in the context of pure random matrix theory @xcite but to our knowledge the present work is the first study carried out for a more realistic system .
if the l - z behavior holds it is more or less straigthforward to write the corresponding diffusion equation to quantify the dissipation mechanism ( for further details see @xcite ) .
we specially thank the hospitality of the center emile borel where part of this work has been done .
e.v acknowledges support from the service de physique thorique cea - saclay .
+ we would like to thank o.bohigas and p. leboeuf for useful suggestions . + m.j.s is supported by a thalmann fellowship of the universidad de buenos aires , argentina , and is on leave from conicet , argentina .
99 e. vergini and m. saraceno , phys .
e * 52 * , 2204 ( 1995 ) .
j. r. nix and j. sierk , _ proc .
int . conf . on nuclear physics _ , bombay 1984 ( world scientific , singapore , 1985 ) , p. 365 .
j. blocki , y. boneh , j. r. nix , j. randrup , m. robel , a. j. sierk and w. j. swiatecki , ann . of phys .
* 113 * , 330 ( 1978 ) .
h. hofmann , p. j. siemens , nuc .
phys . a * 257 * , 1655 ( 1976 ) ; s. e. koonin , r. l. hatch and j. randrup , nuc .
a * 283 * , 87 ( 1977 ) ; s. e. koonin and j. randrup , nuc .
a * 289 * , 475 ( 1977 ) . c. yannouleas , m. dworzecka and j. j. griffin , nuc .
phys . a * 339 * , 219 ( 1980 ) ; c. yannouleas , nuc .
a * 439 * , 336 ( 1985 ) .
w. nrenberg , _ new vistas in nuclear dynamics _
p. j. brusserd and j. h. kock ( plenumn press , new york , 1986 ) .
w. j. swiatecki , nuc .
a * 488 * , 375c ( 1988 ) ; j.blocki , j. -j .
shi and w. j. swiatecki , nuc .
a * 554 * , 387 ( 1993 ) . c. jarzynski , phys .
e * 48 * , 4340 ( 1993 ) .
, edited by r. h. g. helleman , g. iooss and r. stora ( north - holland , amsterdam , 1983 ) .
o. bohigas , r. u. haq and a. pandey , in:_nuclear data for science and technology _ , ed .
, k. h. bckhff ( reidel , dordrecht 1983 ) p. 809 .
d. hill and j. wheeler , phys .
rev . * 89 * , 1102 ( 1952 ) . m. wilkinson , j. phys .
a * 21 * , 4021 ( 1988 ) ; m. wilkinson , j. phys .
a * 22 * , 2795 ( 1989 ) .
a. bulgac , g. do dang and d. kusnezov , ann . phys . * 242 * , 1 ( 1995 ) ; phys .
264 * , 67 ( 1996 ) .
l. d. landau and e. m. lifshitz , _ quantum mechanics _
, vol.3 ( pergamon press ltd .
paris , 1958 ) .
, edited by b. l. altshuler , p. a. lee and r. a. webb ( north - holland , amsterdam , 1991 ) . c. e. zener , proc .
a * 137 * , 696 ( 1932 ) .
diabatic basis have been used also in quantum chemistry to obtain a basis for which the coupling terms vanish ( see for example f.t .
smith , phys . rev . * 179 * , 111 ( 1969 ) ) .
d. bohm , _ quantum theory _ , ( prentice - hall , new york , 1951 ) .
r. balian and m. veneroni , ann . phys . * 135 * , 270 ( 1981 )
. l. a. bunimovich , commun . math
. phys . * 65 * , 295 ( 1979 ) .
m. wilkinson , phys .
rev . a * 41 * , 4645 ( 1990 ) . m. v. berry , ann .
* 131 * , 163 ( 1981 ) .
e. j. heller , in _ wavepacket dynamics and quantum chaology _ ,
proceedings of the les houches summer school of theoretical physics , les houches , 1989 , edited by m. j. giannoni , a. voros , and j. zinn - justin ( north - holland , amsterdam , 1989 ) .
h. p. baltes and e. r. hilf , _ spectra of finite systems _ , ( bibliographisches institut , mannheim , 1976 )
. m. v. berry and m. tabor , j. phys . a * 10 * , 371 ( 1977 ) . | this paper is concerned with the study of one - body dissipation effects in idealized models resembling a nucleus . in particular , we study the quantum mechanics of a free particle that collides elastically with the slowly moving walls of a bunimovich stadium billiard
. our results are twofold .
first , we develop a method to solve in a simple way the quantum mechanical evolution of planar billiards with moving walls .
the formalism is based on the _ scaling method _
@xcite which enables the resolution of the problem in terms of quantities defined over the boundary of the billiard .
+ the second result is related to the quantum aspects of dissipation in systems with complex spectra . we conclude that in a slowly varying evolution the energy is transferred from the boundary to the particle through landau@xmath0zener transitions .
+ + + |
many problems of fluid thermo - mechanics involving unbounded domains occur in many areas of applications , e.g. flows of a liquid in duct systems , fluid flows through a thin or long pipe or through a system of pipes in hemodynamics and so on . from a numerical point of view , these formulations are not convenient and quite practical .
therefore , an efficient natural way is to cut off unbounded parts of the domain by introducing an artificial boundary in order to limit the computational work .
then the original problem posed in an unbounded domain is approximated by a problem in a smaller bounded computational region with artificial boundary conditions prescribed at the cut boundaries .
hence , let @xmath1 be a bounded domain in @xmath2 with boundary @xmath3 . in a physical sense
, @xmath4 represents a `` truncated '' region of an unbounded system of pipes occupied by a moving heat - conducting viscous incompressible fluid .
@xmath5 will denote the `` lateral '' surface and @xmath6 represents the open parts ( cut boundaries ) of the piping system .
it is physically reasonable to assume that in / outflow pipe segments extend as straight pipes .
more precisely , @xmath7 and @xmath6 are @xmath8-smooth open disjoint not necessarily connected subsets of @xmath9 such that @xmath10 , @xmath11 for @xmath12 , @xmath13 , @xmath14 , @xmath15 , @xmath16 , @xmath17 , and the @xmath18dimensional measure of @xmath19 is zero and @xmath20 are smooth nonintersecting curves ( this means that @xmath20 are smooth curved nonintersecting edges and vertices ( conical points ) on @xmath3 are excluded ) .
moreover , all portions of @xmath6 are taken to be flat and @xmath5 and @xmath6 form a right angle @xmath21 at all points of @xmath19 ( in the sense of tangential planes ) , see figure [ pipe ] .
the flow of a viscous incompressible heat - conducting fluid is governed by balance equations for linear momentum , mass and internal energy @xcite : @xmath22 here @xmath23 , @xmath24 and @xmath25 denote the unknown velocity , pressure and temperature , respectively .
tensor @xmath26 denotes the symmetric part of the velocity gradient .
data of the problem are as follows : @xmath27 is a body force and @xmath28 a heat source term .
positive constant material coefficients represent the kinematic viscosity @xmath29 , reference density @xmath30 , heat conductivity @xmath31 and specific heat at constant volume @xmath32 . following the well - known boussinesq approximation ,
the temperature dependent density is used in the energy equation and to compute the buoyancy force @xmath33 on the right - hand side of equation .
everywhere else in the model , @xmath34 is replaced by the reference value @xmath35 .
change of density @xmath34 with temperature is given by strictly positive , nonincreasing and continuous function , such that @xmath36 the energy balance equation takes into account the phenomena of the viscous energy dissipation and adiabatic heat effects .
for rigorous derivation of the model like we refer the readers to @xcite .
rigorously derived asymptotic models describing stationary motion of heat - conducting incompressible viscous fluid through pipe - like domains can be found in @xcite . to complete the model , suitable boundary and initial conditions have to be added .
concerning the boundary conditions of the flow , it is a standard situation to prescribe a homogeneous no - slip boundary condition for the velocity of the fluid on the fixed walls of the channel , i.e. @xmath37 since nothing is known in advance about the flow through the open parts , it is really not clear what type of boundary condition for the velocity should be prescribed on @xmath6 .
the condition frequently used in numerical practice for viscous parallel flows is the most simple outflow boundary condition of the form @xmath38 which seems to be natural since it does not prescribe anything on the cut cross - section of an in / outlets of the truncated region .
therefore , this condition is usually called the `` do nothing '' ( or `` free outflow '' ) boundary condition . in , @xmath39 is the outer unit normal vector to @xmath40 , @xmath41 , while quantities @xmath42 are given functions .
in particular , for time - dependent flows , @xmath42 are given functions of time .
boundary condition results from a variational principle and does not have a real physical meaning . for further discussion on theoretical aspects as well as practical difficulties of this boundary condition and the physical meaning of the quantities
@xmath42 we refer to @xcite .
assume that @xmath42 are given smooth functions of time on @xmath40 , @xmath41 , and consider the smooth extension @xmath43 on @xmath44 such that @xmath45 .
introducing the new variable @xmath46 this amounts to solving the problem with the homogeneous `` do nothing '' boundary condition transferring the data from the right - hand side of the boundary condition to the right - hand side of the linear momentum balance equation .
hence , for simplicity , we assume throughout this paper , without loss of generality , that @xmath47 , i.e. @xmath48 concerning the heat transfer through the walls of pipes we consider the newton boundary condition @xmath49 in which @xmath50 designates the heat transfer coefficient , @xmath51 is the prescribed temperature outside the computational domain and @xmath52 represents the heat flux imposed on the lateral surfaces . on the open parts of the piping system we use the classical outflow ( `` do nothing '' )
condition @xmath53 initial conditions are considered as the given initial velocity field @xmath54 and the temperature profile @xmath55 over the flow domain @xmath56 obtained results in this paper can be extended to problems with dirichlet or the mixed ( dirichlet - neumann ) boundary conditions for the temperature on the walls .
namely , instead of , we can consider ( @xmath57 , @xmath58 ) @xmath59 the paper is organized as follows . in section [ sec : preliminaries ] , we introduce basic notations and some appropriate function spaces in order to precisely formulate our problem .
furthermore , we rewrite the energy equation by using the appropriate enthalpy transformation . in section [ main_result ] , we present the strong form of the model for the non - stationary motion of viscous incompressible heat - conducting fluids in a system of 3d pipes considered in our work , specify our smoothness assumptions on data and formulate the problem in a variational setting .
we also provide the bibliographic remarks on the subject and indicate what kind of difficulties we should overcome in the process .
the main result , the existence of strong - weak solutions , stated at the end of section [ main_result ] , is proved in section [ sec : proof ] .
the proof rests on application of schauder fixed point theorem .
first , we present basic results on the existence and uniqueness of solutions to auxiliary problems , the decoupled initial - boundary value problems for the non - stationary stokes system with mixed boundary conditions and the parabolic convection - diffusion equation with the nonlinear boundary condition . in the proof of the main result we rely on the energy estimates for auxiliary problems , regularity of stationary solutions to the stokes problem and interpolations - like inequalities .
vectors and vector functions are denoted by boldface letters . throughout the paper , we will always use positive constants @xmath60 , @xmath61 , @xmath62 , @xmath63 , @xmath64 , which are not specified and which may differ from line to line but do not depend on the functions under consideration .
throughout this paper we suppose @xmath65 $ ] , @xmath66 denotes the conjugate exponent to @xmath67 , @xmath68 .
@xmath69 denotes the usual lebesgue space equipped with the norm @xmath70 and @xmath71 , @xmath72 ( @xmath73 need not to be an integer , see @xcite ) , denotes the usual sobolev - slobodecki space with the norm @xmath74 . in case of vector - valued functions we shall use the notation @xmath75 and similarly for other function spaces . recall that @xmath76 . in the paper
we shall use the following embedding theorems ( see @xcite ) : @xmath77 for every @xmath78 .
by @xcite , @xmath79 ( the symbol `` @xmath80 '' denotes the compact embedding ) .
further , there exists the continuous operator @xmath81 , such that @xmath82 in what follows we often omit @xmath4 in notations of spaces and norms if it causes no ambiguity . unless specified otherwise , we use einstein s summation convention for indices running from @xmath83 to @xmath84 .
further , let @xmath85 and @xmath86 be the closure of @xmath87 in the norm of @xmath88 , @xmath89 and @xmath90 .
then @xmath86 is a banach space with the norm of the space @xmath88 .
further , define the space @xmath91 equipped with the norm @xmath92 where @xmath93 and @xmath27 are corresponding functions via .
[ lemma : emb_d ] the following embedding holds for the space @xmath94 : @xmath95 for the proof see ( * ? ? ? * lemma 2.1 and appendix a ) .
@xmath96 in what follows we apply a so - called enthalpy transformation @xmath97 , defined by @xmath98 in view of , @xmath99 is strictly increasing , so the transformation is bijective with @xmath100 lipschitz continuous , @xmath101 .
there exists a positive constant @xmath102 such that @xmath103 and @xmath104 now we set @xmath105 for @xmath106 and introduce the following notation @xmath107 in terms of @xmath108 , the energy equation takes the form @xmath109 where @xmath110 here we have , in view of , @xmath111 to simplify mathematical formulations , we introduce the following notations : @xmath112 in all functions @xmath113 are smooth enough , such that all integrals on the right - hand sides make sense . in , @xmath114 denotes the components of the tensor @xmath26 defined by @xmath115 let @xmath116 be fixed throughout the paper , @xmath117 .
let @xmath118 be a banach space , by @xmath119 we denote the bochner space ( see @xcite ) .
the following compactness result for spaces involving time was established by j.p .
aubin ( see @xcite ) and will be crucial to prove the main proposition of the paper .
[ aubin_compact ] let @xmath120 , @xmath121 , @xmath122 be three banach spaces where @xmath120 , @xmath122 are reflexive .
suppose that @xmath120 is continuously imbedded into @xmath121 , which is also continuously imbedded into @xmath122 , and imbedding from @xmath120 into @xmath121 is compact .
for any given @xmath123 , @xmath124 with @xmath125 , let @xmath126 then the imbedding from @xmath127 into @xmath128 is compact . to simplify the notation , in the sequel we normalize material constants @xmath50 , @xmath30 , @xmath29 , @xmath31 and @xmath32 to one .
in view of , system for the new variables @xmath129 transforms to @xmath130 we suppose that all functions in are smooth enough . at this point
we can formulate our problem in a variational sense : suppose that @xmath131 find a pair @xmath132 $ ] such that @xmath133 and the following system @xmath134 holds for every @xmath135 \in { v}_{{\gamma_2}}^{1,2 } \otimes { w}^{1,2}$ ] and for almost every @xmath136 and @xmath137 the pair @xmath138 $ ] is called the strong - weak solution to the system .
the main advantage of the formulation of the navier - stokes equations in free divergence spaces is that the pressure @xmath24 is eliminated from the system .
having @xmath93 in hand , this unknown can be recovered in the same way as in @xcite .
let us briefly describe some difficulties we have to solve in our work .
the equations represent the system with strong nonlinearities ( quadratic growth of @xmath139 in dissipative term @xmath140 ) without appropriate general existence and regularity theory . in @xcite , frehse presented a simple example of discontinuous bounded weak solution @xmath141 of the nonlinear elliptic system of the type @xmath142 , where @xmath143 is analytic and has quadratic growth in @xmath144 .
however , for scalar problems , such existence and regularity theory is well developed ( cf .
@xcite ) . nevertheless , the main ( open ) problem of the system consists in the fact that , because of the boundary condition , we can not prove that @xmath145 .
consequently , we are not able to show that the kinetic energy of the fluid is controlled by the data of the problem and solutions of need not satisfy the energy inequality .
this is due to the fact that some uncontrolled `` backward flow '' can take place at the open parts @xmath6 of the domain @xmath4 and one is not able to prove global existence results . in @xcite@xcite , kramar and neustupa prescribed an additional condition on the output ( which bounds the kinetic energy of the backward flow ) and formulated steady and evolutionary navier - stokes problems by means of appropriate variational inequalities . in @xcite , kuera and skal '
ak proved the local - in - time existence and uniqueness of a variational solution of the navier - stokes equations for iso - thermal fluids , such that @xmath146 under some smoothness restrictions on @xmath54 and @xmath147 . in @xcite ,
the same authors established similar results for the boussinesq approximations of the heat conducting incompressible fluids . in @xcite ,
kuera supposed that the `` do nothing '' problem for the navier - stokes system is solvable in suitable function class with some given data .
the author proved that there exists a unique solution for data which are small perturbations of the original ones . in case of isothermal flows , in @xcite ,
the first author of the present paper proved local - in - time existence and uniqueness of regular solutions to isothermal navier - stokes flows for newtonian fluids in three - dimensional non - smooth domains with various types of boundary conditions , such that @xmath148 , @xmath149 , which is regular in the sense that solutions possess second spatial derivatives . in @xcite ,
the same author proved the local - in - time existence , global uniqueness and smoothness of the solution of an initial - boundary - value problem for boussinesq flows in three - dimensional channel - like domains excluding viscous dissipation and considering constant density in the energy balance equation . in case of corresponding stationary flows , the existence , uniqueness and regularity of the solution
has been recently proved in @xcite . in this paper
, we extend the existence result from @xcite to non - stationary problems .
the following theorem represents the main result of this paper .
[ theorem : existence_result ] assume @xmath150 and @xmath151 then there exists the strong - weak solution @xmath138 $ ] to problem . in , @xmath152 is some specific constant depending on @xmath4 ( cf . ) .
to prove theorem [ theorem : existence_result ] we apply the following [ th : schauder ] let @xmath153 be a closed convex set in a banach space @xmath154 and let @xmath155 be a continuous mapping of @xmath153 into itself such that the image @xmath156 is precompact .
then @xmath155 has a fixed point . before we proceed to prove the main result of this paper , let us establish the following well - possedness results for the auxiliary problems , in particular , the existence of the unique regular solution to the non - stationary stokes problem with the mixed boundary conditions and the existence and uniqueness of the weak solution to the parabolic convection - diffusion equation with the nonlinear boundary condition .
[ stokes_problem ] let @xmath157 and @xmath158
. then there exists the unique function @xmath159 , @xmath160 , such that @xmath161 holds for every @xmath162 and for almost every @xmath136 and @xmath163 moreover , @xmath164 the proof can be handled in exactly the same way as the proof of ( * ? ? ?
* theorem 3.4 ) concerning , in particular , a two - dimensional problem .
the proof is based on the galerkin approximation with spectral basis and the uniform boundedness of approximate solutions in suitable spaces .
in fact , the proof is independent of dimension , therefore we skip it here .
@xmath96 [ linear_heat_problem ] let @xmath165 , @xmath166 , @xmath167 , @xmath168 .
then there exists the unique function @xmath169 , @xmath170 , such that @xmath171 holds for every @xmath172 and for almost every @xmath136 and @xmath173 first , let us introduce the following reflexive banach spaces @xmath174 and @xmath175 respectively , equipped with the norms @xmath176 and @xmath177 let us present some properties of @xmath178 and @xmath179
. let @xmath180 . raising and integrating the interpolation inequality @xmath181 over @xmath182
we get , recall @xmath183 , @xmath184 hence , we have @xmath185 in view of the imbedding relations and for sobolev spaces we have @xmath186 and applying theorem [ aubin_compact ] we get @xmath187 further , raising and integrating the interpolation inequality ( cf .
* theorem 5.2 ) ) @xmath188 from @xmath189 to @xmath190 we get @xmath191 hence , @xmath192 note that @xmath193 now and and theorem [ aubin_compact ] yield the compact embedding @xmath194 finally , and imply the compact embedding @xmath195 moreover , @xmath196 for every @xmath197 .
now , and theorem [ stokes_problem ] immediately yield the following for an arbitrary fixed couple @xmath199 \in x \otimes { l^{2}(i;l^{2})}\ ] ] we now consider the nonlinear problem @xmath200 for every @xmath135\in { v}_{{\gamma_2}}^{1,2 } \otimes { w}^{1,2}$ ] and for almost every @xmath136 and @xmath201 let us estimate the right - hand side in . applying sobolev embeddings together with the young s inequality we deduce @xmath202 now , by virtue of and , we can write @xmath203 hence , for @xmath204 \in x \otimes { l^{2}(i;l^{2})}$ ] we have @xmath205 and theorem [ stokes_problem ] yields the existence of the unique solution @xmath93 to problem and , such that @xmath206 now with @xmath93 in hand , we define @xmath108 to be the solution of problem and , i.e. @xmath108 satisfying @xmath207 for every @xmath208 and for almost every @xmath136 and @xmath209 using it is easy to see that the dissipative term on the right hand side of satisfies the estimate @xmath210 where we have used and the embeddings @xmath211 hence , for given @xmath212 and @xmath213 we have @xmath214 and the existence of the unique weak solution @xmath169 , @xmath215 to follows from theorem [ linear_heat_problem ] .
let @xmath155 denote the mapping defined by the equation @xmath216 = \mathcal{t}([\tilde{{\mbox{\boldmath{$u$}}}},\tilde{e}]).\ ] ] we have shown that the mapping @xmath217 is well defined .
we now prove that @xmath155 is continuous .
let @xmath218 \in x \otimes { l^{2}(i;l^{2})}\ ] ] and @xmath219 $ ] be a sequence in @xmath220 such that @xmath221 \rightarrow [ \tilde{{\mbox{\boldmath{$u$}}}},\tilde{e } ] \textmd { in } x \otimes { l^{2}(i;l^{2})}.\ ] ] let @xmath132 = \mathcal{t}([\tilde{{\mbox{\boldmath{$u$}}}},\tilde{e}])$ ] and @xmath222 = \mathcal{t}([\tilde{{\mbox{\boldmath{$u$}}}}_n,\tilde{e}_n])$ ] .
writing for @xmath223 $ ] and @xmath219 $ ] separately and subtracting their respective equations we get ( in view of ) @xmath224 now , let us simply modify the right - hand side to obtain @xmath225 for convective terms on the right - hand side in we can write @xmath226 and @xmath227 combining with the latter estimates we deduce @xmath228 in addition , in view of , we have @xmath229 in and , @xmath230 converges to zero by the lebesgue dominated convergence theorem .
now , estimates and yield @xmath231 provided @xmath221 \rightarrow [ \tilde{{\mbox{\boldmath{$u$}}}},\tilde{e } ] \textmd { in } x \otimes { l^{2}(i;l^{2})}.\ ] ] we now turn , for a moment , to the energy balance equation .
using the same procedure as before , writing for @xmath223 $ ] and @xmath219 $ ] , respectively , and subtracting both resulting equations yield @xmath232 the next step is to use @xmath233 in order to obtain @xmath234 let us estimate all terms on the right - hand side in to get @xmath235 and @xmath236 further @xmath237 and finally @xmath238 in view of , choosing @xmath239 sufficiently small and combining together with the estimates we deduce @xmath240 applying the gronwall s inequality to the estimate and the fact that @xmath241 , we arrive at @xmath242 for all @xmath243 , where @xmath244 here , by the lebesgue dominated convergence theorem and , @xmath245 for all @xmath243 as @xmath246 and by we deduce @xmath247 finally , in view of and we conclude @xmath248 = \mathcal{t}([\tilde{{\mbox{\boldmath{$u$}}}}_n,\tilde{e}_n ] ) \rightarrow \mathcal{t}([\tilde{{\mbox{\boldmath{$u$}}}},\tilde{e } ] ) = [ { { \mbox{\boldmath{$u$}}}},{e}].\ ] ] hence , @xmath155 is continuous . using theorem [ aubin_compact ] and the embeddings
@xmath249 we have @xmath250 by continuity of @xmath155 and compact embeddings and , @xmath155 is completely continuous .
we conclude the proof by deriving some estimates of @xmath93 and @xmath108 . applying to linear problem with and taking into account and we can write for @xmath93 the estimate @xmath251 further , following corollary [ corolary_stokes ] we have @xmath252 let @xmath253 and @xmath254 . combining and we get @xmath255 hence , @xmath256 . let us turn , for a moment , to the equation with the initial condition and derive some estimates of @xmath108 .
one is allowed to use @xmath257 as a test function in to obtain latexmath:[\[\begin{gathered } \label{est:400 } \frac{1}{2}\frac{d}{dt}\|e(t)\|^2_{l^2 } + a_{e}(\kappa(\tilde{e}(t)),{e}(t),e(t ) ) + \gamma(\beta({e}(t)),e(t ) ) \\
\leq everywhere in @xmath182 .
we are going to estimate all terms on the right - hand side of the latter inequality .
evidently , we have @xmath259 can be estimated as in to obtain @xmath260 the last term in can be handled using the interpolation inequality and the young s inequality to get @xmath261 by virtue of and the fact that @xmath101 we have @xmath262 .
hence , choosing @xmath239 sufficiently small in and combining we arrive at @xmath263 using the gronwall s inequality yields @xmath264\end{gathered}\ ] ] for all @xmath265 .
in view of and the fact that @xmath254 we have @xmath266 therefore @xmath267 hence , combining we can write @xmath268 \exp\left ( c_1 + c_2 t \right).\end{gathered}\ ] ] now , combining , and , we deduce that @xmath108 satisfies the _ a priori _ estimate of the general form @xmath269 we note that , by the _ a priori _ estimate , @xmath108 is bounded in @xmath270 independently of @xmath271 and @xmath272 .
we have previously shown that @xmath254 implies @xmath256 .
hence , there exists a fixed ball @xmath273 defined by @xmath274\in x \otimes l^{2}({i};{l}^{2 } ) , \ ; \|{\mbox{\boldmath{$v$}}}\|_x \leq \frac{1}{2 c_s t^{1/8 } } , \ ; \|z\|_{l^{2}({i};{l}^{2 } ) } \leq r \right\}\ ] ] ( @xmath275 sufficiently large ) such that @xmath276 , where the operator @xmath277 is completely continuous .
now , by theorem [ th : schauder ] , there exists the strong - weak solution @xmath138 $ ] to problem . this completes the proof of the main result . @xmath96
the first author of this work has been supported by the project gar 13 - 18652s .
the second author of this work has been supported by the _ croatian science foundation _
( scientific project 3955 : _ mathematical modeling and numerical simulations of processes in thin or porous domains _ ) .
bene , m. , kuera , p. : _ solutions to the navier - stokes equations with mixed boundary conditions in two - dimensional bounded domains . _
submitted for publication , available online at : http://arxiv.org/abs/1409.4666 ladyzhenskaya , o.a . ,
solonnikov , v.a . ,
uraltseva , n.n . : _ linear and quasilinear equations of parabolic type . _ translations of mathematical monographs 23 , american mathematical society , providence , r.i . | we study an initial - boundary - value problem for time - dependent flows of heat - conducting viscous incompressible fluids in a system of three - dimensional pipes on a time interval @xmath0 .
here we are motivated by the bounded domain approach with `` do - nothing '' boundary conditions . in terms of the velocity , pressure and enthalpy of the fluid ,
such flows are described by a parabolic system with strong nonlinearities and including the artificial boundary conditions for the velocity and nonlinear boundary conditions for the so called enthalpy of the fluid .
the present analysis is devoted to the proof of the existence of weak solutions for the above problem .
in addition , we deal with some regularity for the velocity of the fluid . |
the compelling evidence for the presence of dm in the different structures of the universe @xcite has motivated numerous efforts to search for dm by means of astrophysical observations .
if dm is made of wimps annihilating into standard model particles , indirect searches for dm annihilations with vhe @xmath0-rays provide a very promising way to constrain the nature of the dm particle : @xmath0-rays are free of any propagation effects on short distances ( @xmath1 1mpc ) and dm particle annihilation is predicted to give a unique @xmath0-ray spectrum .
such searches are often conducted using pointed observations toward nearby dm overdensities , because the annihilation rate is proportional to the squared dm density .
popular targets include the galactic center @xcite , satellite galaxies of the milky - way ( mw ) @xcite , globular clusters @xcite and clusters of galaxies @xcite .
+ the dsphs of the local group best meet the criteria for a clear and unambiguous detection of dm .
they are gravitationally bound objects and contain up to @xmath2(@xmath3 ) more mass in dm than in visible matter .
as opposed to the galactic center , they are environments with a favorably low astrophysical @xmath0-ray background .
neither astrophysical @xmath0-ray sources ( supernova remnants , pulsar wind nebulae , etc ) nor gas acting as target material for cosmic rays have been observed in these systems @xcite .
furthermore , their relative proximity and high galactic latitude make them the best astrophysical targets for high signal - to - noise detection .
+ this paper gives a status of the dsph observational program carried out by the ground - based vhe @xmath0-ray observatory veritas .
section [ sec2 ] presents the instrument and summarizes the veritas dsph observations and data analysis .
section [ sec3 ] interprets the results in terms of constraints on dm models , focusing on models which include a sommerfeld enhancement .
finally , section [ sec4 ] is devoted to the conclusion .
veritas is an array of four 12-meter imaging atmospheric cherenkov telescopes ( iacts ) located at the base camp of the f. l. whipple observatory in southern arizona .
each veritas telescope consists of a large optical reflector which focuses the cherenkov light emitted by particle air showers onto a camera of 499 photomultiplier tubes .
the large effective area ( @xmath4 ) , in conjunction with the stereoscopic imaging of air showers , enables veritas to be sensitive over a wide range of energies ( from 100 gev to 30 tev ) with an energy and angular resolution of 15 - 20% and @xmath5 per reconstructed @xmath6-ray , respectively .
veritas is able to detect a point source with 1% of the crab nebula flux at a statistical significance of 5 standard deviations above background ( @xmath7 ) in approximately 25 hours of observations . for further details about veritas ,
see , e.g. , @xcite . [ cols="^,^,^,^,^,^",options="header " , ] the absence of signal in any of these observations can be used to derive constraints on various dark matter models .
the @xmath0-ray flux from the annihilations of dm particles , of mass @xmath8 , in a spherical dm halo is given by a particle physics term times an astrophysics term : @xmath9 the particle physics term contains all the information about the dm particle : its mass @xmath8 , its total velocity - weighted annihilation cross - section @xmath10 and the differential @xmath0-ray spectrum from all final states weighted by their corresponding branching ratios , @xmath11 .
the astrophysical factor @xmath12 is the square of the dm density integrated along the line of sight ( s ) and over the solid angle @xmath13 : @xmath14)ds.}\label{fap}\ ] ] the solid angle is given here by the size of the signal search region defined previously in the analysis , i.e. @xmath15 @xmath1 0.12 deg .
the estimate of the astrophysical factor requires a modeling of the dsph dm distribution .
each dsph dm distribution has been modeled with a navarro , frenk & white ( nfw ) profile @xcite , except segue 1 for which the dm distribution has been modeled with an einasto profile @xcite .
the parameters of a dsph profile are constrained using its star kinematics and have each been taken from the literature @xcite .
the resulting astrophysical factors of draco , ursa minor , bootes 1 , willman 1 and segue 1 are presented in table [ tab2 ] , along with their distance and profile modeling .
+ once the astrophysical factors have been estimated , the uls in the number of @xmath6-ray candidates for each dsph can be translated into uls on the total annihilation cross - section : @xmath16 where @xmath17 is the total observation time and @xmath18 is the effective area as a function of energy , zenith angle and the offset of the source from the pointing position .
+ figure [ fig1 ] shows the veritas 95% cl uls on the velocity - weitghted annihilation cross - section @xmath19 as a function of the dm particle mass for each of the five observed dsph listed in table [ tab1 ] .
the draco , ursa minor , bootes 1 and willman 1 limits have been obtained using a composite dm spectrum with a 90 % @xmath20 branching ratio and a 10% @xmath21 branching ratio @xcite .
the segue 1 limits have been obtained assuming a pure @xmath22 dm annihilation spectrum , which is similar to the composite @xmath20 - @xmath21 spectrum .
the limits are in the range @xmath23 , depending on the considered dsph .
the best limits are obtained for the segue 1 dsph , which has the deepest exposure and one of the highest astrophysical factor among the five dsphs ( see table [ tab2 ] ) .
the 95% cl upper limit on @xmath24 is at the level of @xmath25 around 1 tev and is among the best reported so far with dsph vhe observations . as a function of the dm particle mass for the five dsphs observed with veritas .
the limits are computed using a composite dm annihilation spectrum ( 10% @xmath21 + 90% @xmath20 ) , except for the case of segue 1 in which a pure @xmath22 channel was considered .
the grey band area represents a range of generic values for the annihilation cross - section for thermally produced dm.,width=288 ] in this section , we consider dm models where the annihilation cross - section can be boosted with the sommerfeld enhancement effect .
we use the segue 1 data to constrain such models .
the sommerfeld enhancement is a particle physics effect that was suggested to explain the cosmic ray lepton anomalies measured by atic and pamela @xcite .
it arises when the two dm particles interact through an attractive potential , mediated by the exchange of a massive boson @xcite .
the annihilation cross - section enhancement is particularly effective when the dm particle relative velocity is very small .
depending on the mass @xmath26 and the coupling @xmath27 of the exchanged boson , the sommerfeld enhancement can exhibit a series of resonances for specific values of the dm particle mass , giving very large boost factors up to @xmath28 .
in such scenarios , the annihilation cross - section is given by : @xmath29 where @xmath30 is the sommerfeld boost factor averaged over the dm particle velocity distribution and depends on the dm particle mass .
@xmath31 is the thermal wimp annihilation cross - section before freeze - out .
the sommerfeld enhancement is of particular interest for cold dm halos like dsphs , where the mean star velocity dispersion can reach a few @xmath32 . + we focus here on two models comprising a sommerfeld enhancement to the annihilation cross - section . the first model ( hereafter model i , @xcite ) assumes that the dark matter particle is a wino - like neutralino @xmath33 ( arising in susy extensions of the standard model ) , which annihilates through the exchange of a @xmath34 boson ( @xmath35 , @xmath36 ) , thus leading to a sommerfeld enhancement . the second model ( hereafter model ii )
introduces a new force in the dark sector @xcite , carried by a light scalar field @xmath37 predominantly decaying into leptons and with a mass @xmath38 . in such models , dark matter annihilates to a pair of @xmath37 scalar particles , with an annihilation cross - section boosted by the sommerfeld enhancement .
the coupling @xmath27 of the light scalar particle @xmath37 to the dark matter particle is determined assuming that @xmath39 is the only channel that regulates the dark matter density before freeze - out @xcite .
figure [ fig2 ] shows the veritas constraints on each of these models , derived with the observations of segue 1 .
the left panel of figure [ fig2 ] shows the constraints on model i. the sommerfeld enhancement exhibits two resonances in the considered dark matter particle mass range , for @xmath40 and @xmath41 , respectively .
veritas excludes these resonances , which boost the annihilation cross - section far beyond the canonical @xmath42 .
the right panel of figure [ fig2 ] shows the veritas constraints on model ii , for a scalar particle with mass @xmath43 .
two channels in which the scalar particle decays either to @xmath44 or @xmath45 have been considered .
veritas observations start to disfavor such models , especially for the @xmath46 channel where some of the resonances are beyond @xmath42 .
this result holds for @xmath37 particle masses up to a few gev .
since its commissioning phase , the veritas observatory has been actively observing the best dsph targets visible from the northern hemisphere , with data currently taken toward the draco , ursa minor , bootes 1 , willman 1 and segue 1 galaxies .
the non - detection of a vhe @xmath0-ray signal in any of these observations has lead to limits on the velocity - weighted annihilation cross - section @xmath10 of the order of @xmath23 .
the segue 1 data have provided the best limits on @xmath10 reported so far with any dsph observations , though these limits are at least two orders of magnitude away from the canonical @xmath42 .
the segue 1 data have also been used to constrain models with a sommerfeld enhancement , disfavoring them as being a good interpretation of the cosmic ray lepton excesses measured by atic and pamela .
this research was supported by grants from the u.s .
department of energy , the u.s . national science foundation and the smithsonian institution , by nserc in canada , by science foundation ireland and by the science and technology facilities council in the uk .
we acknowledge the work of the technical support staff at the fred lawrence whipple observatory .
bertone g. , et al .
, * 405 * , 279 ( 2005 ) bergstrom l. , et al . ,
* 9 * , 137 ( 1998 ) kosack k. , et al . ,
apjl , * 608 * , 97 ( 2004 ) aharonian f. , et al . , prl , * 97 * , 221102 ( 2006 ) abramowski a. , et al . , prl , * 106 * , 161301 ( 2011 ) evans n.w . , et al . ,
prd , * 69 * , 123501 ( 2004 ) acciari v. a. , et al . , apj , * 720 * , 1174 ( 2010 ) aharonian f. , et al . ,
phys . , * 29 * , 55 ( 2008 ) aharonian f. , et al . ,
apj , * 691 * , 175 ( 2009 ) abramowski a. , et al .
, astropart .
* 34 * , 608 ( 2011 ) aliu e. , et al . , apj , * 697 * , 1299 ( 2009 ) j. aleksic , et al . ,
jcap , * 06 * , 035 , ( 2011 ) wood m. , et al . , apj , * 678 * , 594 ( 2008 ) abramowski a. , et al . , arxiv:1104.2548 ( 2011 ) aleksic j. , et al . , apj , * 710 * , 634 ( 2010 ) mateo m. , annu
, * 36 * , 435 ( 1998 ) j. holder et al .
, astropart .
phys . , * 25 * , 391 ( 2006 ) acciari v. a. , et al . , apj , * 679 * , 1427 ( 2008 ) t. li and y. ma , apj . ,
* 272 * , 317 ( 1983 ) w.a .
rolke et al . , nucl .
methods , * a551 * , 493 ( 2005 ) j.f .
navarro , et al , apj , * 490 * , 493 ( 1997 ) j.f .
navarro , et al , mnras , * 402 * , 21 ( 2010 ) essig r. , et al . , prd , * 80 * , 023506 ( 2009 ) essig r. , et al . , prd , * 82 * , 123503 ( 2010 ) j. chang , et al
. , nature , * 456 * , 362 ( 2008 ) o. adriani , et al . , nature ,
* 458 * , 607 ( 2009 ) lattanzi m. and silk j. , prd , * 79 * , 083523 ( 2009 ) n. arkani - hamed , et al . ,
prd , * 79 * , 015014 ( 2009 ) j. l. feng , m. kaplinghat & h - b .
yu , prd , * 82 * , 083525 ( 2010 ) | in the cosmological paradigm , cold dark matter ( dm ) dominates the mass content of the universe and is present at every scale .
candidates for dm include many extensions of the standard model , with a weakly interacting massive particle ( wimp ) in the mass range from 50 gev to greater than 10 tev .
the self - annihilation of wimps in astrophysical regions of high dm density can produce secondary particles , including very high energy ( vhe ) @xmath0-rays , with energies up to the dark matter particle mass .
the veritas array of cherenkov telescopes , designed for the detection of vhe @xmath0-rays in the 100 gev-10 tev energy range , is an appropriate instrument for the detection of dm and is complementary to fermi - lat .
dwarf spheroidal galaxies ( dsphs ) of the local group are potentially the best targets to search for the annihilation signature of dm due to their proximity and large dm content .
we report on the latest veritas observations of dsphs and discuss the results in the framework of wimp models . |
metastability in dynamical systems refers to the existence of subdomains , known as _ metastable sets _ or _ almost - invariant sets _ , within which trajectories are confined for long periods of time .
metastability can be driven by small amounts of noise , for example , in a gradient system @xmath0 where @xmath1 is a double well potential , the noise enables sample paths to move between the metastable wells @xcite .
alternatively , metastability can arise from purely deterministic means if trajectories pass between two or more subdomains sufficiently infrequently .
applications of metastability or almost - invariance include molecular dynamics @xcite , where the metastable sets are stable molecular conformations ; astrodynamics @xcite , where the metastable sets are regions from which asteroid escape is rare ; physical oceanography @xcite , where metastable regions are stable structures such as gyres and eddies ; and atmospheric science @xcite , where vortices in the stratosphere form time - dependent metastable regions
. in the deterministic setting one considers a nonsingular map @xmath2 on a smooth manifold @xmath3 .
the perron - frobenius operator @xmath4 is the natural push - forward for densities @xmath5 and is a key tool for studying the metastability of @xmath6 .
the existence of metastability or almost - invariance in deterministic systems is often linked to the existence of isolated spectral values of the perron - frobenius operator @xcite . fixed points of @xmath7 are invariant densities for @xmath6 , and in suitable banach spaces , the presence of real isolated sub - unit eigenvalues of the perron - frobenius operator correspond to eigenfunctions whose decay rates are _ slower _ than the exponential separation of nearby trajectories @xcite .
these eigenfunctions have been used to heuristically decompose the domain into metastable regions , linking the slow exponential decay of the eigenfunctions with slow exchange of trajectories and one has @xmath8 for an isolated @xmath9 , the simplest way to decompose the system domain is to partition it into two sets @xmath10 and @xmath11 . ] , see e.g. @xcite .
the slow exponential decay of such eigenfunctions may also be linked with slow exponential _
escape_. one may create an _
open _ dynamical system by restricting the dynamics to a metastable set .
trajectories will stay in the metastable set for a while , but eventually leave .
one may ask how the _ rate of escape _ from a metastable set is related to the second ( say ) eigenvalue @xmath12 of the perron - frobenius operator @xmath13 acting on the full domain . we answered this question in @xcite : if the metastable set is defined as @xmath14 where @xmath8 , then the rate of escape from this set is slower than @xmath15 .
the concept of metastability and almost invariant sets was extended to random dynamical systems in @xcite , where these sets are referred to as _ coherent structures_. to move from deterministic to random ( or time - dependent ) concepts of metastability , @xcite and @xcite introduced the _ lyapunov spectrum _ for cocycles of random perron - frobenius operators , replacing the spectrum of a single deterministic perron - frobenius operator @xmath7 .
the associated random _
oseledets subspaces _ now play the role of eigenfunctions when determining the random metastable sets . numerical algorithms and experiments based on this theory
were detailed in @xcite .
other work on metastability of random or perturbed dynamical systems includes papers of colonius _ et al .
_ @xcite .
our goal in this paper is to link the slow decay of random functions induced by the perron - frobenius cocycle with escape rates from random metastable sets .
studies of escape rates for random dynamical systems have , to our knowledge , only been concerned with escape from _ fixed _ ( @xmath16-invariant )
sets under random or randomly perturbed maps ( see for example @xcite ; for more recent work see @xcite ) . in this paper
we work with a more general concept of escape from a _ random _ set under a _ random _ map .
we extend the results of @xcite to random dynamical systems by showing a relationship between lyapunov spectrum and the corresponding random escape rates from metastable sets .
more precisely , in section [ sec : general ] , we show in a rather general setting ( @xmath3 polish , nonsingular dynamics ) that given a lyapunov exponent @xmath17 of @xmath18 and @xmath19 , one can define a metastable random set @xmath20 along the orbit of @xmath16 as @xmath21 .
our first main result states that the escape rate from @xmath22 is slower than @xmath23 . in section [ sec : oseledets ] we extend these results to random dynamical systems that admit an oseledets splitting and , in particular , to rychlik random dynamical systems where the dynamics are given by random expanding piecewise @xmath24 interval maps and @xmath7 acts on @xmath25)$ ] . this setting has historically been a standard testbed for spectral analysis of chaotic dynamical systems .
for example @xcite proved that if @xmath26 has bounded variation and if @xmath27 then @xmath28 is quasi - compact .
soon after , keller @xcite proved that @xmath29 is the essential spectral radius .
as @xmath29 is intimately connected with the average expansion experienced along orbits , @xmath30 spectral points of @xmath7 larger than @xmath29 in magnitude can not be explained by local expansion of @xmath6 and must be due to the influence of global structures such as almost - invariant and metastable sets , producing decay rates _ slower than the average local expansion rate_. in the random setting , froyland _ et al . _
@xcite proved a result parallel to @xcite for random rychlik maps .
our main result in section [ sec : oseledets ] relates the escape rate from random almost - invariant sets to isolated values in the lyapunov spectrum of @xmath7 . in section [ sec : rsft ] we adapt the techniques of section [ sec : general ] to partition a random shift of finite type into two disjoint random subshifts , each with a topological entropy that is large , relative to the topological entropy of the original shift .
the more metastable the random shift , the larger the topological entropy of the subshifts .
more precisely , we show how to constructively decompose a random shift @xmath31 into two complementary random shifts @xmath32 with disjoint alphabets so that @xmath33 , where @xmath34 is the second largest lyapunov exponent of the adjacency matrix cocycle for @xmath31 .
we use the notation of arnold @xcite .
let @xmath35 be an invertible measure - preserving transformation of a probability space @xmath36 .
we will call @xmath35 the _ base transformation _ and the tuple @xmath37 the _ base dynamical system_. by @xmath38 we denote the set of endomorphisms ( or transformations ) of a space @xmath39 , which preserves whatever structure @xmath39 may have ( e.g. linearity or measurability ) .
a _ random dynamical system _ is a tuple @xmath40 where @xmath41 is a family of transformations of @xmath39 indexed by @xmath42 .
a random dynamical system defines a discrete _ cocycle _
@xmath43 by @xmath44 with the property * @xmath45 ; * @xmath46 . the function @xmath47 is referred to as the cocycle while @xmath48 is called the _ generator_. we will study two types of cocycles : _ measurable map cocycles _ and their _ perron - frobenius operator cocycles_. firstly , let us assume that @xmath49 , a polish space metrized with a metric @xmath50 .
further , we equip @xmath3 with the borel @xmath51-algebra @xmath52 and a finite measure @xmath53 so that @xmath54 is a measure space
. also let @xmath55 be the borel @xmath51-algebra on @xmath56 ( with respect to the discrete topology ) .
a _ measurable ( map ) cocycle _ is a cocycle @xmath57 such that the mapping @xmath58 is @xmath59-measurable and each @xmath60 is non - singular ( @xmath61 ) . on the other hand ,
supposing that @xmath62 ( or a subspace of @xmath63 ) , we may define a _
perron - frobenius operator cocycle _ as follows .
let @xmath64 be a measurable map cocycle over @xmath65 .
the corresponding _ perron - frobenius operator cocycle _ is a linear cocycle @xmath66 whose generator @xmath67 is given by @xmath68 define a _
random set _ to be any set - valued function @xmath69 such that for each @xmath70 the mapping @xmath71 is measurable ( see @xcite for more details on random sets ) .
one is then able to define _ rate of escape _ from a given random set under the random dynamics of the cocycle .
[ def : escape ] let @xmath72 be a measurable cocycle over @xmath73 and let @xmath69 be a measurable random set .
the _ random escape rate _ with respect to @xmath53 is the non - negative valued function @xmath74 given by @xmath75 where @xmath76 [ prop : er_cae ] assume that the base system @xmath77 is ergodic and that for almost every @xmath18 the radon - nikodyn derivative , @xmath78 is bounded . for any fixed random set @xmath79 , @xmath80 is constant almost everywhere .
we begin with the observation that @xmath81 ( see ) so that @xmath82 . using the boundedness of
the radon - nikodym derivative one sees that @xmath83 .
we will now show that @xmath84 for almost any @xmath18 .
assume otherwise .
as @xmath22 is a random set , @xmath85 is measurable and there exists @xmath86 such that the set @xmath87 has @xmath88 . since @xmath89 and @xmath35 preserves @xmath90 we must have @xmath91 a.e . but
this can not be since @xmath90 is ergodic .
let @xmath92 be a perron - frobenius operator cocycle corresponding to a measurable map cocycle @xmath57 .
for any @xmath93 , and @xmath18 the _ lyapunov exponent _ is defined to be @xmath94 we also define the _ lyapunov spectrum _ to be the set of all lyapunov exponents : @xmath95 . as
each @xmath96 is a markov operator we have @xmath97 and therefore @xmath98 $ ] . with the definition of the perron - frobenius cocycle in ,
it is natural to use the @xmath99-norm for calculating the lyapunov spectrum .
however we will see later in section [ sec : oseledets ] that when working with subspaces of @xmath99 other norms are sometimes more informative .
our main theorem relates the lyapunov exponents of a perron - frobenius operator cocycle @xmath100 to the rates of escape from particular random sets under the corresponding measurable map cocycle @xmath101 .
[ thm : main ] let @xmath102 be a measurable map cocycle over @xmath103 and let @xmath92 be the corresponding perron - frobenius cocycle as defined in . fix @xmath104 and suppose that there exists an @xmath105 such that @xmath106 .
let @xmath107 and @xmath108 be two random sets such that @xmath109 .
then one has @xmath110 .
the proof of theorem [ thm : main ] follows after a preliminary lemma .
[ lem : main ] in the notation of theorem [ thm : main ] we have for every @xmath111 and all @xmath18 @xmath112 firstly we will show that @xmath113 for all @xmath114 .
from one can see that @xmath115 preserves integrals over all of @xmath3 therefore @xmath116 , a constant for all @xmath114 .
this implies that @xmath117 for all @xmath114 .
suppose that @xmath118 .
then @xmath119 this is a contradiction as @xmath106 , therefore @xmath120 .
now we have @xmath121 and @xmath122 adding equations and gives the required result . for
any set @xmath123 denote the indicator function of @xmath124 by @xmath125 .
let @xmath126 be integers such that @xmath127 and let @xmath128 .
using we derive the following : @xmath129 hence @xmath130 now letting @xmath131 ( defined as in ) we have for all @xmath132 @xmath133 where we have used the relation @xmath134 , easily obtainable from . by considering all @xmath135
we arrive at the following series of inequalities : @xmath136 hence @xmath137 where the equality above is due to lemma [ lem : main ] , and the second inequality holds because @xmath138 . by taking logarithms , dividing by @xmath139 and taking limit supremum as @xmath140 we arrive at the required inequality @xmath141 .
the inequality for @xmath142 is obtained similarly by considering @xmath143 in place of @xmath144 .
[ rem : conditional ] we note that if for a random set @xmath69 one defines a _ conditional operator cocycle _ @xmath145 by @xmath146 for all @xmath18 and @xmath19 , then lebesgue escape rate is given by @xmath147 , which equals the absolute value of the lyapunov exponent of a constant function with respect to this conditional cocycle .
theorem [ thm : main ] presents a method of finding pairs of disjoint random sets with low escape rates for a large class of random dynamical systems .
for the remainder of this section we will investigate some of the consequences of this result .
lemma [ lem : fullspectrum ] will show that in a very general setting one may choose any @xmath148 , find an appropriate @xmath149 with lyapunov exponent @xmath150 and obtain two random sets with escape lower than @xmath151 .
in particular @xmath152 may be arbitrarily close to @xmath153 , however , as we will see later choosing such @xmath152 often results in highly irregular random metastable sets .
a mapping @xmath154 is said to be a _ random @xmath99-function _ if it is measurable with respect to @xmath155 and the borel @xmath51-algebra on @xmath63 .
if each @xmath156 is a density in @xmath63 , it is called a _
random density_. such a density is said to be _ preserved _ by a perron - frobenius operator cocycle @xmath7 if @xmath157 for almost every @xmath18 .
[ lem : fullspectrum ] let @xmath158 be a perron - frobenius operator cocycle ( of a measurable map cocycle @xmath6 ) over @xmath65 that preserves a positive random density @xmath159 .
suppose that there exists a random function @xmath160 so that @xmath161 for almost all @xmath104 .
then for every @xmath162 $ ] there exists a random function @xmath163 such that @xmath164 for almost every @xmath18 .
we modify an argument from baladi @xcite ( theorem 1.5 ( 7 ) ) .
define @xmath144 so that @xmath165 for every @xmath18 .
note that @xmath144 is measurable and each @xmath166 . for any @xmath128 we have @xmath167
thus @xmath168 almost everywhere .
now for @xmath169 let @xmath170 . since @xmath171 is measurable , the set @xmath172 is also measurable .
fix @xmath173 sufficiently small so that @xmath174 .
the poincar recurrence theorem asserts that @xmath90-almost surely there is a sequence @xmath175 such that @xmath176 .
hence @xmath177 from which we obtain @xmath178 an application of theorem [ thm : main ] to @xmath144 implies that for any negative @xmath152 , arbitrarily close to zero , there exist complementary random sets whose rate of escape is slower than @xmath151 .
[ ex : fullspectrum ] let @xmath103 be the full two - sided 2-shift on @xmath179 equipped with the @xmath51-algebra @xmath155 generated by cylinders , and with bernoulli probability measure @xmath90 .
let @xmath180)$ ] be the generator of a cocycle @xmath101 , constant on cylinders , given by @xmath181 where @xmath182 for @xmath183 , @xmath184 .
it is easy to check that the corresponding perron - frobenius operator cocycle @xmath100 satisfies lemma [ lem : fullspectrum ] with @xmath185 for all @xmath16 and @xmath186 where @xmath187 after applying lemma [ lem : fullspectrum ] we conclude that any @xmath188 $ ] is a lyapunov exponent with an essentially bounded lyapunov function , hence by theorem [ thm : main ] , there exist complementary random sets with arbitrarily low escape rate . for a numerical demonstration we set @xmath189 and @xmath190 .
we choose @xmath191 such that @xmath192 for all @xmath193 and @xmath194 equals the @xmath195 digit in the fractional part of the binary expansion of @xmath196 for @xmath197 .
the first few central elements of @xmath198 , with the zeroth element underlined , are : @xmath199 numerical approximations of @xmath200 for some values of @xmath152 are shown in figure [ fig : lyapunovfunctions ] . for @xmath201 applying the construction in theorem [ thm : main ] we see from the graph of @xmath200 that @xmath202 and @xmath203 $ ] .
as @xmath152 becomes closer to @xmath153 we can see more oscillations in the @xmath200 and subsequently higher disconnectedness of the corresponding sets @xmath204 . corresponding to different lyapunov exponents in example [ ex : fullspectrum ] .
note the increased irregularity as @xmath152 approaches zero.,width=491 ]
in this section we extend theorem [ thm : main ] to apply in a banach space @xmath205 , with @xmath206 , in which the perron - frobenius cocycle admits an _ oseledets splitting_. we then apply these new results to expanding maps of the unit interval , @xmath207 $ ] , where the banach space @xmath208 is replaced by the space of functions of bounded variation @xmath209 .
a linear operator cocycle @xmath210 over @xmath65 is said to be _ quasi - compact _ if for almost every @xmath16 there exists an @xmath211 such that the set @xmath212 is finite co - dimensional .
we will denote the infimal such @xmath213 by @xmath214 .
quasi - compact cocycles have the property that lyapunov exponents larger than @xmath214 are isolated . for an isolated lyapunov exponent @xmath215 ,
let @xmath169 be small enough so that @xmath216 .
if the co - dimension of @xmath217 in @xmath218 is @xmath50 then we call @xmath219 a lyapunov exponent of _ multiplicity _ @xmath50
. there are at most countably many of these and we refer to them as _ exceptional lyapunov exponents_. the _ exceptional lyapunov spectrum _ is the set of pairs of exceptional lyapunov exponents and their multiplicities , @xmath220 . from now on ,
we retain the assumption that the base system @xmath221 is ergodic , which ensures that both @xmath222 and @xmath223 are constant almost everywhere . by @xmath224 and @xmath225
we will denote the subspaces of the grassmannian @xmath226 of @xmath227 consisting only of subspaces of dimension @xmath50 and codimension @xmath228 respectively ; see the appendix for more details on grassmannians and their topology .
a quasi - compact linear operator cocycle @xmath229 over @xmath65 with exceptional spectrum @xmath230 , @xmath231 , admits a _ lyapunov filtration _ over a @xmath35-invariant set @xmath232 of full measure , if there exists a collection of maps @xmath233 , such that for all @xmath234 and all @xmath235 * @xmath236 * @xmath237 ; * @xmath238 ; * @xmath239 if and only if @xmath240 .
an _ oseledets splitting _ for @xmath7 is a lyapunov filtration with an additional family of maps @xmath241 such that for all @xmath234 and @xmath235 * @xmath242 ; * @xmath243 ; * @xmath244 if @xmath245 .
a lyapunov filtration is measurable if each @xmath246 is @xmath247-measurable .
an oseledets splitting is measurable if its lyapunov filtration is measurable and each of the maps @xmath248 is measurable . for more details on measurability
we refer the reader to the appendix . in order to connect the @xmath227-lyapunov spectrum to escape rates ,
we first need to relate the @xmath227-lyapunov exponents to the @xmath99-lyapunov exponents used in theorem [ thm : main ] . for this
we shall require a certain relation between the two norms .
[ thm : equiv_norms ] let @xmath249 be a quasi - compact linear operator cocycle over @xmath250 with exceptional spectrum @xmath230 and a measurable oseledets splitting @xmath251 on @xmath252 .
let @xmath253 be a second norm on @xmath227 such that @xmath254 for some @xmath255 .
then for almost any @xmath234 , @xmath256 and any @xmath257 , we have @xmath258 ; that is , the lyapunov exponents with respect to the two norms are equal almost everywhere .
firstly note that scaling a norm by a constant factor does not change the lyapunov exponent , hence without loss of generality we may assume that @xmath259 .
fix @xmath260 . since @xmath261 the inequality @xmath262 for all @xmath263 follows
now for the reverse inequality . define a function @xmath264 by @xmath265 where @xmath266 is as in lemma [ lem : usc ] .
since @xmath267 is @xmath268-measurable and @xmath269 is @xmath270-measurable , the function @xmath228 is @xmath271-measurable . for a positive integer @xmath272 let @xmath273 be the indicator function of the ( measurable ) set @xmath274 . given any @xmath234 , the function @xmath275 is finite so @xmath276 for all @xmath277 .
thus @xmath278 pointwise . by lebesgue s dominated convergence theorem
, we see that @xmath279 as @xmath280 .
thus we may choose an @xmath272 large enough so that @xmath281 .
by poincar recurrence there almost surely exists a sequence @xmath282 such that @xmath283 .
then @xmath284 which completes the proof
. by reversing the appropriate inequalities in the proof of theorem [ thm : equiv_norms ] and a similar modification of lemma [ lem : usc ] one can see that the same result holds when the two norms satisfy the relation @xmath285 for some @xmath255 .
in particular theorem [ thm : equiv_norms ] is satisfied when the two norms are equivalent . a direct consequence of theorem [ thm : main ] and theorem [ thm : equiv_norms ] is : [ cor : rychlik ] let @xmath102 be a measurable map cocycle over @xmath103 and let its perron - frobenius cocycle be @xmath286 , where @xmath287 and @xmath288 .
suppose that @xmath100 is quasi compact , with exceptional spectrum @xmath289 , and admits a measurable oseledets splitting @xmath290 . for almost any @xmath234 and any @xmath291 if @xmath292 are random sets such that @xmath293 , then @xmath294 .
this result extends the application of theorem [ thm : main ] to perron - frobenius cocycles on banach spaces for which the cocycle is quasi - compact and the banach space norm dominates the @xmath99 norm .
we now focus on the unit interval , @xmath207 $ ] , one - dimensional map cocycles @xmath295 , and their perron - frobenius operators .
in @xcite it is shown that the perron - frobenius cocycle is quasi - compact if the _ index of compactness _ ( a quantity corresponding to the essential spectral radius in the deterministic setting ) @xmath296<++ > this formula for @xmath297 suggests that any lyapunov spectral points lying between @xmath297 and @xmath153 ( the latter corresponding to the random invariant density ) are associated with large - scale structures responsible for rates of mixing slower than the local expansion of trajectories can account for .
these structures are commonly referred to as _ coherent structures , coherent sets _ or _
( random ) almost invariant sets _ @xcite .
we apply the results of corollary [ cor : rychlik ] to show that these sets also posses a slow rate of escape , bounded by the corresponding exponent in the lyapunov spectrum .
let @xmath298 be the unit interval @xmath299 $ ] with borel @xmath51-algebra and lebesgue measure .
recall that _
variation _ of a function @xmath300 is given by @xmath301 where @xmath302 means that @xmath303 is an @xmath99 version of @xmath144 and the supremum is taken over all finite sets @xmath304 .
define @xmath305 to be the space of functions of bounded variation , equipped with the norm @xmath306 .
let @xmath307 be a shift space on @xmath308 symbols with the left shift map @xmath309 given by @xmath310 . furthermore ,
suppose @xmath155 is the borel @xmath51-algebra generated by cylinders in @xmath42 and suppose that @xmath90 is an ergodic shift - invariant probability measure on @xmath42 .
a _ rychlik map cocycle _ is a cocycle @xmath311 obtained from a collection of @xmath308 _ rychlik maps _
@xmath312 where the generator @xmath313 is given by @xmath314 .
rychlik maps @xcite , a generalisation of lasota - yorke maps , form a large class of almost - everywhere @xmath315 maps of the unit interval whose reciprocal of the modulus of the derivative has bounded variation . we will denote the corresponding perron - frobenius operator cocycle @xmath316 . for more details
we refer the reader to @xcite .
in corollary 28 @xcite it is shown that the perron - frobenius cocycle of any rychlik map cocycle that is _ expanding - on - average _ ( that is , @xmath317 ) admits a @xmath90-continuous ( and therefore measurable ) oseledets splitting in @xmath30 .
we combine this result with corollary [ cor : rychlik ] to obtain the following .
[ cor : rychlik2 ] let @xmath318 be a rychlik map cocycle which is expanding on average and let @xmath319 be its perron - frobenius operator cocycle , which admits a measurable oseledets splitting on a set of full measure @xmath232
. for any isolated lyapunov exponent @xmath320 and almost any @xmath234 there exists a partition of @xmath321 into random sets @xmath292 such that @xmath322 . since @xmath323 , a direct application of corollary [ cor : rychlik ] shows that random sets @xmath292 satisfying @xmath324 have escape rate lower than @xmath325 .
moreover , by an application of proposition 2.1 @xcite to @xmath30 functions we see that each @xmath326 may be written as a _ countable union of closed sets _ ( including possibly singleton sets ) .
thus , the random sets @xmath292 , from which we are bounding the rate of escape , have a relatively simple topological form . [ ex : rychlik ] this example is borrowed from @xcite ( p746 ) and we refer the reader to the original article for additional details .
it is easy to check that the cocycle @xmath6 described below is rychlik and expanding - on - average .
the base dynamical system is given by a shift @xmath35 on sequence space @xmath327 with transition matrix @xmath328,\ ] ] equipped with the @xmath51-algebra generated by @xmath329-cylinders and the markov probability measure @xmath90 determined by the stochastic matrix @xmath330 .
the map cocycle @xmath101 is generated by maps @xmath331 given by @xmath332 where @xmath333 is a collection of six lebesgue - preserving , piecewise affine , markov expanding maps of the interval , which share a common markov partition ( see figure [ fig : flq09 ] ) .
, reproduced from @xcite ( figure 1).,title="fig:",width=415 ] , reproduced from @xcite ( figure 1).,title="fig:",width=415 ] the map cocycle @xmath6 has been designed so that at each step , a particular ( random ) interval of length 1/3 ( selected from @xmath334,[1/3,2/3]$ ] and @xmath335 $ ] ) is approximately shuffled ( with some escape ) to another of these three intervals . for example , the map @xmath336 approximately shuffles @xmath334 $ ] to @xmath337 $ ] .
_ these particular random intervals are the metastable sets or coherent sets for this random system from which we show the escape rate is slow .
_ , spanning second oseledets subspaces @xmath338 for @xmath339.,width=491 ] a test sequence @xmath191 is obtained in the following way .
let @xmath340 be such that @xmath341 , @xmath342 is the @xmath343 digit in the binary expansion of the fractional part of @xmath196 while @xmath344 is the @xmath345 digit of the same expansion , @xmath346 .
let @xmath347 be such that @xmath348 observe that @xmath349 is three - to - one and that we may uniquely choose @xmath350 that satisfies @xmath351 .
shown below are some of the central elements of @xmath198 , with the zeroth element underlined : @xmath352 it is shown in @xcite that @xmath353 \cup \{\lambda_2(\omega^*)\ } \cup \{0\}$ ] where @xmath354 , approximated using the algorithm on p745 @xcite .
the functions @xmath355 spanning the corresponding oseledets subspaces @xmath356 are shown in figure [ fig : rychliksubspaces ] .
one can see that , when compared to those in figure [ fig : lyapunovfunctions ] , these functions are more regular ( i.e. lower variation ) .
we also determine the random metastable sets or coherent sets @xmath357 for the first eight values on the forward orbit of @xmath358 : @xmath359 , & \quad & { a}_{-}(\omega^ { * } ) & = \left [ 3/9 , 1 \right),\\ { a}_{+}({\vartheta}\omega^ { * } ) & = \left[3/9 , 6/9 \right ] , & \quad & { a}_{-}({\vartheta}\omega^ { * } ) & = \left[0 , 3/9 \right)\cup \left ( 6/9,1 \right],\\ { a}_{+}({\vartheta}^2\omega^ { * } ) & = \left [ 6/9 , 1 \right ] , & \quad & { a}_{-}({\vartheta}^2\omega^ { * } ) & = \left[0 , 6/9 \right),\\ { a}_{+}({\vartheta}^3\omega^ { * } ) & = \left[0 , 4/9 \right ] , & \quad & { a}_{-}({\vartheta}^3\omega^ { * } ) & = \left(4/9 , 1 \right],\\ { a}_{+}({\vartheta}^4\omega^ { * } ) & = \left [ 6/9 , 1 \right ] , & \quad & { a}_{-}({\vartheta}^4\omega^ { * } ) & = \left[0 , 6/9 \right),\\ { a}_{+}({\vartheta}^5\omega^ { * } ) & = \left [ 0 , 3/9 \right ] , & \quad & { a}_{-}({\vartheta}^5\omega^ { * } ) & = \left [ 3/9 , 1 \right),\\ { a}_{+}({\vartheta}^6\omega^ { * } ) & = \left[3/9 , 6/9 \right ] , & \quad & { a}_{-}({\vartheta}^6\omega^ { * } ) & = \left[0 , 3/9 \right)\cup \left ( 6/9,1 \right],\\ { a}_{+}({\vartheta}^7\omega^ { * } ) & = \left [ 0 , 3/9 \right ] , & \quad & { a}_{-}({\vartheta}^7\omega^ { * } ) & = \left [ 3/9 , 1 \right ) .
\end{array}\ ] ] as per the discussion in remark [ rem : conditional ] we can approximate the rates of escape from @xmath107 and @xmath108 by computing the largest lyapunov exponent of the matrix approximations of the corresponding conditional cocycle ( we use @xmath360 and @xmath361 for parameters @xmath362 and @xmath272 in the algorithm on p745 @xcite ) .
we then find that @xmath363 and @xmath364 .
this is in agreement with corollary [ cor : rychlik2 ] as both escape rates are less than the previously computed @xmath365 . by inspecting @xmath366
we see that @xmath367 is mostly mapped onto @xmath368 , @xmath369 .
_ this phenomenon is the cause of the slow escape from the random set @xmath370_. by corollary [ cor : rychlik2 ] , _ the presence of a lyapunov spectral value close to 0 forces the existence of a random set with escape rate slower than that spectral value . _
in this section we use our machinery to obtain some results on random shifts of finite type that exhibit metastability , extending some of the results of @xcite to the random shift setting . for a more detailed description of random shifts of finite type see for example @xcite .
we begin by defining random transition matrices , the corresponding random shifts of finite type and some important properties such as aperiodicity .
we alter some of our notation to match the notation usually applied to shifts . throughout
we assume @xmath103 is an abstract ergodic base dynamical system . for any integer @xmath371 , a _
random transition matrix _ is defined to be a measurable @xmath372 transition - matrix - valued function @xmath373 . for
@xmath18 and @xmath374 write @xmath375 .
let @xmath376 be the space of all two - sided @xmath377-valued sequences and let @xmath378 be a random transition matrix over the ergodic dynamical system @xmath379 .
let @xmath51 denote the left shift on @xmath3 and define @xmath380 the bundle random dynamical system determined by the family of maps @xmath381 is called a _ random shift of finite type_. we will call the space @xmath382 a _ random shift space_. a random transition matrix @xmath383 is _ aperiodic _ if for almost every @xmath18 there exists @xmath384 such that @xmath385 .
if @xmath272 is independent of @xmath16 then @xmath22 is said to be _
uniformly aperiodic_. we also use the terms ` aperiodic ' and ` uniformly aperiodic ' to describe a random shift space @xmath386 .
let @xmath387 \ : \
a_{x_i x_{i+1}}({\vartheta}^i\omega ) = 1 \ \text{for all } 0\le i < n-2\}$ ] be the set of all @xmath139-cylinders of @xmath388 beginning at position @xmath153 .
the following limit exists and is constant @xmath90-almost everywhere : @xmath389 observe that @xmath390 , thus the sequence @xmath391 is subadditive . by kingman s subadditive ergodic theorem ( see e.g. theorem 3.3.2 of @xcite ) there exists a measurable function @xmath392 such that @xmath393 and @xmath394 almost everywhere . as @xmath77 is ergodic , @xmath144 is constant almost everywhere . the quantity @xmath395 is called the _ topological entropy _ of @xmath388 .
denote by @xmath396 the constant where @xmath397 almost everywhere .
@xmath398 for every @xmath399 .
[ prop : blocks ] the proof of this result is largely identical to the proof of its deterministic analogue ( see for example proposition 2.2.12 in @xcite ) .
let @xmath400 and @xmath401 be two random shifts of finite type with common base dynamical system @xmath250 .
@xmath402 is a _ subshift _ of @xmath403 if @xmath404 a subshift may not utilise all the symbols of its parent shift for different values of @xmath16 .
this may be thought of as either a subshift whose alphabet , while finite , changes with @xmath16 or as a subshift on all of the alphabet of its parent shift , but possibly containing isolated vertices in the associated adjacency graph .
we now introduce the notion of a _ complementary subshift_. let @xmath386 be a random shift of finite type and let @xmath405 a subshift of @xmath386 .
the _ complementary _ subshift of @xmath403 to @xmath402 is the subshift @xmath406 whose elements @xmath407 if and only if @xmath408 and @xmath409 for all @xmath410 .
we state a recent extended version of the classical oseledets multiplicative ergodic theorem which guarantees the existence of an oseledets splitting of @xmath411 even when the adjacency matrices @xmath412 are not invertible .
this is the case in many interesting examples , including some shifts .
the met is a central piece of machinery which we use to determine complementary subshifts with large topological entropy .
later we will see that the leading lyapunov exponent @xmath413 yields the topological entropy of the shift , while the second lyapunov exponent @xmath34 , if close to @xmath413 , indicates the presence of metastability and the ability to form complementary subshifts with large entropy relative to that of the original shift .
* theorem 4.1 , specialized to adjacency matrices ) suppose @xmath250 is an ergodic base dynamical system and consider a random transition matrix @xmath373 .
there exists a forward @xmath35-invariant full @xmath90-measure subset @xmath414 , numbers @xmath415 and @xmath416 satisfying @xmath417 such that for all @xmath418 : * there exist subspaces @xmath419 , @xmath420 , @xmath421 ; * @xmath422 for @xmath234 ; * @xmath423 with equality if @xmath424 ; * for @xmath425 the limit @xmath426 exists and equals @xmath427 .
[ thm : flqmatrices ] under the hypothesis of theorem [ thm : flqmatrices ] , for all @xmath234 and all vectors @xmath428 one has @xmath429 . if , in addition , @xmath22 is uniformly aperiodic , then for all @xmath234 and for all vectors
@xmath430 one has @xmath429 .
[ lem : positive_quadrant ] let @xmath431 satisfy @xmath432 . since @xmath433 , we must also have @xmath434 , where the absolute values and the inequalities are taken element - wise .
thus , the leading exponent @xmath413 is achieved by a nonnegative vector , namely , @xmath435 .
suppose first that in fact @xmath436 .
for any @xmath428 , there exist positive constants @xmath228 and @xmath437 such that @xmath438 and therefore for any @xmath374 we have @xmath439 .
we conclude that @xmath440 for all positive @xmath441 .
secondly , we consider the case where @xmath442 is merely non - negative and non - zero .
since @xmath22 is uniformly aperiodic , for every @xmath16 there exists an integer @xmath272 such that @xmath443 is positive and therefore @xmath444 is also positive . using the argument above for positive vectors and the fact that @xmath445 we obtain @xmath446 for all @xmath430 .
[ cor : entropy ] for all @xmath263 one has @xmath447 . from proposition [ prop : blocks ] , clearly @xmath448 , thus @xmath449 . by lemma [ lem : positive_quadrant ]
, this equals @xmath413 for all @xmath234 .
our main result in this section is : let @xmath403 be a uniformly aperiodic random shift of finite type with corresponding random adjacency matrix @xmath450 .
fix @xmath234 .
let @xmath451 , @xmath452 .
for all @xmath114 define @xmath453 and @xmath454 .
let @xmath455 be a random transition matrix such that @xmath456 then @xmath457 is a subshift of @xmath386 and we have @xmath458 . if @xmath459 is the complementary subshift to @xmath402 then @xmath460 .
[ thm : rsmain ] firstly , we will show by induction that for all @xmath461 @xmath462 let @xmath463 denote the decomposition of a vector @xmath441 into nonnegative and nonpositive parts ( dependent on @xmath16 ) .
then we have @xmath464 so holds for @xmath465 . assuming that is true for some @xmath466
, we proceed with the inductive step @xmath467 where we have used the inductive hypothesis to obtain inequality .
thus holds for all @xmath466 and all @xmath468 . noting that for @xmath469 both sides of are positive , we have @xmath470 thus , @xmath471 next we will show that @xmath472 to finally obtain that @xmath473 . we need to have some control over the relative size of the positive and negative parts of @xmath441 along the orbit of @xmath16 . to continue the proof of theorem [ thm : rsmain ] we first state and prove the following claim .
_ claim : _ @xmath474 [ lem : bound ] as @xmath22 is a @xmath475 matrix , then @xmath476 and since it is uniformly aperiodic @xmath477 .
the proof of is by contradiction .
suppose that @xmath478 .
then for every @xmath479 we have @xmath480 therefore @xmath481 is a positive vector , but this is a contradiction because the lyapunov exponent of any positive vector equals @xmath482 .
the inequality @xmath483 is proven similarly .
we continue the proof of the theorem as follows @xmath484 theorem [ thm : rsmain ] may be applied to decompose a metastable random shift space into two complementary random subshifts , each with large topological entropy .
one chooses @xmath485 corresponding to the second largest lyapunov exponent @xmath34 and partitions according to the positive and negative parts of the push - forwards of @xmath441 by the random matrix @xmath22 .
we illustrate this with the following example .
let @xmath486 and let @xmath309 be the full two sided shift on two symbols . consider the random matrix @xmath487 given by @xmath488 where @xmath489 \quad \text{and } { a}_{1 } = \left [ \begin{array}{rrrr } 1 & 1 & 0 & 0\\ 1 & 1 & 1 & 0\\ 0 & 1 & 1 & 1\\ 0 & 0 & 1 & 1 \end{array } \right].\ ] ] we consider a generic point @xmath490 where @xmath491 is the @xmath492 digit of the fractional part of the binary expansion of @xmath196 for @xmath493 and ( @xmath494 for @xmath495 ) .
the first few elements of @xmath198 , with the zeroth element underlined , are given below : @xmath496 using the algorithm on p745 @xcite ( with @xmath497 ) we approximate the largest lyapunov exponent @xmath498 , which , by corollary [ cor : entropy ] , equals the topological entropy of the random shift , that is @xmath499 .
the second lyapunov exponent of this system is @xmath500 .
thus , by theorem [ thm : rsmain ] _ we can decompose the shift @xmath31 into two complementary subshifts , each with topological entropy larger than @xmath501_. moreover , _ the decomposition is given by the oseledets subspaces for @xmath34_. these oseledets subspaces are spans of the vectors @xmath502 , whose graphs are shown in figure [ fig : subspaces ] .
the sets of indices @xmath503 and @xmath504 have the following values on the first four points in the forward orbit of @xmath198 : @xmath505 we construct the matrix @xmath506 of the random subshift according to in theorem [ thm : rsmain ] . on the first three elements of the forward orbit of @xmath358
, @xmath506 takes the following values : @xmath507 , \ \
b({\vartheta}\omega^ * ) = \left [ \begin{array}{rrrr } 0 & 0 & 0 & 0\\ 1 & 0 & 1 & 0\\ 0 & 0 & 0 & 0\\ 0 & 0 & 1 & 0 \end{array } \right],\ \ b({\vartheta}^2\omega^ * ) = \left [ \begin{array}{rrrr } 1 & 1 & 0 & 0\\ 0 & 0 & 0 & 0\\ 0 & 1 & 0 & 0\\ 0 & 0 & 0 & 0 \end{array } \right].\ ] ] similarly the adjacency matrices of the complementary subshift @xmath406 begin : @xmath508 , \ \
b'({\vartheta}\omega^ * ) = \left [ \begin{array}{rrrr } 0 & 1 & 0 & 0\\ 0 & 0 & 0 & 0\\ 0 & 1 & 0 & 1\\ 0 & 0 & 0 & 0 \end{array } \right],\ \ b'({\vartheta}^2\omega^ * ) = \left [ \begin{array}{rrrr } 0 & 0 & 0 & 0\\ 0 & 0 & 1 & 0\\ 0 & 0 & 0 & 0\\ 0 & 0 & 1 & 1 \end{array } \right].\ ] ] the graphs of @xmath405 and @xmath406 for the first four elements of the forward orbit of @xmath358 are shown in figure [ fig : subgraph ] . using the algorithm on p745 in @xcite ( with @xmath361 and @xmath360 ) we estimate the largest lyapunov exponents , and hence the topological entropies , of these two subsifts to be @xmath509 and @xmath510 .
both are larger than @xmath511 , as predicted by theorem [ thm : rsmain ] .
let @xmath205 be a banach space .
a subspace @xmath512 of @xmath227 is said to be _ closed complemented _ if it is closed and there exists a closed subspace @xmath513 of @xmath227 such that @xmath514 and @xmath515 , where ` @xmath516 ' denotes the direct sum ; that is , any non - zero element of @xmath227 can be uniquely written as @xmath517 with @xmath518 and @xmath519 .
a natural linear map on @xmath227 is the _ projection _ onto @xmath513 along @xmath512 , defined by @xmath520 .
the _ grassmannian _ @xmath226 of the space @xmath227 is the set of all closed complemented subspaces of @xmath227 . for any @xmath521 there
exists at least one @xmath522 such that @xmath523 , where ` @xmath524 ' now denotes topological direct sum .
every such @xmath525 defines a _ neighbourhood _ @xmath526 of @xmath527 by @xmath528 .
furthermore , on every such neighbourhood we can define a @xmath529-_local norm _ by @xmath530
this induces a topological structure of a banach manifold on @xmath226 . in particular ,
given a suitable topology on @xmath42 , the continuity of maps @xmath531 is well - defined . in a similar fashion , by taking the corresponding borel @xmath51-algebra @xmath532 and @xmath533 ( or @xmath155 ) , we may also define measurability of such maps .
recall that a function @xmath534 is said to be _ upper semi - continuous _ at @xmath535 if for every @xmath169 there is an open neighbourhood @xmath536 such that @xmath537 for all @xmath538 .
the following lemma is used in the proof of theorem [ thm : equiv_norms ] .
let @xmath50 be a fixed integer and let @xmath539 and @xmath205 be banach spaces where @xmath540 .
the function @xmath541 defined as @xmath542 is upper semi - continuous and therefore measurable .
[ lem : usc ] each @xmath543 is finite - dimensional .
since all norms on finite dimensional spaces are equivalent , the function @xmath269 is well defined and @xmath544 for all @xmath545 . for any @xmath546 ,
let @xmath547 be such that @xmath548 .
for any @xmath549 let @xmath550 be neighbourhood of @xmath551 such that for all @xmath552 , @xmath553 take any @xmath554 .
for any @xmath555 write @xmath556 where @xmath557 and @xmath558
. then @xmath559 and @xmath560 .
now @xmath561 the right hand side converges to @xmath562 as @xmath563 .
as @xmath527 and @xmath173 are arbitrary , this establishes upper semi - continuity of @xmath269 on all of @xmath224 .
m. dellnitz , g. froyland , c. horenkamp , k. padberg - gehle , and a. sen gupta . seasonal variability of the subpolar gyres in the southern ocean : a numerical investigation based on transfer operators . , 16(6):655663 , 2009 . c. schtte , w. huisinga , and p. deuflhard
. transfer operator approach to conformational dynamics in biomolecular systems . in _
ergodic theory , analysis , and efficient simulation of dynamical systems _ ,
pages 191223 .
springer , berlin , 2001 . | we explore the concept of metastability in random dynamical systems , focussing on connections between random perron - frobenius operator cocycles and escape rates of random maps , and on topological entropy of random shifts of finite type .
the lyapunov spectrum of the random perron - frobenius cocycle and the random adjacency matrix cocycle is used to decompose the random system into two disjoint random systems with rigorous upper and lower bounds on ( i ) the escape rate in the setting of random maps , and ( ii ) topological entropy in the setting of random shifts of finite type , respectively . |
it is well known that despite the perturbative nonrenormalizability of general relativity , it is possible to compute gravitational quantum corrections with no assumptions other than those of effective theory @xcite .
however , it is also known that power - counting non - renormalizable theories may in some cases lead to quantum theories which are finite in the uv through non - perturbative dynamics .
one well understood case is the gross - neveu model in @xmath1 dimensions @xcite .
there is a growing amount of evidence that gravity is another theory featuring non - perturbative renormalizability or asymptotic safety @xcite .
the existence of fixed points in higher dimensional gravity @xcite , and their preservation upon coupling to matter @xcite , suggest that extra dimensional models with brane confined matter might also feature asymptotic safety . in this paper
we will implement a framework for computing one - loop observables in the asymptotic safety scenario and compare with previous approaches from effective field theory .
this follows a series of papers looking at calculable and measurable effects of asymptotically safe gravity in models with large extra dimensions @xcite , for example at the lhc @xcite .
recent one - loop computations for the back reaction of gravity on the running of the gauge coupling have devoted a sizeable effort to the issue of gauge dependency in the final result @xcite .
similarly , calculations for precision observables in low scale gravity / brane models have produced some debate on gauge fixing @xcite .
it was found that at the amplitude level these calculations produce a prohibitively large ir enhancement .
these effects are gauge dependent and must disappear when calculated in proper observables @xcite . since the graviton couples universally and ubiquitously , the set of one - loop observables must be carefully quantified .
however , our interest is not in the ir spectrum of gravity , and thus we will follow ref .
@xcite and work exclusively in the dedonder gauge where ir divergences are omitted from the outset . the physical interpretation of this gauge is that brane fluctuations decouple from the gravitational theory entirely and may be studied independently @xcite . using the dedonder gauge
we will compute corrections to the gauge boson propagators and find , as expected , that these are sensitive to the cutoff in both the momentum integral and kaluza - klein ( kk ) summation . at this point
we will implement asymptotic safety in an attempt to quantify possible effects occurring at and above this scale .
it was argued previously that in extra dimensional models the physical scale identified with the renormalization group scale @xmath2 should be the ( @xmath3)-dimensional momentum of a kk graviton @xcite . for the case of asymptotic safety
this leads to finite , cut - off independent tree - level scattering amplitudes .
we will demonstrate through explicit computation that one - loop corrections to gauge boson self energies are also finite in this set - up .
most importantly this means that we have no need to apply an explicit uv cutoff . with a finite calculation for a
well defined observable in our possession , we would like to explore two questions .
first , how does the full computation in asymptotic safety compare with an effective theory approach .
it is not clear _ a priori _ which will lead to stronger constraints as there are two competing effects . in effective theory
the kk integral is cut - off separately in the kk mass @xmath4 and the graviton 4-momentum @xmath5 , and thus ( after an analytic continuation ) may include modes with @xmath6 .
the region outside this circle is suppressed by uv effects in our framework . on the other hand , modes with @xmath7
are not included in the effective theory computation , while these modes will contribute in our computation albeit with the uv suppression provided by asymptotic safety .
the results may then also be compared with constraints on extra dimensions from collider @xcite , astrophysical @xcite , and precision tests of gravity @xcite .
second , we would like to have some sense of the ratio of uv to ir contributions as a function of the number of extra dimensions .
we know that tree - level amplitudes can become uv dominated for @xmath8 and more extra dimensions @xcite . in this paper we will show that at one - loop uv domination may begin at lower dimensions .
our task in this section is , starting from the conventional picture of large extra dimensions , to compute the self energy amplitudes using standard perturbative techniques , and thus reduce the general tensor integral into a closed set of scalar integrals .
we will restrict the brane dimension to strictly ( @xmath9 ) , we will not use dimensional regularization .
thus , we consider a gravitational theory in ( @xmath3 ) space - time dimensions with the coupling to matter @xmath10 where @xmath11 is the fundamental scale of gravity .
the normalization factor matches the notation of ref . @xcite . for a realistic model
we require some mechanism which generates the observed planck mass and generically this will involve a volume factor @xmath12 of the extra dimensions .
thus , in the case of a torus compactification we have the for observed 4-dimensional gravitational coupling @xmath13 .
the utility of these conventions is that the integral and kk sum associated with each graviton loop is @xmath14 which is merely a loop integral over the ( @xmath3)-dimensional momentum .
the sum is shorthand for the individual sums over the kk occupation numbers .
later , we will need the 4-dimensional brane projection of @xmath15 , which we denote as @xmath2 .
the gauge fixing and kk reduction is standard @xcite . on the brane we have simply kk gravitons and kk scalars in the spectrum ,
minimally coupled to the standard model . in the kk reduction ,
one also in general obtains kk vector fields , which do not couple to matter fields at this order in the expansion .
the graviton and scalar propagators are required for our computations and are derived by inverting the kinetic terms of the einstein - hilbert action . in the classical regime we expect the scalar part for both to be @xmath16 , but at high energies we anticipate different behavior .
the vertices coupling the scalar and graviton to the brane confined sm can be obtained by expanding the metric in the massive gauge boson action @xmath17 the sum is over the discrete momenta in the theory . implicitly in this step we have assumed a compact space and enforced periodic conditions .
the ellipses stand for terms , in addition to higher order in @xmath18 , from the vector kk modes and brane fluctuations ( branons ) which do not contribute in our computations @xcite .
we start with the seagull diagram , @xmath19 in agreement with the amplitude for a massless gauge boson . here
, @xmath20 is the external momentum , @xmath15 is the bulk loop momentum , and @xmath21 the scalar part of the graviton propagator .
furthermore , the gravi - scalar contribution in four dimensions is proportional to the trace of the energy - momentum tensor and thus can only couple to the mass term
. however we find in the metric expansion that there is no term of order @xmath22 and thus the only contribution from extra dimensions is eq.([eq : seagull ] ) .
we will not simplify the loop integral at this point .
the rainbow diagram is @xmath23 where @xmath24 contains additionally terms proportional to @xmath25 and @xmath26 .
the exact form is given in the appendix [ append ] .
we have also introduced @xmath2 , the brane only component of the full loop momentum . at this point
we could introduce feynman parameters and perform a cut - off or dimensional regularization .
we expect that the leading uv divergences would cancel with similar terms from the seagull diagram .
however , the amplitude would retain power and logarithmic sensitivity to the cut - off scale @xcite . in both cases
the amplitude is uv sensitive , and the effects of quantum gravity are explicitly cut - off .
as noted in the introduction , the fixed point scaling which we will employ leads to a finite amplitude .
therefore , we will tensor reduce this amplitude without making specific reference to the regulator .
additionally , the amplitude in eq.([eq : rainbow ] ) can be reduced to scalar integrals without any knowledge of the higher dimensional theory . using lorentz invariance on the brane
, we project the tensor integral onto the scalar integrals @xmath27 the notation is meant to evoke the similarity to the normal basis integrals obtained in passarino - veltman reduction @xcite .
the seagull contribution simply reads @xmath28 while the rainbow self energy in terms of the basis integrals is @xmath29 + \eta^{\mu\nu } m_v^4 \left [ 10+\frac{n } { 2+n } \right ] \tilde{b}_0 \right ) .
\label{raintot}\end{aligned}\ ] ] it should be noted that we have not taken into account brane fluctuations , coupling proportional to the brane tension @xmath30 .
it is not clear how one would incorporate this parameter within the framework of asymptotic safety .
moreover , we are physically justified to neglect this contribution if we restrict ourselves to rigid branes or more precisely the energy hierarchy described in ref .
quantum corrections to a massless propagator can be simply accounted for by the wave function renormalization @xmath31 . in particular , we define the renormalization group improved scalar propagator @xmath32 for a massless graviton .
the specific form of @xmath33 encodes , in general in a non - perturbative way , the quantum effects in the theory . for practical purposes
we may only calculate the fixed point behavior of @xmath34 in certain approximations , with a truncated action or at fixed order in perturbation theory . for this study
we will use three approximations for @xmath33 motivated by asymptotic safety .
first , we define the linear @xcite @xmath35^{-1 } \ ; , \ ] ] and quadratic approximation @xmath36 \ ; .\ ] ] the numerical value of @xmath37 is numerically related to @xmath38 , the fixed point value of the dimensionless coupling , which itself carries gauge and cut - off dependency .
therefore , for our purposes @xmath37 is treated as input although related to the fundamental scale @xmath11 by a parameter of @xmath39 . the factor two in eq.([quadnew ] ) is convenient when matching with the one - loop perturbative result @xcite , and thus is left explicit rather than being absorbed into the scale @xmath37 . as opposed to the quenched approximation @xcite @xmath40 \dfrac{1}{q^2 } \ ; \dfrac{\mtrans^{2+n } } { |q|^{2+n } } & q^2 > \mtrans^2 \end{array } \right . \ ; , \label{eq : gprop}\ ] ] the graviton propagator in ( [ linear ] ) and ( [ quadnew ] ) does not contain additional poles and thus has only the simple poles corresponding to the classical propagator .
these modified propagators falls off with sufficient power suppression for large values of @xmath41 , so the contour integral can be wick rotated as usual .
finally , we note that recent studies indicate that the running of the gauge couplings induced by the gravitational coupling vanish once a consistent regulator is chosen which respects the appropriate symmetries @xcite . therefore , while the graviton propagator is strongly altered above @xmath37 , we assume that the gauge boson propagator remains classical
. we can heuristically show the finiteness of our calculation in the asymptotic safety framework .
we apply both the quadratic ( [ quadnew ] ) and quenched ( [ eq : gprop ] ) approximations to account for the large anomalous dimension in the uv .
first of all , we consider the uv portion of the seagull integral in eq.([eq : seagull ] ) , assuming the wave - function renormalization in the the quenched approximation eq.([eq : gprop ] ) .
the momentum integral is then trivial @xmath42 for the rainbow diagram we first write the brane momenta as the 4-dimensional projection of the ( @xmath43)-dimensional momentum @xmath44 .
we will only display an explicit form of @xmath45 at a later point , since it is not required for showing uv finiteness .
considering small values of the external momenta , the leading divergence is @xmath46 once we properly perform the tensor reduction the seagull and rainbow diagrams acquire pre - factors @xmath47 with opposite signs , and the sum of eq.([eq : seaquench ] ) and eq.([eq : rainquench ] ) becomes independent of @xmath48 and hence finite .
it is instructive as well to consider the same integrals for the quadratic approximation . for the seagull
we find @xmath49 we obtain the term in @xmath48 by expanding eq.([eq : seaquad ] ) and find the same leading logarithmic dependence on @xmath48 as in the quenched case .
the rainbow we evaluate using the techniques described above , and when summed with eq.([eq : seaquad ] ) the result is again finite . had either of these computations provided sub - leading divergences of any type , we would not have a sufficiently regulated theory , so the fact that the only divergences are logarithmic serves as evidence of finiteness at one - loop .
we have not mentioned terms in the amplitude proportional to the @xmath25 and @xmath26 . by power
counting the later can not produce a divergence , while the former can at worst admit a term proportional to @xmath50 and not @xmath51 .
this can be seen by examining the momentum structure of the vertices .
these terms are individually finite in asymptotically safe gravity .
the lack of sensitivity to @xmath48 in the self energy amplitude is of course only valid for the dedonder gauge and will not be true for any other choice .
however , also in a less appropriate gauge any physical observable must be independent of @xmath48 . to emphasize this point we outline the related computation in unitary gauge .
the amplitude for the seagull diagram is @xmath52 with a similar , albeit lengthy , expression for the rainbow .
there are no higher uv divergences since upon substitution as described in the following sections the additional terms are proportional only to the ratio of the angular parts , with no @xmath15 dependence
. however , the cancellation of the additional logarithmic terms must also be checked . for genuine physical observables , such as
the one presented later in section [ sec : res ] , this cancellation does indeed take place and the result is finite . to illustrate the geometrical picture of our loop integral we consider the same integral with a ( euclidean ) 2 dimensional brane and a single extra dimension . in momentum space
we can easily picture the integral over brane and bulk momenta as a three dimensional integral in terms of spherical polar coordinates . for a rainbow like diagram we have @xmath53 where we denote the 2-dimensional brane analogy of the full momentum by @xmath54 .
the spherical coordinates we orient such that @xmath20 is located at @xmath55 and the dot product occurs between the brane projection @xmath56 with the brane external momentum @xmath20 .
this integral is evaluated numerically , or in this simple case analytically over the angular coordinates .
the full result in ( @xmath57 ) dimensions is merely a higher dimensional generalization of eq.([eq : twoplusone ] ) .
having established the finiteness of our computation in both the quenched and quadratic approximation , we can numerically evaluate the rainbow and seagull diagrams .
it is clear at this point that we would like to perform the ( @xmath57)-dimensional momentum integral over a radial coordinate @xmath58 , but the loop momentum in the gauge boson propagator depends on only the brane projection of @xmath58 .
thus our integral in ( @xmath57 ) dimensions must retain some angular dependence as the bulk and brane momentum are not interchangeable in the rainbow diagram .
a straightforward solution is to define a ( @xmath57 ) dimensional euclidean vector @xmath59 in polar coordinates . treating the brane momentum as the last entries in this ( @xmath57)-vector allows us to make the variable change @xmath60 which in turn requires the projection @xmath61 where the last step is simply a shorthand notation .
the physical interpretation of the angular variables should be clear as they measure the @xmath62-dimensional projection of @xmath15 .
the configuration with all @xmath63 corresponds to momenta confined to the brane , and the corresponding integral is the conventional 4-dimensional case .
this simplifies our loop integrals and is easily evaluated numerically . under the coordinate change the measure in our ( @xmath57)-dimensional loop integration should be replaced as @xmath64 the pre - factor @xmath65 in front is the result of integrating over the 2-additional brane angular coordinates @xmath66 and @xmath67 which our amplitude does not depend on . for the angular measure related to @xmath45 we write @xmath68 , so
the basis integrals from eq.([eq : scalarint ] ) with the renormalization group improved scalar propagator in eq.([eq : renorm ] ) become @xmath69 } \int dq\ , q^{n+1 } z^{-1}\end{aligned}\ ] ] @xmath70 @xmath71 @xmath72 using these basis integrals we can numerically compute the gauge bosons self energies in eq.([eq : seagull ] ) and eq.([raintot ] ) .
the sum of the two integrals does not depend on @xmath48 . in the next section
we will see how this method of computing the kk integral compares with an effective theory analysis .
in order to constrain new physics based on precision measurements it is typical to use either the oblique parameter set @xmath73 @xcite or the @xmath74 parameterization @xcite .
however , the former is not a good parameter set for gravitons , which in general will modify more than just gauge boson self - energies .
the later can be related to the weak mixing angle @xmath75 to define the leading corrections @xmath76 quantifying the violation of custodial symmetry @xcite .
the individual observables @xmath77 and @xmath78 in the context of gravity require additional computations beyond the self - energies , but the specific linear combination @xmath79 has to be smaller than roughly @xmath80 , assuming a light standard model higgs boson @xcite .
the exact central value depends on the higgs mass , but the uncertainty of @xmath81 also ranges around @xmath80 , which makes this value a generic upper bound on new physics effects .
the effective theory result for gravity contributions @xcite in our notation reads @xmath82 for @xmath83 . here , @xmath84 is the effective theory cut - off scale . in figure
[ fig : res ] we compare this result to our computation as described in the previous section .
the numerical results are similar for @xmath85 , where fundamental planck masses below @xmath86 tev are forbidden . in asymptotically safe gravity
the limits are significantly weakened when we increase the number of extra dimensions , which is not the case for the effective theory computation in eq .
( [ eq : eft ] ) . in other words , once we take the higher dimensional quantum effects seriously there are essentially no limits on large extra dimensions with @xmath87 from electroweak precision data . the explanation for this discrepancy is the following : in effective theory , the amplitudes are dominated by modes with both @xmath2 and @xmath88 near the cut - off , in the corner of a square in the @xmath2 vs @xmath88 plane . however , in our picture of asymptotic safety with @xmath89 this region is outside the central circle which means it is suppressed . for higher numbers of extra dimensions this effect
is more pronounced , as shown by the lessening contribution to @xmath90 as @xmath91 increases .
( -174,100)@xmath92 ( 0,0)@xmath11 ( -61,107)(2,0)2.5 ( -57,107)(2,0)2.5 ( -53,107)(2,0)2.5 ( -45,105)eft ( -61,97)(2,0)10 ( -45,95)fp ( -170,100)@xmath93 ( 0,0)@xmath94 ( -61,107)(2,0)2.5 ( -57,107)(2,0)2.5 ( -53,107)(2,0)2.5 ( -45,105)@xmath95 ( -61,97)(2,0)1 ( -59,97)(2,0)1 ( -57,97)(2,0)1 ( -55,97)(2,0)1 ( -53,97)(2,0)1 ( -45,95)@xmath96 ( -61,87)(2,0)10 ( -45,85)@xmath97 to confirm our observation of an increasing impact from the uv regime we need to study the distribution of the momentum modes contributing to @xmath90 .
more specifically , we would like to know the size of the contribution to the observable for uv momentum modes with @xmath98 . in the quenched approximation
we can estimate this fraction from eq.([eq : gprop ] ) . neglecting terms of order @xmath99
the leading momentum integrals are of the form @xmath100 this expansion in @xmath99 is well validated , for example compared with lhc tree level virtual graviton exchange where @xmath101 can easily become @xmath102 @xcite . for @xmath103 approximately @xmath104 of the combined integral comes from the quantum gravity regime , and for a larger number of extra dimensions this fraction increases .
numerically , figure [ fig : res2 ] shows that the exact results follow the pattern delineated in eq.([eq : simpint ] ) .
the uv region becomes the dominant contribution already for @xmath105 extra dimensions .
this can be contrasted with the case of lhc tree level graviton exchange , where in an identical limit the uv graviton contribution becomes dominant for @xmath106 @xcite .
this is not surprising though as the momentum integral in the tree - level exchange is over four less directions in momentum space , and thus we see the rough equivalence of the @xmath107 case in tree - level with the @xmath108 case at one - loop .
asymptotically safe gravity allows us to compute quantum gravitational effects on relevant observables .
the ultraviolet or quantum gravity regime does not require any modified treatment or cut - off procedure , to ensure finiteness of our predictions . in models with large extra dimensions @xcite gravitational effects
should be measurable , since the fundamental planck scale does not lead to a significant suppression , as compared to other tev scale physics or electroweak loop effects . in this framework ,
asymptotic safety is particularly useful because it provides measurable predictions for the lhc @xcite or , as we have shown in the work , for electroweak precision measurements .
we first introduced a method for evaluating gravitational loop integrals in extra - dimensions . defining these in terms of the full ( @xmath3)-dimensional momentum provided a straight - forward regularization scheme . similar to the usual passarino - veltman reduction into scalar one - loop integrals
, we defined and numerically evaluated a set of basis integrals which can be used for a wide class of observables . in this paper
we studied gravitational effects on custodial symmetry in the standard model , the @xmath0 parameter or @xmath81 .
we find that the bounds from electroweak precision data based on asymptotic safety are roughly equivalent to the direct / indirect bounds obtained by previous methods for @xmath96 extra dimensions . for higher numbers of extra dimensions the bounds from virtual gravitons
are irrelevant compared with more direct measurements . compared to an effective field theory ( or cut - off ) prescription our limits are weaker for more than two extra dimensions
this is consistent with the general observation that for larger numbers of extra dimensions the relative contribution from the trans - planckian ( and hence strongly suppressed ) regime become more and more dominant .
no matter what kind of ultraviolet completion of gravity should be chosen by nature , this paper shows that it has to be modeled properly and quantitatively taken into account .
i am grateful to tilman plehn for continuous encouragement , assisting with the write up of this work and , along with the institut fr theoretische physik at heidelberg university , for on - going hospitality .
moreover , i would like to thank daniel litim , jan pawlowski and einan gardi for helpful discussion and theoretical insights .
in eq.([eq : rainbow ] ) we defined @xmath24 which is given here for convenience .
@xmath109 @xmath110\end{aligned}\ ] ] @xmath111 @xmath112 j. f. donoghue , phys .
d * 50 * , 3874 ( 1994 ) more recent c. p. burgess , living rev .
* 7 * , 5 ( 2004 ) d. j. gross and a. neveu , phys .
d * 10 * ( 1974 ) 3235 .
s. weinberg , _ in * hawking , s.w .
, israel , w. : general relativity * , 790 - 831_. m. reuter , phys . rev .
d * 57 * , 971 ( 1998 ) .
m. niedermaier and m. reuter , living rev .
* 9 * , 5 ( 2006 ) ; m. niedermaier , class .
* 24 * , r171 ( 2007 ) ; d. f. litim , arxiv:0810.3675 [ hep - th ] ; a. codello , r. percacci and c. rahmede , annals phys . *
324 * , 414 ( 2009 ) ; r. percacci , arxiv:0709.3851 [ hep - th ] .
r. percacci and d. perini , phys .
d * 68 * , 044018 ( 2003 ) ; a. codello and r. percacci , phys .
lett . * 97 * , 221301 ( 2006 ) ; d. f. litim , phys .
. lett . * 92 * , 201301 ( 2004 ) ; d. f. litim , aip conf .
proc . *
841 * , 322 ( 2006 ) .
p. fischer and d. f. litim , phys .
b * 638 * , 497 ( 2006 ) ; p. fischer and d. f. litim , aip conf .
proc . * 861 * , 336 ( 2006 ) . c. wetterich , phys .
b * 301 * , 90 ( 1993 ) ; j. berges , n. tetradis and c. wetterich , phys .
rept . * 363 * , 223 ( 2002 ) .
n. arkani - hamed , s. dimopoulos and g. r. dvali , phys .
b * 429 * , 263 ( 1998 ) ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. r. dvali , phys .
b * 436 * , 257 ( 1998 ) ; n. arkani - hamed , s. dimopoulos and g. r. dvali , phys .
d * 59 * , 086004 ( 1999 ) .
e. gerwick and t. plehn , arxiv:0912.2653 [ hep - ph ] ; d. f. litim and t. plehn , phys .
lett . * 100 * , 131301 ( 2008 ) ; j. hewett and t. rizzo , jhep * 0712 * , 009 ( 2007 ) ; e. gerwick , d. litim and t. plehn , arxiv:1101.5548 [ hep - ph ]
. s. p. robinson and f. wilczek , phys .
lett . * 96 * , 231601 ( 2006 ) ; d. j. toms , phys .
d * 76 * , 045015 ( 2007 ) ; d. j. toms , phys .
* 101 * , 131301 ( 2008 ) ; d. j. toms , phys .
d * 80 * , 064040 ( 2009 ) ; d. ebert , j. plefka and a. rodigast , phys .
b * 660 * , 579 ( 2008 ) ; a. r. pietrykowski , phys .
lett . * 98 * , 061801 ( 2007 ) ; y. tang and y. l. wu , arxiv:0807.0331 [ hep - ph ] ; j. e. daum , u. harst and m. reuter , jhep * 1001 * , 084 ( 2010 ) ; i. gogoladze and c. n. leung , phys .
b * 645 * ( 2007 ) 451 .
t. han , j. d. lykken and r. j. zhang , phys .
d * 59 * , 105006 ( 1999 ) .
p. das and s. raychaudhuri ,
arxiv : hep - ph/9908205 .
t. han , d. marfatia and r. j. zhang , phys .
d * 62 * , 125018 ( 2000 ) .
m. l. graesser , phys .
d * 61 * , 074019 ( 2000 ) .
r. akhoury and j. j. van der bij , arxiv : hep - ph/0005055 .
r. contino , l. pilo , r. rattazzi and a. strumia , jhep * 0106 * , 005 ( 2001 ) .
j. a. r. cembranos , a. dobado and a. l. maroto , phys .
d * 73 * ( 2006 ) 035008 .
h. c. cheng , k. t. matchev and m. schmaltz , phys .
d * 66 * , 036005 ( 2002 ) .
s. folkerts , d. f. litim , j. m. pawlowski , [ arxiv:1101.5552 [ hep - th ] ] .
r. sundrum , phys .
d * 59 * , 085009 ( 1999 ) .
a. abulencia _ et al . _
[ cdf collaboration ] , phys .
* 97 * , 171802 ( 2006 ) ; v. m. abazov _ et al . _
[ d0 collaboration ] , phys .
* 101 * , 011601 ( 2008 ) ; k. m. cheung and g. l. landsberg , phys .
d * 62 * , 076003 ( 2000 ) ; d. gerdes , s. murgia , j. carlson , r. e. blair , j. houston and d. berebitsky , phys .
d * 73 * , 112008 ( 2006 ) ; v. m. abazov _ et al . _
[ d0 collaboration ] , phys .
lett . * 102 * , 051601 ( 2009 ) ; l. vacavant and i. hinchliffe , arxiv : hep - ex/0005033 ; l. vacavant and i. hinchliffe , j. phys .
g * 27 * , 1839 ( 2001 ) .
s. cullen and m. perelstein , phys .
lett . *
83 * , 268 ( 1999 ) ; v. d. barger , t. han , c. kao and r. j. zhang , phys . lett .
b * 461 * , 34 ( 1999 ) ; c. hanhart , d. r. phillips , s. reddy and m. j. savage , nucl . phys .
b * 595 * , 335 ( 2001 ) ; s. hannestad and g. g. raffelt , phys .
d * 67 * , 125008 ( 2003 ) [ erratum - ibid .
d * 69 * , 029901 ( 2004 ) ] .
e. g. adelberger , b. r. heckel and a. e. nelson , ann .
nucl . part .
sci . * 53 * , 77 ( 2003 ) ; d. j. kapner , t. s. cook , e. g. adelberger , j. h. gundlach , b. r. heckel , c. d. hoyle and h. e. swanson , phys .
* 98 * , 021101 ( 2007 ) ; e. g. adelberger , b. r. heckel , s. a. hoedl , c. d. hoyle , d. j. kapner and a. upadhye , phys .
* 98 * , 131104 ( 2007 ) .
d. ebert , j. plefka and a. rodigast , jhep * 0902 * , 028 ( 2009 ) .
j. l. hewett , phys .
lett . *
82 * , 4765 ( 1999 ) .
g. passarino and m. j. g. veltman , nucl .
b * 160 * , 151 ( 1979 ) .
m. e. peskin and t. takeuchi , phys .
d * 46 * , 381 ( 1992 ) .
g. altarelli and r. barbieri , phys .
b * 253 * , 161 ( 1991 ) .
j. d. wells , arxiv : hep - ph/0512342 .
r. tenchini and c. verzegnassi , _ hackensack , usa : world scientific ( 2008 ) 419 p _ [ aleph collaboration and delphi collaboration and l3 collaboration ] , phys .
rept . * 427 * , 257 ( 2006 ) ; w. m. yao _ et al . _
[ particle data group ] , j. phys .
g * 33 * , 1 ( 2006 ) ; the lep electroweak working group , http://lepewwg.web.cern.ch/lepewwg/ | asymptotic safety offers a field theory based uv completion to gravity . for low planck scales , gravitational effects on low - energy precision observables can not be neglected .
we compute the contribution to the @xmath0 parameter from asymptotically safe gravitons and find that in contrast to effective theory , constraints on models with more than three extra dimensions are significantly weakened .
the relative size of the trans - planckian contribution increases proportional to the number of extra dimensions . |
it has long been appreciated that localized magnetic moments in metals interact through indirect rkky interactions mediated by the conduction electrons.@xcite in a clean metal , the rkky interaction has a long - range oscillatory part , with an amplitude which decreases as a power law of the distance between the impurities , @xmath0 ( @xmath1 is the dimensionality of the system and @xmath2 the fermi wave vector ) .
this behavior is well understood to be a direct consequence of the existence of a sharp fermi surface characterizing itinerant electrons . in presence of impurities and disorder , this behavior may be substantially modified .
it is well known that sufficiently strong disorder can lead to multiple - scattering processes which can trap the electrons through the processes of anderson localization . in this regime , one may expect that the long - range character of rkky interactions should be suppressed , reflecting the reduced mobility of conduction electrons .
the essential physical question is how this process precisely takes place as the disorder strength is gradually increased and the system approaches the localized regime .
the influence of weak nonmagnetic disorder on the rkky interactions has been studied in considerable detail,@xcite and is by now very well understood .
these studies have established that the main effect of weak disorder is to randomly modify the phase of the rkky oscillations due to impurity - induced phase shifts of the electronic wave functions . as a result ,
the rkky interaction decreases exponentially when averaged over disorder , @xmath3 , where @xmath4 is the mean free path .
early work @xcite thus predicted that the range of rkky interactions becomes essentially cut off at @xmath5 , and can be neglected at larger distances .
more careful consideration @xcite discovered that this naive argument is incorrect , essentially because most relevant quantities do not depend on the average value of the interaction , but instead on the typical value of its amplitude .
since this quantity is not affected by the random phase shifts , it is essentially unaffected by weak disorder , and decreases with the distance @xmath6 in the same power - law fashion as in the clean system .
therefore , the presence of weak disorder is not expected to lead to any significant changes in the physical properties which are dominated by the rkky interactions .
the influence of stronger nonmagnetic disorder on rkky interactions has been , however , so far poorly explored . in an important study , lerner examined the probability distribution of the rkky interactions in the metallic phase in the presence of strong disorder within generalized nonlinear @xmath7 model and @xmath8 expansion.@xcite it was found that the quantum interference corrections do not change the power - law decay of all the even moments of the interaction distribution , which remains the same as in the pure metal , but make the coefficients attached to these moments increase critically with disorder .
as the anderson localization regime is approached , the higher moments increase much faster than the variance , which therefore no longer represents a good characterization of the typical interaction strength .
although this work clearly points out to the importance of strong fluctuations of the rkky interactions , it does not explain how does the distribution function evolve as one enters the regime of anderson localization . within an anderson insulator
the electrons are bound to impurities , and thus can hardly be expected to generate the long - range part of the rkky interaction .
how can we have at the same time large moments of the distribution of rkky interactions , along with its fast decay with the distance ? at first glance
these two arguments seem inconsistent and the situation confusing and paradoxical .
clarifying these issues is an interesting and important problem , since one expects the magnetic correlations to play a crucial role in the physics of disorder - driven metal - insulator transitions in general .
the resolution of this puzzle is in fact quite simple , as we explain in this paper .
we find that in presence of strong disorder and localization , the distribution function develops a strongly non - gaussian form featuring long tails . in such cases
it is well known that all moments of the distribution can assume very large values , while at the same time a typical width of the distribution can remain very small . instead of the arithmetic average ( i.e. standard deviation of the distribution )
we find that the _ typical _ value of the interaction is better characterized by the geometric average of the distribution , @xmath9^{2}% } \rangle}$ ] .
this quantity is exponentially suppressed in the presence of anderson localization , explaining how the long - range part of the rkky interaction is suppressed . to illustrate these ideas and obtain quantitative and reliable results
, we numerically study the distribution of the rkky interactions within the anderson insulator phase .
although our numerical results are obtained within one dimensional model , we argue that the same concept of the typical value of the rkky interaction can be used to physically explain the qualitative change of the form of rkky interactions from long ranged to short ranged during the disorder driven metal - insulator transition in general dimensions . in the rest of this paper we present a detailed description of the model , followed by the numerical results and discussion .
the interaction energy between two local moments @xmath10 and @xmath11 embedded in the metallic host at @xmath12 and @xmath13 , respectively , is given by the hamiltonian @xmath14 where @xmath15 is the zero frequency nonlocal electronic susceptibility and @xmath16 is the exchange coupling constant .
therefore the calculation of the interaction between diluted magnetic impurities reduces to the calculation of the susceptibility which , expressed through the matsubara green s functions , is given by @xmath17 here @xmath18 is the fermionic matsubara frequency , and @xmath19 is the inverse temperature . in the clean case , i.e. in the absence of the nonmagnetic disorder , in three dimensional ( 3d ) system , and for @xmath6 much larger than the lattice spacing , the susceptibility is equal to @xmath20 where @xmath21 and @xmath22 is the effective mass . in the presence of weak disorder
the phase of @xmath23 becomes random and @xmath23 averaged over the disorder configurations is exponentially suppressed @xmath24 where @xmath4 is the mean free path.@xcite the second moment ( variance ) of the probability distribution , however , remains long ranged and has the same power law dependence as in the clean system@xcite @xmath25 ^{2 } \frac{1}{r^{6}}.\ ] ] jagannathan _ _ et al.__@xcite have found that the square root of the forth moment of the distribution , @xmath26 , is comparable in magnitude to the second moment .
therefore , the susceptibility distribution is non - gaussian , but its typical value is well characterized by the square root of its second moment . the same remains true in @xmath27 as well.@xcite this statement is , however , not valid in the regime of strong disorder . as shown by lerner,@xcite using the generalized nonlinear @xmath7 model and performing a @xmath8 expansion , further increase of the disorder results in very rapid increase of all the even cumulants of the distribution .
more precisely , for the cumulant @xmath28 of the order @xmath29 @xmath30 where the parameter @xmath31 as the disorder is increased and the system approaches the anderson transition.@xcite in this case the typical value of the distribution can not be determined by the value of its moments .
in order to examine the form of the rkky interactions in the regime of strong disorder , we proceed to a numerical study .
we consider a tight binding model @xmath32 with nearest neighbor hopping @xmath33 and on - site random potential @xmath34 , which is distributed uniformly in the interval @xmath35 $ ] .
we calculate the interaction between the magnetic impurities embedded into this system at distance @xmath6 . from eqs .
( [ eq1 ] ) and ( [ eq2 ] ) we see that all the information that we need is contained in the single particle green s functions @xmath36 . in the matrix notation the green function @xmath37 and the problem reduces to the numerical summation of the corresponding matrix elements over matsubara frequencies @xmath18 which will be done in the zero temperature limit . in order to obtain good statistics with large number of disorder realizations and
to reduce the finite size effects , we concentrate in the following to a one dimensional system . the numerical results for the clean system , fig . 1 , reproduce the well known oscillatory form of the electronic susceptibility with the power law decay with the distance , @xmath38 . here
the power law exponent is equal to @xmath39 since we are working within one dimensional ( 1d ) model .
( the disorder strength @xmath40 and the chemical potential @xmath41 are measured in units of half the bandwidth @xmath42 , and the distance @xmath6 in units of the lattice spacing .
our system had @xmath43 lattice sites . )
we then consider the susceptibility @xmath44 in presence of weak disorder averaged over hundreds of disorder configurations .
the average susceptibility weakens as the disorder is increased . in fig . 2
we remove the @xmath45 dependence by multiplying the average value @xmath46 by @xmath6 . for weak disorder @xmath47
follows an exponential decay as predicted long time ago by de gennes.@xcite however , what we are really interested in are the probability distributions of the electronic susceptibility in the presence of stronger disorder .
3 shows the probability distribution of the scaled susceptibility @xmath48 in the presence of strong disorder , and for several values of @xmath6 . as in the remaining part of the paper ,
the results are obtained by averaging over hundreds of disorder configurations on the lattice with @xmath49 lattice sites .
the width of the distributions are dependent on @xmath6 .
more importantly , cursory examination of the data indicates that the distributions are distinctly non - gaussian .
this feature is important for the following analysis , since the non - gaussian shape of the distribution prevents us from using the standard deviation as a measure of the the width of the distributions .
instead , we define a typical value of the width as the geometrical average @xmath50 fig . 4 shows a comparison of standard deviation and @xmath51 for describing the width of the same distribution .
we can see that the standard deviation is influenced by the long tails , and is far too large to be a useful description of the width .
@xmath51 , however , gives a good estimate of the width of the distribution .
examining @xmath51 for different strengths of disorder ( figs . 5 and 6 ) , we find that for each value of the disorder @xmath40
@xmath52 for sufficiently large r , where @xmath53 defines the localization length .
therefore , if we define an adjusted susceptibility , @xmath54 the distributions of @xmath55 for large r will be entirely independent of r within each disorder strength .
comparing figs .
3 and 7 shows how the use of this adjusted susceptibility causes these distributions to collapse to a single scaling function .
we then combine the data for several distances @xmath6 , and thereby obtain more precise distributions for each value of disorder @xmath40 . for each disorder strength there now is a single characteristic distribution ( independent on @xmath6 ) as shown in fig .
interestingly , the probability distributions are quite asymmetric in the presence of stronger disorder , ferromagnetic interactions being much more probable than the antiferromagnetic ones . in the strongly localize regime
we expect in fact the interactions to be ferromagnetic for r smaller or of the order of the localization length .
the obtained distributions show this behavior even for r much larger that the localization length .
we then plot the tails of the distributions and find that for very strong disorder they become very long , as shown in fig .
the tails of the distribution appear to converge to a universal power law form in the limit of strong disorder .
the form of the tails is qualitatively the same for positive and negative side of the distribution .
[ fig11 ] the existence of such long tails indicates that the width of the distributions for higher disorder can not be accurately characterized by their moments .
the moments of such a distribution are extremely large , while the typical value is , in fact , very small .
therefore , the long range part of the rkky interactions is strongly suppressed in the strongly localization regime .
we have also compared our result for the mean free path @xmath4 , obtained from eq .
( [ eq4 ] ) , and the localization length @xmath53 , from eq .
( [ eq3.2 ] ) . plotting @xmath4 and @xmath53 as a function of @xmath40 ,
see fig .
10 , we find that @xmath4 and @xmath53 are both proportional to @xmath56 , and find that @xmath57 , which is in a good agreement with the analytical result @xmath58.@xcite this analysis further confirms the consistency of our interpretation of @xmath59 as determined from the decay of the typical rkky interaction amplitude , with the localization length of the electronic system .
in this paper we have examined how the distribution function for rkky interactions becomes modified due to anderson localization effects .
we demonstrated that the essential effect of localization is to exponentially suppress the typical amplitude of rkky interactions on distances longer then the localization length , in agreement with intuitive expectations .
the distribution , nevertheless , remains broad in the sense that it develops long tails which dominate the statistics .
our numerical results thus confirm the analytical predictions of lerner that all even moments of the distribution diverge within an anderson insulator .
our results portray an interesting physical picture with potentially far - reaching consequences . in the metallic regime
the rkky interactions remain long - ranged even in presence of weak disorder , and thus a given magnetic moment effectively interacts with many others . in this regime
one may expect a well developed collective behavior of the spin system , leading to magnetic ordering at low temperature . when the anderson localized regime is approached , the situation is quite different .
the rkky interaction between a typical pair of distant spins is now significantly suppressed or even negligibly small .
very occasionally , a pair of distant spins will interact strongly , due to rare disorder configurations producing long tails in the distribution function .
if the resulting rkky interaction is antiferromagnetic , then such a pair can be expected to lock in a tightly bound singlet thus forming an essentially inert molecule that practically detaches from the rest of the spin system .
such process are precisely what one expects within the random singlet phases @xcite which feature quantum griffiths phase anomalies.@xcite . alternatively
, a ferromagnetic interaction will lead to the formation of a bound triplet state , essentially a spin @xmath60 magnetic moment .
it is interesting to note a degree of asymmetry of the distribution of the rkky interactions , that we have found at strong disorder .
this finding seems to indicate that ferromagnetic correlations may effectively compete with the tendency for singlet formation , possibly leading to nano - scale ferromagnetism@xcite coexisting with a random singlet phase .
we thus anticipate that anderson localization processes generically destabilize spin glass ordering in disordered metals , which is instead replaced by an appropriate quantum griffiths phase . precisely how these processes take place in the vicinity of realistic metal - insulator transitions remains a fascinating open direction for future study .
we thank e. abrahams , r. bhatt , s. chakravarty , e. miranda , and b. spivak for useful discussions . this work was supported by the national high magnetic field laboratory , nsf grants dmr-0234215 ( v.d . and d.t . ) and dmr-0542026 ( v.d . ) , center for integrating research and learning stuff ( j.a.s . ) , and serbian ministry of science , project no .
141014 ( d.t . ) . | we study the influence of strong nonmagnetic disorder on the ruderman - kittel - kasuya - yosida ( rkky ) interactions between diluted magnetic moments in metals .
we find that the probability distribution for the rkky interactions assumes strongly non - gaussian form featuring long tails .
since such distributions can not be characterized by its moments , we define a _ typical _ value of the interaction amplitude , which we find to be exponentially suppressed in presence of anderson localization .
our results present a plausible and physically transparent picture describing how anderson localization effectively eliminates the long range nature of the rkky interactions . |
in recent years , multi - tip scanning tunneling microscopes ( stm ) have become popular instruments @xcite .
the most common experiments reported in literature using those instruments are multi - probe measurements that allow to determine the microscopic electrical properties of the samples under investigation .
however , such devices can be further extended with scanning tunneling potentiometry ( stp ) techniques which allow to map the electric potential on the nanoscale with spatial and potential resolution of few and v , respectively @xcite .
combination of stp with multi - probe point measurements creates a powerful tool for the electrical sample characterization capable to give insight into fundamental transport properties , such as the influence of defects on the local electric transport .
this applies especially for surface dominated transport as present e.g. in topological insulators and allows the investigation of transport phenomena such as the landauer dipole @xcite . since the invention of stp by murlat and pohl @xcite several different technical implementations of stp
have been reported mostly in single - tip stm setups @xcite for which the transport field is usually applied to macroscopic sample contacts up to several mm away from each other . as a result ,
the current density in the region of measurement can be low , resulting in a bad signal - to - noise ratio of the measured electric potential distribution throughout the sample surface . to circumvent this problem , patterning metallic contacts on the sample surface
was reported to increase the current density across the structure under investigation @xcite . however , the patterning process often results in contamination of the sample surface .
in contrast , multi - tip stm setups have the advantage of flexible _ in situ _ contacting the sample with several electrodes , e.g. molecular beam epitaxy grown films and nanostructures without additional sample processing @xcite . as a result , by variation of the spacing of the electrodes the current density in the surface and also through the bulk
can be measured and controlled @xcite .
furthermore , the direction of injected current can be varied by repositioning of the electrodes enabling the quantification of transport anisotropies @xcite .
multi - tip stp implementations have been reported that facilitate the above mentioned advantages over single - tip setups @xcite . however , these implementations are based on additional hardware such as external power sources and feedback electronics that need to be added to existing setups in order to enable potentiometry measurements @xcite .
in contrast , we present here a software based multi - tip stp technique that can be readily implemented into existing multi - tip setups without installation of additional hardware .
we use a home - built room temperature four - tip stm in besocke / beetle - based design with electrochemically etched tungsten tips .
each tip has an individual ring for xy - motion and a tube piezo for z - motion @xcite .
atomic resolution on si@xmath1 is achieved with each tip .
the stm is equipped with a scanning electron microscope ( sem ) for precise positioning of the individual tips .
each of the four tips is connected to one of four instances of equivalent electronics which can be switched between current or voltage measurement , as also done in other four - tip stm setups @xcite . for current measurement ,
the voltage applied at each tip can be set individually .
furthermore , the sample can be either on floating or ground potential . in stp measurements three of the four tips are in current probe mode , two current injection tips ( tip 1 and 2 ) and the scanning tip ( tip 3 ) .
the fourth probe is operated in voltage probe mode and is used as a reference voltage probe ( tip 4 ) .
this tip is optional such that the presented implementation is also applicable for three - tip stm setups .
a schematic of the setup is shown in fig . [ 1 ] .
the four biasing electronics units are connected to a commercial multi - tip stm controller with individual digital - to - analog converters ( dac ) outputs to address the bias voltage for each tip and analog - to - digital converters ( adc ) to read out the current or voltage at each tip .
figure [ 2 ] shows a schematic of the biasing electronics which is typical for a multi - tip setup @xcite .
the present setup is based on a femto dlpca-200 current amplifier @xcite and allows variable - gain current measurement or voltage measurement . for current measurement
the tip is connected to the input of the current amplifier which is floating on bias potential @xmath2 . from the output of the current amplifier ,
the bias voltage is subtracted before the signal is fed to the output @xmath3 . for voltage measurement
the tip is connected to a high input impedance ( @xmath4 ) voltage follower with its output connected to the @xmath5 output of the electronics . in high resolution stp measurements , the control of @xmath2 of the scanning tip in the range of v
is required .
this can be achieved by a voltage divider ( e.g. @xmath6 ) which is applied just in front of the @xmath2 input of the scanning tip biasing electronics .
this decreases the dynamic range of the voltage which can be applied to this tip accordingly . not to be limited by this ,
the topography tip - sample tunneling bias voltage is applied to the sample rather than the tip as described below .
is applied via the differential input in front of which the optional voltage divider for the scanning tip is located . ]
is plotted as a function of the lateral coordinate @xmath7 . during topography feedback , the solid green line holds . for potentiometry feedback @xmath8
is shifted by @xmath9 resulting in the dashed blue line . moving the scanning tip by an amount @xmath10 results in a change of the potential under the tip by @xmath11 .
note : the movement of the scanning tip is largely exaggerated in this graph . ]
the multi - tip stp implementation presented here is based on the interrupted feedback loop single - tip stp implementation reported in ref . .
in this technique the surface is scanned point - by - point as in regular stm measurements .
however , at each scanning point after the topography value @xmath12 is acquired , the tip is held at constant height and the bias voltage between scanning tip and sample is shifted .
now the potentiometry feedback is run for a time @xmath13 .
the final value of the potentiometry feedback is saved as the sample potential at the momentary position of the scan @xmath14 . after restoring the topography bias voltage ,
the topography feedback is re - enabled and the scan is continued by moving to the next scan position @xcite .
the flow chart of the measurement procedure is shown in fig . [ 3 ] . in this way topography and potentiometry signal
are acquired point - by - point across the sample surface .
depending on the feedback time @xmath15 and the number of scan points a whole scan takes up to several hours .
the potentiometry feedback is a pi - feedback algorithm just like the one used for the topography feedback .
however , it adjusts the tip voltage as control variable instead of the tip height ( in topography feedback ) and the process variable set point ( i.e. the tunneling current ) is set to @xmath16 during the potentiometry feedback .
the corresponding feedback constants determine the bandwidth of the potentiometry feedback and have to be adjusted for each measurement .
in order to perform a potentiometry measurement , first the sample ( connected to ground ) needs to be contacted by the two current injecting tips ( tip 1 and 2 ) .
after coarse approaching all four tips to the sample surface , the current injecting tips are brought into contact to the surface under tunneling conditions of e.g. @xmath17 ; @xmath18 .
one after the other , the current injection tips are advanced further by a few nm with disabled topography feedback until a stable current ( @xmath191% variation per minute ) of up to several a is measured between that tip and the sample .
typically , there is a sharp increase of the current as the tips come into contact to the surface after which the tips should not be advanced further to prevent damage to the tips or the sample surface .
however , if at this point the resulting current is unstable , advancing the tips a bit further can help to stabilize the contact .
next , the optional reference voltage probe ( tip 4 ) is brought into contact to the sample surface in the same way ( in current measurement mode ) .
when a stable contact is achieved , the electronics of this tip is switched to voltage measurement .
this tip serves to measure the reference voltage which can be used during the measurement preparations as described below but also to compensate small fluctuations in the current injecting contacts during measurements @xcite : changes of the contact resistance in the current injecting contacts result in a change of the sample potential at the point of current injection .
this will directly influence the measured potential at the position of the scanning tip and can be compensated by subtracting the potential of a static reference voltage probe from the measured potential of the scanning tip .
now the sample backside contact is disconnected such that the electric potential of the sample depends only on the potential of the two current injecting tips . to approach the scanning tip ( tip 3 ) ,
the voltage of both injection tips is set to 0 v while that of the scanning tip is set to e.g. @xmath17 .
then the tip is approached to the sample surface until it is in tunneling contact ( @xmath20 ) .
subsequently , a topography scan should be performed to check if the scan area is suitable for potentiometry measurements . to apply a transport field , first the scanning tip ( tip 3 ) is retracted and its voltage is set to @xmath21 .
now the voltage of the injection tips is set such that a lateral current @xmath22 a results in a potential distribution @xmath14 over the sample surface .
subsequently , @xmath14 is shifted until the sample potential under the scanning tip is approximately @xmath23 .
this can be done using the information of the voltage probe . in order to adapt @xmath14 at the position below the scanning tip , the voltage of both current injecting tips is shifted identically , maintaining the injected current @xcite .
next , the bias voltage for the topography measurement @xmath9 is added to the voltage of the current injecting tips . for high resolution measurements the voltage divider
is now applied to @xmath2 of the scanning tip . following the described preparations , the tunneling contact of tip 3 is re - engaged and the stp scan can be started . figure [ 4 ] shows a schematic of the potentiometry measurement setup with the voltages applied to the tip and sample indicated , respectively . in stp , the voltage difference between the scanning tip and sample @xmath24 usually referred to as bias voltage at the scan position @xmath25 is the superposition of the lateral transport field , the offset voltage and the momentary voltage applied to the tip during topography feedback @xmath26 ( solid green line in fig . [ 4 ] ) . when starting the stp scan , after acquiring the topography at the momentary scan position , the topography feedback is deactivated and the injection tip voltages are shifted by @xmath27 leaving only @xmath14 applied to the sample resulting in @xmath28 ( dashed blue line in fig .
subsequently , the potentiometry feedback loop is activated and adjusts @xmath2 to the local sample potential under the tip @xmath29 .
after the feedback time @xmath15 , the potentiometry feedback is deactivated .
@xmath2 is recorded as the sample potential at this scan position and is maintained until the next activation of the potentiometry feedback loop . after restoring the topography bias voltage @xmath26 ,
the topography feedback is re - activated and the tip is moved to the next scan position . note that , for neighboring scan positions ( @xmath30 ) , the local sample potential varies by an amount of @xmath11 during the topography feedback . in detail , during the topography feedback it holds @xmath31 . however , in usual measurements @xmath32 such that the change in @xmath8 for neighboring scan positions can be neglected @xcite . for the potentiometry feedback to work in an optimal way it is recommended that the tunneling @xmath33 curve of the scanning tip is linear , such that a change in the control variable ( tunneling voltage ) will have a linear effect on the process variable ( tunneling current ) @xcite .
furthermore , due to the fundamental noise limit of the potential measurement which is the sum of thermal ( johnson ) noise and the amplifier noise the limit of the potentiometry resolution is @xcite @xmath34 with @xmath35 and @xmath36 here @xmath37 is the noise level of the current amplifier @xcite .
low tunneling voltages and high tunneling current setpoint depending on the bandwidth @xmath38 of the setup are necessary for low noise potentiometry measurement @xcite . for transport measurements with several mv of voltage drop across the scan area , we determine @xmath39 as the feedback loop speed for which a good signal - to - noise ratio is achieved while a relatively high scanning speed is maintained . for this bandwidth ,
the noise limit for three different tunneling resistances is given in table [ table 1 ] . .noise
limit for transport measurements in the present setup at three different tunneling junction resistances and a bandwidth of @xmath39 .
[ table 1 ] [ cols=">,>",options="header " , ]
( a ) topography of an island on the ag / si@xmath0 surface at @xmath40 and @xmath41 with a domain boundary indicated by the green arrow . ( b )
the corresponding potential map shows an enhanced potentiometry signal at the edges of the islands and domain boundaries .
the potentiometry feedback time is @xmath42 at a bandwidth of @xmath43 .
( c ) topography of an island with the scanning tip under laser irradiation .
( d ) corresponding potential map of ( c ) . the amplitude of the potentiometry signal is increased by a factor of around 10 compared to ( b ) .
( e ) potential distribution of a featureless surface area ( black square in ( b ) ) .
the standard deviation is @xmath44v .
( f ) section along the black line in the inset that corresponds to the black rectangle in ( d ) . a standing wave pattern with a periodicity of @xmath45 is observed .
the section is averaged over 3 lines . ]
the ag / si@xmath0 reconstruction is a well - studied sample system when it comes to scanning tunneling microscopy , spectroscopy and also potentiometry @xcite .
therefore , this system can be used as a benchmark system to test the performance of the presented stp implementation . for the preparation of the ag / si@xmath0 surface , n - si(111 )
@xmath46 wafers were flash annealed and an equivalent amount of @xmath47 of ag was deposited at @xmath48 following the procedure outlined in ref . .
the surface of the prepared sample is known to have a high 2d conductivity while the bulk of the sample is almost insulating compared to the surface @xcite . as a first test
, the performance of the potentiometry feedback is demonstrated by measurement of thermoelectric effects .
the thermovoltage between the scanning tip and the sample surface is usually undesired in transport measurements and the identification and reduction of these effects gives a good first insight in the investigated sample system and hinder unwanted effects in later transport measurements . to measure the thermovoltage , a potentiometry scan is performed during which the sample is on ground potential and no transport field i.e.
lateral current is applied .
the result of such a scan is shown in fig . [ 5 ] ( a ) and ( b ) as topography and potential map , respectively . in the topography
, we observe a one atomic layer high island and domain boundaries between different domains of the ag / si@xmath0 reconstruction on the otherwise flat surface .
the corresponding potential map shows an increased potentiometry signal located at the step edges of the island and the domain boundaries , in agreement with the literature @xcite .
the minimum resolution of the potentiometry is determined as the standard deviation of the measured potential of a featureless surface area ( square in fig . [ 5 ] ( b ) and respective histogram in fig . [ 5 ] ( e ) ) , in this case @xmath44v .
note that for this measurement a low bandwidth of @xmath49 was used resulting in a fundamental noise limit to @xmath50v . in a next step , to verify the observation of the thermovoltage features the tip was heated by a laser ( @xmath51 ) and the measurement was repeated .
a strong increase in the potentiometry signal by a factor of around 10 is observed , confirming its nature to be due to thermoelectric effects as seen in the topography and potential map in fig . [ 5 ] ( c ) and ( d ) , respectively @xcite . in the potential map ( fig . [ 5 ] ( d ) ) oscillations in the sample potential can be observed .
these oscillations appear pronounced in fig . [ 5 ] ( f ) and can be identified as a standing wave pattern which can be explained by thermoelectric effects @xcite . when two materials are in tunneling contact , the thermovoltage present at the tunneling junction is proportional to the difference of the squared temperatures of the two materials @xcite .
the leading term of the local thermovoltage is given by @xmath52 where @xmath53 and @xmath54 are the temperature of the tip and sample , respectively .
the observed standing wave pattern result from modulations of the local density of states in the sample @xmath55 and were below the resolution of the setup without laser heating of the tip @xcite .
the resolution of a standing wave pattern confirms the proper function of the potentiometry feedback .
to perform transport measurements a lateral current is applied through the sample surface .
a peculiarity of the ag / si@xmath0 surface prepared as described above are the 3d bulk ag islands visible in the sem images ( fig .
[ 6 ] ( a))@xcite . the bulk ag island thickness is much larger than that of the actual ag / si@xmath0 reconstruction and can be easily contacted allowing a stable injection of a lateral current . as a result , for this measurement
the optional voltage reference probe is not used . to contact the islands the current injecting tips
are approached into tunneling contact and then advanced by few tens of nm with the tunneling feedback deactivated pressing the tips into the islands .
subsequently , the scanning tip is approached to the sample surface as described above and a potentiometry scan is performed .
the results of the stp measurements for two opposing directions of the injected current are shown in fig .
[ 6 ] ( b - e ) . the topography of the ag / si@xmath0 surface ( fig .
[ 6 ] ( b ) ) is dominated by approximately parallel monoatomic substrate steps ( @xmath56 ) with an average terrace width of @xmath57 .
one atomic layer high islands are observed on the terraces .
the cross sections in fig . [
6 ] ( e ) shows large voltage drops at positions of steps in the topography . across the terraces
an almost linear voltage gradient can be observed with no direct influence of the ad - islands on the voltage drop .
a detailed analysis shows that domain boundaries between different domains of the @xmath58 reconstruction provide the largest contribution to the voltage drop across the terraces besides the gradient resulting from the undisturbed surface .
this observation is in agreement with literature @xcite . for reverse current directions
we observe an inverted voltage drop indicating a symmetric current response .
the given sample topography of a parallel step array allows a direct evaluation of the step resistivity and terrace resistivity which was not the case in previous stp studies of the ag / si@xmath0 system @xcite . from the potential cross sections ( averaged over 3 lines )
the average voltage drop located at the topographic steps is determined to be @xmath59 .
the current density at the position of the scanning tip is @xmath60 which is determined from the measurement geometry where @xmath61 m is the distance of the injection tips and @xmath62 the absolute current injected into the sample @xcite . from this
, the step resistivity can be determined to be @xmath63@xmath64 .
the voltage drop across the terraces is determined from the residual resistance after subtracting the sum of the step resistances from the total resistance and gives @xmath65 corresponding to a resistance of @xmath66 .
these values are in agreement with previous values from the literature @xcite . for the present sample ,
the steps contribute 91% of the voltage drop across the ag / si@xmath0 surface while the terraces contribute only to 9% . .
( a ) topography of the surface showing parallel steps with an average terraces width of @xmath67 .
( b ) corresponding potential image .
a sharp voltage drop is observed at the step edges and a small voltage drop across the terraces .
( c ) line profiles indicated in ( a ) and ( b ) .
the average voltage drop at the steps is @xmath68 .
the average slope of the terraces is @xmath69 .
the inset shows the sem image of the tip setup on the si@xmath1 surface .
the distance between the current injecting tips 1 and 2 is @xmath70 .
tunneling conditions are @xmath71 and @xmath72 . ]
the si@xmath1 reconstruction is frequently reported in literature since the beginning of stm studies but is still subject of current investigations @xcite . up to
now there are no stp transport studies on this surface reported in literature . for the measurements
, we have used the same n - si(111 ) @xmath46 substrates as above .
the two current injecting tips were brought carefully into direct contact with the freshly prepared si@xmath1 reconstructed surface ( flash annealed to @xmath73 ) in order to keep the si@xmath1 terminated surface as intact as possible . as a result , the reference voltage probe is again not necessary .
in contrast to the ag / si@xmath0 the si@xmath1 reconstruction on n - si(111 ) @xmath46 can not be regarded as a purely 2d conducting material : the bulk and space charge layer of the underlying substrate can have a significant contribution to the measured conductivity @xcite . in order to achieve accurate results using stp ,
a detailed analysis of the conduction through the surface and bulk is mandatory .
such an analysis can be performed by variable distance four - probe measurements as shown recently @xcite .
a combined analysis of the sample ( four - probe measurements + stp ) can be performed with the setup presented .
for a current injection tip distance of @xmath74 m , at the position of the scanning tip an amount of 69(6)% of the lateral current is transmitted by the surface conductivity associated with the @xmath75 reconstructed sample surface @xcite .
the resulting current density at the position of the scanning tip is @xmath76 .
next , the step and terrace conductivities can be determined as before for the ag / si@xmath0 sample .
the corresponding measurements are shown in fig .
[ 7 ] . in the case of si@xmath1
we observe a variation of the voltage drop at different steps .
we attribute this to variations in the shapes and sizes of the terraces and also domain boundaries which results in changes of the current flow paths .
the average voltage drop at the steps is determined to be @xmath68 and across the terraces @xmath77 .
the resulting resistivity is @xmath78 and @xmath79 , respectively while the average step distance in the measurement is @xmath67 .
these results are in agreement with previously reported results @xcite . for this surface
the steps contribute 82% of the measured surface resistance while the terraces contribute 18% .
we present a multi - tip scanning tunneling potentiometry realization which can be implemented into present multi - tip scanning tunneling microscopes without hardware changes .
the resulting setup allows _ in situ _ contacting and stp measurements with flexible positioning of two current injecting electrodes such that the current density in the sample surface and the current direction with respect to the sample under investigation can be controlled with no need of contact patterning . to map the local potential of the sample on the nanoscale
, a third tip is scanning the sample surface between the current injecting electrodes by an interrupted feedback technique , recording the local topography and electric potential .
a fourth tip is optional and can be used as a reference voltage probe .
thermovoltage measurements on the ag / si@xmath0 surface demonstrate the performance of the potentiometry feedback by resolving a standing wave pattern .
the resolution of the presented setup is determined to be @xmath44v .
the ability to map the local transport field is shown on ag / si@xmath0 where sharp voltage drops located at step edges and symmetrical potential characteristics are observed for reverse current directions . from the investigated sample geometry of a 2d conductor with a parallel step array , conductivities of the terraces and steps can be extracted easily .
this straightforward method can further be applied to samples with similar geometry . for mixed systems of 2d and 3d conduction
, a combination of four - probe measurements together with stp allows to first disentangle the bulk and surface conductivity and then determine the contribution of the terraces and steps on the sample surface to the measured conductivity .
this approach is demonstrated and gives a comprehensive view of the conduction in si@xmath1 on the micro and nanoscale and can be further applied to a wide range of samples that are not limited to pure 2d conductors on an insulating bulk as the case in previous stp studies .
19ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1063/1.2735593 [ * * , ( ) ] link:\doibase 10.1063/1.2821610 [ * * , ( ) ] link:\doibase 10.1002/adma.201200257 [ * * , ( ) ] link:\doibase 10.1063/1.2968111 [ * * , ( ) ] link:\doibase 10.1063/1.1140428 [ * * , ( ) ] link:\doibase 10.1063/1.3469809 [ * * , ( ) ] link:\doibase 10.1006/spmi.1997.0533 [ * * , ( ) ] link:\doibase 10.1038/ncomms7399 [ * * , ( ) ] link:\doibase 10.1063/1.96491 [ * * , ( ) ] link:\doibase 10.1063/1.2953097 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.236802 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.066801 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.91.036805 [ * * , ( ) ] link:\doibase 10.1063/1.3694990 [ * * , ( ) ] http://www.mprobes.com/ [ `` '' ] ( ) http://www.femto.de/en/ [ `` '' ] ( ) due to the voltage spikes occurring when the potentiometry bias voltage is set before the potentiometry feedback and when the topography bias voltage is set before re - activating the topography feedback a short delay time of a few ms is recommended to be implemented respectively to allow the system to settle . if the voltage probe is not present , it is assumed that the transport field is symmetrical , with the option to correct the voltages later .
link:\doibase 10.1143/jjap.51.125202 [ * * , ( ) ] surface photoelectric effects were ruled out by switching off the laser for a short time and still observing the increased potentiometry signal .
link:\doibase 10.1103/physrevb.42.9214 [ * * , ( ) ] link:\doibase 10.1038/nmat3170 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.246802 [ * * , ( ) ] | we present a multi - tip scanning tunneling potentiometry technique that can be implemented into existing multi - tip scanning tunneling microscopes without installation of additional hardware .
the resulting setup allows flexible _ in situ _ contacting of samples under uhv conditions and subsequent measurement of the sample topography and local electric potential with resolution down to and v , respectively .
the performance of the potentiometry feedback is demonstrated by thermovoltage measurements on the ag / si@xmath0 surface by resolving a standing wave pattern .
subsequently , the ability to map the local transport field as a result of a lateral current through the sample surface is shown on ag / si@xmath0 and si@xmath1 surfaces . |
discovered almost 20 years ago , high-@xmath0 copper oxides still remain a challenge of the modern condensed matter physics .
it is not only due to unconventional superconducting state with a highest superconducting transition temperature @xmath0 ever observed .
also they reveal evolution from an undoped antiferromagnetic ( afm ) insulator to an almost conventional , though highly correlated @xcite , fermi liquid system at the overdoped side of the phase diagram . between these two regimes
the system exhibits strongly correlated , or so - called `` pseudogap '' metallic behavior up to an optimal doping concentration @xmath6 .
recent significant improvements of experimental techniques , especially of the angle - resolved photoemission spectroscopy ( arpes ) and the scanning tunneling microscopy ( stm ) , revealed many exciting features of this doping dependent evolution .
first of all , the fermi surface ( fs ) at low doping concentrations @xmath2 has been measured @xcite . together with the previous measurements on optimally and overdoped samples ( see e.g. review @xcite and references therein ) , these observations provide a unified picture of the doping dependent fs , which changes from the `` fermi arcs '' @xcite in the underdoped compounds to the `` large '' fermi surface in the overdoped systems . though this change is smooth , it occurs around the optimal doping concentration .
also , the observed evolution is consistent with the hall coefficient @xmath7 measurements @xcite .
the pseudogap behavior observed in arpes is also tracked in the transport measurements .
in particular , the resistivity curvature mapping over @xmath8 phase diagram clearly demonstrates crossover between underdoped and overdoped regimes @xcite .
and the in - plane resistivity @xmath9 shows @xmath10-linear dependence only around @xmath11 .
the drastic change of the quasiparticle dynamics around the optimal doping was found by the time - resolved measurements of the photoinduced change in reflectivity for bi@xmath12sr@xmath12ca@xmath13dy@xmath14cu@xmath12o@xmath15 ( bscco ) @xcite .
namely , the spectral weight shifts expected for a bcs superconductor can account for the photoinduced response in overdoped , but not underdoped bscco .
this agrees with the observed difference of the low - frequency spectral weight transfer in normal and superconducting states on under- and over - doped samples @xcite .
meanwhile , the integral characteristics of the system demonstrate more smooth behavior upon increase of the doping @xmath2 .
the dependence of the chemical potential shift @xmath16 on @xmath2 shows pinning at @xmath17 , and evolves smoothly at higher doping concentrations @xcite .
the measured nodal fermi velocity @xmath18 is almost doping - independent within experimental error of 20% @xcite .
the experiments involving combination of dc transport and infrared spectroscopy revealed an almost constant effective electron mass @xmath19 in the underdoped and slightly overdoped la@xmath20sr@xmath21cuo@xmath22 ( lsco ) and yba@xmath12cu@xmath23o@xmath14 ( ybco ) @xcite .
low-@xmath2 effective mass dependence contradicts predictions of the brinkman - rice metal - insulator transition theory @xcite , predicting divergence of the @xmath24 at the point of transition .
one of the main drawbacks in this theory is that the magnetic correlations were neglected .
thus the discrepancy with the experiment emphasizes the importance of these correlations in high-@xmath0 copper oxides . from the theoretical point of view
, the description of the crossover between almost localized picture and the fermi liquid regime is very difficult .
starting from the fermi liquid approach one may discuss the overdoped and , partly , optimally doped region , while for underdoped and undoped samples this approach is not applicable .
the strong - coupling gutzwiller approximation @xcite for the hubbard model provides a good description for the correlated metallic system .
this approximation is equivalent @xcite to the mean - field saddle - point solution within a slave - boson approach @xcite . at the same time , as shown within @xmath25 expansion @xcite , with @xmath26 being the dimensionality of the lattice , the gutzwiller approximation is equivalent to the case of @xmath27 .
obviously , for the quasi - two - dimensional systems such as high-@xmath0 copper oxides this is not a proper limit .
the same applies to the dynamical mean - field theory ( dmft ) @xcite , which is exact only for @xmath27 . in this limit
the short - range magnetic fluctuations are excluded .
it is not a good starting point for the system with long - range afm order at low @xmath2 and short - range afm correlations in the underdoped region . to describe the doping dependent evolution of the low - energy excitations we develop a strong - coupling mean - field theory for the paramagnetic non - superconductive phase of high-@xmath0 copper oxides starting from the local limit . to go beyond the usual hubbard - i approximation a self - consistently calculated static spin - spin and kinematic correlation functions
are taken into account .
within this approximation in the framework of the effective @xmath1 model with _ ab initio _ calculated parameters we obtain a doping - dependent evolution of the fs , effective mass and nodal fermi velocity .
the analysis of the low - energy excitations behavior for p - type cuprates yields two quantum phase transitions associated with the change of the fs topology . for n - type cuprates
we observe only single quantum critical concentration .
the key aspect in these findings is an adequate description of the electron scattering by the short range magnetic fluctuations accompanied by the three - site correlated hoppings .
the paper is organized as follows . in section [ section : model ] the effective model and the approximations
are described .
the results of the calculations for p - type cuprates are presented in section [ section : ptype ] .
also , the critical comparison of the @xmath1 and @xmath28 models is made , and the role of short range magnetic order is discussed .
section [ section : ntype ] contains results for n - type cuprates .
the last section summarize this study , and the main points are discussed .
high-@xmath0 cuprates belong to a class of strongly correlated systems where the standard local density approximation ( lda ) schemes and the weak - coupling perturbation theories yield an inappropriate results . to overcome this difficulty recently we have developed an lda+gtb method @xcite . in this method the _ ab initio _ lda calculation is used to construct the wannier functions and to obtain the single electron and coulomb parameters of the multiband hubbard - type model . within this model
the electronic structure in the strong correlation regime is calculated by the generalized tight - binding ( gtb ) method @xcite .
the latter combines the exact diagonalization of the model hamiltonian for a small cluster ( unit cell ) with perturbative treatment of the intercluster hopping and interactions . for undoped and weakly doped lsco and nd@xmath20ce@xmath21cuo@xmath22 ( ncco ) this scheme results in a charge transfer insulator with a correct value of the gap and the dispersion of bands in agreement with the experimental arpes data ( see ref .
@xcite for details ) . then this multiband hubbard - type hamiltonian was mapped onto an effective low - energy model @xcite .
parameters of this effective model were obtained directly from the _ ab initio _ parameters of the multiband model .
the low - energy model appears to be the @xmath1 model ( @xmath28 model with the three - site correlated hoppings ) for n - type cuprates and the _ singlet - triplet _ @xmath1 model for p - type systems .
however , for @xmath29 in a phase without a long - range magnetic order the role of the triplet state and the singlet - triplet hybridization is negligible @xcite .
therefore , the triplet state could be omitted and in the present paper we will describe the low - energy excitations in the single - layer p- and n - type cuprates within the @xmath1 model . to write down the model hamiltonian we use the hubbard @xmath30-operators @xcite : @xmath31 .
here index @xmath32 enumerates quasiparticle with energy @xmath33 , where @xmath34 is the @xmath35-th energy level of the @xmath36-electron system .
the commutation relations between x - operators are quite complicated , i.e. two operators commute on another operator , not a @xmath37-number
. nevertheless , depending on the difference of the number of fermions in states @xmath35 and @xmath38 it is possible to define quasi - fermi and quasi - bose type operators in terms of obeyed statistics . in this notations
the hamiltonian of the @xmath1 model in the hole representation have the form : @xmath39 here @xmath40 is the chemical potential , @xmath41 is the spin operator , @xmath42 , @xmath43 , @xmath44 , @xmath45 is the number of particles operator , @xmath46 is the exchange parameter , @xmath47 is the charge - transfer gap . in the notations of ref
. @xcite the hopping matrix elements @xmath48 corresponds to @xmath49 and @xmath50 for p- and n - type cuprates , respectively , and @xmath51 .
hamiltonian @xmath52 contains the three - site interaction terms : @xmath53 there is a simple correspondence between @xmath30-operators and single - electron annihilation operators : @xmath54 , where the coefficients @xmath55 determines the partial weight of the quasiparticle @xmath56 with spin @xmath57 and orbital index @xmath58 .
these coefficients are calculated straightforwardly within the gtb scheme . in the considered case
there is only one quasi - fermi - type quasiparticle , @xmath59 , with @xmath60 , and the hamiltonian in the generalized form in momentum representation is given by : @xmath61 the fourier transform of the two - time retarded green function in the energy representation , @xmath62 , can be rewritten in terms of the matrix green function @xmath63_{\alpha \beta } = \left<\left < x_{\vec{k}}^\alpha \left| { x_{\vec{k}}^\beta}^\dag \right . \right>\right>_e$ ] : @xmath64 the diagram technique for hubbard x - operators has been developed @xcite and the generalized dyson equation @xcite in the paramagnetic phase ( @xmath65 ) reads : @xmath66^{-1 } \hat{p}(\vec{k},e).\end{aligned}\ ] ] here , @xmath67 is the exact local green function , @xmath68 $ ] , @xmath69 and @xmath70 are the self - energy and the strength operators , respectively .
the presence of the strength operator is due to the redistribution of the spectral weight between the hubbard subbands , that is an intrinsic feature of the strongly correlated electron systems .
it also should be stressed that @xmath69 in eq .
( [ eq : d ] ) is the self - energy in @xmath30-operators representation and therefore it differs from the self - energy entering dyson equation for the weak coupling perturbation theory for @xmath71 . within hubbard - i approximation
@xcite the self - energy @xmath69 is equal to zero and the strength operator @xmath70 is replaced by @xmath72 , where @xmath73 is the occupation factor .
taking into account that in the considered paramagnetic phase @xmath74 , @xmath75 , with @xmath2 being the doping concentration , after all substitutions and treating all @xmath76-independent terms as the chemical potential renormalization , the generalized dyson equation for the hamiltonian ( [ eq : h ] ) becomes : @xmath77^{-1 } \frac{1+x}{2}.\end{aligned}\ ] ] to go beyond the hubbard - i approximation we have to calculate @xmath78 .
for this purpose we use an equations of motion method for the @xmath30-operators @xcite .
the exact equation of motion for @xmath79 is : @xmath80 = ( \varepsilon_0-\mu ) x_{\vec{k}}^{\alpha } + l_{\vec{k}}^{\alpha } \nonumber\\ & = & ( \varepsilon_0-\mu ) x_{\vec{k}}^{\alpha } + { l_{\vec{k}}^{\alpha}}^{(0 ) } + ( l_{\vec{k}}^{\alpha } - { l_{\vec{k}}^{\alpha}}^{(0 ) } ) .\end{aligned}\ ] ] here @xmath81 is the linearized and decoupled in hubbard - i approximation operator @xmath82 , @xmath83 let us define @xmath84 and linearize it with respect to @xmath85 : @xmath86 where @xmath87 are the coefficients of the linearization .
all effects of the finite quasiparticle lifetime are contained in the irreducible part @xmath88 . in this paper
we neglect it , @xmath89 .
since the exact equation for the green function is given by @xmath90 , it is straightforward to find that in our approximation @xmath91 corresponds to the self - energy : @xmath92 introducing notations for the static spin - spin correlation functions @xmath93 and for the kinematic correlation functions @xmath94 the expression for the quasiparticle self - energy becomes : @xmath95 k_{\vec{q } } \nonumber\\ & - & \frac{2}{1+x } \frac{1}{n } \sum\limits_{\vec{q } } \left [ t_{\vec{k}-\vec{q } } - \frac{1-x}{2 } \left ( j_{\vec{q } } - \frac{\tilde{t}_{\vec{k}-\vec{q}}^2}{e_{ct } } \right ) - \frac{1+x}{2 } \frac{2 \tilde{t}_k \tilde{t}_{\vec{k}-\vec{q}}}{e_{ct } } \right ] \frac{3}{2 } c_{\vec{q}}.\end{aligned}\ ] ] here @xmath36 is the number of vectors in momentum space . until now we have made two major approximations .
first , we neglected irreducible part of the self - energy , @xmath88 , thus allowing quasiparticles to have infinite lifetime .
second , and this is not so obvious , in eq .
( [ eq : d_h ] ) for the strength operator @xmath70 we drop out corrections beyond hubbard - i approximation .
these dropped corrections also lead to the finite quasiparticle lifetime
. the consequence of these approximations will be discussed later .
kinematic correlation functions ( [ eq : k ] ) are calculated straightforwardly with the help of green function ( [ eq : d_h ] ) .
the spin - spin correlation functions for the @xmath96 model with three - site correlated hoppings @xmath52 were calculated in ref .
@xcite . in this paper
the equations of motion for the spin - spin green function @xmath97 were decoupled in the rotationally invariant quantum spin liquid phase , similar to refs .
the results of this approach for a static magnetic susceptibility for the @xmath96 model are similar to those obtained by other methods @xcite .
higher - order correlation functions appearing due to the @xmath52 term are decoupled in the following way : @xmath98 thus , higher - order kinematic and spin - spin scattering channels are decoupled .
after the terms proportional to @xmath99 being neglected , the expression for fourier transform of the spin - spin green function becomes : @xmath100 where @xmath101,\end{aligned}\ ] ] and magnetic excitations spectrum @xmath102 represented by the eq .
( 26 ) of paper @xcite .
the following results were obtained by self - consistent calculation of the chemical potential @xmath40 , the spin - spin correlation functions ( [ eq : c ] ) using green function ( [ eq : spingf ] ) , and the kinematic correlation functions ( [ eq : k ] ) using green function ( [ eq : d_h ] ) with the self - energy ( [ eq : sigma ] ) .
for lsco the lda+gtb calculated parameters are ( in ev ) : @xmath103 .
all figures below are in electron representation . .
the position of the chemical potential is denoted by the horizontal ( green ) line .
results within the hubbard - i approximation are shown by dashed ( red ) and dash - dotted ( cyan ) curves for the model with ( @xmath1 model ) and without ( @xmath28 model ) three - site correlated hoppings , respectively .
bold solid ( blue ) curves represent the results for the @xmath1 model with the short range magnetic order . ]
first of all we would like to stress the important effects caused by the three - site correlated hoppings @xmath52 and the renormalizations due to the short range magnetic order .
previously , the importance of the three - site correlated hoppings in the normal and superconducting phases has been demonstrated in refs .
@xcite . in fig .
[ fig : models ] we present our results for 16% hole doping within different approximations . evidently , introduction of three - site interaction terms results in the change of the position of the top of the valence band .
therefore , this will become important at small @xmath2 .
in afm phase of the @xmath96 model there is a symmetry around @xmath104 point . in the paramagnetic phase this symmetry is absent . due to the scattering on the short range magnetic fluctuations with afm wave vector @xmath105 the states near the @xmath106 point
are pushed below the fermi level ( see fig.[fig : models ] ) , thus totally changing the shape of the fs . in other words ,
the short range magnetic order `` tries '' to restore the symmetry around @xmath104 point . in our calculations
the short range magnetic fluctuations are taken into account via the spin - spin correlation functions ( [ eq : c ] ) . model for p - type cuprates .
index @xmath35 enumerates real space vectors connecting neighboring sites : @xmath107 for nearest - neighbors , @xmath108 for the next nearest neighbors , and so on . in ( c ) and ( d ) the doping dependence of the nodal fermi velocity ( in units of ev@xmath109 with @xmath110 being a lattice constant ) and the effective mass are shown . ]
our results for the doping dependence of the kinematic and spin - spin correlation functions are shown in fig .
[ fig : tj3pc ] .
note , the kinematic correlation functions @xmath111 possess a very nontrivial doping dependence . for low concentrations , @xmath112 , due to the strong magnetic correlations the hoppings to the nearest and to the next - nearest neighbors are suppressed leading to the small values of @xmath113 and @xmath114 , while @xmath115 is not suppressed . upon increase of the doping concentration above @xmath116 ,
magnetic correlations decrease considerably and nearest - neighbor kinematic correlation function @xmath113 increase .
next major change sets at @xmath117 when the system possesses almost fermi liquid behavior : @xmath113 becomes largest of all @xmath111 s , while the magnetic correlation functions @xmath118 and the kinematic correlation function @xmath115 are strongly suppressed . within the @xmath1 model for p - type cuprates .
] so , we can clearly define two points of the crossover , namely @xmath116 and @xmath117 .
the system behavior is quite different on the different sides of these points , although there is no phase transition with symmetry breaking occurs . to understand the nature of these crossovers we consider the fs evolution with doping concentration , presented in fig.[fig : tj3pbands ] . at low @xmath2
the fs has the form of the hole pockets centered around @xmath119 point
. then these pockets enlarge and at @xmath3 all of them merge together , forming the two fs contours .
up to @xmath4 the fs topologically equivalent to the two concentric circles with the central one shrinking toward @xmath120 point . for @xmath4
the central fs contour shrinks to the single point and vanishes , leaving one large hole - type fs
. apparently , the topology of the fs changes drastically upon doping .
in particular , it happens at @xmath3 and at @xmath4 .
for the first time the `` electronic transition '' accompanying the change in the fs topology , or the so - called lifshitz transition , was described in ref .
now such transitions referred as a quantum phase transitions with a co - dimension@xmath121 ( see e.g. paper @xcite ) .
note , when the fs topology changes at quantum critical concentrations @xmath122 and at @xmath123 the density of states at the fermi level also exhibit significant modifications .
this results in the different behavior of the kinematic and magnetic correlation functions on the different sides of these crossover points . and
the changes in the density of states at the fermi level will also result in the significant changes of such observable physical quantities as the resistivity and the specific heat . also , from the obtained quasiparticle dispersion we calculated the doping dependence of the nodal fermi velocity @xmath18 and the effective mass @xmath124 ( see fig .
[ fig : tj3pc](c ) and ( d ) ) .
nodal fermi velocity does not show steep variations with increase of the doping concentration in agreement with the arpes experiments @xcite .
effective mass @xmath24 increase with decreasing @xmath2 and reveals tendency to the localization in the vicinity of the metal - insulator transition .
but this increase is not very large and overall @xmath124 doping dependence agrees quite well with the experimentally observed one @xcite .
note , the non - monotonic doping dependence of both these quantities reflects the presence of the critical concentrations @xmath125 and @xmath126 . , but within the model for p - type cuprates without three - site correlated hoppings ( @xmath28 model ) . ] , but within the model for p - type cuprates without three - site correlated hoppings ( @xmath28 model ) . ] to analyze the effect of the three - site hopping term @xmath52 we also calculated the doping dependence of the band structure and fs within the @xmath28 model .
the behavior of the kinematic and the spin - spin correlation functions , presented in fig .
[ fig : tjpc ] , is quite different from that of the @xmath1 model .
there is only one quantum critical point at @xmath127 and the effective mass becomes very large for @xmath2 approaching zero .
for @xmath128 the evolution of the fs and density of states near the fermi level is smooth , without significant changes ( see fig.[fig : tjpbands ] ) .
most part of the difference to the @xmath1 model stems from the role of @xmath52 in the energy of states near the @xmath106 point , thus determining the topology of the fs and the physics at low doping concentrations ( see fig.[fig : models ] ) .
now let us consider n - type cuprates within the @xmath1 model . for ncco the lda+gtb calculated parameters are ( in ev ) : @xmath129 . , but within the @xmath1 model for n - type cuprates
] , but within the @xmath1 model for n - type cuprates .
] the obtained doping dependence of the kinematic and magnetic correlation functions presented in fig .
[ fig : tj3nc ] .
there is only one crossover point at @xmath130 .
also , in contrast to the p - type results , the most important kinematic correlation function on the left of this point is @xmath114 , rather than @xmath115 . for @xmath131
the system demonstrates fermi liquid - like behavior with magnitude of kinematic correlation function decreasing with the distance , and small values of the magnetic correlations .
the role of the short range magnetic order and three - site hopping terms in n - type cuprates is similar to that of p - type .
in particular , due to the scattering on the magnetic excitations the states near the @xmath106 point pushed _ above _ the fermi level , and the local symmetry around the @xmath104 points is restored , reminding of the short - range afm fluctuations ( see fig . [
fig : tj3nbands ] ) . instead of hole pockets around the @xmath119 point in p - type ,
here at low @xmath2 the electron pockets around @xmath132 and @xmath133 points are present . upon increase of the doping concentration these pockets become larger and merge together at @xmath5 . for higher concentrations the fs appear to be a large hole - like one , shrinking toward @xmath106 point .
therefore , no other changes in the fs topology other than at @xmath5 are present .
referring to the same arguments as in previous section , we claim that in our approach there is only one quantum critical point at @xmath134 in the n - type cuprates .
th non - monotonic change of the effective mass and the nodal fermi velocity is also present at this concentration , as evident from fig . [
fig : tj3nc](c ) and ( d ) .
to summarize , we have investigated the doping - dependent evolution of the low - energy excitations for p- and n - type high-@xmath0 cuprates in the regime of strong electron correlations within the sequentially derived effective model with the _ ab initio _ parameters .
we show that due to the changes of the fermi surface topology with doping the system exhibits drastic change of the low - energy physics .
namely , for p - type cuprates there exist two critical concentrations , @xmath135 and @xmath136 . along the different sides of these concentrations the behavior of the density of states near the fermi level , of the kinematic and magnetic correlation functions , of the effective mass and the nodal fermi velocity , is drastically different .
this let us speak about crossover , or , taking into account the accompanying fs topology changes , about quantum phase transitions at these quantum critical concentrations .
first of all , we would like to comment on the approximations made in this work .
since we use the perturbation theory with hopping @xmath138 and exchange @xmath139 as small parameters , appropriate for the strongly correlated regime , the real part of the corrections to the results obtained will be small to the extent of smallness of the higher powers of @xmath140 and @xmath141 .
this will result in the small changes of the band dispersion and in the fine details of the fs , not changing anything qualitatively .
more concerns give the imaginary part of the neglected corrections to the strength operator @xmath70 and to the self - energy @xmath69 through @xmath88 .
application of the equation of motion decoupling method to the hubbard model with finite quasiparticle lifetime @xcite reveals that the results of the mean - field - like approximation is qualitatively correct .
quantitatively , at low doping the imaginary part of the self - energy leads to the hiding of the fs portions above the antiferromagnetic brillouin zone ( @xmath142 line ) .
this results in fermi arc rather than hole pockets at @xmath17 ( see fig .
[ fig : tj3pbands ] ) . also , we can compare our results to the numerical methods , namely , to the exact diagonalization studies @xcite .
the quasiparticle dispersion of the @xmath143 model in ref .
@xcite can be considered as consisting of two bands . for p - type cuprates intensities of the spectral peaks corresponding to the band situated mostly above the fermi level ( in electron representation )
are very low . this band
is often called a `` shadow '' band and appears due to the scattering on the short range afm fluctuations .
notably , our band dispersion from figs .
[ fig : tj3pbands ] and [ fig : tj3nbands ] reproduce very well the shape of the other , `` non - shadow '' , band .
it is this band where the most part of the spectral weight is residing , thus determining most of the low - energy properties , except for such subtle effects as a so - called `` kink '' in dispersion @xcite . also ,
all renormalizations not included in consideration will change the values of the critical concentrations @xmath125 , @xmath126 , and @xmath144 . comparing with the results of a more rigorous theory in paper @xcite
, we expect these values to decrease .
thus we conclude that our theory captures the most important part of the low - energy physics within the considered ( and justified for cuprates ) model .
this claim is supported by the qualitative agreement with the critical concentrations of crossover observed in the transport experiments @xcite and in the optical experiment @xcite , and even quantitative agreement of the doping dependence of the nodal fermi velocity and of the effective mass @xcite .
although we use a simple mean - field theory , though strong - coupled , the agreement with the experiments is not surprising since we included all necessary for high-@xmath0 copper oxides ingredients .
namely , the short range magnetic order and three - site correlated hoppings .
former is the intrinsic property of the cuprates exhibiting long range afm order at low @xmath2 , while latter results from the sequential derivation of the low - energy effective model .
authors would like to thank d.m .
dzebisashvili and v.v .
valkov for very helpful discussions , i. eremin , p. fulde , n.m .
plakida , a.v .
sherman , and v.yu .
yushankhai for useful comments .
this work was supported by intas ( ys grant 05 - 109 - 4891 ) , siberian branch of ras ( lavrentyev contest for young scientists ) , rfbr ( grants 06 - 02 - 16100 and 06 - 02 - 90537 ) , program of physical branch of ras `` strongly correlated electron systems '' , and joint integration program of siberian and ural branches of ras n.74 .
s. nakamae , k. behnia , n. mangkorntong , m. nohara , h. takagi , s.j.c .
yates , and n.e .
hussey , phys . rev .
b * 68 * , 100502(r ) ( 2003 ) . k.m .
shen , f. ronning , d.h .
lu , f. baumberger , n.j.c .
ingle , w.s .
lee , w. meevasana , y. kohsaka , m. azuma , m. takano , h. takagi , and z .- x .
shen , science * 307 * , 901 ( 2005 ) .
a. damascelli , z. hussain , and z .- x .
shen , rev .
mod . phys . *
75 * , 473 ( 2003 ) .
t. yoshida , x.j .
zhou , t. sasagawa , w.l .
yang , p.v .
bogdanov , a. lanzara , z. hussain , t. mizokawa , a. fujimori , h. eisaki , z .- x .
shen , t. kakeshita , and s. uchida , phys .
rev . lett . * 91 * , 027001 ( 2003 ) .
y. ando , y. kurita , s. komiya , s. ono , and k. segawa , phys .
lett . * 92 * , 197001 ( 2004 ) .
y. ando , s. komiya , k. segawa , s. ono , and y. kurita , phys .
lett . * 93 * , 267001 ( 2004 ) .
n. gedik , m. langner , j. orenstein , s. ono , y. abe , and y. ando , phys .
lett . * 95 * , 117005 ( 2005 ) .
santander - syro , r.p.s.m .
lobo , n. bontemps , w. lopera , d. girata , z. konstantinovic , z.z .
li , and h. raffy , phys .
b * 70 * , 134504 ( 2004 ) .
f. carbone , a.b .
kuzmenko , h.j.a .
molegraaf , e. van heumen , v. lukovac , f. marsiglio , d. van der marel , k. haule , g. kotliar , h. berger , s. courjault , p.h .
kes , and m. li , phys . rev .
b * 74 * , 064510 ( 2006 ) ; n. harima , j. matsuno , and a. fujimori , y. onose , y. taguchi , and y. tokura , phys . rev b * 64 * , 220507(r ) ( 2001 ) .
zhou , t. yoshida , a. lanzara , p.v .
bogdanov , s.a .
kellar , k.m .
shen , w.l .
yang , f. ronning , t. sasagawa , t. kakeshita , t. noda , h. eisaki , s. uchida , c.t .
lin , f. zhou , j.w .
xiong , w.x .
zhao , a. fujimori , z. hussain , and z .- x .
shen , nature * 423 * , 398 ( 2003 ) . a.a .
kordyuk , s.v .
borisenko , a. koitzsch , j. fink , m. knupfer , and h. berger , phys .
rev b * 71 * , 214513 ( 2005 ) .
padilla , y.s .
lee , m. dumm , g. blumberg , s. ono , k. segawa , s. komiya , y. ando , and d.n .
basov , phys .
b * 72 * , 060511(r ) ( 2005 ) .
w.f . brinkman and t.m .
rice , phys .
b * 2 * , 4302 ( 1970 ) .
gutzwiller , phys .
* 10 * , 159 ( 1963 ) ; phys .
b * 134a * , 923 ( 1964 ) ; phys . rev .
b * 137a * , 1726 ( 1965 ) .
f. gebhard , phys .
b * 41 * , 9452 ( 1990 ) .
f. gebhard , phys .
b * 44 * , 992 ( 1991 ) .
g. kotliar and a.e .
ruckenstein , phys .
lett . * 57 * , 1362 ( 1986 ) .
w. metzner and d. vollhardt , phys .
lett . * 62 * , 324 ( 1989 ) .
a. georges , g. kotliar , w. krauth , and m. rozenberg , rev .
phys . * 68 * , 13 ( 1996 ) .
korshunov , v.a .
gavrichkov , s.g .
ovchinnikov , i.a .
nekrasov , z.v .
pchelkina , and v.i .
anisimov , phys .
b * 72 * , 165104 ( 2005 ) .
s.g . ovchinnikov and i.s .
sandalov , physica c * 161 * , 607 ( 1989 ) .
gavrichkov , s.g .
ovchinnikov , a.a .
borisov , and e.g. goryachev , zh .
fiz . * 118 * , 422 ( 2000 ) [ jetp * 91 * , 369 ( 2000 ) ] ; v. gavrichkov , a. borisov , and s.g .
ovchinnikov , phys .
b * 64 * , 235124 ( 2001 ) . m.m .
korshunov , s.g .
ovchinnikov , and a.v .
sherman , phys . met .
metallogr . * 101 * , suppl . 1 , s6 ( 2006 ) .
hubbard , proc .
london a * 277 * , 237 ( 1964 ) .
zaitsev , sov .
jetp * 41 * , 100 ( 1975 ) .
izumov and b.m .
letfullov , j. phys . : condens .
matter * 3 * , 5373 ( 1991 ) .
s.g . ovchinnikov and v.v .
valkov , hubbard operators in the theory of strongly correlated electrons ( imperial college press , london , 2004 ) . j.c .
hubbard , proc .
london a * 276 * , 238 ( 1963 ) .
plakida , v.yu .
yushankhai , and i.v .
stasyuk , physica c * 162 - 164 * , 787 ( 1989 ) .
v.v . valkov and d.m .
dzebisashvili , zh .
fiz . * 127 * , 686 ( 2005 ) [ jetp * 100 * , 608 ( 2005 ) ] .
h. shimahara and s. takada , j. phys .
60 * , 2394 ( 1991 ) ; j. phys .
. jpn . * 61 * , 989 ( 1992 ) .
a. barabanov and o. starykh , j. phys .
. jpn . * 61 * , 704 ( 1992 ) ; a. barabanov and v. m. berezovsky , zh .
fiz . * 106 * , 1156 ( 1994 ) [ jetp *
79 * , 627 ( 1994 ) ] .
a. sherman and m. schreiber , phys .
b * 65 * , 134520 ( 2002 ) .
vladimirov , d. ile , and n.m .
plakida , theor .
fiz . * 145 * , 240 ( 2005 ) [ theor and meth .
physics * 145 * , 1575 ( 2005 ) ] .
valkov , t.a .
valkova , d.m .
dzebisashvili , and s.g .
ovchinnikov , pisma v zhetf * 75 * , 450 ( 2002 ) [ jetp lett . * 75 * , 278 ( 2002 ) ] .
korshunov , s.g .
ovchinnikov , and a.v .
sherman , pisma v zhetf * 80 * , 45 ( 2004 ) [ jetp lett . * 80 * , 39 ( 2004 ) ] .
lifshitz , sov .
jetp * 11 * , 1130 ( 1960 ) .
volovik , acta phys . slov .
* 56 * , 49 ( 2006 ) ; cond - mat/0601372 ( unpublished ) .
plakida and v.s .
oudovenko , cond - mat/0606557 ( unpublished , to appear in physica c ) . t. tohyama , phys . rev .
b * 70 * , 174517 ( 2004 ) . | in this paper a mean - field theory for the spin - liquid paramagnetic non - superconducting phase of the p- and n - type high-@xmath0 cuprates is developed .
this theory applied to the effective @xmath1 model with the _ ab initio _ calculated parameters and with the three - site correlated hoppings .
the static spin - spin and kinematic correlation functions beyond hubbard - i approximation are calculated self - consistently .
the evolution of the fermi surface and band dispersion is obtained for the wide range of doping concentrations @xmath2 . for p - type systems the three different types of behavior
are found and the transitions between these types are accompanied by the changes in the fermi surface topology .
thus a quantum phase transitions take place at @xmath3 and at @xmath4 . due to the different fermi surface topology we found for
n - type cuprates only one quantum critical concentration , @xmath5 .
the calculated doping dependence of the nodal fermi velocity and the effective mass are in good agreement with the experimental data . |
the dynamics of crime and the impact of social relations on the increase of violence has been the object of study in several areas such as social sciences @xcite , criminology @xcite , computing @xcite , economics @xcite and physics @xcite .
when , in the 1950s , naroll and bertalanffy @xcite utilized the concept of allometry a term originally coined in the field of biology to describe scaling laws , _
e.g. _ , the relationship between mass and metabolic rate of organisms @xcite so as to adapt it to the social context , a particularly promising line of research was opened , which today arouses interest of scientists from wide - ranging areas @xcite .
more recently , bettencourt _ et al .
_ @xcite revealed that allometric relationships are statistically present in many aspects of city infrastructures and dynamics .
in particular , they observed a characteristic superlinear relation between the number of serious crimes and the resident population in the united states ( us ) cities which clearly denotes the an intricate social mechanism behind the dynamics of violence .
more recently , melo _ et al . _
@xcite provided strong quantitative evidence showing that an entirely similar behavior can also be observed in brazil . despite the importance of the results presented in the aforementioned studies , their impacts on urban planning , more specifically , on the development of public safety policies ,
are limited due to its purely descriptive nature , which prevents a deeper understanding of the organic causes leading to such a disproportional behavior . in this way ,
a microdynamical approach based on the interactions between local neighborhoods within metropolitan areas certainly represents a more realist view to the problem .
our research motivation is aimed at elucidating issues related to the understanding of the impacts of the influence of social relations on crime . in criminology ,
the role of urban space and its social relations has been previously emphasized to explain the origin of crime @xcite .
particularly , the theory of routine activities , proposed by cohen and felson @xcite , states that crimes , more specifically property crimes , such as robbery and theft , occur by the convergence of the routines of an offender , motivated to commit a crime , and an unprotected victim . in this context , can we explain the occurrence of crimes in different areas of the city based on the current population present in the corresponding urban sub - clusters ?
is this effective population equally important for any type of crime ?
how can we systematically delimit the boundaries of these local neighborhoods so that the social influence is accounted for in a consistent way ? in order to answer these questions , we used actual georeferenced data of crimes committed , and of resident and floating populations for census tracts in the city of fortaleza , brazil .
the concept of resident population has been widely used to understand the effects that the growth of major cities has on social and environmental indicators .
in particular , bettencourt _ et al . _
@xcite showed that the number of homicides scales superlinearly with the population of cities in the us .
subsequently , melo _ et al .
_ @xcite confirmed this behaviour for brazilian cities , but also demonstrated that suicides scale sublinearly with their resident populations .
additionally , oliveira _ et al .
_ @xcite found a superlinear allometric relationship between the resident population and @xmath0 emissions ; this study also raised the issue that the allometric exponents may undergo endogenous trends depending on the respective definition of urban agglomerates .
regarding the floating population , it takes into account the complex dynamics of urban mobility , _ i.e. _ , it has a transient characteristic within the city . to measure social influence quantitatively in urban sub - clusters
, we delimited their boundaries beyond the mere administrative divisions , _ e.g. _ division by neighborhoods or census tracts , by using the city clustering algorithm @xcite .
this model defines these boundaries by population density and the level of commuting between areas of the city . in the present study
, we identified that the incidence of property crimes has a superlinear allometric relationship , with the floating population in certain areas of the city .
this result implies that the increased flow of people in a particular area of the city will take place at the cost of a proportionally greater rate of property crime happening in the region .
more important , this superlinear behavior at the subscale of the city neighborhood provides a plausible explanation for the allometry of serious crimes found in @xcite . precisely , the floating population being systematically larger that the resident one should lead to the disproportional behavior observed for serious crimes and ( resident ) population at the city scale .
we also found a superlinear allometric relationship between the number of crimes of disturbing the peace and the resident population .
this result shows that the effect of social influence must be adequately correlated with resident or floating population , depending on the type of crime considered .
in order to quantify the effects of the social influence on the police calls within a large metropolis , we used three georeferenced datasets for the brazilian city of fortaleza : from the first , we obtain the _ resident population _ ( pop ) , defined by the number of residents per _ census tract _ administrative territorial unit established for the purposes of cadastral control and provided by the brazilian institute of geography and statistics ( ibge ) @xcite . in all
, fortaleza has 3018 census tracts , with approximately 2,400,000 residents , spread over an area of 314 square kilometers ( @xmath1 ) in year 2010 . from the second
, we estimate the _ floating population _ ( flo ) for each census tract through a flow network built on data of the bus system provided by the fortaleza s city hall for the year 2015 @xcite .
flo was measured by the number of people who pass through a census tract in one day .
the city of fortaleza has 2034 buses circulating along 359 bus lines serving approximately 700,000 people who use the city s mass transit system on a daily basis . in the case of fortaleza
, buses still represent the main means of public transportation .
the process of generating of the flow network will be detailed in the supporting information ( see the s1 appendix ) .
finally , we obtain the _ crime dataset _ from the integrated coordination office of public safety operations ( ciops ) @xcite , provides the geographic locations of 81,911 calls to the `` 190 '' ( phone number for emergency ) service about property crimes ( pc ) and 53,849 calls to the same service about disturbing the peace ( dp ) .
these calls were made to police between august 2005 and july 2007 .
the color maps show in figs [ fig1](a)-[fig1](d ) provide the local density in logarithmic scale of pop , flo , dp and pc , respectively , for the city of fortaleza .
we can clearly identify stronger correlations between pop and dp as well as between flo and pc .
in special , there is an evident higher incidence of hot spots in ( b ) and ( d ) than in ( a ) and ( c ) .
also , at the downtown area , highlighted by black circles on each map , the high density of flo is compatible with the high rates of pc rate , while low pop densities seem to explain the low frequency of dp complaints . ) .
( b ) the floating population ( flo ) by @xmath1 .
( c ) the disturbing the peace ( dp ) complaints by @xmath1 .
( d ) the property crimes ( pc ) by @xmath1 .
the black circle highlights the downtown area of the city .
this region has a low density of residents and disturbing the peace calls , and is dense in the flow of people and property crimes.,scaledwidth=100.0% ] in spite of seeming trivial to suggest that there are correlations between pop and dp , as well as flo and pc from the density maps shown in fig [ fig1 ] , the respective scattering plots show uncorrelated behaviors ( fig [ fig2 ] ) . actually , we conjecture that such correlations exist indeed .
the most census tracts have a small area , sometimes the size of one city block , and it is likely that such an agglomeration scale is insufficient to capture the correlations and therefore reveal the impact of social influence on dp and pc . based on this hypothesis , we considered coarse - grainning these spatial properties of fortaleza into clusters , using the census tracts grid as a maximal resolution base .
this clustering aims to find the boundaries of the flow of people and the residential areas of the city .
to define the boundaries beyond administrative delineations , we considered the notion of spatial continuity through the aggregation of census tracts that are near one another using the city clustering algorithm ( cca ) @xcite .
the cca constructs the population boundaries of an urban area considering two parameters , namely , a population density threshold , @xmath2 , and a distance threshold , @xmath3 . for the @xmath4 census tract ,
the population density @xmath5 is located in its geometric center ; if @xmath6 , then the @xmath4 census tract is considered populated .
the length @xmath3 represents a cutoff distance between census tracts to consider them as spatially contiguous , _ i.e. _ , all of the nearest neighboring census tracts that are at distances smaller than @xmath3 are clustered .
hence , a cluster made by the cca is defined by populated areas within a distance less than @xmath3 , as seen schematically in fig [ fig3 ] .
previous studies @xcite have demonstrated that the results produced by the cca can be weakly dependent on @xmath2 and @xmath3 for some range of parameter values . here @xmath3 will be quantified in meters ( m ) and @xmath2 in inhabitants by @xmath1 . ; in contrast , the gray polygons can not be clustered ( @xmath7 ) .
( a ) the red dot represents the geometric center of the @xmath4 census tract and the black circle with radius @xmath3 seeks neighbors belonging to the same cluster .
( b)-(c ) the same search operation is made for the other census tracts and is done until there are no more neighbors within the radius of operation .
( d ) the algorithm finishes running and the cluster is found.,scaledwidth=100.0% ] although the algorithm begins collating an arbitrary seed census tract , it does not produce distinct clusters when varying this seed ; the two factors that are responsible for clustering behavior are the parameters @xmath3 and @xmath2 . in order to determine the effect of the parameterization on the value of the exponent @xmath8
, we sought a range within the parameters where @xmath8 has low sensitivity to this variation .
fig [ fig4 ] shows the behavior of exponent @xmath8 in function of the variation of the cca parameters .
as shown in figs [ fig4]a and [ fig4]b , the value of the parameter @xmath8 obtained from the least - square regressions to the data of pop against dp remains practically insensitive to the cca parameters in the range @xmath9 , regardless of the values of @xmath2 adopted in the estimation process .
moreover , the resulting average @xmath10 provides strong evidence to support a superlinear type of relation between these two variables .
an entirely similar behavior can be observed for flo against dp , but now the exponent @xmath8 remains practically invariant within the range @xmath11 .
the resulting average value of @xmath12 also indicates the presence of a superlinear allometric relation . by varying the parameters of the city clustering algorithm ( cca ) , @xmath13 and @xmath14 . *
( a ) the variation of @xmath8 in correlation between the resident population ( pop ) and the disturbing the peace ( dp ) complaints is illustrated ; ( b ) the variation is illustrated for correlations between the floating population ( flo ) and the property crimes ( pc ) .
both in ( a ) and ( b ) , the x - axis represents @xmath3 , and this parameter was varied from 0 to 800 meters ( m ) ( moment when the largest cluster consumes nearly the entire city ) ; exponent @xmath8 is shown on the y - axis .
the colors of the lines represent the variation of the parameter @xmath2 , which corresponds to the resident population density in ( a ) and the floating population density in ( b ) ; this parameter was varied from 1000 to 8000 .
it was not necessary to use values larger than 8000 because many census tracts start being discarded and the cca can no longer form clusters .
the graphs also show red dashed lines ; between these lines is highlighted the range where , regardless of the parameterization , the exponent @xmath8 has smaller ranges of variation . finally , the dotted black line highlights exponent @xmath15 , in which the relationship between variables is isometric , in both graphs the exponent oscillates to low values of @xmath3 ; in ( a ) , the relationship is superlinear starting at @xmath16 m ; however in ( b ) , superlinearity appears at @xmath17 m.,scaledwidth=100.0% ] at this point , it is necessary to determine a suitable criteria for selecting an adequate value of the parameter @xmath2 . while lower values of @xmath3 lead to the formation of a larger number of cca clusters , reduced values of @xmath2 tend to eliminate fewer census tracts from the map , thereby including a larger portion of in the population under analysis . here
we propose that a proper choice of @xmath2 would be associated with a more homogeneous spatial distribution of the population @xcite .
more precisely , we seek for cca clusters whose areas should scale as close to isometrically as possible with the population data , namely , as @xmath18 , with @xmath19 , where @xmath20 is the population , @xmath21 is the area ( are ) of the clusters @xmath1 , and @xmath22 is constant . when varying the city clustering algorithm ( cca ) parameters @xmath13 and @xmath23 . *
( a ) the variation of @xmath24 in correlations between the resident population ( pop ) and the area ( are ) in square kilometers ( @xmath1 ) of clusters discovered with the cca .
( b ) the variation for correlations between the floating population ( flo ) and are . in ( a ) and ( b )
, the x - axis represents the parameter @xmath3 , and the y - axis represents the exponent @xmath24 .
the line colors represent the variation of the parameter @xmath2.,scaledwidth=100.0% ] in order to follow the procedure previously described , we obtain from fig [ fig5]a that @xmath25 and @xmath26 correspond to the pair of cca parameter values leading to the closest to isometric relation found between are and pop . in the case of are and flo , the values are @xmath27 and @xmath28 , as depicted in fig [ fig5]b ( see the s2 appendix ) .
the census tracts were grouped , using the cca , by pop and flo ( fig [ fig6 ] ) . in the fig [ fig6]a ,
the division achieved by pop is illustrated ; the city was divided using @xmath3 = 270 m and @xmath2 = 6000 resident people per @xmath1 . in the fig [ fig6]b
, we show the division achieved by flo , using @xmath3 = 320 m and @xmath2 = 2000 floating people per @xmath1 per day .
we emphasize that there are bigger gaps in the pop map ( fig [ fig6]a ) than in the flo map ( fig [ fig6]b ) .
the reason for such behavior is the fact that fortaleza has commercial regions , _
i.e. _ , regions where there is not a large presence of residents . regarding floating people ,
there are people moving practically throughout the entire city , both in commercial areas and in residential areas . .
( a ) the population density was used in order to find the boundaries of the clusters with @xmath3 = 270 m and @xmath2 = 6000 resident people per @xmath1 .
( b ) the division found by considering urban mobility is shown ; the map illustrated here was generated for @xmath3 = 320 m and @xmath2 = 2000 floating people per km@xmath29 in one day in fortaleza.,scaledwidth=100.0% ] as compared to the results shown in fig [ fig2 ] , the application of the cca to the data discloses a rather different scenario for the correlations among the variables investigated here .
first , as shown in fig [ fig7]a and [ fig7]b , superlinear relations in terms of power laws , @xmath30 , are revealed between pc and flo as well as between dp and pop , with exponents @xmath31 and @xmath32 , respectively . in contrast , the relations obtained between dp and flo and pc and pop are closer to isometric ( linear ) , with exponents @xmath33 and @xmath34 respectively , although the low values of the corresponding determination coefficients indicate that these results should be interpreted with caution .
, between the floating population ( flo ) and the property crimes ( pc ) .
( b ) a superlinear relationship was also found between the resident population ( pop ) and the disturbing the peace ( dp ) complaints , with exponent @xmath32 .
( c)-(d ) the scatering plots of dp with flo and pc with pop show an isometric relation was found between the variables , but with lower correlations than ( a ) and ( b ) .
the @xmath35 is defined as determination coefficient @xcite.,scaledwidth=100.0% ]
in this paper , we define a methodology to understand the impact that society has on calls to the police finding the boundaries of social influence in a large metropolis through the analysis of georeferenced datasets for the city of fortaleza , brazil .
we used the cca , a clustering algorithm , in order to define intracity clusters , _
i.e. _ spatially contiguous populated areas within a cutoff distance .
the volume of social influence was measured in various localities of the city based on the presence of residents and the flow of passers - by through the census tracts . unlike intercity studies , where social influence was measured only by the presence of residents , we propose that , within a city , urban mobility should be considered to understand the dynamics of social indicators , such as criminal activity .
our results show that the incidence of property crimes grows superlinearly as a power law with the floating population , with allometric exponent @xmath36 .
therefore , the increase of the flow of people in a region of the city leads to a disproportionally higher number of property this type of crime .
our results are in agreement with the routine activities hypothesis @xcite , which states that a crime occurs by the convergence of the routines of three agents , specifically : the presence of a motivated offender , the presence of an unprotected victim , and the absence of a guardian able to prevent the transgression . in other words , the regions with higher incidence of floating population potentiate the meeting of the routines of these three agents .
this result is in clear contrast with the incidence of crimes related with peace disturbance , where an allometric relation can also be detected , but with the resident population ( pop ) instead .
we hope the our results could shed some light on the understanding of crimes inside urban areas , as well as assist eventual violence mitigation policies .
the findings described in this paper bring alternatives to implementing innovative practices to decision makers within cities .
the most obvious of these relates to the fact that , by showing the correlation of different types of crimes with the home population but also with the floating population , it is also clear that the police force allocation strategies should be implemented via the analysis of different cluster configurations that depend on the type of crime .
for example , the allocation of community policing , more appropriate to resolve conflicts that potentially can emerge from the disturbance of people s peace , must be planned from a cluster configuration and a hot spot analysis that were produced from the perspective of the density of resident population .
when it is necessary to establish a policy for allocating a uniformed police in order to mitigate crimes against property , the allocation of the police force must be conducted from an analysis from the movement of people .
in addition to these police allocation strategies , the results described herein provide important indicators for the creation of public policies for land use and environmental design in general . work in this line has been developed such as in ref .
@xcite where a framework has been proposed to associate the physical spaces and the feeling of safety as well as @xcite , who launched the environmental criminology putting focus of criminological study on environmental or context factors that can influence criminal activity .
these include space ( geography ) , time , law , offender , and target or victim .
these five components are a necessary and sufficient condition , for without one , the other four , even together , will not constitute a criminal incident .
the discovery demonstrated in the article that there is a superlinear relationship between crime and population ( resident or floating ) in clusters within cities strengthens the claim that changes in urban form can lead to reduced crime as discussed in ref .
the urban mobility system of a large city is composed of several interconnected networks , such as subway , bus , bicycle , taxicab , and private vehicle networks .
buses are the main means of transportation for the most inhabitants in the city of fortaleza , being used by about 700,000 people daily .
taking this fact into account , we assumed that the urban mobility within the city can be represented by the use of the bus system .
thus , the trajectories of bus users will be used to infer the floating population at the different points of the city . in order to understand the people flow throughout the city
, we used four spatio - temporal datasets related to fortaleza s bus network , which are : bus stops ; bus lines ; gps tracking of vehicles ; and ticket card validation .
all datasets refer to a normal business day , in fact , the march 11th , 2015 - a wednesday . in total , fortaleza has 4,783 bus stops served by 2,034 buses along 359 different routes . the integrated transportation model adopted by fortaleza city hall , called _ bilhete nico _ , allows the registred users to make a bus transfer anywhere in the city , as long as it is within two hours since the last validation of their ticket card .
the validation process is understood as the act of the user swiping his / her ticket card at the turnstile on the bus or at the bus terminal .
usually , such procedure happens at the beginning of the trip , since the turnstile is close to the bus entrance in fortaleza . in this context
, we are able to define the origin - destination matrix ( odm ) for fortaleza s bus network through the following hypoteses @xcite : we can assume that an user s origin point could be represented by the earliest of all first daily ticket validations in the interval of two weeks before march 11th , as well as an user s destination point could also be represented by the earliest of all last daily ticket validations in the interval of one month before march 11th , bearing in mind that an user could have different destination points along the week , _
i.e. _ we have to analyze mondays with mondays , tuesdays with tuesdays , and so on .
hence , we estimated the origin - destination pair for about 40% of the bus users representing the overall behavior of the urban mobility in fortaleza .
finally , we supposed that the trajectories of bus users are defined by the composition of routes of the buses took by them between their origin - destination pair @xcite . in this context , we describe the trajectories of bus users as a directed graph @xmath37 , where @xmath38 and @xmath39 are the set of vertices @xmath40 and edges @xmath41 , respectively .
an edge @xmath41 between the vertices @xmath42 and @xmath43 is defined by the ordered pair @xmath44 . in our approach ,
the vertices represent bus stops and the edges represent the demand of bus users between two consecutive bus stops . for each vertice @xmath42
, we defined a weighted function @xmath45 as the sum of the users passing by @xmath42 .
thus , we calculated the floating population as the sum of all @xmath45 within each census tracts .
in the main text , we observed the emergence of a trend toward an isometric relation between the populations ( pop and flo ) of the cca clusters and their respective areas ( are ) .
the fig [ s2_appendix_fig2 ] shows the closest relations to an isometric behavior for both cases .
the cca parameters used to define the clusters were @xmath46 and @xmath25 for the pop case , illustrated in fig [ s2_appendix_fig2]a , and @xmath47 and @xmath48 for the flo case ( fig [ s2_appendix_fig2]b ) . ) . *
( a ) the correlation between pop and the area of the clusters ( are ) calculated using city clustering algorithm ( cca ) where @xmath46 residents per @xmath1 and @xmath49 meters ( m ) .
( b ) the correlation between flo and are calculated using the cca for @xmath50 floating people per @xmath1 and @xmath48 m. the red line shows the ordinary least square ( ols ) regression applied to the logarithm of the data @xcite , and the blue continuous line indicates the nadaraya - watson kernel regression @xcite . finally , the blue dashed lines delimit the 95% confidence interval estimated by 500 random bootstrapping samples with replacement @xcite . the @xmath35 is defined as the determination coefficient @xcite.,scaledwidth=100.0% ] r. guedes , v. furtado , and t. pequeno , `` multiagent models for police resource allocation and dispatch , '' in _ intelligence and security informatics conference ( jisic ) , 2014 ieee joint _ , pp . 288291 , ieee , 2014 . l. g. alves , h. v. ribeiro , e. k. lenzi , and r. s. mendes , `` distance to the scaling law : a useful approach for unveiling relationships between crime and urban metrics , '' _ plos one _ , vol . 8 , no . 8 , p. e69580 , 2013 .
l. m. bettencourt , j. lobo , d. helbing , c. khnert , and g. b. west , `` growth , innovation , scaling , and the pace of life in cities , '' _ proceedings of the national academy of sciences _ , vol .
104 , no .
17 , pp . 73017306 , 2007 .
h. d. rozenfeld , d. rybski , j. s. andrade , m. batty , h. e. stanley , and h. a. makse , `` laws of population growth , '' _ proceedings of the national academy of sciences _ , vol .
105 , no .
48 , pp . 1870218707 , 2008 .
h. d. rozenfeld , d. rybski , x. gabaix , and h. a. makse , `` the area and population of cities : new insights from a different perspective on cities , '' _ the american economic review _ , vol .
101 , no . 5 , pp .
22052225 , 2011 .
`` coordenadoria integrada de operaes de segurana ( ciops ) . ''
available : http://dados.fortaleza.ce.gov.br/dataset/8e995f96-423c-41f3-ba33-9ffe94aec2a8/resource/de4e876a-ee24-4d6e-9722-db9dc454bbe6/download/policecalls.csv . accessed : 2016 - 10 - 06 .
j. b. kinney , p. l. brantingham , k. wuschke , m. g. kirk , and p. j. brantingham , `` crime attractors , generators and detractors : land use and urban crime opportunities , '' _ built environment _ , vol .
34 , no . 1 ,
pp . 6274 , 2008 .
c. caminha , v. furtado , v. pinheiro , and c. ponte , `` micro - interventions in urban transportation from pattern discovery on the flow of passengers and on the bus network , '' in _ smart cities conference ( isc2 ) , 2016 ieee international _ , pp .
16 , ieee , 2016 . | we investigate at the subscale of the neighborhoods of a highly populated city the incidence of property crimes in terms of both the resident and the floating population .
our results show that a relevant allometric relation could only be observed between property crimes and floating population .
more precisely , the evidence of a superlinear behavior indicates that a disproportional number of property crimes occurs in regions where an increased flow of people takes place in the city . for comparison , we also found that the number of crimes of peace disturbance only correlates well , and in a superlinear fashion too , with the resident population .
our study raises the interesting possibility that the superlinearity observed in previous studies [ bettencourt _ et al .
_ , proc .
natl .
acad .
sci .
usa * 104 * , 7301 ( 2007 ) and melo _ et al .
_ , sci .
rep . * 4 * , 6239 ( 2014 ) ] for homicides versus population at the city scale could have its origin in the fact that the floating population , and not the resident one , should be taken as the relevant variable determining the intrinsic microdynamical behavior of the system . |
fig . 1 : current spin polarization ( upper plots ) and interface resistance ( lower plots ) of fe / znse(001 ) ( left panel ) and fe / gaas(001 ) ( right panel ) , as functions of the schottky barrier thickness .
the solid curves refer to the cases of zn and ga termination , and the dashed to se and as termination at the interface .
the values are obtained by an integration over the whole sbz .
note the different scales in the @xmath26 axes .
2 : local dos of the fe / znse(001 ) at @xmath13 with a zn ( upper plot ) and a se ( lower plot ) terminated interface .
the filled gray lines denote the dos of bulk fe , the solid lines the dos for the interface fe layer , and the dashed lines the dos of the first ml of znse ( zn ml for the upper plot and se ml for the lower ) .
the vertical lines indicate the gap region in the znse and in between is the position of the fermi level . in this plot ,
the potential is in the ground state position with no schottky barrier inserted . | we calculate the current spin polarisation and the interface resistance of fe / gaas and fe / znse ( 001 ) spin injection junctions from first principles , including also the possibility of a schottky barrier . from our results of interface resistance
we estimate the barrier thickness needed for efficient spin injection if the process is non - ballistic .
spin injection : interface resistance in fe / semiconductor junctions calculated from first principles olaf wunnicke , phivos mavropoulos@xmath0 , and peter h. dederichs@xmath0 * key words : * + spin injection ; spin dependent transport ; interface resistance + * running title : * + spin injection in fe / semiconductor junctions + * corresponding author : * + olaf wunnicke , institut fr festkrperforschung , forschungszentrum jlich , d-52425 jlich , germany + email : [email protected] + tel . : + 49 - 2461 - 61 - 4799 , fax : + 49 - 2461 - 61 - 2620 + the achievement of spin - polarised electric transport via the conduction band of semiconductors ( sc ) is one of the central issues in the field of spintronics .
potential technological applications have already been proposed since 1990 @xcite , but the conditions under which this can be realised are not yet clear , neither experimentally nor theoretically .
the bottleneck seems to be the injection of the spin - polarised current from a ferromagnetic metal ( fm ) contact into the sc .
one basic reason for this was revealed by schmidt and collaborators @xcite , who argued that the huge conductivity mismatch of the fm and the sc , in correlation with the much smaller spin - flip mean free path in the fm , will lead to a drastic drop of the spin polarisation of the current in the vicinity of the interface .
there are , in theory , two ways to circumvent the obstacle : the ballistic spin injection and the injection through a spin - selective tunneling barrier .
at least the latter has already been used in praxis with considerable success @xcite .
the possibility of ballistic spin injection was proposed by grundler @xcite .
first - principles calculations @xcite have shown the necessity to consider the full band structure of the fm and sc together with the electronic structure of the interface .
then the difference in symmetry of the fm and sc for the minority - spin wave functions can play the most important role , by essentially disallowing the incoming fm minority electron to pass onto a sc state .
spin polarisations and magnetoresistance ratios of up to the ideal 100% were calculated for fe / gaas(001 ) , fe / znse(001 ) , and fe / inas(001 ) systems . in parallel to this
, there is also the proposal of rashba @xcite and fert and jaffrs @xcite that a tunneling barrier at the fm / sc interface could solve the problem even in the diffusive regime .
the idea is that a tunneling barrier can be highly spin selective , as shown in tmr junctions , and at the same time provides an extremely high interface resistance ; this combination of properties would lead to spin injection even in the troubling diffusive regime . in this contribution
we extend our previous work @xcite to the evaluation of interface resistance for majority and minority electrons in fe / gaas and fe / znse ( 001 ) junctions , with or without a tunneling barrier , in the approach developed in ref .
@xcite , and compare to the limits set by the fert - jaffrs theory .
finally , we discuss the importance of interface resonances in such junctions .
schep and collaborators @xcite , working in the approximation of the resistor model , have deduced the resistance of a ballistic interface of area @xmath1 between two bulk regions a ( in our case fe ) and b ( sc ) in which conduction channels are randomly mixed due to diffuse scattering , as @xmath2 .
\label{eq1}\ ] ] here , @xmath3 is the ballistic transmission probability of the interface .
the diffusion is assumed to occur only in the bulk regions . @xmath4 and @xmath5 represent the number of landauer conduction channels in the two bulk regions , _
i.e. _ they are proportional to the area of the projected fermi surface on the interface plane .
we calculate the electronic structure of the bulk and interface regions in the local density approximation of density - functional theory , using the kkr green function method , and assuming two independent spin channels . from the green function the ballistic scattering properties of the interface
are obtained using a method equivalent to the landauer - bttiker formalism @xcite .
the scattering at the interface is assumed specular , _
i.e. _ the part of the bloch @xmath6-vector which is parallel to the interface , @xmath7 , is conserved ; the presupposition for this is that the interface is defect - free .
the transmission probability can be viewed then as a function of @xmath7 , and @xmath3 is an integral of @xmath8 over the surface brillouin zone sbz : @xmath9 in order to achieve spin injection , the sc conduction band edge @xmath10 must be slightly lower than the fermi level @xmath11 ; in our calculations we model such a situation by rigidly lowering the sc potential in the region far from the interface .
we use a value of @xmath12 as a typical one for an experimental situation .
the potential close to the interface , up to two monolayers ( ml ) , is determined by the self - consistent electronic structure and the metal - induced gap states . from this point (
3rd ml ) on the transition to the bulk - like region would depend on the doping and the exact band offset of the sc , with the possibility of a longer or shorter schottky barrier .
we model the situation by either a gradual lowering of the potential up to the final bulk value , thus describing a schottky tunneling barrier , or by an abrupt lowering of the potential to the final bulk value ; more details are given in ref .
@xcite .
the former situation is relevant for the case of spin injection through a tunneling barrier with an extremely high interface resistance and spin selectivity . as explained in refs .
@xcite , for the evaluation of the current spin polarisation the calculation of @xmath8 at @xmath13 is enough for tunneling or thermal spin injection ; for the interface resistance the integration of eq .
( [ eq2 ] ) must be employed , but we note that it can be limited within a small circle around @xmath13 , because of the tiny fermi sphere in the sc . in the semiconductor part , the low value of @xmath14 means that the corresponding fermi sphere is extremely small ; then the conducting channels of the sc are much fewer than the ones in fe , and in eq .
( [ eq1 ] ) @xmath15 .
next , one should compare @xmath16 with @xmath17 . in the case of a tunneling barrier
, @xmath3 is expected to be so small that @xmath16 will be negligible and the interface resistance will be dominated by the ballistic specular scattering .
this remains true even in the absence of a tunneling barrier , because @xmath8 is of the order of @xmath18 for majority spin electrons ( and orders of magnitude less for minority ) @xcite , making @xmath3 of the order of @xmath19 or less .
the results obtained for fe / znse and fe / gaas are presented in fig . 1 .
in the upper plots the current spin polarisation is shown , as a function of barrier thickness , for all possible terminations of the sc part .
the high degree of polarisation is due to the symmetry mismatch of the incoming minority fe states , as explained in ref .
@xcite .
the values obtained for the polarisation by use of eq .
( [ eq2 ] ) are somewhat smaller than at @xmath13 , because at nonzero @xmath7 the symmetry mismatch for minority is not strict .
in all cases except the se terminated fe / znse system , the current spin polarisation decreases for thicker schottky barriers ; this shows mostly in the zn terminated fe / znse system .
the cause can be traced back to the existence of resonant interface states in the vicinity of the fermi level for the minority spin . as discussed in ref .
@xcite , they can contribute to an increase of the tunneling current .
the density of states for the fe / znse interface at @xmath13 is shown in fig . 2 .
for the zn termination , the interface state lies clearly in the vicinity of the fermi level , while for the se termination it lies higher and does not contribute to the conductance .
however , we must note that effects that are not taken into account here , such as lattice relaxations , a different pinning of the band offset , or a finite bias , could move the resonant states away from or onto the fermi level . in the latter case the polarisation
is strongly affected and can even changes sign , if the peak is at @xmath11 . in the lower plots of fig . 1 ,
the interface resistance is presented for various barrier thicknesses .
note the logarithmic scale used for the resistance . for znse
the constant slope reflects the exponential decay of the wave functions within the barrier . for gaas
the asymptotic behaviour of an exponential decay presents itself for thicker barriers , because of the smaller band gap of this material . in order to estimate the range where spin injection with a barrier is most efficient , we take the criterion proposed in ref .
@xcite .
this reads for the case of spin injection and detection via a second sc / fm interface , @xmath20 with @xmath21 the sc resistance ( normalised to interface area ) , @xmath22 its thickness ( between the barriers ) and @xmath23 the spin - flip mean free path .
typically , @xmath21 can be of the order of @xmath24 and @xmath23 of the order of a @xmath25 @xcite . according to the results presented in fig . 1
, this places the necessary schottky barrier thickness in the range of 70 for znse or 100 for gaas , if @xmath22 is of the order of a @xmath25 . in conclusion , we have performed ab initio calculations of the transmission probability in fe / znse and fe / gaas spin injection devices taking into account the possibility of a schottky barrier .
we have estimated the corresponding interface resistances for several barrier thicknesses and found relevant barrier thicknesses for efficient spin injection , and shown that resonant interface states can lower the spin injection efficiency .
our presentation is relevant for situations where diffusive scattering is present in the bulk .
30 s. datta and b. das , appl .
. lett . *
56 * , 665 ( 1990 ) .
g. schmidt , d. ferrand , l. w. molenkamp , a. t. filip , and b. j. van wees , phys .
rev .
b * 62 * , _ r_4790 ( 2000 ) . h. j. zhu , m. ramsteiner , h. kostial , m. wassermeier , h .-
p .
schnherr , and k. h. ploog , phys .
rev .
lett . * 87 * , 016601 ( 2001 ) ; a. t. hanbicki , b. t. jonker , g. itskos , g. kioseoglou , and a. petrou , appl .
phys .
lett . * 80 * , 1240 ( 2002 ) .
d. grundler , phys . rev .
b * 63 * , r161307 ( 2001 ) .
o. wunnicke , ph .
mavropoulos , r. zeller , p. h. dederichs , and d. grundler , phys .
rev .
b * 65 * , 241306(r ) ( 2002 ) . ph .
mavropoulos , o. wunnicke , and p. h. dederichs , phys .
rev .
b * 66 * , 024416 ( 2002 ) .
m. zwierzycki , k. xia , p. j. kelly , g. e. w. bauer , and i. turek , arxiv : cond - mat/0204422 ( unpublished ) .
e. i. rashba , phys .
rev .
b * 62 * , _ r_16267 ( 2000 ) .
a. fert and h. jaffrs , phys .
rev .
b * 64 * , 184420 ( 2001 ) . k. m. schep , j. b. a. n. van hoof , p. j. kelly , g. e. bauer , and j. e. inglesfield , phys . rev .
b * 56 * , 10805 ( 1997 ) . h. u. baranger and a. d. stone , phys . rev .
b * 40 * , 8169 ( 1989 ) .
o. wunnicke , n. papanikolaou , r. zeller , p. h. dederichs , v. drchal , and j. kudrnovsk phys .
rev .
b * 65 * , 064425 ( 2002 ) |
the number counts in a bin of mass @xmath1 and redshift @xmath2 , can be considered as a monopole of the halo density field : _ ( i_m , i_z ) = _ ( i_m , i_z ) + ^2 m z r^2 _ ( = r| m , z ) cluster counts have been shown as a powerful probe of cosmology , e.g. @xcite has produced constraint on @xmath3 and @xmath4 with sz detected clusters .
the study of the clustering of galaxies may be done with the angular correlation function @xmath5 or its harmonic transform @xmath6 , in tomographic redshift bins : c_^(i_z , j_z ) = k^2 k j_(k r_1 ) j_(k r_2 ) p_(k | z_1,z_2 ) in the following i use @xmath6 instead of @xmath5 for simpler equations , although they can be related by a simple linear transformation .
the cross covariance between these two probes involves the halo - galaxy - galaxy angular bispectrum in the squeezed limit @xcite : ( _ ( i_m , i_z ) , c_^(j_z , k_z ) ) = .|_m_1,z_1 b_0^hgg(m_1,z_123 ) @xmath7 is a projection of the 3d bispectrum , for which we need a non - linear model . in the framework of the halo model + hod ,
i have shown that the bispectrum ( or higher orders ) can be computed with a diagrammatic formalism @xcite .
see the diagrams for this hgg bispectrum in fig .
[ fig : diagrams ] .
i have shown that the equations for the covariance can be rewritten in terms of effective quantities , e.g. @xcite : _ ( _ ( i_m , i_z ) , c_^(j_z , k_z ) ) & = 4 f _ b_1^(k_,z_2)^2 + & b_1^(i_m , z_1 ) p_(k_,z_2 ) _
0,p(z_1,z_2 ) these intermediate quantities are integrated over the halo mass and contain the hod and mass function dependency .
they can be compared to measurement on data or to other modeling , providing the possibility for some model independence .
i have built a fast and efficient code to compute the different terms of the covariance ; the plots in fig .
[ fig : cov ] illustrate the numerical results .
we see that different terms can become important depending on mass and redshift .
the code runs in @xmath8 cpu - second , and is thus adequate to be integrated into an mcmc pipeline .
cluster counts follow a poissonian distribution ( up to sample variance ) , thus one can not assume that the joint likelihood of @xmath9 is gaussian .
+ to tackle this , i expanded the joint likelihood with the gram - charlier series , around a fiducial independent case .
i am then able to resum the expansion into @xcite : ( x ) = ( ) ( w ) this analytic form is well - behaved ( positive ) , can be extended straightforwardly to include sample variance , and has correct asymptotic behaviour at large @xmath10 ( gaussian with the correct covariance matrix ) .
hyperparameters ( hps ) allow to detect over / underestimation of error bars , or inconsistencies between data sets @xcite .
the method is at the moment only adapted to gaussian distributions , thus not for poissonian cluster counts .
it is however mathematically impossible to keep the gaussian properties of hps in the poissonian case ( that is , rescaling the variance while keeping the mean ) .
however i found a prescription which approximately respect them . on fig .
[ fig : poissonhp ] are shown three pdfs , corresponding to three different values of the bayesian hp @xmath11 . +
this will allow the use of hps for the cluster counts - galaxy 2-pt combination , after further extension ( sample variance , correlation with @xmath12 as treated in sect .
[ sect : jointlikely ] ) .
i sketched how to combine cluster counts and galaxy 2-pt measurements for increased cosmological constraints , from the physical modeling to the likelihood . in the context of the halo model ,
i introduced a diagrammatic method allowing elegant computation of the equations involved .
i derived a non - gaussian joint likelihood using gram - charlier series , and showed how to introduce bayesian hps to a poissonian distribution .
further work will be necessary to include experimental effects in the covariance and the likelihood : photo - z errors , purity ... next order derivation of the joint likelihood may also be necessary to solve a small bias issue at low @xmath10 , and the bayesian hyperparameters method need to be extended to cluster sample variance and correlation with galaxies . in the medium term
, i aim to build a full mcmc pipeline of cluster - galaxy combination , for realistic forecasts and application to des data .
further details on the model , method , and forecasts will be available in lacasa & rosenfeld ( in prep . )
m=13 - 14 & z=0.2 - 0.3 ; log@xmath13m=15 - 16 & z=0.2 - 0.3 ; log@xmath13m=13 - 14 & z=0.9 - 1 .
the @xmath14 term can be either negative ( dotted line ) at low @xmath15 when galaxies are antibiased , or positive ( solid line ) at high @xmath15 when galaxies are biased.,title="fig : " ] m=13 - 14 & z=0.2 - 0.3 ; log@xmath13m=15 - 16 & z=0.2 - 0.3 ; log@xmath13m=13 - 14 & z=0.9 - 1 .
the @xmath14 term can be either negative ( dotted line ) at low @xmath15 when galaxies are antibiased , or positive ( solid line ) at high @xmath15 when galaxies are biased.,title="fig : " ] m=13 - 14 & z=0.2 - 0.3 ; log@xmath13m=15 - 16 & z=0.2 - 0.3 ; log@xmath13m=13 - 14 & z=0.9 - 1 .
the @xmath14 term can be either negative ( dotted line ) at low @xmath15 when galaxies are antibiased , or positive ( solid line ) at high @xmath15 when galaxies are biased.,title="fig : " ] | present and future large scale surveys offer promising probes of cosmology . for example the dark energy survey ( des ) is forecast to detect @xmath0300 millions galaxies and thousands clusters up to redshift @xmath01.3 .
i here show ongoing work to combine two probes of large scale structure : cluster number counts and galaxy 2-point function ( in real or harmonic space ) .
the halo model ( coupled to a halo occupation distribution ) can be used to model the cross - covariance between these probes , and i introduce a diagrammatic method to compute easily the different terms involved .
furthermore , i compute the joint non - gaussian likelihood , using the gram - charlier series
. then i show how to extend the methods of bayesian hyperparameters to poissonian distributions , in a first step to include them in this joint likelihood . |
in this section , we list the differences of the dual - v scheme @xcite ( fig .
[ fig_setup](b ) of the main article ) from the @xmath3-type scheme ( fig .
[ fig_setup](a ) of the main article ) , which was the focus of the discussion in the main article . * there are two excited states @xmath120 instead of only one .
there are also two excited states @xmath121 coupled to the state @xmath39 , instead of only one .
* the two classical drives have different polarizations , @xmath122 and @xmath123 , and are therefore two independent running waves @xmath100 instead of forming a standing wave .
the quantum probe fields ( single photon ) and can also have two different polarizations .
the operators for right - moving and left - moving fields with @xmath122 polarization is @xmath124 , where `` @xmath125 '' in the subscript is for the right - moving field and `` @xmath126 '' is for the left - moving field .
similarly , @xmath127 are the operators for the @xmath123 fields .
the @xmath122 fields only couple to the transitions @xmath128 and @xmath129 , while the @xmath123 fields only couple to the transitions @xmath130 and @xmath131 . *
the decay rate @xmath21 is from each of the states @xmath120 ( assumed equal ) . *
two - mode transfer matrices are required to describe the scattering ( more details in sec .
[ sec_scattering_coefficients ] below ) .
* under eit storage , we assume that only one of the classical drives is switched on , so that one of the excited states is decoupled from the dynamics . hence , in this case , the dual - v scheme is completely equivalent to the @xmath3type scheme . after storing a photon into state @xmath38 and transferring it to state @xmath39 by a @xmath32-pulse , for the dual - v scheme , the atom behaves as an effective v - type atom with two excited states @xmath121 , instead of a two - level atom .
* it is in principle possible to store the photon such that it is incident only from one side instead of symmetrically from both sides , as required for the @xmath3-type scheme . *
if one - sided storage is desired , a different setup can be used , where the photons are incident on the ensemble directly ( without placing the ensemble inside a sagnac interferometer ) . to convert between conditional reflection or transmission and conditional phase shift ,
a mirror behind the ensemble can be used .
this setup may require less stabilization of path lengths than a sagnac interferometer .
the unconditional fidelity in this setup is found to be approximately the same , but the conditional fidelity has a worse scaling .
the latter is because in eqs . and of the main article
, the terms linear in @xmath132 do not cancel as in the symmetric case ( also see the fidelity derivations in sec . [ sec_analytical_fidelity ] below ) .
here , we state the expressions required for evaluating the choi - jamiolkowski fidelity and success probability accounting for non - zero bandwidth of the scattered photon @xmath28 .
we will also discuss , how the unconditional and conditional fidelities are related to figures of merit for quantum repeaters .
a more detailed discussion can be found in refs .
@xcite .
the ideal evolution of the controlled - phase gate is defined by its action on the computation basis states @xmath133 to calculate the choi - jamiolkowski fidelity , we have to define the input basis states ( @xmath134 for @xmath135 ) and their ideal evolution into the output states ( @xmath136 ) .
since photon @xmath25 is stored and retrieved , while photon @xmath28 is scattered , it is most natural to describe the former as a temporal wave packet and the latter through a frequency distribution over @xmath137 , where @xmath138 ( see figs . [ fig_setup](a ) and [ fig_setup](b ) of the main article ) is the two - photon detuning .
the ideal evolution is defined such that the output frequency distribution of photon @xmath28 is equal to the input frequency distribution , while the output wave packet of photon @xmath25 is set to be the one obtained after storage and retrieval in the absence of scattering . in terms of the storage and retrieval kernels ,
@xmath139 and @xmath140 respectively ( discussed in sec .
[ eit_storage_retrieval ] below ) , the output wave packet of photon @xmath25 is @xmath141 where @xmath67 is the input wave packet . we define @xmath142 to be unnormalized , and the absolute square of its norm @xmath143 is the eit storage and retrieval efficiency , which is , in general , less than unity due to imperfections .
when defining the output basis states ( @xmath136 ) , we ensure that they are normalized by dividing them by @xmath144 . in the numerical calculations , we use the discrete definition of @xmath142 ( and similar one for @xmath145 below ) instead of the continuum version that was stated in the main article for simplicity .
the choices of @xmath142 and @xmath105 above are natural , but there may exist more optimal choices , which give better fidelities . for the real physical evolution , we also need to calculate the output wave packet for photon @xmath25 , when photon @xmath28 was scattered from the ensemble . in this case , we have @xmath146 defining @xmath145 is necessary , because the reflection coefficient of the ensemble with a stored photon depends on the position of the atom , in which the photon was stored .
the choi - jamiolkowski fidelity is computed by assuming a particular entangled state as the input and applying both the ideal and real evolution to it @xcite .
the fidelity between the output states is then the choi - jamiolkowski fidelity .
it is given by @xmath147 if we neglect the bandwidth of photon @xmath28 by setting @xmath105 to be a delta - function , @xmath148 , in the above expression , we obtain eq . of the main article .
the success probability for having two photons at the output is @xmath149 setting @xmath148 , this reduces to eq . of the main article .
the conditional choi - jamiolkowski fidelity is @xmath75 .
the abstract choi - jamiolkowski fidelity considered above can be related to more concrete figures of merit , such as entanglement swap fidelity and success probability in the setting of quantum repeaters . in refs . @xcite it is shown that the success probability of the entanglement swap operation is exactly the same as the success probability @xmath150 that we have defined above .
furthermore , since @xmath78 measures the probability for the photons to be in the right modes with the right phase , whereas @xmath150 measures whether the photons are coming out , it holds that @xmath151 .
the expressions for @xmath78 stated in the main article thus give a lower bound for @xmath150 . for our particular implementation of the controlled - phase gate ,
most of the error in @xmath78 is caused by photon loss , and hence we have @xmath83 . at the same time , @xmath81 is a lower bound and an approximation for the entanglement swap fidelity @xcite . in fig .
[ fig_cphase_fidelity_f_swap ] , we illustrate the approximate equalities of @xmath78 with @xmath150 and @xmath81 with the entanglement swap fidelity @xmath152 . for deterministic operation of the controlled - phase gate with the success probability @xmath150 for heralded operation of the gate .
( b ) comparison of the conditional choi - jamiolkowski fidelity @xmath81 with the entanglement swap fidelity @xmath152 for heralded operation of the gate . both for
( a ) and ( b ) , dotted green and dash - dotted black curves are calculated with @xmath76 , while the solid blue and dashed red curves are calculated with @xmath30 chosen such that the entanglement swap fidelity @xmath152 is maximal .
all quantities are plotted as functions of the number of atoms @xmath12 with fixed @xmath13 . for storage and retrieval , we use the continuum theory described in sec .
[ continuum_eit_storage_retrieval ] . ]
to find the scattering coefficients for an ensemble of @xmath3type atoms , the single - mode transfer matrix formalism @xcite is sufficient . for the dualv scheme ,
two different polarization modes of the electric field are coupled to the atoms , and hence a generalization to a two - mode transfer matrix formalism is required .
the general multi - mode transfer matrix formalism is developed in ref .
@xcite . here
, we briefly summarize it before applying it to calculate the scattering coefficients of atomic ensembles . in the multi - mode transfer matrix formalism
, electric fields are described by vectors of @xmath153 elements , where @xmath154 is the number of the different modes of the electric field .
the fields propagating to the right and the fields propagating to the left are treated as being distinct , hence there are @xmath154 elements for each propagation direction .
we can write the electric field vectors @xmath155 where @xmath156 is the part of the electric field that propagates to the right ( in the positive direction ) , @xmath157 is the part of the electric field that propagates to the left ( in the negative direction ) .
atoms and free propagation between atoms are described by @xmath158 matrices , which relate the vectors of electric field at one position to vectors of electric field at a different position . in general , the transfer matrix for the whole ensemble is obtained by multiplying the transfer matrices for the atoms and free propagation .
we use this approach for the dual - v atoms , for which we assume placement of the atoms , which is incommensurate with the wavelength of the classical drives ( or even completely random ) . for the @xmath3-type scheme
, the atoms are regularly placed with distance @xmath159 , as shown in fig .
[ fig_lambda_type_kd_pi_half_setup ] .
hence , the ensemble consists of repeated unit cells , and one can instead exponentiate the transfer matrix for a single unit cell to find the transfer matrix for the whole ensemble .
for @xmath3type atoms described by @xmath160 transfer matrices , closed - form expression can be obtained for the transfer matrix for the whole ensemble given in terms of the transfer matrix for the unit cell @xcite .
if the transfer matrix for the unit cell is @xmath161 then the transfer matrix for an ensemble of @xmath162 such unit cells is @xmath163 where @xmath164 is the identity matrix , @xmath165 and @xmath166 fulfills @xmath167 .
physically , @xmath166 is equal to the bloch vector @xmath44 multiplied by the length occupied by the unit cell . in the case of the @xmath3type scheme with inter - atomic spacing @xmath168 below
, the unit cell has length @xmath169 and hence we have @xmath170 .
once the transfer matrix for the ensemble @xmath171 has been obtained , one needs to extract the scattering coefficients from it .
the transfer matrix for the ensemble @xmath171 fulfills the relation @xmath172 for the dualv atoms , the parts of the electric field vector @xmath173 have two elements . to calculate the reflection coefficient of the ensemble , when the field is incident from the left
, we assume input fields @xmath174 ( i.e. only a right - moving @xmath122 field incident from the left ) . then from eq .
we see that @xmath175 as the transmission coefficient , we take the first ( @xmath122 ) element of @xmath176 , and as the reflection coefficient , we take the second ( @xmath123 ) element of @xmath177 . the reflection and transmission coefficients of the whole ensemble in the single - mode case
simplify to @xmath178 and @xmath179 respectively .
the only thing left to describe in the general case , is how the matrix @xmath180 ( and @xmath171 if it is not simply equal to @xmath181 ) is calculated .
it is a product of the matrices @xmath182 describing scattering of the electric fields by the atoms and the matrices @xmath183 describing the free propagation of the electric fields .
they are given by @xmath184 where @xmath164 means the @xmath185 identity matrix ( a scalar equal to unity in the single - mode case ) , @xmath16 is the wave vector , and @xmath101 is the propagation distance . each matrix @xmath182 is given in terms a @xmath185 matrix @xmath186 , which will be defined for the specific cases of @xmath3type and dualv atoms considered below . here
, we derive the scattering coefficients @xmath106 and @xmath107 given by eqs . and of the main article . for the @xmath3-type atoms , we use single - mode transfer matrices and hence the parameter @xmath186 in eq .
is a scalar . we will consider a unit cell that consists of two atoms and two lengths of free propagation ( see fig .
[ fig_lambda_type_kd_pi_half_setup ] ) .
one of the atoms is placed on the anti - node of the standing wave of the classical drive , and the other is placed on the node .
the scattering from the former will be described by a parameter @xcite @xmath187 while the scattering for the latter ( an effective two - level atom ) is described by a parameter @xmath188 the transfer matrix for the unit cell is @xmath189 carrying out the above matrix multiplications results in @xmath190 from eq . , we then have @xmath191 with @xmath166 given by @xmath192 . from this matrix
, we obtain the reflection and transmission coefficients @xmath193 the minima of @xmath106 and maxima of @xmath107 ( see fig .
[ fig_r_and_t_plot ] of the main article ) occur when @xmath194 in eq .
is approximately equal to zero .
however , exact equality is never satisfied , since @xmath166 is complex ( a consequence of @xmath195 ) . in the regime , where losses are small ( @xmath196\ll 1 $ ] )
, the approximate resonance condition is @xmath197/2)}}=0.\end{gathered}\ ] ] in the coefficients and we can approximate @xmath198/2 ) -\frac{i\operatorname{im}[\operatorname{tr}(t_\text{cell})]/2}{\sqrt{(1-\operatorname{re}[\operatorname{tr}(t_\text{cell})]/2)(1+\operatorname{re}[\operatorname{tr}(t_\text{cell})]/2)}}.\end{gathered}\ ] ] from eq .
we have @xmath199/2)=\pi k\end{gathered}\ ] ] for some integer @xmath200 .
since the are interested in the first reflection minimum closest to @xmath201 , we choose @xmath202 . for large @xmath162
, we have @xmath203/2=\cos(\pi(n-1)/n)\approx -1+\pi^2/(2n^2).\end{gathered}\ ] ] hence , eq . can be approximated as @xmath204/(2\pi),\end{gathered}\ ] ] and we also obtain the approximate expressions : @xmath205/(2\pi)\right),\\ & \cos(n\theta)\approx(-1)^{n-1},\\ & \sin(\theta)\approx \pi / n.\end{aligned}\ ] ] with these approximations and using the fact that @xmath196=-4\operatorname{im}[\beta_2\beta_3]$ ] , eqs . and become @xmath206}-\beta_2-\beta_3},\\ \label{t_t_n_approx1 } & t_0\approx \frac{(-1)^{n-1}}{1-\frac{2in^3\operatorname{im}[\beta_2\beta_3]}{i\pi^2}(\beta_2+\beta_3)}.\end{aligned}\ ] ] to determine the dominant terms in eqs . and , we write the approximate expressions for eqs . and in the limit , where @xmath5 is small , and @xmath207 is big ( @xmath208 ) .
we thereby get @xmath209 the second term on the right hand side of eq .
is included for the sole purpose of finding an approximation for @xmath210\approx \operatorname{re}[\beta_2]\operatorname{im}[\beta_3]$ ] . an expression for the detuning @xmath5
can be found using eq . .
expanding its left hand side in @xmath5 around @xmath211 to second order results in @xmath212 under the assumption that @xmath213 , we choose the solution , where @xmath214 : @xmath215 expanding this around the limit of large @xmath162 we find @xmath216 the first term on the right hand side of eq .
could also be derived from the knowledge that the dispersion relation is approximately quadratic for small @xmath5 , i.e. @xmath217 with the mass @xmath218 given in ref .
@xcite , but the other two terms result from higher order corrections to this approximation .
when we calculate the fidelity @xmath78 in sec .
[ sec_analytical_fidelity ] below , we find that it is maximal for a detuning @xmath219 , and the second one is proportional to @xmath220 .
hence , the second one is smaller for large @xmath162 and can be neglected .
we keep the third term , since it depends on @xmath53 and will be important when discussing the bandwidth of the protocol . using only the first term in eq . and inserting it into the first term of eq . , we find @xmath221 upon inserting into the second term of eq .
, we find @xmath222\approx \frac{\gamma_\text{1d}\gamma ' } { 4\delta_\text{c}^2 \left(1-(\pi^2|\omega_0|^2)/(\gamma_\text{1d}^2n^2)\right)^2 } \approx \frac{\gamma_\text{1d}\gamma'}{4\delta_\text{c}^2 } + \frac{|\omega_0|^2\pi^2\gamma'}{2\delta_\text{c}^2\gamma_\text{1d}n^2}.\end{gathered}\ ] ] using the first and the third terms of eq . and inserting them into eq . , we find @xmath223 combining eqs . and and neglecting a term of order @xmath224 results in @xmath225\approx\operatorname{re}[\beta_2]\operatorname{im}[\beta_3 ] \approx \frac{\pi^2\gamma'}{8\delta_\text{c}n^2 } + \frac{|\omega_0|^2\pi^4\gamma'}{8\delta_\text{c}\gamma_\text{1d}^2n^4}.\end{gathered}\ ] ] invoking eq .
again and neglecting the @xmath53 dependent terms for a moment , we see that @xmath226 , @xmath227 and @xmath228\propto n^{1/4}$ ] . hence , @xmath229 ; @xmath230)^{-1}$ ] and we can approximate eqs . and
by @xmath231 \approx \frac{\gamma_\text{1d}\gamma'n}{8\delta_\text{c}^2 } + \frac{|\omega_0|^2\pi^2\gamma'}{4\delta_\text{c}^2\gamma_\text{1d}n},\\ \label{t_0_sm } & t_0\approx(-1)^{n-1}\del{1+i\frac{2n^3}{\pi^2}\beta_2\operatorname{im}[\beta_2\beta_3 ] } \approx ( -1)^{n-1}\del{1-\frac{\gamma_\text{1d}\gamma'n}{8\delta_\text{c}^2 } -\frac{|\omega_0|^2\pi^2\gamma'}{4\delta_\text{c}^2\gamma_\text{1d}n}}.\end{aligned}\ ] ] since the number of atoms is @xmath232 , the above expressions correspond to eqs . and of the main article , except for the removal of the overall phase factor @xmath233 for the transmission coefficient @xmath107 ( discussed in sec .
[ sec_sagnac_interferometer_and_phases ] below ) . in fig .
[ fig_r_and_t_plot_omega](a ) we plot @xmath1 as a function of @xmath53 and show that the analytical expression in eq .
matches the full expression in eq .
evaluated at the resonance frequency . to account for the non - zero bandwidth of the scattered photon
, we also need the width of the resonance . after expanding the reflection coefficient @xmath106 around @xmath109 ,
we get @xmath234 where @xmath235 is given by eq .
and @xmath236 is the width .
since we approximately have that @xmath237 , the width of the transmission resonance is @xmath238 . in fig .
[ fig_r_and_t_plot_omega](b ) we compare eq . with the numerically computed width . and
transmittance without a stored photon @xmath1 of an ensemble of @xmath3type atoms plotted as functions of the rabi frequency of the classical drive @xmath53 and evaluated at the transmission resonance closest to @xmath201 ( see fig .
[ fig_r_and_t_plot ] of the main article ) , i.e. at @xmath239 . both @xmath107 and @xmath51
are calculated either directly from the transfer matrix with no approximations ( `` full '' ) , which for @xmath107 is given by eq . , or from the approximate expressions in eqs .
and respectively ( `` approximate '' ) .
( b ) the width of the transmission resonance calculated either from the full expression as @xmath240{\text{re}\sbr{\sqrt{4/\partial_\delta^2 t_0(\delta)}}}_{\delta=\delta_\text{res}}$ ] or using the approximate expression ( dash - dotted black ) .
the parameters for both ( a ) and ( b ) are : number of atoms @xmath7 , @xmath13 , and @xmath9 .
( the same as in fig .
[ fig_r_and_t_plot ] of the main article except for @xmath53 , which is varied here . ) ] here , we derive the scattering coefficients @xmath51 and @xmath50 given by eqs . and of the main article .
the starting point is the assumption that the photon has been stored in a single atom that is placed at the anti - node of the standing wave of the classical drive ( storing in an atom that is on the node will have a negligible change in the scattering properties , unless @xmath20 is close to unity ) .
the storage of a photon in the atom transfers it from state @xmath37 to state @xmath39 , such that it behaves like a resonant two - level atom ( given by the @xmath241 transition ) that is described by the transfer matrix @xmath242 with @xmath243 hence , the transfer matrix for the unit cell containing the stored photon will be given by @xmath244 instead of eq . .
if we assume that the photon is stored in the unit cell with index @xmath245 , the transfer matrix for the whole ensemble @xmath246 is given by @xmath247 from eq .
we can find the scattering coefficients using @xmath248 where @xmath249 are the elements of the matrix @xmath246 similar to the definition of the elements of the matrix @xmath171 .
for the numerical calculations , eqs . and are used directly , and we set @xmath250 in eq . . for the analytical calculations , we can find approximate expressions for the scattering coefficients , but the procedure is rather involved .
we shall therefore restrict ourselves to a brief discussion of the main steps .
we do several simplifications on the ( very complicated ) expressions resulting from eqs . and .
we use the fact that @xmath251 and the approximate expression @xmath252 . also , while expanding the numerator and denominator around big @xmath162 , we use that @xmath226 ( a consequence of eqs . and ) to determine which terms can be neglected .
then we replace the index of the unit cell with the stored photon @xmath245 by @xmath253 , where @xmath52 is the rescaled position coordinate .
after further approximating @xmath254 and @xmath255 , we get @xmath256 next , we insert the expressions for @xmath257 and @xmath258 with the approximation @xmath259 , expand around @xmath56 , and use @xmath260 to identify which terms are dominant for @xmath261 .
this results in @xmath262 we see that the resulting expressions do not depend on @xmath53 .
this is a consequence of approximating @xmath259 .
if we use @xmath208 together with eq . , we find corrections from the dependence on @xmath53 .
we only need the first term in eq . to find the lowest order correction due to @xmath53 . at @xmath56 , and expanding around large @xmath162 , we have @xmath263 in fig .
[ fig_r_and_t_plot_omega](a ) , we plot @xmath0 as a function of @xmath53 and show that the analytical expression in eq .
matches the full expression calculated using eq .
( evaluated at the resonance frequency ) without doing any approximations . in eq . of the main article
, we include all error terms of eq . and also add the @xmath53 dependent error term from eq . . likewise , in eq
. of the main article
, we include all error terms of eq . and also add the @xmath53 dependent error term from eq . . for the dualv scheme , we need two - mode ( @xmath264 ) transfer matrices to describe the @xmath122 and @xmath123 polarized modes .
the transfer matrices for the atoms in eq .
have the form @xmath265 ( @xmath266 is the index of the atom ) , where @xcite @xmath267 and @xmath268 in the above , we have used the same convention as in eqs . , i.e. that the first element of @xmath269 is the @xmath122 component and the second element is the @xmath123 component .
hence , in eq . , @xmath270 for example means the reflection coefficient , such that the incident @xmath122 field is reflected also as a @xmath122 field .
similarly for the other reflection coefficients .
if the atom is in state @xmath37 ( without a stored photon ) , the elements of @xmath271 are given by @xcite @xmath272\\ & r_{j,-+ } = -\frac{i(\gamma_\text{1d}/2)|\omega_0|^2 } { \delta_\gamma^2\delta -2\delta_\gamma|\omega_0|^2}e^{2ik_0 z_j},\\ & r_{j,+- } = -\frac{i(\gamma_\text{1d}/2)|\omega_0|^2 } { \delta_\gamma^2\delta -2\delta_\gamma|\omega_0|^2}e^{-2ik_0 z_j},\end{aligned}\ ] ] where @xmath273 .
if the atom is in state @xmath39 ( with a stored photon ) , it acts as a resonant v - type atoms , and hence the elements of @xmath271 are given by @xmath274 we will only calculate the reflection and transmission of ensembles of dual - v atoms numerically .
here , we calculate the result of scattering from the sagnac interferometer shown in fig .
[ fig_setup](d ) of the main article .
the sequential picture of the scattering is that the incident field on one of the ports is split by the 50:50 beam splitter , gets scattered by the ensemble , and then the transmitted and reflected parts will again interfere on the same beam splitter .
thus the matrix that relates the outputs to the inputs can be written @xmath275 where the matrix @xmath276 describes the beam splitter , and the matrix @xmath277 describes the ensemble .
we choose the phases of the beam splitter , such that it performs the hadamard operation on the field , i.e. @xmath278 the ensemble can , in general , have different transmission and reflection coefficients depending on , whether the field is incident from the left or right .
therefore , we write @xmath279 where @xmath280 and @xmath281 are respectively the reflection and transmission coefficients when the field is incident from the left ( propagating in the positive direction ) , and @xmath282 and @xmath283 are respectively the reflection and transmission coefficients when the field is incident from the right ( propagating in the negative direction ) . multiplying the matrices ,
we get @xmath284 for a standard ( non - rotating ) sagnac interferometer where @xmath285 and @xmath286 , we recover the well known result , that light always leaves the port in which it is incident @xcite .
because the equality @xmath285 need not be true , the off - diagonal entries of @xmath287 are , in general , not zero and describe the leakage of the incident power to the other port of the sagnac interferometer . however , due to the sequential operation of the gate , this leakage does not introduce any logic errors .
since scattering of photon @xmath28 happens , while photon @xmath25 is stored inside the ensemble , leakage of photon @xmath28 into the rail that encodes state @xmath27 ( see fig . [
fig_setup](c ) of the main article ) is separated in time from the subsequent retrieval of photon @xmath25 and hence can be either absorbed or rerouted along a different path . even though the off - diagonal entries of @xmath287 are non - zero in general , they are strongly suppressed in the ideal limit , since e.g. both @xmath280 and @xmath282 approach the same value ( either @xmath288 or @xmath211 depending on , whether a photon was stored in the ensemble or not ) . as a concrete example , for the @xmath3type scheme , we can use eq . and find @xmath289 where @xmath290 is the same as @xmath106 , and @xmath291 is the reflection coefficient where the field is incident from the right instead of the left . the difference between eqs . and
is only in the sign of the term @xmath292 in the numerator . as discussed above eq .
, this term is much smaller than @xmath257 in the limit of large number of atoms and has therefore been neglected in eq . .
hence , eq . can serve as an approximate expression for both @xmath290 and @xmath291 .
we also note that it can be shown that the transmission coefficient for any @xmath160 transfer matrix is independent of whether the field is incident from one side or the other ( given by eq . for @xmath107 ) .
however , we still account for the possible difference in the scattering coefficients in the numerical calculations ( both for @xmath160 and @xmath264 transfer matrices ) by defining the scattering coefficients @xmath103 and @xmath293 in eqs . and
as @xmath294 where the scattering coefficients with `` @xmath125 '' in the subscript assume incident photon from the left of the ensemble and the scattering coefficients with `` @xmath126 '' in the subscript assume incident photon from the right of the ensemble .
we also need to discuss the choice of the phases in the interferometer . in eq .
there is an implicit assumption that the lengths of free propagation in the interferometer are chosen such that the distance from the beam splitter to either end of the ensemble is a multiple of the wavelength @xmath295 .
hence , the phase of free propagation is equal to unity for these parts .
what we call `` ensemble '' may also contain some length of free propagation to the right of the ensemble to adjust the phases of its scattering coefficients . to illustrate the necessity of phase adjustment , note that there is an overall phase factor @xmath233 in eqs .
, , and compared with eqs . and of the main article .
if this phase factor is not removed , it will directly appear as a factor of @xmath103 , when @xmath296 and @xmath297 .
hence , the ideal value @xmath62 can not be obtained even in the limit of many atoms .
there are similar phase factor considerations for @xmath293 due the phase factor of the reflection coefficient @xmath291 , while @xmath290 is assumed to be at @xmath298 and hence does not acquire phase factors with changing ensemble length .
if the overall phase of the transmission coefficients is adjusted , the correct phase of @xmath103 and @xmath293 is obtained and the controlled - phase gate operation can approach the ideal limit as the number of the atoms is increased . for the dual - v scheme , we can choose any inter - atomic spacing @xmath101 , which is not a multiple of @xmath159 ( see fig .
[ fig_cphase_fidelity_kd ] below ) .
for this general case , we expect that instead of the overall phase factor @xmath233 , the transmission coefficients have the phase factor @xmath299 , which we likewise remove by adding a distance of free propagation @xmath300 to the right of the ensemble chosen such that @xmath301 . multiplying this extra matrix of free propagation
modifies the scattering coefficients according to @xmath302 the physical interpretation of this mathematical result is that , since the free propagation was added on the right of the ensemble , then reflection for the field incident from the left ( @xmath290 ) is unaffected , while the reflection coefficient for fields incident from the right ( @xmath291 ) acquires twice the propagation phase . the transmission coefficients only acquire the propagation phase once .
the phase adjustments above effectively force the interferometer round trip length to be equal to an odd multiple of half wavelengths ( @xmath303 ) .
hence , the phase adjustment can be made independent of the number of atoms .
( `` dispersion '' ) , the fully discrete theory of sec . [ discrete_eit_storage_retrieval ] , and the storage kernel of sec . [ continuum_eit_storage_retrieval ]
( `` kernel '' ) .
contrary to the fidelity calculations , here we assume eit storage from one side ( left ) only to show the influence of random placement of the atoms more clearly .
the common parameters for the two subplots are @xmath13 , @xmath7 , @xmath98 , @xmath304 ( width of the stored gaussian spin wave ) .
( a ) regularly placed atoms are assumed with inter - atomic distance @xmath97 .
all three curves are nearly indistinguishable from each other .
( b ) randomly placed atoms , where the position of each atom is chosen from a uniform distribution over the whole ensemble .
the average density is the same as in ( a ) .
the curve for the fully discrete storage is clearly distinct from the two others and exhibits rapid variation with position . ]
we will model the eit storage and retrieval process in three different ways : 1 . using the dispersion relation ( see for instance ref .
@xcite and sec .
[ dispersion_eit_storage_retrieval ] below ) .
2 . using the fully discrete theory ( see ref .
@xcite and sec .
[ discrete_eit_storage_retrieval ] below ) .
3 . using the storage and retrieval kernels ( see ref .
@xcite and sec .
[ continuum_eit_storage_retrieval ] below ) .
we will consider the so - called adiabatic eit storage @xcite , where a single - photon wave packet is incident on the ensemble and is mapped onto a spin wave ( a superposition of states , where a single atom is in state @xmath38 and the rest are in state @xmath37 ) .
we assume a constant rabi frequency of the classical drive @xmath305 , but choosing a co - propagating classical drive with rabi frequency @xmath306 will only change the spatially - dependent phase factor of the stored spin wave . in the limit of high storage efficiency , the temporal profile of the photon and the spatial profile of the stored spin wave will approximately have the same form .
e.g. if a photon with gaussian temporal wave packet is stored , the resulting spin wave will have a gaussian spatial profile .
this is a consequence of the time - independent rabi frequency of the classical drive and allows us to use the eit dispersion relation to describe the storage and retrieval .
the eit dispersion relation is used to gain intuition about the storage and retrieval process and also for the analytical calculations . in the numerical calculations of the fidelities and success probability , the atoms are always modeled as being discrete . using the fully discrete eit storage and retrieval model ( sec .
[ discrete_eit_storage_retrieval ] ) becomes very costly as the number of atoms increases .
therefore , to be able to calculate fidelities with a large number of atoms , we instead use the less computationally demanding storage and retrieval kernels derived in the continuum model and suitably discretized ( sec . [ continuum_eit_storage_retrieval ] ) . in fig .
[ fig_eit_storage_models_comparison ] , we show that the three models agree very well for the case of regularly placed ensembles with big optical depth @xmath307 . for randomly placed atoms ,
the fully discrete theory gives noticeably different results compared to using the discretized continuum theories ( dispersion relation or storage kernel ) .
since we assume regular placement for the curves in fig .
[ fig_cphase_fidelity_natoms ] of the main article , the discretized storage and retrieval kernels are sufficient .
we also verify this in fig .
[ fig_cphase_fidelity_lambda_check_discrete ] below for the @xmath3-type scheme , and the same conclusion holds for the dual - v scheme . in fig .
[ fig_cphase_fidelity_random_placement ] below , where we show fidelities for randomly placed atoms , we only use the fully discrete model for storage and retrieval , even though it limits the maximal number of atoms that can be used in the calculations .
the adiabatic eit storage and retrieval can be modeled in a particularly simple way if the influence of the interface between the atomic medium and vacuum is ignored . to use the eit dispersion relation , we need to assume that the ensemble is of an infinite extent .
however , to compute the the storage and retrieval efficiency , we need to assume propagation through a finite ensemble . in the calculations below ,
this is reflected in infinite bounds for the integration but a finite propagation length .
the only processes that happen in this model is that the stored photon wave packet broadens in space as it propagates , and its norm decays due to spontaneous emission .
the eit storage and retrieval efficiency will then be the norm of the wave packet that has propagated for the full length of the ensemble @xmath45 ( with a stop at @xmath308 to allow for the second photon to be scattered off the ensemble ) .
the eit dispersion relation is @xcite @xmath309 with @xmath310 in rescaled coordinates @xmath52 and wave vectors @xmath311 the dispersion can be written @xmath312 with @xmath313 we only consider incident photons where the temporal profile is gaussian .
when such a photon is mapped onto a stored excitation , this results in an approximately gaussian spatial profile of the form @xmath314 where the rescaled quantities are @xmath77 , @xmath315 , @xmath316 . by fourier transforming this wave function to get @xmath317 , multiplying the fourier transform by @xmath318 and fourier transforming back we find @xmath319 the norm squared of the wave packet at time @xmath320 is given by @xmath321 the combined storage and retrieval efficiency is given by eq . with @xmath322 , i.e. the time required to pass the whole ensemble .
we thereby get @xmath323 using the intuition about eit from sec .
[ dispersion_eit_storage_retrieval ] , one can implement the numerical simulations of eit storage and retrieval accounting for the discrete nature of atoms .
this approach is very similar to the `` electric field elimination '' approach of ref .
the main difference is that , since storage and retrieval of a single photon only requires calculating the dynamics in the atomic single - excitation manifold , we can eliminate the electric field directly in the schrdinger picture instead of the heisenberg picture like in ref .
@xcite .
the electric field operator can be written @xmath324 , where @xmath325 and @xmath326 are the parts of the field that respectively propagate to the right ( positive direction ) and left ( negative direction ) .
the two parts are assumed to be completely separate fields , and their commutators are @xmath327 = \delta_{\alpha\beta}\delta(z - z'),\end{gathered}\ ] ] where @xmath328 .
the hamiltonian for the ensemble of @xmath3-type atoms coupled to the electric field is @xmath329 , representing the atomic , interaction and photonic hamiltonian respectively .
the three parts of the hamiltonian are given by @xmath330\bigg\ } , \displaybreak[0]\\ & \hat{h}_\text{p } = -i\hbar c\int \sbr{{\hat{\mathcal{e}}}_{+}^\dagger(z)\dpd{{\hat{\mathcal{e}}}_{+}(z)}{z } -{\hat{\mathcal{e}}}_{-}^\dagger(z)\dpd{{\hat{\mathcal{e}}}_{-}(z)}{z}}\dif z,\end{aligned}\ ] ] where @xmath331 is the group velocity in the waveguide .
we assume that the detuning @xmath332 is always set to zero during storage and retrieval ( this was also assumed in sec .
[ dispersion_eit_storage_retrieval ] above ) .
however , if desired , off - resonant ( @xmath333 ) eit storage and retrieval is also possible @xcite , and hence we keep the @xmath332 term in the equations of motion below . on the single - excitation manifold
, the state can be written @xmath334 from the schrdinger equation , we get the equations of motion for the atomic coefficients @xmath335 where @xmath336 . for the electric field coefficients
@xmath337 we have the equations @xmath338 these equations can be formally solved , so that we obtain @xmath339 where @xmath340 are the input fields , and @xmath166 is the heaviside theta function . inserting these solutions into eq . and
approximating @xmath341 @xcite , we find @xmath342 the fidelity calculations in sec . [ sec_fidelity_general ] above are formulated in terms of @xmath343 and @xmath344 , which are respectively the input field to be stored and the retrieved output field .
( there are two different output fields , @xmath345 and @xmath346 , but for the discussion of storage and retrieval , the difference between them is not important . ) we need to account for the beam splitter in the sagnac interferometer .
hence , the relations between the fields in this section and sec .
[ sec_fidelity_general ] are @xmath347 note that storage is done from two directions in order to ensure that no excitations are stored on atoms at the nodes of the standing wave of the classical drive applied during the scattering for the @xmath3-type scheme
. the conditions for this may not necessarily be the same as the conditions derived for scattering from the sagnac interferometer ( see sec .
[ sec_sagnac_interferometer_and_phases ] ) .
if this is an issue , it can be compensated by adjusting the position of the atoms between storage and scattering , e.g. , by adjusting the phases of the trapping lasers or of the classical drives .
for the dual - v scheme , there is no phase requirement during storage and this is not a concern .
as the input wave function , we assume @xmath348 where the width and central time @xmath349 are defined in terms of the eit group velocity .
the factor @xmath350 ( a real number , since @xmath351 is imaginary ) in the definition of @xmath352 is introduced to compensate for the spin wave becoming wider as it propagates inside the ensemble ( see eq . ) .
this particular factor is chosen such that the stored gaussian spin wave ( centered at the position @xmath308 ) has width @xmath353 . in the end , since we optimize over @xmath353 , this adjustment has no effect on the final values of the numerically calculated fidelities and success probability .
however , it ensures that the optimal @xmath353 in the numerical calculations is similar to the optimal value found by neglecting broadening of the spin wave under propagation . in the fully discrete model , we do not explicitly calculate the storage and retrieval kernels that appear in eqs . and .
instead , we calculate the action of these kernels on respectively a specific @xmath343 or a spin wave given by the coefficients @xmath354 . for storage , this amounts to numerically solving eqs . and for a given @xmath343 and the initial condition @xmath355 at @xmath356 .
we take the stored spin wave to be the coefficients @xmath354 at @xmath357 .
this final time is the sum of the time for propagation through vacuum and ( half of ) the eit medium . for retrieval ,
are solved with @xmath358 under the initial conditions that at @xmath356 the coefficients @xmath359 are set to the spin wave that is to be retrieved , and @xmath360 . at each time step , we calculate @xmath344 using eq . with @xmath361 along with eq .
we assume that the retrieval happens until @xmath362 , i.e. the time that it takes for the eit polaritons to move through the whole ensemble . as an approximation to the fully discrete theory of sec .
[ discrete_eit_storage_retrieval ] , one can use the continuum theory of ref .
the continuum approximation allows the derivation of explicit expressions for the linear maps ( given in terms of integration with a particular kernel ) describing storage and retrieval . before doing the continuum approximation , we first rewrite the hamiltonian in terms of the collective operators @xmath363 where @xmath364 , and @xmath365
is the average density .
these collective operators have the equal time commutation relation @xmath366 & = \frac{1}{n_0}\delta(z - z ' ) ( \delta_{\beta,\alpha'}\hat{\sigma}_{\alpha\beta ' } -\delta_{\beta',\alpha}\hat{\sigma}_{\alpha'\beta}).\end{aligned}\ ] ] using the collective atomic operators , the hamiltonian is @xmath367\dif z\\ & \hat{h}_\text{i}=-\hbar n_0\int \bigg\{\left[\hat{\sigma}_{bc}(z)\omega_0 + \text{h.c.}\right ] + g\sqrt{2\pi}\left[\hat{\sigma}_{ba}(z){\hat{\mathcal{e}}}(z ) + \text{h.c.}\right]\bigg\}\dif z\\ & \hat{h}_\text{p}=-i\hbar c\int\bigg[{\hat{\mathcal{e}}}_{+}^{\dagger}(z)\dpd{{\hat{\mathcal{e}}}_{+}(z)}{z } -{\hat{\mathcal{e}}}_{-}^{\dagger}(z)\dpd{{\hat{\mathcal{e}}}_{-}(z)}{z}\bigg]\dif z.\end{aligned}\ ] ] instead of the state , we use @xmath368 note that if , for example , the excitation is entirely in the metastable states at a time @xmath369 , we have @xmath370 \dif z = \frac{1}{l}\int \int |s(z , t)|^2\dif z,\end{gathered}\ ] ] where we have used the continuum approximation @xmath371 @xcite together with the low excitation approximation @xmath372 to get @xmath373 and @xmath374 .
. also gives the normalization condition for @xmath277 in the continuum model ( and similarly for @xmath375 ) .
the equations for the coefficients are @xmath376 as an extension to the theory of ref .
@xcite , we want to consider an input field which is , in general , incident from both sides instead of only one . the approach that we use is to consider the parts of the single photon excitation incident from both sides as being stored separately from each other . when doing this we ignore the fact that the two parts have opposite spatial phases @xmath377 , which interfere inside the ensemble to produce a spatially modulated spin wave with amplitude proportional to @xmath36 .
this spatial modulation of the stored spin wave is very important for the @xmath3scheme , since the part of the excitation that is stored on the nodes of the standing wave of the classical drive does not significantly change the scattering properties of the ensemble ( see sec .
[ lambda_r_t_with_stored_photon ] ) . expressed in terms of the notation introduced in the fidelity calculations ( see eq . ) we have @xmath378 for odd @xmath266 with atoms placed at positions @xmath18 ( @xmath19 ) , i.e. @xmath379 being an odd multiple of @xmath380 . with the photon incident from both sides ,
we ensure that there is no amplitude on these atoms , since @xmath381 for odd @xmath266 .
this is correctly reproduced by the fully discrete model of sec .
[ discrete_eit_storage_retrieval ] , since it always accounts for the phases of free propagation . on the other hand , due to the removal of the rapidly varying spatial phases in the continuum theory ,
this factor @xmath382 is not present in the two separate parts of the stored spin wave . to compensate for this
, we set @xmath383 for odd @xmath266 . as we show in fig .
[ fig_cphase_fidelity_lambda_check_discrete ] , this phenomenological adjustment of the reflection coefficient gives results that are essentially indistinguishable from the results produced by the fully discrete model of sec .
[ discrete_eit_storage_retrieval ] . for the dualv scheme , no such adjustment of the reflection coefficients
is needed neither in the continuum nor the discrete model , which also produce indistinguishable results . type scheme with different models used for eit storage and retrieval plotted as functions of the number of atoms @xmath12 with fixed @xmath13 .
the `` discrete '' plots use the fully discrete theory of sec .
[ discrete_eit_storage_retrieval ] , and the `` kernel '' plots use the discretized storage and retrieval kernels derived in sec .
[ continuum_eit_storage_retrieval ] . in both cases , the optimal @xmath42 and @xmath79 ( width of the stored gaussian spin wave )
are found by doing numerical optimization using the discretized storage and retrieval kernels , since the fully discrete model is much more computationally demanding . ]
the photons incident from the left and right couple to different components of the atomic coefficients , which can be written @xmath384 after inserting these definitions into eqs .
, and and separating the components , we get @xmath385 to solve for @xmath386 , @xmath387 and @xmath388 , the approach in ref .
@xcite can be directly used .
it consists of transforming into the coordinates @xmath52 and @xmath389 , laplace transforming in space , solving the algebraic equations of the laplace transforms ( under the adiabatic approximation @xmath390 ) and then laplace transforming back . as a minor modification , we also transform back from the co - propagating time coordinate @xmath389 to the original @xmath369 ( by replacing all @xmath391 by @xmath369 in the final expressions ) . solving for @xmath392 , @xmath393 and @xmath394
is simply a spatial reflection of the original problem around @xmath56 .
we find that the stored spin wave is @xmath395 where the storage kernel ( in the adiabatic approximation ) is @xmath396 written in terms of @xmath397 ( related to the optical depth by @xmath398 ) and the modified bessel function of the first kind @xmath399 .
the retrieved field is @xmath400 where the retrieval kernel is @xmath401 the stored spinwaves are chosen such that they are centered at @xmath56 and become narrower in space for increasing optical depth ( see the fidelity derivations in sec .
[ sec_analytical_fidelity ] below ) .
the input and output photon wave functions are centered around times @xmath402 . inserting these mean values into eqs . and
, we see that the argument of @xmath399 becomes very big , since @xmath403 and @xmath404 .
this allows us to use the asymptotic expansion @xmath405 , which is valid for @xmath406 and @xmath407 . in this limit ,
the kernels become @xmath408 these formulas have a better numerical behavior compared to eqs . and , since one does not need to multiply the value of the @xmath399 ( exponentially large ) with an exponentially small factor .
hence , we always use eqs . and in the numerical calculations . the relations between the fields in this section and sec
[ sec_fidelity_general ] are @xmath409 the results of this section assume that the atoms can be modeled as a continuum , but the scattering coefficients in eq .
are only given at the discrete atom positions .
hence , we need to sample the resulting continuum solutions at the discrete positions of the atoms . in the numerical calculations ,
the continuum solutions are always sampled as if the atoms were placed regularly independent of the actual placement . to justify regular sampling , we note that instead of the rescaled position coordinate @xmath52 , one could use @xmath410 @xcite , where @xmath411 is the local density of the atoms . for an average density @xmath365 ,
this rescaled coordinate is equivalent to @xmath52 . for the local density @xmath412 ,
the rescaled position becomes @xmath413 , where @xmath166 is the the heaviside theta function . with the convention @xmath414 , each @xmath415
is transformed into @xmath416 regardless of the actual value of @xmath415 . as shown in fig .
[ fig_eit_storage_models_comparison](b ) for random placement of the atoms , even though the continuum solutions seem to be able to reproduce the results of the fully discrete theory to some degree , there are still significant differences , which we believe to be caused by reflection of parts of the propagating excitation due to disorder , which are not accounted for in the continuum theory . hence , with randomly placed atoms ,
the fully discrete model is required .
having in mind both the separation of the spin waves into two independent parts and the sampling of the continuum solutions at regular intervals , we can define the storage and retrieval kernels that will be used in eqs . and .
we define the vector representing the spin wave to have @xmath417 elements for two separately stored spin waves that have fields incident either from the left or from the right as the input . using the same storage time @xmath418 as for the discrete model in sec .
[ discrete_eit_storage_retrieval ] , the storage kernel is @xmath419 where we have assumed that the coefficients corresponding to @xmath388 are stored in the part of the vector with indices @xmath19 , and the coefficients corresponding to @xmath394 are stored in the part of the vector with @xmath420 .
the retrieval kernel is @xmath421 as a consequence of having a spin wave vector with @xmath417 elements , in eq .
we define @xmath422 for @xmath420 .
here , we derive [ f_cj_1,f_cj_cond_1,f_cj_2,f_cj_cond_2 ] of the main article .
we also derive the bandwidth dependent correction to eq . of the main article , which is used in the discussion of the gate time .
from [ r_ensemble_n , t_ensemble_n , r_imp_ensemble_n , t_imp_ensemble_n ] of the main article , we find the approximate reflection coefficients of the combined system of the atomic ensemble and the sagnac interferometer .
they are @xmath423 note that in the symmetrized reflection coefficient @xmath424 , the linear terms proportional to @xmath425 ( which are present in eqs . and of the article ) cancel each other . using the expression for the spin wave given by eq .
with @xmath426 we get @xmath427 using eq . from the main article ,
an approximation for the unconditional cj fidelity is @xmath428 where @xmath429 , and all the error terms ( including @xmath430 ) are assumed to be small . in principle , if we want to optimize the above expression , we should optimize with respect to @xmath42 and @xmath79 simultaneously ( we do this in the numerical calculations ) . here
, we use an approximate optimization procedure , which ignores the fact that one of the error terms depends on the product of @xmath42 and @xmath79 . as we will see , however , this error term is smaller than the error terms , which only depend on @xmath42 for fixed @xmath20 and large @xmath12 . therefore , we first optimize @xmath78 over @xmath42 separately and then use the optimal value of @xmath42 to optimize over @xmath79 .
the optimal value of @xmath42 is determined by the condition that the third and fourth error terms on the right hand side of eq .
are equal , i.e. @xmath431 this results in @xmath432 inserting this value of @xmath42 into eq .
we obtain @xmath433 we use this expression to optimize over @xmath79 .
the optimal @xmath79 is obtained when @xmath434 from this condition we get @xmath435 and with this value of @xmath79 , eq
. becomes @xmath436 we see that for @xmath437 , the last term in this expression approaches zero faster than the second one .
hence , in eqs .
and of the main article we have omitted this term .
the difference between eq .
and eq . is whether we set respectively @xmath438 or @xmath439 .
next we calculate the conditional fidelity @xmath81 . since it is given by @xmath75 , the expansion of the ratio will contain higher order error terms than the expansion of the unconditional fidelity @xmath78 . hence , we need an expansion of @xmath78 with more terms than in eq . .
including the second order terms and dividing out @xmath440 ( since it gets canceled in @xmath81 ) , we get @xmath441 using eq . , we also get the success probability @xmath442 where @xmath443 using the above expressions , the conditional fidelity can be written @xmath444 where @xmath445 using the @xmath42 from eq .
, we get @xmath446 if we choose @xmath438 , @xmath447 is the dominant error term with the value @xmath448 in which case @xmath449 can be neglected .
if we choose @xmath450 , we get @xmath451 , and we need to keep @xmath449 .
the value of @xmath449 depends on the width of the stored gaussian @xmath79 . for simplicity
, we use the value of @xmath79 given by eq . , which makes the unconditional fidelity maximal .
with this choice , the conditional fidelity is given by eq .
of the main article .
a comparison of the analytical formulas and the numerical results for @xmath78 and @xmath81 is shown in fig .
[ fig_cphase_fidelity_big_natoms ] .
type scheme plotted as functions of the number of atoms @xmath12 with fixed @xmath13 .
dotted green curves are the numerically calculated fidelities with @xmath76 .
dash - dotted black curves are the approximate analytical results given by eq .
( unconditional ) and eq .
( conditional ) of the main article .
solid blue curves are the numerically calculated fidelities with @xmath30 chosen such that the entanglement swap fidelity ( which is approximately equal to the conditional choi - jamiolkowski fidelity as shown in fig
. [ fig_cphase_fidelity_f_swap ] ) is maximal .
the dashed red curves are the approximate analytical results given by eq .
( unconditional ) and eq .
( conditional ) of the main article . for storage and retrieval in the numerical calculations ,
we use the discretized continuum theory described in sec .
[ continuum_eit_storage_retrieval ] . ]
now we derive the correction to the unconditional fidelity due to non - zero bandwidth of the scattered photon @xmath28 .
the general expression for the fidelity is given by eq . .
as discussed in the main article , we ignore the non - zero bandwidth in the last term since the variation of @xmath51 and @xmath50 with frequency around the resonance detuning @xmath109 is smaller than variation of @xmath106 and @xmath107 .
hence , we approximate @xmath452 such that the expression for the fidelity becomes @xmath453 using eqs . and together with expressions @xmath237 and @xmath54 and defining the spectral width of photon @xmath28 by @xmath454 , we get @xmath455 using the optimal @xmath42 from eq . , this becomes @xmath456 to minimize the sum of the two last error terms in this expression , we optimize over @xmath53 and get @xmath457 for the optimal @xmath458 i.e. compared to eq . , there is an extra error term that depends on @xmath459 .
while the @xmath3-type scheme is highly sensitive to the exact placement of the atoms , the dual - v scheme is much less sensitive . to verify this
, we numerically evaluate the gate performance for various placements of the atoms .
first , we investigate the dependence for regularly placed dual - v atoms with different interatomic spacings @xmath101 . in fig .
[ fig_cphase_fidelity_kd ] , we see that for @xmath460 different from multiples of @xmath380 , both the conditional and unconditional fidelities are approximately constant .
second , we consider randomly placed dual - v atoms . in fig .
[ fig_cphase_fidelity_random_placement ] , the fidelities with regular and random placement are seen to have qualitatively the same behavior , but the performance with random placement is slightly worse .
times the photon wave vector @xmath16 for fixed number of atoms @xmath7 and @xmath13 .
for storage and retrieval , we use the discretized storage and retrieval kernels derived in sec .
[ continuum_eit_storage_retrieval ] . ] with fixed @xmath13 .
the regularly placed atoms have have positions @xmath461 for @xmath97 and @xmath462 .
the positions of the randomly placed atoms are chosen from the uniform distribution over the whole ensemble and then sorted in increasing order . for random placement
, we average over 1000 ensemble realizations . for storage and retrieval , the fully discrete model of sec .
[ discrete_eit_storage_retrieval ] is used . for regular placement
, it gives the same results as the discretized storage and retrieval kernels derived in sec .
[ continuum_eit_storage_retrieval ] , but for random placement , the fully discrete model has significant deviations from the discretized storage and retrieval kernels as shown in fig .
[ fig_eit_storage_models_comparison ] . the optimal @xmath42 and @xmath79 ( width of the stored gaussian spin wave )
are found by doing numerical optimization using the discretized storage and retrieval kernels , since the fully discrete model is much more computationally demanding . ]
11ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physreva.77.063823 [ * * , ( ) ] @noop @noop link:\doibase 10.1103/physreva.71.062310 [ * * , ( ) ] link:\doibase 10.1103/physreva.52.1394 [ * * , ( ) ] @noop http://stacks.iop.org/0953-4075/39/i=5/a=001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.103902 [ * * , ( ) ] http://stacks.iop.org/1367-2630/13/i=2/a=023003 [ * * , ( ) ] http://stacks.iop.org/1367-2630/17/i=11/a=113001 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.033805 [ * * , ( ) ] | we propose a method to induce strong effective interactions between photons mediated by an atomic ensemble . to achieve this
, we use the stationary light effect , which makes the atomic ensemble behave as an optical cavity and thereby increases the interaction strength .
regardless of the optical depth per atom , the interaction strength between the photons can be enhanced by increasing the total number of atoms ( total optical depth ) . for sufficiently many atoms
, the setup can be viable as a controlled - phase gate for photons .
we derive analytical expressions for the fidelities for two modes of gate operation : deterministic and heralded conditioned on the presence of two photons at the output .
the latter mode can be used to improve the secret key rate for quantum repeaters compared to using only linear optics .
optical photons are ideal carriers of quantum information over long distances , and such quantum communication may enable a wealth of applications @xcite .
quantum information processing with photonic qubits is , however , severely limited by the lack of efficient two - qubit gates . in principle , such gates could be realized by strongly coupling photons to a single atom @xcite .
experiments have pushed towards realizing such strong coupling , e.g. in cavity qed structures @xcite and optical waveguides @xcite , but the realization of gates remains challenging .
an alternative approach to strong light - matter interactions is to use atomic ensembles where a large number of atoms compensates for a small coupling efficiency of the individual atoms @xcite , but this approach typically does not enhance the nonlinear interactions required for quantum gates . in recent years
, experiments in e.g. tapered optical fibers @xcite and hollow core photonic - crystal fibers @xcite have realized an intermediate regime where the coupling of the individual atoms is sizeable , but still not sufficient to realize photonic gates based on single atoms .
it is thus an open question to which degree such moderate couplings enable processing of quantum information .
, @xmath0 ) and transmittances ( @xmath1 , @xmath2 ) of an ensemble of @xmath3type atoms without ( @xmath4 , @xmath1 ) and with ( @xmath0 , @xmath2 ) a stored photon for different frequencies ( two - photon detunings ) @xmath5 in units of the total linewidth @xmath6 .
the vertical dotted line marks the operation point .
the parameters are : number of atoms @xmath7 , @xmath8 , @xmath9 , and @xmath10 .
( b ) secret key rate @xmath11 per repeater station as a function of the number of atoms @xmath12 with fixed @xmath13 for dual - v atoms and a communication distance of 1000 km .
we compare the protocol of ref .
@xcite ( `` linear '' ) with a modified protocol where the entanglement swapping ( and also initial entanglement generation if it improves @xmath11 ) is performed with the proposed stationary light cp gate ( `` sl '' ) .
we consider two different source repetition rates : 100 mhz and 1 mhz .
we assume an attenuation length of 22 km in the fibers and an optical signal speed of @xmath14 km / s .
the ensemble storage and retrieval efficiency increases with @xmath12 and is set to the same value in the original protocol as for the modified one . the photodetector efficiency is assumed to be 90% .
the steps in the curves occur when the fidelity of the cp gate allows additional swap levels . ] in this letter , we propose a method to exploit the moderate coupling regime to realize quantum gates between photons .
we show that by using sufficiently many atoms it is possible to compensate for the limited coupling strength of the individual atoms and achieve ideal gate operation @xcite .
the proposed gate is based on _ stationary light _
@xcite where polaritons ( coupled light - matter excitations ) have very low group velocity due to counter - propagating classical drives .
reflections from the ends of an ensemble experiencing stationary light result in a cavity - like transmission spectrum with transmission resonances whenever the polaritons form standing waves inside the ensemble @xcite .
we show that the storage of a single photon completely changes the transmission spectra of the ensemble , see fig .
[ fig_r_and_t_plot](a ) , because the cavity structure created by the remaining atoms enhances the interaction with the stored excitation .
this can be used to mediate a gate between photons , which can be either deterministic or heralded where successful operation is conditioned on subsequent detection of two photons . as a particular application
we show that the gate can be directly used to enhance the communication rate of ensemble based quantum repeaters @xcite ( fig .
[ fig_r_and_t_plot](b ) ) .
_ overview .
_ we consider two different level schemes for the atoms in the ensemble : @xmath3type and dualv ( figs .
[ fig_setup](a ) and [ fig_setup](b ) , respectively ) .
the linear properties of these two schemes are described in detail in ref .
@xcite . in the @xmath3type scheme ,
the two classical drives have the same polarization , which results in a standing wave of the rabi frequency @xmath15 , where @xmath16 is the wave vector of the classical drive , assumed to be the same as the wave vector of the probe field @xmath17 ( single photon ) . for the @xmath3type scheme , we assume that the atoms are placed at positions @xmath18 with @xmath19 to achieve the lowest possible group velocity @xcite .
this scheme is easier to analyze , and all our analytical results are obtained for this scheme .
the dualv scheme @xcite is less sensitive to the exact positions of the atoms and hence easier to implement in practice . for the dualv scheme ,
we only verify numerically that the gate errors have the same scaling as for the @xmath3type scheme .
the coupling of the atoms to the waveguide is characterized by the parameter @xmath20 ( half of the optical depth per atom ) , where @xmath21 is the decay rate from state @xmath22 for the @xmath3type scheme into both right - moving and left - moving guided modes ( assumed to be equal ) , @xmath23 is the decay rate into all the other modes , and @xmath24 is the total decay rate . in the dual - rail encoding of photonic qubits shown in fig .
[ fig_setup](c ) , two identical atomic ensembles are required , where the upper one only functions as a memory . alternatively , the single - rail encoding can also be implemented with one atomic ensemble @xcite , but the dual - rail encoding allows heralded operation , which has better fidelity .
each ensemble is placed inside a sagnac interferometer ( fig .
[ fig_setup](d ) ) .
the operation of the cp gate is sequential .
first , photon @xmath25 is stored either in the upper ( @xmath26 ) or the lower ( @xmath27 ) ensemble using electromagnetically induced transparency ( eit ) @xcite
. then photon @xmath28 is scattered from the lower ensemble under conditions of stationary light ( @xmath29 ) or passes through a beam splitter with transmission coefficient @xmath30 ( @xmath31 ) .
the role of this beam splitter will be explained below .
the sagnac interferometer can be set up such that most of the incident power in each of its two input ports is reflected back through the same port , regardless of whether the ensemble is reflective or transmissive .
reflection or transmission of the ensemble instead controls the phase of the reflected field .
the scattering of photon @xmath28 can be arranged such that if there is no stored photon in the lower ensemble ( photon @xmath25 is in the state @xmath26 ) , the atomic ensemble is completely transmissive in the ideal case , and photon @xmath28 is reflected from the sagnac interferometer with no additional phase .
if there is a stored photon ( photon @xmath25 is in state @xmath27 ) , photon @xmath28 is reflected from the interferometer with a @xmath32 phase shift .
the latter case performs the desired controlled - phase gate operation @xmath33 , while the rest of the basis states are unchanged .
finally , photon @xmath25 is retrieved using eit .
_ storage and retrieval .
_ under eit storage and retrieval , the incident photon is resonant with the classical drive , and the classical drive is incident from one side and is assumed to be resonant with its transition for simplicity . after entering the sagnac interferometer , photon @xmath25 will be split into two halves by the 50:50 beam - splitter ( see fig .
[ fig_setup](d ) ) , which upon reaching the ensemble from the opposite sides will have opposite spatial phase factors @xmath34 and @xmath35 and interfere inside the ensemble resulting in a spatially modulated @xmath36 stored spin wave .
such storage procedure is necessary for the @xmath3-type scheme , since the part of the excitation that is stored on the nodes of the standing wave of the classical drive ( that is applied during scattering of photon @xmath28 ) does not change the scattering properties of the ensemble . before the eit storage
, all atoms are initialized in state @xmath37 , and after storage , the incident photon is mapped onto an atom being in state state @xmath38 .
to produce an optical non - linearity , we assume that state @xmath38 is subsequently transferred to state @xmath39 using a @xmath32-pulse . _ reflection and transmission .
_ we use the ( multi - mode ) transfer matrix formalism @xcite to model the scattering process . to illustrate the scattering behavior
, we assume that photon @xmath25 was stored in the center of the atomic ensemble at an anti - node of the classical drive . the reflectances and transmittances of an ensemble of @xmath3type atoms
are plotted in fig .
[ fig_r_and_t_plot](a ) as functions of two - photon detuning @xmath40 , where @xmath41 ( @xmath42 ) is the detuning of the probe field ( classical drive ) .
the reflectance @xmath4 ( @xmath0 ) and transmittance @xmath1 ( @xmath2 ) are for an ensemble without ( with ) a stored photon .
the ensemble is seen to have transmittance resonances with a large transmittance and a small reflectance , similar to the transmission of a cavity .
these resonances occur when the standing wave condition is fulfilled , i.e. @xmath43 , where @xmath44 is the bloch vector of the stationary light polaritons and @xmath45 is the length of the ensemble @xcite . when a photon is stored in the ensemble , an atom changes from state @xmath37 to @xmath39 . in state @xmath39 ,
the atom acts as a two - level atom that is resonant with the incident photon ( see fig .
[ fig_setup](a ) ) . since the effective interaction is enhanced by the cavity - like behavior of the ensemble
, this single two - level atom can have a strong effect of an incident photon , even though the atom in itself has a limited coupling .
we focus on the behavior at the resonance nearest @xmath46 ( vertical dotted line in fig . [ fig_r_and_t_plot](a ) ) . in the limit of large atom number @xmath12 and for @xmath47
, this resonance is at a two - photon detuning @xmath48 , for which we obtain @xcite @xmath49\\ \label{t_ensemble_n } & \begin{aligned } & t_0\approx 1-r_0 , \end{aligned } \displaybreak[0]\\ \label{r_imp_ensemble_n } &
\begin{aligned } & r_1(\tilde{z})\approx 1 -\frac{4\pi^2\gamma'\delta_\text{c}^2}{\gamma_\text{1d}^3 n^2 } + \frac{32\pi^4 \delta_\text{c}^2 \gamma ' |\omega_0|^2}{\gamma_\text{1d}^5 n^4}\\ & -\frac{4i\pi^2\delta_\text{c}}{\gamma_\text{1d } n } \left(\tilde{z}-\frac{1}{2}\right )
-\frac{4\pi^4\delta_\text{c}^2(2\gamma_\text{1d}+\gamma')}{\gamma_\text{1d}^3 n^2 } \left(\tilde{z}-\frac{1}{2}\right)^2 , \end{aligned } \displaybreak[0]\\ \label{t_imp_ensemble_n } & \begin{aligned } & t_1(\tilde{z})\approx \frac{4\pi^2\delta_\text{c}^2\gamma'}{\gamma_\text{1d}^3 n^2 } -\frac{32\pi^4 \delta_\text{c}^2 \gamma ' |\omega_0|^2}{\gamma_\text{1d}^5 n^4}\\ & + \frac{8\pi^4\delta_\text{c}^2\gamma'}{\gamma_\text{1d}^3 n^3 } \left(\tilde{z}-\frac{1}{2}\right ) + \frac{4\pi^4\delta_\text{c}^2\gamma'}{\gamma_\text{1d}^3 n^2 } \left(\tilde{z}-\frac{1}{2}\right)^2 . \end{aligned}\end{aligned}\ ] ] here , @xmath50 and @xmath51 , were obtained by solving the discrete problem , where a photon is stored in a single discrete atom and then taking the continuum limit such that the index of the atom is replaced by its position inside the ensemble @xmath52 .
the @xmath53 dependent terms are only relevant for determining the gate time and will be ignored for now . by aligning the interferometer , the reflection coefficients of the combined interferometer - ensemble system
are given by @xmath54 and @xmath55 @xcite .
if we take @xmath56 and a detuning @xmath57 , we have @xmath58 , @xmath59 , and @xmath60 .
hence , regardless of the value of @xmath61 we can achieve an ideal cp gate @xmath62 , @xmath63 with sufficiently many atoms .
_ fidelity .
_ to quantify the errors of the gate , we calculate the choi - jamiolkowski ( cj ) fidelity @xcite .
the eit storage is described using the storage @xmath64 and retrieval @xmath65 kernels derived in ref .
@xcite ( suitably modified to take into account storage from both directions @xcite ) .
when photon @xmath25 is stored and retrieved without a scattering taking place , its wave function is @xmath66 , where @xmath67 is the wave function of the input photon @xmath25 . the efficiency of the storage and retrieval is then @xmath68 .
if photon @xmath28 was reflected from the interferometer , while photon @xmath25 was stored in the ensemble ( computational basis state @xmath69 ) , the wave function of the retrieved photon @xmath25 is instead @xmath70 . neglecting bandwidth effects of photon @xmath28
, we obtain the cj fidelity @xmath71 where @xmath72 .
if the gate is conditioned on the presence of two photons after the gate operation @xcite , we find that the success probability is @xmath73 with @xmath74 , and the conditional cj fidelity is @xmath75 . to optimize the performance of the gate , we set @xmath76 and optimize @xmath42 and the width of the stored spin wave @xmath77 such that @xmath78 is maximal
afterwards , for fixed optimal @xmath42 and @xmath79 we explore the effect of @xmath80 .
as shown below , @xmath81 can be substantially improved by choosing a particular @xmath80 at the cost of increasing @xmath82 by a constant factor . whether this is a desirable trade off , depends on the particular application . in fig .
[ fig_cphase_fidelity_natoms](a ) we plot the numerically calculated @xmath83 and @xmath81 for the @xmath3-type scheme , where photon @xmath25 was chosen to have a gaussian temporal profile , and photon @xmath28 is centered on @xmath84 and assumed to be narrow in frequency compared to the resonance width . as seen in the figure ,
both @xmath78 and @xmath81 approach their ideal value of unity for large @xmath12 , but @xmath81 approaches it much faster .
for large @xmath12 , we can find analytical expressions for the curves in fig .
[ fig_cphase_fidelity_natoms](a )
. the stored spin wave will be approximately gaussian so that it has the form @xmath85
. consequently , @xmath86 @xcite . neglecting distortions of photon @xmath25 under storage and retrieval , but still accounting for errors due to the spatial extent of the stored excitation , we approximate @xmath87 and @xmath88 . here , @xmath89 is a symmetrized version of @xmath90 , which accounts for storage and scattering from both sides of the ensemble due to the sagnac interferometer . for fixed @xmath21 and large @xmath12 , after choosing @xmath91 , @xmath92 , and @xmath93 , such that @xmath78 is maximal , we get @xmath94 if @xmath95 , @xmath81 is maximal , and we get @xmath96 these expressions confirm that the gate fidelity improves with @xmath12 and that the conditional fidelities have better scaling .
type scheme . for @xmath76 , @xmath83 and @xmath81
are shown by dotted green and dash - dotted black curves respectively .
for @xmath80 chosen such that @xmath81 is maximal , @xmath83 and @xmath81 are shown by solid blue and dashed red lines respectively .
( b ) the same as in ( a ) , but using the dualv instead and interatomic spacing @xmath97 .
the common parameters are @xmath8 , and @xmath98 . under eit ( storage and retrieval ) , @xmath99 . under stationary light ( scattering ) , @xmath15 and @xmath100 for @xmath3type and dualv respectively . ] in fig .
[ fig_cphase_fidelity_natoms](b ) we plot the numerically calculated fidelities @xmath78 and @xmath81 for the dualv scheme . in the simulation ,
the distance between the atoms was set to be incommensurate with the wavelength of the classical drive @xmath97 .
the results are , however , almost independent of @xmath101 , and the gate can function even with completely random placement of the atoms @xcite . the dualv scheme is seen to have a very similar behavior to the @xmath3type scheme . _ gate time . _
the gate time will be set both by the storage and retrieval time of photon @xmath25 and the necessity of the scattered photon @xmath28 to be narrow in frequency .
the eit storage and retrieval time is limited by @xmath102 @xcite and decreases with larger @xmath12 .
we therefore focus on the scattering of photon @xmath28 .
due to non - zero bandwidth of photon @xmath28 , the reflection coefficient @xmath103 ( at @xmath84 ) in eq . should be replaced by @xmath104 @xcite , where @xmath105 is the frequency distribution of photon @xmath28 .
since @xmath51 and @xmath50 vary much slower than @xmath106 and @xmath107 around @xmath84 ( vertical dotted line in fig .
[ fig_r_and_t_plot](a ) ) , we ignore a similar modification to @xmath108 in eq . .
we first expand the reflection coefficient @xmath106 around @xmath109 .
we thereby obtain @xmath110 with @xmath111 being the resonance width . defining the spectral width of photon @xmath28 by @xmath112 ,
we get @xmath113 .
we now include the @xmath53 dependent terms in eqs . and . for optimal @xmath114
, we get @xmath115 . requiring the error from finite bandwidth to be proportional to the error in eq . , we find that the required time is @xmath116 .
hence , the gate time decreases with @xmath12 . _
repeater secret key rate .
_ as a direct application of the proposed cp gate , we consider quantum repeaters based on atomic ensembles @xcite .
we modify one of the fastest known repeater protocols for atomic ensembles @xcite by implementing the proposed cp gate instead of linear optics for entanglement swapping using the setup in fig .
[ fig_setup](c ) . the secret key rate per repeater station
is calculated as described in ref .
@xcite and compared to the results of the original protocol ( see fig . [ fig_r_and_t_plot](b ) ) .
this analysis is similar to the cp gate in ref .
@xcite with the difference that we also consider the possibility of generating the initial entanglement using the cp gate . for a fair comparison , we consider equal storage and retrieval efficiencies for both protocols . as seen in fig .
[ fig_r_and_t_plot](b ) for @xmath13 , the proposed gate allows improving the rate of quantum repeaters if @xmath117 .
_ conclusion .
_ we have shown , how stationary light can be used to create a cp gate between photons .
most importantly , the gate can exploit a large number of atoms @xmath12 to compensate for a limited single - atom coupling efficiency to light .
in particular , the gate can have a rapid convergence as @xmath118 towards unit fidelity , if it is operated in a heralded fashion . the gate is ideally suited to be employed in setups currently under development @xcite , where there is a moderate coupling efficiency to light @xmath119 .
the gate can be directly employed to improve the communication rate of quantum repeaters based on atomic ensembles and may in general serve as a tool for photonics based quantum information processing .
the research leading to these results was funded by the european union seventh framework programme through siqs ( grant no .
600645 ) and erc grant qios ( grant no .
306576 ) .
j.b . acknowledges funding from the carlsberg foundation .
35ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1038/nature07127 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.127902 [ * * , ( ) ] link:\doibase 10.1038/nphys708 [ * * , ( ) ] http://dx.doi.org/10.1038/nphoton.2013.48 [ * * , ( ) ] link:\doibase 10.1126/science.1238169 [ * * , ( ) ] http://dx.doi.org/10.1038/nature13188 [ * * , ( ) ] http://dx.doi.org/10.1038/nature18592 [ * * , ( ) ] http://dx.doi.org/10.1038/nphoton.2014.253 [ * * , ( ) ] link:\doibase 10.1038/nature06230 [ * * , ( ) ] link:\doibase 10.1038/nphoton.2009.287 [ * * , ( ) ] http://dx.doi.org/10.1038/ncomms4808 [ * * ( ) ] http://dx.doi.org/10.1038/nnano.2015.159 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.82.1041 [ * * , ( ) ] link:\doibase 10.1103/physreva.70.063403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.203603 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.033603 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.263603 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.180503 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.203902 [ * * , ( ) ] link:\doibase 10.1103/physreva.94.043833 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.062301 [ * * , ( ) ] link:\doibase 10.1103/physreva.85.013822 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.89.143602 [ * * , ( ) ] link:\doibase 10.1038/nature02176 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.063823 [ * * , ( ) ] link:\doibase 10.1038/35106500 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.83.33 [ * * , ( ) ] @noop @noop link:\doibase 10.1103/physreva.93.040303 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.033805 [ * * , ( ) ] link:\doibase 10.1103/physreva.52.1394 [ * * , ( ) ] link:\doibase 10.1103/physreva.72.032333 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.114.110502 [ * * , ( ) ] link:\doibase 10.1103/physreva.92.012307 [ * * , ( ) ] * supplemental material for + `` controlled - phase gate for photons based on stationary light '' * @subsection @subsubsection |
the purpose of this paper is to begin to advance the work of deligne - rapoport , katz - mazur , edixhoven , and others regarding the stable model of @xmath6 . in the simplest case of @xmath7 ( @xmath8 ) , the minimal resolution over @xmath9
is always semi - stable and was explicitly described in ( * ? ? ?
* vi.6.16 ) .
edixhoven later worked out the minimal resolution of the katz - mazur model for @xmath6 over @xmath9 ( special case of @xcite ) , but this model is never semi - stable when @xmath10 .
however , in the @xmath11 case , edixhoven did go on to work out a semi - stable model for @xmath12 over a finite extension of @xmath13 ( @xcite ) .
unfortunately , as edixhoven states in the introduction , his methods do not generalize sufficiently to calculate the stable model when @xmath14 because of `` wild ramification . ''
recently , more progress toward understanding the stable model of @xmath6 has been made by coleman using a moduli - theoretic approach . in @xcite coleman showed that for @xmath5 the ordinary region of @xmath6 has exactly @xmath15 connected components ( over @xmath16 ) , whose reductions can be described using igusa curves
. he also was able to give a moduli - theoretic interpretation of the horizontal components of edixhoven s model for @xmath12 .
in particular , the points on these components correspond to pairs @xmath17 such that @xmath18 $ ] is in the language of @xcite too - supersingular , i.e. that @xmath18 $ ] has no canonical subgroup . in the first half of this paper we begin with the model for @xmath0
constructed in @xcite , and apply techniques of rigid analysis to construct a semi - stable covering .
from this point on , all such statements will always mean over @xmath19 , the ring of integers in @xmath16 .
paraphrasing @xcite , this means that we will find a finite set of disjoint affinoids @xmath20 with good reduction , such that the complement in @xmath0 of @xmath21 is simply the disjoint union of annuli .
by @xcite this is equivalent to determining a semi - stable model for the curve .
the goal of the remaining sections , then , is to provide a moduli - theoretic interpretation of the components that leads to a conjecture regarding the stable model of @xmath4 . to do this we first map the affinoids in the semi - stable covering down to @xmath3 via an appropriate moduli - theoretic map ( using formulas from @xcite ) .
then we show by explicit calculations that the images contain the @xmath2-invariants of curves with a certain type of complex multiplication . along with similar data for @xmath22 and @xmath23 ,
this leads to two conjectures which could be considered the main results of the paper .
conjecture [ conjecture : mainconjecture1 ] concerns the distribution of cm curves inside a given @xmath24 disk of @xmath3 , and conjecture [ conjecture : mainconjecture2 ] generalizes the stable model description of @xmath0 to a description of @xmath4 .
these conjectures will be at least partially proven in an upcoming joint work with robert coleman .
in this section we determine the stable model for @xmath0 ( genus @xmath25 ) by essentially constructing in the language of @xcite a semi - stable covering for the curve .
the initial model will be the one determined in @xcite , namely the following system of equations .
@xmath26 @xmath27 as modular functions , the @xmath28-expansions at infinity of @xmath29 , @xmath30 , and @xmath31 can be expressed in terms of the dedekind eta - function as follows .
@xmath32 in @xcite it is shown that @xmath29 and @xmath30 are functions on @xmath0 which are fixed by the atkin - lehner involution , @xmath33 .
therefore equation actually describes the genus two quotient curve , @xmath34 then @xmath31 is a third function on @xmath0 ( actually a pullback of a function on @xmath35 ) which generates , by equation , the degree @xmath36 extension from @xmath37 up to @xmath0 . in later sections we will want to interpret the various components of our model in moduli - theoretic terms in the hopes of conjecturing what happens in general ( for @xmath4 , @xmath38 ) . for this section , however , we will stick to bare - bones rigid analysis .
first we show that @xmath37 has good reduction by finding an explicit good reduction model .
then we show that the @xmath39 ramification points in the degree two extension lie in two equidistant sets of @xmath40 , all within a wide - open annulus .
the result of this information is that the stable model of @xmath0 has @xmath40 components , four of genus @xmath36 and one of genus @xmath41 . in particular , two components with the same ( stable ) reduction as @xmath37
will be switched under the atkin - lehner involution .
the other two genus @xmath36 affinoids are fixed by atkin - lehner and have trivial quotients . at the end of @xcite
there is a proof that @xmath37 has good reduction which is based on a hyperelliptic model for the genus @xmath36 curve . in order to understand the extension up to @xmath0
, however , this hyperelliptic model is not optimal . for this reason we will now offer a proof using a different model , one which will be more appropriate for the full analysis of @xmath0 .
essentially , we will show that with respect to the parameter @xmath30 the affinoid @xmath42 is a genus @xmath36 affinoid ( in the sense of @xcite ) with good reduction .
then in the section that follows we will see that the @xmath39 ramification points in the degree two extension up to @xmath0 are actually separated from this affinoid by a wide open annulus .
this is a key point in understanding the stable model of @xmath0 .
we begin by choosing any @xmath43 satisfying @xmath44 and making the change of variables @xmath45 .
plugging into @xmath46 we then obtain a polynomial @xmath47 whose coefficients are polynomials in @xmath48 and are given below in table [ x0-y - model ] .
the entry in the @xmath49 row and @xmath50 column is the coefficient of @xmath51 .
if true , the preceeding conjecture would provide at least the beginning of a moduli - theoretic description of the components in the stable model for @xmath0 .
however , at the end of the paper we will seek to propose a general conjecture for the stable model of @xmath4 .
therefore , we will want to know how conjecture [ conjecture : cmp5 ] generalizes to other primes .
for this reason we have also checked a number of examples for @xmath22 and @xmath23 , and have found that the data consistently supports conjectures [ conjecture : cmp7 ] and [ conjecture : cmp13 ] .
all three conjectures will be tied together with a general conjecture in next and final section of the paper . for a conjecture regarding the components in the stable model of @xmath4 ,
it is at least clear what `` types '' of components are suggested by the @xmath0 data .
we will naturally conjecture that in general , for each supersingular elliptic curve ( at least defined over @xmath65 ) , we have components that look like @xmath66 , @xmath67 , the four genus @xmath36 components , and the trivial component which intersects the others .
in other words , we will conjecture that there are components which have the same moduli - theoretic properties that the components in our semi - stable covering of @xmath0 have been proven or conjectured to have . of the various ways in which to improve such a conjecture , one way would certainly be to conjecture the number of such components and give an explanation of the number .
this we will do with the help of lemma [ lemma : ap ] .
conjecture [ conjecture : mainconjecture1 ] uses the lemma to generalize the conjectures of the previous section regarding the placement of cm @xmath2-invariants .
conjecture [ conjecture : mainconjecture2 ] then takes this into account in describing the stable model of @xmath4 .
finally , we conclude the paper with a guess which goes one step farther and predicts the genera of the components .
the action is simply conjugation , after identifying @xmath82 with @xmath83 $ ] of course .
now , for part ( 1 ) let @xmath84 and @xmath85 .
then by definition , @xmath86 solving and setting @xmath87 and @xmath88 coefficients equal , this leads to the following system of linear equations in @xmath89 and @xmath90 .
@xmath91 when @xmath92 , i.e. when @xmath93 , the system is trivially satisfied for all @xmath94 . when @xmath95 , however , the determinant is @xmath96 and therefore @xmath97 is the only solution .
similarly , it is a straightforward exercise in congruences to verify half of part ( 2 ) , namely that @xmath98 implies @xmath99 .
this means that as soon as @xmath100 has one solution , it must have at least @xmath101 solutions by part ( 1 ) . but
@xmath102 always has a solution since the set @xmath103 has @xmath104 elements and therefore must contain a quadratic residue .
therefore we are done since there are @xmath105 choices for @xmath106 , @xmath101 solutions to @xmath102 , and only @xmath107 nonzero elements of @xmath108 to begin with .
the reason that this lemma is relevant is the following .
let @xmath109 be a supersingular curve so that @xmath110 .
the isomorphism is non - canonical , but it suffices to choose an isomorphism once and for all .
now suppose that @xmath111 is a cm curve with @xmath112 and such that @xmath113 .
identify @xmath114 with a subring of @xmath16 via the canonical embedding coming from the action of @xmath114 on holomorphic differentials .
then for every isomorphism @xmath115 we obtain an embedding @xmath116 .
furthermore , part ( 2 ) of lemma [ lemma : ap ] implies that for any uniformizer @xmath117 , the conjugacy class of @xmath118 inside @xmath110 is independent of @xmath119 .
so without taking @xmath119 into account , there are only @xmath101 options for the image of @xmath118 inside @xmath120 .
to illustrate this point , consider the following example .
[ example : p7 ] let @xmath53 be a curve with @xmath121 $ ] , and let @xmath109 be the unique supersingular elliptic curve in characteristic @xmath122
. from @xcite we may take @xmath123 to be a maximal order in the quaternion algebra @xmath124 $ ] with @xmath125 , @xmath126 , and @xmath127 .
let @xmath31 be the uniformizer @xmath128 and suppose that for a particular isomorphism @xmath115 we have @xmath129 . since @xmath130 we must immediately have @xmath131 and @xmath132 ( mod @xmath122 ) . therefore we have @xmath133 , which means that for this example the image of @xmath118 inside @xmath134 must lie in the conjugacy class ( of @xmath25 elements ) containing @xmath135 however , the precise image of @xmath118 is _ not _ independent of @xmath119 .
so we can not yet use this image to put an equivalence relation on the set of curves with a given endomorphism ring . to obtain an invariant which is independent of @xmath119 , we first note that @xmath136 acts transitively on the set of isomorphisms from @xmath137 to @xmath123 via conjugation , with @xmath138 as the kernel in all cases . so letting @xmath139 we see that the conjugacy class defined by @xmath31 inside of @xmath120 can be broken down into @xmath140 subsets , so that the _ subset _ containing the image of @xmath118 is now independent of @xmath119 .
again , we illustrate this point by revisiting the previous example .
[ example : p7b ] the unique supersingular curve @xmath109 in characteristic @xmath122 has @xmath141 . in terms of the above identification , the automorphism group is generated by the ( invertible ) element @xmath142 .
so for example , if @xmath143 for some @xmath119 , then @xmath144 .
this means that if we want to associate to @xmath31 an element of @xmath134 which is _ independent _ of @xmath119 , we can _ not _ distinguish between @xmath145 .
similarly , conjugation by @xmath142 breaks down the entire set of eight elements into four subsets of order @xmath36 , namely @xmath146 so independent of @xmath119 , it makes sense to say that there are @xmath147 options for the image of @xmath118 in @xmath134 .
the result of the preceeding argument is that we have shown how to put an equivalence relation with @xmath140 ( possibly empty ) classes on the set of cm curves @xmath148 reducing to a fixed supersingular curve @xmath109 and with a fixed endomorphism ring such that @xmath112 . indeed , with the same argument we may even generalize the relation by requiring only that @xmath149 be fixed . since there are only two ramified quadratic extensions of @xmath150 ,
this makes a total of @xmath151 classes into which all such cm curves must fall .
what we have _ not _ shown , however , is that this equivalence relation is reflected somehow in the geometric placement of the @xmath2-invariants of these cm curves inside of @xmath3 .
so we have provided evidence for conjecture [ conjecture : mainconjecture1 ] , which would explain conjectures [ conjecture : cmp5 ] , [ conjecture : cmp7 ] , and [ conjecture : cmp13 ] with a general theory .
however , we do not yet have a proof of the result at this time .
[ conjecture : mainconjecture1 ] let @xmath109 be a supersingular elliptic curve defined over @xmath65 and let @xmath152 .
+ ( 1 ) the @xmath2-invariants of all cm curves @xmath111 such that @xmath153 and @xmath154 lie in @xmath151 residue disks inside the corresponding atkin - lehner + circle , @xmath140 for each ramified quadratic extension of @xmath150 .
+ + ( 2 ) two such curves , @xmath66 and @xmath67 , lie in the same residue disk iff @xmath155 and for any uniformizer @xmath156 of the common image we have + @xmath157 for some @xmath158 , @xmath159 . [ conjecture : mainconjecture2 ] there is a semi - stable covering of @xmath4 defined over @xmath19 such that for each supersingular elliptic curve @xmath109 defined over @xmath65 there is a connected component of the supersingular locus containing ( only ) the following .
+ ( 1 ) one component lying via @xmath160 over the atkin - lehner circle ( of @xmath109 ) , + ( 2 ) two components , @xmath66 and @xmath67 , which are @xmath161 and @xmath162 of the + too - supersingular disk ( of @xmath109 ) , and + ( 3 ) @xmath151 components lying via @xmath160 over the cm disks of + conjecture [ conjecture : mainconjecture1 ] .
+ the intersections of these components with each other and with the @xmath163 ordinary components of @xmath4 are as pictured in figure [ figure : stablep3graph ] .
it is important to acknowledge that conjecture [ conjecture : mainconjecture2 ] is based on just one very well understood example .
ironically , however , if we make a more precise conjecture it is easy to obtain far more corroborating data . in particular , from looking more closely at the @xmath0 example one might go so far as to guess the following .
the component lying over al is trivial , @xmath66 and @xmath67 are analytically isomorphic copies of edixhoven s horizontal component , and each component lying over a cm disk has genus @xmath164 ( and is hyperelliptic ) .
these statements also hold for @xmath0 , and with the guess it becomes possible to generate complete graphs of conjectural semi - stable models for @xmath4 .
the genera of each @xmath66 and @xmath67 come directly from @xcite .
the genera of the ordinary components come from @xcite and the fact ( from @xcite ) that the four nontrivial ordinary components of @xmath4 are isomorphic copies of the two nontrivial ordinary components of @xmath12 .
using this approach we have generated graphs for @xmath165 , @xmath166 , and @xmath167 ( figures [ figure : graph7 ] , [ figure : graph13 ] , and [ figure : graph17 ] ) . in each of these cases , and in fact even in the _ general _ case , it is easy to show that the total genus of the curve is at least correct .
while it is perhaps too early to make the guess an official conjecture , the corroborating data of this genus calculation seems to be very promising .
999 k. buzzard , _ analytic continuation of overconvergent eigenforms _ , j. amer .
* 16 * ( 2003 ) , no .
. s. bosch , u. gntzer , r. remmert , _ non - archimedean analysis _ , grundlehren der mathematischen wissenschaften * 261 * , springer - verlag ( 1984 ) .
r. coleman , _ on the components of @xmath6 _ , j. number theory , to appear .
r. coleman , _
stable maps of curves _ , doc . math .
, extra volume : `` kazuya kato s fiftieth birthday '' ( 2003 ) , 217225 . b. edixhoven , _ minimal resolution and stable reduction of @xmath171 _ , ann .
fourier ( grenoble ) * 40 * ( 1990 ) , no .
1 , 3167 .
p. deligne , m. rapoport , _ schemas de modules de courbes elliptiques _ , lecture notes in math .
* 349 * ( 1973 ) , 143316 .
n. katz , b. mazur , _ arithmetic moduli of elliptic curves _ , annals of mathematics studies * 108 * , princeton university press ( 1985 ) .
s. lang , _ elliptic functions _ , graduate texts in mathematics * 112 * , springer - verlag ( 1987 ) .
k. mcmurdy , _ explicit parameterizations of ordinary and supersingular regions of @xmath6 _ , progress in mathematics * 224 * ( 2004 ) , 165179 .
a. pizer , _ an algorithm for computing modular forms on @xmath172 _ , journal of algebra * 64 * ( 1980 ) , 340390 .
a. robert , _ a course in @xmath173-adic analysis _ ,
graduate texts in mathematics * 198 * , springer - verlag ( 2000 ) . | in this paper we determine the components in the stable model of @xmath0 over @xmath1 by constructing in the language of @xcite an explicit semi - stable covering .
we then offer empirical data regarding the placement of certain cm @xmath2-invariants in the supersingular disk of @xmath3 over @xmath1 which suggests a moduli - theoretic interpretation for the components of the stable model .
the paper then concludes with a conjecture regarding the stable model of @xmath4 for @xmath5 , which is as yet unknown .
[ subsection ] [ subsection ] [ subsection ] [ subsection ] [ subsection ] this is an unofficial version of the paper .
the definitive version has been published in the lms journal of computation and mathematics , volume 7 , pp .
2136 .
we are very grateful to the lms for their careful refereeing , editing , and publication of the official version . |
the system under consideration is a @xmath86 diameter micropillar .
the sample has been fabricated via molecular beam epitaxy ( mbe ) and contains a @xmath87-thick cavity surrounded by two distributed bragg reflectors ( dbrs ) .
the dbr structure consists of 18 ( 5 ) bottom ( top ) alas / gaas mirror pairs @xcite . a modulation doped low density in(ga)as qd layer ( @xmath88x@xmath89 @xmath90 )
has been grown in the middle of the cavity .
the fundamental mode of the cavity appears at @xmath91 mev cavity and has a full width half maximum of @xmath92 corresponding to a q - factor of @xmath21 .
the faraday geometry magnetic field was introduced by the use of a permanent ring magnet , which creates a homogeneous magnetic field of @xmath93 perpendicular to the plane of the qd .
single photons are input via a @xmath94 objective lens and collected into six @xmath95 core single mode fibers .
detection events were recorded by a home built six channel field programmable gate array ( fpga ) with @xmath96 timing resolution .
this allows us to extract the qd behaviour over a range of short time windows . | solid - state quantum emitters have long been recognised as the ideal platform to realize integrated quantum photonic technologies .
we use a self - assembled negatively charged qd in a low q - factor photonic micropillar to demonstrate for the first time a key figure of merit for deterministic switching and spin - photon entanglement : a shift in phase of an input single photon of @xmath0 with values of up to @xmath1 ( @xmath2 ) demonstrated .
this @xmath3 ( @xmath4 ) measured value represents an important threshold : above this value input photons interact with the emitter deterministically .
a deterministic photon - emitter interaction is the only viable scalable means to achieve several vital functionalities not possible in linear optics such as quantum switches and entanglement gates .
our experimentally determined value is limited by mode mismatch between the input laser and the cavity , qd spectral fluctuations and spin relaxation .
we determine that up to @xmath5 of the collected photons have interacted with the qd and undergone a phase shift of @xmath6 . quantum dots ( qds )
have long been proposed as key components in quantum optical circuits . in principle
they are capable of producing single photons on demand @xcite , and acting as a mediator " qubit that allows indirect interactions between two photonic qubits .
dramatic progress has been made very recently with qd sources , with record - breaking brightness and indistinguishable single photons @xcite .
however qds can be exploited not just as sources , but as quantum switches , memories and entanglers @xcite .
one of the most useful potential functionalities is as a spin - photon interface . in the absence of a magnetic field
a qd charged with an excess electron has two degenerate spin - dependent transitions that couple to opposite circularly polarized light states . by maximising the interaction of the qd with light
, one may use the qd transition to switch " the phase , @xmath7 , of a single photon by up to @xmath6 .
this interaction is maximum when directly on resonance with the transition , and reduces with spectral detuning .
however , it is by combining this with the selection rules for each spin transition that one may achieve more powerful operations .
one may selectively phase shift one component of a linearly polarized photon ( e.g. the right circular component @xmath8 only ) by initialising the spin ( e.g. @xmath9 ) .
interference between the interacting and non - interacting components induces a giant " faraday rotation of the linear polarisation up to @xmath10 ( e.g. from vertical to horizontal ) @xcite .
this switches the photon from one polarisation to an orthogonal one , controlled by the spin .
spin - photon entanglement naturally ensues by preparing the spin into an equal superposition state : @xmath11
. however , for this functionality one requires deterministic " interactions , i.e. the ability of the quantum emitter to switch the photon between orthogonal polarization states .
this represents the ultimate limit in strong non - linear interactions and requires significant control of the photonic environment around the qd .
several approaches to deterministic interactions have been taken , most notably in the achievement of strong coupling " via high q - factor , low modal volume cavities @xcite as inspired by atom - cavity physics @xcite .
in such qd systems a @xmath12 phase shift was observed @xcite .
recent demonstrations have measured a @xmath6 phase switch on the qd spin , upon the successful detection of a rotated photon @xcite . however , in that work as in others , the high quality - factor ( q - factor ) design of the cavity means out of plane scattering limits the input and output coupling efficiency .
this limits the phase shift imparted on reflected photons and thus the probability of observing a rotation , making subsequent spin - photon entanglement protocols inefficient .
we now demonstrate the next step to a fully deterministic spin - photon interface , showing a phase switch imparted on the photon instead of the spin .
this is made possible by taking a different approach , where we use a very low q - factor micropillar cavity with well - defined input and output modes .
key here is that the coupling of the input and output modes with the qd should be higher than the leaky " modes .
this is parametrised by the @xmath13-factor " , related to the ratio of these two ( see supplementary ) .
it has been theoretically established that very weak intensity narrow bandwidth photons input and collected from such a mode will undergo deterministic @xcite interaction with the qd if @xmath14 @xcite , in contrast for a @xmath15 not all photons interact with the qd thus a maximum phase of only @xmath16 can maximally be detected ( see fig . [ fig:1n](c ) ) .
one can show that under these conditions , the scattered light always undergoes a phase shift of @xmath6 in the absence of pure dephasing effects , the threshold for this behaviour is @xmath17 with a discontinuous jump from @xmath16 to @xmath6 for @xmath14 . interference between
the incoming and scattered light leads to a deterministic scattering of the light backwards , with the intensity ( @xmath18 ) of backscattered light scaling with @xmath13-factor as @xmath19 @xcite .
our previous work @xcite has shown phase shifts @xmath20 from a micropillar system when a single frequency laser was tuned through the resonant frequency of the qd .
in that device ( @xmath21 ) , the @xmath13-factor was estimated to be @xmath22 @xcite .
thus we expect to be in the regime where a @xmath6 phase shift should be observed for reflected photons .
the low phase shift actually measured in ref .
@xcite results from two experimental limitations : the limit in precise mode - matching of the input laser to the micropillar mode resulting in a fraction of laser light that reflects without entering the cavity , and the presence of spectral jitter " which leads to a change in frequency of the qd transition over a timescale shorter than the laser scan time .
this detunes the qd relative to the laser ( @xmath23 ) and reduces the observed phase shift . in this work
we demonstrate that by measuring the phase shift faster than the spectral jitter time , one may obtain a measure of the phase shift which is limited only by the mode - matching limitations . by using a specially designed two - channel heralding technique , we explore time windows when the qd remains close to resonance with the laser , and measure the phase shift of light reflected from the qd - micropillar system .
we measure all the light reflected from the pillar , without using polarization filtering .
we consistently evaluate phase shift values @xmath3 when the qd is close to resonance ( @xmath24 ) where @xmath25 is the linewidth of the transition .
we also demonstrate that in the rare cases where we remain precisely on resonance ( @xmath26 , we measure phase shifts of up to @xmath27 . we therefore conclude that we reach the @xmath3 threshold , confirming that @xmath14 , and that only mode - matching limits our measured phase shift .
beamsplitter ( bs ) splits the reflected signal into two channels .
one which sends photons to detectors @xmath28 to perform the phase measurement defined in the main text , and a second channel which is split by a polarising beamsplitter ( pbs ) into co - poarised and cross - polarised components which are measured by detectors @xmath29 and @xmath30 respectively.(*b * ) energy level diagram of a negatively charged qd in a faraday magnetic field .
measurements are performed on light that is resonantly scattered from the spin down transition .
( * c * ) with the black circles the dependence of the maximum possible phase shift on the @xmath13-factor with @xmath31 in units of @xmath6.,scaledwidth=48.0% ] the setup used is shown in fig .
[ fig:1n ] , and is similar to that in ref .
@xcite .
the sample is placed in a @xmath32 magnetic field along the optical axis ( faraday geometry ) inducing an energy splitting of @xmath33 between the @xmath34 and @xmath9 ground state ( fig 1b ) .
vertically linearly polarized coherent light ( @xmath35 ) from a single frequency laser of linewidth @xmath36 ( over @xmath37 ) is focused on the micropillar with an @xmath38 objective .
it is then spectrally tuned to maximise the cross - polarised count - rate , upon reflection from the spin @xmath9 transition , which interacts with @xmath8 ( @xmath39 ) light .
the reflected signal is split into two with a non - polarizing beamsplitter ( bs1 ) : one arm is sent to the heralding " channel , apd - s 5 ( vertical-@xmath40 ) and 6 ( horizontal-@xmath41 ) ; the other arm is sent to the phase measurement arm , where four components of the field are measured , apd1 ( @xmath40 ) , apd-2 ( @xmath41 ) , apd3 ( @xmath42 ) and apd4 ( @xmath43 ) . to calculate the phase shift , we monitor the swing " in counts from @xmath40 to @xmath41 .
the light emerging from the qd - micropillar will both have undergone a faraday rotation ( @xmath44 ) and show some ellipticity .
this is because for @xmath45 , there are qd scattering losses affecting only the interacting , @xmath39 light , while @xmath46 of the non - interacting @xmath47 light is reflected back into the mode ( assuming a single sided cavity ) . to normalise to the elipticity we measure the phase shift using @xmath48
this is an estimate of the true phase accounting for ellipticity caused by losses ( see supplementary ) .
our previous measurements revealed that sub - second timescale spectral jitter of the qd transition broadened the fourier transform limited @xmath49 transition to the measured linewidth ( @xmath50 ) , and is a commonly reported problem that is thought to occur due to spectral wander ( nuclear spin and charge fluctuations ) @xcite . to overcome this
, we measure the phase shift in @xmath51 time intervals , smaller than the expected jitter and electron spin relaxation timescales , but sufficiently long to collect reliable statistics .
we then monitor the cross - polarised detector ( apd-6 , @xmath41 ) , where we measure an average count - rate of @xmath52 .
cross - polarized counts are only observed when the laser is on resonance with the qd and undergoes a faraday rotation with a non - zero projection onto h. in non - cavity experiments this is the channel where cross - polarised resonant scattering is detected @xcite . since the qd is seldom found close to resonance with the laser ,
the cross - polarised time - averaged count - rate is low .
fig .
[ fig:1](a ) shows a histogram of the distribution of the count - rate in @xmath51 time bins for the cross - polarized channel ( apd-6 ) . despite the vast majority ( @xmath53 ) of time bins containing low count - rates
, the histogram shows that high count - rates are found for some @xmath51 time windows , with a second peak in the distribution at @xmath54 , and outliers that show up to @xmath55 count - rates .
these occur when the qd and laser are close to resonant , as one would expect that the higher the count - rate , the closer the qd is to exact resonance .
we expect that when the cross - polarized count - rates in the heralding arm are high , the phase shift of input photons will also be significant .
fig .
[ fig:1](a ) also shows the phase shift measured for each value of the herald apd-6 .
one sees that as the cross - polarized herald count - rates increase , the phase shift does also , with a value saturating at @xmath56 for apd-6@xmath57 .
this represents @xmath51 average response when the input laser is within the linewidth of the transform limited qd transition .
surprisingly , high cross - polarized herald counts do not correlate with a steadily increasing measured phase .
this is because the number of cross - polarized heralds arriving at apd-6 is affected by power fluctuations of the input laser which is normalised out when we measure the phase .
one can gain more information by including measurements of the co ( apd-5 ) channel .
efficient coherent scattering from the qd will result in light rotating from @xmath58 , causing a significant intensity reduction in the directly reflected channel , i.e. the co - polarized counts should decrease .
this should also occur just as the cross - polarized counts increase .
this anti - correlation gives a more accurate measure of when the qd is on resonance , as it reduces the influence of uncorrelated background counts between the two channels .
to explore this fig .
[ fig:1](b ) shows a 2d histogram revealing the correlation in the count distribution between the co and cross channels of the herald arm ( apd-5,6 ) , again , for @xmath51 time bins . a clear pattern emerges .
by far the largest majority of time bins populate a region where the co - polarized counts are high , and the cross - polarized counts are low .
the maximum in the distribution occurs for apd-5 ( co ) @xmath591mhz , and apd-6 ( cross ) @xmath5950khz . at this point
, we expect that the qd is not close to resonance with the laser , and that the phase shift measured will be low .
this is confirmed by the plot in fig .
[ fig:1](c ) , showing the measured phase associated with the time bins for each co / cross point in fig .
[ fig:1](b ) . for all values where co @xmath60400khz , the cross value is always @xmath61250khz , and the measured phase is @xmath62 . however , it is also clear that there is a distribution of time bins where cross count - rate values are larger than the co count - rates .
fig .
[ fig:1](b ) shows a region for cross @xmath60250khz and co @xmath61250khz , where a hot - spot " occurs with a clear distribution centered at 420khz in the cross - polarised channel and 140khz in the co - polarised channel .
the probability of seeing counts in this region is @xmath63 .
the probability of seeing such events due to the poissonian statistics alone is negligible : @xmath64 .
the complete reversal in the distribution of count - rates between apd-5 , and 6 on its own already indicates a phase shift exceeding @xmath16 .
a very crude estimate of the phase shift obtained by using the ratio @xmath65 would give a value of @xmath66 .
however , one needs to be careful in using these values : using selected heralded data introduces a bias to the measurements , and one must also take into account the spin - selective loss in the cavity , as noted before .
fig.[fig:1]c shows the average unbiased phase shift measurements from the phase shift arm plotted as a function of the heralding count - rates in apd-5 and -6 .
the average phase shift for the hot - spot region ( apd-5 @xmath61250khz , apd-6 @xmath60250khz ) gives a consistent phase shift of @xmath670.45@xmath6 .
further , there are clearly pockets " in the data where the phase shift is higher than @xmath16 .
in particular one observes the highest phase shifts for very low @xmath68 ratios in the herald arm .
this implies that there are specific time windows that allow the highest phase shifts to be measured .
we calculate that to measure a @xmath69 phase shift , the laser should be within @xmath70 of the peak of the @xmath71ev linewidth transition .
we calculate that this should occur @xmath72 of the time ( assuming inhomogeneous broadening of a 0.7@xmath73ev qd line to 5@xmath73ev ) .
we also require that enough photon counts can be gathered in the time window in question to overcome poissonian statistics , and that the @xmath74 spin lifetime of the transition is longer than the time bin used . by varying the time bin width
we determine a spin @xmath74 time of 250@xmath73s and a spectral jitter time of @xmath75 ( see supplementary ) .
finding the instances where near exact resonance conditions are fulfilled , will allow the highest phase shift possible to be measured . the highest phase shift measured in the 100@xmath73s bins in fig .
[ fig:1](c ) corresponds to @xmath76 ( @xmath77 ) .
however by analysing a range of time bin lengths we find several regions with phase shifts @xmath3 .
for a shorter time bin of @xmath78 we find a maximum phase shift of @xmath79 ( @xmath80 ) .
we believe that the limit in the phase shift measured is due to background counts , i.e. that some photons have reflected from the top of the pillar , due to mode mismatch between the mode of the laser into the microscope objective and the micropillar cavity mode .
a phase shift of @xmath81 implies that @xmath5980% of the photons input into the pillar are collected and have interacted with the qd .
thus the numerical aperture / spatial overlap ( mode - matching ) between the input laser and cavity mode is @xmath53 .
while @xmath5 interaction efficiency is good , it is also clear that extremely good mode - matching is required to reduce background counts ( @xmath82 mode - matching is required for @xmath83 ) .
this could possibly be achieved using an integrated fibre approach @xcite .
we conclude that using a low q - factor , high efficiency qd microcavity system , one may achieve deterministic photon - spin interactions , by inducing a @xmath3 phase shift on a narrow bandwidth single photon .
we measure by far the largest ever photon phase shift from a solid state quantum emitter . while this qd - micropillar system is somewhat lossy ,
@xmath5 of the reflected photons in the collection channel undergo a deterministic @xmath6 phase shift .
two challenges need to be surmounted before the qd - micropillar can be used as a single photon switch or entangler .
firstly , spectral jitter of the qd transition arising from charge fluctuations around the qd is problematic as only @xmath84 of the time photons are in resonance with the transition , where they meet our heralding condition .
however , as the spectral jitter occurs on a ms timescale , electric field gates applied as part of a feedback circuit , as shown in @xcite are viable .
secondly , exceptional reduction in background scatter and @xmath85 optical mode matching is required to achieve a near deterministic @xmath6 phase shift .
nevertheless , this device demonstrates that deterministic and efficient spin - induced photon switching and spin - photon entangling operations are feasible with a solid - state emitter .
the authors would like to thank h.f .
hofmann for helpful discussions .
this work was funded by the future emerging technologies ( fet)-open fp7 - 284743 [ project spin photon angular momentum transfer for quantum enabled technologies ( spangl4q ) ] and the german ministry of education and research ( bmbf ) and engineering and physical sciences research council ( epsrc ) ( ep / m024156/1 , ep / n003381/1 and ep / m024458/1 ) . |
magnetoelectric multiferroics , insulating magnets exhibiting simultaneous magnetic and ferroelectric order , are widely investigated in large part due to their potential applications for developing novel devices , including magnetic sensors and multistate memory , among others [ ] . the cross - control of these distinct order parameters , such as adjusting the magnetization using an applied electric field or vice - versa , is expected to provide an extra degree of freedom in developing new types of spin - charge coupled devices , such as voltage switchable magnetic memories [ ] . additionally , there are a number of fundamental materials questions surrounding the development of multiferroic order .
magnetic and ferroelectric order are generally contraindicated in the same phase , as ferromagnetism in transition metal systems typically requires partially filled d - orbitals while ferroelectric distortions are promoted in a d@xmath2 electronic configuration [ ] . despite this apparent restriction , a rather large number of single phase systems have been identified as magnetoelectric multiferroics [ ] . a number of microscopic mechanisms have been proposed for the development of multiferroic order , including , a magnetic jahn - teller distortion [ ] for tbmn@xmath3o@xmath4 , [ ] bond and site ordering having distinct centers of inversion symmetry , [ ] a microscopic mechanism leading to a spin - current interaction [ ] , the dzyloshinskii - moriya interaction , [ ] a general anisotropic exchange striction , [ ] a spin - phonon interaction , [ ] and a strain induced ferroelectricity . [ ] phenomenologically , magnetically - induced ferroelectric order developing in systems having multiple magnetic phases can be understood by considering a trilinear term in the magnetoelectric free energy , f@xmath5 , coupling the electric polarization with two distinct order parameters @xmath6 and @xmath7 which together break inversion symmetry so that f@xmath8p@xmath9 [ ] . since the free energy must transform as a scalar , there are strong symmetry restrictions on the allowed representations for @xmath6 and @xmath7 ; in particular , the product @xmath10 must be antisymmetric under spatial inversion .
this trilinear coupling also predicts electric field control of a magnetic order parameter [ ] .
a general discussion of the symmetry of the magnetoelectric coupling in multiferroics is considered for the specific case of in the following section .
investigations on multiferroic thin films , in which such a trilinear coupling is believed to be responsible for the multiferroic order [ ] , have established that the multiferroic transition temperature can be varied through the application of either ( or both ) magnetic and electric fields [ ] , confirming the strong coupling between magnetic and dielectric degrees of freedom .
higher order magnetoelectric coupling terms quadratic in both magnetic and ferroelectric terms will give rise to magnetization induced shift in the dielectric response [ ] . such coupling has been investigated both theoretically and experimentally in a range of materials including mn@xmath11o@xmath12 [ ] , cocr@xmath3o@xmath12 [ ] , bamnf@xmath12 [ ] , and secuo@xmath11 and tecuo@xmath11 [ ] . because this magnetodielectric coupling is also expected to depend strongly on the symmetry of the magnetic phase , it has been suggested that changes in this coupling may be used to probe changes in the ordered spin structure [ ] .
triclinic iron vanadate , , has recently been identified as a multiferroic system having the p@xmath13 space group [ ] .
magnetic , thermodynamic , and neutron diffraction studies on single crystal and ceramic samples have shown that transitions from a paramagnetic phase into a collinear incommensurate ( ci ) phase at @xmath14=22 k and then into non - collinear incommensurate ( nci ) phase at @xmath0=15 k. [ ] ferroelectric order in develops in this non - collinear spiral magnetic phase .
the onset of ferroelectric order with the development of a second magnetic phase suggests that a symmetry - based approach may be useful in exploring the multiferroic properties in this system .
we present a full landau theory for this system , specifically considering the allowed magnetoelectric coupling terms .
our result is that a nonzero induced spontaneous polarization @xmath15 requires having a magnetic spiral [ ] described by two order parameters which are out of phase with respect to one another .
[ ] in this low symmetry structure there are no restriction on the orientation of @xmath15 based on symmetry arguments , unlike the majority of similar magnetically - induced multiferroics [ ] .
this paper is organized as follows . in sec .
ii we present a symmetry analysis of fevo@xmath12 based on landau theory . here
we analyze the symmetry of the magnetoelectric interaction . in sec .
iii we present the results of a number of experiments designed to probe the structure of these magnetoelectric interactions . in sec .
iv we briefly summarize our results .
motivated by this general discussion of the possibility of magnetically driven ferroelectric order in , we now present a landau theory for with some details of the construction relegated to the the appendix . as discussed in detail in ref .
[ ] , the fourier transform of the spin ordering is proportional to the critical eigenvector of the inverse susceptibility matrix at the ordering wave vector @xmath16 . in the appendix
we analyze the constraint of spatial inversion in the p@xmath13 space group of the paramagnetic phase , with the following results .
the structure consists of six @xmath17=5/2 fe@xmath18 spins in the unit cell at locations @xmath19 . for @xmath20 , @xmath21 and @xmath22 .
then inversion symmetry ( @xmath23 ) implies that spin fourier transform obeys @xmath24 where , as defined in the appendix , @xmath25 is the spatial fourier transform of the thermally averaged spin operator .
as explained in the appendix , this relation implies that the spin distribution is inversion - symmetric about some origin .
we find that inversion symmetry implies that @xmath26 & = & \sigma_n [ x_1^ * , y_1^ * , z_1^ * , x_2^*,y_2^ * , ... z_6^ * ] \ \end{aligned}\ ] ] where all the components are complex valued with @xmath27 , @xmath28 , @xmath29 , are normalized by @xmath30 = 1 $ ] , and the wave vector argument is implicit .
the amplitude @xmath31 is the complex valued magnetic order parameter , which obeys @xmath32 as noted in the appendix , this relation implies that each @xmath33 is inversion invariant about a lattice point ( which depends on @xmath34 ) where the order parameter wave has its origin .
as the temperature is lowered one passes from the paramagnetic phase into a phase with an order parameter @xmath35 and then , at a lower temperature into a phase where two order parameters @xmath35 and @xmath36 are nonzero , both of which obey eq .
( [ isym ] ) , but which have different centers of inversion symmetry .
the total magnetoelectric free energy , @xmath37 can be written as @xmath38 where @xmath39 is the purely magnetic free energy , @xmath40 is the dielectric potential which we approximate as @xmath41 , where @xmath42 is the dielectric susceptibility ( whose crystalline anisotropy is neglected ) , and to leading order in @xmath33 , the magnetoelectric coupling term is given by : @xmath43 p_\gamma \ , \end{aligned}\ ] ] where @xmath34 and @xmath44 label order parameter modes and @xmath45 labels the cartesian component of @xmath46 .
terms linear in @xmath33 are prohibited because they are not time reversal invariant and also can not conserve wave vector .
the magnetoelectric interaction @xmath47 has to be inversion invariant and the appendix shows that the @xmath48 coefficients are pure imaginary , so that @xmath49 p_\gamma = 2 \sum_\gamma r_\gamma |\sigma_1(\vec q ) \sigma_2(\vec
q)| \sin(\phi_2 - \phi_1 ) p_\gamma \ , \end{aligned}\ ] ] where @xmath50 .
there is no restriction on the direction of the spontaneous polarization , so that all components of @xmath46 will be nonzero .
however , if the magnetic structure is a spiral , then the arguments of mostovoy [ ] might be used to predict the approximate direction of @xmath46 .
the result of eq .
( 6 ) is quite analogous to that for ni@xmath11v@xmath3o@xmath51 [ ] or for tbmno@xmath11 [ ] , in that it requires the presence of two modes @xmath52 and @xmath53 which are out of phase with one another : @xmath54 .
then the order parameter wavefunctions have different origins and will therefore break inversion symmetry . if , as stated in ref . , the eigenvector is _ not _ inversion invariant as implied by eq .
( [ isym ] ) , then one would conclude that the magnetic ordering transition is not continuous . however , the most likely scenario is that the ordering transitions are continuous and that the spin distribution for each @xmath55 is inversion symmetric as obtained in this derivation .
the acentric distribution found in ref .
differs only slightly from being inversion symmetric for reasons that are obscure .
[ ] as discussed in detail in ref .
the fourier transform of the distribution just below a continuous magnetic ordering transition is proportional to the critical eigenvector of the inverse susceptibility matrix .
( the critical eigenvector is the one whose eigenvalue first approaches zero , i. e. which first becomes unstable , as the temperature is lowered through the ordering transition . ) we introduce the inverse susceptibility as follows .
the thermally averaged spin at the site at position @xmath88 in the unit cell at @xmath89 , @xmath90 is defined as @xmath91 \ , \end{aligned}\ ] ] where @xmath92 is the quantum spin operator at site @xmath93 and @xmath94 is the density matrix : @xmath95\ , \end{aligned}\ ] ] where @xmath96 is the hamiltonian of the system .
the fourier transform of the spin distribution is given by @xmath97 where @xmath98 is the actual position @xmath99 of the spin and @xmath100 is the total number of unit cells in the system . following landau
, we write the free energy , @xmath101 as an expansion in powers of @xmath102 as @xmath103 \ , \label{feq } \end{aligned}\ ] ] where the matrix @xmath104 is the hermitian inverse susceptibility matrix . of course
, we do not know or wish to consider the exact form of @xmath96 and we do not attempt to construct the inverse susceptibility from first principles .
but we can analyze how symmetry influences the structure of the inverse susceptibility . in what follows
we assume that the wave vector @xmath16 at which ordering occurs has been established experimentally and therefore we focus only on that wave vector .
we now consider the case of fevo@xmath12 which hase six spin sites within the unit cell of the space group p@xmath105 .
the only point group symmetry element is spatial inversion about the origin @xmath23 , so that the six sites consist of three pairs of sites @xmath106 and @xmath107 , with @xmath20 .
since the group of the wave vector contains only the identity element , the standard analyses based on this group would indicate that an allowed spin distribution function is a basis function of the identity irrep and therefore that symmetry places no restriction on the form of the spin distribution function .
however , since @xmath23 is a symmetry of the system when all the spins are zero , the free energy of the system for a configuration with an arbitrary distribution of @xmath108 is the same as that for a configuration obtained by inversion applied to the distribution @xmath109 .
so we consider the effect of inversion on @xmath109 .
the effect of @xmath23 is to move a spin , without changing its orientation ( because spin is a pseudo vector ) , from an initial location @xmath98 , to a final location @xmath110 .
this means that @xmath111 where , for @xmath20 , @xmath112 it then follows that @xmath113 because we have 6 spins in the unit cell each having three cartesian spin components the matrix @xmath104 is an 18 @xmath114 18 matrix which we write in terms of 9 @xmath114 9 submatrices ( for @xmath20 and @xmath115 , respectively ) as @xmath116 \ .\end{aligned}\ ] ] now we consider the invariance of the free energy under spatial inversion : @xmath103 \nonumber \\ & = & \frac{1}{2 } \sum_{\vec q , \alpha , \tau , \beta , \tau ' } f_{\alpha , \tau ; \beta , \tau ' } ( \vec q ) [ { \cal i } s_\alpha ( { \vec q } , \tau ) ^ * ] [ { \cal i } s_\beta ( { \vec q } , \tau ' ) ] + { \cal o } [ s({\vec q})^4 ] \nonumber \\ & = & \frac{1}{2 } \sum_{\vec q , \alpha , \tau , \beta , \tau ' } f_{\alpha , \tau ; \beta , \tau ' } ( \vec q ) s_\alpha ( { \vec q } , \overline \tau ) s_\beta ( { \vec q } , \overline \tau ' ) ^ * + { \cal o } [ s({\vec q})^4 ] \nonumber \\ & = & \frac{1}{2 } \sum_{\vec q , \alpha , \tau , \beta , \tau ' } f_{\alpha , \tau ; \beta , \tau ' } ( \vec q)^ * s_\alpha ( { \vec q } , \overline \tau ) ^ * s_\beta ( { \vec q } , \overline \tau ' ) + { \cal o } [ s({\vec q})^4 ] \nonumber \\ & = & \frac{1}{2 } \sum_{\vec q , \alpha , \tau , \beta , \tau ' } f_{\alpha , \overline \tau ; \beta , \overline \tau ' } ( \vec q)^ * s_\alpha ( { \vec q } , \tau ) ^ * s_\beta ( { \vec q } , \tau ' ) + { \cal o } [ s({\vec q})^4 ] \ . \label{feq2 } \end{aligned}\ ] ] the next - to - last equality follows because the free energy is real . the last equality is obtained by interchanging the roles of the dummy variables @xmath117 and @xmath118 and the roles of @xmath119 and @xmath120
. we now compare eq .
( [ feq ] ) and the last line of eq .
( [ feq2 ] ) .
since these forms have to be equal irrespective of the values of the @xmath17 s , we must have that @xmath121 this equality relates ( for @xmath122 ) the submatrices @xmath123 and @xmath124 and ( for @xmath125 and @xmath126 ) @xmath127 and @xmath128 . as a result
we see that @xmath129 , so that @xmath127 is symmetric and @xmath130 .
thus @xmath131 \ .\end{aligned}\ ] ] then the eigenvectors @xmath132 $ ] , written in terms of the nine component vectors @xmath133 and @xmath134 satisfy @xmath135 for instance @xmath136 \nonumber \\ \phiv & = & [ s_x(\vec q , 4 ) , \ s_y(\vec q , 4 ) , \ s_z(\vec q , 4 ) , \
s_x(\vec q , 5 ) , \
s_y(\vec q , 5 ) , \
s_z(\vec q , 5 ) , \
s_x(\vec q , 6 ) , \
s_y(\vec q , 6 ) , \
s_z(\vec q , 6 ) ] \ .\end{aligned}\ ] ] the second equation of eq .
( [ mateq ] ) can be written as @xmath137 so if @xmath138 $ ] is an eigenvector with eigenvalue @xmath139 , then so is @xmath140 $ ] . in principle , these could be two independent degenerate eigenvectors .
but if one considers the simple case when @xmath141 and @xmath142 , where @xmath48 , and @xmath143 are scalars and @xmath144 is the unit matrix , one sees that these two solutions are , apart from a phase factor , the same . only for special values of the matrices
are these two eigenvectors distinct degenerate solutions .
this is an example of an accidental degeneracy whose existence we exclude .
therefore the condition that these two solutions only differ by a phase factor leads to the result that @xmath145 thus the @xmath34th eigenvector is @xmath146 = e^ { i\rho_n /2 } [ e^{i \rho_n/2 } \phiv^ * , e^{-i \rho_n/2 } \phiv ] \ , \end{aligned}\ ] ] which we write in canonical form as @xmath147 \ \end{aligned}\ ] ] where @xmath148 is a complex - valued amplitude and @xmath149 is normalized : @xmath150_j|^2 \ .\end{aligned}\ ] ] since the inverse susceptibility matrix is 18 dimensional , there are 18 eigenvectors , each of this canonical form .
we identify @xmath55 as the order parameter which characterizes order of the @xmath34th eigenvector . as the temperature
is lowered one such solution ( which we label @xmath151 ) becomes critical and at a lower temperature a second solution ( which we label @xmath152 ) becomes critical . as we shall see in a moment , the magnitudes of the associated order parameters @xmath153 and their relative phase are fixed by the fourth order terms in the free energy which we have so far not considered .
( [ eq5 ] ) we see that @xmath154 = [ \sigma_n(\vec q)^ * \thetav_n^ * , \sigma_n(\vec q)^ * \thetav_n ] = \sigma_n(\vec q)^ * \lambda_n \ , \label{inveq } \end{aligned}\ ] ] which indicates that the order parameter transforms under inversion as @xmath155 also , under spatial translation , @xmath156 , we have that @xmath157 note that eq .
( [ ieq ] ) does _ not _ imply that the @xmath34th eigenvector is invariant under inversion _ about the origin_. however , as we now show , it does imply that the @xmath34th eigenvector is invariant about an origin which depends on the choice of phase of the @xmath34th eigenvector .
( it is obvious that a cosine wave is only inversion invariant about one of its nodes which need not occur at the origin . )
if @xmath158 denotes inversion about the lattice vector @xmath159 , then we have @xmath160 let @xmath161 . then if we choose @xmath89 so that @xmath162 , then @xmath163 so , eq . ( [ ieq ] ) implies inversion symmetry about a point which can be chosen to be arbitrarily close to a lattice point for an infinite system .
thus the contribution to the free energy from these order parameters @xmath55 at wave vector @xmath16 can be written as @xmath164 \nonumber \\ & & \ + \sum_{n < m } c_{nm } |\sigma_n(\vec
q ) \sigma_m(\vec q)|^2 + \sum_{n
< m } \biggl ( d_{nm } [ \sigma_n(\vec q ) \sigma_m(\vec q)^*]^2 + d_{nm}^ * [ \sigma_n(\vec q)^ *
\sigma_m(\vec q)]^2 \biggr ) \ , \label{ffeq } \end{aligned}\ ] ] where translational invariance indicates that for an incommensurate wave vector the free energy is a function of @xmath165 , @xmath166 , @xmath167 , and @xmath168 .
in writing this free energy we have assumed that the wave vectors of @xmath6 and @xmath7 are locked to be the same , as discussed in ref . .
the generic situation in multiferroics is that as one lowers the temperature an order parameter @xmath6 first becomes nonzero and then , at a lower temperature , a second order parameter @xmath7 becomes nonzero .
in many cases , such as ni@xmath11v@xmath3o@xmath51 [ ] or tbmno@xmath11 [ ] @xmath6 and @xmath7 have different nontrivial symmetry . here
all the order parameters have the symmetry expressed by eqs .
( [ ieq ] ) and ( [ teq ] ) .
( the phase @xmath169 of the second order parameter is fixed relative to that , @xmath170 , of the first order parameter by the term in @xmath171 in eq .
( [ ffeq ] ) . ) finally , we consider the magnetoelectric coupling , @xmath47 , in the free energy which is responsible for the appearance of ferroelectricity ( for which @xmath172 , where @xmath46 is the electric polarization ) . we write @xmath173 where @xmath39 is the purely magnetic free energy of eq .
( [ ffeq ] ) , @xmath40 is the dielectric potential which we approximate as @xmath41 , where @xmath42 is the dielectric susceptibility ( whose crystalline anisotropy is neglected ) , and to leading order in @xmath33 @xmath43 p_\gamma \ , \end{aligned}\ ] ] where @xmath34 and @xmath44 label order parameter modes and @xmath45 labels the cartesian component of @xmath46 .
terms linear in @xmath33 are prohibited because they can not conserve wave vector .
terms of order @xmath174 or higher can exist .
[ ] the interaction @xmath47 has to be inversion invariant . since @xmath175 and @xmath176
, we see that the terms with @xmath177 are not inversion invariant and hence are not allowed . thus @xmath178 p_\gamma \
.\end{aligned}\ ] ] using @xmath179 and eq .
( [ ieq ] ) we see that inversion invariance implies that @xmath180 , where @xmath181 is real
. then @xmath49 p_\gamma = 2 \sum_\gamma r_\gamma |\sigma_1(\vec q ) \sigma_2(\vec
q)| \sin(\phi_2 - \phi_1 ) p_\gamma \ .
\label{eqres } \end{aligned}\ ] ] note that there is no restriction on the direction of the spontaneous polarization , so that all components of @xmath46 will be nonzero .
however , if the magnetic structure is a spiral , then the arguments of mostovoy [ ] can be used to predict the approximate direction of @xmath46 .
the result of eq .
( [ eqres ] ) is quite analogous to that for ni@xmath11v@xmath3o@xmath51 [ ] or for tbmno@xmath11 [ ] , in that it requires the two modes @xmath52 and @xmath53 to be out of phase with one another , in other words that @xmath54 . if , as stated in ref . , the eigenvector is _ not _ inversion invariant as implied by eq .
( [ inveq ] ) , then one would conclude that the magnetic ordering transition is not continuous . however , the differences between the diffraction patters of the structure of ref . and that suggested here are subtle enough [ ] that our suggested structure seems probably the correct one .
m. kenzelmann , a. b. harris , a. aharony , o. entin - wohlman , t. yildirim , q. huang , s. park , g. lawes , c. broholm , n. rogado , r. j. cava , k. h. kim , g. jorge , and a. p. ramirez , phys .
b * 74 * , 014429 ( 2006 ) . see sec . vc .
motivated by eq . ( 6 ) , which predicts that the magnetic structure defined by @xmath35 and @xmath56 is coupled to the electric polarization @xmath57 , we experimentally investigated the nature of the higher order magnetoelectric coupling in .
bulk single phase polycrystalline iron vanadate ( ) ceramic samples were prepared using standard solid state reactions . because eq .
6 predicts that all components of the polarization vector are nonzero , and previous measurements on ceramic have found clear evidence for multiferroic behaviour [ ] we focused our study on polycrytalline samples . a stoichiometric ratio of iron oxide ( fe@xmath3o@xmath11 ) and vanadium pentaoxide ( v@xmath3o@xmath4 ) solid solutions were thoroughly mixed and ground to produce a homogeneous mixture .
this mixture was slowly heated to 600@xmath58c for 4 hours in air .
intermediate grindings followed by thermal annealing in air were repeated several times to complete the solid state reaction and ensure a fully reacted and uniform composition .
this homogeneous solid solution was finally annealed in air at 800@xmath58c for 4 hours , yielding a yellowish brown powder identified as a single phase iron vanadate by x - ray diffraction and raman spectroscopy . in order to apply large electric fields to
we also prepared thin film samples .
these were fabricated from a phase pure stoichiometric iron vanadate target .
the powder used for the sputtering target was prepared by the method described above .
approximately 30 g of powder was mixed with 15 ml of 2 mole percent polyvinyl alcohol as a binder .
the dried powder was pressed into a circular disc having a diameter of approximately 50 mm with a thickness of roughly 3.5 mm followed by air annealing at 600@xmath58c for 4 hours to burn off the residual organics .
a final thermal annealing was done at 800@xmath58c for 4 hours to produce the dense pellet used for the sputtering target .
films were deposited at room temperature using rf magnetron sputtering onto conducting silicon substrates .
the working pressure was held at 1.5x10@xmath59 torr , with the atmosphere consisting of a mixture of approximately 1.5x10@xmath60 torr partial pressure of oxygen as the reactive gas and approximately 1.35x10@xmath59 torr partial pressure of argon as the sputtering gas .
these as deposited films , prepared over a time of 4 hours , were amorphous . after air annealing at 700@xmath58c for 4 hours the films were indexed as single phase polycrystalline .
we investigated the structural , magnetic , and electronic properties of these samples using a number of different techniques .
we used a rigaku ru200 powder x - ray diffractometer and horiba triax raman spectrometer to study the crystalline structure of these samples .
we used a hitachi scanning electron microscope ( sem ) to investigate the surface morphology of the thin film samples and an associated energy dispersive x - ray ( edx ) assembly to probe the chemical composition of both samples .
we measured the temperature dependent magnetization of the powder sample using a quantum design mpms squid magnetometer , although the very small magnetic anomalies associated with the transitions could not be clearly distinguished from background in the thin films samples .
we conducted temperature and field dependent dielectric and pyrocurrent measurements using the temperature and field control provided by a quantum design ppms system used in conjunction with an agilent 4284a lcr meter and a keithley 6517 electrometer .
these measurements were done on a cold pressed pellet of bulk with the top and bottom electrodes fashioned using silver epoxy and on the thin films with room sputtered gold ( au ) used as the top electrode and the si substrate serving as the bottom electrode .
-2@xmath61 x - ray diffraction ( xrd ) pattern of thin film , ( b ) surface scanning electron micrograph of thin film , ( c ) cross - sectional sem image of , thin film and ( d ) room temperature raman spectrum on bulk powder and thin film samples .
the peak at 560 @xmath62 ( indicated by an asterisk ) arises from the silicon substrate , title="fig:",width=396 ] + the structure of the ceramic sample was practically identical to that previously presented for a bulk sample prepared using a different technique [ ] . the x - ray diffraction ( xrd ) pattern for the thin film is shown in fig 1(a ) .
these diffraction peaks are consistent with the expected xrd pattern for [ jcpds no .
38 - 1372 ] .
the surface morphology the thin film sample is shown in fig 1(b ) .
this sem micrograph indicates that the film consists of grains with various orientations , as well as a number of pinhole defects .
we calculated the thickness of these thin films to be roughly 200 nm , using the cross - sectional sem micrograph , fig 1(c ) .
this value is very consistent with estimates from well defined interference fringes observed in reflection spectra ( not shown ) .
edx analysis of both the bulk and thin film samples show a 1:1 iron to vanadium ratio .
we carried out room temperature raman vibrational spectroscopy to further probe the microstructures of both bulk and thin films .
the identification of raman active modes and their detailed temperature dependent analysis on bulk sample is discussed elsewhere [ ] . here we plot the room temperature raman spectrum of both bulk and thin film in fig1(d ) .
we are able to identify all the raman active modes for thin films , which are observed in bulk [ ] with a small shift in the raman peaks for the thin films .
the raman peak arising from the silicon substrate is indicated by an asterisk .
the temperature dependent magnetization for the bulk sample ( not shown ) was practically identical to that measured previously on a different ceramic sample prepared using a different technique [ ] . in particular , the magnetization showed the usual two anomalies associated with the two incommensurate transitions in this system .
we plot the zero - field dielectric constant and resistivity for bulk over a broad range of temperatures in fig .
the dielectric constant exhibits a sharp peak near @xmath0 , arising from the development of ferroelectric order in the incommensurate spiral magnetic phase . as shown in fig .
2(a ) , above 35 k , the dielectric constant for shows a gradual decrease on cooling , typical of many insulating materials [ ] . below roughly 30 k , the dielectric constant increases smoothly with further cooling .
since the resistivity of increases monotonically with decreasing temperature ( except for a small anomaly at @xmath0 ) , also shown in fig .
2(a ) , we attribute this increase in the dielectric constant to a quartic magnetoelectric coupling , @xmath63 .
although does not order magnetically until cooled below @xmath14=22 k , heat capacity measurements suggest the presence of short range spin correlations developing well above this temperature [ ] .
it has been suggested in a number of other systems , including tecuo@xmath11 [ ] and mn@xmath11o@xmath12 [ ] , that short - range magnetic correlations can produce magnetodielectric corrections ; we propose that the same mechanism is responsible for the non - monotonic temperature dependence of the dielectric constant of in the paramagnetic phase .
the fourth order magnetoelectric coupling contains terms quadratic in @xmath46 and @xmath55 .
if @xmath35 is the order parameter that develops at @xmath14 , then this coupling is probably dominated by @xmath64 at high temperatures .
the finite spin correlations developing above @xmath14 cause @xmath65 to be nonzero so that the coupling term is in effect @xmath66 , where @xmath67 . this term produces a shift in the dielectric constant in the paramagnetic phase , as seen in fig .
2(a ) below approximately 30 k. below @xmath14 , approximately 22 k , when @xmath35 acquires a finite expectation value , a trilinear coupling term @xmath68 is allowed .
this term will lead to mode mixing so that the critical mode approaching @xmath0 is not @xmath36 but @xmath69 where @xmath70 is of order @xmath71 [ ]
. then the divergence in this variable as @xmath0 is approached will lead to a simultaneous divergence ( with a very much reduced amplitude ) in the observed dielectric constant .
this mode mixing is therefore expected to lead to a slight increase in the magnetodielectric shift below @xmath14 [ ] . this is seen clearly in fig .
2b , where the dashed line shows the extrapolation of the magnetodielectric shift from @xmath72 to @xmath73 , an extrapolation that does not include any corrections arising from the trilinear magnetoelectric term . to further investigate spin - charge coupling in , we plot the temperature dependent dielectric constant measured at different magnetic fields in fig 2(c ) .
we find that the dielectric anomaly signaling the onset of ferroelectric order shifts to lower temperatures with increasing magnetic field , with the reduction in transition temperature reaching 0.7 k in a magnetic field of @xmath74=80 koe .
this result is expected , as the ferroelectric order producing the dielectric anomaly is associated with the incommensurate spiral transition , which typically show a reduction in transition temperature in applied magnetic fields .
we conducted additional measurements of the dielectric response of the bulk sample while sweeping the magnetic field at fixed temperature .
these results are shown in fig .
3 , plotted as @xmath75 versus @xmath74 with data measured at different temperatures offset vertically for clarity . at @xmath76=17 k , which is intermediate between @xmath14 and @xmath0 , there is a small negative magnetocapacitance , with the dielectric constant being reduced by approximately 0.03% in a field of @xmath74=80 koe .
as the temperature approaches the multiferroic transition at @xmath0 , the magnetodielectric coupling shows qualitative changes . by @xmath76=15
k the magnetocapacitive shift is positive for small fields , with a shift in dielectric constant on the order of 0.02% at high magnetic fields .
the magnetocapacitive response is maximal near @xmath76=14.5 k , with the dielectric constant being reduced by approximately 0.1% in a field of @xmath74=80 koe . at still lower temperatures the magnitude of the magnetocapacitive shift becomes smaller .
perhaps the most dramatic feature in the isothermal magnetocapacitance curves presented in fig .
3 is the presence of clear maxima , which vary as a function of temperature and magnetic field .
these maxima appear first at small fields at @xmath76=14.5 k , then shift to larger fields as the temperature is reduced .
we believe that these anomalies do not reflect the suppression of the multiferroic transition temperature in a magnetic field , as discussed in the context of fig 2(b ) .
these isothermal dielectric anomalies persist to temperatures 2 k or 3 k below @xmath0 , while the maximum suppression of @xmath0 was only 0.7 k over the field range studied , as determined from the measurements in fig .
3 . we propose that this dielectric anomaly may indicate a spin - reorientation transition in .
the magnetodielectric coupling is expected to depend on the symmetry of the magnetically ordered state [ ] , so a field induced spin reorientation crossover could potentially produce the low temperature dielectric anomalies observed in fig .
similar magnetic field - induced dielectric anomalies have been observed in other materials including mn@xmath11o@xmath12 [ ] , although the specific mechanisms responsible remain unclear .
one possibility is that the external magnetic field serves to reduce the slight geometrical frustration present in [ ] , allowing a different spin structure to emerge .
alternatively , the spin orientation could be a spin - flop transition as seen in tbmno@xmath11 .
[ ] we note , however , that remains ferroelectric at high magnetic fields [ ] , so the modified spin structures would still need to transform as defined by eq .
3 . the trilinear magnetoelectric coupling that produces multiferroic order , given in eq . 6 , also results in an electric field ( @xmath79 ) dependence of the magnetic structure through the coupling term in the free energy @xmath80 .
we first confirmed that these thin film samples were also multiferroic , through measurements of the dielectric constant and pyrocurrent , illustrated in fig .
the dielectric constant for the thin film is slightly higher than that found for the ceramic sample .
we attribute this discrepancy mainly to the uncertainty in accurately determining the geometrical factor for these thin films .
the dielectric response for these thin film samples is approximately independent of measuring frequency and the loss for these films is tan @xmath81 0.01 , which may be due to the presence of pinhole defects in the thin film sample as seen in the sem micrograph in fig .
the zero - field temperature dependent dielectric constant , measured at @xmath82=30 khz , is plotted in fig .
there is a sharp peak near @xmath0=15 k , associated with the development of ferroelectric order in these thin film samples .
we note that , unlike the measurements on bulk shown in fig .
2(a ) , the background dielectric constant for decreases monotonically with decreasing temperature .
this behaviour can be associated with the much larger conductivity of the thin film sample , arising from the presence of the pinhole defects , which obscures the low temperature increase in dielectric constant observed in bulk ( fig .
2(a ) ) .
we confirmed that the low temperature phase of the thin film is ferroelectric by integrating the pyrocurrent after poling at positive and negative fields to yield the spontaneous polarization .
these results are shown in the inset to fig .
4(a ) and indicate a spontaneous polarization of 6 @xmath83c / m@xmath84 , consistent with previous measurements on polycrystalline bulk [ ] measurements of the dielectric response for thin films under applied magnetic fields ( not shown ) yield a suppression of the multiferroic transition temperature very similar to that observed in bulk ( see fig .
2(b ) ) . to probe the electric field control of the multiferroic phase transition temperature , expected from the nature of the magnetoelectric coupling , we measured the temperature dependent dielectric response in the thin film sample as a function of bias voltage .
focusing on thin film samples allows the application of relatively large electric fields ( on the order of mv / m ) with small applied bias voltages .
we chose to probe the transition through dielectric measurements as the magnetic anomaly at @xmath0 can not be clearly discerned in these thin film samples .
we plot the temperature dependent dielectric constant measured at e=0 and e=3.75 mv / m in fig .
with the application of an electric field , the dielectric peak shifts upwards in temperature , by approximately 0.25 k in a field of e=3.75 mv / m .
we note that any sample heating , which is expected to be negligible in any case because of the low dissipation , would raise the sample temperature relative to the thermometer temperature , leading to an apparent decrease in transition temperature , rather than the increase seen in fig .
this increase in transition temperature is consistent with an external electric field promoting the development of ferroelectric order , and is similar to what has been observed previously in multiferroic films [ ] .
the relatively small increase of the ferroelectric transition under such large applied electric fields can be directly attributed to the very small polarization in .
we confirmed that the dielectric anomaly in fig .
4(b ) can still be associated with the multiferroic transition , even in the presence of an electric field , by measuring the response under the simultaneous application of magnetic and electric fields ( not shown ) .
although the dielectric peak broadens considerably , the continuing presence of a single peak under such crossed fields is strong evidence that this anomaly reflects the multiferroic transition in . and
( b ) magnetic field dependence of multiferroic transition temperature together with the proposed magnetic field induced spin reorientation cross over . here , nci@xmath85 indicates the proposed phase having a spin reoriented structure , title="fig:",width=566 ] + we summarize the results of these magnetoelectric and magnetodielectric studies on in figs 5(a ) and 5(b ) .
we plot the e - field and h - field dependence of the multiferroic transition temperature in ( @xmath0 ) in fig .
5(a ) , where ci and nci represent the incommensurate magnetic structures below @xmath14 and @xmath0 respectively .
this transition temperature is monotonically suppressed in an applied magnetic field , decreasing by approximately 0.7 k in an applied field of @xmath74=80 koe .
the transition temperature , however , increases systematically with increasing bias voltage , shifting upwards by 0.25 k in an electric field of roughly 4 mv / m . as discussed in ref .
[ ] , the magnetic field dependence of the ci - nci phase boundary is expected to follow @xmath86 , while the electric field dependence should be @xmath87 .
this ability to control the transition temperature using either magnetic or electric fields is a key feature for a number of proposed applications for multiferroic materials .
although the size of the transition temperature shifts in are likely too small to be of any practical use , these results , taken in conjunction with previous studies on thin films , provide important evidence that this behaviour is generic among multiferroic materials .
the magnetodielectric coupling in allows us to tentatively identify the onset of short - range magnetic correlations , as indicated by the increase in dielectric constant below @xmath76=30 k in fig .
2(a ) , and also to propose the onset of a spin reorientation crossover , based on the field dependent dielectric anomalies in fig .
3 . using the data from fig .
3 , we plot this proposed spin reorientation cross - over boundary line in fig .
5(b ) , together with the magnetic field dependence of @xmath0 ( similar to that shown in fig .
the high - field putative spin - reorientation structure is labeled as nci@xmath85 . as the two boundaries do not coincide , the dielectric anomalies in fig .
3 are not likely to be associated with the @xmath14 to @xmath0 magnetic transition , but may potentially be attributed to a change in magnetic structure . magnetic field dependent specific heat measurements ( not shown ) do not show any additional anomalies at this proposed cross - over , suggesting there is a negligible change in entropy between the two spin structures .
we have presented a model for the development of multiferroic order in in the context of landau theory , which we have used to develop constraints on the possible magnetic structures based on symmetry considerations alone .
one of the noteworthy predictions of this model is that the electric polarization in the multiferroic phase is able to develop along any direction . to further investigate the higher order magnetoelectric coupling in this system ,
we have investigated the ferroelectric and dielectric response in to applied magnetic and electric fields .
the multiferroic phase transition temperature can be tuned by applying electric or magnetic fields , in line with the predicted trilinear magnetoelectric coupling .
we find evidence for a shift in dielectric constant well above the magnetic transition temperature @xmath14 , which is expected to develop from a fourth order magnetoelectric coupling term when short range spin correlations develop in the paramagnetic phase .
the dielectric constant shows a small , but distinct , increase below the first magnetic order transition , which is consistent with the contribution from a trilinear magnetoelectric coupling term .
we find evidence for magnetic field induced dielectric anomalies in the non - collinear incommensurate magnetic phase of , which we attribute to a spin reorientation transition that does not suppress the ferroelectric structure .
these studies on demonstrate the rich spin - charge coupling present in many multiferroic materials , emphasize the importance to considering higher order expansions of the magnetoelectric coupling to adequately explain the properties of these materials , and illustrate how dielectric spectroscopy can be a valuable tool for probing the magnetic structures in such systems . | we have investigated the magnetoelectric and magnetodielectric response in , which exhibits a change in magnetic structure coincident with ferroelectric ordering at @xmath0@xmath115 k. using symmetry considerations , we construct a model for the possible magnetoelectric coupling in this system , and present a discussion of the allowed spin structures in .
based on this model , in which the spontaneous polarization is caused by a trilinear spin - phonon interaction , we experimentally explore the magnetoelectric coupling in thin films through measurements of the electric field induced shift of the multiferroic phase transition temperature , which exhibits an increase of 0.25 k in an applied field of 4 mv / m . the strong spin - charge coupling in is also reflected in the significant magnetodielectric shift , which is present in the paramagnetic phase due to a quartic spin - phonon interaction and shows a marked enhancement with the onset of magnetic order which we attribute to the trilinear spin - phonon interaction .
we observe a clear magnetic field induced dielectric anomaly at lower temperatures , distinct from the sharp peak associated with the multiferroic transition , which we tentatively assign to a spin reorientation cross - over .
we also present a magnetoelectric phase diagram for . |
in this paper , we outline a framework for modelling a proximity network from transportation data , and propose an analytical method to estimate the infection probability of each individual from stochastic simulations .
the prevalence , i.e. the average number of recovered agents who were once infected before the spread of the epidemic subsides , is of particular interest .
we can construct an individual - based network using a dataset of human mobility .
the _ people flow data _
@xcite record the one - day movement of individuals living in the kanto region of japan , which includes the tokyo metropolitan area .
details of the dataset are given in the methods section .
we randomly chose the mobility track of @xmath0 individuals from this dataset .
time - dependent proximity networks were then constructed by connecting individuals when they came within a certain geographical proximity , defined by a distance threshold @xmath1 ( fig .
[ fig : map ] ) .
let us denote the adjacency matrix of the proximity network as @xmath2 .
all results in the main text were obtained with @xmath3 m and @xmath4 . note that we observed the percolation - like transition and the scaling relation of the giant cluster component size ratio as @xmath5 . using this scaling relation ,
@xmath3 m with @xmath4 corresponds to @xmath6 m with the real population in the corresponding region , where @xmath7 .
a detailed discussion can be found in supplementary information si1 .
we simulated the disease propagation using the agent - based sir model @xcite on the above - mentioned network , which assigns susceptible ( s ) , infected ( i ) , or recovered ( r ) states to each agent .
we assumed that each agent repeats the same trip pattern every day in the same way as recorded in the people flow data . when a susceptible agent is connected with an infected one , the former
is infected with probability @xmath8 for a time interval @xmath9 .
if a susceptible agent is connected with @xmath10 infected agents , they are infected with probability @xmath11 @xcite .
further details of the stochastic simulation are given in the methods section .
in this study , @xmath8 is sufficiently small which enables us to perform the simulation with the discretised time step .
an infected agent recovers with probability @xmath12 over this time interval .
once @xmath2 , @xmath13 , and @xmath14 are given , stochastic simulations can be conducted to determine the epidemic dynamics .
we denote the probabilities that agent @xmath15 is in state s , i , or r at time @xmath16 as @xmath17 , @xmath18 , and @xmath19 , respectively ; the equality @xmath20 holds .
we define the infection probability of agent @xmath15 as @xmath21 , because a recovered agent has experienced the infected state before recovering .
the epidemic dynamics can be characterized by three different stages @xcite : the initial stage , in which stochastic fluctuations are dominant , the exponential growth stage , and the final stage , where nonlinearity suppresses the further spread of the disease .
the latter two stages representing the dynamics of the outbreak are of particular interest . in these stages ,
the time evolution of the epidemic can be approximated by the following deterministic differential equations : @xmath22 here , we approximate the time - dependent adjacency matrix @xmath2 with its time averaged form @xmath23 @xcite .
this averaging gives a sufficient approximation of the time evolution of @xmath24 , @xmath25 , and @xmath26 if the parameters @xmath13 and @xmath14 are sufficiently small , as discussed in supplementary information si2 .
we also assume that node @xmath15 being susceptible and node @xmath27 being infected are statistically independent events @xcite .
although this assumption does not hold for small networks @xcite , we have verified that the numerical solution of equation ( [ eq : mf ] ) gives a good approximation of the dynamics if @xmath1 is much greater than the percolation transition point and the giant cluster component is sufficiently large .
one of the most common methods used to analyse the contagion processes within networks is hmf @xcite .
this approach assumes that , in a statistical sense , nodes grouped by the degrees behave in the same way . under this assumption
, one can derive the equation for the infection probability of each agent , and , in the absence of degree correlation , the epidemic threshold , i.e. the critical value of the epidemic s spread , can be represented as a function of the first and second moments of the degree distribution @xcite .
one question that persists is whether such an approximation is valid in realistic networks of human contact . actually , the accuracy of hmf depends on factors such as the mean degree and first - neighbour degree @xcite . for the simplest case of the absence of degree correlation , where hmf is often applied ,
the component of the first eigenvector is proportional to the degree @xcite .
however , this assumption does not hold for the proximity network ( fig .
[ fig : series]a ) .
this highlights the need to develop an analytical framework to identify people with a high infection probability along with the overall infection propagation pattern . in this study , we have developed a method based on spectral decomposition and mode truncation which demonstrates that a small number of dominant modes , which may not necessarily have the largest eigenvalues , give a higher contribution in the prevalence of epidemics .
in this section , we present a method for analysing the final epidemic size based on the spectral analysis of the averaged matrix @xmath23 , and elucidate the relevant dynamics of the epidemic s spread .
the node - wise probabilities @xmath17 , @xmath18 , and @xmath19 can be expanded as @xmath28 where @xmath29 denotes the @xmath30th eigenvector of @xmath23 associated with the eigenvalue @xmath31 .
all eigenvalues of a real symmetric matrix @xmath23 are real , and can be labeled in descending order as @xmath32 . the normalization condition @xmath33 is adopted . since @xmath34 is symmetric ,
the left and right eigenvectors coincide , and the expansion coefficients can be obtained with @xmath35 , @xmath36 , and @xmath37 , respectively .
equation ( [ eq : mf ] ) can be rewritten in terms of these expansion coefficients ( supplementary information si3 ) .
the time evolution of the average number of infected and recovered agents , @xmath38 and @xmath39 , is plotted in fig . [
fig : series]b .
the time series of @xmath18 is converted to that of @xmath40 as in fig .
[ fig : series]c . in the exponential stage of the epidemic
spreading , @xmath41 , @xmath25 , and @xmath26 are small .
the exponential growth of each coefficient @xmath42 can be described by the following linearised equation @xcite @xmath43 ( see supplementary information si4 for the derivation of this expression from the nonlinear mode coupling equation ) .
hence , @xmath44 holds .
it is evident from this equation that a mode with a larger eigenvalue @xmath45 grows faster in this stage . in particular
, the exponential growth rate of the number of infected agents is characterised by @xmath46 .
the stochastic simulation of the agent - based model verifies that @xmath47 is indeed dominant in the initial stage ( fig .
[ fig : series]c ) .
in addition , the epidemic threshold is related to @xmath48 as @xmath49 ( fig .
[ fig : series]d ) @xcite .
however , as shown in fig .
[ fig : series]c , the coefficients @xmath50 ( @xmath51 ) are comparable to @xmath52 in the later stages of the epidemic , where @xmath53 is the node - wise average of a vector @xmath54 .
the importance of each mode needs to be assessed to ascertain its dominance in the spreading mechanism .
this can be achieved by introducing a quantifier of the _ contribution _ of each mode to the prevalence as @xmath55 the definition of this contribution enables us to account for the prevalence as the sum of the contributions of all modes , @xmath56 .
it is worth introducing the contribution @xmath57 for the case where all agents are infected , i.e. @xmath58 for all @xmath15 .
since the coefficients satisfy @xmath59 in this case , one can easily verify that @xmath60 holds .
interestingly , the contribution @xmath57 does not monotonically increase with the eigenvalue , although a positive correlation can be observed ( fig .
[ fig : final_size]a ) .
since @xmath61 and @xmath57 are functions of the average of the eigenvector components , a mode associated with a smaller eigenvalue may contribute more than one with a larger eigenvalue .
although the larger eigenvalue denotes faster growth in the exponential regime , it does not necessarily mean a larger contribution to the prevalence .
it is interesting to note that the `` hidden '' important structures , corresponding to modes that make a large contribution to the prevalence but have smaller eigenvalues , are unravelled through the spectral analysis .
it is still possible to detect such modes when the prevalence is much smaller than the all - infected case ( fig .
[ fig : final_size]b ) .
the heterogeneity in the contribution is evident from the cumulative contribution @xmath62 and @xmath63 ( fig .
[ fig : final_size]c ) . here ,
the modes are sorted such that @xmath64 , where @xmath65 denotes the permutation of the index set @xmath66 .
note that @xmath67 holds , because the infection pattern is represented using all modes .
figure [ fig : final_size]c suggests @xmath68 for the all - infected case , indicating that approximately 90% of the prevalence is described by the top 10% of modes for the present proximity network .
this figure indicates that the number of modes needed to describe the prevalence decreases if the prevalence is smaller than @xmath69 .
the above observations led us to the idea of describing the epidemic dynamics with a small number of dominant modes that give a high contribution . in the case of the agent - based sir dynamics
, we can derive the truncated equation for the contributions at the final time with @xmath70 ( @xmath71 ) modes , @xmath72 , as @xmath73 \frac{\phi_j^{(\sigma(b ) ) } } { \langle \phi^{\sigma ( b ) } \rangle } \right\ } , \label{eq : self - conf_eig } \end{gathered}\ ] ] where @xmath74 .
this equation is derived in the methods section .
this transcendental equation approximates the epidemic dynamics with fewer effective degrees of freedom .
if we use all the modes , i.e. @xmath75 , the solution of equation ( [ eq : self - conf_eig ] ) coincides with the solutions of equation ( [ eq : mf ] ) for @xmath76 .
furthermore , the solution of this equation computed with newton s method provides a good description of the contribution obtained by the stochastic simulation ( fig .
[ fig : final_size]d ) .
we further note that considering fewer than @xmath0 modes suffices to approximate the prevalence as obtained by the stochastic simulation ( figs .
[ fig : series]d , [ fig : final_size]c ) . hence
, this procedure replaces the need to solve an ordinary differential equation in @xmath77 independent variables ( [ eq : mf ] ) with the requirement to use information from the @xmath70 dominant modes . as discussed above
, the essence of theoretical analyses of epidemic spreading within networks is to infer the dynamical information from the network topology .
conventionally , hmf @xcite uses degree distributions and degree correlations as topological information .
in contrast , our approach uses spectral information , which can take more general properties of the network topology into account . as described in the methods , the final size equation in hmf @xcite
is derived from equation ( [ eq : self - conf_eig ] ) under the following simple assumptions : ( i ) degree correlation is absent , and ( ii ) the first mode almost perfectly governs the dynamics . as shown in fig .
[ fig : series]a , the first assumption is not satisfied in the proximity network .
moreover , it is evident from figs .
[ fig : series]d and [ fig : final_size]a that considering only the first mode is not sufficient to account for the whole epidemic dynamics , as @xmath78 is less than 6% .
the contribution of other dominant modes must be taken into account to quantify the epidemic dynamics ( fig .
[ fig : final_size]d ) .
thus , hmf with no degree correlation is not a sufficient approximation for the proximity network .
it would be possible to improve the approximation by considering the degree correlation in hmf @xcite , but it is generally difficult to include information about the second - nearest neighbours .
in contrast , the proposed approach provides the flexibility to improve the approximation by using an arbitrary number of dominant modes .
another feature of epidemic spreading in proximity networks is the strong heterogeneity of the infection risk in space and time .
it is important to identify such locations that have a high probability of infection , namely `` hot spots '' , for the mitigation and management of the epidemic .
the spatial distribution of the hot spots can be mapped to provide an insight into the nature of epidemic spreading in urban areas ( fig .
[ fig : risk ] ) .
more importantly , to uncover the relevant epidemic dynamics , such identification should be conducted with fewer parameters than the number of agents . for the present model to be predictive ,
let us infer the infection probability of a test agent from the mobility pattern of other agents .
the probability of infection of this agent depends on the presence of other infectious agents in the vicinity , which in turn varies from region to region and at different times of day , @xmath79 .
let @xmath80 be a position vector and @xmath81 be the position of agent @xmath15 at time @xmath16 .
we introduce a kernel function @xmath82 signifying the presence or absence of any interaction between two agents based on the distance between them .
since agents are connected within the threshold @xmath1 , the heaviside step function @xmath83 is adopted as the kernel function .
this allows us to define a spatio - temporal risk factor of infection @xmath84 as @xmath85 which can be interpreted as the weighted sum of probabilities @xmath21 . integrating @xmath86 along the trajectory of the test agent @xmath87 ,
one can obtain the infection probability of this agent .
details of the derivation are discussed in supplementary information si5 .
figure [ fig : risk]a depicts the spatial distribution of the risk factor @xmath84 across the kanto region of japan . in the present model , which assumes homogeneous parameter values of @xmath13 and @xmath14 , the high - risk area during
the daytime is the central business district , where many workers are located . in the nighttime ,
certain suburban areas from where people commute to the centre become high risk .
furthermore , people travelling to the central business district from their suburban residences allow the disease to spread over the whole metropolitan area .
figure [ fig : risk]c shows that the infection probability of the test agents can be estimated from the integration of @xmath86 , obtained from stochastic simulation , along its trajectory .
instead of counting the infection probability @xmath21 in the neighbourhood of the region of interest , the spectral properties of the averaged adjacency matrix can be incorporated into the risk factor by choosing the first few high - contribution modes .
namely , one can approximate @xmath21 and @xmath84 as @xmath88 and @xmath89 using @xmath70 modes .
hazard maps based on the mode truncation at different times are similar to those given by the stochastic simulation ( fig .
[ fig : risk]a ) . moreover , fig .
[ fig : risk]d shows that the integration of @xmath90 along the path of the test agents predicts their infection probability .
thus , we conclude that mode truncation enables us to describe the spatio - temporal infection risk with a small number of modes . in other words , few modes that have high contributions
dominate the spread of the epidemic .
data on human mobility , particularly contact information of urban residents , play a key role in the analysis of an epidemic spreading . such high - resolution data allow detailed individual - level modelling , rather than coarse - grained metapopulation studies . at the same time
, novel theoretical tools can be used to uncover the dominant dynamics at work in data - driven models without the need for oversimplification .
linear analysis characterises the exponential growth in the initial stage ; nonetheless , this study clarified that the final size of the epidemic is provided by the analysis of various modes .
the mode truncation described here revealed that the prevalence of epidemics is actually dominated by a small number of modes as compared to the number of agents , i.e. the effective number of degrees of freedom is much smaller than the system size .
the proposed method practically limits the expanse of mitigation policies such as targeted vaccination , and provides specific alerts by following the spreading trail of infection .
this also quantifies the spatio - temporal risk associated with empirical travel patterns .
the analytical framework and realistic mobility data presented here provide a promising starting point for the study and possible confrontation of epidemics in the real world .
future studies should examine ways to facilitate the rapid decay of dominant modes using dynamic intervention strategies .
people flow data were generated based on person
trip surveys made by japan s ministry of land , infrastructure , transport , and tourism . the surveys were based on questionnaires that asked about basic individual information such as gender , age , occupation , places visited , and trip modes on one day . for the sake of privacy ,
spatial information was recorded with a resolution such that the complete location could not be determined .
the questionnaires allowed us to access advanced information such as transportation modes , travel times , transfer points , and trip purposes .
traces of the trips between two places reported in the raw data were interpolated using geographical information systems methods @xcite .
the proximity network was constructed by placing a link between two individuals at time @xmath16 if they were within the threshold distance @xmath1 of one another ( fig . [ fig : map]b ) .
we added another constraint by only placing a link between individuals using the same transportation mode ; it is impractical to connect individuals walking along the road with those traveling in an automobile or train because of the improbability of the transmission of infection .
the connectivity of the network can be described by the time - dependent adjacency matrix @xmath2 , where @xmath91 if individuals @xmath15 and @xmath27 are connected at time @xmath16 and @xmath92 otherwise .
this allowed us to obtain networks for @xmath0 agents at every @xmath9 time interval .
proximity networks are spatially embedded temporal networks , which appear in various forms of transportation and infrastructure networks @xcite .
the agent - based sir model simulation was performed with the same time - dependent adjacency matrix @xmath2 .
the time evolution of the probabilities @xmath17 , @xmath18 , and @xmath19 was computed using 500 independent stochastic simulation runs with different random initial conditions .
we also assumed that the human mobility patterns had a periodicity of 24 h owing to the circadian nature and economic situation of people going to their place of work and back home .
parameter values were taken as @xmath93 , @xmath94 , and @xmath95 , except for fig .
[ fig : series]d , in which @xmath13 varies .
one of the main results of this article , the equation for the asymptotic solution of equation ( [ eq : self - conf_eig ] ) , is derived as follows .
substituting the expansion ( [ eq : expansion ] ) into equation ( [ eq : mf ] ) , we obtain @xmath96 let us define @xmath97 \phi_j^{(a ) } , \label{eq : psi - r}\end{aligned}\ ] ] so that the solution of this equation is given as @xmath98 or equivalently , @xmath99 for @xmath100 , @xmath101 should vanish , and the conservation of probability condition is given as @xmath102 .
this is equivalent to @xmath103 substituting equation ( [ eq : s_evo ] ) into this expression , we obtain @xmath104 \phi_j^{(b ) } \right\}. \label{eq : self - conf_eig_orig } \end{aligned}\ ] ] this is a transcendental equation for @xmath105 , and one can determine the infection probability by solving this equation instead of solving the differential equation .
we choose @xmath70 ( @xmath106 ) modes and neglect the other modes to obtain @xmath107 \phi_j^{(\sigma(b ) ) } \right\ } , \label{eq : self - conf_eig_orig2 } \end{aligned}\ ] ] where @xmath65 and @xmath108 denote arbitrary permutations of modes .
finally , we multiply both sides by @xmath109 , and obtain equation ( [ eq : self - conf_eig ] ) by choosing the top @xmath70 modes with respect to their contribution .
the relationship between the equation for the expansion coefficients ( [ eq : self - conf_eig_orig ] ) and that obtained with hmf can be described as follows .
we concentrate on the simplest case , where the probability @xmath110 that a node with degree @xmath10 and one with degree @xmath111 are connected is proportional to the product of their degrees , i.e. @xmath112 , where @xmath113 denotes the degree distribution of the network satisfying @xmath114 . in this case ,
the largest eigenvalue of the adjacency matrix is @xmath115 , and the component of the corresponding normalised eigenvector is @xmath116 . here
, we denote the degree of node @xmath15 as @xmath117 .
we assume that the probability of nodes with the same degree is identical , i.e. @xmath118 holds . substituting the expression for the first eigenvector into equation ( [ eq : self - conf_eig_orig ] ) and neglecting other modes with the initial condition @xmath119
, we get @xmath120 , \label{eq : final_dd}\end{aligned}\ ] ] where the sum in the second equation is taken for degree @xmath10 .
note that @xmath121 holds , where we have defined @xmath122 substituting this equation into equation ( [ eq : final_dd ] ) and taking the limit @xmath100 , we obtain @xmath123 this is the equation for @xmath124 in hmf derived in @xcite . 50 peiris , j. s. m. _ et al . _ .
17671772 ( 2003 ) .
fraser , c. _ et al .
_ pandemic potential of a strain of influenza a ( h1n1 ) : early findings .
15571561 ( 2009 ) .
gire , s. k. _ et al .
_ genomic surveillance elucidates ebola virus origin and transmission during the 2014 outbreak .
13691372 ( 2014 ) .
eubank , s. _ et al .
_ modelling disease outbreaks in realistic urban social networks . 180184 ( 2004 ) .
saito , m. m. _ et al .
_ enhancement of collective immunity in tokyo metropolitan area by selective vaccination against an emerging influenza pandemic .
e72866 ( 2013 ) .
ferguson , n. m. _ et al .
_ strategies for containing an emerging influenza pandemic in southeast asia
. 209214 ( 2005 ) .
ajelli , m. & merler , s. an individual - based model of hepatitis a transmission . 478488 ( 2009 ) .
merler , s. & ajelli , m. the role of population heterogeneity and human mobility in the spread of pandemic influenza . 557565 ( 2010 ) .
ferguson , n. m. _ et al .
_ strategies for mitigating an influenza pandemic .
448452 ( 2006 ) .
halloran , m. e. _ et al .
_ modeling targeted layered containment of an influenza pandemic in the united states .
46394644 ( 2008 ) .
wang , b. , cao , l. , suzuki , h. & aihara , k. epidemic spread in adaptive networks with multitype agents .
035101 ( 2011 ) .
colizza , v. , pastor - satorras , r. , & vespignani , a. reaction diffusion processes and metapopulation models in heterogeneous networks . , * 3 , * 276282 ( 2007 ) .
balcan , d. & vespignani , a. phase transitions in contagion processes mediated by recurrent mobility patterns . 581586 ( 2011 ) .
poletto , c. , tizzoni , m. , & colizza , v. heterogeneous length of stay of hosts movements and spatial epidemic spread . 476
( 2012 ) .
colizza , v. , barrat , a. , barthlemy , m. & vespignani , a. the role of the airline transportation network in the prediction and predictability of global epidemics .
20152020 ( 2006 ) .
colizza , v. , barrat , a. , barthelemy , m. , valleron , a .-
j . & vespignani .
a. modeling the worldwide spread of pandemic influenza : baseline case and containment interventions . e13 ( 2007 ) .
balcan , d. _ et al . _ multiscale mobility networks and the spatial spreading of infectious diseases .
2148421489 ( 2009 ) .
sattenspiel , l. & dietz , k. a structured epidemic model incorporating geographic mobility among regions . 7191 ( 1995 ) .
, a. epidemic outbreaks on structured populations .
125129 ( 2007 ) .
watts , d. j. , muhamad , r. , medina , d. c. & dodds , p. s. multiscale , resurgent epidemics in a hierarchical metapopulation model . 1115711162 ( 2005 ) .
rohani , p. , earn , d. j. d. & grenfell , b. t. opposite patterns of synchrony in sympatric disease metapopulations .
968971 ( 1999 ) .
rvachev , l. a. & longini jr , i. m. a mathematical model for the global spread of influenza . 322 ( 1985 ) .
riley , s. large - scale spatial - transmission models of infectious disease .
12981301 ( 2007 ) .
grais , r. f. , ellis , j. h. & glass , g. e. assessing the impact of airline travel on the geographic spread of pandemic influenza . 10651072 ( 2003 ) .
tellier , r. aerosol transmission of influenza a virus : a review of new studies .
s783s790 ( 2009 ) .
onnela , j. p. _ et al .
_ 73327336 ( 2007 ) .
cattuto , c. _ et al . _ e11596 ( 2010 ) .
sekimoto , y. , shibasaki , r. , kanasugi , h. , usui , t. & shimazaki , y. pflow : reconstructing people flow recycling large - scale social survey data . 2735 (
2011 ) .
masuda , n. & holme , p. predicting and controlling infectious disease epidemics using temporal networks . 6 ( 2013 ) .
danon , l. _ et al . _ networks and the epidemiology of infectious disease . 284909 ( 2011 ) .
pastor - satorras , r. , castellano , c. , van mieghem , p. & vespignani , a. epidemic processes in complex networks . 925979 ( 2015 ) .
kitsak , m. _ et al . _ .
888893 ( 2010 ) .
brockmann , d. & helbing , d. 13371342 ( 2013 ) .
salath , m. _ et al .
_ a high - resolution human contact network for infectious disease transmission .
2202022025 ( 2010 ) .
kitchovitch , s. & li , p. risk perception and disease spread on social networks . 23452354 ( 2010 ) .
pastor - satorras , r. & vespignani , a. epidemic spreading in scale - free networks .
32003203 ( 2001 ) .
moreno , y. , pastor - satorras , r. & vespignani , a. epidemic outbreaks in complex heterogeneous networks .
521529 ( 2002 ) .
bogun , m. , pastor - satorras , r. & vespignani , a. epidemic spreading in complex networks with degree correlations . in _ statistical mechanics of complex networks notes in physics , springer ,
berlin _ , 2003 .
yang , r. _ et al .
_ epidemic spreading on heterogeneous networks with identical infectivity .
189193 ( 2007 ) . keeling , m. j. the effects of local spatial structure on epidemiological invasions .
859867 ( 1999 ) .
eames , k. t. d. & keeling , m. j. modeling dynamic and network heterogeneities in the spread of sexually transmitted diseases .
1333013335 ( 2002 ) .
boots , m. & sasaki , a. parasite - driven extinction in spatially explicit host - parasite systems . 706713 ( 2002 ) .
youssef , m. & scoglio , c. an individual - based approach to sir epidemics in contact networks .
136 144 ( 2011 ) .
barrat , a. , barthlemy , m. & vespignani , a. ( cambridge univ .
press , 2008 ) .
stehl , j. , _
et al . _
simulation of an seir infectious disease model on the dynamic contact network of conference attendees .
87 ( 2011 ) .
newman , m. ( oxford univ .
press , 2010 ) .
gleeson , j. p. , melnik , s. , ward , j. a. , porter , m. a. & mucha , p. j. accuracy of mean - field theory for dynamics on real - world networks .
026106 ( 2012 ) .
wang , y. , chakrabarti , d. , wang , c. & faloutsos , c. epidemic spreading in real networks : an eigenvalue viewpoint . in _ proc .
22nd international symposium on reliable distributed systems _ 2534 ( 2003 ) .
barthlemy , m. spatial networks
. 1 101 ( 2011 ) .
n.f . is grateful for the stimulating discussions with t. takaguchi , t. aoki , y. sughiyama , and y. yazaki .
this research is supported by the aihara project , the first program from jsps , initiated by cstp , and by crest , jst .
this research is the result of the joint research with center for spatial information science , the university of tokyo ( no . 315 ) .
is supported by jsps kakenhi grant number 15k16061 .
all authors designed the research .
n.f . and a.r.s .
carried out the numerical simulations .
n.f . , a.r.s , and k.i .
analysed the results .
, k.i . , and k.a .
contributed to the analytical calculations .
n.f . and a.r.s
wrote the bulk of the manuscript .
k.i . and k.a .
edited the manuscript .
the authors declare no competing financial interests .
the data size of the people flow data of @xmath125 is much smaller than the population of the corresponding tokyo metropolitan area , @xmath7 .
therefore , it is natural to ask whether it makes sense to consider a model with a small number of agents . a clue to answering this question is to use scaling .
we found that this system exhibits a percolation - like transition : the order parameter of the giant cluster component size ratio @xmath126 has a second - order phase transition ( fig .
[ fig : gcc ] left ) .
although it is unclear whether the percolation - like transition takes place for large @xmath0 ( because of the spatial resolution of the data for privacy reasons ) , we found that @xmath126 obeys the scaling law ( fig .
[ fig : gcc ] right ) .
far from the percolation transition point , @xmath127 holds , where @xmath128 is a scaling function .
if we assume that this scaling holds for the size of the real population , the parameter values @xmath4 , @xmath129 m used in the simulations in the main text correspond to @xmath130 m in the real population .
scaling of the giant cluster component size ratio for different @xmath0 .
the right - hand figure shows that this is a function of the scaled parameter @xmath131 .
, scaledwidth=95.0% ] scaling of the giant cluster component size ratio for different @xmath0 .
the right - hand figure shows that this is a function of the scaled parameter @xmath131 .
, scaledwidth=95.0% ]
we discuss an approximation to replace the time - dependent adjacency matrix with the time - averaged one , and determine the condition under which this approximation is valid .
let us consider a general example where a dynamical variable @xmath132 evolves via contact with other nodes as @xmath133 where the parameter @xmath13 represents the coupling strength and @xmath2 is the instantaneous adjacency matrix of a temporal network at time @xmath16 .
equation ( [ eq : matrix_ode ] ) can be solved as @xmath134 \bm x(0 ) , \label{eq : tprod}\end{aligned}\ ] ] where @xmath135 denotes the time - ordered exponential .
if @xmath2 changes with the time interval @xmath136 , equation ( [ eq : tprod ] ) can be rewritten as @xmath137 where @xmath138 is the time - averaged adjacency matrix .
the above equation can be interpreted as follows .
the terms of @xmath139 in the power series expansion in equation ( [ eq : tprod ] ) consist of time - ordered products of @xmath140 adjacency matrices of the temporal network .
these terms collect paths of length @xmath140 , namely all possible indirect interactions occurring @xmath140 times between two nodes at different times taking into account the time order of the interactions .
if we consider the case of a small value of @xmath13 , the time evolution can be approximated up to the first order .
this corresponds to assuming that the interaction only takes place by direct contact between agents , and all multi - step interactions during the time interval @xmath141 are negligible . in such a case ,
the time - averaged network is sufficient to describe the dynamical processes in the temporal network .
note that the time - order information from one day does not appear in the first - order terms .
one can derive the mode coupling equation from the following differential equations ( 1 ) in the main text : @xmath142 let us substitute equation ( 2 ) in the main text , @xmath28 into this equation .
taking into account the eigenvalue equation for the eigenvector @xmath143 , @xmath144 it is easy to verify that @xmath145 & = - \beta \bigg ( \sum_{b=1}^n \hat s_b \phi^{(b)}_j \bigg ) \bigg [ \sum_{k=1}^n \overline{a}_{jk}(t ) \bigg ( \sum_{c=1}^n \hat i_c \phi^{(c)}_k \bigg ) \bigg]\\ & = - \beta \bigg ( \sum_{b=1}^n \sum_{c=1}^n \lambda_c \phi^{(b)}_j \phi^{(c)}_j \hat s_b \hat i_c \bigg),\\ % \frac { d } { dt } \bigg[\sum_{a'=1}^n \hat i_{a ' } \phi^{(a')}_j \bigg ] & = \beta \bigg ( \sum_{b=1}^n \hat s_b \phi^{(b)}_j \bigg ) \bigg [ \sum_{k=1}^n \overline{a}_{jk}(t ) \bigg ( \sum_{c=1}^n \hat i_c \phi^{(c)}_k \bigg ) \bigg ] -\mu \sum_{a'=1}^n \hat i_{a ' } \phi_j^{(a ' ) } \\ & = \beta \bigg ( \sum_{b=1}^n \sum_{c=1}^n \lambda_c \phi^{(b)}_j \phi^{(c)}_j \hat s_b \hat i_c \bigg ) -\mu \sum_{a'=1}^n \hat i_{a ' } \phi_j^{(a ' ) } , \\
% \frac{d } { dt } \bigg[\sum_{a'=1}^n \hat r_{a ' } \phi^{(a')}_j \bigg ] & = \mu \sum_{a'=1}^n \hat i_{a ' } \phi_j^{(a ' ) } .
\label{eq : mf2 } \end{split}\end{aligned}\ ] ] multiplying @xmath143 on both sides of the above equations and taking the sum over @xmath15 to use the orthonormal condition @xmath146 where @xmath147 is kronecker s delta , we obtain @xmath148 where @xmath149 . note that each element of @xmath150 consists of the product of the eigenvector components and is time - independent .
it is easy to extend this equation to the case where the static adjacency matrix @xmath23 is asymmetric .
let us define the left and right eigenvectors of @xmath23 for the eigenvalue @xmath45 as @xmath151 and @xmath143 , and set the normalisation condition @xmath152 .
then , equation ( [ eq : mf3 ] ) holds if we define @xmath153 . by considering @xmath70 ( @xmath106 ) modes and neglecting the other modes in equation ( [ eq : mf3 ] )
, we obtain the transient dynamics of the epidemic from a small number of modes .
equations ( [ eq : mf3 ] ) contain mode coupling terms , but the modes are decoupled in the linearised equation .
let us take @xmath154 and focus on the initial stage of the epidemic spreading .
assume that @xmath155 and @xmath18 are small , and approximate @xmath156 by neglecting the product @xmath157 .
since @xmath24 appears only as the product @xmath158 , the linearised approximation allows us to put @xmath159 . under this assumption
, we can take @xmath160 substituting this into the evolution equation for @xmath161 in equation ( [ eq : mf3 ] ) , the linearised equation becomes @xmath162 by taking the summation over @xmath163 using equation ( [ eq : approx_mode ] ) , we obtain @xmath164 in this way , we derive equation ( 3 ) in the main text .
we now discuss how to estimate the infection probability of a test agent .
let @xmath165 be the position of agent @xmath15 at time @xmath16 .
since the infection probability depends on the distance between agents , we introduce the kernel function @xmath82 . in this study
, we used the heaviside step function @xmath166 , where @xmath167 for @xmath168 and @xmath169 otherwise .
then the element of the adjacency matrix @xmath170 is given as @xmath171 .
we want to estimate the risk of infection of a test agent @xmath175 who travels inside the urban area with an arbitrary trajectory and returns to the original location after time @xmath176 .
equation ( 1 ) in the main text can be solved as @xmath98 where @xmath177 . \end{aligned}\ ] ] for @xmath100 , @xmath101 should vanish , and the conservation of probability condition is given as @xmath178 . therefore , the infection probability of this agent is @xmath179 \right\}. \label{eq : final}\end{aligned}\ ] ] this equation is transcendental and can not be solved analytically . however ,
if @xmath0 is sufficiently large and the contribution of the test agent to @xmath84 is negligible , we can approximate the infection probability of the test agent @xmath175 using @xmath84 .
since the time - averaged adjacency matrix is given in terms of the risk factor as @xmath180 we obtain @xmath181 , \label{eq : final_add}\end{aligned}\ ] ] by taking @xmath182 and @xmath183 .
therefore , the final infection probability of a susceptible agent traveling along an arbitrary trajectory is given by the time integral of the risk factor along its trajectory .
if the mode truncation with @xmath70 ( @xmath106 ) modes @xmath184 approximates @xmath19 , the risk factor is given by the superposition of @xmath70 independent variables as @xmath185 the infection probability is approximated with the mode truncation @xmath186 , \end{aligned}\ ] ] by replacing @xmath187 with @xmath188 in equation ( [ eq : final_add ] ) . | recently developed techniques to acquire high - quality human mobility data allow large - scale simulations of the spread of infectious diseases with high spatial and temporal resolution .
analysis of such data has revealed the oversimplification of existing theoretical frameworks to infer the final epidemic size or influential nodes from the network topology .
here we propose a spectral decomposition - based framework for the quantitative analysis of epidemic processes on realistic networks of human proximity derived from urban mobility data .
common wisdom suggests that modes with larger eigenvalues contribute more to the epidemic dynamics .
however , we show that hidden dominant structures , namely modes with smaller eigenvalues but a greater contribution to the epidemic dynamics , exist in the proximity network .
this framework provides a basic understanding of the relationship between urban human motion and epidemic dynamics , and will contribute to strategic mitigation policy decisions .
epidemics of infectious diseases in the human population , e.g. the sars outbreak in 20022003 @xcite , 2009 h1n1 influenza pandemic @xcite , and ebola outbreak @xcite , can be a serious factor in human mortality , and have a significant socio - economic impact in terms of reducing a population s healthy years because of morbidity . in recent years
, the mathematical modelling of the outbreak and spread of infectious diseases has mainly been performed using two approaches : agent - based models @xcite and structured metapopulation models @xcite .
both techniques incorporate real or synthetic data on the long - range mobility and migration of populations , but vary in their granularity of the population , from individuals to sub - groups of society .
various spatial scales have been studied , ranging from specific geographic locales such as cities @xcite to nations @xcite and intra-/inter - continental regions @xcite .
in particular , it is imperative to study the mechanisms of an epidemic s spread in urban scenarios , considering restrictions on available space and transportation methods ( e.g. traffic fluxes , travel routes , timescales of modes of transport ) , which result in high density and heterogeneous contact patterns among residents .
these factors make the populace vulnerable to epidemics of diseases that spread through exhaled aerosols , e.g. @xcite .
to understand the details of epidemic dynamics within cities , it is crucial to exploit the heterogeneous contact patterns derived from human mobility data .
fortunately , recent technical advances in data acquisition methods enable us to incorporate human mobility data obtained from various sources @xcite into mathematical models @xcite .
network epidemiology @xcite is a useful means of uncovering the dependence of the contagion dynamics on the heterogeneous contact pattern .
the approach of tracking connections between people provides a comprehensive framework with which to study the effects of the underlying network of people in the region of interest ( e.g. cities ) .
the present understanding of epidemic or social contagion processes has been bolstered by many network - based studies @xcite , which treat cities as social petri - dishes @xcite .
analytical approaches such as heterogeneous mean field ( hmf ) theory @xcite and moment closure approximation @xcite provide an important bridge between the network topology and the spread of epidemics in complex networks .
for example , the epidemic threshold is absent in scale - free networks with no degree correlation @xcite
. however , certain simplifications made to obtain analytical results , e.g. neglecting the degree correlation in the hmf @xcite , could be superfluous , and these approximations may not necessarily hold in realistic networks . therefore , a novel analytical approach is required to better understand the spread of epidemics in heterogeneous populations . |
the milky way contains about 150 globular clusters ( gcs ) with masses of @xmath6 to over @xmath7 @xmath8 and tens of thousands of open clusters containing from 100 to over @xmath6 stars . however , young massive clusters containing @xmath9 of stars are rare , with only a handful known @xcite .
while no gcs have formed in the milky way within the last 5 gyr , open clusters that survive many crossing times continue to form .
a few of these clusters have stellar masses greater than @xmath6 m@xmath10 and therefore qualify as young massive clusters ( ymcs ; * ? ? ?
ymcs must either form from clumps having masses greater than and sizes comparable to the final cluster or be formed from a larger , more diffuse reservoir , in which case massive protocluster clumps may be rare or nonexistent @xcite .
massive proto - clusters ( mpcs ) are massive clusters ( m@xmath11 @xmath8 ) in the process of forming from a dense gas cloud . in @xcite , we examine the theoretical properties of mpcs : mpcs are assumed to form from massive , cold starless clumps analagous to pre - stellar cores @xcite . in this paper
, we refer to two classes of objects : starless mpcs , which have very low luminosity and do not contain ob stars , and mpcs , which are gas - rich but have already formed ob stars .
the only currently known starless mpc is g0.253 + 0.016 , which lies within the dense central molecular zone and is subject to greater environmental stresses than similar objects in the galactic plane @xcite .
because massive clusters contain many massive stars , at some point during their evolution ionization pressure will prevail over protostellar outflows as the dominant feedback mechanism .
other sources of feedback are less than ionization pressure up until the first supernova explosion @xcite .
these proto - clusters must have masses m@xmath12 , . ] or about @xmath13 @xmath8 for an assumed sfe=30% ( an upper limit on the star formation efficiency ) , confined in a radius @xmath3 pc , in order to remain bound against ionization feedback .
these properties motivate our search for proto - clusters in the bolocam galactic plane survey ( bgps ; * ? ? ?
* http://irsa.ipac.caltech.edu/data/bolocam_gps/ ) .
the distinction between relatively short - lived ` open clusters ' and long - lived ( @xmath14 gyr ) bound clusters occurs at about @xmath6 @xmath8@xcite .
clusters with @xmath15 will be destroyed by interactions with giant molecular clouds over the course of a few hundred million years after they have dispersed their gas @xcite , while clusters with @xmath16 may survive @xmath17 gyr .
closer to the galactic center , within approximately a kiloparsec , all clusters will be destroyed on shorter timescales by strong tidal forces or interactions with molecular clouds . in the galaxy , there are few known massive clusters .
@xcite catalogs a few of them , of which ngc 3603 , trumpler 14 , and westerlund 1 and 2 are the likely descendants of the objects we investigate .
these clusters have @xmath18 pc , @xmath19 @xmath8 , and ages @xmath20 myr .
we present a census of their ancestral analogs .
the bgps is a 1.1 mm survey of the first quadrant of the galactic plane in the range @xmath21 with resolution @xmath22 sensitive to a maximum spatial scale of @xmath23 @xcite .
the bgps ` bolocat ' v1.0 catalog includes sources identified by a watershed decomposition algorithm and flux measurements within apertures of radius 20 , 40 , and 60@xcite .
we searched the bgps for candidate mpcs in the 1st quadrant ( @xmath24 ; 5991 sources ) .
the inner 6 degrees of the galaxy are excluded because physical conditions are significantly different from those in the rest of the galaxy @xcite and the bgps is confusion - limited in that region .
we identify a flux - limited sample by setting our search criteria to include all sources with @xmath25 @xmath8 in a 20 radius out to 26 kpc , or a physical radius of 2.5 pc at that distance .
the radius cutoff is motivated by completeness and physical considerations : the cutoff of 26 kpc includes the entire star forming disk in our targeted longitudes , and @xmath26 pc corresponds to the radius at which a @xmath13 @xmath8 mass has an escape speed @xmath27 @xmath28 , i.e. ionized gas will be bound .
the maximum radius and minimum mass imply a minimum mean density @xmath29 , which implies a maximum free - fall time @xmath30 myr . using the bolocat v1.0 catalog
, we first set a flux limit on the sample by assuming the maximum distance of @xmath31 kpc and imposing a mass cutoff of @xmath32 @xmath8 inside a 20 ( 2.5 pc ) radius aperture .
following equation 19 in @xcite : @xmath33 and assuming @xmath34k , the implied flux cutoff is 1.13 jy , above which 456 ` flux - cutoff ' candidates were selected in the bolocat v1.0 catalog .
cutoffs of 4.3 jy for the 40 and 10.2 jy for the 60 bolocat v1.0 apertures were used to select more nearby candidates inside the same physical radius , but no sources were selected based on these larger apertures . the bgps is insensitive to scales larger than 120@xcite . as a result ,
the survey is incomplete below a distance @xmath35 from the sun . within this radius
, alternate methods must be sought to determine the total mass within @xmath36 pc .
although the sample is incomplete for @xmath37 kpc , sources that have sufficient mass despite the 120 spatial filtering are included .
distances to bgps - selected candidates were determined primarily via literature search . where distances were unavailable , we used velocity measurements from @xcite and assumed the far distance for source selection .
we then resolved the kinematic distance ambiguity towards these sources by searching for associated near - infrared stellar extinction features from the ukidss gps @xcite .
most literature distances were determined using a rotation curve model and some method of kinematic distance ambiguity resolution .
because the literature used different rotation curve models , there is a @xmath38 systematic error in distance resulting in a @xmath39 systematic error in mass .
we used the larger 40 radius apertures to determine the flux for sources at @xmath40 kpc and 60 radius apertures for sources at @xmath41 kpc ( corresponding to @xmath42 pc ) .
the masses were computed assuming a temperature @xmath34k , opacity @xmath43 , and gas - to - dust ratio of 100 @xcite.k is more appropriate for a typical pre - star - forming clump than an evolved hii - region hosting one ( * ? ? ?
however , because we are interested in cold progenitors as well as actively forming clusters , the selection is based on @xmath34k , which is more inclusive . ]
the mass estimate drops by a factor of @xmath44 if the temperature assumed is doubled to @xmath45k .
@xcite notes that significant free - free contamination , as high as 80% , is possible for some 1.1 mm sources , so the selected candidates may prove to be more moderate - mass and evolved proto - clusters .
we used the nrao vla archive survey ( nvas ; * ? ? ? * ) to estimate the free - free contamination for the sample . for most sources ,
the free - free contamination inferred from the vla observations is small ( @xmath46 ) , but for a subset the contamination was @xmath47 assuming that the free - free emission is optically thin .
corrected masses using the measured free - free contamination and higher dust temperatures are listed in table [ tab : candidates ] ; these are reasonable lower limits on the total mass of these regions .
all of the contamination estimates are technically lower limits both because of the assumption that the free - free emission is optically thin and because the vla filters out large - scale flux .
however , in most cases , the emission is likely to be dominated by optically thin emission ( evolved hii regions tend to be optically thin and bright , while compact hii regions are optically thick but relatively faint ; * ? ? ?
* ) and for most sources vla c or d - array observations were used , and at 3.6 and 6 cm the largest angular scale recovered is 180 - 300 , greater than the largest angular scale in the bgps . applying a cutoff of m@xmath48 @xmath8 left 18protocluster candidates out of the original 456 .
the more stringent cut m@xmath49 @xmath8 leaves only 3 mpcs .
the final candidate list contains only sources with @xmath50 ( the completeness limit ; see table [ tab : candidates ] ) .
the table lists their physical properties , their literature distance , their mass ( assuming @xmath51 and a free - free subtracted lower - limit ) , and their inferred escape speed ( @xmath52 ) assuming a radius equal to the aperture size at that distance .
the table also includes measurements of the iras luminosity in the 60 and 100 @xmath53 m bands within the source aperture .
these 18 candidates include some overlapping sources .
there are two clumps in w51 separated by about 1.5 pc and 4.5 @xmath28 along the line of sight that are each independently massive enough to be classified as mpcs , but are only discussed as a single entity because they are likely to merge if their three - dimensional separation is similar to their projected distance .
the candidates in w49 are more widely separated , about 4.4 pc and 7 @xmath28 along the line of sight , but could still merge . additionally , 9 of the 18 are within 8.7 kpc , so the mass estimates are lower limits .
these are promising candidates for follow - up , but can not be considered complete for population studies . if our radius restriction is dropped to 1.5 pc , the minimum complete distance drops to 5.2 kpc and the three lowest - mass sources in table [ tab : candidates ] no longer qualify , but otherwise the source list remains unchanged .
our analysis is therefore robust to the selection criteria used .
the massive clumps in table [ tab : candidates ] can be used to constrain the galactic formation rate of massive clusters ( mcs ) above @xmath54 if we assume that the number of observed proto - clusters is a representative sample .
the region surveyed covers a fraction of the surface area of the galaxy @xmath55 assuming the star forming disk has a radius of 15 kpc . ] .
the cumulative cluster formation rate above a cluster mass @xmath56 is given by @xmath57 where @xmath58 myr is the assumed cluster formation timescale .
, the time from the start of star formation until gas expulsion , is a poorly understood quantity , but is reasonably constrained to be @xmath59 myr from the age spread in the orion nebula cluster @xcite and @xmath60 myr because the most massive stars will go supernova by that time . ] with the measured @xmath61 proto - clusters , we infer a galactic formation rate @xmath62 this cluster formation rate is statistically weak , with poisson error of about 3.5 @xmath63 and can be improved with more complete surveys ( e.g. , hi - gal , * ? ? ? * ) .
this formation rate is an upper limit because all of the estimated masses are upper limits as discussed in section [ sec : selection ] .
we use cluster observations in m31 from @xcite to infer the massive cluster formation rate in m31 .
they observe 2 clusters with @xmath64 and ages @xmath65 myr in 15% of the m31 star - forming disk .
the implied cluster formation rate in andromeda is @xmath66 @xmath63 .
given m31 s total star formation rate @xmath67 lower than the galactic rate ( andromeda @xmath68 , milky way @xmath69 @xmath8@xmath63 ; * ? ? ?
* ; * ? ? ?
* ) , the predicted galactic cluster formation rate is @xmath70 @xmath63(assuming the cfr scales linearly with the sfr ; * ? ? ?
the scaled - up andromeda cluster formation rate matches the observed galactic cluster formation rate .
the samples are small , but as a sanity check , the agreement is comforting . in the sample of potential proto - clusters , all have formed massive stars based on a literature search and iras measurements .
a few of the low mass sources , g012.209 - 00.104 , g012.627 - 00.016 , g019.474 + 00.171 , and g031.414 + 00.307 have relatively low iras luminosities ( @xmath71 ) and little free - free emission .
however , _ all _ are detected in the radio as h ii regions ( some ultracompact ) and have luminosities indicating early - b type powering stars .
non - detection of ` starless ' proto - cluster clumps implies an upper limit on the starless lifetime . for an assumed @xmath72 myr , the @xmath73 upper limit on the starless proto - mc clump is @xmath74 assuming poisson statistics and using all 18 sources .
this limit is consistent with massive star formation on the clump free - fall timescale ( @xmath75 myr ) .
it implies that massive stars form rapidly when these large masses are condensed into cluster - scale regions and hints that massive stars are among the first to form in massive clusters .
assuming a lower limit 30% sfe and t@xmath76 , 3 candidates in table [ tab : candidates ] will become massive clusters like ngc 3603 : g010.472 + 00.026 , w51 , and w49 ( g043.169 + 00.01 ) .
even if t@xmath77 , w49 is still likely to form a @xmath78 @xmath8 mc , although g10.47 would be too small .
w51 , which is within the spatial - filtering incompleteness zone , passes the cutoff and is likely to form a pair of massive clusters .
however , if the dust in w51 is warm and the free - free contamination is considered , the total mass in each of the w51 clumps is below the 3 @xmath8 cutoff .
the bgps covers about 30% of the galactic star - forming disk in the range 1 kpc @xmath79 kpc .
we can extrapolate our 3 detections to predict that there are @xmath8010 ( @xmath81 ) proto - clusters in the galaxy outside of the galactic center .
the agreement between the sfr - based prediction from m31 and our observations implies that we have selected genuine massive proto - clusters ( mpcs ) .
these most massive sources have escape speeds greater than the sound speed in ionized gas , indicating that they can continue to accrete gas even after the formation of massive stars .
assuming they are embedded in larger - scale gas reservoirs , we are measuring lower bounds on the ` final ' clump plus cluster mass .
all of the young massive proto - clusters candidates observed are within the solar circle despite our survey covering more area outside of the solar circle .
the outer radius limit for massive cluster formation is consistent with the observed metallicity shift noted at the same radius by @xcite .
they identify the solar circle as the corotation radius of pattern speed and orbits within the galaxy ( within this radius , stars orbit faster than the spiral pattern ) .
the fact that this radius also represents a cutoff between the inner , massive - cluster - forming disk and the outer , massive - cluster - free disk hints that gas crossing spiral arms may be the triggering mechanism for massive cluster formation .
however , given the small numbers , the detected clusters are consistent with a gaussian + exponential disk distribution following that described by @xcite .
future work should include a census for mpcs within @xmath82 kpc using the herschel hi - gal survey @xcite and in the southern plane with atlasgal @xcite .
some surveys have already identified proto - clusters in these regions ( e.g. * ? ? ?
* ; * ? ? ?
* ) , but they are not complete
. a complete survey of distances will be essential for continuum surveys to be used .
there are two modes of massive cluster formation consistent with our observations that can be observationally distinguished .
either a compact starless massive proto - cluster phase does occur and is short , or the mass to be included in the cluster is accumulated from larger volumes over longer timescales . extending our proto - cluster survey to the southern sky , e.g. using the atlasgal and hi - gal surveys ,
will either discover starless mpcs or strengthen the arguments that there is no starless mpc phase . if instead massive clusters form by large scale ( @xmath83 pc ) accretion , substantial reservoirs of gas should surround these most massive regions and be flowing into them .
signatures of this accretion process should be visible : mpcs should contain molecular filamentary structures feeding into their centers ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . alternatively , lower mass clumps may merge to form massive clusters @xcite , in which case clusters of clumps - which should be detectable in extant galactic plane surveys - are the likely precursors to massive clusters .
finally , massive clusters may form from the global collapse of structures on scales larger than we have probed , which could also produce clusters of clumps .
using the bgps , we have performed the first flux - limited census of massive proto - cluster candidates .
we found 18 candidates that will be part of the next generation of open clusters and 3 that could form massive clusters similar to ngc 3603 ( @xmath84 @xmath8 ) . we have measured a galactic massive cluster formation rate @xmath85 assuming that clusters are equally likely to form everywhere within the range 1 kpc @xmath86 15 kpc .
the observed mpc counts are consistent with observed cluster counts in andromeda scaled up by @xmath87 assuming a formation timescale of 2 myr . despite this survey being
the first sensitive to pre - star - forming mpc clumps , none were detected .
this lack of detected pre - star - forming mpcs suggests a timescale upper limit of about @xmath5 myr for the pre - massive - star phase of massive cluster formation , and hints that massive clusters may never form highly condensed clumps ( @xmath88 ) prior to forming massive stars .
it leaves open the possibility that massive clusters form from large - scale ( @xmath89 pc ) accretion onto smaller clumps over a prolonged ( @xmath90 myr ) star formation timescale .
observations are needed to distinguish competing models for mc formation : birth from isolated massive proto - cluster clumps , either compact and rapid or diffuse and slow , or from smaller clumps that never have a mass as large as the final cluster mass .
this sample of the 18 most massive proto - cluster clumps in the first quadrant ( where they can be observed by both the vla and alma ) presents an ideal starting point for these observations .
we thank the referee for thorough and very helpful comments that strengthened this letter .
this work was supported by nsf grant ast 1009847 . ,
j. h. , sjouwerman , l. o. , fomalont , e. b. , & radziwill , n. m. 2008 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
7016 , society of photo - optical instrumentation engineers ( spie ) conference series 1 : @xcite , 2 : @xcite , 3 : @xcite , 4 : @xcite , 5 : @xcite , 7 : @xcite , 8 : @xcite , 9 : @xcite , 10 : @xcite , 11 : @xcite , 12 : @xcite .
6 : the distances to g030.704 was determined using the near kinematic distance from the velocity of the hht - observed hco+ line @xcite .
@xmath91 : the minimum likely mass , @xmath92 .
@xmath93 : the fraction of flux from free - free emission ( as opposed to dust emission ) at @xmath0 mm | we search the @xmath0 mm bolocam galactic plane survey for clumps containing sufficient mass to form @xmath1 star clusters .
18 candidate massive proto - clusters are identified in the first galactic quadrant outside of the central kiloparsec .
this sample is complete to clumps with mass m@xmath2 and radius @xmath3 pc .
the overall galactic massive cluster formation rate is @xmath4 , which is in agreement with the rates inferred from galactic open clusters and m31 massive clusters .
we find that all massive proto - clusters in the first quadrant are actively forming massive stars and place an upper limit of @xmath5 myr on the lifetime of the starless phase of massive cluster formation .
if massive clusters go through a starless phase with all of their mass in a single clump , the lifetime of this phase is very short . |
scanning tunneling microscopy and spectroscopy ( stm / sts ) are extensively used to probe the quasiparticle spectra of high temperature superconductors , especially of the cuprate materials such as bi@xmath0sr@xmath0cacu@xmath0o@xmath1 ( bi2212)@xcite .
these studies have contributed immensely in obtaining insight into superconducting and pseudogap phases of these materials owing to the very high spatial and energy resolution of the spectroscopy .
the full experimental potential of stm / sts is , however , far from exploited , since the chemical and spatial resolution as well as the directional selectivity could , in principle , be improved by overcoming various challenges related to intrinsic properties of stm . the first challenge is related to the fact that the signal from superconducting layers is filtered by insulating oxide layers .
in addition , the signal does not arrive at the microscope tip directly from the site below the tip but rather through symmetry driven tunneling channels .
hence , experimentally observed spectral features are not only distorted by filtering , but ascertaining their spatial origin becomes harder the deeper within the sample is their source .
another factor to be aware of is that the electronic and geometrical properties of the microscope tip may also affect stm images . in standard interpretation
, the tip is assumed symmetric to rotations around the vertical axis ( ` s - wave ' ) .
the relative bluntness of the tip suggests this assumption and , furthermore , the metallic nature of common tip materials does not favor oriented `` dangling '' orbitals , and specific effects of tip geometry and electronic structure are usually ignored .
however , this interpretation has been challenged by a variety of sophisticated theoretical studies @xcite , which , on the one hand show that the closer the tip is to the surface , the more important the symmetry of tip orbitals is to the stm image and , on the other hand , they indicate that s - wave tips might not be able to produce images with atomic resolution .
notably , the relative importance of tip orbitals with different symmetries in the stm signal can be controlled by varying the tip - surface distance , or alternatively , by functionalizing the tip by attaching particular atoms / molecules at the apex of the tip . in various experimental studies@xcite ,
the tip properties have been deliberately modified for better spatial or chemical selectivity by picking up a molecule , e.g. co , o@xmath0 or n@xmath0 , from the surface to the apex of the tip after which the plain tip results have been compared to the results with a functionalized tip .
this has proven to improve chemical selectivity in adsorbate on a solid surface system since each surface species responds in an individual way to the change of tip symmetry due to functionalization . even in the case of relatively complicated molecules , such as pentacene attached to the tip
, the contrast of differential conductance maps revealing the geometric shape of molecular orbitals is surprisingly sharpened @xcite by the enhanced directionality of the probe - sample overlap due to the dominance of the @xmath2 or @xmath3 type localized orbitals closest to the sample surface .
the idea of utilizing a functionalized tip is further encouraged by the fact that the resolution of atomic force microscopy ( afm ) studies is enhanced by using co molecule attached to the metallic tip @xcite , although improving the resolution and selectivity in the case of afm has a different physical basis from stm .
the preceding work on molecular adsorbate systems provides motivation for considering functionalized tips for investigating cuprate materials , where the interesting physics is found at layers not immediately on the surface .
we can foresee many uses of functionalized tips in probing these materials .
an obvious application would be a more accurate probe for the surface properties . in the case of bi2212 ,
to date one has mostly observed the bi atoms of the surface layer in topographic stm images .
information from the oxygen atoms of the surface layer has been difficult to obtain , and hence stm provides limited ability to observe the real symmetry of the surface .
a more involved application is related to the di / dv spectra . as shown in our previous studies@xcite ,
the spectra are mappings of the electronic structure of the cuprate layer , filtered by the bio and sro layers , which sit on top of the cuprate layers .
although the signal can be theoretically resolved into partial signals of different origin , the sts experiment measures a signal combined from different sources and , furthermore , the information from different directions in momentum space is integrated into a single outgoing signal .
a possible means of decomposing the sts spectrum experimentally is to measure topographic spectral maps .
in practice , real space and fourier transformed maps differentiating electron and hole states have been produced in order to recognize direction dependent underlying electronic order @xcite
. a third possible application would be investigation of symmetry properties related to nematicity transitions in the subsurface electronic structure at low values of doping , where the oxygen atoms of cuo@xmath4 layers in different directions appear non - equivalent below the pseudogap transition temperature @xmath5 as shown in ref . .
+ in fact , functionalized tips may well have been used inadvertently in studies of bi2212 .
it has been reported that tip resolution can improve dramatically after use , and it has been suggested that this may be due to the attachment of a bi atom to the tunneling tip @xcite . here
we will explore the effects of tips , which are dominated by a single atomic orbital of varying symmetry .
we find that such tips can not only couple with different surface orbitals , but can also affect the interference between tunneling channels from the cuprate layer to the microscope tip .
this suggests possibilities for investigating irregularities within the cuprate layers , which are not directly accessible to the stm tip , and thus reveal features invisible to stm / sts experiments with a symmetric tip . in this way ,
our study advances the understanding of matrix element effects in stm.@xcite this article is organized as follows .
section ii briefly describes the model utilized in this study .
section iii presents calculations of the effect of tip symmetry on topographical maps , stm corrugations and di / dv spectra at various high symmetry sites .
in section iv we show how a hidden irregularity within the cuprate layer can be studied by comparing results obtained with different tip symmetries , among other results .
finally , conclusions are given in section v.
although the model used is essentially the same as in ref . , we provide an overview here for the sake of completeness .
the simulation supercell consists of 8 unit cells that have been rotated @xmath6 degrees as shown in fig .
the unit cell geometry is adapted from ref . .
in total there are @xmath7 atoms and @xmath8 orbitals accounting for both spins .
the topmost 7 layers below the cleavage plane are included , which are in descending order : @xmath9 , @xmath10 , @xmath11 , @xmath12 , @xmath11 , @xmath10 and @xmath9 . depending on the atom type different sets of orbitals
are included . for @xmath13 , @xmath14 and @xmath12
the orbital set is @xmath15 and for @xmath16 the orbital set is @xmath17 .
@xmath18 has a single @xmath19 orbital .
the green s function is computed in a supercell brillouin zone at @xmath20 equally distributed * k*-points corresponding to @xmath21 * k*-points in the unit cell . as shown in fig .
[ topo ] , we define the position of cu ( or surface bi ) atoms as top site ( @xmath22 ) , the position of bonding o atoms of the cuprate layer as bridge site ( @xmath23 ) , and the position of o(bi ) as the hollow site ( @xmath24 ) . in this choice of geometry ,
the @xmath25 and @xmath26 orbitals point along the bond between the neighbouring cu atoms , i.e. , in the direction from the top site to the bridge site ( @xmath27 ) , whereas the linear combinations @xmath28 point from the top site towards the hollow site ( @xmath29 ) .
concerning the oxygen atoms related to the cuprate layer , the _ apical oxygen _ is at the top site , while the _ bonding oxygens _ between cu atoms are at bridge sites . layer , while the right panel depicts the bio and sro layers on top of the cuo@xmath4 layer ; ( b)-(e ) topographic stm images for different tip symmetries .
( b ) @xmath19-orbital , ( c ) @xmath30-orbital , ( d ) @xmath31-orbital , ( e ) @xmath32-orbital .
indicated also are bi and o positions and the top ( t ) , bridge ( b ) and hollow ( h ) sites . in ( d ) and
( e ) the tip symmetry is further indicated with a schematic of the corresponding p - orbital.,scaledwidth=50.0% ] the tip is modeled as a single orbital in order to highlight the effect of each type of tip symmetry on the stm / sts spectrum .
we consider four different tip symmetries which can be divided into two groups .
these consist of @xmath19- and @xmath30-orbitals , which are symmetric to rotation around z - axis , accompanied with two choices of horizontal @xmath33-orbitals .
we use a shorthand notation @xmath32 for a @xmath33-orbital along the bond between two neighbouring cu atoms , and @xmath31 for an orbital rotated by @xmath34 , i.e. , aligned in the direction connecting next nearest neighbouring cu atoms .
these two horizontal tip symmetries can also be connected to the nodal ( @xmath31 ) and antinodal ( @xmath32 ) directions .
the hamiltonian is the same as that used previously in our stm studies of optimally doped b2212@xcite , where superconductivity is modeled with a pairing matrix @xmath35 between electrons and holes with opposite spins : @xmath36\\ & + & \sum_{\alpha \beta \sigma } \left[\delta_{\alpha \beta } c^{\dagger}_{\alpha \sigma } c^{\dagger}_{\beta -\sigma } + \delta_{\beta \alpha}^{\dagger } c_{\beta -\sigma } c_{\alpha \sigma } \right ] \end{array } \label{hamiltonian}\ ] ] with real - space creation ( annihilation operators ) @xmath37 ( @xmath38 ) .
indices are combination of site and spin indices .
the d - wave symmetry of the superconducting gap is modeled by taking @xmath39 to be non - zero only between the neighbouring @xmath40 orbitals , and with opposite signs for @xmath41- and @xmath42-directions , which leads to a d - wave symmetric superconducting gap @xcite .
the tight - binding parameters @xmath43 are taken to be doping - independent in the spirit of the rigid band model.@xcite the tunneling current is computed within the todorov - pendry approximation @xcite as @xmath44 the hopping integrals @xmath45 depend on the symmetry of tip and surface orbitals as well as the distances involved through the slater - koster coefficients@xcite . the tip to surface distance is usually greater than the distance between the atoms in the bulk material . to allow for this the standard slater - koster method
is enhanced by an exponential decay at large distances @xcite .
the slater - koster method allows one to take the directionality of both the tip and surface orbitals into account .
it can be considered as a special case of the more general `` derivative rule '' of chen@xcite and a similar approach of sacks and noguera @xcite who utilize derivatives of the real space tip wave function .
there is an alternative way of writing eq . :
@xmath46 where @xmath47 the difference between the two formulae is that eq .
emphasizes the coupling between the tip and the surface orbitals expressed by the hopping integrals @xmath48 , whereas eq . makes more transparent how tunneling takes place through the filtering layers between the tip and the superconducting layers .
of course the symmetry of the tip itself is described by @xmath48 , but in order to recognize the relevant surface orbitals , it is essential to consider the filtering matrix elements , @xmath49 , as well . the spectral function @xmath50 depends on both regular and anomalous green functions as shown below @xmath51 to conclude , the formulae given above provide not only the total @xmath52 spectrum , but also an explicit form of the filtering effect on the cuprate layer , and decompose the density matrix of the cuprate orbitals into regular and anomalous spectral terms .
the topographic maps of fig . [ topo](b - e ) are calculated using eq . in constant current mode ,
i.e. , there is a feedback loop varying the tip - surface distance in order to retain the tunneling current constant . for bias voltage
we choose @xmath53 we also follow the common practice of accompanying topographic maps with corrugation curves in fig .
[ corrugation ] , since the latter reveal the absolute scale of the contrast otherwise obtained in a rather arbitrary way from topographic maps . ) shown in the insert .
the vertical scale represents the height change the tip experiences along the route .
curves for @xmath31 and @xmath32 are similar in overall shape yet they differ strongly in relative amplitude .
this shows that they connect at least partially to different surface orbitals .
note that each curve was shifted downwards by its own average tip surface distance thus making all curves zero centered.,scaledwidth=40.0% ] experimental topographic maps typically resemble the calculated maps for s - wave tip with a high intensity at the position of the surface bi atoms . within the tersoff - hamann type analysis @xcite
, a high intensity signal is attributed to a high local density of electrons . as fig .
[ ldos](a ) indicates , the surface bi atoms have a large density of states over a wide range of energies around the fermi energy .
this was demonstrated already in our earlier work @xcite where we further related this to the coupling between the @xmath30 orbitals of bi and the @xmath54 orbital of the cu atom below through the intermediate @xmath30 orbital of the apical oxygen .
in contrast , orbitals of o(bi ) contribute to the bio bands , which contain a gap around the fermi - energy . in fig .
[ ldos](a ) these states are broadened since we have invoked a fermi liquid type self - energy in the noncuprate layers @xcite . as seen in figs .
[ topo ] ( b ) and [ topo ] ( c ) , tips with @xmath19- as well as @xmath30-symmetry expose the bi atoms as bright spots .
this is hardly surprising , especially in the case of the rather strongly oriented @xmath30 tip orbital .
our computations also suggest that @xmath30 symmetry slightly sharpens the contrast of the topographic map , but as shown in the corrugation curves of fig .
[ corrugation ] , the effect may not be experimentally significant . tips with in - plane p - symmetry show striking changes in topographical maps . in the case where the tip orbital lies along the top - hollow ( nodal ) direction , @xmath31 ( fig .
[ topo](c ) ) , the contrast is inverted so that o(bi ) atoms appear as bright spots .
however , as discussed further in section iv below , visibility of oxygen atoms is spurious since the tunneling signal is still dominated by overlap between the horizontal tip orbital and the @xmath30 orbitals of bi atoms .
this is also the obvious reason why the bright spots in the topographical image lose their sharpness .
the change in contrast becomes even more dramatic in the case of the tip orbital lying along the top - bridge ( anti - nodal ) direction , @xmath32 . as seen in fig .
[ topo ] ( d ) , the fourfold symmetry is now broken , and bright lines appear perpendicular to the orbital direction .
in addition , intensity maxima are seen at bridge sites when the bonding oxygen orbital points to the same direction as the tip orbital .
this indicates that in this case also the nature of the topographic image is driven the overlap between the tip orbital and the @xmath30 orbital of bi atoms .
the intensity patterns in topographic images can be quantified as corrugation curves , such as those shown in fig .
[ corrugation ] .
interestingly , the horizontal tip orbitals not only invert the contrast of the topographic image , but they are also more sensitive to the horizontal position than tips with rotational symmetry around the z - axis .
this is seen in the more strongly varying corrugation as the tip is scanned between the top , bridge and hollow sites .
these differences , originating from the symmetry dependent nature of the tip - surface overlap , might be useful in pursuing a better lateral resolution in stm imaging .
unfortunately , our theoretical predictions can not be compared directly with measurements at this time since we are not aware of any experimental stm literature on bi2212 where a deliberately functionalized tip has been deployed .
however , some insight can be obtained by considering a study such as that of ref . , which discusses the effects of attaching o@xmath4 functionalizing molecules to a tungsten tip for investigating a metal phtalocyanine monolayer on an au(111 ) surface .
ref . finds selectivity with respect to molecular orbitals between normal and o@xmath4-functionalized tips , and observes the stm pattern of the adsorbate monolayer to rotate as different tips are used to probe the surface .
the shape and intensity of a @xmath52 spectrum depend on the tip symmetry as well as its position . in figs .
[ ldos](b - d ) , the @xmath52 curves are calculated at @xmath55 vertical distance between the tip and the surface atoms . at this tip - surface distance
, the gap features are clearly observed , and calculations indicate that the signal predominantly comes through the @xmath30 orbitals of the bi atoms .
there are clear changes in the vhs features at @xmath56 - 0.15 to -0.4 ev with tip site and orbital symmetry : they seem to vanish for the s - wave tip at a hollow site , and for both the horizontal p - wave tip orbitals at a top site .
the apparent reason for this is the increasing spectral weight of the bio bands with increasing binding energy .
these configurations seem to favour a relatively large overlap between the tip orbital and the p - orbitals of o(bi ) atoms . for an @xmath19-wave tip on the top site ,
the @xmath52 curve resembles the bi @xmath57 ldos , as expected .
however , this is also true for the horizontal @xmath33-wave tips at the bridge or hollow site .
these results support the conventional assumption that the experimental tips are predominantly @xmath19-wave .
we note also a trend in the overall intensities of the spectra .
as expected , horizontal p - symmetry of the tip orbital tends to decrease spectral intensity due to decreasing overlap . for s - wave symmetry , the intensity is highest for the top site and the lowest for the hollow site . for the horizontal p - symmetry ,
the top site gives the lowest intensity due to antisymmetry between @xmath30 of bi and the tip - orbital .
moving the tip either to bridge or hollow site allows an increased overlap between the tip orbital and @xmath30 of bi . of bi , @xmath58 of o(bi ) and @xmath30 of o(bi ) .
the first orbital hybridizes with the cuo@xmath0 bands , the second is dominant in bio bands ( bi pocket states ) , and the third is rather localized ; ( b ) , ( c ) and ( d ) show the di / dv curves with different tip orbitals at the high symmetry sites b , h and t , respectively.,scaledwidth=100.0% ] of bi , @xmath58 of o(bi ) and @xmath30 of o(bi ) .
the first orbital hybridizes with the cuo@xmath0 bands , the second is dominant in bio bands ( bi pocket states ) , and the third is rather localized ; ( b ) , ( c ) and ( d ) show the di / dv curves with different tip orbitals at the high symmetry sites b , h and t , respectively.,scaledwidth=100.0% ] + of bi , @xmath58 of o(bi ) and @xmath30 of o(bi ) .
the first orbital hybridizes with the cuo@xmath0 bands , the second is dominant in bio bands ( bi pocket states ) , and the third is rather localized ; ( b ) , ( c ) and ( d ) show the di / dv curves with different tip orbitals at the high symmetry sites b , h and t , respectively.,scaledwidth=100.0% ] of bi , @xmath58 of o(bi ) and @xmath30 of o(bi ) . the first orbital hybridizes with the cuo@xmath0 bands ,
the second is dominant in bio bands ( bi pocket states ) , and the third is rather localized ; ( b ) , ( c ) and ( d ) show the di / dv curves with different tip orbitals at the high symmetry sites b , h and t , respectively.,scaledwidth=100.0% ] di / dv spectra are also used to detect and analyze hidden subsurface irregularities .
these can include substitutional impurities , such as zn or ni replacing a cu atom @xcite , or dopant atoms like oxygen , the fingerprints of which have been discussed both experimentally @xcite and theoretically @xcite .
however , perhaps the most fascinating potential application may be the @xmath59 nematic intracell symmetry breaking of the subsurface electronic structure observed at very low values of hole doping in the pseudogap phase@xcite .
this phase transition in the electronic order is not observed in the topographic images , but rather in @xmath52 maps and in the so called z - maps derived from the differential conductance data .
the relevance to the present theoretical considerations arises from the fact that the subsurface bonding oxygen atoms in cuo@xmath4 appear non - equivalent when this low - symmetry phase is mapped .
we propose that breaking the symmetry of the tip by functionalization would give additional information about the nematic phase .
note that the observed patterns are difficult to interpret due to the filtering effect of the bio and sro layers .
it is interesting therefore to explore whether different tip symmetries can provide insight by probing different tunneling channels . in earlier
work@xcite we showed that the bonding oxygens of the cuo@xmath4 layers are invisible to a spherically symmetric tip placed right above that atom .
hence we choose a local perturbation at this bonding oxygen .
the simplest kind of perturbation is a change of the onsite energy , which could be induced by a be nearby dopant atom .
figure [ channels](d ) shows a tunneling tip positioned above a bonding oxygen in the cuprate layer , equivalent to the bridge site of fig . 1 .
to simulate the effect of a nearby impurity , we modify the onsite energy of the @xmath33-orbital along the bond by -2.0 ev.@xcite in fig . [
channels](c ) we compare its ldos ( blue line ) to that of the regular site ( red line ) .
the di / dv spectra from the bridge site with and without onsite modification are plotted for @xmath19 ( fig .
[ channels ] a ) and @xmath32 ( [ channels](b ) ) type tips to demonstrate how the impurity effect modifies the spectrum in different tunneling channels .
although the difference between the two ldos results is striking ( fig .
[ channels](c ) ) , the spectra obtained by a rotationally symmetric tip do not reflect this difference .
this can be explained by the results of ref . , where we showed that two tunneling channels in antiphase are open between the oxygen and the tip ( fig .
[ channels](d ) ) , causing destructive interference between the corresponding signals . [ the observed signal comes from different channels involving the cu atoms . ] in fig .
[ channels](d ) we show the di / dv curve measured by a tip with @xmath32 orbital , and a clear difference is seen between the spectra of the regular and perturbed oxygens .
the shapes of ldos ( b ) and the spectra ( d ) are not identical , but the relative change of the spectrum follows that of ldos , especially over the range between @xmath60 the change in the spectrum of fig .
[ channels](d ) is caused by a reversal of the relative phase between the two tunneling channels in fig .
[ channels](a ) due to p - type symmetry of the tip , a point to which we will return below .
this result suggests that experiments probing the breakdown of @xmath61 symmetry , such as those in ref .
, could be further clarified by varying the tip symmetry .
a proper understanding of the observed topographical images and spectra requires an analysis of the filtering matrix elements @xmath49 which describe the coupling between the tip and the surface orbitals , propagation across the oxide layers , and coupling between the superconducting cuprate layer and the covering oxide layers . starting from the latter ,
the cuprate layer is coupled to the oxide layers mainly through the overlap between the @xmath54 orbitals of the cu atoms and the @xmath30 orbitals of the apical oxygen above .
the propagation through the oxide layers is mediated by @xmath30 orbitals of the apical oxygen and the bi atom above on the surface . when the tip symmetry is varied , the relevant factor is that @xmath30 of the surface bi is the dominant surface orbital mediating the signal from the cuprate layer to the microscope tip .
hence , the analysis of the observed results can be built on the overlap between the tip orbital and the @xmath30 orbitals of the bi atoms in the vicinity of the tip .
while other surface orbitals are present , they are not strongly coupled to the cuprate bands , and hence they have a relatively low spectral weight in the vicinity of the fermi energy .
thus , the horizontal p - orbitals of bi and o(bi ) atoms are the dominant orbitals of the bio bands related to the `` bismuth pocket '' states above fermi - level and the corresponding valence bands below the `` spaghetti '' of bands ( fig .
[ ldos](a))@xcite .
since only the bi @xmath57 surface orbital is important , the analysis of the various tip orbitals is rather straightforward .
the @xmath62 and @xmath30-orbitals couple best on the top site , as seen in fig .
[ ldos](b ) . for the horizontal tip orbitals ,
the overlap between the tip and @xmath30 of a surface bi is lowest when the tip is directly above the bi atom . in the case of @xmath31
both the bridge and hollow sites seem to give a large overlap between tip orbital and @xmath30 of bi , but the hollow sites appear brighter since the overlap between the horizontal p - orbitals contributing to bio bands is larger at this site . for @xmath32 the fourfold symmetry
is broken .
since the orbital is oriented along the bonds between cu atoms , there is a zero overlap line when following the bonding direction perpendicular to the orbital orientation . on the other hand ,
@xmath32 overlaps strongly with the two bi atoms at the bridge site in the antinodal direction of @xmath32 .
moving towards the vicinal hollow sites preserves the strong overlap . at the hollow site ,
a strong overlap with four bi atoms can be found .
since the distance is @xmath63-fold as compared to the bridge site , overall the overlap remains roughly the same when moving from the bridge site to the hollow site , resulting in bright and dark stripes in fig .
[ topo](e ) .
we consider the possibility of analyzing subsurface irregularities by functionalizing the tip .
this may be possible if the interference between two different tunneling channels can be controlled .
as an example , let us consider the situation where the tip is at the bridge site ( see fig .
[ channels ] ( d ) ) , i.e. , above an oxygen atom in the cuo@xmath0 layer at an equal distance from the two neighbouring bi atoms ( @xmath64 and @xmath65 ) .
this opens two channels from the cuo@xmath0 layer ( @xmath66 orbitals of cu atoms @xmath67 and @xmath68 ) with filtering matrix elements @xmath69 and @xmath70 which are equal apart from a phase factor .
it is obvious that @xmath71 and @xmath72 are equal in the case of @xmath19- or @xmath30-symmetric tip , since in that case @xmath73 , but the opposite sign is obtained with horizontal @xmath33-orbitals .
-symmetric tip above the oxygen ; ( c ) ldos of the @xmath33-orbital of the bonding oxygen in the cuprate layer .
_ regular _ case ( red ) is compared with an orbital with a _ modified _ onsite energy ( blue ) ; ( d ) two competing channels from the bonding oxygen ( @xmath74 ) through @xmath75 orbitals ( @xmath67 or @xmath68 ) , apical oxygens ( @xmath76 and @xmath77 ) and the surface bismuths ( @xmath64 and @xmath65 ) to the tip ( @xmath78).,scaledwidth=100.0% ] -symmetric tip above the oxygen ; ( c ) ldos of the @xmath33-orbital of the bonding oxygen in the cuprate layer .
_ regular _ case ( red ) is compared with an orbital with a _ modified _ onsite energy ( blue ) ; ( d ) two competing channels from the bonding oxygen ( @xmath74 ) through @xmath75 orbitals ( @xmath67 or @xmath68 ) , apical oxygens ( @xmath76 and @xmath77 ) and the surface bismuths ( @xmath64 and @xmath65 ) to the tip ( @xmath78).,scaledwidth=100.0% ] + -symmetric tip above the oxygen ; ( c ) ldos of the @xmath33-orbital of the bonding oxygen in the cuprate layer .
_ regular _ case ( red ) is compared with an orbital with a _ modified _ onsite energy ( blue ) ; ( d ) two competing channels from the bonding oxygen ( @xmath74 ) through @xmath75 orbitals ( @xmath67 or @xmath68 ) , apical oxygens ( @xmath76 and @xmath77 ) and the surface bismuths ( @xmath64 and @xmath65 ) to the tip ( @xmath78).,scaledwidth=100.0% ] -symmetric tip above the oxygen ; ( c ) ldos of the @xmath33-orbital of the bonding oxygen in the cuprate layer .
_ regular _ case ( red ) is compared with an orbital with a _ modified _ onsite energy ( blue ) ; ( d ) two competing channels from the bonding oxygen ( @xmath74 ) through @xmath75 orbitals ( @xmath67 or @xmath68 ) , apical oxygens ( @xmath76 and @xmath77 ) and the surface bismuths ( @xmath64 and @xmath65 ) to the tip ( @xmath78).,scaledwidth=100.0% ] to consider interference effects , we also need to compare elements of the spectral function @xmath79 where @xmath74 denotes the p - orbital of the oxygen along the bond between the cu atoms .
it is evident that the p - symmetry leads to an antiphase between elements @xmath80 and @xmath81 , and hence the two first spectral terms are equal to each other , but the last one has the opposite sign . finally , using equations , and we obtain the total current from orbital @xmath74 to the tip : @xmath82 it is clear that - sign on the right hand side of eq .
( 6 ) applies to the case with @xmath73 , i.e. , the @xmath19- or @xmath30-symmetry .
hence the oxygen below the tip would be invisible in stm . on the other hand , constructive interference is obtained for horizontally @xmath33-symmetric tip .
this should emphasize any variations in the horizontal p - orbital of the bonding oxygen .
the situation is of course reversed if the bonding oxygen is replaced by an @xmath19-wave symmetric perturbation , i.e. there is a constructive interference for @xmath19- and @xmath30-symmetric tips
. this explains our comparison of a spectrum at the bridge site for unperturbed system with a vacancy at the site of the bonding oxygen . with regard to the @xmath59 transition
, the main practical conclusion of our analysis is that obtaining information related to the non - equivalent bonding oxygens can be enhanced by manipulating the tip symmetry by functionalizing molecules .
this suggests that an s - wave symmetric tip gathers the signal from a neighbourhood of the bonding oxygen below , while a p - wave tip should be more selective to the on - site oxygen orbitals .
we have investigated the effect of different tip symmetries on topographic stm images and di / dv spectra .
the change in tip symmetry can lead to contrast inversion and even to the breaking up of the @xmath61 symmetry due to strong differences in the nature of the tip - surface overlap .
furthermore , the corrugation curves show that the horizontal p - orbitals at the apex of the tip would be more sensitive to position of the surface atoms .
the di / dv curves reveal the leading sample states ( cuo@xmath0-band , bio - band or some more localized states ) to which different symmetries tend to couple .
more strikingly , changing the symmetry of the tip can be used to change the relative phase between two open tunneling channels .
this could be applied to detect subsurface irregularities which would otherwise be invisible to stm .
* acknowledgments * this work is supported by the us department of energy grant de - fg02 - 07er46352 and benefited from the allocation of supercomputer time at nersc and northeastern university s advanced scientific computation center ( ascc ) .
i.s would like to thank ulla tuominen foundation for financial support . this work benefited from resources of institute of advanced computing , tampere .
discussions with matti lindroos are gratefully acknowledged .
matrix element effects are well - known to be important in other highly resolved spectroscopies such as angle - resolved photoemission@xcite , resonant inelastic x - ray scattering(rixs)@xcite , compton scattering @xcite and positron - annihilation @xcite .
s. sahrakorpi , m. lindroos , r. s. markiewicz , and a. bansil , phys .
95 * , 157601 ( 2005 ) ; m. lindroos and a. bansil , phys .
lett . * 77 * , 2985 ( 1996 ) ; a. bansil , m. lindroos , s. sahrakorpi , and r. s. markiewicz , phys .
b * 71 * , 012503 ( 2005 ) ; m. c. asensio , j. avila , l. roca , a. tejeda , g. d. gu , m. lindroos , r. s. markiewicz , and a. bansil , phys . rev .
b * 67 * , 014519 ( 2003 ) . r.s .
markiewicz and a. bansil , phys .
lett . * 96 * , 107005 ( 2006 ) ; y. w. li , d. qian , l. wray , d. hsieh , y. xia , y. kaga , t. sasagawa , h. takagi , r. s. markiewicz , a. bansil , h. eisaki , s. uchida , and m. z. hasan , phys .
b * 78 * , 073104 ( 2008 ) .
a. bansil , r. s. rao , p. e. mijnarends , and l. schwartz , phys .
b * 23 * , 3608 ( 1981 ) ; p. e. mijnarends and a. bansil , phys .
b * 13 * , 2381 ( 1976 ) ; g. stutz , f. wohlert , a. kaprolat , w. schlke , y. sakurai , y. tanaka , m. ito , h. kawata , n. shiotani , s. kaprzyk , and a. bansil , phys .
b * 60 * , 7099 ( 1999 ) .
mijnarends , a.c .
kruseman , a. van veen , h. schut , and a. bansil , j. phys . : condens .
matter * 10 * , 10383 ( 1998 ) ; l. c. smedskjaer , a. bansil , u. welp , y. fang and k. g. bailey , physica c * 192 * , 259 ( 1992 ) .
we expect the rigid band model to be a good approximation for doping away from the cuo@xmath0 planes
. it will be interesting to further examine doping effects by using supercell or other first principles approaches @xcite .
a. bansil , zeitschrift naturforschung a * 48 * , 165 ( 1993 ) ; h. asonen , m. lindroos , m. pessa , r. prasad , r.s .
rao , and a. bansil , phys .
b * 25 * , 7075 ( 1982 ) ; r. prasad and a. bansil , phys .
b21 , 496 ( 1980 ) .
this equation can also be derived as a special case of landauer - bttiker approach as expressed in ref .. on the other hand , eq .
can be converted into tersoff - hamann@xcite formalism in the limiting of a spherical tip orbital in real space@xcite .
note that the density matrices in eq .
concern , in principle , an interacting system . an alternative approach would be to use noninteracting density matrices and replace the hopping terms by a t - matrix , @xmath83 however , for the tip distance used in this work , the tip - sample hopping terms are of the order of @xmath84 , while the interaction terms within the sample are of the order of @xmath85 [ e.g. , the hopping between @xmath32 of the bonding oxygen and the @xmath40 of the neighbouring cu atom is @xmath86.$ ] hence , it is safe to utilize noninteracting density matrices in calculating the conductance .
lda calculations tend to give conducting bio bands the bottom of which are slightly below fermi - level . however , one never observes these states in arpes or sts , and hence it is possible that doping the bi2212 samples may lift these states off the fermi level @xcite . | we discuss how variations in the scanning tunneling microscope ( stm ) tip , whether unintentional or intentional , can lead to changes in topographic images and di / dv spectra .
we consider the possibility of utilizing functionalized tips in order to improve the sensitivity of stm experiments to local irregularities at the surface or hidden below the surface layers .
the change in the tip symmetry can radically alter the contrast of the topographic image due to changes in tip - surface overlap .
the di / dv curves change their shape according to which sample bands the tip orbital tends to overlap . in addition , relative phases between competing tunneling channels can be inverted by changing the tip symmetry , which could help reveal the origin of a local irregularity in tunneling spectrum . |
cyclic phenomena are often ignored when studying epidemiological and evolutionary processes , but nevertheless they can have important consequences , e.g. we contract most infective diseases only once in our lifetime , because our immune system has `` memory '' .
vaccines are designed based on this knowledge , and they work quite efficiently .
however , it is widely accepted that some viruses , such as the flu virus , mutate in order to make themselves unrecognizable by our immune system , and thus be able to reinfect us .
immunity to pertussis ( whooping cough ) is temporary , and decreases as the time after the most recent pertussis infection increases .
the chicken pox ( agent vzv varicella zoster virus ) also repeats , because immunity wanes with time .
our model is motivated by this type of situation . in epidemiological models populations
are often categorized in three states : susceptible @xmath10 , infected @xmath11 and recovered @xmath12 .
there exists a vast literature about the so - called @xmath13 models , where the `` loss of immunity '' step is not considered ( e.g. the classic texts by bailey @xcite , anderson and may @xcite , the review by hethcote @xcite ) .
we consider the case when the `` loss of immunity '' step is present , otherwise referred to as @xmath14 models @xcite .
recently it has been shown that many species of bacteria are able to produce toxic substances that are effective against bacteria of the same species which do not produce a resistance factor against the toxin @xcite .
these bacteria exist in colonies of three possible types : sensitive @xmath10 , killer @xmath15 and resistant @xmath12 .
the killer type produces the toxin , and the resistance factor that protects it from its own toxin , at a metabolism cost .
the resistant strain only produces the resistance factor ( at a smaller metabolism cost ) .
the sensitive strain produces neither .
clearly , the @xmath16 colony can be invaded by a @xmath17 colony , the @xmath17 can be invaded by @xmath18 , and the @xmath18 can be invaded by @xmath16 . in our first model
the population is categorized into three `` species '' : @xmath19 , @xmath20 , @xmath21 , and the rules are such that when an @xmath19 meets a @xmath20 , it becomes @xmath20 , when @xmath20 meets @xmath21 , it becomes @xmath21 , when @xmath21 meets @xmath19 , it becomes @xmath19 .
the total population size is conserved .
otherwise this could be seen as a three - party voter model , when the follower of a certain party `` converts '' a follower of another certain party when they meet .
the reaction is similar to the cyclical `` rock - paper - scissors game '' , of which a biological example has been found recently @xcite .
it is played by the males of a lizard species that exist in three versions : the blue throat male defends a territory that contains one female , the orange throat male defends a territory with many females , and the male with a throat with yellow stripes does not defend his own territory , but can sneak into the territory of the orange males and mate with their females .
hence , a `` blue '' population can be invaded by `` orange '' males , while an `` orange '' population is vulnerable to `` yellow '' males , who on their turn are at an disadvantage against the `` blue '' males , who defend their territory very well .
many spatial models that relate to such systems have been built @xcite .
the second model is a three component version of the famous kimura model of neutral genetic drift @xcite . in that case ,
when individual from two species meet , the offspring may be either of the first or the second species , with equal probability .
consider a system in which three species @xmath19 , @xmath20 , @xmath21 are competing in a way described by the reaction : @xmath1 , @xmath22 , @xmath3 . the rate equations for this system will be : @xmath23 with @xmath24 const .
( we assume the rates are the same , in which case a time
rescale will remove them from the equations . )
these equations can be rewritten as : @xmath25 which leads to the second conservation rule : @xmath26 = const .
the above model contrasts with the neutral drift model ( @xmath27 or @xmath5 , @xmath28 or @xmath7 , @xmath29 or @xmath9 ) for which : @xmath30 assuming that the system is subject to stochastic noise due to poisson birth and death processes ( intrinsic noise ) we get the master equation for the cyclic model : @xmath31@xmath32@xmath33\ ] ] for the neutral drift case the master equation is : @xmath34@xmath35@xmath36@xmath37@xmath38\ ] ] now we introduce the `` shift '' operators , defined by @xmath39 and @xmath40 and likewise for b and c operators .
the master equation for the cyclic model now reads : @xmath41 p(a , b , c , t)\ ] ] and for the neutral drift one : @xmath42@xmath43 p(a , b , c , t)\ ] ] next we transform to the intensive quantities @xmath44 we use the system size expansion of horsthemke and brenig @xcite .
however , the notation and style is closer to van kampen @xcite . the shift operators become @xmath45 further we use the rules of transformation of random variables to define : @xmath46 before we go ahead with the expansion , a few comments are necessary . from the rate equations ,
it is clear that our system does not have one single steady state or limit cycle . instead
, the mean - field solution is completely dependent on the initial conditions , and we do not have a macroscopic solution to expand about , but rather an infinity of neutrally stable cycles ( for the competition case ) or points ( for the neutral model . ) the above expansion , otherwise known as kramers - moyal expansion , is discussed by van kampen @xcite .
it is a risky expansion , and it does not work in most cases , mainly because higher derivatives are not small themselves .
also , with time , large fluctuations may occur , and since our system presents itself as a fluctuations driven one , it becomes even more important to watch out for this sort of complications . in the case of neutral drift ,
the second order term is the leading one , so we would expect this ansatz to work for large system sizes . in the cyclic competition model , the first order term does not drive the system out of the ( neutrally ) stable trajectories , and again it would be the second order term to essentially determine the fate of the system .
so , we set out to investigate the applicability of the kramers - moyal expansion to the cyclic competition and neutral drift three - species systems .
we transform the master equation into an equation for @xmath47 by substituting the expressions for the shift operators in it , and further by grouping together the terms of the same order in @xmath48 .
the terms of order @xmath49 cancel , and the first terms in the expansion are of order @xmath50 ; keeping only those gives us the first order equation for the cyclic case : @xmath51 w\ ] ] for the neutral drift case the first - order term is identically zero . the second order term is obtained by considering the terms of order @xmath52 in the expansion of the master equation : @xmath53 @xmath54w\ ] ] and is identical for both cyclic and neutral drift cases .
the concentrations @xmath55 , @xmath56 , @xmath57 of a three component system are commonly represented by the distances of a point inside an equilateral triangle of unit height from its sides ( fig .
1 ) : for the neutral drift case the first order equation tells us that in the mean field approximation the system will remain in the initial state forever . after some algebra , the first order equation [ 13 ] for the cyclic system becomes : @xmath58 at the centre of the triangle ( @xmath59 ) all three expressions above are zero , so in the first order approximation the system will stay at that state forever .
it can be verified that the product @xmath60 is a solution of the first - order equation , as is any function of that product .
the lines @xmath61 const . will then represent possible trajectories of the system , which are closed , since there is no term to drive the system out of those trajectories . in this ( mean - field ) approximation ,
the cyclic model exhibits global stability with constant system size @xmath48 @xcite .
the populations of each species oscillate out of phase with one - another , and the amplitude remains constant @xcite .
these are the same trajectories that are obtained when one solves the rate equations [ 1 ] .
2 shows some of those closed trajectories .
the second order term [ 14 ] , which represents the `` diffusion term '' must be added to the first order equation to give the fokker - planck equation for our system .
this term is identical for both the cyclic ( competing species ) and neutral genetic drift cases .
as we will see , this makes the long ( evolutionary ) time fate of both systems essentially the same , at least for reasonably large system sizes .
the trajectories of a given realisation of the system can easily be obtained by simulation : they initially spiral out of the centre , with the populations of each species oscillating out of phase with one another , so that the total size of the population is conserved .
the amplitude drifts until one of the populations becomes zero ( species goes extinct ) . in other words ,
the slow second order term makes the system cross from a closed ( mean - field ) trajectory to a neighbouring one .
this would suggest an adiabatic approximation . in the neutral drift model ,
the fluctuations drive the system from one point
solution of the mean - field equations to a neighbouring one .
the computer simulations of both systems start with an equal number of @xmath19 , @xmath20 , @xmath21 , ( start at the centre ) and generate times for the next possible reaction event with exponential distribution as @xmath62 , ( for @xmath63 reaction , and similarly for the other two reactions , ) where @xmath64 is a random variable with uniform distribution in @xmath65 $ ] ( this means the events are really independent ) @xcite .
the reaction which occurs first is then picked and the system is updated .
the process is repeated until one of the species , and then another one , go extinct .
3 shows the variation with time of the number of @xmath19 for a realization of the system in the cyclic competition scenario , and fig . 4 shows the evolution of the number of @xmath19 population in the neutral drift case ( in both realizations
the total population size is @xmath66 . ) from the time series for the population numbers it looks as if the neutral drift model is some sort of `` adiabatic '' approximation of the cyclic competition model . if in the neutral drift model the population number is random walking , in the cyclic competition one the amplitude of oscillations is random walking ! in other words , if we average the cyclic competition model over the cyclic orbits , we get essentially the same behaviour as the neutral drift model .
then the fluctuations drive both models from a neutrally stable cycle ( point ) to the neighbouring ones , and these fluctuations remain quite small by virtue of the vicinity of the neighbouring cycle ( point . )
it is quite funny how we all tend to go around in circles in life , even though the point we end up to is still the same !
we investigated the probability distribution of first extinction times .
for that we ran fifty thousands copies of the system for each ( different ) size of the system , and the number of survivors was plotted vs. time . in the semilog axes we get a straight line , except for the few `` rare events ''
the slope of this straight line is -3.01 for system size 3000 , -2.98 for system size 6000 , -2.99 for system size 9000 .
5 shows the plots for these three different system sizes : 3000 , 6000 , and 9000 .
the time dependence of survival probability scales approximately with @xmath67 , but there is still a very weak dependence left for small @xmath48 .
this can be justified by looking at the fokker - planck equation for this system , which contains both a term of order @xmath50 and @xmath52 . to further investigate the behaviour of the system
, we looked at the cummulative , conditional on being alive ( which is equivalent to a normalization condition , ) probability distribution for @xmath68 .
if @xmath69alive@xmath70 increases linearly with @xmath71 for small @xmath71 , the extinction rate will be non - zero , and we will get exponential decay .
this corresponds to a constant distribution for @xmath72alive@xmath70 .
6 shows these plots for @xmath73 , @xmath74 , @xmath75 , and @xmath76 in linear axes . for the neutral drift case the data from the simulations give us a slope of -2.992 for system size 6000 , -2.991 for system size 3000 .
the data collapse when time is rescaled with @xmath67 , and that is in agreement with the fact that there is only a term of the order @xmath67 present in the fokker - planck equation . in fig .
7 we show the number of survivors vs. time plots for cyclic ( competition ) and drift cases together , for system size @xmath77 .
they clearly agree .
in order to get an expression for @xmath78 for long times , we looked at our `` experimental '' data .
it is possible to find a complete set of eigenfunctions of the laplace equation , which vanish at the boundary .
having this complete set of eigenfunctions for the equilateral triangle @xcite , the probability density can be written : @xmath79 where @xmath80 are the abovementioned symmetric eigenfunctions .
the non - normalized eigenfunctions are then obtained from the general expression : @xmath81\ ] ] where we are using the relation @xmath82 and summation over the index pair @xmath83 means summation for @xmath84 , @xmath85 , @xmath86 , @xmath87 , @xmath83 , @xmath88 @xcite .
we get a symmetric eigenfunction when @xmath89 is a multiple of 3 , @xmath90 is also a multiple of 3 , but @xmath91 , and @xmath92 . , of all the symmetric eigenfunctions , pinsky only considers the non - degenerate ones .
there are also two - fold degenerate symmetric eigenfunctions , which we verified to be orthogonal to the original eigenfunctions . ] in order to calculate the expansion coefficients , we took snapshots of the system at given times , and then used the relation : @xmath93 where @xmath94 is the number of experimental points .
the time - evolution of the expansion coefficients is given in fig .
the doubly degenerate functions die out quite soon , while the lam symmetric functions persist . with the coefficients obtained this way we
can then construct the probability density at different times .
some snapshots at the @xmath95 are shown in fig .
9 . at @xmath96
the @xmath97 is a @xmath98-function . as time passes , for @xmath99
we see the @xmath78 starts to spread , and it spreads even more for @xmath100 , becoming a `` cake '' for very long times ( @xmath74 , ) for both the cyclic system and the neutral drift one .
the `` cake '' obtained this way shows some gibbs oscillations , which are due to the inclusion of only a few terms in the expansion , and the presence of the absorbing boundary . for a uniform probability distribution
the expansion coefficients are proportional to @xmath101 , where @xmath90 is the first index in the pair @xmath83 .
it can be seen that our `` experimental '' expansion coefficients approach those values .
the second order fokker
planck equation for the drift case accepts solutions of the form : @xmath102 with @xmath103 the smallest eigenvalue of the spatial equation .
we solved the eigenvalue problem by using the galerkin method , with the help of maple , and obtained a value of @xmath104 for the smallest eigenvalue , when we keep the first six eigenfunctions in the expansion .
this is in very good agreement with our `` experimental '' data .
now consider the cyclic ( competition ) case when the rates of the three equations are not the same , i. e. the rate equations read like : @xmath105 we can always rescale the @xmath106 so that the matrix @xmath107 be symmetric , and the equations read : @xmath108 again , @xmath109const . by manipulating the rate equations , we can find the second integral of motion to be @xmath110const .
the new centre will then be not at the point ( 1/3 , 1/3 , 1/3 ) , but at @xmath111 . the lines along which the quantity @xmath112 is constant are shown in fig .
we need to check whether the exponential decay behaviour is universal when the symmetry is broken this way . for this
we ran fifty thousand simulations for different combinations of @xmath113 , @xmath114 , and @xmath115 , for system size @xmath116 , starting from the new centre ( @xmath117 , @xmath118 , @xmath119 . ) the number of survivors vs. time plots were again obtained , and those plots show that the exponential decay behaviour is robust .
the time scale is now dependent on the initial distance from the boundary , with the equal rates case having the largest distance , and so the largest time scale ( and the smallest slope . ) some of those plots are shown in fig
. 11 . in that figure ,
the times for the equal rates case ( for which the sum of rates is 3 ) are multiplied by three , to make them comparable to the times for the unequal rates case , for which the sum of rates is 1 .
we have considered an @xmath0 model in both the cyclic competition and neutral genetic drift versions , and studied the long - term behaviour of such a model .
the number of the @xmath106 species in the cyclic competition case oscillates with a drifting with time amplitude , until one of the species ( and then the next one in the cycle ) goes extinct . in the neutral drift case
the number of the @xmath106 species drifts , until one of them becomes zero . in both scenarios
the number of survivors vs. time plots show an exponential decay , with the same exponent .
the result is verified by writing and solving the fokker
planck equation for the second model .
finally , its robustness is checked against variations in the rate of the three different reactions in the system .
it is very interesting to note that there is no difference in the time scale for the ensemble of species in cyclic competition and the case of neutral genetic drift .
there is growing concern about the effects of habitat fragmentation in the survival of the species @xcite .
if the population of a certain species goes extinct in one patch ( e.g. a herd , school , swarm ) while it still survives in other patches , then what is known as `` rescue effect '' can prevent global extinction @xcite .
otherwise , the species is doomed to go extinct altogether .
the models with individuals of `` species '' @xmath19 , @xmath20 , and @xmath21 distributed in a lattice have been studied by szbo et al . @xcite .
if we try to model habitat fragmentation as an ensemble of patches , each of those patches can be considered as one copy of our non - spatial system . however , with continuous migration in and out of the patch , the non - spatial picture described in this paper would be seriously perturbed .
one aspect of this immigration and emigration has been recently discussed by togashi and kaneko @xcite . in a first approximation ,
the broad range of extinction times could relate to the persistence of species that exist in different `` versions '' such as certain lizards , or even some kinds of bacteria .
applied to its epidemiological scenario , it would relate to the endemicity of certain diseases , namely those with mutating pathogen , and then the outbreaks of epidemics in certain patches ( areas ) .
the many patches will then form a network @xcite .
we have work in progress which places our @xmath0 ensemble in one of these small worlds with a scale - free topology .
we thank nicolaas g. van kampen for constructive criticism , and michael dbeli for helpful and interesting discussions and suggestions .
99 bailey n. t. , _ the mathematical theory of infectious diseases _ , 2nd edition , london : griffin ( 1975 ) .
anderson r. m. and may r. m. , _ infectious diseases of humans _ , oxford university press , london ( 1991 ) .
hethcote h. w. , siam review , * 42 * , 599 ( 2000 ) .
cooke k. l. , calef d. f. , and level e. v. , _ nonlinear systems and its applications _ , academic press , new york , 73 ( 1977 ) .
longini i. m. , mathematical biosciences , * 50 * , 85 ( 1980 ) .
reeves p. , _ the bacteriocins _ , springer verlag , new york ( 1972 ) .
james r. , lazdunski c. , and pattus f. , ( editors ) _ bacteriocins , microcins and lantibiotics _ , springer verlag , new york ( 1991 ) .
sinervo b. and lively c. , nature , * 380 * , 240 ( 1996 ) .
maynard smith j. , nature , * 380 * , 198 ( 1996 ) .
szab g. , and czrn t. , phys .
e , * 63 * , 061904 ( 2001 ) .
szab g. , santos m. a. , and mendes j. f. f. , phys .
e , * 60 * , 3776 ( 1999 ) .
frachebourg l. , krapivsky p. l. , and ben - naim e. , phys .
rev e , * 54 * , 6186 ( 1996 ) .
pimm s. l. , nature , * 393 * , 23 ( 1998 ) .
blasius b. , huppert a. , and stone l. , nature , * 399 * , 354 ( 1999 ) .
kimura m. , and weiss g. h. , genetics , * 49 * , 561 ( 1964 ) .
weiss g. h. , and kimura m. , j. appl .
prob . , * 2 * , 129 ( 1965 ) .
horsthemke w. , and brenig l. , zeitschrift fr physik b , * 27 * , 341 ( 1977 ) .
horsthemke w. , malke - mansour m. , and brenig l. , zeitschrift fr physik b , * 28 * , 135 ( 1977 ) .
van kampen n. g. , adv .
phys . * 34 * , 245 ( 1976 ) .
van kampen n. g. , _
stochastic processes in physics and chemistry ( revised edition ) _ , north - holland ( 1997 ) .
ruijgrok th . , and ruijgrok m. , j. stat .
phys . , * 87 * , 1145 ( 1997 ) .
szab g. , and hauert c. , phys .
, * 89 ( 11 ) * , 118101 ( 2002 ) .
gibson m. a. , and bruck j. , j. phys .
a , * 104 * , 1876 ( 2000 ) .
lam m. g. , _ leons sur le thorie mathmatique de lelasticit des corps solides _ , paris , bachelier ( 1852 ) .
pinsky m. a. , siam j. math .
, * 11 * , 819 ( 1980 ) .
brown j. h. , and kodric - brown a. , ecology , * 58 * , 445 ( 1977 ) .
earn d. j. , levin s. a. , and rohani p. , science , * 290 * , 1360 ( 2000 ) .
albert r. , and barabsi a .-
, * 74 * , 47 ( 2002 ) .
hethcote h. w. , math .
, * 28 * , 335 ( 1976 ) .
hethcote h. w. , _ three basic epidemiological models _ , in applied mathematical ecology , l. gross , t. g. hallam , and s. a. levin , eds . , springer - verlag , berlin ( 1989 ) .
togashi y. , and kaneko k. , phys .
, * 86 ( 11 ) * , 2459 ( 2001 ) . | we study the @xmath0 model ( @xmath1 , @xmath2 , @xmath3 ) , and its counterpart : the three component neutral drift model ( @xmath4 or @xmath5 , @xmath6 or @xmath7 , @xmath8 or @xmath9 . ) in the former case , the mean field approximation exhibits cyclic behaviour with an amplitude determined by the initial condition .
when stochastic phenomena are taken into account the amplitude of oscillations will drift and eventually one and then two of the three species will become extinct . the second model remains stationary for all initial conditions in the mean field approximation , and drifts when stochastic phenomena are considered .
we analyzed the distribution of first extinction times of both models by simulations of the master equation , and from the point of view of the fokker - planck equation .
survival probability vs. time plots suggest an exponential decay . for the neutral model
the extinction rate is inversely proportional to the system size , while the cyclic model exhibits anomalous behaviour for small system sizes . in the large system size
limit the extinction times for both models will be the same .
this result is compatible with the smallest eigenvalue obtained from the numerical solution of the fokker - planck equation .
we also studied the behaviour of the probability distribution .
the exponential decay is found to be robust against certain changes , such as the three reactions having different rates . |
starting with the work of @xcite , it has been appreciated that the technique of localisation provides a powerful tool for performing exact non - perturbative computations in supersymmetric field theories defined on curved manifolds . while @xcite computed the exact path integral of four - dimensional @xmath5 gauge theories on a round four - sphere ,
later the attention shifted to three - dimensional theories .
the authors of references @xcite computed the partition function of general @xmath5 superconformal gauge theories , with chern
simons terms and matter in arbitrary representations , placed on the round three - sphere .
if the theory has an r - symmetry , then it is possible to define it on more general manifolds , while preserving supersymmetry , by turning on a background non - dynamical gauge field coupled to the r - symmetry current @xcite .
the first examples of such backgrounds were considered in @xcite , where the exact partition function of supersymmetric gauge theories defined on different squashed three - spheres was computed .
it was later shown in @xcite that the `` ellipsoid '' partition function of @xcite extends to a very general class of supersymmetric three - dimensional geometries .
an asymptotically locally ads ( alads ) supergravity solution dual to the construction of @xcite was presented in @xcite , and various extensions were obtained in @xcite . in this paper
we will start exploring the gauge / gravity duality for four - dimensional @xmath2 superconformal field theories defined on curved manifolds , with the aim of comparing gravity calculations with exact field theory results .
this problem in four ( boundary ) dimensions turns out to be more complicated than in one dimension lower for different reasons .
when the new minimal or conformal supergravity formulations of rigid supersymmetry are used , unbroken supersymmetry in euclidean signature implies that the field theory must be defined on a hermitian manifold @xcite . however , exact calculations of the path integral using localisation in field theories on such manifolds have been less developed than their three - dimensional analogues so far ( but see the recent @xcite )
. moreover , in four dimensions the path integral generically has logarithmic divergences , and extracting physical information from its finite part is more subtle than in three dimensions due to a number of ambiguities .
these issues are under better control if one considers supersymmetric manifolds with topology @xmath6 . in these cases the path integral with periodic boundary conditions
is expected to be proportional to a supersymmetric index @xcite , refined with fugacities related to a choice of complex structure on @xmath6 @xcite . for superconformal theories , the index
may be equivalently defined as the generating function of operators weighted by their fermion number , so that the contributions from the long multiplets cancel out , and as shown in @xcite this may be obtained by putting the theory on @xmath7 . in @xcite the superconformal index
was successfully compared with the spectrum of kaluza
klein modes dual to the protected operators being counted by the index . however , from the perspective of the euclidean path integral , it is more natural to compare the field theory partition function with the holographically renormalised supergravity on - shell action @xcite . in the large @xmath8 limit
the index scales like @xmath9 @xcite , while the on - shell action scales like @xmath10 .
therefore , it would appear that the latter can not capture new information about the field theory , apart from the weyl anomaly @xcite . when the field theory is placed on a round ( i.e. conformally flat ) @xmath3 ,
the gravity dual is simply global ads@xmath0 with euclideanized time , and the renormalised gravity action has been argued to reproduce the _ casimir energy _ of the field theory super yang mills , and the casimir energy can be computed at weak coupling . ] on @xmath11 @xcite .
more generally , the on - shell action is expected to capture the ratio between the path integral and the index , which has been dubbed `` index casimir energy '' in references @xcite .
this quantity is sensitive to the regularisation of the path integral , and will be discussed in more detail towards the end of the paper . on the gravity side ,
constructing five - dimensional supersymmetric alads solutions is technically more difficult than obtaining analogous solutions in one dimension less .
one reason is that in odd dimensions the asymptotic form of the fields ( in a fefferman
graham coordinate system @xcite ) contains logarithmic terms , suggesting that analytic solutions are much harder to find .
@xcite presented an alads solution that is special in this regard , as the logarithmic terms do not arise .
this solution has a conformal boundary with the topology of @xmath12 , however its isometry group is @xmath13 , with the non - compact time being generated by a null ( in the boundary ) killing vector , hence it is not possible to analytically continue it to euclidean signature , while keeping a real metric .
although in order to compare with localisation results we are ultimately interested in euclidean conformal boundaries , we will start working with minimal five - dimensional gauged supergravity in lorentzian signature .
this will allow us to construct a supersymmetric solution utilising the formalism of @xcite , thus avoiding complications arising in euclidean signature , where we would have to deal with a complex graviphoton field from the outset .
the solution that we will present is a one - parameter family of supersymmetric deformations of ads@xmath0 , whose conformal boundary includes a biaxially squashed three - sphere with @xmath4 symmetry , as well as a non - trivial gauge field .
it preserves two of the eight supercharges of minimal gauged supergravity .
globally , the space - time is equivalent to ads@xmath0 , with topology @xmath14 , where the first factor is a time direction , and the @xmath15 factor arises from the three - sphere of the boundary smoothly shrinking to zero size in the interior . in particular , the space - time does not have a horizon .
the metric , gauge field , and killing spinor of the solution restricted to the asymptotic boundary reduce precisely to a background @xcite solving the charged conformal killing spinor equation @xcite , as well as the new minimal version of rigid supersymmetry in curved space @xcite .
therefore we interpret our solution as the gravity dual to an @xmath2 superconformal gauge theory on a curved , non conformally - flat , background preserving two supercharges .
another feature of this background is that the aforementioned logarithmic divergences do not appear in the partition function @xcite .
our five - dimensional solution can be uplifted either to type iib or to eleven - dimensional supergravity , with the specific dual field theory depending on the choice of internal manifold @xmath16 . in the uplift to type iib supergravity @xcite , @xmath16 can be a sasaki einstein five - manifold , as well as a more general manifold arising from the classification of @xcite . in the uplift to eleven - dimensional supergravity @xcite
, @xmath16 can be any of the manifolds in the classification of @xcite .
a main feature of the solution is that although it has been constructed using a combination of perturbative and numerical analysis , all the quantities of physical interest , namely the on - shell action and the holographic conserved charges , have been obtained analytically as a function of the boundary squashing parameter . upon a simple wick rotation , and compactification of the euclidean time ,
the boundary metric describes a squashed @xmath3 , and the background gauge field becomes complex , consistently with the results of @xcite . while the boundary metric remains real , in the bulk both the gauge field and the metric become _
complex_. however , the renormalised on - shell action is real and we show that it reduces to a boundary term .
moreover , it receives a crucial contribution from the chern simons term , with the result depending on the gauge choice for the graviphoton .
we will explain that demanding the killing spinors ( on the boundary as well as in the bulk ) to be invariant with respect to a specific choice of killing vector @xmath17 generating time translations , fixes the gauge uniquely .
this is necessary in order to have well - defined spinors after performing the wick rotation , and on the field theory side it is required for defining the euclidean path integral with supersymmetric boundary conditions @xcite .
we will show that the holographic charges obey relations consistent with the rigid supersymmetry algebras of the field theories defined on the boundary geometry .
moreover , in the euclidean setting , we will see that the renormalised on - shell action reproduces a suitably defined casimir energy , arising from the supersymmetric path integral , up to terms that we will propose to be removable by a choice of finite local counterterms .
in particular , the casimir energy depends linearly on the complex structure modulus @xmath18 of the boundary manifold and the @xmath19 anomaly coefficient , while we will show that there exists a choice of finite local counterterms , constructed from the supergravity fields and the complex structure , which cancels the remainder terms .
the rest of the paper is organised as follows . in section [ susysec ]
we review the derivation of the relevant ode stemming from supersymmetry , following @xcite . in section [ newsol_section ]
we present our new solution , combining analytical methods and a numerical integration of the ode .
section [ holorensection ] is devoted to holographic renormalisation and the computation of conserved charges . in section [ summarysol ]
we summarize the main features of the solution . in section [ fielddisc ]
we discuss key properties of the dual field theories , and compare these with the gravity results .
section [ disc ] concludes summarising our findings and outlining directions for future work .
two appendices contain technical details on the solution and the holographic quantities .
we begin by presenting the supersymmetry equations of minimal five - dimensional gauged supergravity , adopting the formalism of @xcite . in particular , we will be interested in supersymmetric solutions of the time - like class , derived from an ansatz possessing ( locally ) @xmath20 symmetry .
this leads precisely to the same equations obeyed by the supersymmetric black hole solutions found in @xcite , and in this section we will follow closely the derivation therein . in the next section , we will present a new solution to these equations .
the reader familiar with eq .
below , or not interested in its derivation , can safely skip to the next section . in the conventions of @xcite , the action for the bosonic sector of minimal five - dimensional gauged supergravity reads
s_bulk = .
[ sugraction ] here @xmath21 denotes the ricci scalar of the five - dimensional metric @xmath22 , and @xmath23 . , and the ricci tensor is @xmath24.[footnoteconventionscurvature ] ]
the graviphoton @xmath25 is an abelian gauge field with field strength @xmath26 .
moreover , @xmath27 is the five - dimensional newton constant and @xmath28 is a parameter with dimensions of length , related to the cosmological constant .
the equations of motion derived from ( [ sugraction ] ) read [ einsteineq ] r _ + 2 f_f^_+ g _ ( + f_f^ ) & = & 0 , + [ maxwelleq ] * f + f f & = & 0 .
for most of this paper we will work in lorentzian signature @xmath29 , although we will also discuss an analytic continuation later .
a solution is supersymmetric if there is a non - trivial dirac spinor @xmath30 satisfying the equation [ killingspeqdirac ] = 0 , where @xmath31 generate the clifford algebra cliff@xmath32 , so @xmath33 .
any such solution uplifts ( locally ) to a supersymmetric solution of type iib supergravity @xcite or of eleven - dimensional supergravity @xcite .
although global aspects of this uplift can be subtle @xcite , for the solution that we will discuss the gauge field @xmath25 will be a _ globally defined _ one - form on the five - dimensional space , and therefore there are no global obstructions to uplifting it to ten or eleven dimensions .
reference @xcite showed that solutions to ( [ killingspeqdirac ] ) possess a killing vector @xmath34 ( constructed as bilinear in the spinor @xmath30 ) , and fall into two classes , depending on whether this is everywhere null , or timelike at least in some open set .
we will focus on the timelike class , which is further specified by a khler metric @xmath35 on the four - dimensional `` base '' @xmath36 transverse to @xmath34 .
thus , in coordinates such that @xmath37 , the five - dimensional metric takes the form s^2 = - f^2 ( y + ) ^2 + f^-1s^2_b , [ generalggmetric ] where @xmath38 and @xmath39 are a positive function and a transverse one - form , respectively .
the base is characterized by a khler form @xmath40 , inducing a complex structure @xmath41 , and a complex two - form @xmath42 , which is @xmath43 with respect to the complex structure and satisfies , where @xmath44 is the ricci one - form potential .
namely , given the ricci curvature two - form @xmath45 , where @xmath46 is the riemann tensor of the khler metric , the ricci potential is defined by @xmath47 .
equivalently , splitting @xmath42 into its real and imaginary parts as , we have a triplet of real two - forms @xmath48 , @xmath49 .
choosing @xmath50 as orientation on @xmath36 , the @xmath51 are anti - self - dual : @xmath52 , where @xmath53 denotes the hodge star on @xmath36 .
also satisfy the algebraic relations . here
@xmath54 denote tangent space indices on the khler base . ] the gauge field of a supersymmetric solution in the timelike class is specified by the geometry and reads [ gaugefieldgg ] f = .
moreover , denoting by @xmath55 the ricci scalar of @xmath35 and introducing @xmath56 , supersymmetry also implies that [ relfr ] f^-1 = - r_b , and g^+ = - ( r - x^1 ) [ g+istracelessricci ] , while @xmath57 may be determined using the maxwell equation . , where @xmath58 is also defined in @xcite . in the specific case of interest for us , this is automatically satisfied . ] for more details we refer to @xcite .
similarly to @xcite , we will consider the following @xmath59 symmetric ansatz for the metric s^2_b = ^2 + a^2()(_1 ^
2+_2 ^ 2 ) + b^2()_3 ^ 2 , where @xmath60 are @xmath61 left - invariant one - forms parameterized by coordinates @xmath62 , and @xmath63 is a radial coordinate . specifically
, we define _ 1 & = & - + , + [ 1 mm ] _ 2 & = & + , + [ 1 mm ] _ 3 & = & + .[defleftinv1forms ] note that these satisfy @xmath64 , @xmath65 and @xmath66 .
the hat on @xmath67 ( and the consequent one on the left - invariant one - forms ) serves to distinguish this coordinate from a new one , @xmath68 , that will be introduced later . while at this stage @xmath67 should be regarded just as a local coordinate , @xmath68 will be an actual euler angle on @xmath11 .
the main motivation for working with this ansatz is that the supersymmetry equations reduce to a system of ode s , which ultimately can be expressed as a single sixth - order ode for one function @xcite .
as we are interested in alads solutions , @xmath63 is related ( but , as we will see , not identical ) to a fefferman
graham coordinate .
the boundary geometry will necessarily have locally @xmath69 isometry , therefore supersymmetry implies that the asymptotic metric and gauge field ( see section [ uvanalysis ] for their expression ) must be that of the example in section 4 of @xcite .
we fix the orientation on @xmath36 defining @xmath70 , while the five - dimensional orientation is given by @xmath71 . the @xmath59 invariant ansatz is completed by taking x^1 & = & - a^2 ( ) _ 1 _ 2 + b ( ) _ 3 , + [ 1 mm ] & = & a()(_1 + _ 2)(+ b()_3 ) , + [ 1 mm ] & = & ( ) _ 3 , + [ 1 mm ] p & = & p ( ) _ 3 , where the function @xmath72 is chosen positive .
the gauge field becomes [ gaugefieldgr ] f = .
the khler conditions then imply that @xmath73 and @xmath74 , where a prime denotes derivative with respect to @xmath63 , while eq .
yields f^-1 = [ 4 ( a)^3 + 7 a a a - a + a^2 a ] .[invfgeneral ] imposing implies that @xmath75 obeys the differential equation - = , [ psisusy ] where we defined g = - - 3 - + 4 . finally , combining the supersymmetry condition with the maxwell equation
, one can solve for @xmath75 in terms of @xmath76 , [ exprpsigeneral ] = - ( ^2 f^-1 + 8 ^-2 f^-2 - + f^-1 g ) , and eventually derive the equation governing @xmath76 @xcite : [ sixthordereq ] ( ^2 f^-1 + 8 ^-2f^-2 - + f^-1 g ) + = 0 , where @xmath77 is the laplacian on @xmath36 .
after using the expressions for @xmath38 and @xmath78 given above , one obtains a non - linear , sixth - order equation for @xmath72 with no explicit dependence on @xmath63 .
any solution to eq . gives a supersymmetric solution to minimal five - dimensional gauged supergravity , preserving at least one quarter of the supersymmetry , namely two real supercharges . in @xcite
, a simple family of solutions was found , [ grsolution ] a ( ) = ( / ) , where @xmath79 is a parameter . for @xmath80
this yields just ads@xmath0 ( of radius @xmath28 ) with a vanishing maxwell field , while for @xmath81 one obtains a charged , rotating black hole with a regular horizon . in order to obtain a geometry with a horizon , the authors of @xcite imposed specific boundary conditions on the function @xmath72 . here
we will impose different conditions , in particular we will require that in the interior the space closes off smoothly like @xmath82 , where the first factor is a time coordinate , precisely as in global ads@xmath0 .
another key difference with respect to the solution of @xcite is that the latter is asymptotically ads , with a conformally flat boundary @xmath83 , while our solution will be asymptotically _ locally _ ads , with a non conformally flat boundary comprising a squashed @xmath11 .
correspondingly , while the gauge field strength in @xcite vanishes asymptotically , in our case it will remain non - trivial at the boundary .
in this section we study solutions to the sixth - order equation .
we start by obtaining a general asymptotic solution satisfying the alads condition .
then we analyse the equation at finite @xmath63 and impose that the solution closes off smoothly in the interior .
this will give a new one - parameter family of solutions .
we will show that a solution connecting the asymptotic and the interior regions exists both by presenting a linearised solution around the ads@xmath0 background and by providing numerical evidence .
we will set @xmath84 for simplicity ; factors of @xmath28 can easily be restored by dimensional analysis ( in particular by sending @xmath85 , @xmath86 and @xmath87 ) . in the following , we study the sixth - order equation perturbatively at large positive @xmath63 ( i.e. in the `` uv region '' ) .
we assume an asymptotic expansion for the unknown function @xmath76 of the type a ( ) & = & a_0 e^ + [ 2 mm ] & = & a_0 e^ , [ uvexpansion ] with @xmath88 .
terms weighted by odd negative powers of @xmath89 could be included in the square bracket , but would be set to zero by the differential equation .
we solved the latter perturbatively up to order @xmath90 ( included ) , and found that there are five free coefficients that determine all the others . renaming them for convenience , these free coefficients are @xmath91 , @xmath92 , @xmath93 , @xmath94 and @xmath95 .
we display the first few terms of the solution : a ( ) & = & a_0 e^ + ( a_2 + c ) + ( a_4 + c-23 c^2 ^2 ) + & & + ( a_6 + c + [ 2 mm ] & & -c^2 ^ 2 + c^3 ^ 3 ) + o(e^-6 ) .[uvsola ] as discussed in the previous section , this determines completely the large-@xmath96 form of the five - dimensional metric and the gauge field @xmath97 . in the following we provide their expression at leading order . before presenting this
, we introduce new coordinates @xmath98 that will prove particularly natural from a boundary perspective .
these are given by [ changepsi ] = - t , y = t. then the five - dimensional metric takes the asymptotic form [ asymptoticmetric ] s^2 = ^2 + e^2 s^2_bdry+ , with boundary metric [ bdrymetric ] s^2_bdry = ( 2a_0)^2 , while the gauge field reads [ asymptotica ] a = a_bdry + o(e^- ) , with boundary value [ bdrya ] a_bdry = . here
, we have introduced the parameter [ defsquashing ] v^2 = 1-4c.the one - forms @xmath99 are defined in the same way as @xmath100 in , with @xmath101 replaced by the new coordinate @xmath102 .
we take @xmath103 , @xmath104 and @xmath105 , so that @xmath106 are euler angles on @xmath11 .
then from we see that the boundary is a direct product geometry including a squashed @xmath11 , with radius @xmath107 and squashing parameter @xmath108 .
namely , if we regard @xmath11 as a hopf fibration over @xmath109 , we have that the radius of the @xmath110 fiber generated by @xmath111 is rescaled with respect to the @xmath109 radius @xmath107 by a factor of @xmath108 .
we will denote this squashed three - sphere by @xmath112 .
so in these coordinates the boundary metric describes a direct product @xmath113 ; this will allow us to perform a simple wick - rotation later on by just analytically continuing @xmath114 .
note that we need @xmath115 in order to have a riemannian metric on @xmath11 and avoid closed timelike curves in the boundary .
when @xmath116 , then @xmath117 and there is no squashing ; namely , @xmath118 becomes conformally flat and describes an @xmath119 boundary with a round @xmath11 of radius @xmath107 .
correspondingly , all @xmath120 terms in vanish in this case
. however , these terms are crucial for our purposes , as we are interested in non conformally flat boundaries .
the only other free parameter of the uv solution appearing in the leading order expression of the supergravity fields is @xmath91 , which controls the overall size of the boundary .
since equation has no explicit dependence on @xmath63 , we could set @xmath91 to any chosen non - zero value by performing a constant shift of @xmath63 , followed by a redefinition of the other parameters in .
however , we will not fix this shift symmetry here as it will be convenient to use it in the next section , when we will study the solution in the interior .
an important comment about the gauge potential @xmath25 is in order .
this is determined by supersymmetry only up to a gauge transformation : from eq .
we see that a gauge shift @xmath121 just transforms the supersymmetry parameter as @xmath122 . in , , we chose the gauge by imposing @xmath123 and that the @xmath20 symmetry is respected ; then we fixed the residual freedom to shift @xmath25 by a term proportional to @xmath124 by imposing that the spinor @xmath30 is _ independent of the time coordinate _ @xmath114 .
this last property will be very important for us in the following .
solves equation . that the boundary spinor is constant with @xmath25 chosen as in follows from the discussion in ( *
* sect . 4 ) .
our gauge choice is different from the one of @xcite , which is @xmath125 . ]
it will prove crucial in section [ holorensection ] when we will evaluate the supergravity action on - shell , as the chern
simons term is not invariant under ( large ) gauge transformations on a space with boundary , hence the on - shell action changes when different gauge choices are made . in appendix [ detailssol ]
we provide more details about the five - dimensional uv solution , including the first few sub - leading terms of the metric and the gauge field , in which the remaining free parameters @xmath126 , @xmath127 , @xmath128 appear . in the same appendix
, we also show that the five - dimensional spacetime is alads by recasting the five - dimensional fields in fefferman graham form .
having determined the asymptotic behavior of the solution at large @xmath63 , we should now study how this ends in the interior ( i.e. the `` ir region '' ) .
we will require that the spacetime shrinks smoothly to zero size , with no horizon being formed . by exploiting the freedom to shift @xmath63
, we can assume with no loss of generality that this happens at @xmath129 .
we also assume that @xmath72 can be expanded in a taylor series around @xmath130 as [ iransatz ] a ( ) = a_0^ir + a_1^ir + a_2^ir ^2 + a_3^ir ^3 + .
although more general anstze may be considered , we will see that is enough to describe a new one - parameter family of regular solutions . since we need @xmath131 as @xmath129
, we will choose @xmath132 . then , expanding equation at small @xmath96 , we find that this requires @xmath133 , @xmath134 and a_4^ir = 0 .
we will choose @xmath135 .
and @xmath136 leads to an expansion involving both odd and even powers of @xmath63 .
the corresponding first terms in the expansion of the function @xmath38 read f ( ) = ^2 - a_3^ir ^4 - a_4^ir ^5 + o(^6),so @xmath137 when @xmath129 .
this is compatible with the presence of a horizon , because @xmath38 is the norm of the killing vector @xmath34 , and this should become null at the horizon .
the fact that @xmath138 remains finite is compatible with a horizon of finite size .
we will not study this any further in the present paper .
] at the following order in the @xmath63 expansion we get = 0 . vanishing of either the first or the second factor leads to two distinct families of solutions : if the first factor is zero then by going on in the perturbative expansion we reconstruct , namely the solution of @xcite , while if it is the second factor that vanishes we obtain a new family of solutions .
so in the following we fix @xmath139 and focus on the latter .
is straightforwardly related to the one with @xmath140 by sending @xmath141 , @xmath142 , @xmath143 .
] at higher orders , we find that all even coefficients @xmath144 are zero , while @xmath145 and @xmath146 are free parameters , and all the successive coefficients are determined ( at least up to the order @xmath147 , where we stopped the analysis ) .
actually , of the two free parameters @xmath145 and @xmath146 , only one is physical . to see this , note @xcite that under a rescaling of the coordinates @xmath148 , @xmath149 , a solution @xmath72
is transformed into another solution @xmath150 .
this also implies @xmath151 .
the transformation has the effect of rescaling @xmath145 by @xmath152 and @xmath146 by @xmath153 .
so we could rescale e.g. @xmath145 at will without changing the five - dimensional solution .
however , for the moment we keep @xmath145 arbitrary ; later on we will tune it in order to match the uv solution described in section [ uvanalysis ] , where the freedom to rescale @xmath63 has already been fixed by assuming that asymptotically the solution goes like @xmath154 rather than @xmath155 ( this also gives @xmath156 in the uv ) .
it will be convenient to trade @xmath146 for a new parameter , @xmath157 , invariant under rescaling of the radial coordinate : a_5^ir = ( 1 + 3)(a_3^ir)^2 , where the numerical factors are chosen for convenience .
thus @xmath157 parameterizes a one - parameter family of solutions .
we find that the ir solution is a double series in @xmath158 and @xmath157 , of the form : a ( ) = _ n 0_0k c_n , k ^k ^2n+1 .
we display the first few terms : [ irsolution ] a ( ) = .
if @xmath159 ( and choosing @xmath160 ) , matches the small @xmath63 expansion of the exact solution yielding ads@xmath0 , which is given by with @xmath80 . hence switching @xmath157 on corresponds to a deformation of ads@xmath0 , showing up in @xmath72 at order @xmath161 .
the ir solution for the function @xmath38 is [ solfir ] f ( ) = , and takes the form of an alternate series with even powers only : for @xmath162 , all the @xmath163 terms are positive while all the @xmath164 are negative . while we omit the expression of @xmath75 , it is instructive to look at the form of the five - dimensional metric in the ir : s^2 & = & 12(1 + 6 ^2)^2 + 12 ^2 + ^2 t_3 +
[ 2 mm ] & & - t^2 + .
[ favelesse ] here , both the uv parameter @xmath108 and the ir parameter @xmath145 are functions of @xmath157 .
the former appears because we are performing the change of coordinates setting the boundary metric in a direct product form .
we know that when @xmath159 , then @xmath165 and @xmath166 .
in the next section we will determine both @xmath167 and @xmath168 at linear order in @xmath157 .
we will see that the relation is such that at first order in @xmath157 the mixed term @xmath169 disappears .
we see that in the limit @xmath129 , there is a _ round _
@xmath11 shrinking to zero , while the @xmath170 component of the metric remains finite .
this implies that in a neighbourhood of @xmath171 the solution is non - singular and the topology of the space - time is @xmath172 .
the gauge field is also smooth as can be seen from its ir expansion a = .
[ favealburro ] in particular , notice that while the @xmath173 component is finite in the @xmath129 limit , the @xmath174 component vanishes smoothly at the origin of @xmath15 .
the linearised solution to be discussed next will demonstrate that the solution is smooth for any finite value of @xmath63 , at least when the deformation parameter @xmath175 is small , while the numerical analysis of section [ numerics_sec ] will provide evidence that this persists for all the allowed values of the deformation parameter . in this section
, we study the solution analytically at first order in an expansion around global ads@xmath0 .
this establishes the existence of a regular solution connecting the uv and ir asymptotics , at least when this is a small perturbation around ads@xmath0 . for small values of the parameter @xmath157
, our solution is a perturbation of the ads@xmath0 solution @xmath176 .
we can thus linearise the sixth - order equation around this background and solve for the perturbation .
of the six integration constants that arise , we find that three are fixed by regularity at @xmath177 and by the requirement @xmath178
. a fourth one is determined by imposing that at large @xmath63 the solution diverges as @xmath179 ( specifically , this removes a term going as @xmath180 ) .
so at the linearised level we obtain a two - parameter family of solutions deforming ads@xmath0 .
however , the ir analysis in the previous section demonstrates that this can not hold at the non - linear level , where only two distinct one - parameter families of solutions survive , the first being the one of @xcite , and the other being the new family parameterized by @xmath157 . in order to determine the latter
analytically at the linearised level , we match the two - parameter linearised solution with the ir solution parameterized by @xmath157 .
as we have imposed that the linearised solution goes like @xmath181 at large @xmath63 , the comparison also fixes the ir parameter @xmath145 ( see the discussion in the previous section ) .
we find : [ rela3irwithxi ] a_3^ir = ( 1 - ) + ( ^2 ) . in this way
, we obtain the solution [ solafirstorderxi ] a ( ) = ( 1 + ) + + ( ^2 ) .
it follows a particularly simple expression for @xmath38 : f = 1 + + ( ^2 ) , while the one for @xmath75 is : = -(1 + ) ^2- + o(^2 ) .
we also provide the corresponding expression of the five - dimensional supergravity fields . in the coordinates @xmath182 bringing the boundary metric to the direct product form @xmath183 ,
the bulk metric is [ metric5dlinearxi ] s^2 & = & ^2 - ^2 t^2 + [ 2 mm ] & + & ( _ 1 ^ 2+_2 ^ 2 ) + [ 2 mm ] & + & _ 3 ^ 2 + [ 2 mm ] & + & o(^2 ) , while the gauge field reads [ a5dlinearxi ] a = t - _ 3 + o(^2 ) .
these solve the equations of motion and at order @xmath184 .
moreover , one can see that near @xmath185 they reduce to the metric ( [ favelesse ] ) and gauge field ( [ favealburro ] ) , as required . for @xmath186
, we correctly retrieve the ads@xmath0 metric in global coordinates , s^2_ads = ^2 - ^2t^2 + ^2 ( _ 1 ^ 2 + _ 2 ^ 2 + _ 3 ^ 2 ) , together with a flat gauge field @xmath187 .. ] since @xmath188 , from the einstein equation we see that the ricci tensor of the metric satisfies @xmath189 so at linear order in @xmath157 this is still an einstein space , of the same einstein constant as ads@xmath0 .
however , the riemann tensor is different from the one of ads@xmath0 , and matches it only asymptotically , as required for an alads space .
we also checked that the five - dimensional space is not conformally flat , as the weyl tensor of is proportional to @xmath157 .
the boundary metric reads s^2_bdry & = & - t^2 + ( _ 1 ^ 2+_2 ^ 2 ) + & & + _ 3 ^ 2 + o(^2 ) . finally , expanding at large @xmath96 and comparing with the uv perturbative solution
, we can determine the relation between the uv parameters @xmath190 and the ir parameter @xmath157 , at linear order in @xmath157 .
we find : a_0 & = & 14 + ( 1 - 4 2 ) + ( ^2 ) , a_2 = - - ( 1 + 4 2 ) + ( ^2 ) , + [ 2 mm ] a_4 & = & ( - 2 ) + ( ^2 ) , a_6 = ( - 72 ) + ( ^2 ) , + c & = & + ( ^2 ) . [ uvparfromirpar ] it might be possible to solve the sixth - order equation beyond linear order , for example expressing the higher order terms in the perturbative expansion in a recursive fashion , similarly to the expansion of the solution of @xcite obtained in @xcite . in this respect , the linearised solution of this section is analogous the solution found in @xcite . in this section
we present a numerical study of the one - parameter family of solutions .
our primary scope is to show the existence of a regular solution connecting the ir and uv asymptotics beyond linear level in the deformation .
secondly , we wish to determine how the ir parameter @xmath157 and the uv parameters @xmath91 , @xmath126 , @xmath127 , @xmath128 , @xmath191 are related beyond the small @xmath157 approximation leading to .
let us briefly describe our procedure .
we fix the ir initial conditions around @xmath177 as in section [ sec : irsolution ] and integrate the sixth - order equation numerically towards the uv . to do this , we have to assign a value to the ir parameters @xmath157 and @xmath192 . as discussed in section [ sec : irsolution ] , only @xmath157 is physical , while @xmath192 can be changed by rescaling the radial coordinate , and will be fixed by requiring that asymptotically the solution goes like @xmath193 . at linear order in @xmath157
this is achieved by imposing , but beyond that we do not have an analytic expression and need to resort to numerics .
recalling the discussion in section [ sec : irsolution ] , @xmath192 can be fixed as follows : for any choice of @xmath157 , we integrate a first time choosing the reference value @xmath194 ; , yielding @xmath195 and therefore @xmath196 asymptotically . ] then from the uv behavior of the numerical solution we read off the rescaling @xmath197 to be performed so that @xmath198 ; this is most easily achieved by evaluating the function @xmath38 and finding the rescaling @xmath199 ensuring that it goes to 1 asymptotically .
then @xmath200 will produce the wanted uv behavior ; thus we fix this value and repeat the numerical integration .
the results we will present have this procedure implemented . recalling , the final value of @xmath145 can be read in the plots below from the behavior of @xmath38 at @xmath177 . to the sixth - order equation , rescaled by @xmath201 .
its square equals @xmath202 .
asymptotically , it gives the parameter @xmath91 , controlling the overall size of the boundary .
the different values of the ir parameter @xmath157 are indicated on the curves.,scaledwidth=60.0% ] and @xmath203 components of the metric .
asymptotically , this gives the value of the parameter @xmath204 , controlling the squashing of the boundary @xmath11.,scaledwidth=60.0% ] + and the uv parameter @xmath204 controlling the squashing of the boundary @xmath11 .
the squashing ranges between 0 and @xmath205 for the dots are effectively calculated values while the blue , continuous line is an interpolation .
the red , dashed line represents the relation obtained from the linearised analysis around the ads solution at @xmath186 ( which is denoted by the slightly larger , red dot).,scaledwidth=58.0% ] + in figure [ fig : aplot ] we display the solution @xmath72 to the sixth - order equation , for different choices of the ir parameter @xmath157 .
figures [ fig : squashingplot ] and [ fig : differentplotsmetricanda ] provide further plots representing components of the metric , written in the form s^2
= g _ ^2 + g _ ( _ 1 ^ 2 + _ 2 ^ 2 ) + g__3 ^ 2 + g_ttt^2 + 2 g_t _ 3 t , and of the gauge field a = a_t t + a _ _ 3 .
these are evaluated using and . and
@xmath173 we also need to know the value of the uv parameter @xmath204 as a function of @xmath157 ; this is determined as we will explain . ]
as one can see , the solution appears perfectly smooth within the numerical approximation . given a value of the ir parameter @xmath157
, we then proceed to compute the parameters @xmath91 , @xmath126 , @xmath127 , @xmath128 and @xmath206 controlling the uv solution .
this is done by matching the numerical solution @xmath76 obtained integrating from the ir with the uv solution given in , for some reasonably large value of the radial coordinate @xmath63 . in practice
, we compare the numerical and the uv solutions at several points in the interval @xmath207 , and determine the uv parameters via a best fit . the results are presented in figures [ fig : plotvsquaredvsxi ] and [ fig : otheruvparametersplots ] , where a comparison with the expressions from the linearised analysis around the ads solution at @xmath186 is also made .
figure [ fig : plotvsquaredvsxi ] displays the relation between the squashing parameter @xmath204 and @xmath157 .
one can see that for @xmath157 running between @xmath208 and @xmath209 , the squashing spans all values between @xmath210 and @xmath205 . from an ads / cft perspective , it is actually more natural to regard the family of solutions as parameterized by the boundary parameter @xmath204 rather than by the ir parameter @xmath157 .
for this reason , figure [ fig : otheruvparametersplots ] gives the other uv parameters as functions of @xmath204 .
we recall that although here @xmath91 is determined by @xmath204 , this is only because for ease of analysis we imposed that the solution ends at @xmath177 ; in fact , @xmath91 can be set to any desired non - zero value simply by shifting @xmath63 .
relevant physical properties of the solution can be inferred from the figures .
[ fig : squashingplot ] allows to exclude the existence of closed timelike curves .
these appear whenever the @xmath211 component of the metric , evaluated from , becomes negative .
however , from fig .
[ fig : squashingplot ] we see that as long as @xmath211 is positive asymptotically , it remains positive in the interior as well . , but @xmath203 can not be negative , see . ] imposing this uv boundary condition is the same as requiring that the boundary metric has one negative eigenvalue only , and is very natural from an ads / cft perspective .
we conclude that within this assumption the solution is free from closed timelike curves .
moreover , note from fig .
[ fig : gttplot ] that for sufficiently large @xmath157 there is a region in the bulk where the vector @xmath212 , generating time translations with respect to the asymptotic rest frame , becomes spacelike .
hence for sufficiently large @xmath157 the solution presents an ergoregion although there is no horizon . on the other hand
, we should note that the supersymmetric vector [ susyvector ] v = = + , associated to time translations with respect to a boosted asymptotic frame , is everywhere timelike in the bulk ( this follows from eq . and fig .
[ fig : fplot ] ) ; so if @xmath34 generates time translations then there is no ergoregion .
however , @xmath34 is null with respect to the boundary metric , so from an ads / cft perspective one would rather take @xmath212 as the generator of time translations both in the boundary and in the bulk .
similar features appear in the asymptotically ads black hole of @xcite .
the ir analysis has established a one - parameter family of solutions , labeled by the ir parameter @xmath157 .
the uv parameters controlling the asymptotic behavior of the solution can not remain arbitrary , and should be related between them and to @xmath157 . in section [ linearsec ]
we worked out this relation analytically at linear order in @xmath157 , while in section [ numerics_sec ] we studied it numerically . in the following ,
we make analytic progress beyond linear level by showing that a global property satisfied by our solution allows to determine one of the uv parameters in terms of the others .
let us consider the following integral on the three - sphere at infinity , @xmath213 , [ pagecharge ] _
s^3_bdry ( * _ 5 f + a f ) .
note that @xmath213 is the boundary of a cauchy surface ( namely , a hypersurface at constant @xmath114 ) @xmath214 that closes off smoothly in the interior ; using the ir expansions of section [ sec : irsolution ] one can also check that the integrand goes to zero smoothly on @xmath215 when @xmath216 . also recalling that @xmath25 is globally defined , we can apply stokes theorem and write _ s^3_bdry ( * _ 5 f + a f ) = _ c ( * _ 5 f + a f ) = 0 , where we used the maxwell equation .
we conclude that in our solution we have [ starftoawedgef ] _ s^3_bdry * _ 5 f = - _ s^3_bdry a f = - ^2 c^2 = - ( 1-v^2)^2 , where in the second equality we used the boundary gauge field given in , and in the third just the definition . is specified by taking the volume form of the unit , round three - sphere to be @xmath217 .
although this is a slightly unusual orientation , of course we have @xmath218 . ]
on the other hand , starting from the uv expansion of @xmath219 given in , we obtain an expression for @xmath220 that once compared with allows to solve for the parameter @xmath127 in terms of @xmath126 and @xmath221 .
we obtain [ sola4 ] a_4 = ( 5 + 64 a_2 - 256 a_2 ^ 2 + 32 c- 160 a_2 c- 104 c^2 ) . later on
we will be able to increase our analytic control on the uv behavior of the solution by also determining the parameter @xmath128 .
we note that the integral is a _ page charge _ and therefore should be quantized .
it is thus a natural notion of electric charge to be used when discussing black holes and other solitonic objects . in the next section
we will see that this is _ different _ from the notion of holographic charge obtained from integrating the holographic @xmath21-current , _ cf .
_ ( [ defelcharge ] ) below .
in this section we perform holographic renormalisation and compute the on - shell action as well as the holographic charges in terms of the squashing parameter @xmath108 .
the on - shell supergravity action is divergent due both to the infinite volume of the spacetime and to terms going to infinity while approaching the boundary .
the divergences can be removed implementing holographic renormalisation @xcite ( see @xcite for the inclusion of a maxwell field ) .
our alads spacetime @xmath222 can be seen as a foliation of timelike hypersurfaces homeomorphic to the boundary @xmath223 , labeled by the radial coordinate @xmath63 .
we denote by @xmath224 the hypersurface at fixed @xmath63 and by @xmath225 its interior region . if we write the five - dimensional metric as [ bulkmetric ] s^2 = f^-1 ( ) ^2 + h_ij(,x ) x^ix^j , with @xmath226 , then @xmath227 , is the induced metric on @xmath228 .
is obtained by contracting the bulk volume form with the unit , outward pointing vector @xmath229 , which yields @xmath230 , with @xmath231 .
[ orientationhypersurface ] ] holographic renormalisation regulates the action by evaluating it on @xmath225 , and prescribes the addition of suitable counterterms which cancel the divergences arising when @xmath224 is sent to the conformal boundary @xmath223 , so that the action remains finite in the limit .
the renormalised on - shell action is given by [ renormalizedaction ] s _ = _ ( s_bulk + s_gh + s_ct ) , where @xmath232 is the bulk supergravity action evaluated on - shell on @xmath225 .
@xmath233 is the gibbons hawking boundary term , needed to make the variational problem with dirichlet boundary conditions well - definite ; it reads s_gh = _ m _ ^4x k , where @xmath231 and @xmath234 is the trace of the extrinsic curvature @xmath235 of the hypersurface @xmath224 .
finally , the counterterm action @xmath236 is given by [ counterterm ] s_ct = -_m _ ^4x , where in this formula the ricci tensor @xmath237 and the ricci scalar @xmath21 are those of @xmath238 , and the indices are raised with @xmath239 .
note that in this section we are reinstating the ads radius @xmath28 .
the first two terms @xcite are local covariant expressions on @xmath224 , designed to cancel terms that diverge with a power - law ( when the solution is written in fefferman
graham coordinates ) .
the term quadratic in the curvatures @xcite only depends on the metric and gauge field at the conformal boundary and cancels possible `` logarithmic '' divergences , sub - leading terms . ] which are proportional to the weyl anomaly of the dual conformal field theory @xcite .
however , the results of @xcite imply that this anomaly and correspondingly the logarithmic divergence in the on - shell supergravity action vanishes in the case of interest for us ( see section 4 therein ) .
this can also be checked directly : from the expressions we find that asymptotically , which accounts for the different numerical factor in the equation below with respect to _ e.g. _ equation ( 1.4 ) in @xcite . ]
[ contribconformalanomaly ] h ( r_ijr^ij - r^2 ) = h f_ijf^ij
= ( 1-v^2)^2 , hence the gravitational and the gauge field contributions to the logarithmic term , though non - vanishing , cancel against each other in .
the full expression is nevertheless needed in order to derive the counterterms renormalising the energy - momentum tensor and the r - current to be introduced below .
the counterterms provide a `` minimal subtraction '' scheme ensuring that all divergences are cancelled .
however , one could include additional _ finite _ counterterms in @xmath236 , and these would affect the result for the on - shell action .
we postpone a discussion of these extra counterterms and the ensuing ambiguities to section [ fielddisc ] , when we will compare our gravity results with the field theory side .
let us now manipulate the five - dimensional integral computing the bulk action . although we do not have full analytic control on the solution , we will be able to evaluate @xmath232 exactly , by showing that it reduces to a boundary term . using the maxwell equation to eliminate the chern simons term and then the trace of the einstein equation , the bulk action can be written as s_bulk = -_m _ ^5 x g - _ m _ ( a_5f ) . therefore ,
when the solution is non - singular and @xmath240 goes to zero sufficiently fast in the ir , the last term reduces to an integral on the boundary @xmath224 . remarkably , we find that this is true for the first integral as well .
to see this , we need to recall some notions introduced in section [ susysec ] and exploit the following chain of equalities showing that the five - dimensional volume form is exact : ^5 x g = r_b t x^1x^1 = t r x^1 = - ( t p x^1 ) . in the first equality we used the general form of the five - dimensional metric together with and the fact that the volume form of the four - dimensional khler metric is related to the khler form @xmath40 by @xmath241 . the second equality is obtained noting that @xmath242 is the primitive part of the ricci form @xmath243 , namely the piece of @xmath243 that has non - zero wedging with @xmath40 .
the last expression follows from @xmath244 and @xmath245 .
if the solution is regular and @xmath246 goes to zero in the ir then we can apply stokes theorem and write the bulk on - shell action as [ sbulkasboundaryintegral ] s_bulk = _ m _ ( y p x^1 - 2 a_5f ) .
this formula holds for any regular supersymmetric solution to minimal five - dimensional gauged supergravity of the time - like class satisfying the conditions to apply stokes theorem .
this is the case for our solution , hence we can pass to evaluate @xmath232 explicitly .
the intermediate steps of the computation are more readable when presented in the fefferman
graham radial coordinate @xmath247 introduced in appendix [ detailssol ] .
we will also implement the change of coordinates . plugging the expansions in
, we find for the contribution from the five - dimensional volume ( up to terms vanishing in the limit @xmath248 ) : & & _ m_t p x^1 = - a^2 ( ) t + [ 2 mm ] & = & -t , + [ sbulk_vol5_fg ] where in the first equality we used the expressions of the ricci potential @xmath44 and khler form @xmath40 given in section [ susysec ] , while in the second equality we plugged the uv solution in and passed to the fefferman graham coordinate @xmath247 using .
since time translations are a symmetry of our solution , integration over time just gives an overall factor .
using the expansions , the gauge field contribution yields [ afcontribsbulk ] - _ m _
= - t. in both equations and we invoked the global constraint discussed in section [ sec : globalconstraint ] and used to eliminate @xmath127 .
note that both the logarithmic divergences and the terms proportional to @xmath126 cancel out when the two contributions to @xmath232 are added up , leaving just power - law divergences and finite terms depending on the boundary parameter @xmath249 .
the evaluation of the remaining contributions to the on - shell action just involves the uv solution and is thus straightforward .
the gibbons hawking term is found to contain only divergences , s_gh = - t , while the counterterms contain both a divergence and a finite piece , [ sct_fg ] s_ct = - ( c^2 + 12 a_0 ^ 4 r^4 ) t. adding up the contributions , and writing the result in terms of the squashing parameter @xmath191 , we obtain for the renormalised on - shell action [ resultsonshell ] s_ren = -t .
it is crucial to stress that the action is gauge - dependent , so this result is sensitive to the gauge chosen . from and we see that under a gauge transformation @xmath250 , where @xmath251 is a constant ( possibly depending on @xmath108 ) , the on - shell action changes by [ gaugeshiftsonshell ] s_ren = - t _ s^3_bdry * _ 5f = a_t ( v^2 - 1)^2 t .
our motivation for fixing the gauge as in is that with this choice the supersymmetry parameter does not depend on the time coordinate @xmath114 ( neither on the boundary , nor in the bulk ) when the obvious left - invariant frame following from is used .
in particular , note that the term in @xmath252 in , which will play an important role in section [ fielddisc ] , directly comes from the gauge field contribution
. of course , the lorentzian time should be non - compact and therefore @xmath253 in the formulae above is just a formal writing .
however after performing an analytic continuation @xmath254 we can compactify the time coordinate . then the boundary topology becomes @xmath6 , the boundary metric becomes euclidean and the boundary gauge field acquires an imaginary , flat component .
we remark that both the bulk metric and the bulk gauge field become complex , however this does not affect the on - shell action , which remains real . denoting by @xmath255 the
analytically continued on - shell action , and by @xmath256 the finite period of @xmath257 , we obtain by requiring that this is related to the analytic continuation of the lorentzian action as @xmath258 .
note that we regard the lorentzian time as @xmath259 and the analytically continued time as @xmath260 , so the lorentzian volume form on @xmath224 is related to the analytically continued one as @xmath261 . ] [ resultsonshelleucl ] i = .
this exact expression for the renormalised on - shell action in terms of the squashing of the boundary @xmath112 is a main result of this paper .
the regularised action @xmath262 gives rise to the quasi - local energy - momentum tensor [ defenmom ] t_ij = - . in order to obtain the dual field theory energy - momentum tensor @xmath263 , one needs to rescale this before taking the limit @xmath264 , because in alads space the metric of the background on which the field theory lives is multiplied by the divergent conformal factor @xmath265 .
this gives the formula [ holoenmomtensor ] t_ij & = & - _
e^2/ , where the curvatures are computed and the indices are raised with the induced metric @xmath238 , and @xmath266 denotes the bach tensor ( see eq . ) . with respect to the metric
this appears noting that in @xmath267 , and recalling that the variation of the euler density @xmath268 with respect to the metric vanishes .
one may expect the term @xmath269 to appear from the variation of @xmath270 ; this indeed shows up by using the identity @xmath271 . ] the first line yields a finite contribution , while the second line just cancels `` logarithmic '' divergences , here going as @xmath272 .
in addition to the energy - momentum tensor , we can also introduce a current by varying the regularised action with respect to the gauge field @xmath273 at the boundary @xmath228 : j^i = .
the possible contributions arise from the bulk action and from the counterterm action . varying the bulk action with respect to @xmath25 and using the maxwell equation to eliminate the bulk integral we are left with the boundary integral s_bulk = - _ m _ a ( * _ 5 f + a f ) .
note that here we are assuming that the spacetime is smooth and that @xmath224 is the only boundary .
the counterterm contribution cancels a `` logarithmic '' divergence arising from @xmath219 .
since the graviphoton @xmath25 couples to the dual field theory r - symmetry current , sending @xmath264 after appropriate rescaling yields a finite expectation value @xmath274 for the latter .
thus we arrive at the expression [ expressionrcurrent ] j^i = _ = _ e^4/\ { * _ 4 - _ j f^ji } , where @xmath275 is computed using the induced metric @xmath238 and it is understood that @xmath276 is restricted to the boundary @xmath277 .
the last term comes from the variation of @xmath236 in . from the holographic energy - momentum tensor and r - current one
can construct a set of conserved charges . in order to do so
, one chooses a spacelike , compact hypersurface @xmath278 in the boundary @xmath223 , with metric @xmath279 , @xmath280 , and writes the boundary metric in adm form : s^2_bdry = - n^2t^2 + _
( x^+ n^t ) ( x^+ n^t).[bdrymetricadm ] let @xmath281 be the unit , timelike vector orthogonal to @xmath278 .
as a one - form , it reads @xmath282 .
then from any killing vector @xmath283 of the boundary metric one can construct a conserved charge as [ conservedqgeneral ] q_z = _ ^3x u^i t_ij z^j .
moreover , from the current @xmath284 one obtains an electric charge @xmath285 , corresponding to the vev of the dual field theory r - charge operator , by evaluating the integral q = _ ^3 x u_i j^i . assuming that @xmath286 on @xmath224 , so that the divergent pieces drop from ( this is satisfied in the case of interest for us ) , it is straightforward to see that q & = & _ ^3 x u_i j^i = _ ( * _ 5 f + a f).[defelcharge ] the charge @xmath285 is invariant under small gauge transformations of @xmath25 , and it is conserved provided the r - current is non - anomalous , which is true in the case we are studying here . on the boundary @xcite .
] however , we remark that this is different from the quantized page charge , as we already remarked in section [ sec : globalconstraint ] .
the two definitions coincide only when the chern
simons contribution @xmath287 vanishes .
we now compute the conserved charges for our solution .
the spacelike boundary hypersurface @xmath278 is the squashed three - sphere @xmath213 , and from the boundary metric we see that @xmath288 to compute the holographic r - charge we can take advantage of formula ; then gives [ resultq ] q = - _ s^3_bdry a f = - ( v^2 - 1)^2 .
the other relevant charges are the _ total energy _
@xmath289 , associated with the generator of time translations @xmath212 , [ formulaformass ] e
= _ s^3_bdry u^it_it vol(s^3_bdry ) , and the angular momentum @xmath290 for the @xmath110 right - isometry generated by @xmath291 , [ defangmom ] j = _ s^3_bdry u^it_i vol(s^3_bdry ) , while one can check that the charges associated with the @xmath61 left - isometries vanish .
note that @xmath292 and @xmath290 are defined also on non conformally flat boundaries , differently from the ashtekar
das conserved quantities @xcite . @xmath289 and @xmath290 can be evaluated by inserting the uv solution in eq . for @xmath293 , and performing the integrals above .
this yields expressions depending on the uv parameters on @xmath126 , @xmath128 and @xmath204 ( recall that @xmath127 was determined in as a consequence of the smoothness of the solution ) .
ultimately all charges have to be functions of a single independent parameter of the solution , that we identify with the squashing @xmath108 .
although we do not have full analytical control on the relations between these uv parameters , in a moment we will introduce an extra information allowing to deduce the exact expressions of the charges in terms of @xmath108 . already at this stage , by using the values of the on - shell action in ( [ resultsonshelleucl ] ) , the charge @xmath285 in ( [ resultq ] ) , and the energy - momentum tensor obtained from ( [ holoenmomtensor ] ) , one can check that the following relation is satisfied : ) one obtains @xmath294 as a function of @xmath295 ( and @xmath91 ) that we have not presented .
the point is that after using the dependence on @xmath296 drops out of the combination @xmath297 . in lorentzian signature @xmath256
denotes simply a time interval . ] [ bpsrelation ] e+ j + q = .
we note that the combination @xmath298 is the holographic charge @xmath299 associated with the supersymmetric killing vector @xmath34 given in , and that the coefficient in front of @xmath285 is the boundary value of @xmath300 .
so can be written as [ smarrgeneral ] q_v + ( v^a_^bdry ) q = . recalling and
, we observe that under a gauge transformation the on - shell action ( that we can define formally in any gauge , provided we do not compactify the time direction ) shifts by a term proportional to the electric charge : @xmath301 .
this implies that the relation above remains valid independently of the gauge chosen .
note that in the gauge @xmath302 , the relation is satisfied by the black hole of @xcite , with @xmath303 identified with the ashtekar
das mass @xcite .
in order to obtain @xmath289 and @xmath290 as functions of @xmath108 we will exploit the fact that for any continuous parameter @xmath304 of the _ boundary _ geometry , one can derive a corresponding `` ward identity '' by simply applying the chain rule @xcite s_ren & = & _ m ^4 x ( - t_ij + j^i ) , [ wardidentitygeneral ] where @xmath305 is the finite metric on the conformal boundary : @xmath306 .
the parameter we will to consider is the one that can be introduced in the boundary metric and gauge field by rescaling the coordinate @xmath114 as @xmath307 . varying with respect to @xmath308 we get t_ij = u^i t_it , j^i = - u_i j^i a_t .
so gives s_ren = t _ s^3_bdry vol(s^3_bdry ) ( -u^i t_it - u_i j^i a_t ) = -(e+ a_t q ) t . recalling that @xmath309 , and turning to the analytically continued on - shell action ( so that @xmath310 becomes @xmath311 ) , we arrive at the relation [ relfromchainrule ] e+ q = .
denoting @xmath312 , this can be re - written in the covariant form [ relfromchainruleabstract ] q_u + ( u^a_^bdry ) q = , which is manifestly gauge - invariant .
one can also consider a variation of the on - shell action with respect to the parameter that can be introduced in the background fields by rescaling the original time coordinate @xmath313 .
this yields precisely the relation , thus providing a general way to derive it . in section [ algebsection ]
we will discuss the field theory interpretation of the conserved charges and of this relation . finally , one can check that varying with respect to the squashing @xmath204 does not yield new information . inserting the expressions for @xmath285 and for the action in
, we obtain the total energy as a function of the squashing parameter only : [ resultmass ] e = ( + - v^2 + v^4 ) .
we can compare this expression with the one obtained from the definition using the holographic energy - momentum tensor , which depends on the uv parameters @xmath126 , @xmath128 and @xmath204 . in this way
, we obtain an equation that determines @xmath128 , see eq . in the appendix for the explicit expression .
the angular momentum can now be deduced from , which yields the following simple expression : [ resultj ] j = ( v^2 -1)^3 .
we have thus obtained all relevant charges as exact expressions of the squashing parameter @xmath108 .
a check of our results so far is provided by the limit @xmath314 , in which the boundary three - sphere @xmath112 becomes round , and the bulk solution is just ads@xmath0 . in this case , @xmath285 and @xmath290 vanish , while @xmath289 and @xmath255 remain finite and reproduce the known expressions for pure ads@xmath0 , namely @xmath315 and @xmath316 @xcite . for completeness , in appendix [ enmomtandrcurrent ]
we give the explicit expressions for the energy - momentum tensor and for the r - symmetry current . before moving on to discussing the dual field theories
, it may be useful to provide a summary of the one - parameter family of solutions that we constructed . in section [ uvanalysis ]
( see also appendix [ detailssol ] ) , we determined the general asymptotic expansion of an alads solution to the ode ( [ sixthordereq ] ) governing @xmath20 invariant supersymmetric solutions to minimal five - dimensional gauged supergravity .
this expansion is characterized by five parameters @xmath91 , @xmath126 , @xmath127 , @xmath128 and @xmath108 .
two of them , @xmath91 and @xmath108 , appear explicitly in the boundary metric ; this describes the direct product @xmath317 with overall size proportional to @xmath91 , where @xmath318 is a three - sphere with @xmath4 isometry and squashing parameter @xmath108 . for @xmath319
this is a non conformally flat boundary and , correspondingly , the boundary gauge field is non - trivial .
the three remaining uv parameters @xmath126 , @xmath127 , @xmath128 appear in sub - leading components of the metric and of the gauge field . the ir analysis of section [ sec : irsolution ] established the existence of a one - parameter family of solutions closing off smoothly in the interior . by means of a linearised and a numerical analysis ( sections [ linearsec ] and [ numerics_sec ] respectively ) we showed that this parameter can be directly related to the squashing @xmath108 at the boundary ( see fig.@xmath320[fig : plotvsquaredvsxi ] ) , and that for any value of the solution is everywhere regular and free from closed time - like curves ( figs.@xmath320[fig : aplot][fig : differentplotsmetricanda ] ) .
in addition , we determined the relations between the different uv parameters corresponding to our one - parameter family ( fig.@xmath320[fig : otheruvparametersplots ] ) . for two of the uv parameters , namely @xmath127 and @xmath128
, we obtained an analytic expression in terms of @xmath108 and @xmath126 . and
@xmath321 , given in ( [ sola6 ] ) , with the numerical results , finding perfect agreement . ]
recalling that @xmath91 can be chosen freely by shifting the radial coordinate , the only uv parameter that we could not determine analytically as a function of @xmath322 is @xmath126 ; however , this drops out of the on - shell action as well as of all the holographic charges . despite the lack of full analytic control on the solution , in the present section we evaluated the on - shell action as well as the holographic conserved charges exactly , in terms of @xmath108 .
the on - shell action could be evaluated by showing that it reduces to a boundary integral , and is given in eq . .
we remarked that this is a gauge - dependent result , and gave it in a gauge such that the supersymmetry parameter is independent of the time coordinate @xmath114 . on the other hand , the holographic charges are gauge - independent , and the relevant ones are the energy @xmath289 , the angular momentum @xmath290 and the electric charge @xmath285 .
their evaluation involved a ward identity ( following simply from applying the chain rule to the variation of the renormalised on - shell action ) , and their expressions are given in , and .
in this section we discuss the field theory duals to our supergravity solution .
we first recall some aspects of rigid superymmetry at the conformal boundary and its relation with supersymmetry in the bulk , and then discuss the relation of the localised partition function computed in our background with the supersymmetric index and the casimir energy .
we will end comparing the latter with the holographically renormalised on - shell gravity action . here
we will assume that the five - dimensional solution is uplifted @xcite to a ten - dimensional solution @xmath323 of type iib supergravity , where @xmath324 is a sasaki einstein manifold .
in this case , when the squashing parameter is trivial so that @xmath325 ads@xmath0 ( in global coordinates ) , the field theory dual is generically an @xmath2 superconformal quiver gauge theory , with matter in bi - fundamental or adjoint representations of the gauge group .
for example , one can consider @xmath326 super yang
mills @xcite , the klebanov
witten quiver gauge theory @xcite , or the @xmath327 family of quivers @xcite .
our solution should then correspond to a relevant deformation of the flat - space lagrangian of the given field theory , obtained through couplings to a non - trivial background metric and a gauge field sourcing the r - symmetry current @xcite . in the following we discuss some general properties of the superconformal field theories living at the non conformally flat boundary of our solution .
we already mentioned that the weyl invariance and r - symmetry are preserved also at the quantum level , since the associated anomalies vanish on our background @xcite . here
we focus on the algebra associated with the supersymmetry being preserved .
for definiteness we work in lorentzian signature , however most of what we will discuss can be repeated in euclidean signature .
a superconformal field theory can be coupled to curved space by considering conformal supergravity and freezing the background . though this is a general procedure , here we are interested in the @xmath328 case with a minimal set of background fields .
case . ] the field theory energy - momentum tensor couples to the background metric while the @xmath110 r - current couples to the gauge field sitting in the gravity multiplet of conformal supergravity , that we will denote by @xmath329 . for the background to be bosonic the gravitino must be set to zero , and imposing that its supersymmetry variation be zero too yields the differential equation ( * ? ? ?
* eq.@xmath320(16.10 ) ) _ i^a _ + - _ i _ - = 0 , where @xmath330 , and @xmath331 is the levi - civita connection constructed with the background metric . here , @xmath332 is the positive - chirality part of the majorana spinor parameter associated with the @xmath285-supersymmetry transformations of conformal supergravity , while @xmath333 is the negative - chirality part of the parameter of special @xmath334-transformations
. these can be disentangled by taking the gamma - trace of the equation above : [ susycond_csg ] _
+ = _ i ^j_j^a _ + , _ - = ^i_i^a _ + .
therefore we see that the supersymmetry preserved on a bosonic background of conformal supergravity generically is a combination of a @xmath285- and an @xmath334-transformation .
the first equation in is a charged version of the _ conformal killing spinor ( cks ) equation_. in lorentzian signature , it admits a solution with no zeros if and only if the four - dimensional background has a null conformal killing vector @xmath335 @xcite . this can be written as a spinor bilinear as [ defz ] z^i = 14 ^i _ + . in euclidean signature
, the charged cks equation has been studied in @xcite and leads to the requirement that the manifold be hermitian .
precisely the same conditions are retrieved in a holographic setup @xcite .
in fact , one can show that on alads spaces the supersymmetry transformations of five - dimensional minimal gauged supergravity reduce at the boundary to the transformations of conformal supergravity @xcite .
so the fields and supersymmetry parameter of any supersymmetric alads solution satisfy asymptotically , with @xmath336 and @xmath337 corresponding to the leading and first sub - leading components of the asymptotic five - dimensional supersymmetry parameter .
the conformal supergravity gauge field @xmath329 is identified with the boundary value of the bulk gauge field @xmath25 appearing in the previous sections ; in our normalizations , we have @xmath338 .
the bulk supersymmetry time - like killing vector @xmath34 introduced in section [ susysec ] reduces to the null ( conformal ) killing vector @xmath335 on the boundary ; see @xcite for more details on the dictionary between bulk and boundary quantities .
given a solution @xmath332 to the cks equation and its charge conjugate @xmath339 , the associated superalgebra can be obtained from the conformal supergravity algebra , which gives and a special supersymmetry @xmath334 .
our conformal supergravity gauge field is related to the one of @xcite as @xmath340 . ]
[ superalgebracks ] [ _ _ + , _ _ - ] = 2 + , where @xmath341 is a generic field in the theory , @xmath247 is its r - charge , @xmath342 is the lie derivative along @xmath335 and the ellipsis denotes a scale transformation and a special conformal transformation proportional to the imaginary part of the complex one - form @xmath343 .
the latter is defined through [ defvnm ] ^i _ i^a _ + = 2 _ i ^i _ + , where we are assuming that the spinor has no zeros so that the definition holds everywhere .
in fact , the imaginary part of @xmath343 is related to the failure of the conformal killing vector @xmath335 to be killing @xcite .
so when @xmath335 satisfies the killing condition @xmath344 , @xmath345 is real and the terms in the ellipsis in vanish .
locally , a conformal killing vector can always be made killing by a conformal transformation of the metric ( which in a holographic setup is implemented by a change of radial coordinate ) .
when @xmath335 is a killing vector , one obtains the same algebra by coupling the field theory to the _ new minimal _ formulation of off - shell supergravity . indeed , modulo the conformal transformation just discussed , the cks equation in is equivalent to the supersymmetry condition that is obtained from the rigid limit @xcite of new minimal supergravity @xcite .
this reads ( _ i - a^_i + v^_i + v^_j ^j_i ) _
+ = 0 , [ nmkse ] where the gauge field @xmath346 and the global one - form @xmath347 , satisfying @xmath348 , constitute the new minimal set of auxiliary fields .
the equivalence is easily seen by identifying = v^,a^ = a^- v^ , see @xcite for more details .
the definition leaves the component of @xmath343 along @xmath335 undetermined , and this can be used to arrange for @xmath349 . in the new minimal variables , ( [ superalgebracks ] )
reads simply [ superalgebranm ] [ _ _ + , _
_ - ] = 2 ( l_z - r z^i a^_i ) , in agreement with @xcite .
this equivalence with new minimal supergravity may be useful for concrete field theory computations in ads / cft context , as typically the superconformal field theories dual to gravity solutions are defined as the strongly coupled ir fixed point of a lagrangian theory in the uv , which is non - conformal .
therefore , as long as one is interested in supersymmetric quantities that are independent of the coupling , one may take advantage of the new minimal formulation of rigid supersymmetry in order to couple the uv lagrangian to the background of interest .
let us now focus on the @xmath350 background defined at the boundary of our bulk supergravity solution .
working in the frame e^0 = t , e^1 = a_0 _ 1 , e^2 = a_0_2 , e^3 = a_0 v_3 , for the boundary metric , one can see that the cks equation is solved by a constant spinor @xmath332 satisfying the projection @xmath351 , with ( recall ) [ acs ] a^ = - a_bdry = - t -(v^2 - 1 ) _ 3 .
the new minimal equation is solved by the same @xmath332 , fixing the background fields as v^&= & _ 3 + z , + [ 2 mm ] a^&= & a^+ v^ = - t + ( 2+v^2 ) _ 3 + z,[vnmanm ] where @xmath352 parameterizes the part of @xmath353 and @xmath354 left undetermined by the equation , and can be any function consistent with @xmath348 . from it
follows that the vector @xmath335 is killing and can be written as z = + , where we have conveniently fixed the complex constant parameterizing @xmath332 . as a one - form
, @xmath335 reads z = ( 2a_0)^2 ( -t + _ 3 ) . noting that @xmath355
, we see that the supersymmetry algebra depends manifestly on the gauge choice of @xmath329 ( but does not depend on @xmath352 ) , through the term @xmath356 .
we now illustrate some distinguished choices of gauge .
the following parallels a discussion in @xcite , extending it from the round case to our @xmath357 geometry .
we start from the gauge adopted so far , in which @xmath329 is given by . denoting by @xmath358 , @xmath359 the supercharges associated with @xmath332 , @xmath360 , and by @xmath361 , @xmath290 and @xmath21 the abstract operators associated to @xmath362 , @xmath363 and r - symmetry transformations , respectively , in this case the algebra takes the form [ abstractsuperalgebra ] \ { q_+ , q_- } = h + j - r. the commutators of @xmath364 with @xmath361 , @xmath290 and @xmath21 can be inferred from the fact that the conformal killing spinor @xmath332 has r - charge @xmath365 and satisfies along a vector @xmath366 is defined by @xmath367 l _ _
+ = 0,l _ _ + = _ + .
therefore we have @xmath368 = \pm\mathcal q_\pm\,$ ] , @xmath369 = \mp\frac{1}{2 } \mathcal q_\pm\,$ ] and @xmath370 = 0\,$ ] .
these clearly identify @xmath290 as the generator of right @xmath110 angular momentum , and @xmath21 as the generator of @xmath371 . the fact that @xmath361 commutes with the supercharges identifies this uniquely as the operator relevant for the euclidean path integral . in the round limit @xmath372 , this is a two supercharges sub - algebra of the one given in eq .
( 5.9 ) of @xcite .
a second special gauge is @xmath373 , which can be reached by implementing the gauge transformation @xmath374 .
the conformal killing spinor acquires a phase @xmath375 , inducing the same dependence of the supercharges on @xmath114 .
this modifies the commutator with the generator of time translations and changes the factor in front of @xmath21 in .
namely , denoting now by @xmath376 the operator associated with @xmath377 , the algebra becomes @xmath378 = \mp \frac{3{v}^2 - 2}{2{v}^2 } \mathcal q_\pm$ ] and \
{ q_+ , q_- } = + j - r , [ abstractsuperalgebra2 ] with the other commutation relations unchanged
. for @xmath165 , this reduces to a sub - algebra of the one in eq .
( 6.11 ) of @xcite , and can be embedded in the superconformal algebra on @xmath379 given _ e.g. _ in @xcite , where @xmath376 is identified with the conformal hamiltonian generating dilatations .
a third special gauge choice is @xmath380 ( that is @xmath381 ) , which can be reached by making the gauge transformation @xmath382 , starting from ( [ acs ] ) .
then the spinor acquires a phase @xmath383 ; denoting by @xmath384 the generator associated with @xmath385 , the modified commutation relations are @xmath386 = \pm\frac{1}{{v}^2}\mathcal q_\pm$ ] and \
{ q_+ , q_- } = p_0 + j , [ abstractsuperalgebra3 ] which for @xmath372 reduces to a sub - algebra of the one in eq .
( 5.6 ) of @xcite .
as all the anticommutators in ( [ abstractsuperalgebra ] ) , ( [ abstractsuperalgebra2 ] ) , ( [ abstractsuperalgebra3 ] ) vanish on a supersymmetric state , each of them yields a bps relation .
however , noting that the algebras can be mapped into each other by identifying [ relationshamiltonians ] h
-r = p_0 = - r , we see that these are all equivalent , for example to h + j - r & = & 0 .
[ bpscond ] when @xmath372 , this is precisely the shortening condition obeyed by the multiplets contributing to the supersymmetric index @xcite . inserting either of the anticommutators in the path integral one notices that ( at least formally )
, the bps condition ( [ bpscond ] ) is obeyed by the vacuum expectation values of the operators , _
i.e. _ @xmath387 , etcetera .
we can therefore attempt to compare this relation with the relation among the holographic charges that we derived in section [ sec : holocharges ] .
since here we are interpreting the holographic charges as expectation values of operators in the field theory , we will denote them as j = ( v^2 -1)^3 , r = - q = ( v^2 - 1)^2 .
the interpretation of the holographic total energy @xmath292 is somewhat more subtle , as this contains the expectation value of a hamiltonian operator plus a contribution from the vacuum energy .
as noted around equation ( [ smarrgeneral ] ) , this is most familiar in the context of asymptotically ads black holes @xcite , where the non - vacuum energy can be interpreted as the ashtekar
although for general alads solutions with non conformally flat boundary we do not have such a definition of mass , we will use the black hole example as a guide for interpreting the total energy @xmath292 in our case .
we begin writing the relation ( [ bpsrelation ] ) in the form e + j = - ( z a^ ) r , [ comparetm ] where the left hand side is manifestly gauge - invariant , and so is the right hand side ( albeit less explicitly ) .
since @xmath388 is a gauge - invariant energy , and does not contain the contribution of the on - shell action ( which will be discussed in the next subsection ) , it is natural to identify it with the vacuum expectation value of a hamiltonian operator . comparing ( [ comparetm ] ) with the field theory bps relation , we see that the natural identification is p_0 = e - + ( z a^ ) r , or equivalently = e - + ( z a^ ) r . with this identification
the field theory bps relation and relation ( [ bpsrelation ] ) match .
the reason why we have not referred to as `` bps relation '' the latter is that we have not derived it from the supersymmetry algebra of gauged supergravity .
the black hole solution @xcite obeys the bps relation m _ + j- r & = & 0 , extracted from the ordinary ads@xmath0 superconformal algebra , which is presumably valid for any asymptotically ads solution . therefore the ashtekar
das mass @xmath389 may be identified as the vev of the conformal hamiltonian @xmath390 . although we noted that relation reduces to this in the limit @xmath391 and in the gauge @xmath392 , this is only formal , as in the case @xmath391 our solution reduces to pure ads@xmath0 , so that @xmath393 .
nevertheless , we expect that the generalized ( _ i.e. _ @xmath394-deformed ) bps relation ( [ bpscond ] ) can be derived from the supersymmetry algebra of five - dimensional gauged supergravity , analysed in the context of alads solutions .
the explicit expression of @xmath390 in terms of @xmath108 reads = ( v^2 - 1)^2(2+v^2 ) , with the expressions for @xmath395 and @xmath396 following from .
these , however , are predictions for ( universal ) one - point functions in strongly coupled superconformal field theories at large @xmath8 , which are not easily computable . in the next section we will turn our attention to the on - shell action in the specific gauge leading to the algebra ; we will interpret it as a casimir energy , and then compare it with a corresponding quantity on the field theory side . in this section
we wish to discuss the field theories from the point of view of the path integral , therefore we will concentrate on the euclidean signature version of the background , with periodically identified time . in order to have well - defined killing spinors ( at the boundary as well as in the bulk ) we must fix the gauge in which these are time - independent ,
namely @xmath329 is the one in , thus justifying _ a posteriori _ the choice of gauge for the graviphoton field @xmath397 made in section [ uvanalysis ] . in this case the generator of time translations in the field theory @xmath361 commutes with the supercharges , @xmath398=0\,$ ] .
the ads / cft master equation @xcite relates the holographically renormalised supergravity action evaluated on a classical five - dimensional geometry @xmath399 to the euclidean path integral in the field theory defined on the four - dimensional boundary @xmath400 , namely e^-s _ [ m_5 ] & = & z _ [ m_4 ] n .
[ master ] in particular , when @xmath401 ( topologically ) the path integral on @xmath402 is performed with _
periodic _ boundary conditions for the fermions on @xmath403 , and therefore it is related to the trace over all the states of the field theory @xcite as z _ [ m_4 ] & & [ ( -1)^f e^-h ] i ( ) , [ notquite ] where @xmath97 is the fermion number , @xmath18 is the circumference of @xmath403 , and @xmath361 is in general an operator commuting with the supercharges preserved by the background . let us first consider the case @xmath404 . then @xmath361 is the hamiltonian commuting with the supercharges defined on @xmath405 , and the trace is the supersymmetric index , which may also be defined in radial quantisation @xcite .
casimir energy _ of a quantum field theory on @xmath11 may be defined from the limit of large @xmath406 radius of the partition function as e _ & & - _ z _ [ s^1_s^3 _ ] .
[ usualcasimir ] in the semiclassical gravity approximation , when we can use ( [ master ] ) , the relation @xmath407= \beta { e_\mathrm{casimir}}$ ] is valid for any @xmath18 ; when one introduces additional background fields , as long as translation invariance along the @xmath403 is preserved , the renormalised on - shell action is guaranteed to remain linear in @xmath18 @xcite . however , while in the large @xmath8 limit @xmath408 , the supersymmetric index @xmath409 is of order @xmath9 @xcite , which appears to contradict the relation ( [ master ] ) together with ( [ notquite ] ) .
the caveat is that the precise version of ( [ notquite ] ) includes a factor @xmath410 , often omitted in the literature z _ [ s^1_s^3 _ ] & = & e^-f ( ) .
this implies that , at least for field theories with a freund rubin gravity dual in type iib supergravity with @xmath8 units of five - form flux , we must have @xmath411 , and using ( [ master ] ) we see that ) has been argued in @xcite to be proportional to the anomaly coefficient @xmath412 . ]
f = s _ = e_n . in a related context , in a calculation of the partition function of a weakly coupled gauge theory on @xmath413with _ anti - periodic _ boundary conditions on @xmath403 , in @xcite
it was shown that @xmath410 is the usual casimir energy of the field theory on @xmath11 .
we shall return to the interpretation of @xmath410 for the path integral with supersymmetric boundary conditions momentarily .
more generally , one can introduce a refined supersymmetric index with _
fugacities _ turned on @xcite , which takes the form ( p , q ) & = & [ ( -1)^f p^j+j- q^j - j- ] , where @xmath414 are complex parameters and @xmath415 denote the cartan generators of @xmath61 and @xmath416 , respectively , and @xmath21 is the r - charge .
this receives contributions only from states obeying the relation @xmath417 , and hence it reduces to the unrefined index @xmath409 upon setting @xmath418 . in @xcite
it has been argued that this index is computed by the path integral on a compact hermitian manifold @xmath419 homeomorphic to @xmath6 , namely a ( primary ) _ hopf surface _
thus z _ [ h_p , q ] & = & e^-f ( p , q ) , where @xmath410 may depend on the complex structure of the manifold , as well as on counterterms local in the background fields .
let us briefly recall the definition of a hopf surface @xmath419 and make contact with our boundary four - dimensional geometry .
a complex manifold homeomorphic to @xmath3 is necessarily a primary hopf surface @xcite , namely a quotient of @xmath420 by an infinite cyclic group . in particular , denoting as @xmath421 the complex coordinates on @xmath422 , the quotient acts as ( z_1,z_2 ) ~(p z_1,q z_2 ) , where @xmath423 are complex numbers with @xmath424 .
the complex structure of the resulting manifold @xmath419 depends on the two parameters @xmath423 . for more details we refer to @xcite and references therein .
it is straightforward to determine the complex structure @xmath423 compatible with a general metric on @xmath3 with @xmath20 isometry , namely [ metric4d ] s^2 = r^2 , where we have included a mixed term in the @xmath403 fibration over @xmath11 , parameterised by the constant @xmath425 .
we will then set @xmath426 to make contact with our solution .
we introduce complex coordinates [ complcoords ] z_1 = e^ b t e^(+ ) , z_2 = e^ b t e^(- ) , with respect to the complex structure @xmath427 associated to the fundamental two - form & = & ( b v t _ 3 + _ 1 _ 2 ) . assuming the identification @xmath428 , from we read off p = q = e^ b _
in particular , taking @xmath426 and @xmath429 as in our solution , we have that @xmath430 .
this is precisely the standard complex structure of the _ round _ @xmath413 , with ratio of radii given by @xmath431 , albeit the metric is _ not _ the round one . in the complex
coordinates , the metric reads .
note also that this metric is different from the metric ( 4.7 ) in @xcite and coincides with it only in the conformally flat limit . ]
s^2 = r^2 , with our metric corresponding to @xmath426 , @xmath432 and @xmath433 .
according to @xcite , the localised partition function of a supersymmetric gauge theory on our background must be proportional to the _ unrefined _ index , depending on the parameter @xmath434 .
let us now return to the interpretation of @xmath410 in this case . should contain information on a supersymmetric casimir energy of the theory placed on an arbitrary hopf surface @xmath419 .
this point will be addressed elsewhere @xcite . ]
the authors of @xcite have conjectured was presented in @xcite .
we thank s. kim for clarifying comments . ] that z_[s^1_s^3_round ] & = & x^_0 ( ) [ ztoind ] where @xmath435 and _ 0 & = & [ ( -1)^f h ] = _ [ h ] - _ [ h ] .
[ eindex ] when a free field theory limit exists this is just the sum of zero - point `` energies '' , namely eigenvalues of @xmath361 , weighted by a sign , and it was therefore referred as to _ index casimir energy _ in @xcite .
this expression is divergent and must be regulated , with the result depending on the choice of regularisation . as @xmath361 commutes with the supercharge(s ) , a natural regularisation @xcite is to weight the terms in the infinite sum with a factor of @xmath436 , sending @xmath437 at the end .
one obtains _ 0 & = & _ x1 ^ - [ ( -1)^f h x^h ] = _
x1 ^ - x _ ( f _ ( x ) + f _ ( x ) ) , [ baggio ] where the letter indices for a chiral multiplet with r - charge @xmath247 and a vector multiplet are given by f _ ( x ) = , f _ ( x ) = , respectively @xcite . in the case of interest to us the symmetry of the theory is broken to @xmath438 explicitly by the squashing , therefore there is no compelling reason for using a regulator that would preserve @xmath439 rotation symmetry .
see @xcite for a related discussion in five dimensions . using ( [ baggio ] ) and recalling that for a superconformal field theory the trace anomaly coefficients are = ( 3 ^ 3 - ) , = ( 9 ^ 3 - 5 ) , where @xmath440 denotes the sum over the ( r - charge)@xmath441 of all fermionic fields , for a quiver gauge theory one obtains @xcite & = & - ( - ) + ( + 3 ) + o ( ) .
after subtracting the first term , which is divergent in the @xmath442 limit ( in any case for superconformal quivers this term vanishes @xcite ) , one obtains f & = & _ 0 = ( + 3 ) . note that although we have taken the limit @xmath443 in ( [ baggio ] ) to derive @xmath444 , there is no contradiction with the definition of the casimir energy given in ( [ usualcasimir ] ) , and we see that from the latter we have @xmath445 , in the limit @xmath446 . for a superconformal quiver we can therefore compare @xmath447 with the gravity side writing e _ & = & = , [ aspetta ] where we used the holographic relation @xmath448 . in section [ holorensection ]
we computed the on - shell gravity action with supersymmetric @xmath357 boundary conditions .
after removing all divergences using the holographic counterterms , this reads [ copyon ] i = .
notice that the first term in the square bracket depends linearly on the complex structure parameter @xmath449 , while the remainder is a polynomial in the squashing parameter @xmath204 .
the former agrees precisely with the field theory expectation ( [ aspetta ] ) and we would like to interpret this as the ( index ) casimir energy , while the remainder polynomial in @xmath204 should be identified with a local counterterm . in the limit @xmath450 the solution becomes ads@xmath0 and ( [ copyon ] ) reduces to @xmath451 , which in @xcite was interpreted as the casimir energy on @xmath11 . on the other hand ,
the @xmath450 limit of our proposal for the casimir energy is @xmath452 .
the ratio between the former and the latter is @xmath453 , in agreement in @xcite and our @xmath18 differ by a factor of @xmath454 . ] with @xcite . in that reference
it was argued that the difference with respect to the usual casimir energy arises from the use of a regularisation not respecting the full symmetry of @xmath6 .
let us now discuss the polynomial in @xmath204 in ( [ copyon ] ) . in five dimensions , holographic
renormalisation suffers from ambiguities associated to local , scale invariant counterterms that remain finite when the uv cut - off is sent to infinity .
this means that after having removed all the divergences , one may still add these finite terms to the action , with arbitrary coefficients .
although it is believed that different choices of counterterms should correspond to different choices of renormalisation scheme in the dual field theory @xcite , a complete understanding of the details is currently lacking .
for a bulk action comprising only the einstein
hilbert term and the cosmological constant , generically one can construct three inequivalent finite counterterms using the metric and covariant derivatives @xcite . these are given by integrals of @xmath455 , @xmath456 and @xmath457 , where @xmath21 , @xmath237 and @xmath458 , are the ricci scalar , ricci tensor and riemann tensor of the boundary metric , respectively . in the presence of a gauge field ,
a fourth finite local invariant is @xmath459 .
equivalently , these may be parametrized by a basis given by @xmath268 , @xmath460 , @xmath455 and @xmath459 , where @xmath268 is the euler density @xmath461 , and @xmath460 is the square of the weyl tensor , @xmath462 . in particular , we have the following standard finite counterterm for the on - shell action : [ ambij ] s = _ m^4x h ( e + c_ijklc^ijkl + r^2 - f_ijf^ij ) , where @xmath79 , @xmath18 , @xmath463 , @xmath464 are arbitrary numerical constants . for our background , @xmath465 and gives a quadratic polynomial in @xmath204 : s ( 4-v^2 ) ^2 + ( 8- ) ( 1-v^2)^2 . tuning the coefficients @xmath466 we can cancel some of the polynomial terms in ( [ copyon ] ) , suggesting that indeed these are ambiguous , but there is no choice cancelling all terms simultaneously .
the main issue here is whether there exist _ new _ types of counterterms that may be used to modify the on - shell action ( [ copyon ] ) . below
we will discuss how new counterterms may emerge naturally taking into account the additional structure arising from supersymmetry .
the terms in ( [ ambij ] ) are , apart for @xmath455 , those appearing in the bosonic action of conformal supergravity .
therefore the only combination that is supersymmetric , with respect to _ rigid supersymmetry _ , is that with @xmath467 and @xmath468 , which vanishes @xcite .
however , the correct notion of supersymmetry in the context of holography is that of the bulk , and should involve also the divergent counterterms ; taking this into account it should be possible to formulate a supersymmetric version of holographic renormalisation . from this point of view
, it is very natural to expect that the supersymmetric counterterms should be derived by expanding asymptotically the @xmath27-structure encoding the bulk supersymmetry .
this was done at leading order in ref .
@xcite , where the @xmath469-structure on the lorentzian boundary was obtained from the @xmath61-structure in the bulk . in euclidean signature ,
the relevant @xmath27-structure is a metric - compatible complex structure @xmath470 .
since on our background @xmath465 and @xmath471 , the ambiguity so far may be parameterized by @xmath455 and @xmath459 . using the boundary complex structure tensor @xmath470 one can construct additional scale - invariant counterterms . here
we will not attempt to classify systematically all these terms and to study how supersymmetry constrains them .
we simply notice that using the complex structure we can introduce the _
ricci form _ of the boundary geometry , @xmath472 , which is invariant under global rescaling of the metric .
then a new finite counterterm is given by [ ricciformterm ] ^4 x h r_ijr^ij = 4 ^2 ( 4 - 3 v^2 ) ^2 .
this polynomial in @xmath204 is independent of those obtained by evaluating the @xmath455 and @xmath459 terms .
other obvious terms such as @xmath473 and @xmath474 are not independent , hence we do not need to consider them . combining these three counterterms with suitably chosen coefficients ,
we can remove the remainder polynomial in @xmath204 in .
namely , we find that [ onshellwithoutpolynomial ] i- _ m^4x h ( r^2 + f_ijf^ij - r_ijr^ij ) = . in appendix [ ambiguousapp ]
we discuss how the addition of these counterterms affects the holographic energy - momentum tensor and the r - symmetry current .
one can see that the value of the r - charge @xmath285 and the angular momentum @xmath290 are not modified , while the total energy @xmath292 is shifted by the same amount that changes the on - shell action .
therefore , the relation ( [ smarrgeneral ] ) remains valid as well as , and the vevs of the gauge - invariant hamiltonians introduced in section [ algebsection ] are not affected . in conclusion , we have shown that the supergravity on - shell action contains a term with the correct linear dependence on the complex structure modulus @xmath18 , that reproduces the ( index ) casimir energy expected from a supersymmetric field theory on our @xmath475 background .
however , it also contains other terms that we suggested to be removable by a choice of local counterterms constructed with the background metric , the background gauge field and the complex structure . from the point of view of the boundary field theory , in addition to the counterterms constructed with the metric and the conformal supergravity gauge field @xmath476
, it is natural to also include counterterms obtained from other combinations of the new minimal auxiliary fields @xmath477 and @xmath478
. it would be interesting to study how these can arise from supersymmetry in the bulk and thus be employed in holographic renormalization .
recently the authors of @xcite constructed a five - dimensional supersymmetric solution and presented evidence that this should be dual to a massive deformation of @xmath5 super yang
mills theory on @xmath479 .
in particular , they showed that the renormalised on - shell action agrees with the logarithm of the localised partition function on @xmath479 @xcite , but only up to terms argued to be ambiguous . in this respect ,
our construction is somewhat analogous to theirs .
in this paper we have constructed a new supersymmetric alads one - parameter family of solutions to minimal five - dimensional gauged supergravity .
this can be uplifted to type iib or to eleven - dimensional supergravity by using the consistent truncations of @xcite .
the parameter @xmath204 deforming ads@xmath0 into our solution may be thought of as a squashing of the boundary metric , together with the addition of a non - trivial graviphoton field , therefore it corresponds to a relevant deformation of the lagrangian in the dual superconformal field theory . the solution is smooth and has the topology of global ads@xmath0 , namely @xmath480 ; for generic values of the parameter it preserves an @xmath20 subgroup of the isometry group @xmath481 of ads@xmath0 .
despite we constructed the solution using a combination of perturbative expansions and numerical integration of the relevant ode , many quantities of physical interest were computed analytically as a function of the parameter @xmath204 .
these include the on - shell action , the total energy , the angular momentum and the electric charge .
we also provided numerical evidence that the solution is free from closed timelike curves .
after a wick rotation of the time coordinate , at the boundary
the metric remains real while the background gauge field becomes complex . in the bulk both the gauge field and
the metric become _ complex _ , but the appearance of complex metrics in quantum gravity should not be surprising ( see _ e.g. _ @xcite ) .
our main motivation for studying this supergravity solution is that via the gauge / gravity correspondence it provides the dual to a class of superconformal field theories defined on a four - dimensional curved manifold with certain background fields . in particular , after the analytic continuation , the riemannian manifold is homeomorphic to @xmath3 , equipped with a non conformally flat hermitian metric , namely it is a particular hopf surface @xcite .
it has been argued in @xcite that the localised partition function of supersymmetric field theories on a hopf surface with complex structure specified by two parameters @xmath423 is proportional to the refined supersymmetric index with @xmath423 fugacities .
the proportionality factor is expected to capture a supersymmetric _ casimir energy _ of the field theory , which may be expressed as a linear combination of the @xmath482 and @xmath483 anomaly coefficients @xcite .
since our background has complex structure @xmath484 , we have compared the renormalised on - shell supergravity action , leads to the antiperiodic spin structure on the boundary , and therefore it can not contribute @xmath485 @xcite . see @xcite for a related discussion in the context of asymptotically locally ads@xmath486 solutions . ] to the partition function associated with the unrefined index , obtaining agreement with the field theory expectation , up to terms that may be removed by a finite local counterterm .
we plan to come back to a more systematic analysis of the allowed supersymmetric counterterms in the future . in @xcite
it was noticed that the large @xmath8 behavior of the index could not be matched with the entropy of the supersymmetric black hole of @xcite , which must scale like @xmath487 .
it may be possible to revisit this puzzling feature in the light of the discussion in the present paper .
there are a number of possible extensions of this work that it is natural to study .
for simplicity , in this paper we have restricted our attention to a four - dimensional boundary with direct product metric , while as discussed in @xcite by a simple change of local coordinates it is straightforward to include a mixed term ( a non - zero parameter @xmath425 in the metric ( [ metric4d ] ) ) . as seen in section [ indexandcasimir ]
this deformation corresponds to complexifying the complex structure parameter @xmath488 of the hopf surface .
it would be nice to study this modification of our solution in more detail ( notice that this applies also to the unsquashed case @xmath372 ) .
a more complicated generalisation consists in reducing the symmetry of the ansatz , for example down to a @xmath489 sub - group . in this case
the supersymmetry conditions will lead to partial differential equations , making the problem of finding solutions much more difficult .
a further interesting direction is that of constructing gravity solutions dual to supersymmetric field theories on four - dimensional hermitian manifolds with topology different from @xmath3 , preserving generically only one supercharge . this is a very hard problem for a number of reasons .
firstly , while in the presence of two supercharges with opposite r - charges ( that we considered presently ) there exist al least two commuting killing vectors , a generic hermitian background is not guaranteed to possess any isometry .
we expect that in order to construct solutions with only one supercharge it will be necessary to start from euclidean supergravity with general complex gauge field , and possibly complex metric .
moreover , there exist topological obstructions for constructing smooth five - dimensional manifolds filling a four - dimensional boundary ; for example , it is not possible to construct a smooth five - dimensional solution whose boundary is ( topologically ) @xmath490 , while the first del pezzo surface @xmath491 is not obstructed @xcite .
we would like to thank b. assel , z. komargodski , s. murthy and p. west for insightful discussions , as well as s. ferrara , s. giusto , s. kim , j. sparks , a. tomasiello and a. tseytlin for useful correspondence .
is supported by the stfc grant st / j002798/1 .
d.m . is supported by the erc starting grant n. 304806 , `` the gauge / gravity duality and geometry in string theory '' , and also acknowledges partial support from the stfc grant st / j002798/1 .
in this appendix we provide more details on the uv solution .
we set @xmath492 . here
we give some more details about the solution at large @xmath96 . from the discussion in section [ susysec ] ,
the five - dimensional metric has the form s^2 = - f^2 ( y + _ 3)^2 + f^-1 , while the gauge field is given by [ formulafappendix ] f = .
the uv expansion of the functions @xmath38 and @xmath75 is obtained by plugging the uv solution for @xmath76 into and .
we obtain [ solftrho ] f & = & 1 + + + o(e^-5 ) and & = & -2a_0 ^ 2 e^2 + + 4 a_2 - 2c+ 4c+ + o(e^-3 ) . passing from the coordinates @xmath493 to the coordinates @xmath494 introduced in , namely = + t , y = t , = - , the five - dimensional metric can be expressed as s^2 = g _ ^2 + g _ ( _ 1 ^ 2 + _ 2 ^ 2 ) + g__3 ^ 2 + g_ttt^2 + 2 g_t _ 3 t , with g _ & = & f^-1 , g _ = f^-1a^2 , g _
= - f^2 ^2 + f^-1(2aa)^2 , + [ 2 mm ] g_tt & = & - f^2(1 + ) ^2 + ^2 f^-1 ( 2aa)^2 , g_t = -f^2 ( 1+)+ f^-1 ( 2aa)^2.[metriccompsimplicit ] by substituting the uv solution for @xmath76 , @xmath38 , @xmath75 , these metric components read : g _ & = & 1 - + ( e^-3 ) , + g _ & = & a_0 ^ 2 e^2 + + + ( e^- ) , + [ 2 mm ] g _ & = & ( 1 - 4c)+ ( e^- ) , + g_tt & = & - e^2 + + ( e^- ) , + [ 2 mm ] g_t & = & ( e^-2).[uvmetric ] in order to see the parameters @xmath127 , @xmath128 appear in the metric one needs to go one order further .
we will not present this here , as the expressions become cumbersome , but will do it in the next subsection when we will turn to fefferman graham coordinates . in a gauge ensuring that the five - dimensional supersymmetry parameter @xmath30 does not depend on @xmath114
, the gauge potential can be written as [ formofa ] a = a_t t + a _ _ 3 , with a_t & = & ( f + f+ ) , + a_&= & ( f+ ) .[acompimplicit ] by plugging the uv expansion in , these read a_t & = & + + o(e^-3 ) , + [ 2 mm ] a_&= & -c+ + [ 1 mm ] & & + o(e^-3 ) .
the leading order behavior of the metric and of the gauge field determines the background fields for the dual field theory living on the boundary , as displayed in eqs . and . in the main text
, we will also need the leading order expression of @xmath219 .
starting from , this is found to be _ 5 f & = & - t _ 3 + [ 2 mm ] & & + t_1_2 + [ 2 mm ] & & + _ 1 _ 2 _ 3 .
[ starf_uv ] in the following , we show that the five - dimensional metric and gauge field can be expressed in a fefferman
graham asymptotic expansion , which implies that the uv solution determined above is alads .
the general fefferman graham form of the metric is s^2 = + g_ij(x , r ) x^ix^j , where [ fgexpansion ] g(x , r ) = r^2 , while in the gauge @xmath495 the maxwell field has the form a(x , r ) = a^(0 ) + + .
since at leading order in @xmath496 the uv solution presented above in this appendix satisfies @xmath497 , @xmath498 and @xmath499 , it is immediately compatible with the fefferman
graham form upon setting @xmath500 . however , this is not as obvious at subleading orders , as @xmath501 is a non - trivial function of @xmath63 , and the @xmath502 components of the metric naively seem to have `` logarithmic '' ( when @xmath500 ) terms too early in the expansion . in order to show that the solution can be put in fefferman graham form , we need perform a change of radial coordinate , so that f^-1/2 = . by solving this equation perturbatively at large @xmath96 ,
we find [ changeradialcoord ] a_0 ^ 2 r^2 & = & a_0 ^ 2e^2 + + + + ( e^-3 ) .
one can now check that the five - dimensional metric is consistent with the fefferman graham expansion . using the coordinates @xmath503
, we obtain s^2 = + g _ ( _ 1 ^ 2 + _ 2 ^ 2 ) + g__3 ^ 2 + g_ttt^2 + 2 g_t t _ 3 , where @xmath203 , @xmath211 , @xmath170 and @xmath504 depend only on @xmath247 and have an expansion of the form .
specifically , we find : g^(0 ) _ & = & a_0 ^ 2,g^(2 )
_ = -,g^(4 ) _ = , + [ 2 mm ] g^(4 ) _ & = & , & & g_^(0 ) = a_0 ^ 2(1 - 4c),g_^(2 ) = , g_^(4 ) = - , + [ 2 mm ] & & g_^(4 ) = -,g_t^(0 ) = g_t^(2 ) = g_t^(4 ) = 0 , g_t^(4 ) = -2g_^(4 ) - 2g_^(4 ) + .
[ solgamma ] g_tt^(0 ) & = & -,g_tt^(2 ) = - , g_tt^(4 ) = 0 , + [ 2 mm ] g_tt^(4 ) & = & 8g_^(4 ) + 4()^2 g_^(4 ) - . + according to the standard ads / cft rules , the @xmath505 coefficients are to be interpreted as source background fields for the dual field theory , while the @xmath506 are related to the expectation value of the dual energy - momentum tensor
. let us count how many free parameters we have at this stage .
we already discussed that the boundary metric has two independent parameters ( plus the trivial one obtained by rescaling @xmath114 ) .
indeed , @xmath507 can always be set to zero by a change of local coordinates , as done here , and @xmath508 can be set to any value by rescaling @xmath114 .
we can thus identify the free parameters with @xmath509 and @xmath510 , which are arbitrary because @xmath91 and @xmath511 are . as expected from general considerations about solutions on alads backgrounds
, we see that the sources fix @xmath512 and @xmath513 . regarding the parameters related to vacuum expectation values , we find that three of them are arbitrary . indeed ,
@xmath514 , @xmath515 and @xmath516 are arbitrary because they are independent functions of the free parameters @xmath126 , @xmath127 , @xmath128 , while @xmath517 is fixed in terms of the other expectations values and the sources as in .
the regularity condition imposed in the ir will constrain these vevs so that no freedom is left once the sources are fixed . the gauge field in the new radial variable @xmath518 is a = t + _ 3 + ( r^-3 ) , with a_t^(0 ) & = & , a_t^(2 ) = 0 , + a_t^(2 ) & = & , + [ 2 mm ] a_^(0 ) & = & -c , a_^(2 ) = , + [ 2 mm ] a_^(2 ) & = & .
because of supersymmetry , the solution for the gauge field does not contain any additional free parameter with respect to the metric .
we have thus proved that the uv solution can be put in fefferman
graham form .
we can also check that our findings agree with general results on alads backgrounds .
for instance , an asymptotic analysis of the five - dimensional einstein equations in the presence of a cosmological constant @xcite and a maxwell field @xcite shows that the first subleading component of the metric is determined as [ g(2)general ] g_ij^(2 ) = - ( r_ij - r g^(0)_ij ) , while the trace of @xmath506 is fixed by [ traceg(4)general ] tr g_(4 ) = g_(2)^2 + f_(0)^2 . here
, @xmath237 is the ricci tensor of @xmath519 , @xmath21 its curvature and the indices are raised using @xmath505 . and @xmath520.[comparisontaylor ] ] the traceless part of @xmath506 is instead not constrained by the einstein equation .
we checked that these conditions are indeed satisfied by our solution .
in the following we give the explicit expressions for the holographically renormalised energy - momentum tensor and r - current of our family of solutions .
these are computed using and , and turn out to depend on the uv parameters @xmath91 , @xmath126 , @xmath127 , @xmath128 and @xmath191 .
however , we know that the full solution only has one independent parameter . in the main text we showed how one can eliminate two of them : @xmath127 is determined by the global smoothness of the solution as discussed in section [ sec : globalconstraint ] , while @xmath128 is fixed by a ward identity as explained in section [ sec : holocharges ] .
we find that their explicit expressions are a_4 & = & ( 5 + 64 a_2 - 256 a_2 ^ 2 + 32 c- 160 a_2 c- 104 c^2 ) , + [ 2 mm ] a_6 & = & [ sola6 ] .
the relation between the remaining parameters has been determined numerically , see fig .
[ fig : otheruvparametersplots ] . for the energy - momentum tensor
we find t_ijx^i
x^j = t_(_1 ^ 2+_2 ^ 2 ) + t__3 ^ 2 + 2 t_tt _ 3 + t_ttt^2 , with , using , t_&= & , + [ 2 mm ] t_&= & , + [ 2 mm ] t_t & = & , + [ 2 mm ] t_tt&= & ( + - + v^2 ) . [ allthestress ] it is easy to check that @xmath263 is covariantly conserved and traceless with respect to the boundary metric .
the renormalised electric current , corresponding to the dual field theory r - current , reads j = and is covariantly conserved with respect to the boundary metric .
we can verify that evaluating the associated charges via the definitions in section [ sec : holocharges ] gives the results presented in the main text . recall that @xmath521 , and that the volume form on the three - sphere at the boundary is ( s^3_bdry ) = ^3a_0 ^ 3 v _
1_2_3 = -^3a_0 ^ 3 v . with our choice of orientation , @xmath522 .
so integrating we have @xmath523 , which when @xmath372 is the volume of a round 3-sphere of radius @xmath524 .
then for the energy and the angular momentum we find e&= & _
s^3_bdry u^it_it vol(s^3_bdry ) = ( + - v^2 + v^4 ) , + [ 2 mm ] j & = & _ s^3_bdry u^it_i vol(s^3_bdry ) = ( v^2 -1 ) ^3 , while the electric charge is q = _ s^3_bdry u_i j^i vol(s^3_bdry ) = - ( v^2 -1)^2 .
these are indeed the expressions appearing in the main text .
consider the standard finite counterterms for the on - shell action .
these lead to the following ambiguity in the holographic energy - momentum tensor [ ambiguitytij ] t_ij = - = , where we used the fact that the metric variation of the euler density vanishes identically in four dimensions .
the tensor @xmath525 is given by h_ij = - ^4 x r^2 = 2 _ i _ j r - 2 h_ij r + 12 h_ij r^2 - 2 r r_ij , while @xmath266 is the bach tensor following from the variation of the weyl square term : [ bachtensor ] b_ij & = & - ^4 x c_ijklc^ijkl + [ 2 mm ] & = & 13_i_j r - r_ij + 16 h_ij r - 2 r _ ikjlr^kl + 23 r r_ij + 12 h_ij ( r_klr^kl - 13 r^2 ) .these can also be deduced from the expressions given in @xcite , after adapting them to @xmath526 signature and to the curvature tensor conventions in our footnote [ footnoteconventionscurvature ] .
both @xmath525 and @xmath266 are covariantly conserved , and @xmath527 . including the new counterterm also affects the energy - momentum tensor , as ^4 x h r_klr^kl = 2 r_ik r_j^k - h_ijr_kl r^kl + 2 _ l_k ( r_i^k _ j^l ) , where the variation is done at fixed @xmath470 .
specializing these formulae to our background , one can check that the ambiguities in the energy - momentum tensor affect the energy @xmath289 but not the angular momentum @xmath290 .
specifically , @xmath289 is shifted by the same amount as the on - shell action @xmath528 , in such a way that both relations and continue to hold .
finally , we note that there is also an ambiguity in the r - current : the variation of the counterterms with respect to the boundary gauge field @xmath529 yields @xmath530 .
however , this does not modify the temporal component of the current and therefore does not affect the charge @xmath285 defined in .
v. pestun , _ localization of gauge theory on a four - sphere and supersymmetric wilson loops _ , commun
. math .
phys . * 313 * , 71 ( 2012 ) [ arxiv:0712.2824 [ hep - th ] ] .
a. kapustin , b. willett and i. yaakov , _ exact results for wilson loops in superconformal chern - simons theories with matter _ , jhep * 1003 * , 089 ( 2010 ) [ arxiv:0909.4559 [ hep - th ] ] .
d. l. jafferis , _ the exact superconformal r - symmetry extremizes z _ , jhep * 1205 * , 159 ( 2012 ) [ arxiv:1012.3210 [ hep - th ] ] .
n. hama , k. hosomichi and s. lee , _ notes on susy gauge theories on three - sphere _
, jhep * 1103 * , 127 ( 2011 ) [ arxiv:1012.3512 [ hep - th ] ] . g. festuccia and n. seiberg , _ rigid supersymmetric theories in curved superspace _ ,
jhep * 1106 * ( 2011 ) 114 [ arxiv:1105.0689 [ hep - th ] ] . c. klare ,
a. tomasiello and a. zaffaroni , _ supersymmetry on curved spaces and holography _ , jhep * 1208 * ( 2012 ) 061 [ arxiv:1205.1062 [ hep - th ] ] .
t. t. dumitrescu , g. festuccia and n. seiberg , _ exploring curved superspace _
, jhep * 1208 * , 141 ( 2012 ) [ arxiv:1205.1115 [ hep - th ] ] .
d. cassani , c. klare , d. martelli , a. tomasiello and a. zaffaroni , _ supersymmetry in lorentzian curved spaces and holography _ , commun .
* 327 * ( 2014 ) 577 [ arxiv:1207.2181 [ hep - th ] ] . c. closset , t. t. dumitrescu , g. festuccia and z. komargodski , _ supersymmetric field theories on three - manifolds _ , jhep * 1305 * ( 2013 ) 017 [ arxiv:1212.3388 [ hep - th ] ] . n. hama , k. hosomichi and s. lee , _ susy gauge theories on squashed three - spheres _ ,
jhep * 1105 * , 014 ( 2011 ) [ arxiv:1102.4716 [ hep - th ] ]
. y. imamura and d. yokoyama , _
@xmath5 supersymmetric theories on squashed three - sphere _ phys .
d * 85 * , 025015 ( 2012 ) [ arxiv:1109.4734 [ hep - th ] ] .
l. f. alday , d. martelli , p. richmond and j. sparks , _
localization on three - manifolds _ , jhep * 1310 * ( 2013 ) 095 [ arxiv:1307.6848 [ hep - th ] ] .
d. martelli , a. passias and j. sparks , _ the gravity dual of supersymmetric gauge theories on a squashed three - sphere _ , nucl .
b * 864 * , 840 ( 2012 ) [ arxiv:1110.6400 [ hep - th ] ] .
d. martelli and j. sparks , _ the gravity dual of supersymmetric gauge theories on a biaxially squashed three - sphere _ , nucl .
b * 866 * ( 2013 ) 72 [ arxiv:1111.6930 [ hep - th ] ] .
d. martelli , a. passias and j. sparks , _ the supersymmetric nuts and bolts of holography _ , nucl .
b * 876 * , 810 ( 2013 ) [ arxiv:1212.4618 [ hep - th ] ] .
d. martelli and a. passias , _ the gravity dual of supersymmetric gauge theories on a two - parameter deformed three - sphere _ , nucl .
b * 877 * , 51 ( 2013 ) [ arxiv:1306.3893 [ hep - th ] ] . c. closset and i. shamir , _
the @xmath531 chiral multiplet on @xmath532 and supersymmetric localization _ , arxiv:1311.2430 [ hep - th ] .
o. aharony , j. marsano , s. minwalla , k. papadodimas and m. van raamsdonk , _ the hagedorn - deconfinement phase transition in weakly coupled large n gauge theories _ , adv .
* 8 * , 603 ( 2004 ) [ hep - th/0310285 ] . c. romelsberger ,
_ counting chiral primaries in n = 1 , d=4 superconformal field theories _ , nucl .
b * 747 * , 329 ( 2006 ) [ hep - th/0510060 ] .
j. kinney , j. m. maldacena , s. minwalla and s. raju , _ an index for 4 dimensional super conformal theories _ ,
phys . * 275 * , 209 ( 2007 ) [ hep - th/0510251 ] . c. closset , t. t. dumitrescu , g. festuccia and z. komargodski , _ the geometry of supersymmetric partition functions _ , jhep * 1401 * ( 2014 ) 124 [ arxiv:1309.5876 [ hep - th ] ] .
a. gadde , l. rastelli , s. s. razamat and w. yan , _ on the superconformal index of @xmath2 ir fixed points : a holographic check _ ,
jhep * 1103 * , 041 ( 2011 ) [ arxiv:1011.5278 [ hep - th ] ] .
r. eager , j. schmude and y. tachikawa , _ superconformal indices , sasaki - einstein manifolds , and cyclic homologies _ , arxiv:1207.0573 [ hep - th ]
. s. s. gubser , i. r. klebanov and a. m. polyakov , _ gauge theory correlators from noncritical string theory _ ,
b * 428 * , 105 ( 1998 ) [ hep - th/9802109 ] .
e. witten , _ anti - de sitter space and holography _ , adv .
* 2 * ( 1998 ) 253 [ hep - th/9802150 ] .
m. henningson and k. skenderis , _ the holographic weyl anomaly _ , jhep * 9807 * ( 1998 ) 023 [ hep - th/9806087 ] .
v. balasubramanian and p. kraus , _ a stress tensor for anti - de sitter gravity _
, commun .
* 208 * ( 1999 ) 413 [ hep - th/9902121 ] .
s. kim , _ the complete superconformal index for n=6 chern - simons theory _ , nucl .
b * 821 * , 241 ( 2009 ) [ erratum - ibid .
b * 864 * , 884 ( 2012 ) ] [ arxiv:0903.4172 [ hep - th ] ] .
h. -c .
kim and s. kim , _
m5-branes from gauge theories on the 5-sphere _ , jhep * 1305 * , 144 ( 2013 ) [ arxiv:1206.6339 [ hep - th ] ]
. h. -c .
kim , s. kim , s. -s .
kim and k. lee , _ the general m5-brane superconformal index _ , arxiv:1307.7660 [ hep - th ] .
c. fefferman , r. graham , _ conformal invariants _ , in lie cartan et les mathmatiques daujourdhui , astrisque ( 1985 ) , 95 .
j. p. gauntlett , j. b. gutowski and n. v. suryanarayana , _ a deformation of @xmath533 _ , class .
* 21 * ( 2004 ) 5021 [ hep - th/0406188 ] .
j. p. gauntlett and j. b. gutowski , _ all supersymmetric solutions of minimal gauged supergravity in five - dimensions _ ,
d * 68 * ( 2003 ) 105009 [ erratum - ibid .
d * 70 * ( 2004 ) 089901 ] [ hep - th/0304064 ] .
d. cassani and d. martelli , _ supersymmetry on curved spaces and superconformal anomalies _
, jhep * 1310 * ( 2013 ) 025 [ arxiv:1307.6567 [ hep - th ] ] .
a. buchel and j. t. liu , _ gauged supergravity from type iib string theory on @xmath327 manifolds _ , nucl .
b * 771 * ( 2007 ) 93 [ hep - th/0608002 ] .
j. p. gauntlett and o. varela , _ consistent kaluza - klein reductions for general supersymmetric ads solutions _ , phys . rev .
d * 76 * , 126007 ( 2007 )
[ arxiv:0707.2315 [ hep - th ] ] . j. p. gauntlett , d. martelli , j. sparks and d. waldram , _ supersymmetric ads@xmath0 solutions of type iib supergravity _
, class .
* 23 * , 4693 ( 2006 ) [ hep - th/0510125 ] .
gauntlett , e. o colgain and o. varela , _ properties of some conformal field theories with m - theory duals _ ,
jhep * 0702 * ( 2007 ) 049 [ hep - th/0611219 ] .
j. p. gauntlett , d. martelli , j. sparks and d. waldram , _
supersymmetric ads(5 ) solutions of m theory _
, class .
* 21 * ( 2004 ) 4335 [ hep - th/0402153 ] .
j. b. gutowski and h. s. reall , _
supersymmetric ads@xmath0 black holes _ ,
jhep * 0402 * , 006 ( 2004 ) [ hep - th/0401042 ] .
p. figueras , c. a. r. herdeiro and f. paccetti correia , _ on a class of 4d khler bases and ads@xmath0 supersymmetric black holes _ ,
jhep * 0611 * ( 2006 ) 036 [ hep - th/0608201 ] .
a. butti , m. grana , r. minasian , m. petrini and a. zaffaroni , _ the baryonic branch of klebanov - strassler solution : a supersymmetric family of @xmath534 structure backgrounds _ , jhep * 0503 * , 069 ( 2005 ) [ hep - th/0412187 ] . j. gaillard , d. martelli , c. nunez and i. papadimitriou , _ the warped , resolved , deformed conifold gets flavoured _ , nucl .
b * 843 * , 1 ( 2011 ) [ arxiv:1004.4638 [ hep - th ] ] .
s. s. gubser , c. p. herzog and i. r. klebanov , _ symmetry breaking and axionic strings in the warped deformed conifold _ , jhep
* 0409 * , 036 ( 2004 ) [ hep - th/0405282 ] . s. giusto , s. d. mathur and a. saxena , _ dual geometries for a set of 3-charge microstates _ , nucl .
b * 701 * ( 2004 ) 357 [ hep - th/0405017 ] .
v. jejjala , o. madden , s. f. ross and g. titchener , _ non - supersymmetric smooth geometries and d1-d5-p bound states _ , phys .
d * 71 * ( 2005 ) 124030 [ hep - th/0504181 ] .
s. de haro , s. n. solodukhin and k. skenderis , _ holographic reconstruction of space - time and renormalization in the ads / cft correspondence _ , commun .
phys . * 217 * ( 2001 ) 595 [ hep - th/0002230 ] .
m. taylor , _ more on counterterms in the gravitational action and anomalies _ , hep - th/0002125 .
m. bianchi , d. z. freedman and k. skenderis , _ holographic renormalization _
b * 631 * ( 2002 ) 159 [ hep - th/0112119 ] .
i. papadimitriou and k. skenderis , _ thermodynamics of asymptotically locally ads spacetimes _ , jhep * 0508 * ( 2005 ) 004 [ hep - th/0505190 ] .
a. ashtekar and s. das , _ asymptotically anti - de sitter space - times : conserved quantities _ , class .
* 17 * ( 2000 ) l17 [ hep - th/9911230 ] .
j. m. maldacena , _ the large n limit of superconformal field theories and supergravity _ , adv .
* 2 * , 231 ( 1998 ) [ hep - th/9711200 ] .
i. r. klebanov and e. witten , _ superconformal field theory on three - branes at a calabi - yau singularity _
b * 536 * , 199 ( 1998 ) [ hep - th/9807080 ] .
s. benvenuti , s. franco , a. hanany , d. martelli and j. sparks , _ an infinite family of superconformal quiver gauge theories with sasaki - einstein duals _ , jhep * 0506 * , 064 ( 2005 ) [ hep - th/0411264 ] .
e. s. fradkin and a. a. tseytlin , _ conformal supergravity _ , phys .
rept . * 119 * , 233 ( 1985 ) .
d. z. freedman and a. van proeyen , _ supergravity _ , cambridge university press ( 2012 ) .
c. klare and a. zaffaroni , _ extended supersymmetry on curved spaces _ , jhep * 1310 * ( 2013 ) 218 [ arxiv:1308.1102 [ hep - th ] ] . v. balasubramanian , e. g. gimon , d. minic and j. rahmfeld , _ four - dimensional conformal supergravity from ads space _ , phys . rev .
d * 63 * ( 2001 ) 104009 [ hep - th/0007211 ] .
m. f. sohnius and p. c. west , _ an alternative minimal off - shell version of n=1 supergravity _ , phys .
b * 105 * , 353 ( 1981 ) . c. p. herzog and k. -w .
huang , _ stress tensors from trace anomalies in conformal field theories _ , phys .
d * 87 * , 081901 ( 2013 ) [ arxiv:1301.5002 [ hep - th ] ] .
k. kodaira , _ complex structures on @xmath3 _ , proceedings of the national academy of sciences of the united states of america * 55 * , 240 ( 1966 ) .
b. assel , d. cassani and d. martelli , _ localization on hopf surfaces _ , arxiv:1405.5144 [ hep - th ] . s. benvenuti and a. hanany , _ new results on superconformal quivers _ , jhep * 0604 * , 032 ( 2006 ) [ hep - th/0411262 ] . n. d. birrell and p. c. w. davies , _ quantum fields in curved space _ , cambridge university press ( 1982 ) .
c. closset , t. t. dumitrescu , g. festuccia , z. komargodski and n. seiberg , _ contact terms , unitarity , and f - maximization in three - dimensional superconformal theories _ , jhep * 1210 * , 053 ( 2012 ) [ arxiv:1205.4142 [ hep - th ] ] . c. closset , t. t. dumitrescu , g. festuccia , z. komargodski and n. seiberg , _ comments on chern - simons contact terms in three dimensions _ ,
jhep * 1209 * , 091 ( 2012 ) [ arxiv:1206.5218 [ hep - th ] ] .
s. cecotti , s. ferrara , m. porrati and s. sabharwal , _ new minimal higher derivative supergravity coupled to matter _ , nucl .
b * 306 * , 160 ( 1988 ) .
n. bobev , h. elvang , d. z. freedman and s. s. pufu , _
holography for @xmath535 on @xmath479 _ , arxiv:1311.1508 [ hep - th ] .
j. d. brown , e. a. martinez and j. w. york , jr . , _ rotating black holes , complex geometry , and thermodynamics _ , annals n. y. acad .
sci . * 631 * , 225 ( 1991 ) .
j. d. brown , e. a. martinez and j. w. york , jr .
, _ complex kerr - newman geometry and black hole thermodynamics _ , phys .
lett . * 66 * , 2281 ( 1991 ) . | we present a new one - parameter family of supersymmetric solutions deforming ads@xmath0 .
this is constructed as an asymptotically locally anti de sitter ( alads ) solution of five - dimensional minimal gauged supergravity , with topology @xmath1 and a non - trivial graviphoton field , and can be uplifted to ten or eleven dimensional supergravities .
an analytic continuation of this solution yields the gravity dual to a class of four - dimensional @xmath2 supersymmetric gauge theories on a curved manifold with topology @xmath3 , comprising an @xmath4-symmetric squashed three - sphere , with a non - trivial background gauge field coupling to the r - symmetry current .
we compute the holographically renormalised on - shell action and interpret it in terms of the casimir energy of the dual field theory .
we also determine the holographic conserved charges of the solution and discuss relations between them .
2.3 cm * the gravity dual of supersymmetric gauge theories + on a squashed @xmath3 * davide cassani and dario martelli _ department of mathematics , king s college london , + [ 1 mm ] the strand , london wc2r 2ls , united kingdom + _ 2 cm 6 mm |
the _ golden ratio _ has fascinated people of diverse interests for at least 2400 years ( see , e.g. , @xcite ) . in mathematics ( and the arts ) two quantities @xmath21 and @xmath22 are in the golden ratio if the ratio of the sum of the quantities @xmath23 to the larger quantity @xmath21 is equal to the ratio of the larger quantity @xmath21 to the smaller quantity @xmath22 .
this amounts to setting @xmath24 and solving @xmath25 which yields @xmath20 apart from being abundant in nature , and finding diverse applications ranging from architecture to music , the golden ratio has also found more recent uses in _ technical analysis _ of asset prices ( in strategies such as _ fibonacci retracement _ representing an ad - hoc method for determining _ support _ and _ resistance _ levels ) . despite its universal presence and canonical role in diverse applied areas , we are not aware of any more _ exact _ connections between the golden ratio and _ stochastic processes _ ( including any proofs of optimality in particular).=-1 one of the aims of the present paper is to disclose the appearance of the golden ratio in an optimal stopping strategy related to the radial part of three - dimensional brownian motion . more specifically , denoting the radial part by @xmath14 it is well known that @xmath14 is transient in the sense that @xmath4 as @xmath26 . after starting at some @xmath27 , the ultimate minimum of @xmath14 will therefore be attained at some time @xmath8 that is not predictable through the sequential observation of @xmath14 ( in the sense that it is only revealed at the end of time ) .
the question we are addressing is to determine a ( predictable ) stopping time of @xmath14 that is as close as possible to @xmath8 .
we answer this question by showing that _ the time at which the excursion of @xmath14 away from the running minimum @xmath28 and the running minimum @xmath28 itself form the golden ratio _ is as close as possible to @xmath8 in a normalized mean deviation sense .
we consider this problem by embedding it into transient bessel processes of dimension @xmath29 and in this context we derive similar optimal stopping rules .
we also disclose further / deeper extensions of these results to transient diffusion processes .
the relevance of these questions in financial applications is motivated by the problem of optimal trading in the presence of bubbles . in this context
we show that the golden ratio rule offers a rigorous optimality argument for the choice of the well - known _ golden retracement _ in technical analysis of asset prices . to our knowledge
this is the first time that such an argument has been found / given in the literature .
the problem considered in the present paper belongs to the class of optimal prediction problems ( within optimal stopping ) .
similar optimal prediction problems have been studied in recent years by many authors ( see , e.g. , @xcite ) . once the `` unknown '' future is projected to the `` known '' present , we find that the resulting optimal stopping problem takes a novel integral form that has not been studied before
. the appearance of the minimum process in this context makes the problem related to optimal stopping problems for the maximum process that were initially studied and solved in important special cases of diffusion processes in @xcite and @xcite .
the general solution to problems of this kind was derived in the form of the maximality principle in @xcite ; see also section 13 and chapter v in @xcite and the other references therein .
more recent contributions and studies of related problems include .
close three - dimensional relatives of these problems also appear in the recent papers @xcite and @xcite where the problems were effectively solved by guessing and finding the optimal stopping boundary in a closed form .
the maximality principle has been extended to three - dimensional problems in the recent paper @xcite . although the structure of the present problem is similar to some of these problems , it turns out that none of these results is applicable in the present setting .
governed by these particular features in this paper we show how the problem can be solved when ( i ) no closed - form solution for the candidate stopping boundary is available and ( ii ) the loss function takes an integral form where the integrand is a functional of both the process @xmath14 and its running minimum @xmath28 .
this is done by extending the arguments associated with the maximality principle to the setting of the present problem and disclosing the general form of the solution that is valid in all particular cases .
the key novel ingredient revealed in the solution is the replacement of the diagonal and its role in the maximality principle by a nonlinear curve in the two - dimensional state space of @xmath14 and @xmath28 .
we believe that this methodology is of general interest and the arguments developed in the proof should be applicable in similar two / multi - dimensional integral settings .
\1 . we consider a nonnegative diffusion process @xmath0 solving @xmath30 where @xmath31 and @xmath2 are continuous functions satisfying ( [ 24 ] ) and ( [ 25 ] ) below , and @xmath32 is a standard brownian motion . by @xmath33
we denote the probability measure under which the process @xmath14 starts at @xmath34 . recalling that the scale function of @xmath14 is given by @xmath35 and the speed measure of @xmath14 is given by @xmath36 we assume that the following conditions are satisfied : @xmath37 from ( [ 24 ] ) we read that @xmath14 is a transient diffusion process in the sense that @xmath4 @xmath33-a.s . as @xmath26 , and from ( [ 25 ] )
we read that @xmath38 is an entrance boundary point for @xmath14 in the sense that the process @xmath14 could start at @xmath38 but will never return to it ( implying also that @xmath14 will never visit @xmath38 after starting at @xmath27 ) .
the main example we have in mind is the @xmath39-dimensional bessel process @xmath14 solving @xmath40 where @xmath29 . recalling that the scale function is determined up to an affine transformation we can choose the scale function ( [ 22 ] ) and hence the speed measure ( [ 23 ] ) to read @xmath41 for @xmath27 .
it is well known that when @xmath42 one can realize @xmath14 as the radial part of @xmath39-dimensional standard brownian motion .
similar interpretations of ( [ 26 ] ) are also valid when @xmath43 ( with an addition of the local time at zero ) and @xmath44 but @xmath14 is not transient in these cases ( but recurrent ) , and hence the problem considered below will have a trivial solution . other examples of ( [ 21 ] ) are obtained by composing bessel processes solving ( [ 26 ] ) with strictly decreasing and smooth functions .
this is of interest in financial applications and will be discussed below .
there are also many other examples of transient diffusion processes solving ( [ 21 ] ) that are not related to bessel processes .
\3 . to formulate the problem to be studied below
consider the diffusion process @xmath14 solving ( [ 21 ] ) , and introduce its running minimum process @xmath45 by setting @xmath46 for @xmath7 . due to the facts that @xmath14 is transient ( converging to @xmath47 ) and
@xmath38 is an entrance boundary point for @xmath14 , we see that the ultimate infimum @xmath48 is attained at some random time @xmath8 in the sense that @xmath49 with @xmath33-probability one for @xmath27 given and fixed ( the case @xmath50 being trivial and therefore excluded ) . it is well known that @xmath8 is unique up to a set of @xmath33-probability zero ( cf .
@xcite , theorem 2.4 ) .
the random time @xmath8 is clearly unknown at any given time and can not be detected through sequential observations of the sample path @xmath51 for @xmath7 . in many applied situations of this kind ,
we want to devise sequential strategies which will enable us to come as `` close '' as possible to @xmath8 .
most notably , the main example we have in mind is the problem of optimal trading in the presence of bubbles to be addressed below . in mathematical terms
this amounts to finding a stopping time of @xmath14 that is as `` close '' as possible to @xmath8 .
a first step toward this goal is provided by the following lemma .
we recall that stopping times of @xmath14 refer to stopping times with respect to the natural filtration of @xmath14 that is defined by @xmath52 for @xmath7 .
[ lemma1 ] we have @xmath53 for all stopping ( random ) times @xmath13 of @xmath14 .
the identity is well known ( see , e.g. , @xcite , page 450 ) and can be derived by noting that @xmath54 for all stopping ( random ) times @xmath13 of @xmath14 as claimed .
\4 . taking @xmath55 on both sides in ( [ 211 ] )
yields a nontrivial measure of error ( from @xmath13 to @xmath8 ) as long as @xmath56 for @xmath27 given and fixed .
the latter condition , however , may not always be fulfilled .
for example , when @xmath14 is a transient bessel process of dimension @xmath29 it is known ( see @xcite , lemma 1 ) that @xmath57 as @xmath58 .
hence we see that @xmath59 if and only if @xmath60 or equivalently @xmath61 .
it is clear from ( [ 211 ] ) , however , that the pointwise minimization of the euclidean distance on the left - hand side is equivalent to the pointwise minimization of the integral on the right - hand side . to preserve the generality
we therefore `` normalize '' @xmath62 on the left - hand side by subtracting @xmath8 from it . after taking @xmath55 on both sides of the resulting identity
, we obtain @xmath63 for all stopping times @xmath13 of @xmath14 ( for which the right - hand side is well defined ) . the optimal prediction problem
therefore becomes @xmath64 where the infimum is taken over all stopping times @xmath13 of @xmath14 ( with finite mean ) and @xmath27 is given and fixed .
note that the problem ( [ 214 ] ) is equivalent to the problem of minimizing @xmath65 over all stopping times @xmath13 of @xmath14 ( with finite mean ) whenever @xmath66 . to tackle the problem ( [ 214 ] ) we first focus on the right - hand side in ( [ 213 ] ) above .
[ lemma2 ] we have @xmath67 for all stopping times @xmath13 of @xmath14 ( with finite mean ) and all @xmath68 . using a well - known argument ( see , e.g. ,
@xcite , page 450 ) , we find that @xmath69 \bigr ) \,dt \nonumber\\ & = & \int_0^\infty{\mathsf e}_x \bigl ( i(t <
\tau ) { \mathsf e}_x \bigl[\bigl ( 2 i(\theta\le t ) - 1 \bigr ) \vert{{\mathcal}f}_t^x \bigr ]
\bigr ) \,dt \\ & = & { \mathsf e}_x \int_0^\tau\bigl ( 2 { \mathsf p } _ { x } ( \theta\le t \vert{{\mathcal}f}_t^x ) - 1 \bigr ) \,dt \nonumber\\ & = & { \mathsf e}_x \int_0^\tau\bigl ( 1 - 2 { \mathsf p } _ { x } ( \theta > t \vert { { \mathcal}f}_t^x ) \bigr ) \,dt\nonumber\end{aligned}\ ] ] for any stopping time @xmath13 of @xmath14 ( with finite mean ) and any @xmath68 given and fixed . setting @xmath70 and recalling that @xmath71 , we find by the markov property that @xmath72\\[-8pt ] & = & { \mathsf p } _ { x}(i_\infty \circ\theta_t <
i \vert{{\mathcal}f}_t^x ) \vert_{i = i_t } = { \mathsf p}_{x_t}(i_\infty <
i ) \vert_{i = i_t}\nonumber\end{aligned}\ ] ] for @xmath73 . to compute the latter probability we recall that @xmath74 is a continuous local martingale and note that @xmath75 if and only if @xmath76 where @xmath77 .
this shows that the set @xmath78 coincides with the set @xmath79 which in turn can be expressed as @xmath80 where @xmath81 .
taking @xmath82 we see that the continuous local martingale @xmath83 is bounded above by @xmath38 and bounded below by @xmath84 with @xmath85 under @xmath33 .
it follows therefore that @xmath86 is a uniformly integrable martingale and hence by the optional sampling theorem we find that @xmath87 + { \mathsf e}_x [ m_\infty i(\sigma= \infty ) ] \nonumber\\[-8pt]\\[-8pt ] & = & l(i ) { \mathsf p } _ { x } ( \sigma < \infty)\nonumber\end{aligned}\ ] ] upon using that @xmath88 @xmath33-a.s . on @xmath89 . combining ( [ 218 ] ) with the previous conclusions we obtain @xmath90 for @xmath91 in @xmath1 . from ( [ 217 ] ) and ( [ 219 ] ) we see that @xmath92 for all @xmath34 and @xmath7 ( for the underlying three - dimensional law , see @xcite , theorem a ) . inserting this expression back into ( [ 216 ] ) we obtain ( [ 215 ] ) and the proof is complete . \5 . from ( [ 213 ] ) and ( [ 215 ] ) we see that the problem ( [ 214 ] ) is equivalent to @xmath93 where the infimum is taken over all stopping times @xmath13 of @xmath14 ( with finite mean ) and @xmath27 is given and fixed . passing from the initial diffusion process @xmath14 to the scaled diffusion process
@xmath94 we see that there is no loss of generality in assuming that @xmath95 in ( [ 21 ] ) or equivalently that @xmath96 for @xmath27 ( with @xmath97 @xmath33-a.s . as @xmath5 ) .
note that the time of the ultimate minimum @xmath8 is the same for both @xmath14 and @xmath94 since @xmath3 is strictly increasing .
note also that @xmath13 is a stopping time of @xmath14 if and only if @xmath13 is a stopping time of @xmath94 . to keep the track of the general formulas throughout
we will continue with considering the general case ( when @xmath31 is not necessarily zero and @xmath3 is not necessarily the identity function ) .
this problem will be tackled in the next section below .
\6 . for future reference
we recall that the infinitesimal generator of @xmath14 equals @xmath98 for @xmath27 . throughout
we denote @xmath99 and set @xmath100 for @xmath101 in @xmath1 .
it is well known that @xmath102 for @xmath103 in @xmath1 .
the green function of @xmath14 is given by @xmath104\\[-8pt ] & = & \frac{(l(b ) - l(x ) ) ( l(y ) -
l(a))}{l(b ) - l(a ) } \qquad\mbox{if } a \le y \le x \le b.\nonumber\end{aligned}\ ] ] if @xmath105 is a measurable function , then it is well known that @xmath106 for @xmath103 in @xmath1 .
this identity holds in the sense that if one of the integrals exists , so does the other one , and they are equal .
it was shown in the previous section that the optimal prediction problem ( [ 214 ] ) is equivalent to the optimal stopping problem ( [ 221 ] ) .
the purpose of this section is to present the solution to the latter problem .
using the fact that the two problems are equivalent this also leads to the solution of the former problem . in the setting of ( [ 21])([25 ] )
consider the optimal stopping problem ( [ 221 ] ) .
this problem is two - dimensional and the underlying markov process equals @xmath107 .
setting @xmath108 for @xmath7 enables @xmath107 to start at @xmath109 under @xmath33 for @xmath91 in @xmath1 , and we will denote the resulting probability measure on the canonical space by @xmath110 .
thus under @xmath110 the canonical process @xmath107 starts at @xmath109 .
the problem ( [ 221 ] ) then extends as follows : @xmath111 for @xmath91 in @xmath1 where the infimum is taken over all stopping times @xmath13 of @xmath14 ( with finite mean ) .
in addition to @xmath112 and @xmath3 from ( [ 21 ] ) and ( [ 22 ] ) above , let us set @xmath113 for @xmath91 in @xmath1 .
the main result of this section may then be stated as follows .
[ theor3 ] the optimal stopping time in problem ( [ 31 ] ) is given by @xmath114 where the optimal boundary @xmath15 can be characterized as the minimal solution to @xmath115 } \int_i^{f(i ) } \frac{c_i'(i , y ) [ l(y ) - l(i)]}{\sigma^2(y ) l'(y ) } \,dy\ ] ] staying strictly above the curve @xmath17 for @xmath18 ( in the sense that if the minimal solution does not exist , then there is no optimal stopping time ) .
the value function is given by @xmath116}{(\sigma^2 /2)(y )
l'(y ) } \,dy\ ] ] for @xmath117 and @xmath118 for @xmath119 with @xmath18 .
it is evident from the integrand in ( [ 31 ] ) that the excursions of @xmath14 away from the running minimum @xmath28 play a key role in the analysis of the problem . in particular , recalling definition ( [ 32 ] ) , we see from ( [ 31 ] ) that the process @xmath107 can never be optimally stopped in the set @xmath120 where we let @xmath121 denote the state space of the process @xmath107 . indeed ,
if @xmath122 is given and fixed , then the first exit time of @xmath107 from a sufficiently small ball with the centre at @xmath109 ( on which @xmath123 is strictly negative ) will produce a value strictly smaller than @xmath38 ( the value corresponding to stopping at once ) . defining @xmath124 for @xmath18
we see that @xmath125 for @xmath126 and @xmath127 for @xmath128 whenever @xmath91 in @xmath1 are given and fixed .
note that the mapping @xmath129 is increasing and continuous as well as that @xmath130 for @xmath18 with @xmath131 and @xmath132 .
this shows that @xmath133 .
note in particular that @xmath134 contains the diagonal @xmath135 in the state space .
before we formalize further conclusions , let us recall that the general theory of optimal stopping for markov processes ( see @xcite , chapter 1 ) implies that the continuation set in the problem ( [ 31 ] ) equals @xmath136 , and the stopping set equals @xmath137 .
it means that the first entry time of @xmath107 into @xmath138 is optimal in problem ( [ 31 ] ) whenever well defined .
it follows therefore that @xmath134 is contained in @xmath139 , and the central question becomes to determine the remainder of the set @xmath139 . since @xmath4 @xmath33-a.s . as @xmath5 it follows that @xmath97 @xmath33-a.s . as @xmath58 so that the integrand in ( [ 31 ] ) becomes strictly positive eventually , and this reduces the incentive to continue ( given also that the `` favorable '' set @xmath134 becomes more and more distant ) .
this indicates that there should exist a point @xmath140 at or above which the process @xmath14 should be optimally stopped under @xmath110 where @xmath91 in @xmath1 are given and fixed .
this yields the following candidate : @xmath141 for an optimal stopping time in ( [ 31 ] ) where the function @xmath142 is to be determined .
free - boundary problem .
_ to compute the value function @xmath143 and determine the optimal function @xmath144 , we are led to formulate the free - boundary problem @xmath145 for @xmath18 , where @xmath146 is the infinitesimal generator of @xmath14 given in ( [ 222 ] ) above . for the rationale and further details regarding free - boundary problems of this kind ,
we refer to @xcite , section 13 , and the references therein ( we note , in addition , that the condition of normal reflection ( [ 39 ] ) dates back to @xcite )
. \4 . _
nonlinear differential equation .
_ to solve the free - boundary problem ( [ 38])([311 ] ) consider the stopping time @xmath147 defined in ( [ 37 ] ) and ( formally ) the resulting function @xmath148 for @xmath149 in @xmath1 given and fixed .
applying the strong markov property of @xmath107 at @xmath150 and using ( [ 223])([225 ] ) , we find that @xmath151 it follows from ( [ 313 ] ) that @xmath152\\[-8pt ] & & { } - \frac{l(f(i ) ) - l(i)}{l(f(i ) ) - l(x ) } \int_i^{f(i ) } c(i , y ) g_{i , f(i)}(x , y ) m(dy).\nonumber\end{aligned}\ ] ] using ( [ 310 ] ) and ( [ 311 ] ) we find after dividing and multiplying with @xmath153 that @xmath154 moreover , it is easily seen by ( [ 224 ] ) that @xmath155\\[-8pt ] & & \qquad = \int_i^{f(i ) } c(i , y ) [ l(y ) - l(i ) ] m(dy).\nonumber\end{aligned}\ ] ] combining ( [ 314])([316 ] ) we see that @xmath156 m(dy).\ ] ] inserting this back into ( [ 313 ] ) and using ( [ 224 ] ) and ( [ 225 ] ) , we conclude that @xmath157 m(dy)\ ] ] for @xmath149 in @xmath1 . finally , using ( [ 39 ] ) we find that @xmath158}\nonumber\\[-8pt]\\[-8pt ] & & \hspace*{6.5pt}{}\times\int_i^{f(i ) } c_i'(i , y ) [ l(y ) - l(i ) ] m(dy)\nonumber\end{aligned}\ ] ] for @xmath18 .
recalling that @xmath134 is contained in @xmath139 , we see that there is no restriction to assume that each candidate function @xmath144 solving ( [ 319 ] ) satisfies @xmath159 for all @xmath18 .
in addition we will also show below that all points @xmath160 belong to @xmath139 for @xmath18 so that ( at least in principle ) there would be no restriction to assume that each candidate function @xmath144 solving ( [ 319 ] ) also satisfies @xmath161 for all @xmath18 .
these candidate functions will be referred to as admissible .
we will also see below , however , that solutions to ( [ 319 ] ) `` starting '' at @xmath162 play a crucial role in finding / describing the solution . summarizing the preceding considerations , we can conclude that to each candidate function @xmath144 solving ( [ 319 ] ) , there corresponds the function ( [ 318 ] ) solving the free - boundary problem ( [ 38])([311 ] ) as is easily verified by direct calculation .
note , however , that this function does not necessarily admit the stochastic representation ( [ 312 ] ) ( even though it was formally derived from this representation ) .
the central question then becomes how to select the optimal boundary @xmath144 among all admissible candidates solving ( [ 319 ] ) . to answer this question
we will invoke the subharmonic characterization of the value function ( see @xcite , chapter 1 ) for the three - dimensional markov process @xmath163 where @xmath164 for @xmath165 .
fuller details of this argument will become clearer as we progress below .
it should be noted that among all admissible candidate functions solving ( [ 319 ] ) only the optimal boundary will have the power of securing the stochastic representation ( [ 312 ] ) for the corresponding function ( [ 318 ] ) . this is a subtle point showing the full power of the method ( as well as disclosing limitations of the optimal stopping problem itself ) .
\5 . _ the minimal solution .
_ motivated by the previous question we note from ( [ 318 ] ) that @xmath166 is decreasing over admissible solutions to ( [ 319 ] ) .
this suggests to select the candidate function among admissible solutions to ( [ 319 ] ) that is as far as possible from @xmath162 .
the subharmonic characterization of the value function suggests to proceed in the opposite direction , and this is the lead that we will follow in the sequel . to address the existence and uniqueness of solutions to ( [ 319 ] ) , denote the right - hand side of ( [ 319 ] ) by @xmath167 . from the general theory of nonlinear differential equations ,
we know that if the direction field @xmath168 is ( locally ) continuous and ( locally ) lipschitz in the second variable , then the equation ( [ 319 ] ) admits a ( locally ) unique solution .
for instance , this will be the case if , along a ( local ) continuity of @xmath169 , we also have a ( local ) continuity of @xmath170 .
in particular , we see from the structure of @xmath171 that equation ( [ 319 ] ) admits a ( locally ) unique solution whenever @xmath172 is ( locally ) continuously differentiable .
it is important to realize that the preceding arguments apply only away from @xmath162 since each point @xmath160 is a singularity point of equation ( [ 319 ] ) in the sense that @xmath173 when @xmath174 due to @xmath175 for @xmath18 . in this case
it is also important to note that the preceding arguments can be applied to the equivalent equation for the inverse of @xmath176 since this singularity gets removed ( the derivative of the inverse being zero ) . and @xmath15 to the differential equation ( [ 34 ] ) from theorem [ theor3 ] .
the optimal stopping boundary @xmath15 is the minimal solution staying strictly above the curve @xmath162 .
this is a genuine drawing corresponding to the golden ratio rule of corollary [ coro5 ] when @xmath14 is the radial part of three - dimensional brownian motion and the optimal stopping boundary @xmath15 is linear . ] to construct the minimal solution to ( [ 319 ] ) staying strictly above @xmath162 , we can proceed as follows ( see figure [ fig1 ] ) . for any @xmath177 such that @xmath178 as @xmath179 , let @xmath180 denote the solution to ( [ 319 ] ) on @xmath181 such that @xmath182 .
note that @xmath183 is singular at @xmath184 and that passing to the equivalent equation for the inverse of @xmath180 , this singularity gets removed as explained above .
( note that the solution to the equivalent equation for the inverse can be continued below @xmath185 as well until hitting the diagonal at some strictly positive point at which the derivative is @xmath186 .
this yields another solution to ( [ 319 ] ) staying below @xmath187 and providing its `` physical '' link to the diagonal .
we will not make use of this part of the solution in the sequel . )
note that the right - hand side of equation ( [ 319 ] ) is positive for @xmath161 so that @xmath183 is strictly increasing on @xmath188 . by the uniqueness of the solution
we know that the two curves @xmath180 and @xmath189 can not intersect for @xmath190 , and hence we see that @xmath191 is increasing .
it follows therefore that @xmath192 exists on @xmath1 . passing to an integral equation equivalent to ( [ 319 ] )
it is easily verified that @xmath193 solves ( [ 319 ] ) wherever finite .
this @xmath15 represents the minimal solution to ( [ 319 ] ) staying strictly above the curve @xmath162 on @xmath1 .
we will first consider the case when @xmath15 is finite valued on @xmath1 .
_ stochastic representation .
_ we show that the function ( [ 318 ] ) associated with the minimal solution @xmath15 admits the stochastic representation ( [ 312 ] ) . for this , let @xmath183 be the solution to ( [ 319 ] ) on @xmath181 such that @xmath194 for @xmath177 with @xmath178 as @xmath195 . consider the function @xmath196 defined by ( [ 318 ] ) for @xmath197 and @xmath198 with @xmath199 given and fixed .
recall that @xmath200 solves the free - boundary problem ( [ 38])([311 ] ) for @xmath201 .
consider the stopping time @xmath202 where @xmath203 and @xmath204 .
applying it s formula and using ( [ 38 ] ) , we find that @xmath205\\[-8pt ] & = & v_{f_n}(i , x ) + \int_0^{\tau_n } { \mathbb l}_x(v_{f_n } ) ( i_t , x_t ) \,dt\nonumber\\ & & { } + \int_0^{\tau_n } \sigma(x_t ) \,\frac{\partial v_{f_n}}{\partial x}(i_t , x_t ) \,db_t \nonumber\\ & = & v_{f_n}(i , x ) - \int_0^ { \tau_n } c(i_t , x_t ) \,dt + m_{\tau_n},\nonumber\end{aligned}\ ] ] where we also use ( [ 39 ] ) to conclude that the integral with respect to @xmath206 is equal to zero and @xmath207 is a continuous local martingale for @xmath7 .
since the process @xmath107 remains in the compact set @xmath208 up to time @xmath209 under @xmath210 , and both @xmath112 and @xmath211 are continuous ( and thus bounded ) on this set , we see that @xmath212 is a uniformly integrable martingale , and hence by the optional sampling theorem we have @xmath213 .
taking @xmath214 on both sides of ( [ 320 ] ) , we therefore obtain @xmath215\\[-8pt ] & = & v_{f_n}(i_n , i_n)\ : { \mathsf p}_{i , x } ( \tau_{i_n } < \tau_{f_n } ) + { \mathsf e}_{i , x } \int_0^ { \tau_n } c(i_t , x_t ) \,dt\nonumber\end{aligned}\ ] ] since @xmath216 on @xmath217 and @xmath218 on @xmath219 . using that @xmath220 we find by ( [ 317 ] ) , ( [ 223 ] ) and ( [ 36 ] ) that @xmath221\\[-8pt ] & & \qquad\le \vert l(h(i_n ) ) - l(i_n ) \vert\int_{i_n}^{h(i_n ) } m(dy)\ : \frac{l(f_*(i ) ) - l(x)}{l(f_*(i ) ) - l(i_n ) } \nonumber\\ & & \qquad= \frac{1}{2 } \vert l(i_n ) \vert\frac{l(f_*(i ) ) - l(x ) } { l(f_*(i ) ) - l(i_n ) } \int_{i_n}^{h(i_n ) } m(dy ) \rightarrow0\nonumber\end{aligned}\ ] ] as @xmath179 since @xmath222 and @xmath223 , so that @xmath224 , due to ( [ 25 ] ) above . hence letting @xmath195 in ( [ 321 ] ) and using that @xmath225 by the monotone convergence theorem , as well as that @xmath226 since @xmath227 and @xmath228 , we find noting that @xmath229 and using the dominated convergence theorem that @xmath230 for all @xmath231 in @xmath1 as claimed .
_ nonpositivity .
we show that for every solution @xmath144 to ( [ 319 ] ) such that @xmath232 on @xmath1 and the function @xmath233 defined by ( [ 318 ] ) above , we have @xmath234 for all @xmath231 in @xmath1 .
clearly , since @xmath235 for @xmath236 in ( [ 318 ] ) , it is enough to prove ( [ 324 ] ) for @xmath15 and @xmath237 with @xmath18 .
for this , consider the stopping time @xmath238 and note that @xmath239 .
hence by the strong markov property of @xmath107 applied at @xmath240 we find using ( [ 323 ] ) that @xmath241 \\ & = & { \mathsf e}_{i , x } \int_0^{\tau_h } c(i_t , x_t ) \,dt + { \mathsf e}_{i , x } { \mathsf e}_{i_{\tau_h},x_{\tau_h}}\biggl [ \int_0^{\tau_{f _ * } } c(i_t , x_t ) \,dt \biggr ] \nonumber\\ & = & { \mathsf e}_{i , x } \int_0^{\tau_h } c(i_t , x_t ) \,dt + { \mathsf e}_{i , x } v_{f_*}(i_{\tau_h } , x_{\tau_h } ) \le0,\nonumber\end{aligned}\ ] ] where the final inequality follows from the facts that @xmath242 for all @xmath243 $ ] and @xmath244 due to @xmath245 upon recalling ( [ 318 ] ) as already indicated above .
this completes the proof of ( [ 324 ] ) .
_ optimality of the minimal solution .
_ we will begin by disclosing the subharmonic characterization of the value function ( [ 31 ] ) in terms of the solutions to ( [ 319 ] ) staying strictly above @xmath162 .
for this , let @xmath176 be any solution to ( [ 319 ] ) satisfying @xmath161 for all @xmath18 .
consider the function @xmath246 defined by ( [ 318 ] ) for @xmath247 in @xmath1 and set @xmath248 for @xmath249 in @xmath1 .
let @xmath91 in @xmath1 be given and fixed .
due to the `` double - deck '' structure of @xmath233 , we can apply the change - of - variable formula from @xcite that in view of ( [ 311 ] ) reduces to standard it s formula and gives @xmath250\\[-8pt ] & = & v_f(i , x ) + \int_0^t { \mathbb l}_x(v_f)(i_s , x_s ) \,ds \nonumber\\ & & { } + \int_0^t \sigma(x_s)\ , \frac{\partial v_f}{\partial x}(i_s , x_s ) \,db_s,\nonumber\end{aligned}\ ] ] where we also use ( [ 39 ] ) to conclude that the integral with respect to @xmath251 is equal to zero .
the process @xmath252 defined by @xmath253 is a continuous local martingale . introducing the increasing process @xmath254 by setting @xmath255 and using the fact that the set of all @xmath256 for which @xmath257 equals @xmath258 is of lebesque measure zero
, we see by ( [ 38 ] ) that ( [ 326 ] ) can be rewritten as follows : @xmath259 from this representation we see that the process @xmath260 is a local submartingale for @xmath7 .
let @xmath13 be any stopping time of @xmath14 ( with finite mean ) .
choose a localization sequence @xmath261 of bounded stopping times for @xmath212 . then by ( [ 324 ] ) and
( [ 329 ] ) we can conclude using the optional sampling theorem that @xmath262 \nonumber\\[-8pt]\\[-8pt ] & \ge & v_f(i , x ) + { \mathsf e}_{i , x } m_{\tau\wedge \sigma_n } = v_f(i , x).\nonumber\end{aligned}\ ] ] letting @xmath179 and using the dominated convergence theorem ( upon recalling that @xmath220 as already used above ) we find that @xmath263 taking first the infimum over all @xmath13 , and then the supremum over all @xmath144 , we conclude that @xmath264 upon recalling that @xmath166 is decreasing over @xmath232 so that the supremum is attained at @xmath15 . combining ( [ 332 ] ) with ( [ 323 ] )
we see that ( [ 33 ] ) and ( [ 35 ] ) hold as claimed . note that ( [ 330 ] ) implies that the function @xmath265 is subharmonic for the markov process @xmath163 where @xmath164 for @xmath7 .
recalling that @xmath266 is decreasing over @xmath232 , and that @xmath267 for all @xmath91 in @xmath1 by ( [ 324 ] ) above , we see that selecting the minimal solution @xmath15 staying strictly above @xmath162 is equivalent to invoking the subharmonic characterization of the value function ( according to which the value function is the largest subharmonic function lying below the loss function ) . for more details on the latter characterization in a general setting
we refer to @xcite , chapter 1 .
it is also useful to know that the subharmonic characterization of the value function represents the dual problem to the primal problem ( [ 31 ] ) ( for more details on the meaning of this claim including connections to the legendre transform see @xcite ) .
consider finally the case when @xmath15 is not finite valued on @xmath1 . since @xmath193 is increasing we see that there is @xmath268 such that @xmath269 for all @xmath270 when @xmath271 and @xmath272 for all @xmath273 with @xmath274 when @xmath275 . if @xmath271 , then the proof above can be applied in exactly the same way to show that ( [ 33 ] ) and ( [ 35 ] ) hold as claimed under @xmath210 for all @xmath91 in @xmath1 with @xmath276 .
if @xmath273 with @xmath274 when @xmath275 , then the same proof shows that ( [ 35 ] ) still holds with @xmath277 in place of @xmath278 , however , the stopping time ( [ 33 ] ) can no longer be optimal in ( [ 31 ] ) .
this is easily seen by noting that the value in ( [ 35 ] ) is nonpositive ( it could also be @xmath186 ) for any @xmath279 for instance , while the @xmath210-probability for @xmath14 hitting @xmath280 before drifting away to @xmath277 is strictly smaller than @xmath281 so that the @xmath210-expectation over this set in ( [ 31 ] ) equals @xmath277 ( since the integrand tends to @xmath281 as @xmath282 tends to @xmath277 ) showing that the stopping time ( [ 33 ] ) can not be optimal .
the proof above shows that the optimality of ( [ 35 ] ) in this case is obtained through @xmath283which play the role of approximate stopping times ( obtained by passing to the limit when @xmath284 tends to @xmath277 in ( [ 321 ] ) above ) .
this completes the proof of the theorem .
in this section we show that the minimal solution to ( [ 34 ] ) admits a simple closed - form expression when @xmath14 is a transient bessel process ( theorem [ theor4 ] ) . in the case when @xmath14 is the radial part of three - dimensional brownian motion this leads to the golden ratio rule ( corollary [ coro5 ] ) .
we also show that @xmath14 stopped according to the golden ratio rule has what we refer to as the golden ratio distribution ( corollary [ coro8 ] ) . in the setting of ( [ 26])([28 ] ) consider the optimal prediction problem ( [ 214 ] ) . recall that this problem is equivalent to the optimal stopping problem ( [ 221 ] ) which further extends as ( [ 31 ] ) .
the main result of this section can now be stated as follows .
[ theor4 ] if @xmath14 is the @xmath39-dimensional bessel process solving ( [ 26 ] ) with , then the optimal stopping time in ( [ 214 ] ) is given by @xmath285 where @xmath286 is the unique solution to @xmath287 belonging to @xmath288 .
the value function ( [ 31 ] ) is given explicitly by @xmath289 \qquad\mbox{if } d \ne4 \nonumber\\[-8pt]\\[-8pt ] & = & \biggl [ x^2 \biggl ( \frac{1}{2 } + \biggl(\frac{i}{x } \biggr)^2 \biggr ) \biggl ( \biggl(\frac{\lambda i}{x}\biggr)^2 - 1 \biggr ) \nonumber\\ & & \hspace*{2.6pt } { } - \frac{x^2}{4 } \biggl ( \biggl ( \frac{\lambda i}{x } \biggr)^4 - 1 \biggr ) - 2 i^2 \log\biggl ( \frac { \lambda i}{x } \biggr ) \biggr ] \qquad\mbox{if } d = 4\nonumber\end{aligned}\ ] ] for @xmath290 and @xmath118 for @xmath291 with @xmath18 . by the result of theorem [ theor3 ]
we know that the optimal stopping time @xmath292 is given by ( [ 33 ] ) above where the optimal boundary @xmath15 can be characterized as the minimal solution to ( [ 34 ] ) staying strictly above the curve @xmath293 for @xmath18 . using ( [ 27 ] ) and ( [ 32 ] )
it can be verified that ( [ 34 ] ) reads as follows : @xmath294}\biggr)\nonumber\\ & & \hspace*{0pt}{}\times\biggl({\biggl ( \biggl(\frac{f(i)}{i}\biggr)^{d-2 } - 1 \biggr ) \biggl ( \biggl ( \frac{f(i)}{i}\biggr)^{d-2 } - 2 \biggr)}\biggr)^{-1 } \qquad\mbox{if } d \ne4 \\ & = & \frac{2 ( { f(i ) } /{i } ) [ ( { f(i)}/{i})^2 -2 \log({f(i)}/ { i } ) - 1 ] } { ( ( { f(i)}/{i})^2 - 1 ) ( ( { f(i)}/{i})^2 - 2 ) } \qquad\mbox{if } d = 4\nonumber\end{aligned}\ ] ] and @xmath295 for @xmath18 . hence it is enough to show that @xmath296 is the minimal solution to ( [ 45 ] ) staying strictly above the curve @xmath295 for @xmath18 . to show that @xmath15 is a solution to ( [ 45 ] ) staying strictly above @xmath162 , insert @xmath297 into ( [ 45 ] ) with @xmath298 to be determined . multiplying both sides of the resulting identity by @xmath299 ( to be able to derive the factorization ( [ 47 ] ) below ) it is easy to see that this yields the equation @xmath300 where we set @xmath301\\[-8pt ] & = & \lambda^4 - 5 \lambda^2 + 4 \log \lambda+ 4 \qquad\mbox{if } d = 4\nonumber\end{aligned}\ ] ] for @xmath302 .
after some algebraic manipulations we find that @xmath303 for @xmath302 and @xmath29 .
hence we see that the equation @xmath304 has two roots @xmath305 and @xmath306 where @xmath307 .
it is easy to check that @xmath308 and @xmath309 showing that @xmath310 has a local maximum at @xmath311 and @xmath310 has a local minimum at @xmath312 . noting that @xmath313 , @xmath314 and @xmath315 this shows that ( i ) @xmath310 is strictly increasing on @xmath316 with @xmath313 and ; ( ii ) @xmath310 is strictly decreasing on @xmath317 with @xmath314 and ; and ( iii ) @xmath310 is strictly increasing on @xmath318 with @xmath315 .
it follows therefore that the equation @xmath300 has exactly three roots @xmath319 where @xmath320 and @xmath321 . setting @xmath322
this shows that @xmath296 is a solution to ( [ 45 ] ) staying strictly above the curve @xmath295 for @xmath18 as claimed . to show that @xmath15 is the minimal solution satisfying this property , set @xmath323 and note that ( [ 45 ] ) can then be rewritten as follows : @xmath324 for @xmath18 . since @xmath325 for @xmath326 we see from ( [ 48 ] ) that @xmath327 is increasing for @xmath328 .
noting that ( [ 48 ] ) implies that @xmath329 it follows therefore that the integrand on the left - hand side is bounded by a constant ( not dependent on @xmath280 ) as long as @xmath328 for @xmath330 with any @xmath331 given and fixed . letting then @xmath332 in ( [ 49 ] ) we see that the left - hand side remains bounded while the right - hand side tends to @xmath277 leading to a contradiction . noting that @xmath328 if and only if @xmath333 , we can therefore conclude that there is no solution @xmath144 to ( [ 45 ] ) satisfying @xmath334 for @xmath18 .
thus @xmath15 is the minimal solution to ( [ 45 ] ) staying strictly above @xmath162 and the proof is complete .
[ coro5 ] if @xmath14 is the radial part of three - dimensional brownian motion , then the optimal stopping time in ( [ 214 ] ) is given by @xmath335 where @xmath20 is the golden of three - dimensional brownian motion . ]
ratio ( see figure [ fig2 ] ) .
in this case @xmath336 and equation ( [ 42 ] ) reads @xmath337 for @xmath298 . solving the latter quadratic equation and choosing the root strictly greater than @xmath281 , we find that @xmath338 where @xmath20 is the golden ratio .
the optimality of ( [ 410 ] ) then follows from ( [ 41 ] ) and the proof is complete .
returning to the result of theorem [ theor3 ] above we now determine the law of the transient diffusion process @xmath14 stopped at the optimal stopping time @xmath292 ( for related results on the skorokhod embedding problem see @xcite , pages 269277 , and the references therein ) .
[ prop6 ] in the setting of theorem [ theor3 ] we have @xmath339 for @xmath340 with @xmath27 .
note that @xmath341 where we set @xmath342 , @xmath343 and @xmath344 for @xmath7 .
let @xmath27 be given and fixed
. for @xmath345 set @xmath346 where @xmath347 \rightarrow{\mathbb r}$ ] is a continuously differentiable function with bounded support .
using the fact that @xmath348 when @xmath349 it is easily verified by it s formula that the process @xmath350 defined by @xmath351 is a continuous local martingale .
moreover , since @xmath352 is continuous and has bounded support , we see that @xmath353 is bounded and therefore uniformly integrable . by the optional sampling theorem we thus find that @xmath354 \nonumber\\ & = & \int_{-\infty}^{l(x ) } g(j ) \,df(j ) + { \mathsf e}_x \bigl [ \bigl(f_*^l ( i_{\tau_*}^l ) - i_{\tau_*}^l\bigr ) g'(i_{\tau_*}^l ) \bigr]\nonumber\\ & = & g(j ) f(j ) \vert_{-\infty}^{l(x ) } - \int_{-\infty } ^{l(x ) } f(j ) \,dg(j)\nonumber\\[-8pt]\\[-8pt ] & & { } + \int_{-\infty}^{l(x ) } \bigl(f_*^l(j ) - j\bigr ) g'(j ) \,df(j ) \nonumber\\ & = & g^l(x ) - \int_{-\infty}^{l(x ) } f(j ) g(j ) \,dj\nonumber\\ & & { } + \int_{-\infty}^{l(x ) } \bigl(f_*^l(j ) - j\bigr ) g(j ) \,df(j),\nonumber\end{aligned}\ ] ] where we set @xmath355 and @xmath310 denotes the distribution function of @xmath356 under @xmath33 .
since ( [ 415 ] ) holds for all functions @xmath357 of this kind , it follows that @xmath358 for @xmath359 with @xmath360 .
solving ( [ 416 ] ) under this boundary condition we find that @xmath361 for @xmath362 . recalling that @xmath343 and substituting @xmath363 it follows that @xmath364 for @xmath91 in @xmath1 .
hence we find that @xmath365\\[-8pt ] & = & \exp\biggl ( - \int_{f_*^{-1}(y)}^x \frac{dl(z)}{l(f_*(z ) ) - l(z ) } \biggr)\nonumber\end{aligned}\ ] ] for @xmath340 with @xmath27 .
this completes the proof . specializing this result to the @xmath39-dimensional bessel process @xmath14 of theorem [ theor4 ] we obtain the following consequence .
[ coro7 ] in the setting of theorem [ theor4 ] we have @xmath366 for @xmath367 with @xmath27 . in this case
@xmath296 for @xmath18 where @xmath286 is the unique solution to either ( [ 42 ] ) when @xmath368 or ( [ 43 ] ) when @xmath369 and @xmath3 is given by ( [ 27 ] ) . inserting these expressions into the right - hand side of ( [ 412 ] )
it is easily verified that this yields ( [ 420 ] ) .
specializing this further to the radial part @xmath14 of three - dimensional brownian motion in corollary [ coro5 ] we obtain the following conclusion .
[ coro8 ] in the setting of corollary [ coro5 ] we have @xmath370 for @xmath371 with @xmath27 . in this case @xmath336 and @xmath372 so that @xmath373 .
hence we see that ( [ 420 ] ) reduces to ( [ 421 ] ) .
note from ( [ 421 ] ) that the density function of @xmath374 under @xmath33 is given by @xmath375 for @xmath376 with @xmath27 and equals zero otherwise .
we refer to ( [ 421 ] ) and ( [ 422 ] ) as the _ golden ratio distribution_. it is easy to see that @xmath377 for @xmath27 .
the fact that this number is strictly greater than @xmath378 ( the initial point corresponding to stopping at once ) is not surprising since @xmath14 is a submartingale .
it needs to be recalled moreover that the aim of applying the golden ratio rule @xmath292 is to be as close as possible to the time @xmath8 at which the ultimate minimum is attained .
we will see in the next section that the golden ratio distribution provides insight as to what extent the golden ratio rule has the power of capturing the ultimate maximum of a strict local martingale.=1
in this section we present some applications of the previous results in problems of optimal trading .
we also outline some remarkable connections between such problems and the practice of technical analysis .
these applications and connections rest on three basic ingredients that we describe first .
_ fibonnaci retracement .
_ we begin by explaining a few technical terms from the field of applied finance . _
technical analysis _ is a financial term used to describe methods and techniques for forecasting the direction of asset prices through the study of past market data ( primarily prices themselves plus the volume of their trade ) .
_ support _ and _ resistance _ are concepts in technical analysis associated with the expectation that the movement of the asset price will tend to cease and reverse its trend of decrease / increase at certain predetermined price levels .
a _ support_/_resistance level _ is a price level at which the price will tend to find support / resistance when moving down / up .
this means that the price is more likely to bounce off this level rather than break through it .
one may also think of these levels as turning points of the prices .
_ fibonacci retracement _ is a method of technical analysis for determining support and resistance levels .
the name comes after its use of _ fibonacci numbers _
@xmath379 for @xmath199 with @xmath380 and @xmath381 .
fibonacci retracement is based on the idea that after reversing the trend at a support / resistance level , the price will retrace a predictable portion of the past downward / upward move by advancing in the opposite direction until finding a new resistance / support level , after which it will return to the initial trend of moving downwards / upwards .
fibonacci retracement is created by taking two extreme points on a chart showing the asset price as a function of time and dividing the vertical distance between them by the key fibonacci ratios ranging from @xmath382 ( start of the retracement ) to @xmath383 ( end of the retracement representing a complete reversal to the original trend ) .
the other key fibonacci ratios are @xmath384 ( shallow retracement ) , @xmath385 ( moderate retracement ) and @xmath386 ( golden retracement ) .
they are obtained by formulas @xmath387 , @xmath388 and @xmath389 , respectively ( see the next paragraph ) .
these retracement levels serve as alert points for a potential reversal at which traders may employ other methods of technical analysis to identify and confirm a reversal . despite its widespread use in technical analysis of asset prices , there appears to be no ( rigorous ) explanation of any kind as to why the fibonacci ratios should be used to this effect .
we will show below that the golden ratio rule derived in the previous section offers a rigorous optimality argument for the choice of the golden retracement ( @xmath386 ) . to our knowledge
this is the first time that such an argument has been found / given in the literature .
golden ratio and fibonacci numbers .
_ the link between the two is well known and is expressed by binet s formula @xmath390 where @xmath391 and @xmath392 .
it follows that @xmath393 this fact is used in the description of fibonacci retracement above .
\3 . _ the cev model .
_ one of the simplest / tractable models for asset price movements that is capable of reproducing the implied volatility smile / frown effect and the ( inverse ) leverage effect ( both observed in the empirical data ) is the _ constant elasticity of variance _ ( cev ) model in which the ( nonnegative ) asset price process @xmath394 solves @xmath395 where @xmath396 is the appreciation rate , @xmath2 is the volatility coefficient , and @xmath397 is the elasticity parameter . if @xmath398 then @xmath399 is a geometric brownian motion which was initially considered in @xcite and @xcite . for @xmath400 this model
was firstly considered in @xcite for @xmath401 and then in @xcite for @xmath402 . due to its predictive power and tractability ,
the cev model is widely used by practitioners in the financial industry , especially for modeling prices of equities and commodities .
if @xmath401 then the model embodies the _ leverage effect _ ( commonly observed in equity markets ) where the volatility of the asset price increases as its price decreases . if @xmath402 then the model embodies the _ inverse leverage effect _
( often observed in commodity markets ) where the volatility of the asset price increases when its price increases .
for example , it is reported in @xcite that the elasticity coefficient @xmath403 for gold on the london bullion market in the period from 2000 to 2007 was approximately 0.49 .
similar elasticity coefficients have also been observed for other precious metals ( such as copper for instance ) . in the remainder of this section we focus on the case
when @xmath95 and .
it is well known ( cf .
@xcite ) that @xmath399 solving ( [ 53 ] ) is a strict local martingale ( a local martingale which is not a true martingale ) in this case due to the fact that @xmath404 is strictly decreasing on @xmath405 for any @xmath406 .
this also implies that @xmath399 does not admit an equivalent martingale measure so that the cev model may admit arbitrage opportunities .
one way of looking at the models of this type is to associate them with asset price bubbles ( see @xcite ) . after soaring to a finite ultimate maximum ( bubble ) at a finite time
, the asset price will tend to zero as time goes to infinity , and the central question for a holder of the asset becomes when to sell so as to be as close as possible to the time at which the ultimate maximum is attained .
more precisely , introducing the running maximum process @xmath407 associated with @xmath399 by setting @xmath408 and recalling that @xmath409 as @xmath58 , we see that the ultimate supremum @xmath410 is attained at some random time @xmath8 in the sense that @xmath411 with @xmath412-probability one for @xmath406 given and fixed .
the optimal selling problem addressed above then becomes the optimal prediction problem @xmath413 where the infimum is taken over all stopping times @xmath13 of @xmath399 ( with finite mean ) and @xmath406 is given and fixed .
we will now show that due to the well - known connection between cev and bessel processes ( dating back to similar transformations in @xcite and @xcite ) the problem ( [ 56 ] ) can be reduced to the problem ( [ 214 ] ) solved above .
\4 . _ the golden ratio rule for the cev process .
_ for @xmath29 given and fixed consider the @xmath39-dimensional bessel process @xmath14 solving ( [ 26 ] ) under @xmath33 with @xmath27 . recall that the scale function @xmath3 is given by ( [ 27 ] ) and set @xmath414 for @xmath27 with @xmath415 given and fixed .
then the process @xmath416 defined by @xmath417 is on natural scale and it s formula shows that @xmath399 solves @xmath418 where @xmath419 and @xmath420 is a standard brownian motion . note that equation ( [ 58 ] ) coincides with equation ( [ 53 ] ) for @xmath95 and @xmath421 . by the uniqueness in law for this equation ( among positive solutions )
it follows that @xmath416 is a cev process . from the properties of
@xmath14 it follows that after starting at @xmath422 , the process @xmath399 stays strictly positive ( without exploding at a finite time ) and @xmath409 with @xmath412-probability one as @xmath58 .
this shows that @xmath8 in ( [ 55 ] ) is well defined .
moreover , due to the reciprocal relationship ( [ 57 ] ) we see that the time of the ultimate maximum @xmath8 for @xmath399 in ( [ 55 ] ) coincides with the time of the ultimate minimum @xmath8 for @xmath14 in ( [ 210 ] ) , and hence the problem ( [ 56 ] ) has the same solution as the problem ( [ 214 ] ) ( note also that the natural filtrations of @xmath399 and @xmath14 coincide so that @xmath13 is a stopping time of @xmath399 if and only if @xmath13 is a stopping time of @xmath14 ) . since @xmath423 if and only if @xmath424 it follows from ( [ 41 ] ) in theorem [ theor4 ] that the optimal stopping time in ( [ 56 ] ) is given by @xmath425 where @xmath286 is the unique solution to either ( [ 42 ] ) or ( [ 43 ] ) belonging to @xmath288 .
in particular , if @xmath336 then we know from ( [ 411 ] ) that @xmath426 so that ( [ 59 ] ) reads @xmath427 this is the _ golden ratio rule _ for the cev process @xmath428 where @xmath14 is the radial part of three - dimensional brownian motion ( see figure [ fig3 ] ) . to relate the golden ratio rule ( [ 510 ] ) to fibonacci retracement discussed above , let @xmath429 denote the larger quantity and let @xmath430 denote the smaller quantity in the golden ratio rule . to determine the percentage of @xmath21 in @xmath23 we need to calculate @xmath431\\[-8pt ] & = & \frac{\varphi}{1 + \varphi } = \frac{1}{\varphi}.\nonumber\end{aligned}\ ] ] multiplying this expression by @xmath432 gives @xmath386 and this is exactly the _ golden retracement _ discussed above . in view of the optimality of ( [ 510 ] ) in ( [ 56 ] ) we see that the golden retracement of @xmath433 for the cev process @xmath428 ( starting close to zero ) where @xmath14 is the radial part of three - dimensional brownian motion can be seen as a rational support level ( in the sense that rational investors who aim at selling the asset at the time of the ultimate maximum will sell the asset at the time of the golden retracement , and therefore the asset price could be expected to raise afterwards ) . to our knowledge
this is the first time that such a rational optimality argument for the golden retracement has been established . | let @xmath0 be a transient diffusion process in @xmath1 with the diffusion coefficient @xmath2 and the scale function @xmath3 such that @xmath4 as @xmath5 , let @xmath6 denote its running minimum for @xmath7 , and let @xmath8 denote the time of its ultimate minimum @xmath9 .
setting @xmath10 we show that the stopping time @xmath11 minimizes @xmath12 over all stopping times @xmath13 of @xmath14 ( with finite mean ) where the optimal boundary @xmath15 can be characterized as the minimal solution to @xmath16 } \int_i^{f(i ) } \frac{c_i'(i , y ) [ l(y ) - l(i)]}{\sigma^2(y ) l'(y ) } \,dy\ ] ] staying strictly above the curve @xmath17 for @xmath18 . in particular , when @xmath14 is the radial part of three - dimensional brownian motion , we find that @xmath19 where @xmath20 is the golden ratio .
the derived results are applied to problems of optimal trading in the presence of bubbles where we show that the golden ratio rule offers a rigorous optimality argument for the choice of the well - known golden retracement in technical analysis of asset prices . , . . |
considering the state of the art @xcite , we propose a new method for multivariate approximation which allows to interpolate large scattered data sets stably , accurately and with a relatively low computational cost .
the interpolant we consider is expressed as a linear combination of some basis or kernel functions . focusing on radial basis functions ( rbfs ) ,
the partition of unity ( pu ) is performed by blending rbfs as local approximants and using locally supported weight functions . with this approach
a large problem is decomposed into many small problems , @xcite , and therefore in the approximation process we could work with a large number of nodes .
however , in some cases , local approximants and consequently also the global one suffer from instability due to ill - conditioning of the interpolation matrices .
this is directly connected to the order of smoothness of the basis function and to the node distribution .
it is well - known that the stability depends on the flatness of the rbf .
more specifically , if one keeps the number of nodes fixed and considers smooth basis functions , then the problem of instability becomes evident for small values of the shape parameter .
of course , a basis function with a finite order of smoothness can be used to improve the conditioning but the accuracy of the fit gets worse .
for this reason , the recent research is moved to the study of more stable bases . for particular rbfs , techniques allowing to stably and accurately compute the interpolant , also in the _ flat limit _
@xmath0 , have been designed in the recent years .
these algorithms , named rbf - qr methods , are all rooted in a particular decomposition of the kernel , and they have been developed so far to treat the gaussian and the inverse multiquadric kernel .
we refer to @xcite for further details on these methods .
a different and more general approach , consisting in computing , via a truncated singular value decomposition ( svd ) stable bases , namely weighted svd ( wsvd ) bases , has been presented in @xcite .
we remark that in the cases where the rbf - qr algorithms can be applied , they produce a far more stable solution of the interpolation problem .
nevertheless , the present technique applies to _ any _ rbf kernel , and to any domain . in this paper , a stable approach via the pu method , named wsvd - pu , which makes use of local wsvd bases and uses compactly supported weight functions ,
is presented .
thus , following @xcite , for each pu subdomain a stable rbf basis is computed in order to solve the local interpolation problem .
consequently , since the local approximation order is preserved for the global fit , the interpolant results more stable and accurate .
concerning the stability , we surely expect a more significant improvement in the stabilization process with infinitely smooth functions than with functions characterized by a finite order of regularity . moreover , in terms of accuracy , the benefits coming from the use of such stable bases are more significant in a local approach than in a global one .
in fact , generally , while in the global case a large number of truncated terms of the svd must be dropped to preserve stability , a local technique requires only few terms are eliminated , thus enabling the method to be much more accurate .
concerning the computational complexity of the algorithm , the use of the so - called block - based space partitioning data structure enables us to efficiently organize points among the different subdomains , @xcite .
then , for each subdomain a local rbf problem is solved with the use of a stable basis .
the main and truly high cost , involved in this step , is the computation of the svd . to avoid this drawback , techniques based on krylov space methods
are employed , since they turn out to be really effective , @xcite .
a complexity analysis supports our findings .
the guidelines of the paper are as follows . in section [ wsvd ]
, we present the wsvd bases , computed by means of the lanczos algorithm , in the general context of global approximation .
such method is used coupled with the pu approach which makes use of an optimized searching procedure , as shown in section [ pum_wsvd ] .
the proposed approach turns out to be stable and efficient , as stressed in section [ compl_anal_wsvd ] . in sections
[ ne_lan ] and [ applicazione ] extensive numerical experiments and applications , carried out with both globally and compactly supported rbfs of different orders of smoothness , support our results .
moreover , all the matlab codes are made available to the scientific community in a downloadable free software package : http://hdl.handle.net/2318/1527447 .
in subsection [ rbf_prelim ] we briefly review the main theoretical aspects concerning rbf interpolation , @xcite , while the remaining subsections are devoted to the efficient computation of the wsvd basis via krylov space methods .
our goal is to recover a function @xmath1 , @xmath2 being a bounded set in @xmath3 , using a set of samples of @xmath4 on @xmath5 pairwise distinct points @xmath6 , namely @xmath7^t$ ] , @xmath8 , @xmath9 . to this end ,
one considers a positive definite and symmetric kernel @xmath10 to construct an interpolant in the form @xmath11 the kernels we consider are always radial , meaning that there exist a positive _ shape parameter _ @xmath12 and a function @xmath13 such that for all @xmath14 , @xmath15 . in table
[ tab_rbf ] we report a list of some strictly positive definite radial kernels with their smoothness degrees .
we remark that gaussian , inverse multiquadric and mat@xmath16rn functions are globally supported and strictly positive definite in @xmath3 for any @xmath17 , whereas wendland ones are compactly supported ( whose support is @xmath18 $ ] ) and strictly positive definite in @xmath3 for @xmath19 ( see @xcite ) .
ll ' '' '' rbf & @xmath20 + ' '' '' gaussian @xmath21 ( ga ) & @xmath22 + ' '' '' inverse multiquadric @xmath21 ( imq ) & @xmath23 + ' '' '' mat@xmath16rn @xmath24 ( m6 ) & @xmath25 + ' '' '' mat@xmath16rn @xmath26 ( m4 ) & @xmath27 + ' '' '' wendland @xmath24 ( w6 ) & @xmath28 + ' '' '' wendland @xmath26 ( w4 ) & @xmath29 + the real coefficients @xmath30^t$ ] in are determined by solving the linear system @xmath31 , where the interpolation ( or kernel ) matrix @xmath32 is given by @xmath33 the so constructed solution @xmath34 is a function of the _ native hilbert space _ @xmath35
uniquely associated with the kernel , and , if @xmath36 , it is in particular the @xmath37-projection of @xmath4 into the subspace @xmath38 although this interpolation method is known to be highly unstable in most cases being the matrix @xmath39 severely ill conditioned , it has been proven ( see @xcite ) that the interpolation operator @xmath40 is stable as an operator in the function space @xmath37 .
this gap has been widely recognized to be caused by the use of the standard basis , and a lot of efforts have been made in recent years to introduce better or perfectly conditioned basis ( see @xcite for a general theoretical treatment of this topic , and @xcite for particular instances of stable basis ; for an overview see the book @xcite ) .
we are interested here in the use of the _ wsvd basis _ introduced in @xcite , thanks to its flexibility with respect to the choice of the kernel @xmath41 .
we recall in the following some relevant properties of this basis , while we refer to the paper @xcite for further details . to construct a basis @xmath42 of @xmath43 it is enough to assign an invertible coefficient matrix @xmath44_{i , j=1}^n$ ] such that @xmath45 or , equivalently , an invertible value matrix @xmath46_{i , j=1}^n$ ] .
the two matrices are related as @xmath47 ( see @xcite ) , and in our situation they are defined as follows ( see @xcite ) .
a wsvd basis @xmath48 is a basis for @xmath49 characterized by the matrices @xmath50 is a singular value decomposition of the scaled kernel matrix @xmath51 , and @xmath52 is a diagonal matrix of positive weights .
we observe that the definition uses a set of positive weights that was employed in the original formulation as cubature weights to construct the basis .
nevertheless , these weights do not change the numerical behavior of the basis , hence we assume from now on @xmath53 , @xmath54 . moreover , for notational convenience , the diagonal elements of @xmath55 are denoted as @xmath56 .
this basis has been introduced to mimic in a discrete sense the properties of the _ eigenbasis_. such basis is constructed starting from the operator @xmath57 , @xmath58(\boldsymbol{x } ) = \int_{\omega } \phi(\boldsymbol{x},\boldsymbol{y } ) f(\boldsymbol{y } ) d\boldsymbol{y},\ ] ] through the following theorem ( see e.g. @xcite ) . if the kernel @xmath41 is continuous and positive definite on a bounded set @xmath59 , the operator @xmath60 has a countable set of eigenfunctions @xmath61 and eigenvalues @xmath62 .
the eigenfunctions are orthonormal in @xmath63 and orthogonal in @xmath37 with @xmath64 .
moreover , the kernel can be expressed in terms of the eigencouples as @xmath65 where the series converges uniformly and absolutely . for its use in interpolation in @xmath37 ,
it is convenient to use the basis @xmath66 , that is normalized in @xmath37 . with this normalization , the basis has the following properties .
the eigenbasis @xmath66 has the following properties : a. it is @xmath37-orthonormal , b. it is @xmath63-orthogonal with squared norm @xmath67 , c. @xmath68 , @xmath69 , d. @xmath70 and @xmath71 as @xmath72 , e. @xmath73 .
as proven in @xcite , the wsvd basis enjoys the same properties when the inner product of @xmath63 is replaced with its discrete version @xmath74 , as summarized in the following statement .
[ discrete properties ] the wsvd basis @xmath75 has the following properties : a. it is @xmath37-orthonormal , b. it is @xmath74-orthogonal with norm @xmath76 , c. [ double ] @xmath77 , @xmath69 , d. @xmath78 , e. [ discrete sum condition ] @xmath79 .
since the interpolation is a @xmath37-projection , we can rewrite the interpolant @xmath34 in terms of the @xmath37-orthonormal wsvd basis as @xmath80 and , thanks to point of property [ discrete properties ] , this can be further rewritten as @xmath81 the latter form of the interpolant shows that @xmath34 is also the solution of the discrete least - squares approximation problem @xmath82 among all functions in @xmath83 . if we instead solve the minimization problem over the subspace @xmath84 , @xmath85 , we find a solution @xmath86 given by the truncation of the interpolant , i.e. , @xmath87 observe that it makes sense to consider the last minimization problem and its solution @xmath86 for some @xmath85 instead of the original interpolation problem , since in this way we leave out the portion of the subspace @xmath88 corresponding to small singular values , and this corresponds to solve the linear system by means of a low - rank approximation of the matrix @xmath39 .
a detailed discussion of this approach can be found in @xcite , but we remark here that this method strictly depends on the behavior of the singular values of @xmath39 .
namely , if we consider smoother rbfs , then by using instead of the standard approach a better stabilization of the interpolation process is expected . on the other hand
, this method has some disadvantages .
first , it is required to compute a singular value decomposition of the ( possibly large ) kernel matrix , and in the end only a few elements of the decomposition are used .
this is computationally expensive , but in the next section we explain how to overcome this problem .
second , this method requires to neglect part of the information to reduce instability , and , in some cases , this removal is too big to obtain a meaningful approximant .
a solution to this problem is provided by the coupling with a localization method , and it is the main topic of this paper .
we present here a way to compute an approximation of the wsvd basis that makes use of the lanczos algorithm .
the method is discussed in @xcite , and it aims at reducing the computational cost of the procedure by approximating the truncated svd of @xmath39 .
we start by a general description of the lanczos algorithm .
further details can be found in @xcite .
let @xmath89 be the krylov subspace of order @xmath90 generated by the matrix @xmath39 and the vector @xmath91 .
the lanczos method computes an orthonormal basis @xmath92 of @xmath93 through a gram - schmidt orthonormalization , i.e. , the lanczos basis @xmath94 , is computed by the following recurrence formula : @xmath95 letting @xmath96 the matrix having the vectors @xmath97 as columns , and letting @xmath98 be the @xmath99 tridiagonal matrix defined as @xmath100 the algorithm can be formulated in matrix form as @xmath101 where @xmath102 is the unit vector and @xmath103 is a scalar value .
once we compute the matrices , the solution of the initial system can be approximated as @xmath104 , where @xmath105 is such that @xmath106 .
the idea is to use the matrix @xmath107 to approximate the svd of @xmath39 , and since @xmath39 has usually a good low - rank approximation , we expect to do so with @xmath108 .
specifically , let @xmath109 be a singular value decomposition of @xmath107 , where @xmath110 , @xmath111 are unitary matrices and @xmath112 with @xmath113 the diagonal matrix with singular values @xmath114 , @xmath115 on the diagonal . since the last row of @xmath116 is the zero vector , the decomposition does not change if we remove this row and the last column of @xmath117 . thus , to simplify the notation we denote by @xmath117 the matrix without the last column , so that the decomposition becomes @xmath118 .
now we want to define an approximation of the wsvd basis using the approximate svd of @xmath39 .
we define a set of functions @xmath119 which shows similarities with the wsvd basis , even if it does not form a basis , since they do not span @xmath120 . anyway , with abuse of notation , we call again this set of functions a _ basis_. the approximate wsvd basis is characterized by the matrices @xmath121 where @xmath122 is the lanczos decomposition of @xmath39 of order @xmath90 and @xmath123 is a singular value decomposition of @xmath107 .
the next statement clarifies the connection between the two basis .
as it is evident by construction , the basis strongly depends on the particular function @xmath124 .
[ approx discrete properties ] the approximate wsvd basis @xmath125 has the following properties : a. it is near @xmath37-orthonormal , meaning that its @xmath37-gramian is the identity matrix plus a rank one matrix , b. it is @xmath74-orthogonal with norm @xmath76 , c. [ approx double ] @xmath126 if @xmath4 is the function used to construct the basis , d. @xmath127 , e. [ m = n ] it coincides with the wsvd basis if @xmath128 .
this basis allows to solve again the least square approximation problem .
namely , if @xmath36 is the function used for the lanczos algorithm , the approximant @xmath129 defined as @xmath130 minimizes the distance @xmath131 for @xmath132 , @xmath85 .
moreover , thanks to property of property [ approx discrete properties ] , @xmath129 can be written in terms of @xmath37-inner products as @xmath133 note that the point of property [ approx discrete properties ] proves that @xmath134 .
approximating @xmath86 with its fast computable version @xmath129 solves efficiently the problems .
we will see in the following sections how to successfully couple this technique with a fast domain decomposition method .
the main idea is to use the stable basis introduced in the previous section in order to generate local stable approximants and accumulate them into the global fit .
let @xmath135 be an open and bounded domain , and let @xmath136 be an open and bounded covering of @xmath2 satisfying some mild overlap condition among the subdomains ( or patches ) @xmath137 .
in other words , the subdomains must form a covering of the domain and , moreover , the overlap must be sufficient so that each interior point @xmath138 is located in the interior of at least one patch @xmath137 . furthermore , let us suppose that the set @xmath139 , for @xmath138 , is uniformly bounded on @xmath2 , with @xmath140 . associated with the subdomains we choose partition of unity weight functions @xmath141
, i.e. a family of compactly supported , nonnegative and continuous functions subordinate to the subdomain @xmath137 , such that @xmath142 on @xmath2 and @xmath143 .
in order to have a better stabilization of the global fit , our goal is to define stable approximants of the form on each subdomain @xmath137 . in other words , for each ( local ) matrix @xmath144 , i.e. the @xmath145-th interpolation matrix associated with the subdomain @xmath137 , a low - rank approximation is computed and thus the so - called wsvd - pu approximant is given by : @xmath146 where @xmath147 is a partition of unity weight function and @xmath148 . as evident
, @xmath149 defines a local stable rbf approximant on @xmath137 of the form : @xmath150 according to @xcite , if we assume to have a @xmath151-stable partition of unity , then the derivatives of the weight functions satisfy @xmath152 is the diameter of @xmath137 and @xmath153 is a constant . as nonnegative functions @xmath154 , we may consider a _
shepard weight _ ,
i.e. , @xmath155 @xmath156 being compactly supported functions with support on @xmath137 , such as the wendland functions @xcite . before computing the global fit by mean of local stable approximants
obtained with the lanczos procedure , we briefly sketch in the sequel some relevant properties . since point of property [ approx discrete properties ] implies @xmath157 , we can recover the pu interpolant , by considering @xmath158 , i.e. : @xmath159 if the functions @xmath160 , @xmath161 , satisfy the interpolation conditions @xmath162 for each @xmath163 , then the global pu approximant inherits the interpolation property of the local interpolants , i.e. @xmath164 in order to be able to formulate error bounds , we consider the _
fill distance _ @xmath165 and some further assumptions on the regularity of @xmath137 .
specifically , we require that an open and bounded covering @xmath166 is _ regular _ for @xmath167 .
this means to fulfill the following properties
@xcite : 1 . for each @xmath168 ,
the number of subdomains @xmath169 with @xmath170 is bounded by a global constant @xmath171 ; 2 .
there exists a constant @xmath172 and an angle @xmath173 such that every subdomain @xmath169 satisfies an interior cone condition with angle @xmath174 and radius @xmath175 ; 3 .
the local fill distances @xmath176 are uniformly bounded by the global fill distance @xmath177 .
the first property ensures that is actually a sum over at most @xmath171 summands .
moreover , it is crucial for an efficient evaluation of the global approximant that only a constant number of local interpolants has to be evaluated .
it follows that it should be possible to locate those @xmath171 indices in constant time .
the second and third properties are significant for estimating errors of rbf interpolants . after defining the space @xmath178 of all functions @xmath179
whose derivatives of order @xmath180 satisfy @xmath181 for @xmath182 , we consider the following convergence result ( see theorem @xmath183 in @xcite or also theorem @xmath184 in @xcite ) .
it is presented for strictly positive definite functions even if it holds more in general for strictly conditionally positive definite functions .
let @xmath185 be open and bounded and suppose that @xmath186 .
let @xmath187 be a strictly positive definite function .
let @xmath166 be a regular covering for @xmath188 and let @xmath189 be @xmath190-stable for @xmath191 .
then the error between @xmath192 , where @xmath193 is the native space of @xmath194 , and its pu interpolant can be bounded by : @xmath195 for all @xmath196 and all @xmath197 . for each subdomain , in order to generate the local stable approximation matrix , the lanczos method
is applied to the matrix @xmath144 and to the function values @xmath198 associated with the subdomain @xmath137 . in this way
the matrix @xmath199 and the lanczos basis @xmath200 are computed for each subdomain . then , for each interpolation problem a local stable basis is formed . by using a different stopping criterion in the lanczos algorithm , with respect to the one employed in @xcite , we can compute stable bases for a wider family of rbfs , both globally defined and compactly supported .
the main problem in the lanczos procedure concerns the stopping criterion , ( see step 3 of the lanczos algorithm ) . from property [ discrete properties ] ( point ) and property [ approx discrete properties ] ( point ) a reliable one is : @xmath201 for a certain fixed tolerance @xmath202 , which is supposed to be equal for all the subdomains .
we point out that , even if @xmath202 is fixed among the subdomains , the left - hand side of depends on the specific patch .
moreover , supposing to have a quasi - uniform node distribution , there are no restrictions in selecting the same tolerance for all the subdomains . on the opposite , if the points are more clustered in several subdomains , one can always keep a fixed tolerance , but techniques enabling us to select suitable sizes of the different subdomains are recommended ( see e.g. @xcite ) . from property
[ approx discrete properties ] ( point ) , the fact that we impose as maximum number of iterations , in the lanczos algorit- + hm at step 2 , exactly the number of nodes in @xmath137 , i.e. @xmath203 , naturally follows .
p12cm*1c 0.01 cm inputs : @xmath203 , number of data in @xmath137 ; @xmath204 , the local interpolation matrix ; 0.08 cm 2.2 cm @xmath205 , the function values associated to @xmath137 ; 0.08 cm 2.2 cm @xmath202 , the tolerance used as stopping criterion ; @xmath206 , the radial basis 0.08 cm 2.2 cm function .
0.08 cm outputs : @xmath207 , the new basis in @xmath137 ; @xmath199 , the tridiagonal matrix .
0.12 cm step 1 : set @xmath208 ; @xmath209 ; @xmath210 .
0.12 cm step 2 : * for * @xmath211 0.08 cm 2.2 cm @xmath212 0.08 cm 2.2 cm @xmath213 0.08 cm 2.2 cm @xmath214 0.08 cm 2.2 cm @xmath215 0.12 cm 2.2 cm step 3 : * if * @xmath216 or @xmath217 0.08 cm 4.2 cm * break * 0.12 cm 2.2 cm @xmath218 + then , once the matrix @xmath199 is found for @xmath137 the stable basis is computed by calculating the singular value decomposition of @xmath199 and a local approximant on each subdomain in the form is computed . then the local fits are accumulated into the global one .
the use of stable bases by decomposing the initial problem into many small ones leads to a larger benefit in terms of accuracy than employing a global approach .
in fact if one uses a global method the approximant results stable , but a large number of terms in the lanczos procedure are neglected .
this surely leads to a decrease of the fit accuracy . whereas the local method turns out to be really accurate since , dealing with small problems , less terms in the computation of the basis
are eliminated to preserve stability .
extensive numerical experiments support our findings .
the key step of the pu method consists in organizing the scattered data among the subdomains . to this aim
the kd - tree partitioning structures are widely used , @xcite .
however , they are not specifically implemented for the pu method . here a novel partitioning procedure ,
specifically the so - called _ block - based partitioning structure _ , built for bivariate and trivariate interpolation in order to determine the points belonging to the different pu subdomains , is considered . even if such partitioning structure is robust enough to work on 2d or 3d irregular domains , @xcite , we present such efficient technique for a scattered data set lying in the unit square , i.e. @xmath219 ^ 2 $ ] . at first ,
a partition of unity structure , composed by @xmath220 circular patches @xmath137 of radius : @xmath221 and whose centres @xmath222 , are a grid of points on @xmath2 , is generated .
as in @xcite , the number of pu subdomains is chosen so that @xmath223 .
this choice and lead to a reliable partition of unity structure since , in this way , patches form a covering of the domain @xmath2 . in order to find the points belonging to the different subdomains and
consequently solve , with the use of stable bases , @xmath220 small interpolation problems , we propose a new partitioning structure .
it leads to a natural searching procedure that turns out to be really cheap in terms of computational complexity . to this aim
we first cover @xmath2 with @xmath224 square blocks , where the number @xmath225 of blocks along one side of the unit square is : @xmath226 in this way the width of blocks is equal to the subdomain radius .
this choice can appear trivial , but on the contrary it enables us to consider in the searching process an optimized number of blocks .
blocks are numbered from @xmath227 to @xmath224 ( bottom to top , left to right ) .
thus , with a repeated use of a quicksort routine the set @xmath228 is partitioned by the block - based partitioning structure into @xmath224 subsets @xmath229 , @xmath230 , where @xmath229 are the points stored in the @xmath190-th _ neighbourhood _ , i.e. in the @xmath190-th block and in its eight neighbouring blocks . in such framework , we are able to get an optimal procedure to find the nearest points .
in fact , given a subdomain @xmath137 , whose centre belongs to the @xmath190-th block , we search for all data lying in the @xmath145-th subdomain only among those lying in the @xmath190-th neighbourhood . the same partitioning structure , in case of compactly supported rbfs ( csrbfs ) , must be considered locally for each subdomain .
in fact , in order to build the @xmath145-th stable approximation matrix , among all points lying in the @xmath145-th subdomain , only those belonging to the support of the csrbf must be considered . among several routines which can be employed to determine the neighboring points , we choose the block - based data structure
anyway , we stress that the algorithm , here proposed , works in any dimension @xmath17 , while the block - based data structure is only implemented for @xmath231 , @xcite . thus in higher dimensions such structure must be replaced by standard routines , such as kd - trees , @xcite . for easiness of the reader ,
the procedure is here described in the unit square , however , following @xcite , any extension to irregular domains is possible .
since the stable wsvd - pu algorithm is characterized by the construction of local rbf stable approximants , we consider the local data sets , composed by @xmath203 points , @xmath232 .
thus , the complexity of this algorithm is influenced by the following computational issues : 1 .
organize by means of a partitioning structure the nodes among the subdomains , 2 .
compute the stable basis on each subdomain concerning the efficient organization of points , an extensive complexity analysis , briefly shacked in subsection [ comp_ps ] , can be found in @xcite .
the cost associated to the computation of a local stable basis is investigated in subsection [ compl_lan ] . performing the lanczos procedure on a matrix @xmath233
requires @xmath234 , where @xmath190 is the number of vectors computed by the algorithm , i.e. @xmath190 is the _ good _ low rank approximation , ( a priori unknown in our case ) , @xcite .
given @xmath144 the interpolation matrix defined on the @xmath137 , the lanczos method forms the matrix @xmath199 for @xmath137 after @xmath235 iterations .
usually we have @xmath236 , but in some cases the maximum number of iterations @xmath203 can be reached and so , in a more general setting , @xmath237 .
this routine requires : @xmath238 time complexity .
thus for each subdomain the upper bound for the computational time of the lanczos procedure is given by the right - hand side of . in case of sparse matrices , such as the ones arising from the use of csrbfs
, the lanczos procedure can be performed in : @xmath239 time complexity , where @xmath240 is the number of non - zero entries .
then a singular value decomposition is applied to the matrix @xmath199 .
we remark that performing a singular value decomposition on a matrix @xmath241 requires @xmath242 time complexity .
the singular value decomposition for each subdomain is applied to the matrix @xmath199 ; once more we stress that @xmath243 .
thus for each subdomain the singular value decomposition can be performed in : @xmath244 time complexity .
let us now focus on the block - based partitioning structure used to organize the @xmath5 data sites in blocks .
we remark that such efficient organization of points is specifically implemented for 2d data sets .
anyway , the proposed wsvd - pu algorithm is robust enough to work in any dimension @xmath17 , provided that a different partitioning structure is performed .
let @xmath245 be the number of data sites belonging to a strip .
the procedure used to store the points among the different subdomains is based on recursive calls to a _ quicksort _ routine which requires @xmath246 , where @xmath247 is the number of elements to be sorted .
thus , letting @xmath248 the average number of points lying in a strip , the computational cost needed to organize the @xmath5 points among the different subdomains is : @xmath249 concerning the searching procedure , for each subdomain a quicksort procedure is used to order distances .
thus observing that the data sites in a neighbourhood are about @xmath250 and taking into account the definitions of @xmath225 and @xmath251 , the complexity can be estimated by : @xmath252 the estimate follows from the fact that we built a partitioning structure strictly related to the size of the subdomains and ad hoc for the pu method .
the same computational cost , in case of csrbfs , must be considered locally for each subdomain , to build the sparse interpolation and evaluation matrices . in such steps
we usually have a relatively small number of nodes @xmath203 , with @xmath253 , where the index @xmath145 identifies the @xmath145-th subdomain .
this section is devoted to point out , by means of extensive numerical simulations , stability and accuracy of the wsvd - pu interpolant . to this
aim comparisons with the standard pu interpolant will be carried out .
experiments are performed considering @xmath254 @xmath255 , uniformly random halton nodes , a grid of @xmath256 subdomain centres and a grid of @xmath257 evaluation points , which are contained in the unit square @xmath258 \times [ 0 , 1]$ ] . in order to show the high stability of the proposed method , we compute the root mean square error ( rmse ) , i.e. @xmath259 for different values of the shape parameter in the range @xmath260 $ ] . moreover , in order to point out the versatility of the proposed method , different kernels with different order of smoothness are considered , see table [ tab_1 ] .
the error is computed using as test function the well - known franke s function : @xmath261+\frac{3}{4}\exp\left[-\frac{(9x_1 + 1)^2}{49}-\frac{9x_2 + 1}{10}\right]\\ & + \frac{1}{2 } \exp\left[-\frac{(9x_1 - 7)^2+(9x_2 - 3)^2}{4}\right]-\frac{1}{5 } \exp\left[-(9x_1 - 4)^2-(9x_2 - 7)^2\right].\end{aligned}\ ] ] in figure [ fig_1 ] we compare the rmses obtained by means of the wsvd - pu interpolant ( solid line ) with the ones obtained performing the classical pu method ( dashed line ) . as tolerance value in we set @xmath262 .
these graphs point out that the use of the wsvd - pu local approach reveals a larger stability than the standard pu interpolant .
moreover , the use of a local method enables us to improve the rmse for the optimal shape parameter in case of flat kernels , see figure [ fig_1 ] and table [ tab_1 ] .
this is consistent with the fact that in a local stable method , differently from @xcite , we have to solve small linear systems and therefore few terms are neglected in .
furthermore , from figure [ fig_1 ] we can note that the wsvd - pu method turns out to be more effective with flat kernels , while for more picked bases the improvement of using stable bases becomes negligible as the order of bases function decreases .
thus , from our numerical experiments , we can observe three kinds of behavior depending on different rbf regularity classes .
specifically , the features of such classes , which differ both in terms of stability and accuracy from the standard basis , can be summarized as : * for @xmath21 kernels : improvement of stability and of the optimal accuracy ; * for @xmath263 kernels , with @xmath264 : improvement of stability and same optimal accuracy ; * for @xmath265 kernels : same stability and same optimal accuracy . for @xmath266 , @xmath24 and @xmath26 kernels .
the classical pu interpolant is plotted with dashed line and the wsvd - pu approximant with solid line . from left to right , top to bottom we consider the ga , imq , m6 , w6 , m4 and w4 kernels , respectively.,title="fig:",width=332 ] -1.6 cm for @xmath266 , @xmath24 and @xmath26 kernels .
the classical pu interpolant is plotted with dashed line and the wsvd - pu approximant with solid line . from left to right , top to bottom we consider the ga , imq , m6 , w6 , m4 and w4 kernels , respectively.,title="fig:",width=332 ] + for @xmath266 , @xmath24 and @xmath26 kernels .
the classical pu interpolant is plotted with dashed line and the wsvd - pu approximant with solid line . from left to right , top to bottom we consider the ga , imq , m6 , w6 , m4 and w4 kernels , respectively.,title="fig:",width=332 ] -1.6 cm for @xmath266 , @xmath24 and @xmath26 kernels .
the classical pu interpolant is plotted with dashed line and the wsvd - pu approximant with solid line . from left to right , top to bottom we consider the ga , imq , m6 , w6 , m4 and w4 kernels , respectively.,title="fig:",width=332 ] + for @xmath266 , @xmath24 and @xmath26 kernels .
the classical pu interpolant is plotted with dashed line and the wsvd - pu approximant with solid line . from left to right , top to bottom we consider the ga , imq , m6 , w6 , m4 and w4 kernels , respectively.,title="fig:",width=332 ] -1.6 cm for @xmath266 , @xmath24 and @xmath26 kernels .
the classical pu interpolant is plotted with dashed line and the wsvd - pu approximant with solid line . from left to right , top to bottom we consider the ga , imq , m6 , w6 , m4 and w4 kernels , respectively.,title="fig:",width=332 ] ' '' '' @xmath5 & method & rmse & @xmath267 & rmse & @xmath267 & rmse & @xmath267 & rmse & @xmath267 + ' '' '' @xmath268 & pu & @xmath269 & @xmath270 & @xmath271 & @xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath276 + & wsvd - pu & @xmath277 & @xmath270 & @xmath278 & @xmath279 & @xmath273 & @xmath274 & @xmath275 & @xmath276 + ' '' '' @xmath280 & pu & @xmath281 & @xmath282 & @xmath283 & @xmath272 & @xmath284 & @xmath285 & @xmath286 & @xmath287 + & wsvd - pu & @xmath288 & @xmath270 & @xmath289 & @xmath279 & @xmath290 & @xmath285 & @xmath291 & @xmath287 + ' '' '' @xmath292 & pu & @xmath293 & @xmath285 & @xmath294 & @xmath295 & @xmath296 & @xmath297 & @xmath298 & @xmath299 + & wsvd - pu & @xmath300 & @xmath270 & @xmath301 & @xmath272 & @xmath302 & @xmath274 & @xmath303 & @xmath276 + moreover , since we are interested in pointing out the efficiency of the proposed wsvd - pu algorithm , in table [ tab_3 ] we also report the cpu times obtained by using our stable interpolation method with the gaussian rbf as local approximant , for each of the three different data sets .
tests have been carried out on a intel(r ) core(tm ) i3 cpu m330 2.13 ghz processor . ' '' ''
@xmath5 & @xmath304 & @xmath305 & @xmath292 + ' '' '' @xmath306 cpu [ s ] & @xmath307 & @xmath308 & @xmath309 +
in this section we focus on an application to earth s topography , which consists in approximating with our algorithm a set of real scattered data . in particular
, we consider the so - called _ glacier _ data set .
it is composed by @xmath310 points representing digitized height contours of a glacier , @xcite .
the difference between the highest and the lowest point is @xmath311 m. such points , differently from the halton data , are not quasi - uniform .
furthermore , they are distributed on an irregular domain @xmath312 ^ 2 $ ] .
a 2d view of such points is plotted in figure [ fig_2 ] ( left ) . among them
we select @xmath313 points for the cross - validation ( plotted in red in the left frame of figure [ fig_2 ] ) .
since the nodes of the glacier data set do not cover the whole unit square , after generating an initial grid of subdomain centres in @xmath314 ^ 2 $ ] , we reduce such points taking only those lying in @xmath2 by means of the technique described in @xcite . to obtain reliable and numerically significant results of the error , in this application
it is more appropriate to use the relative rmse ( rrmse ) : @xmath315 in figure [ fig_2 ] , we show how the rrmses vary with respect to the shape parameter @xmath316 $ ] . in doing so
, we consider the following kernels : ga , w6 and m4 .
as already shown , the results point out once more that the proposed approach is stable and moreover turns to be effective also in applications .
the errors for the optimal shape parameter @xmath12 are shown in table [ tab_2 ] .
since we refer to points with highly varying densities and thus truly ill - conditioned matrices , the classical pu method does not give acceptable approximations .
consequently , we do not report the errors obtained with this standard algorithm . with ga , w6 and m4 kernels for the wsvd - pu approximant.,title="fig:",width=332 ] -1.6 cm with ga , w6 and m4 kernels for the wsvd - pu approximant.,title="fig:",width=332 ] ' '' '' rrmse & @xmath267 & rrmse & @xmath267 & rrmse & @xmath267 + ' '' '' @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath320 +
the first , third and fourth authors are partially supported by the university of torino through research project metodi numerici nelle scienze applicate . the second and fifth authors are partially supported by the funds of the university of padova , project cpda124755 multivariate approximation with application to image reconstruction .
w. pogorzelski , integral equations and their applications .
i , translated from the polish by jacques j. schorr - con , a. kacner and z. olesiak .
international series of monographs in pure and applied mathematics , vol .
pergamon press , oxford , 1966 .
a. safdari - vaighani , a. heryudono , e. larsson , a radial basis function partition of unity collocation method for convection - diffusion equations arising in financial applications , j. sci .
* 64 * ( 2015 ) , 341367 .
h. wendland , fast evaluation of radial basis functions : methods based on partition of unity , in approximation theory x : wavelets , splines , and applications , c. k. chui , l. l. schumaker , j. stckler ( eds . ) , vanderbilt univ . press ,
nashville , tn , 2002 , pp . | in this paper we propose a new stable and accurate approximation technique which is extremely effective for interpolating large scattered data sets .
the partition of unity ( pu ) method is performed considering radial basis functions ( rbfs ) as local approximants and using locally supported weights . in particular , the approach consists in computing , for each pu subdomain , a stable basis . such technique ,
taking advantage of the local scheme , leads to a significant benefit in terms of stability , especially for flat kernels .
furthermore , an optimized searching procedure is applied to build the local stable bases , thus rendering the method more efficient .
meshfree approximation , radial basis functions , partition of unity , scattered data interpolation , numerical stability , krylov space methods .
65d05,65d15,65y20 . |
accreting t tauri stars are observed to be less luminous in x - rays than non - accretors ( ( * ? ? ?
* stelzer & neuhuser 2001 ) ; ( * ? ? ?
* flaccomio , micela & sciortino 2003a ) ; ( * ? ? ?
* flaccomio , damiani , micela , 2003c ) ; ( * ? ? ?
* stassun , ardila , barsony , 2004 ) ; ( * ? ? ?
* preibisch , kim , favata , 2005 ) ; ( * ? ? ?
* flaccomio , micela & sciortino 2006 ) ; ( * ? ? ?
* telleschi , gdel , briggs , 2007a ) ) .
accreting stars appear to be a factor of @xmath0 less luminous , and show a larger variation in their x - ray activity compared to non - accreting stars ( ( * ? ? ?
* preibisch 2005 ) ) . however , it is only in recent years that this result has become clear , with previous studies showing conflicting results ( e.g. @xcite and @xcite ) .
the apparent discrepancy arose from whether stars were classified as accreting based on the detection of excess ir emission ( a disk indicator ) or the detection of accretion related emission lines .
however , with careful re - analysis of archival data ( ( * ? ? ?
* flaccomio 2003a ) ) and recent large x - ray surveys like the _ chandra _ orion ultradeep project ( coup ; ( * ? ? ?
* getman , flaccomio , broos , 2005 ) ) and the xmm-_newton _ extended survey of the taurus molecular cloud ( xest ; ( * ? ? ?
* gdel , briggs , arzner , 2007a ) ) the result is now clear , namely that accreting t tauri stars are observed to be , on average , less luminous in x - rays than non - accreting stars .
although the difference is small it has been found consistently in various star forming regions : taurus - auriga ( ( * ? ? ?
* stelzer & neuhuser 2001 ) ; ( * ? ? ?
* telleschi 2007a ) ) , the orion nebula cluster ( ( * ? ? ?
* flaccomio 2003c ) ; ( * ? ? ?
* stassun 2004 ) ; ( * ? ? ?
* preibisch 2005 ) ) , ngc 2264 ( ( * ? ? ?
* ; * ? ? ?
* flaccomio 2003a , 2006 ) ) and chamaeleon i ( ( * ? ? ?
* flaccomio 2003a ) )
. it should be noted , however , that such observations from ccd detectors are not very sensitive to x - rays that are produced in accretion shocks .
high resolution x - ray spectroscopic measurements have indicated emission from cool and high density plasma , most likely associated with accretion hot spots , in several ( but not all ) accreting stars ( e.g. ( * ? ? ?
* telleschi , gdel , briggs , 2007b ) ; ( * ? ? ?
* gnther , schmitt , robrade , 2007 ) ) . in this work
we only consider coronal x - ray emission such as is detected by ccd measurements .
it is not yet understood why accreting stars are under luminous in x - rays , although a few ideas have been put forward .
it may be related to higher extinction due to x - ray absorption by circumstellar disks , however the coup results do not support this suggestion ( ( * ? ? ?
* preibisch 2005 ) ) .
it may be related to magnetic braking , whereby the interaction between the magnetic field of an accreting star with its disk slows the stellar rotation rate leading to a weaker dynamo action and therefore less x - ray emission ; although the lack of any rotation - activity relation for t tauri stars has ruled out this idea ( ( * ? ? ?
* flaccomio 2003c ) ; ( * ? ? ?
* preibisch 2005 ) ; ( * ? ? ?
* briggs , gdel , telleschi , 2007 ) ) .
a third suggestion is that accretion may alter the stellar structure affecting the magnetic field generation process and therefore x - ray emission ( ( * ? ? ? * preibisch 2005 ) ) .
however , the most plausible suggestion is the attenuation of coronal x - rays by the dense gas in accretion columns ( ( * ? ? ?
* flaccomio 2003c ) ; ( * ? ? ?
* stassun 2004 ) ; ( * ? ? ? * preibisch 2005 ) ; ( * ? ? ?
* gdel , telleschi , audard , 2007b ) ) .
x - rays from the underlying corona may not be able to heat the material within accretion columns to a high enough temperature to emit in x - rays .
field lines which have been mass - loaded with dense disk material may obscure the line - of - sight to the star at some rotation phases , reducing the observed x - ray emission . here
we demonstrate this in a quantitative way by developing an accretion flow model and simulating the propagation of coronal x - rays through the stellar magnetosphere .
in order to model the coronae of t tauri stars we need to assume something about the form of the magnetic field .
observations suggest it is compact and inhomogeneous and may vary not only with time on each star , but also from one star to the next . to capture this behaviour
, we use as examples the field structures of two different main sequence stars , lq hya and ab dor determined from zeeman - doppler imaging ( ( * ? ? ?
* donati , cameron , semel , 2003 ) ) .
although we can not be certain whether or not the magnetic field structures extrapolated from surface magnetograms of young main sequence stars do represent the magnetically confined coronae of t tauri stars , they do satisfy the currently available observational constraints . in future
it will be possible to use real t tauri magnetograms derived from zeeman - doppler images obtained using the espadons instrument at the canada - france - hawaii telescope ( ( * ? ? ?
* donati , jardine , gregory , 2007 ) ) . however , in the meantime , the example field geometries used in this work ( see fig .
[ coronae ] ) capture the essential features of t tauri coronae .
they reproduce x - ray emission measures ( ems ) and coronal densities which are typical of t tauri stars ( ( * ? ? ?
* jardine , cameron , donati , 2006 ) ) .
the surface field structures are complex , consistent with polarisation measurements ( ( * ? ? ?
* valenti & johns - krull 2004 ) ) and x - ray emitting plasma is confined within unevenly distributed magnetic structures close to the stellar surface , giving rise to significant rotational modulation of x - ray emission ( ( * ? ? ?
* gregory , jardine , cameron , 2006b ) ) . for a given surface magnetogram
we calculate the extent of the closed corona for a specified set of stellar parameters .
we extrapolate from surface magnetograms by assuming that the magnetic field @xmath1 is potential such that @xmath2 .
this process is described in detail by ( * ? ? ? * jardine ( 2006 ) ) , ( * ? ? ?
* gregory , jardine , simpson , ( 2006a ) ) and ( * ? ? ?
* gregory ( 2006b ) ) .
we assume that the corona is isothermal and that plasma along field line loops is in hydrostatic equilibrium .
the pressure is calculated along the path of field line loops and is set to zero for open field lines and for field lines where , at some point along the loop , the gas pressure exceeds the magnetic pressure .
the pressure along a field line scales with the pressure at its foot point , and we assume that this scales with the magnetic pressure .
this technique has been used successfully to calculate mean coronal densities and x - ray ems for the sun and other main sequence stars ( ( * ? ? ?
* jardine , wood , cameron , 2002 ) ) as well as t tauri stars ( ( * ? ? ?
* jardine 2006 ) ) .
the ab dor - like coronal field has an x - ray em where @xmath3 and @xmath4 are the coronal density and volume .
the em - weighted density is @xmath5 .
] of @xmath6 ( without considering accretion ) and a mean em - weighted coronal density of @xmath7 , consistent with estimates from the modelling of individual flares ( ( * ? ?
* favata , flaccomio , reale , 2005 ) ) .
the lq hya - like field has a more extended corona and consequently a lower coronal density and em , @xmath8 .
we assume that the structure of the magnetic field remains undistorted by the in - falling material and that the magnetosphere rotates as a solid body .
the accreting field geometries shown in fig .
[ coronae ] are therefore only snap - shots in time , and in reality will evolve due to the interaction with the disk .
the question of where the disk is truncated remains a major problem for accretion models .
it is still unknown if the disk is truncated in the vicinity of the corotation radius , the assumption of traditional accretion models ( e.g. ( * ? ? ? * knigl 1991 ) ) , or whether it extends closer to the stellar surface ( e.g. ( * ? ? ?
* matt & pudritz 2005 ) ) . in this work
we assume that accretion occurs over a range of radii within the corotation radius .
this is equivalent to the approach taken previously by e.g. @xcite who have demonstrated that such an assumption reproduces observed spectral line profiles and variability .
the accretion filling factors are of order @xmath9 , consistent with observationally inferred values ( e.g. ( * ? ? ?
* valenti & johns - krull 2004 ) ) .
we assume that material is supplied by the disk and accretes onto the star at a constant rate . for a dipolar magnetic field accretion flows impact the stellar surface in two rings in opposite hemispheres centred on the poles . in this case ,
half of the mass supplied by the disk accretes into each hemisphere . for complex magnetic fields accretion occurs into discrete hot spots distributed in latitude and longitude ( ( * ? ? ?
* gregory 2006a ) ) .
it is therefore not clear how much of the available mass from the disk accretes into each hot spot .
we use a spherical grid and assume that each grid cell within the disk which is accreting supplies a mass accretion rate that is proportional to its surface area . if an accreting grid cell has a surface area that is @xmath10 of the total area of all accreting grid cells , then this grid cell is assumed to carry @xmath10 of the total mass that is supplied by the disk . therefore ,
as an example , if grid cells which constitute half of the total area of all accreting cells in the disk carry material into a single hot spot , then half of the mass accretion rate is carried from the disk to this hot spot . in this way
the accretion rate into each hot spot is different and depends on the structure of the magnetic field connecting the star to the disk .
we consider a star of mass @xmath11 , radius @xmath12 , rotation period @xmath13 , a coronal temperature of @xmath14 and assume that the disk supplies a mass accretion rate of @xmath15 . in order to model the propagation of coronal x - rays through the magnetosphere we first need to determine the density of gas within accretion columns .
* gregory , wood & jardine ( 2007 ) ) develop a steady state accretion flow model where material accretes from a range of radii within corotation , free - falling along the field lines under gravity .
the resulting density profiles do not depend on the absolute field strength , but instead on how the field strength varies with height above the star .
the density profiles are typically steeper than those derived for accretion flows along dipolar field lines since the strength of a higher order field drops faster with height above the star .
[ flows ] shows the variation of the number density along the paths of a selection of accreting field lines , with those obtained for dipolar field lines shown for comparison ( ( * ? ? ?
* gregory 2007 ) ) . for our assumed accretion rate of @xmath15
the flow densities range from @xmath16 , whilst for a lower accretion rate of @xmath17 the range is @xmath18 .
we model the propagation of coronal x - rays through the magnetosphere by considering absorption by the dense gas within accretion columns . for the radiation transfer we use monte carlo techniques and discretise the emissivity and density onto a spherical polar grid ( e.g. ( * ? ? ?
* jardine 2002 ) ; ( * ? ? ?
* whitney , wood , bjorkman , 2003 ) ) .
the stellar inclination is set to @xmath19 and we assume the x - ray emission from the @xmath14 corona is optically thin , but that the x - rays may be subsequently absorbed in the cool and hence optically thick accretion columns . for the x - ray absorptive opacity we adopt a value of @xmath20 , typical of neutral gas at temperatures below @xmath21 at x - ray energies of a few kev ( e.g. ( * ? ? ?
* krolik & kallman 1984 ) ) . at these energies
the opacity of hot gas ( above @xmath22 ) is several orders of magnitude lower ( e.g. ( * ? ? ?
* krolik & kallman 1984 ) , their fig .
1 ) justifying our assumption that the coronal x - ray emission is optically thin . in the monte carlo x - ray radiation transfer simulations we assume the scattering opacity is negligible , so our results in fig .
[ coronae ] show the effects of attenuation of the coronal emission by the accretion columns .
the second row in fig .
[ coronae ] shows the x - ray images in the absence of attenuation ( i.e. x - ray opacity in the accretion columns is set to zero ) whilst the fourth row shows the same x - ray emission models , but with our adopted value for the soft x - ray opacity in the accretion columns .
the observed x - ray em is reduced by a factor of 1.4 ( 2.0 ) for the ab dor - like ( lq hya - like ) field when accretion flows are considered , where the reduction factor is the average for an entire rotation cycle . for the ab dor - like field
there are large accretion curtains which cross the observers line - of - sight to the star as it rotates ( see fig .
[ coronae ] ) . for the lq hya - like field accretion is predominantly along field lines which carry material into low latitude hot spots , however , one of the brightest x - ray emitting regions is obscured by an accretion column which attenuates the coronal x - rays and produces a large reduction in the observed x - ray emission .
this immediately suggests that the geometry of the accreting field is a contributory factor in causing the large scatter seen in the x - ray luminosities of accreting stars .
we have demonstrated that the suppression of x - ray emission in accreting stars apparent from ccd observations can , at least in part , be explained by the attenuation of coronal x - rays by the dense material in accretion columns .
this suggests that both accreting and non - accreting stars have the same intrinsic x - ray luminosity , with accreting t tauri stars being observed to be less luminous due to the effects of absorbing gas in accretion columns .
the reduction in the observed x - ray emission depends on the structure of the accreting field . for stars where accretion columns rotate across the line - of - sight , x
- rays from the underlying corona are strongly absorbed by the accreting gas which reduces the observed x - ray emission .
a preliminary calculation indicates that the column densities from our simulations are large enough that the softer ( cooler ) coronal spectral components may be substantially , if not completely , absorbed by the accreting gas .
the effect is greater the larger the accretion rate . indeed
* gdel ( 2007b ) ) have recently found that stars in xest which have the largest accretion rates ( and are driving jets ) show a highly absorbed coronal spectral component , which is attributed to attenuation by accreting gas .
this however does not rule out the fact that other mechanisms may also be responsible for reducing the x - ray emission in accreting stars .
* jardine ( 2006 ) ) have demonstrated that some stars ( typically those of lower mass ) have their outer coronae stripped away via the interaction with a disk .
this also reduces the observed x - ray emission and this effect , combined with the radiative transfer calculations presented here , is likely to lead to a larger reduction in the observed x - ray emission .
this would reduce the number of field lines which could be filled with coronal gas , such as is also suggested by ( * ? ? ?
* preibisch ( 2005 ) ) and ( * ? ? ?
* telleschi ( 2007a ) ) , with the observed x - ray emission being further reduced due to obscuration by the accreting gas . | accreting t tauri stars are observed to be less luminous in x - rays than non - accretors , an effect that has been detected in various star forming regions . to explain this we have combined , for the first time , a radiative transfer code with an accretion model that considers magnetic fields extrapolated from surface magnetograms obtained from zeeman - doppler imaging .
such fields consist of compact magnetic regions close to the stellar surface , with extended field lines interacting with the disk .
we study the propagation of coronal x - rays through the magnetosphere and demonstrate that they are strongly absorbed by the dense gas in accretion columns . |
diffraction calculations such as huygens diffraction , fresnel diffraction and angular spectrum method , are important tools in wide - ranging optics @xcite , ultrasonic @xcite , x - ray @xcite and so forth .
its applications in optics include computer - generated - hologram ( cgh ) and digital holography @xcite , phase retrieval , image encryption and decryption and so forth .
fast fourier transform ( fft)-based diffraction calculations according to their convolution or fourier transform form are used in these applications .
the fft - based diffraction calculations , however , can only be applied to planar surfaces in parallel . in order to apply the methods to a non - parallel planar surface ,
many methods have been proposed : for example , non - parallel fresnel diffractions @xcite and non - parallel angular spectrum methods @xcite .
unfortunately , these non - parallel diffractions are limited to planar surfaces .
if we calculate a source surface with arbitrary shape using non - parallel diffractions , we need to approximate the arbitrary shape surface with many small non - parallel planar surfaces , and then , we need to calculate the non - parallel diffraction per the small non - parallel planar surfaces . in this paper , we propose fresnel diffraction calculation on an arbitrary shape surface . without approximating an arbitrary shape surface with small non - parallel planar surfaces ,
this method is capable of calculating fresnel diffraction from a source arbitrary shape surface to a planar destination surface .
although such calculation can be readily calculated by the direct integral of a diffraction calculation , the calculation cost is proportional to @xmath0 in one dimensional or @xmath1 in two dimensional cases , where @xmath2 is the number of sampling points .
however , the calculation cost of the proposed method is @xmath3 in one dimensional or @xmath4 in two dimensional cases using non - uniform fast fourier transform . in section 2 ,
we describe the arbitrary shape surface fresnel diffraction . in section 3
, we present the numerical results .
section 4 concludes this work .
let us begin with huygens diffraction .
huygens diffraction @xcite on a planar surface is expressed as : @xmath5 where , @xmath6 and @xmath7 are planar source and destination surfaces , @xmath8 and @xmath9 are the position vectors on the source and destination surfaces , @xmath10 and @xmath11 are the wavelength and wave number of light , and @xmath12 , where @xmath13 is the distance between the source and destination surfaces .
@xmath14 is the incident wave to the source surface .
if the incident wave is used as a planar wave that is perpendicular to the optical axis , we can treat as @xmath15 .
we expand the huygens diffraction to a source surface with arbitrary shape .
as shown in fig .
[ fig : system ] , the source surface with arbitrary shape @xmath16 is defined by the displacement @xmath17 at the position @xmath8 . note that when the arbitrary shape surface is uniform - sampled , the corresponding coordinate @xmath8 is non - uniform - sampled depending on the slope of @xmath16 at the position @xmath8 .
huygens diffraction on an arbitrary shape surface is expressed by : @xmath18 @xmath19 here , if the incident wave @xmath20 is used as a planar wave that is perpendicular to the optical axis , the incident wave is expressed as @xmath21 . applying fresnel approximation to eq.([eqn : r1 ] ) using @xmath22 , we can obtain the following approximation : @xmath23 and , we also approximate @xmath24 in the integral of eq.([eqn : huy_diff2 ] ) .
therefore , we obtain fresnel diffraction on an arbitray surface : @xmath25 the above eqs.([eqn : huy_diff2 ] ) and ( [ eqn : fre_diff1 ] ) can be treated an arbitrary shape surface .
we can readily calculate by the direct integral with regard to these equations ; however , the calculation cost is @xmath0 in one dimensional or @xmath1 in two dimensional cases , where @xmath2 is the number of sampling points , because we can not calculate them using fourier transform . in order to obtain the fourier form of eq.([eqn : fre_diff1 ] ) , we approximate the third exponential term in the integration as follows : @xmath26 eventually , we obtain the following equation : @xmath27 because the coordinate @xmath8 is sampled by the non - uniform sampling rates , instead of ( uniform ) fourier transform , we can calculate the above equation using non - uniform fourier transform ( nuft ) : @xmath28 } , \end{aligned } \label{eqn : fre_diff_nuf}\ ] ] where , @xmath29}$ ] denotes nuft .
nuft of a function @xmath30 is defined as : @xmath31}=\int \!\ !
\int f({{\bf x}_{1 } } ) \exp(- i \pi { { \bf x}_{1}}{{\bf x}_{2 } } ) d{{\bf x}_{1}}. \end{aligned } \label{eqn : nufourier}\ ] ] although the form of nuft is similar to that of uniform fourier transform , the coordinate @xmath8 is non - uniform - sampled and @xmath9 is uniform - sampled , unlike uniform fourier transform . for the numerical implementation of eq.([eqn : nufourier ] ) , it is necessary to use non - uniform fast fourier transform ( nufft ) which has the complexity of @xmath3 .
many methods for nufft have been proposed over the course of the past twenty years or so @xcite .
nuffts are based on the combination of an interpolation and the uniform fft . in this paper , we used l. greengard and j. y. lee s nufft @xcite . for more details , see @xcite .
, @xmath32 , @xmath33 and @xmath34to @xmath8 , respectively . ( a ) source surface ( b ) diffraction result by eq.([eqn : huy_diff2 ] )
( c ) diffraction result by eq.([eqn : fre_diff1 ] ) ( d ) diffraction result by eq.([eqn : fre_diff_nuf ] ) ( e ) absolute error between ( b ) and ( d).,width=453 ] let us examine our method using two source surfaces with arbitrary shapes in one dimension : a surface composed of four tilted planar surfaces , and quadratic curve .
we used the distance @xmath35 m , the wavelength of 633 nm , and the number of sampling points on source and destination @xmath36 .
the sampling rates on the source and destination surface are @xmath37 m. we used a planar wave as the incident wave that is perpendicular to the optical axis .
figure [ fig : complex - planar ] shows a source surface composed of four small planar surfaces with 128 points , which are tilted @xmath38 , @xmath32 , @xmath33 and @xmath34 to @xmath8 , respectively .
the horizontal axis in ( a ) indicates the position on @xmath8 in metric units .
the horizontal axes in ( b)-(e ) indicates the position on @xmath9 in metric units .
the sampling rates on these small planar surfaces are @xmath39 , however , the sampling rates on @xmath8 are @xmath40 , @xmath41 , @xmath42 and @xmath43 , respectively .
the destination planar surface is not inclined to @xmath9 .
figures [ fig : complex - planar ] ( b)-(d ) show the intensity profiles of the diffraction results by eq.([eqn : huy_diff2 ] ) , eq.([eqn : fre_diff1 ] ) with direct integral and our method ( eq.([eqn : fre_diff_nuf ] ) ) , respectively .
figure [ fig : complex - planar ] ( e ) depicts the absolute error between ( eq.([eqn : huy_diff2 ] ) ) and eq.([eqn : fre_diff_nuf ] ) .
the absolute error falls into within approximately 0.025 . )
( c ) diffraction result by eq.([eqn : fre_diff1 ] ) ( d ) diffraction result by eq.([eqn : fre_diff_nuf ] ) ( e ) absolute error between ( b ) and ( d).,width=453 ] figure [ fig : quad - curve ] shows a source surface with quadratic curve .
the sampling rate on @xmath8 according to the source surface depends on the slope of the quadratic curve .
figure [ fig : quad - curve ] ( b)-(d ) shows the intensity profiles of the diffraction results by eq.([eqn : huy_diff2 ] ) , eq.([eqn : fre_diff1 ] ) and our method ( eq.([eqn : fre_diff_nuf ] ) ) , respectively .
figure [ fig : quad - curve ] ( e ) depicts the absolute error between eq.([eqn : huy_diff2 ] ) and eq.([eqn : fre_diff_nuf ] ) .
the absolute error falls into within approximately 0.005 .
the primary factor of these absolute errors in figs.[fig : complex - planar ] and [ fig : quad - curve ] is the approximations by eqs.([eqn : r2 ] ) and ( [ eqn : fre_diff4 ] ) .
we proposed fresnel diffraction calculation from a source surface with arbitrary shape to a planar destination surface , using one nufft .
we have had only a direct integral for calculating such diffraction thus far .
unfortunately , it is very time consuming because the calculation cost takes @xmath0 in one dimensional or @xmath1 in two dimensional cases .
in contrast , the calculation cost of our method is @xmath3 in one dimensional or @xmath4 in two dimensional cases using nufft .
the method is very useful for calculating a cgh from a three - dimensional object composed of multiple polygons or arbitrary shape surfaces . in our next work
, we will show the fast calculation of a cgh from such 3d objects using this method .
this work is supported by the ministry of internal affairs and communications , strategic information and communications r&d promotion programme ( scope ) , japan society for the promotion of science ( jsps ) kakenhi ( young scientists ( b ) 23700103 ) 2011 , and the nakajima foundation . | fresnel diffraction calculation on an arbitrary shape surface is proposed .
this method is capable of calculating fresnel diffraction from a source surface with an arbitrary shape to a planar destination surface .
although such calculation can be readily calculated by the direct integral of a diffraction calculation , the calculation cost is proportional to @xmath0 in one dimensional or @xmath1 in two dimensional cases , where @xmath2 is the number of sampling points .
however , the calculation cost of the proposed method is @xmath3 in one dimensional or @xmath4 in two dimensional cases using non - uniform fast fourier transform
. 99 j.w.goodman , `` _ introduction to fourier optics ( 3rd ed . ) _ , '' robert & company ( 2005 ) .
okan k. ersoy , `` _ diffraction , fourier optics and imaging _ , '' wiley - interscience ( 2006 ) .
e. g. williams , `` _ fourier acoustics
sound radiation and nearfield acoustical holography _ , '' academic press ( 1999 ) .
d. m. paganin , `` _ coherent x - ray optics _ , '' oxford university press ( 2006 ) .
t. c. poon ( ed . ) , `` _ digital holography and three - dimensional display _
, '' springer 2006 .
d. leseberg and c. frre , `` computer - generated holograms of 3-d objects composed of tilted planar segments , '' appl.opt . ,
* 27 , * 3020 ( 1988 ) . c. frere and d. leseberg , `` large objects reconstructed from computer - generated holograms , '' appl
. opt . * 28 * , 24222425 ( 1989 ) .
l. yu , y. an , and l. cai , `` numerical reconstruction of digital holograms with variable viewing angles , '' opt . express * 10 * , 12501257 ( 2002 ) .
h. sakata and y. sakamoto , `` fast computation method for a fresnel hologram using three - dimensional affine transformations in real space , '' appl . opt . * 48 * , h212h221 ( 2009 ) .
t. tommasi and b. bianco , `` frequency analysis of light diffraction between rotated planes , '' opt .
lett .
* 17 * , 556558 ( 1992 ) .
k. matsushima , h. schimmel , and f. wyrowski , `` fast calculation method for optical diffraction on tilted planes by use of the angular spectrum of plane waves , '' j. opt .
soc .
am .
a * 20 * , 17551762 ( 2003 ) .
g. b. esmer and l. onural , `` computation of holographic patterns between tilted planes , '' proc .
spie * 6252 * , 62521k-1 62521k-6 ( 2006 ) .
l. ahrenberg , p. benzie , m. magnor , and j. watson , `` computer generated holograms from three dimensional meshes using an analytic light transport model , '' appl . opt . * 47 * , 15671574 ( 2008 ) .
a. dutt and v. rokhlin , `` fast fourier transforms for nonequispaced data , '' siam j. sci .
comp .
, * 14 * , 13681393 ( 1993 ) .
q. h. liu and n. nguyen , `` an accurate algorithm for nonuniform fast fourier transforms ( nuffts ) , '' ieee microw . guided wave lett .
, * 8 * , 1820 ( 1998 ) . q. h. liu , n. nguyen , and x. y. tang , `` accurate algorithms for nonuniform fast forward and inverse fourier transforms and their applications , '' proc . ieee geosci .
remote sensing symp .
, * 1 * , 288290 ( 1998 ) .
l. greengard and j. y. lee , `` accelerating the nonuniform fast fourier transform , '' siam rev .
* 46 * , 443454 ( 2004 ) . |
the past decade of precision electroweak experiments has seen outstanding confirmation of the standard model at the per mille level .
but the successes of the standard model have drawn increased attention to its deficiencies , notably its unsatisfactory treatment of the mechanism behind electroweak symmetry breaking .
this phenomenon occurs roughly at the 1 tev scale , which the lhc will access directly .
the merits of any additional machine must therefore be evaluated within the context of the lhc program .
a tev - scale linear collider ( lc ) is a leading candidate for such a facility .
such a machine offers control over the beam energy and polarization , and a clean environment that enables precision event reconstruction . in this paper
i illustrate aspects of the lc physics program with examples drawn from higgs physics , top quark physics , and the study of large extra spacetime dimensions .
more comprehensive reviews of physics at @xmath0 linear colliders are given in refs .
@xcite , @xcite , and @xcite .
well - developed lc designs have been put forward by the slac - kek joint effort ( the nlc / jlc designs)@xcite and by desy ( the tesla design)@xcite .
the two machines differ technologically but achieve similar ends .
the nlc uses warm rf cavities operating in the x - band ( 11.4 ghz ) .
the baseline design assumes initial operation at @xmath1 gev at a luminosity of @xmath2 @xmath3s@xmath4 , with an @xmath5 beam polarization of 80 - 90% .
the linac is designed to allow adiabatic energy upgrades to @xmath6 tev through the addition of klystrons .
the size of the beam - delivery and final focus systems would allow eventual operation at 1.5 tev .
the tesla design uses superconducting rf operating at @xmath7 ghz , and reaches a center of mass energy of 800 - 1000 gev .
initial operation would be in the 200 - 500 gev range , with an @xmath5 polarization of 80% as well as a positron beam polarization of 60% , which would introduce new measureables into physics processes . this design permits operation at very high luminosity , up to @xmath8 @xmath3s@xmath4 .
the two designs have rather different beam characteristics and time structures . a brief comparison of the tesla and nlc parameters is shown in table [ tab : machines ] ; a more complete list can be found in ref .
@xcite . designing a detector for a linear collider
is widely regarded , at least by those who work at hadron colliders , as an easy problem .
certainly the lc does not share the lhc s formidable challenges of high event rates and radiation exposures .
the challenges for a linear collider detector stem from the desire to fully exploit the clean machine environment by building a detector of the highest possible precision .
current designs are evolutionary extensions of the lep and sld detectors and feature ccd pixel vertex detectors , silicon or tpc outer trackers , and a fine - grained em calorimeter located inside the magnet coil .
several designs incorporate the hadronic calorimeter inside the coil as well .
in contrast to the lhc where triggering is a major challenge , at a linear collider the full detector can be read out between bunch trains and triggers formed in software .
.summary of jlc / nlc and tesla accelerator parameters .
luminosities include the pinch enhancement .
[ cols= " < , < , < , < , < , < , < " , ] of still greater interest , however , is a direct measurement of the top - higgs yukawa coupling , @xmath9 .
such a measurement , like those of the higgs branching ratios discussed above , is needed to establish the `` higgsness '' of the higgs , and may also probe the special nature of the top quark . at the lhc , the ratio @xmath10 can be measured to an accuracy of 25% for @xmath11 gev@xcite . for a light higgs ,
@xmath9 can be measured at the lc in @xmath12 production . for higgs masses around 120 gev
, the cross section for this process peaks at about 2.6 fb for @xmath13-@xmath14 gev and then falls off slowly .
this is some three orders of magnitude smaller than the dominant @xmath15 , @xmath16 , and @xmath17 backgrounds .
but the spectacular nature of these events ( @xmath18 if both tops decay hadronically , or @xmath19 + if both tops decay semi - leptonically ) , and their many kinematic constraints , provide enough handles that backgrounds can be acceptably reduced through direct mass reconstruction@xcite or a neural net@xcite . in the latter study ,
the authors assume 1000 fb@xmath4 ( about 3 years of running at @xmath20 ) at @xmath21 gev , and obtain a 5.5% uncertainty on @xmath9 .
outstanding flavor - id is again a prerequisite for this measurement .
the possibility of measuring @xmath9 with such high precision is a strong argument in favor of the highest possible luminosities .
the recent proposal@xcite to resolve the hierarchy problem through a theory of low - scale quantum gravity with large extra spacetime dimensions has generated great interest because of its testable consequences at colliders@xcite . in these models , standard model fields
are confined to the 4-dimensional boundary of a `` bulk '' with @xmath22 compact extra dimensions of characteristic size @xmath23 .
gravitons propagate in the bulk , where they couple with a strength of order the electroweak strength ( hence the elimination of the hierarchy ) .
the apparent weakness of gravity in our 4-dimensional world arises from the geometrical suppression of the gravitational flux lines by a factor proportional to the volume of the compact extra dimensions : @xmath24 where @xmath25 gev is the planck scale , @xmath26 is the volume of the compact extra dimensions , and @xmath27 is the fundamental planck scale in the bulk . if we require @xmath28 to eliminate the hierarchy , we can obtain the characteristic size @xmath23 of the extra dimensions for various values of @xmath22 .
values of @xmath23 as large as a fraction of a millimeter are permitted by current limits from cavendish - type experiments@xcite . for phenomenological purposes , we are most concerned with the effective lagrangian that describes the interactions between gravitons and sm fields in our 4-dimensional world@xcite .
gravitons then appear as a kaluza - klein tower , or series , of closely - spaced massive spin-2 states that can be emitted or exchanged along with sm gauge bosons .
each such state has a very weak coupling to matter , of order @xmath29 , but because of the large number of these states their cumulative effect is comparable to that of standard model processes at energies near @xmath27 . one way to search for the effect of these large extra dimensions at
the lc is through the effect of graviton exchange on fermion pair production@xcite .
this process is extremely well - understood theoretically and is a sensitive probe of many types of new physics , including @xmath30 s , compositeness , and technicolor .
graviton exchange turns out to leave the total cross section and integrated left - right asymmetry unchanged , but modifies the angular distributions in a way that depends on a single parameter , @xmath31 . here
@xmath32 is a dimensionless parameter of order one ( but of either sign ) that depends on model - dependent physics above @xmath27 .
a fit to the angular distribution of @xmath33 , and @xmath34 gives the exclusion reach shown in figure [ fig : hewett](a ) .
a 1 tev lc with 200 fb@xmath4 can exclude @xmath27 up to 6.6 tev , similar to the 6.0 tev achievable with the lhc in 100 fb@xmath4 using @xmath35 and @xmath36 final states only .
however , the lhc may have difficulty distinguishing a graviton signal from some other new physics process , such as a @xmath30 . at the lc , the polarized beams and the ability to observe @xmath37 and @xmath34 final states allow a clear separation between spin-1 and spin-2 exchange for @xmath27 up to about @xmath38 , as shown in figure [ fig : hewett](b )
tev - scale @xmath0 linear colliders offer complementary access to the physics of electroweak symmetry breaking that will be explored initially by the lhc . assuming that both the nlc and tesla designs prove technologically ( and financially ) feasible , the choice of which one to build may depend on the relative importance of high luminosity for the highest precision measurements at lower energies ( tesla ) , versus upgradability to 1 - 1.5 tev for exploratory physics and possible fuller elucidation of the susy spectrum .
more advanced accelerator designs , such as the two - beam clic@xcite design , may open the path to even higher energies in coming decades , ensuring a vibrant future for physics in the post - lhc era .
i would like to thank the organizers for a stimulating and enjoyable conference in a lovely setting .
this work is supported in part by doe contract number de - fg02 - 95er40899 and by nsf career award phy-9818097 .
s. kuhlmann _
et al . _ , hep - ph/9605011 .
h. murayama and m. peskin , ann .
part sci . * 46 * , 533 ( 1996 ) , hep - ex/9606003 .
e. accomando _
et al . _ ,
. rept . * 299 * , 1 ( 1998 ) , hep - ph/9705442 . c. adolphsen
_ , slac-474 ( 1996 ) .
r. brinkmann , g. materlik , j. rossbach , and a. wagner ( eds . ) , desy-1997 - 048 .
see www.slac.stanford.edu/xorg/ilc-trc/ilc-trchome.html .
parameters for the 1.5 tev nlc are given at www-project.slac.stanford.edu/lc/local/accelphysics/nlc_param.htm .
lep electroweak working group , cern - ep-99 - 015 ( 1999 ) .
sau lan wu , these proceedings
. j. hobbs , these proceedings . atlas collaboration , lhcc 99 - 14/15 ( 1999 ) .
garca - abia and w. lohmann , proceedings of the international workshop on linear colliders ( lcws99 ) , sitges , spain , 1999 ; hep - ph/9908065 .
a. juste , proceedings of the international workshop on linear colliders ( lcws99 ) , sitges , spain , 1999 ; hep - ex/9912041 .
m. battaglia , proceedings of the international workshop on linear colliders ( lcws99 ) , sitges , spain , 1999 ; hep - ph/9910271 .
t. barklow _
et al . _ , hep - ph/9704217 , in _ new directions for high energy physics : snowmass 96 _ , d. cassell _ et al .
; e. ruiz morales and m. peskin , proceedings of the international workshop on linear colliders ( lcws99 ) , sitges , spain , 1999 , hep - ph/9909383 .
k. fujii , t. matsui , and y. sumino , phys .
rev . * d50 * , 4341 ( 1994 ) ; p. comas , r. miquel , m. martinez , and s. orteu , in _ linear collisions at 500 gev ( lcws95 )
h. baer , s. dawson , and l. reina , phys . rev . *
d61 * , 013002 ( 2000 ) ; hep - ph/9906419 .
a. juste and g. merino , proceedings of the international workshop on linear colliders ( lcws99 ) , sitges , spain , 1999 ; hep - ph/9910301 .
p. horava and e. witten , nucl
. phys . * b460 * , 506 ( 1996 ) ; _ ibid .
_ , * b475 * , 94 ( 1996 ) ; e. witten , nucl . phys . *
b471 * , 135 ( 1996 ) ; i. antoniadis , phys .
b246 * , 377 ( 1990 ) ; j. lykken , phys . rev . *
d54 * , 3693 ( 1996 ) .
n. arkani - hamed , s. dimopoulos , and g. dvali , phys . lett . *
b429 * , 263 ( 1998 ) , hep - ph/9803315 ; i antoniadis , n. arkani - hamed , s. dimopoulos , and g. dvali , phys
* b436 * , 257 ( 1998 ) , hep - ph/9804398 ; n. arkani - hamed ; n. arkani - hamed , s. dimopoulos , and j. march - russell , hep - th/9809124 ; n. arkani - hamed , these proceedings .
a. mirabelli , m. perelstein , and m. peskin , phys .
82 * , 2236 ( 1999 ) , hep - ph/9811337 .
g. f. guidice , r. rattazzi , and j. d. wells , nucl . phys .
* b544 * , 3 ( 1999 ) ; t. han , j. lykken , and r. j. zhang , phys . rev . * d59 * , 105006 ( 1999 ) .
j. hewett , phys .
lett . * 82 * , 4765 ( 1999 ) , hep - ph/9811356 .
j. c. long , h. w. chan , and j. c. price , nucl .
b539 * , 23 ( 1999 ) , hep - ph/9805217 .
r. bossart _ et al .
_ , cern / ps 99 - 005 ( 1999 ) . | i discuss the motivation and physics potential of an electron - positron linear collider with a center - of - mass energy at the 1 tev scale , in light of what we may expect to learn with the lhc .
the comparison is illustrated with examples drawn from higgs physics , top quark physics , and the search for large extra spacetime dimensions . |
the characterization of accretion disks around young high - mass protostars is one of the main unsolved questions in high - mass star formation research @xcite .
the controversy arises around the difficulty to accumulate mass onto a high - mass protostar when it gets larger than 8m@xmath4 because the radiation pressure of the growing protostar may be strong enough to revert the gas inflow ( e.g. , @xcite ) .
different ways to circumvent this problem are proposed , the main are disk accretion from a turbulent gas and dust core ( e.g. , @xcite ) , competitive accretion and potential ( proto)stellar mergers at the dense centers of evolving high - mass ( proto)clusters ( e.g. , @xcite ) , and ionized accretion flows continuing through the hypercompact hii region phase ( e.g. , @xcite ) . over recent years , much indirect evidence has been accumulated that high - mass accretion disks do exist .
the main argument stems from high - mass molecular outflow observations that identify collimated and energetic outflows from high - mass protostars , resembling the properties of known low - mass star formation sites ( e.g. , @xcite ) .
such collimated jet - like outflow structures are only explainable if one assumes an underlying high - mass accretion disk that drives these outflows via magneto - centrifugal acceleration . from a theoretical modeling approach , recent 2d and 3d magneto - hydrodynamical simulations of high - mass collapsing gas cores
result in the formation of high - mass accretion disks as well @xcite .
although alternative formation scenarios are proposed , there appears to be a growing concensus in the high - mass star formation community that accretion disks should also exist in high - mass star formation .
however , it is still poorly known whether such high - mass disks are similar to their low - mass counterparts , hence dominated by the central protostar and in keplerian rotation , or whether they are perhaps self - gravitating non - keplerian entities .
while the indirect evidence for high - mass accretion disks is steadily increasing , direct observational studies are largely missing .
this lack of observational evidence can be attributed to two main reasons .
the first is the clustered mode of high - mass star formation and the typically large distances , hence spatially resolving and disentangling such structures is a difficult task .
the second difficulty is to choose the right spectral line tracer which allows unambiguous identification and characterization of the disk structure
. many spectral lines are either optically thick ( e.g. , co , hco@xmath5 , cs ) , chemically difficult to interpret ( e.g. , hc@xmath0n , hnco ) or excited in the envelope and disk which causes confusion ( e.g. , hcn , ch@xmath0cn ) .
to overcome these problems , we used the australia telescope compact array ( atca ) at 1.2 cm wavelengths including the most extended baselines ( resulting in a spatial resolution @xmath6 ) , and we aimed at the highly excited nh@xmath0(j , k ) inversion lines ( 4,4 ) and ( 5,5 ) .
nh@xmath0 is known to be a dense core tracer ( e.g. , @xcite ) , and the high ( j , k ) inversion lines with excitation temperatures ( @xmath7 ) of 200 and 295k , respectively , should only be excited in the innermost and warm region close to the central protostars .
similarly , @xcite also modeled the nh@xmath0(4,4 ) emission of the collapsing hot core g31.41 .
furthermore , radiation transfer calculations for 3d hydro - simulations revealed that the 1.2 cm band regime should be particularly well suited for such studies because the inner disk regions may be optically thick at frequencies above about 100ghz @xcite .
this may make future high - mass disk studies of the innermost regions at shorter wavelengths with alma difficult .
therefore , studying these lines at high angular resolution with the atca allows us to penetrate deeply into the natal cores and study the physical properties of the predicted high - mass accretion disks . over the last few years
there have been a lot of `` trial and error '' approaches for high - mass disk studies , but no consistent investigation of a larger sample is public so far .
since the above outlined approach has been proven very successful in the recent atca high-(j , k ) nh@xmath0 study of iras18089 - 1732 @xcite , we are now aiming at a source sample of twelve promising disk candidates mainly identified by previous lower resolution nh@xmath0 studies of 41 sources using the atca @xcite and 60 sources using mopra ( walsh et al .
priv . comm . ) .
the proposed sources comprise the best high - mass - disk - candidate sample for the southern hemisphere to date ( sec .
[ sample ] ) .
our sample of twelve sources ( see table [ tablesample ] ) was observed during six consecutive nights between july 1st and 7th 2008 .
we always observed two sources per night in a track - sharing mode cycling between the gain calibrators and sources .
table [ tablesample ] lists the corresponding gain calibrators ( phase and amplitude ) for each pair of sources .
simultaneously , we observed the nh@xmath0(4,4 ) and ( 5,5 ) inversion lines with the frequencies of the main hyperfine components at 24.139 and 24.533ghz , respectively . the phase reference centers and velocities relative to the local standard of rest ( @xmath8 )
are given in table [ tablesample ] .
bandpass and flux were calibrated with observations of 1253 - 055 and 1934 - 638 .
the spectral resolution of the observations was 62khz corresponding to a velocity resolution of @xmath9kms@xmath10 .
the observations were conducted in the 1.5d configuration including antenna 6 which resulted in a maximum baseline length of 4.3 km .
depending on the source structure ( e.g. , compact versus extended ) and strength we applied different weightings ( robust values between -2 and 2 ) for each source and line .
to better trace faint and/or extended features , we occasionally excluded antenna 6 and hence the longest baselines from the imaging process ( table [ tablesample2 ] ) .
the synthesized beams and rms are given in table [ tablesample2 ] .
.observational parameters ii [ cols= " < , > , > , > , > " , ] [ results ]
we present a large observing campaign to search for rotational motions in high - mass star - forming regions via high - excitation nh@xmath0 line observation with high spatial resolution . while we detect all sources from the original sample ( excluding the one source that was chosen via other selection criteria ) , in more than 50% of them rotational and/or infall motions can be identified .
several more general conclusions can be drawn from this observing campaign .
+ ( a ) high - excitation nh@xmath0 lines are very good tracers of the dense gas within hot - core type young high - mass star - forming regions . in this sample , the nh@xmath0(4,4 ) and ( 5,5 ) emission is always associated with the ch@xmath0oh class ii methanol maser emission , which is a well - known signpost for high - mass star formation .
+ ( b ) we identify rotational / infalling motions in half of the observed sources . + ( c ) the signatures comprise simple velocity gradients as well as more spherical infall signatures
. + ( d ) although the spatial resolution is of order a few 1000au , we do not identify obvious keplerian signatures .
+ ( e ) while high optical depth is an issue , the data nevertheless indicate that accretion disks resembling those of low - mass star - forming regions ( e.g. , @xcite ) have to be typically of smaller sizes , not yet resolved by our observations .
likely such smaller accretion disks could be fed by the rotating envelopes observed here .
+ ( f ) in many fields of view we do detect additional hii regions that are not directly linked to our targets but that reside in the same projected area on the sky .
this is additional evidence for the clustered mode of high - mass star formation .
+ where to go from here ?
clearly there are different paths to be followed .
while our approach for this study was biased toward hot - core type sources based on our spectral line selection criteria , other observational approaches can target sources in different evolutionary stages as well as of different luminosities . in particular , observations at mm wavelengths with broad spectral bandpasses allow us to observe several spectral lines simultaneously which can trace rotational motions in sources of different characteristics ( e.g. , @xcite ) .
for example , qizhou zhang and collaborators started a similarly large project searching for disk signatures toward a sample of young high - mass star - forming regions with the submillimeter array , and they also detected rotational signatures toward all targets ( qizhou zhang et al .
, in prep . ) .
while as outlined in section [ general ] flattened structures do exist on scales exceeding @xmath2au , most data indicate that the keplerian accretion disks should reside on smaller scales , probably @xmath11au .
this is also shown in the recent 3-dimensional radiative hydrodynamic simulations by @xcite where most of the disk - like structures are contained with a radius of @xmath1500au .
furthermore , recent work has also shown that accretion disks exceeding 150au are prone to fragmentation ( e.g. , @xcite ) , hence , it is not even expected to find keplerian signatures on much larger scales .
while we can already achieve tremendous progress with the existing instrumentation , clearly we are still lacking spatial resolution if we want to understand the structure and physical processes of the more proper central accretion disks .
future instruments like alma and evla will allow us to make considerable progress in this direction for the cold gas components .
furthermore , due to the large accretion rates , strong viscous forces and central luminous sources , a considerable fraction of warm / hot gas is expected to exist with the high - mass disks .
therefore , imaging these warm / hot dust and gas components at high spatial resolution with future instruments like miri on jwst will allow us to constrain the properties of these components in much more detail .
it is likely that only a concerted effort at long and short wavelengths combined with hydrodynamic and radiation transfer modeling will give us a real understanding of the working of high - mass accretion disks .
the australia telescope compact array is part of the australia telescope which is funded by the commonwealth of australia for operation as a national facility managed by csiro .
acknowledges financial support by the emmy - noether - program of the deutsche forschungsgemeinschaft ( dfg , grant be2578 ) . | to better understand the physical properties of accretion disks in high - mass star formation , we present a study of a dozen high - mass accretion disk candidates observed at high spatial resolution with the australia telescope compact array ( atca ) in the high - excitation ( 4,4 ) and ( 5,5 ) lines of nh@xmath0 .
all of our originally selected sources were detected in both nh@xmath0 transitions , directly associated with ch@xmath0oh class ii maser emission and implying that high - excitation nh@xmath0 lines are good tracers of the dense gas components in hot - core type targets .
only the one source that did not satisfy the initial selection criteria remained undetected . from the eleven mapped sources ,
six show clear signatures of rotation and/or infall motions .
these signatures vary from velocity gradients perpendicular to the outflows , to infall signatures in absorption against ultracompact hii regions , to more spherical infall signatures in emission .
although our spatial resolution is @xmath11000au , we do not find clear keplerian signatures in any of the sources .
furthermore , we also do not find flattened structures .
in contrast to this , in several of the sources with rotational signatures , the spatial structure is approximately spherical with sizes exceeding @xmath2au , showing considerable clumpy sub - structure at even smaller scales .
this implies that on average typical keplerian accretion disks if they exist as expected should be confined to regions usually smaller than 1000au .
it is likely that these disks are fed by the larger - scale rotating envelope structure we observe here .
furthermore , we do detect 1.25 cm continuum emission in most fields of view . while in some cases weak cm continuum emission is associated with our targets , more typically larger - scale hii regions
are seen offset more than @xmath3 from our sources .
while these hii regions are unlikely to be directly related to the target regions , this spatial association nevertheless additionally stresses that high - mass star formation rarely proceeds in an isolated fashion but in a clustered mode . |
very recently the h dibaryon was put back on the agenda by lattice qcd calculations of different collaborations , nplqcd @xcite and hal qcd @xcite , providing evidence for a bound state at pion masses larger than the physical ones .
it was first proposed thirty years ago @xcite as a spin and flavor singlet composed of six quarks ( @xmath0 ) .
such a proposal emerged on the context of the mit bag model @xcite . when applied to all six quark systems it predicted the existence of only one stable dihyperon with @xmath1 and a mass of 2150 mev , 81 mev below the @xmath2 threshold ,
therefore strongly bound . moreover , being the lightest particle of a two - baryon system with strangeness ( @xmath3 ) 2 , it would be stable against the strong interaction and would necessarily decay weakly .
it was assumed to be a single hadron made of six quarks squeezed in a small region and not a two baryon state bound in an @xmath4 wave like the deuteron . if this were the case it would open the door to the exotic states , i.e. , hadrons that do not fit in the standard @xmath5 or @xmath6 configurations .
otherwise , the fact that the h dibaryon is the most promising candidate to be the second dibaryonic state after the deuteron has generated a lot of expectation . on the experimental side , there are very few data in the @xmath7 sector , coming from the inelastic @xmath8 cross section at a lab momentum of around 500 mev / c , and from the elastic @xmath9 and inelastic @xmath10 cross sections for lab momenta in the range of 500 600 mev / c @xcite . thus , the relevant information we have is indirect and comes from double@xmath11 hypernuclei .
their binding energies , @xmath12 , provide upper limits for that of the h dibaryon , i.e. , @xmath13 .
the first hypernuclear events are quite old and admit several interpretations @xcite . in 2001
it was reported the so called nagara event @xcite , interpreted uniquely as the sequential decay of @xmath14he emitted from a @xmath15hyperon nuclear capture at rest .
the mass and the values of @xmath12 and of the @xmath2 interaction energy , @xmath16 , were determined without ambiguities .
the small value of @xmath16 suggested an attraction weaker than the one previously estimated .
it also gave the most stringent constraint to the mass of the h dibaryon to date , i.e. , @xmath17 mev at a 90% confidence level .
it took almost one decade , but four more double@xmath11 hypernuclear events were reported , from kek e176 and e373 experiments @xcite , still with preliminary results .
all the details are summarized in table [ hip ] .
.reported double hypernuclear events .
[ cols="^,^,^,^",options="header " , ] [ contrib ] in our case the situation is qualitatively similar to refs .
the oge interaction in the three diagonal channels is plotted in fig .
[ gluon ] . the repulsive character of the gluon exchange in @xmath2 exceeds by far the attraction present in @xmath18 , and even more in @xmath19 .
therefore if one looks at the binding energy with only gluon interaction , one finds that the fredholm determinant is larger than 1 , what is associated to a repulsive force .
this is a general conclusion valid for any value of @xmath20 , since it is a multiplicative constant that appears in the three channels and the ratio between the repulsion in @xmath2 and the attraction in @xmath18 is preserved .
indeed , such an approximated ratio ( @xmath21 ) is also found in @xcite , where a su(6 ) quark model is used , in spite of the huge difference in the quark
gluon coupling constant , taken there to be @xmath22 1.52 , almost one order of magnitude larger .
the global effect of the chromomagnetic interaction is thus found to be repulsive and therefore one has to keep looking for the responsible of the binding . in order to appreciate the overall features of any interaction , the @xmath7 ,
@xmath23 two baryon system has been solved in the one two and three channel approaches by including only the interactions indicated in each line of table [ contrib ] .
this gives us information on the character of the interaction involved , whether it is repulsive , weakly attractive or attractive enough to bind the system .
the pseudoscalar interaction generates repulsion , and it contributes only to the @xmath2 channel , being negligible in the others . therefore in a model containing only pse , or oge + pse interactions
there is not a bound state .
the same situation occurs in refs .
@xcite , where a resonance 26 mev above threshold is found in a oge + pse model .
however , the addition of a scalar exchange at the baryonic level provides large attraction so that the binding becomes @xmath24 20 mev . in ref .
@xcite the pseudoscalars also introduce repulsion at short distances , but weaker than in the @xmath25 interaction , in such a way that a bound h particle provided there were medium range attraction as the @xmath26 implies is not discarded .
repulsive pseudoscalar contribution is also found in @xcite , together with an attractive @xmath26 piece , for a @xmath27 configuration . when considering the two cluster configuration with a pure chiral content the bound state does not appear .
the reason for that is not a repulsive character of the chiral meson exchange but the lack of strength .
because of that , an increment of the chiral coupling constant in order to properly describe the @xmath28 phase shifts is successful for binding the system . in our model , the @xmath29 is very attractive and provides by itself the largest values for @xmath30 , from 5.4 mev to 8.3 mev .
let us notice that in a pure @xmath29 model the two transition potentials to @xmath19 would not be possible , since they are due to the scalar meson exchanges .
then the @xmath2 would not be coupled to @xmath19 and so the two channel calculation would give the same result as the pure @xmath2 , as can be checked in table [ contrib ] .
when adding the @xmath29 to the oge + pse terms , an unbound but near threshold h dibaryon is obtained in all cases . as in ref .
@xcite , quite close to our treatment , the @xmath29 is the only attractive piece .
the contribution coming from the exchange of the scalar octet is weakly attractive or repulsive , as can be seen in table [ contrib ] .
however , when supplementing the oge + pse + @xmath29 potential with the octet of scalars , a bound state appears in the two channel calculation with @xmath31 1.6 mev and @xmath32 7.0 mev in the three channel case .
it is exactly at this point where the important role of the scalars in our model lies , or more precisely , that of the @xmath33 exchange .
this exchange piece manages to enlarge the binding energy without being attractive itself because it is the main piece in the @xmath2@xmath34 and @xmath34@xmath18 transitions .
it redistributes part of the flux to the much more attractive @xmath34 and @xmath18 channels , originating thus an appreciable increase in the binding .
similar results have been found in the quark model framework of ref .
@xcite .
as deduced from our discussion above , the binding energy of the dibaryon depends mostly on the @xmath29 and @xmath33 exchange potentials .
therefore it is interesting to evaluate the dependence of the binding on the parameters involved in these two interactions , more precisely on their masses and coupling constants .
varying the @xmath29 and particularly the @xmath33 coupling constants along a reasonable range changes the binding energy as seen in fig .
[ depg ] , being these changes larger in the first case . in both cases
@xmath30 grows with the coupling constant .
indeed it can be checked that the change on @xmath35 , that affects the exchange of every meson , produces quite the same results as the change on @xmath36 exclusively .
the dependence on the masses is opposite , see fig .
[ depm ] for details .
the binding decreases as the masses get larger , due to the shorter range of the attraction .
this behavior is more pronounced in the case of @xmath37 .
@xmath30 ranges approximately from 4 mev to 12 mev for @xmath38@xmath39 .
both dependence features ( masses and couplings ) are general , and similar qualitative results were also found in @xcite . the flavor symmetry breaking ( fsb ) is expected to be an important effect in the h dibaryon due to the presence of two strange quarks .
the present study has been made under the assumption that su(3 ) is broken .
several sources of symmetry breaking are taking part in our work .
first of all , the strange quark has been assigned a mass different to the light quark masses ( @xmath40 = 555 mev @xmath41 313 mev ) .
it enters in the expressions of the oge and pseudoscalar potentials through the factor @xmath42 .
however , the differentiation of the @xmath43quark in the wave function through the harmonic oscillator parameter @xmath44 gives the larger source of fsb . besides giving different values for the orbital matrix elements ,
it implies the suppression of a number of diagrams because light and strange quarks , as distinguishable particles , can not be exchanged . in order to roughly evaluate the effect of fsb on the h dibaryon
, the binding energy has been recomputed when some sources of symmetry breaking have been eliminated , i.e. , for @xmath45 313 mev and @xmath46 .
we have variated in fig .
[ par ] the harmonic oscillator parameter as well as the mass of the strange quark , between their values for exact su(3 ) and their true values in our model .
one can see how the binding energy increases as one approaches the exact su(3 ) limit . as a consequence ,
the binding is increased by around 45@xmath47 , giving @xmath24 10.2 mev .
thus , we can conclude that fsb lowers the attraction .
a similar conclusion was obtained in ref .
@xcite , through a rgm calculation with only oge and confinement potentials . there , the binding energy drastically changed from 38 mev for exact su(3 ) to 26 mev above threshold when the fsb effect was incorporated .
a similar qualitative result was found by ref .
@xcite , where the authors concluded that the symmetry breaking diminished the strong attraction coming from the chromomagnetic interaction . from chiral effective field theory @xcite ,
also the bound state found for exact su(3 ) disappears when using physical values for the baryon and meson masses .
a closer inspection of every component of our potential in all the six channels contributing to h allows us to conclude that the effect of fsb is to diminish the strength of the potential , be it repulsive or attractive .
therefore , depending on the combination of the contributions in any channel , the total potential will be larger or smaller than the one with fsb , but no other general conclusion can be inferred . in our particular case , it turns out that @xmath2 and @xmath18 hardly change , whereas @xmath34 gains attraction when approaching su(3 ) . as for the transitions , the only appreciable change takes place in the @xmath18@xmath34 interaction , that acquires twice as much attraction for the exact su(3 ) case
this result is in contrast with the one found in @xcite , where the @xmath2@xmath34 interaction is strongly dependent on a fsb parameter . at first sight
there are two types of possible configurations for the h dibaryon : the six quark cluster and the two baryon cluster . as we have already mentioned the first one
was employed in the original work @xcite , getting a bound h. however latter calculations found that a dibaryon with only the @xmath27 configuration would not bind @xcite .
the spatial distribution was also investigated in ref .
@xcite , where an extended resonating group method was employed in order to account for the possibility of a change in the baryon wave functions .
again , the dibaryon did not bind in the @xmath27 configuration , being thus necessary to consider less compact configurations , due to the medium range attraction .
when enlarging the size of the wave function , the dibaryon became more attractive , being @xmath48 18.2 mev for the stable solution .
the probabilities however did not drastically change from the more compact ( @xmath27 ) to the more spread ( two cluster ) configuration . in the latter case
they were found to be @xmath49 0.216 , @xmath50 0.543 and @xmath51 0.242 .
this result is slightly different from that in the flavor singlet state , eq .
( [ hbcon ] ) .
the small difference justifies the perturbative treatment employed , since the wave function was first postulated to be the flavor singlet when performing the rgm calculation . usually , the flavor singlet wave function , eq .
( [ hbcon ] ) , is first postulated and a perturbative variational calculation is afterwards performed when making use of the rgm treatment .
this is not the case of our work , since the coefficients of the baryon baryon components of the flavor wave function are obtained as an output of the calculation , without making further assumptions .
the probabilities we got are @xmath49 0.177 , @xmath52 0.446 and @xmath51 0.377 .
they are quantitatively similar to those of the flavor singlet , from what we can infer that in our model the baryon
baryon wave function is at first approximation su(3 ) symmetric , being the difference between both due to the flavor symmetry breaking effects . in order to properly understand the @xmath53 sector ,
the analysis performed on the h dibaryon and the results obtained should be complemented studying the other spin isospin channels . from such a complete analysis , more general and powerful conclusions may arise . unlike the @xmath54 h dibaryon channel , the @xmath55
is found to be very weakly attractive .
only the @xmath19 interaction is present .
the @xmath56 system does not couple to @xmath57 and the wave function of @xmath2 and @xmath18 is antisymmetric , and therefore vanishes , for @xmath58 even . since @xmath2 is forbidden , there are no lighter states and therefore the @xmath19 state could decay weakly . the @xmath59 potential , fig .
[ t0](@xmath60 ) , exhibits a repulsive core , due to the strong oge interaction that compensates by far the attraction coming from the @xmath29 . at intermedium range
the attractive feature survives but is not enough to form a bound state . the weak attraction in @xmath19 can be enhanced if the tensor force is turned on , however a bound state is still not possible .
this qualitative behavior coincides with that of refs .
@xcite and nijmegen models d and f @xcite .
the big difference from the h channel is an example of the importance of the flavor dependence interaction that model ii incorporates .
a summary of the predictions for the scattering lengths of the @xmath7 two - baryon systems with modern versions of the nijmegen potential @xcite , the jlich potential @xcite , and the work by fujiwara _
et al . _
@xcite can be found in table 2 of ref .
@xcite .
the potentials in the two @xmath61 channels are all attractive . regarding the @xmath62 , both @xmath19 and @xmath56 interactions are strongly attractive , as can be seen in fig . [ t1](@xmath63 ) .
such a strong attraction provides for the coupled system a binding energy of 4.8 mev .
however this is not the case in ref .
@xcite , where all the potentials with @xmath61 are repulsive . as for the @xmath64 channel ,
there are three coupled interactions : @xmath19 , @xmath56 and @xmath18 .
the three of them are moderately attractive and look very similar , as can be seen in fig .
[ t1](@xmath60 ) .
the nijmegen models d and f @xcite also predict attractive interactions .
however , no bound state appears in our model .
we have performed the first calculation of the h dibaryon in a model constrained by the elastic and inelastic @xmath65 , @xmath66 and @xmath2 cross sections .
special interest has been devoted to analyze the contribution of the different pieces of the interaction and to the effect that the addition of channels to the @xmath2 system produces on the binding .
we obtained a bound h dibaryon , with @xmath24 7 mev , compatible with a plausible extrapolation of recent lattice qcd results and with the nagara event , the most stringent restriction known so far .
the scalar octet exchange , although not giving an attractive contribution by itself , plays a key role as it is the main ingredient of the transition potentials that connect @xmath2 to the more attractive @xmath19 and @xmath18 states .
the probabilities @xmath49 0.177 , @xmath52 0.446 and @xmath51 0.377 are quite similar to those of the flavor singlet so that our wave function is at first approximation su(3 ) symmetric .
finally , a bound state has also been found in the @xmath7 , @xmath67 channel , with a binding energy of 4.8 mev , thus smaller than @xmath30 .
the abundance of events foreseen in a near future and/or the improvement on the lattice calculations will help us to advance in our knowledge of the mechanisms that play important roles in the dynamics of the h dibaryon .
for such a task , a detailed theoretical study as the one presented here could be relevant .
s. r. beane _ et al . _
[ nplqcd collaboration ] , phys .
* 106 * , 162001 ( 2011 ) .
t. inoue _ et al .
[ hal qcd collaboration ] , phys .
lett . * 106 * , 162002 ( 2011 ) .
r. l. jaffe , phys .
lett . * 38 * , 195 ( 1977 ) [ erratum - ibid .
* 38 * , 617 ( 1977 ) ] .
a. chodos , r. l. jaffe , k. johnson , c. b. thorn , and v. f. weisskopf , phys .
d * 9 * , 3471 ( 1974 ) ; t. a. degrand , r. l. jaffe , k. johnson , and j. e. kiskis , phys . rev .
d * 12 * , 2060 ( 1975 ) .
j. k. ahn _
et al . _ ,
phys . lett .
b * 633 * , 214 ( 2006 ) .
d. j. prowse , phys .
* 17 * , 782 ( 1966 ) .
s. aoki _ et al .
_ , prog .
* 85 * , 1287 ( 1991 )
. h. takahashi _ et al .
_ , phys .
lett . * 87 * , 212502 ( 2001 ) .
k. nakazawa [ kek - e176 and j - parc - e07 collaboration ] , nucl .
phys . a * 835 * , 207 ( 2010 ) .
u. straub , z. y. zhang , k. brauer , a. faessler , and s. b. khadkikar , phys .
b * 200 * , 241 ( 1988 ) ; u. straub , z. y. zhang , k. brauer , a. faessler , s. b. khadkikar , and g. lbeck , nucl .
a * 508 * , 385c ( 1990 ) ; a. faessler and u. straub , prog .
* 24 * , 323 ( 1990 ) . | we perform a coupled channel calculation of the h dibaryon within a chiral constituent quark model .
the problem is solved within a quark model constrained by the experimental data of strangeness 1 and 2 two baryon systems .
we examine in detail the role played by the different contributions of the interacting potential as well as the number of coupled channels considered .
special attention has been payed to the parameter dependence , flavor symmetry breaking and spatial configurations .
the value extracted for the binding energy of the h dibaryon , being compatible with the restrictions imposed by the nagara event , falls within a plausible extrapolation of recent lattice qcd results . |
plasma oscillations in heterostructures with a two - dimensional electron gas ( 2deg ) and some devices using the excitation of these oscillations have been experimentally studied over decades ( see , for instance , refs .
there are also many theoretical papers on different aspects of these plasma waves and their potential applications .
the gated and ungated 2deg channels in heterostructures can serve as resonant cavities for terahertz ( thz ) electron plasma waves .
@xcite the resonant properties of such channels can be utilized in different thz devices , in particular , resonant detectors .
@xcite the mechanism of the thz detection observed in transistor structures @xcite might be attributed to the nonlinearity of the plasma oscillations as suggested previously .
@xcite the variation of the conductivity of 2deg with periodic gate system can be possibly explained by the heating of 2deg by absorbed thz radiation ( heating or bolometric mechanism ) .
@xcite recently , a concept of resonant detectors in which a plasma resonant cavity is integrated with a schottky junction has been proposed and substantiated .
@xcite as shown , when the frequency of incoming thz radiation is close to one of the plasma resonant frequencies , the ac potential drop across the barrier at the schottky junction becomes resonantly large .
this leads to rather large values of the rectified component of the current through the junction , which is used as output signal of the detector .
since the resonant excitation of the plasma oscillations by absorbed thz radiation leads to the heating of 2deg , the heating ( bolometric ) mechanism can also contribute to the rectified current over the schottky barrier .
thus , two mechanisms can be responsible for the operation of this detector : the dynamic mechanism considered previously @xcite and the heating mechanism .
similar mechanisms can work in the detectors utilizing the excitation of plasma oscillations but having the barrier of another origin .
in particular , the barrier formed by an additional gate which results in an essential depletion under it was used in recent experiments .
@xcite in this paper , we develop a device model for a thz resonant detector based on a heterostructure with the 2deg channel and the barrier region utilizing the plasma oscillations excitation .
we consider a device with three gates ( see fig .
1 ) : the 2deg under two extreme gates forms the plasma resonant cavities , while the central gate , to which a sufficiently large negative bias voltage is applied , forms the barrier .
the excitation of the plasma oscillations by the incident thz radiation is assumed to be associated with an antenna connected with the source and drain contacts to the 2deg channel . in the following we demonstrate that each of the mechanisms in question can dominate in different temperature ranges ( the dynamic mechanism is predominant at elevated temperatures , whereas the heating mechanism provides larger contribution at low temperatures , @xmath0 k ) .
the device structure under consideration differs from that studied in ref .
@xcite in which the thz radiation input was realized by a periodic grating
. however , the physical mechanisms of detection are the same .
we assume that the net drain - source voltage @xmath1 includes both the dc bias voltage @xmath2 and the ac component @xmath3 induced by incoming thz radiation via an antenna connected with the source and drain contacts .
thus , @xmath4 , where @xmath5 is the thz radiation frequency . the bias voltage @xmath6 is applied between the side gates and the pertinent contacts .
this voltage affects the electron density in the quasi - neutral portions of the 2deg channel , so that the dc electron density is given by @xmath7 where @xmath8 is the electron charge , @xmath9 is the gate layer thickness , and @xmath10 is the dielectric constant . it is also taken to be that the potential drop , @xmath11 , between the central gate and the source contact is negative and its absolute value is sufficiently large to deplete some portion of the 2deg channel underneath this gate and create a potential barrier .
this implies that @xmath12 , where @xmath13 is the 2deg threshold voltage .
the ac component of the electric potential in the channel @xmath14 satisfies the boundary conditions which follow from the fact that at the source and drain contacts its values coincide with the ac components of the incoming signal voltage : @xmath15 assuming that at the conductivity of the quasi - neutral sections is much larger than that of depletion region ( including the thermionic current over the barrier and the displacement current associated with the capacitances shown in the equivalent circuit of fig .
1 ) one may use the following additional conditions : @xmath16 here @xmath17 is the length of the 2deg channel between the source and drain contacts , @xmath18 is the length of the depletion region under the central gate , and @xmath19 is the length of the latter ( see fig . 1 ) .
the quantity @xmath20 is determined by the potential of the central gate @xmath11 . definitely , the conductivity of the quasi - neutral sections markedly exceeds the real part of the depleted region admittance .
the role of the displacement current across the depleted region is discussed in viii .
the electron transport in the channel is governed by the hydrodynamic equations coupled with the poisson equation for the self - consistent electric potential @xcite . at sufficiently small intensities of thz radiation
, this system of equations can be linearized and reduced to the following equation @xcite valid in the quasi - neutral regions of the channel ( @xmath21 and @xmath22 ) : @xmath23 where @xmath24 is the frequency of electron collisions with impurities and phonons , @xmath25 is the plasma wave velocity , and @xmath26 is the electron effective mass .
the effect of electron pressure results in some renormalization of the plasma wave velocity : @xcite @xmath27 , where @xmath28 is the velocity of the electron `` sound '' ( usually @xmath29 ) .
assuming that the central gate length is comparable with the gate layer thickness , the shape of the barrier under the central gate can be considered as parabolic : @xmath30 - e(v_d + \delta v^{b})/2 $ ] , where @xmath31 is the barrier height in equilibrium ( it depends on potential , @xmath11 , of the central gate ) , and @xmath32 is the lateral potential drop across the depletion ( barrier ) region associated with the dc bias voltage @xmath2 and the ac voltage @xmath33 .
the variation of the barrier height @xmath34 due to the lateral potential drop is given by @xmath35 \simeq - \frac{e(v_d + \delta v^{b})}{2}.\ ] ] hence , at @xmath36 the net source - drain current can be calculated using the following formula : @xmath37 - 1\biggr\},\ ] ] where @xmath38 is the maximum current ( for a 2deg @xmath39 ) and the ac potential drop , @xmath33 , across the depletion ( barrier ) region is given by @xmath40 and @xmath41 is the electron temperature , @xmath42 is the lattice temperature , and @xmath43 is the variation of the electron temperature due to thz irradiation . in this case , at @xmath44 , eq .
( 6 ) yields the following equation for the dc source - drain current @xmath45 : @xmath46\end{aligned}\ ] ] at @xmath47 and @xmath48\end{aligned}\ ] ] at @xmath49 .
here , @xmath50\ ] ] is the dc source - drain current without thz irradiation ( dark current ) .
solving eq .
( 4 ) with boundary conditions ( 2 ) and ( 3 ) and taking into account eq . ( 7 ) , we obtain @xmath51 } + \frac{e^{- i\omega\ , t}}{\cos [ q_{-\omega}(l - l)]}\biggr\}.\ ] ] here @xmath52 equation ( 11 ) yields @xmath53^{-1},\ ] ] where @xmath54 is the characteristic plasma frequency . due to the dependence ( given by eq .
( 1 ) ) of the dc electron density on the gate voltage @xmath6 , the characteristic plasma frequency @xmath55 can be tuned by this voltage . because of the dependence of @xmath55 on the length , @xmath56 of the quasi - neutral sections of the getated 2deg channel ( serving as the resonant plasma cavities ) and , hence , on the potential of the central gate @xmath11
, @xmath55 is somewhat varied with varying @xmath11 .
the power absorbed by the channel can be calculated using the following equation : @xmath57 considering eq .
( 15 ) , the energy balance equation governing the averaged electron temperature can be presented as @xmath58 where @xmath59 is the electron energy relaxation time .
the factor @xmath60 is associated with the nonuniformity of the ac electric field under the side gates .
this factor is given by @xmath61.\end{aligned}\ ] ] at @xmath62 , eq .
( 17 ) yields @xmath63,\ ] ] so that one can use the following simplified formula : @xmath64
using eqs . ( 8) and ( 15 ) , we obtain @xmath65 where the term @xmath66 is associated with the contribution of the electron heating to the rectified current . combining eqs .
( 13 ) and ( 20 ) , we arrive at @xmath67 } \biggl(\frac{ev_{\omega}}{2k_bt_0}\biggr)^2.\ ] ] since @xmath68 is proportional to the incoming thz power , the detector responsivity as a function of the signal frequency and the structural parameters ( except the antenna parameters ) can be in the following form @xmath69 } \frac{(1 + h)}{\sqrt{k_bt_0}}.\ ] ] as follows from eq .
( 23 ) , the responsivity as a function of the frequency of incoming thz radiation exhibits sharp peaks at the plasma resonant frequencies @xmath70 , where @xmath71 is the resonance index , provided that the quality factor of resonances @xmath72 .
as seen from eq . ( 22 ) , the ratio of the heating and dynamic components of the rectified dc source - drain current is given by the quantity @xmath73 ( heating parameter ) . at @xmath62 , considering eqs .
( 18 ) and ( 19 ) , the quantity @xmath73 can be estimated as @xmath74 considering that @xmath75 , eq .
( 24 ) can be presented in the following form : @xmath76 the first factor in eq .
( 25 ) is usually small , while the second one can be large , so that @xmath73 can markedly exceed unity , particularly , at low temperatures .
it is natural to assume that in a wide range of the temperature ( from liquid helium to room temperatures ) the electron collision frequency ( the inverse momentum relaxation time ) is determined by the electron interaction with acoustic phonons and charged impurities , while the electron energy relaxation is due to the interaction with both acoustic and polar optical phonons .
hence , @xmath77 and @xmath78 $ ] , where @xmath79 and @xmath80 are related to the electron scattering on acoustic phonons and impurities ( charged ) , respectively , whereas @xmath81 and @xmath82 are the electron energy relaxation times associated with the acoustic and optical phonons . as a result
, the product @xmath83 can be presented as @xmath84 \tau_{\varepsilon}^{(ac)}\tau_{\varepsilon}^{(op)}}{[\tau_{\varepsilon}^{(ac ) } + \tau_{\varepsilon}^{(op)}]}.\ ] ] we shall use the following formulas for the temperature dependences of the parameters in this equation : @xmath85 @xmath86,\ ] ] @xmath87 where @xmath88 ps , @xmath89 ps , and @xmath90 ps @xcite are the characteristic scattering times for the acoustic phonon scattering ( polar and deformation mechanisms , respectively ) and for the optical phonon scattering , and @xmath91 is the optical phonon energy .
the temperature dependence of the collision frequency of 2d electrons with charged impurities is assumed to be as follows : @xmath92 here the function @xcite @xmath93 characterizes a decrease in the collision frequency with increasing thickness of the spacer , @xmath94 , between the 2deg channel and the charged donor sheet , and @xmath95 @xmath96 .
the characteristic frequency @xmath97 is proportional to the donor sheet density .
we set @xmath98 s@xmath99 . at @xmath100 k , these values correspond to @xmath101s@xmath99 and the electron mobilities @xmath102 @xmath103/v@xmath104s . when @xmath105 , eq .
( 30 ) yields @xmath106 while at sufficiently thick spacers , one obtains @xmath107 . in the case of relatively thick electron channels ,
in which the quantization is insignificant , @xmath107 . .
the inset shows the temperature dependences of the electron collision frequencies associated with impurity and acoustic scattering mechanism.,width=321 ] figure 2 shows the temperature dependence of the `` heating '' parameter @xmath73 , which determines the relative contribution of the heating mechanism to the detector responsivity , calculated for different values of parameter @xmath108 , i.e. , different doping levels .
it is assumed that for a gaas channel @xmath109 k. @xcite we set @xmath110 cm / s , @xmath111 cm / s , the barrier height @xmath112 mev , and the bias voltage @xmath113 mv , so that the effective barrier height @xmath114 mev . as seen from fig . 2
, the heating mechanism can provide significantly larger contribution to the detector responsivity at low temperatures ( @xmath115 k ) .
however at elevated temperatures , this mechanism becomes relatively inefficient .
this is attributed to weak electron heating due to strong energy relaxation on optical phonons at elevated temperature .
as mentioned in sec .
v , the detector resonsivity exhibits sharp peaks at the plasma resonant frequencies @xmath70 .
the sharpness of these peaks depends on the electron collision frequency @xmath24 , which in turn depends on the temperature . as follow from eq .
( 23 ) , the maximum values of the detector responsivity at the fundamental resonance ( @xmath116 ) is given by @xmath117 as a result , for the value of the ratio of the detector responsivity at the resonance and the dark current one obtains @xmath118 figure 3 shows the temperature dependences of the maximum ( resonant ) values of the dynamic and heating contributions to the detector responsivity devided by the dark current value calculated using eq .
( 32 ) for the barrier height @xmath119 mev , the fundamental plasma frequencies @xmath120 thz . and different values of the parameter @xmath108 .
the inset shows the temperature dependence of max @xmath121 , which accounts for both mechanisms .
, of the detector responsivity maximum ( resonant ) value and the dark current .
the inset shows the net value @xmath122 as a function of temperature.,width=321 ] for different values of the effective barrier height @xmath123.,width=321 ] taking into account that the detector detectivity @xmath124 , we arrive at the following formula : @xmath125 } \frac{(1 + h)}{(k_bt_0)^{5/4}}.\ ] ] figure 4 shows the temperature dependences of the detector detectivity calculated for different values of the effective barrier height @xmath126 . as follows from the above formulas , the detector responsivity and detectivity increase with decreasing effective barrier height
however , this quantity can not be set too small ( at a given temperature @xmath42 ) in the frame of our model , which is valid if @xmath127 , and when the 2deg channel can be partitioned into the quasi - neutral and deleted sections .
the displacement current across the depleted region can be , in principle , essential . to take this current into consideration one
needs to modify eq . ( 3 ) .
@xcite introducing the admittance , @xmath128 , of the barrier region , we can use the following boundary conditions at the edge of the quasi - neutral sections of the 2deg channel ( for the asymmetrical plasma modes for which @xmath129 ) : @xmath130 here , @xmath131 $ ] is the ac conductivity of the quasi - neutral sections of the channel accounting for the electron collisions and inertia .
since the real part of the conductivity of the depleted ( barrier ) region is definitely small in comparison with the conductivity of the qusi - neutral sections , one can take into account only the capacitive component of the current across the depleted region . in this case ,
@xmath132 , where @xmath133 is the pertinent capacitance , which is determined by the capacitances @xmath134 and @xmath135 : @xmath136 ( see the equivalent circuit in fig . 1 ) .
this equivalent circuit accounts for the inductance of the quasi - neutral sections @xmath137 , due to the inertia of the electron transport along the channel , the channel resistance due to the electron scattering , the resistance of the barier region @xmath138 , and different capacitances .
equation ( 34 ) can be presented in the form : @xmath139 where @xmath140 if the parameter @xmath141 , eqs . ( 3 ) and ( 35 ) practically coincide ( except the case @xmath142 ) . according to ref .
@xcite @xmath143 and @xmath144 , where @xmath145 and @xmath145 are logarithmic factors which are deternined by the geometry of the planar conducting areas ( the gates and quasi - neutral sections of the channel ) .
these factors can be estimated as : @xcite @xmath146 .
hence , @xmath147 , where @xmath148 .
the factor @xmath149 somewhat exceeds unity although it is not too large .
thus , taking into account that @xmath150 , one can conclude that for real device structures @xmath141 .
the condition of the smallness of @xmath151 can be also presented in the form @xmath152 , where @xmath153 is the capacitance of the side gates ( see fig . 1 ) .
relatively small capacitances @xmath135 and @xmath134 can , to some extent , affect the plasma oscillations .
however , their role reduces mainly to a small modification in the resonant plasma frequencies .
@xcite if @xmath151 would be large , the ac potential near the edges of the quasi - neutral regions is small as well as the ac potential drop across the depleted region .
this implies that the boundary condition ( 34 ) could be @xmath154 .
in such a case , the fundamental plasma frequency is doubled and dynamic mechanism is weakened . however , the details of the ac potential distribution near the edges of the quasi - neutral regions are not crucial for the heating mechanism . actually , there is some delay in the electron transit across the barrier .
the delay time can be estimated as @xmath155 , where @xmath156 is the thermal electron velocity .
the delay in the electron transit and , therefore , the electron transit time effects can be disregarded if @xmath157 .
@xcite assuming @xmath120 thz and @xmath158 cm / s , from the last inequality we obtain @xmath159 m .
we developed a device model for a resonant detector of thz radiation based on the gated 2deg channel with an electrically induced barrier .
the model accounts for the resonant excitation of the plasma oscillations in the gated 2deg channel as well as the rectified dc current through the barrier . as shown , this current comprises two components : the dynamic component associated with the ac potential drop across the barrier and the heating ( bolometric ) component due to a change in the electron temperature which stems from the resonant plasma - assisted absorption of thz radiation . using our model , we calculated the frequency and temperature dependences of the detector responsivity .
the detector responsivity exhibits sharp resonant peaks at the frequencies of incoming thz radiation corresponding to the plasma resonances in the gated 2deg channel .
the plasma resonances can tuned by the gate voltage @xmath11 and , to some extent , by the potential of the central gate @xmath11 .
it is demonstrated that the dynamic mechanism dominates at elevated temperatures , whereas at low lattice temperatures ( @xmath115 k for algaas / gaas based detectors ) , the heating mechanism prevails .
this can be attributed to a marked increase in the electron energy relaxation time @xmath59 with decreasing lattice temperature . at rather low temperatures
when the ratio @xmath160 is large , the thermionic electron current over the barrier can be surpassed by the tunneling current .
the rectified portion of this current can be associated with both dynamic and heating mechanisms ( the thermo - assisted tunneling current in the case of the latter mechanism ) .
this , however , requires a separate detailed study . in the case of detection of thz radiation
modulated at some frequency @xmath161 , the relative contributions of the dynamic and heating mechanisms to the responsivity , @xmath162 , characterising the detector response at the frequency @xmath163 can be different than those considered above .
this is due to different inertia of these mechanisms . in particular , if @xmath164 ( but @xmath161 ) , the variation of the electron effective temperature averaged over the thz oscillations @xmath165 , i.e. , @xmath166 . as a result , @xmath167
this implies that even at low temperatures , the heating mechanism can be inefficient if @xmath164 . both dynamic and heating mechanisms might be responsible for the thz detection in the plasmonic resonant detectors utilizing another barrier structure ( for instance , with the lateral schottky barrier @xcite or with electron transport through the gate barrier @xcite ) and another types of the plasma resonant cavity ( with ungated quasi - neutral regions as in high - electron mobility transistors with relatively long ungated source - gate and gate - drain regions ) , as well as in those with another methods of the excitation of plasma oscillations ( utilizing periodic gate structures @xcite ) .
this work was supported by the grant - in - aid for scientific research ( s ) from the japan society for promotion of science , japan .
the work at rpi was partially supported by the office of naval research , usa .
f. teppe , w. knap , d. veksler , m. s. shur , a. p. dmitriev , v. yu . kacharovskii , and s. rumyantsev , appl .
. lett . * 87 * , 052105 ( 2005 ) .
m. hanabe , t. otsuji , t. ishibashi , t. uno , and v. ryzhii , jpn . j. appl .
phys.*44 * , 3842 ( 2005 ) .
d. veksler , f .
teppe , a .
p. dmitriev , v .
kacharovskii , w. knap , and m. s. shur , phys .
b * 73 * , 125328 ( 2006 ) a. el fatimy , f. teppe , n. dyakonova , w. knap , d. seliuta , g. valusis , a. shcherepetov , y. roelens , s. bollaert , a. cappy , and s. rumyantsev , appl
. lett . * 89 * , 131926 ( 2006 ) .
j. torres , p. nouvel , a. akwaoue - ondo , l. chusseau , f. teppe , a. shcherepetov , and s. bollaert , appl .
phys . lett.*89 * , 201101 ( 2006 ) a. satou , n. vagidov , and v. ryzhii , ieice techn . rep .
ed2006 - 197 , p. 77
( 2006 ) a. satou , v. ryzhii , i. khmyrova , m. ryzhii and m. s. shur , j. appl .
* 95 * , 2084 ( 2004 ) .
v. ryzhii , a. satou , w. knap , and m. s. shur , j. appl .
* 99 * ( 2006 ) 084507 .
v. f. gantmakher and y. b. levinson , _ carrier scattering in metals and semoconductors _ ( north - holland , amsterdam , 1987 ) .
a. shik , _ quantum wells _ ( world scientific , singapore , 1997 ) .
v. ryzhii , a. satou , i. khmyrova , m. ryzhii , t. otsuji , v. mitin , and m. s. shur , j. phys .
: conf . ser . *
38 * , 228 ( 2006 ) . | we analyze the operation of a resonant detector of terahertz ( thz ) radiation based on a two - dimensional electron gas ( 2deg ) channel with split gates .
the side gates are used for the excitation of plasma oscillations by incoming thz radiation and control of the resonant plasma frequencies .
the central gate provides the potential barrier separating the source and drain portions of the 2deg channel .
two possible mechanisms of the detection are considered : ( 1 ) modulation of the ac potential drop across the barrier and ( 2 ) heating of the 2deg due to the resonant plasma - assisted absorption of thz radiation followed by an increase in thermionic dc current through the barrier . using the device model
we calculate the frequency and temperature dependences of the detector responsivity associated with both dynamic and heating ( bolometric ) mechanisms .
it is shown that the dynamic mechanisms dominates at elevated temperatures , whereas the heating mechanism provides larger contribution at low temperatures , @xmath0 k. |
unlike what happens in ordinary fluids , vorticity in quantum fluids is constrained to vortex filaments of quantized circulation .
recent experimental , numerical and theoretical studies @xcite have revealed the surprising result that quantum turbulence , despite the discrete nature of the vorticity , obeys the same kolmogorov scaling which is observed for homogeneous isotropic turbulence in ordinary fluids .
more precisely , the superfluid kinetic energy spectrum scales as @xmath2 in the ` hydrodynamical ' range @xmath3 , where @xmath4 is the wavenumber , @xmath5 the size of the system , and @xmath6 the average distance between vortex lines .
( what happens at length scales smaller than @xmath6 is very interesting - issues which are debated are the kelvin wave cascade and the possible existence of an energy flux bottleneck - but is outside the scope of this work . )
evidence for kolmogorov scaling has been found in both bosonic @xmath0he and in fermionic @xmath1he - b , at both high temperatures ( where helium acquires a two - fluid nature due to the presence of the normal fluid ) and low temperatures .
it has also been found , numerically @xcite and experimentally @xcite , that , when averaged over scales larger than @xmath6 , the turbulent velocity components obey the same gaussian statistics of classical turbulence .
the current @xcite interpretation of these results is that quantum turbulence represents the ` skeleton ' of ordinary turbulence .
however , there is also evidence for a very different spectral nature , in which the largest eddies are weak , most of the energy is contained in the intermediate scales , and the large wavenumber range has the @xmath7 dependence of isolated vortex lines .
such features suggest a tangle of randomly oriented vortex lines whose velocity fields tend to cancel each other out
. this second form of turbulence , named ` vinen ' or ` ultraquantum ' turbulence to distinguish it from the previous ` kolmogorov ' or ` quasiclassical ' turbulence @xcite , has been identified both numerically @xcite and experimentally ( in low temperature @xmath0he @xcite and in @xmath1he - b @xcite ) , and also ( numerically ) in high temperatures @xmath0he driven by a heat current @xcite ( thermal counterflow ) .
the natural question is whether either vinen turbulence is some new form of disorder ( if so , why it has not been observed in classical turbulence ? ) or there are physical mechanisms which prevent the development of the classical kolmogorov spectrum and the energy cascade . in this report we shall argue that the latter is the solution of this important puzzle .
in this section we briefly summarize the kolmogorov phenomenology for classical turbulence ( undoubtedly , the content of this section is well known and can be found in every standard text on classical turbulence @xcite ) .
this will make our derivations of the conditions for / against the existence of the kinetic energy cascade and of the associated kolmogorov scaling in quantum turbulence more transparent . in classical homogeneous and isotropic turbulence
( away from the boundaries ) the richardson energy cascade takes place if there exists an interval of length scales ( known as the ` inertial range ' ) such that , at every scale @xmath8 within this range , the dissipation time @xmath9 ( where @xmath10 is the kinematic viscosity ) is much longer than the eddie turnover time @xmath11 : @xmath12 with the velocity at the length scale @xmath8 given by @xmath13 where @xmath14 is the energy dissipation rate and @xmath15 the energy per unit mass ( we assume that the fluid has constant density ) . condition ( [ eq : tau ] ) for the existence of the cascade can be reformulated as @xmath16 , where @xmath17 is the scale - by - scale reynolds number .
inequality ( [ eq : tau ] ) becomes invalid at the kolmogorov length scale , @xmath18 , where , by definition , @xmath19 ( that is , viscous and inertial forces are comparable ) .
the energy spectrum , @xmath20 , is defined by @xmath21 where @xmath22 is the volume . within the inertial range , the kolmogorov scaling of the energy spectrum
is obtained assuming that @xmath20 depends only on @xmath23 and @xmath4 .
simple dimensional analysis then yields the famous result that , within the inertial range @xmath24 , the energy spectrum is @xmath25 where @xmath26 is a dimensionless constant of order unity .
we now turn to quantum turbulence . unlike classical turbulence , where vorticity is continuous and eddies have arbitrary shape and strength , quantum turbulence consists of individual vortex lines .
each vortex line carries one quantum of circulation @xmath27 , where @xmath28 is planck s constant and @xmath29 the mass of the relevant boson ( one atom for @xmath0he , one cooper pair for @xmath1he - b ) . in analogy with classical turbulence , we are interested in the existence of a richardson energy cascade in the hydrodynamic range @xmath30 , which is the regime of vortices interacting with each other ; we are not interested in the @xmath31 regime of isolated vortex lines ( an important but different physical problem with no direct relation to classical fluid dynamics ) .
we , therefore , consider the length scales @xmath8 such that @xmath32 where the mean intervortex distance , @xmath6 can be inferred from the observed vortex line density @xmath33 , defined as the total length of vortex lines per unit volume , as @xmath34 however , in the zero - temperature limit this relation holds only for the smoothed line density ; owing to the presence of high frequency kelvin waves undamped by the mutual friction , in the zero - temperature limit the actual vortex line density exceeds the smoothed line density , see ref .
@xcite for details which will also be discussed below in the penultimate section of this report .
similar to the kolmogorov phenomenology for classical turbulence , the conditions for existence of the richardson cascade and , therefore , the @xmath35 kolmogorov scaling of the superfluid kinetic energy spectrum , can be obtained by comparing the time scale of dissipation with the turnover time of the macroscopic eddies . in the next two sections
we shall make this comparison ( distinguishing between high temperature and low temperature regimes ) for turbulent states which , anomalously , do not follow the kolmogorov scaling . for the sake of brevity , we shall call _ non - cascading _ the turbulence which does not exhibit scale - by - scale energy transfer . thus , for a homogeneous isotropic system , the energy spectrum does not scale as @xmath35 at large wavenumbers @xmath4 ; a noticeable feature of such non - cascading turbulence is that kinetic energy is concentrated at some intermediate wavenumbers , giving the spectrum @xmath20 the shape of a ` bump ' followed by @xmath7 behaviour at large @xmath4 .
we start with quantum turbulence at high temperatures such that @xmath36 for @xmath0he , where @xmath37 is the superfluid transition temperature , or @xmath38 for @xmath1he - b , where @xmath39 is the critical temperature of the superfluid transition . at high temperatures
the dissipation is caused by the mutual friction between the normal fluid and superfluid vortices .
an anomalous , non - cascading superfluid energy spectrum has been predicted @xcite for @xmath0he counterflow and for grid turbulence in @xmath1he - b in the presence of very viscous , stationary normal fluid @xcite .
similar conclusions were reached in refs .
@xcite , where it has been shown that , in the case where the normal fluid is either stationary or non - turbulent , the strong mutual friction , which is dissipative at all length scales , prevents the emergence of the inertial range and the @xmath40 energy spectrum in the superfluid component .
our aim here is to reveal the physical mechanism which prevents the formation of the classical kolmogorov spectrum . to find the timescale associated with the mutual friction , we develop the following simple macroscopic model of thermal counterflow in @xmath0he .
the coarse - grained ( that is , averaged over a scale much larger than the intervortex distance ) superfluid and normal fluid velocities , @xmath41 and @xmath42 , are governed by the equations @xcite @xmath43 where @xmath44 is the pressure , @xmath45 the entropy , @xmath46 the temperature , @xmath47 and @xmath48 the superfluid and normal fluid densities , @xmath49 the total density , @xmath50 the viscosity , and @xmath51 the convective derivative . the mutual friction force per unit volume is @xmath52 where @xmath53 is the temperature - dependent mutual friction coefficient ( for @xmath0he its values , obtained from the counterflow measurements , are tabulated in ref .
the quantum of circulation is @xmath54 .
we consider for simplicity steady one - dimensional flow along the @xmath55-direction of a long channel which is closed at one end and open to the helium bath at the other end . at the closed end ,
an electrical resistor dissipates a given heat flux @xmath56 , which is carried away by the normal fluid at the velocity @xmath57 .
superfluid flows in the opposite direction to maintain the condition of zero mass flux , @xmath58 . in this way a relative motion ( counterflow ) @xmath59 is set up beween the normal fluid and the superfluid , which is proportional to the applied heat flux , @xmath60 .
the importance of this flow configuration can not be understated , as it is used to study the exceptional heat conducting properties of liquid helium as a coolant in engineering applications . provided that @xmath61 exceeds a small critical value ,
the superfluid becomes turbulent , and a tangle of vortex lines fills the channel with vortex line density @xmath33 .
we assume that @xmath61 is not so large that the normal fluid becomes turbulent .
let @xmath62 , @xmath63 and @xmath64 be the values of superfluid velocity , normal fluid velocity and vortex line density , respectively , in the steady - state regime at given temperature @xmath46 and heat flux @xmath65 . to simplify the problem and to highlight the role of the friction
, we neglect the viscous term in eq .
( [ eq : vn ] ) ( which would cause a pressure drop along the channel ) , and obtain the pressure and temperature gradients induced by the quantum turbulence : @xmath66 and @xmath67 eq .
( [ eq : dp ] ) is known as the allen - reekie rule and has been verified in the experiments .
experiments and numerical simulations show that @xmath68 where @xmath69 is a temperature dependent parameter @xcite ; from eq .
( [ eq : dt ] ) , this means that the temperature gradient @xmath70 is proportional to @xmath71 , in agreement with experiments .
the next step is consider the effect of superfluid velocity fluctuations @xmath72 on top of the steady flow @xmath73 . proceeding with a one - dimensional model , we write @xmath74 for the sake of simplicity , and , again , to bring in evidence the role of the mutual friction
, we assume that the velocity of the normal fluid remains constant , and linearize the friction , neglecting fluctuations of @xmath33 , @xmath46 and @xmath44 , so that @xmath75 . after subtracting the steady state
, the superfluid equation reduces to @xmath76 having assumed that the channel is long , we change to the moving reference frame @xmath77 in which the equation for the superfluid velocity fluctuations becomes @xmath78 where the friction timescale is @xmath79 in deriving eq .
( [ eq : vs - tau ] ) we have considered a region of the fluid where the vortex line density , @xmath33 is essentially constant apart from small fluctuations which cause fluctuations of the friction which are much quicker than the evolution of @xmath72 . in this approximation
we can assume that @xmath33 is constant , so the equation for @xmath72 is linear , and , fourier trasforming , we have @xmath80 where @xmath81 is the fourier component of the superfluid velocity @xmath82 at time @xmath83 and wavenumber @xmath84 .
the important assumption is that @xmath85 does not depend on @xmath4 , so that the linear equation for the velocity @xmath72 in physical space becomes a linear equation for the fourier component of the velocity @xmath86 .
there are two conditions for the existence of the cascade .
the first , given above by inequality ( [ eq : rggl ] ) , is that the scale @xmath8 must be larger than the intervortex spacing , that is to say we are in the hydrodynamic regime of many vortices , not in the regime of isolated vortex lines .
the second , as for the classical cascade , see condition ( [ eq : tau ] ) , is that @xmath87 using eq .
( [ eq : tauf ] ) and eq .
( [ eq : ur ] ) , which is applicable in the considered hydrodynamic regime of many vortices , eq . ( [ eq : f ] ) becomes @xmath88 the energy dissipation rate is @xmath89 where we have taken into account that the mean value of velocity fluctuations is zero .
using now eqs .
( [ eq : vs - tau ] ) and ( [ eq : tauf ] ) , the energy dissipation rate can be estimated from eq .
( [ eq : dissipation ] ) as @xmath90 where @xmath91 is the rms of the superfluid velocity fluctuations .
we also make the assumption that the average amplitude of the fluctuations , @xmath92 is only a fraction of the mean superflow , @xmath93 with @xmath94 or less . in the counterflow produced by a strong heat current such that the normal fluid is highly turbulent , the turbulent intencity in the latter was recently measured @xcite by means of the technique using triplet - state he@xmath95 molecular tracers .
the fluctuations in the normal fluid were found to be about 0.25 of the normal fluid s mean velocity .
the fluctuations in the superfluid have not been measured , but the same level of turbulence intencity can be anticipated . however , in the regime where the normal fluid is laminar or stationary , as in the case considered in the current work , velocity fluctuations in the superfluid component should be much smaller . since @xmath96 , combining eqs .
( [ eq : gamma ] ) and ( [ eq : eps - rms ] ) we write criteria ( [ eq : rggl ] ) and ( [ eq : f ] ) in the form @xmath97 where @xmath98 for temperatures between 1.5 and @xmath99 typical of the counterflow , @xmath100 is between 0.1 and 0.75 and @xmath53 varies from 0.074 to 0.5 @xcite , while @xmath69 is of the order of @xmath101 @xcite , so that @xmath102 is of the order unity and it is impossible to satisfy simultaneously both conditions in inequalities ( [ eq : conditions ] ) .
it follows , then , that for the values of parameters typical of the thermal counterflow the kolmogorov scaling of the energy spectrum should not be expected .
( note that this conclusion has been obtained assuming that the normal fluid velocity is constant .
this assumption is violated , so that our model is no longer valid , if the normal fluid itself becomes turbulent , as in the counterflow at a sufficiently large heat current . in this case , owing to the mutual friction between the normal fluid and quantized vortices , the superfluid energy spectrum acquires the @xmath35 kolmogorov scaling . ) a typical superfluid kinetic energy spectrum obtained from our numerical simulation of the counterflow turbulence is shown in figure [ fig : counterflow ] .
( arbitrary units ) vs wave number @xmath4 ( @xmath103 ) in thermal counterflow .
the hydrodynamical range is from @xmath104 to @xmath105 .
notice the lack of energy at the largest scales and the slope , which is rather different from kolmogorov s @xmath35 scaling ( indicated by the dashed line ) . ] the numerical method and procedure are described in our earlier work @xcite .
the parameters are : temperature @xmath106 , counterflow velocity @xmath107 , friction coefficient @xmath108 , vortex line density @xmath109 corresponding to @xmath110 , in good agreement with ref . @xcite . the calculation is performed in a cubic periodic domain of size @xmath111 ; numerical discretization along the vortex lines is @xmath112 .
as seen from figure [ fig : counterflow ] , the energy spectrum has a broad peak at intermediate wave numbers , without much energy at the largest scales ( smallest @xmath4 ) . at large @xmath4
, the spectrum follows the typical @xmath7 scaling of smooth isolated vortex lines , not kolmogorov s @xmath35 scaling .
in fact , using the parameters of the simulation , inequality ( [ eq : conditions ] ) becomes @xmath113 , which can not be satisfied .
since the spectrum of an isolated vortex line scales as @xmath7 , the observation of @xmath114 in a turbulent tangle of vortex lines suggests that far - field effects cancel each other out , in other words that the vortex lines are randomly oriented , and that the only length scale of the turbulence is @xmath6 .
a similar argument based on the comparison between the time scale of friction with the turnover time of eddies was used to justify the absence of the kolmogorov energy spectrum in @xmath1he - b grid turbulence in the presence of the stationary normal fluid @xcite . on the other hand ,
the presence of the @xmath35 spectrum in @xmath0he coflows was clearly demonstrated by the experimental measurements @xcite and various numerical simulations @xcite .
there is strong experimental evidence in both @xmath0he and @xmath1he - b for the existence of kolmogorov energy spectra at temperatures so low that the normal fluid , hence the friction , is negligible .
this evidence is the observed decay law of the vortex line density , @xmath115 , which can be related to an underlying kolmogorov spectrum .
numerical simulations performed using the gross - pitaevskii equation and the vortex filament model confirmed this result by directly measuring the energy spectra .
however , there is also experimental and numerical evidence for a different turbulent regime in which the kolmogorov spectrum is absent and the resulting decay of turbulence is @xmath116 . as for the high temperature regime discussed in the previous section , the existence of this second regime ( called _ vinen _ , or _
ultraquantum _ turbulence in the literature to distinguish it from the _ quasiclassical _ kolmogorov regime ) , presents us with a puzzle .
our aim is to develop a model which would yield conditions necessary for emergence of the quasiclassical quantum turbulence at very low temperatures such that the presemce of the normal fluid can be neglected .
a model developed below is based on the experiments reported in refs .
@xcite in which turbulence was generated by injecting a jet of negative ions ( electrons in a bubble state ) . at low temperatures each injected electron dresses itself into a quantized vortex ring .
the rings then collide and reconnect to produce a vortex tangle which gradually fills the experimental cell .
a regime of quantum turbulence was then identified by the scaling with time of the decay of vortex line density after the ion injection has stopped .
numerical analysis of the decay of quasiclassical and ultraquantum regimes of turbulence can be found in our paper @xcite . in refs . @xcite it was found that the regime of turbulence , generated by injection of the ion jet , depends on a duration , @xmath117 of the ion injection : the ultraquantum regime was generated when the duration of injection was relatively short , and
the quasiclassical regime was observed if the injection time was longer .
our model is outlined in the following three paragraphs .
assume that vortex rings ( whose sizes are narrowly distributed around some mean value ) are injected into the experimental cell filled with @xmath0he at very low temperature .
the rings collide and reconnect , gradually forming the vortex tangle .
part of the energy injected by vortex rings is fed into the kelvin waves and ultimately dissipated by the phonon radiation .
the vortex line density , @xmath33 grows until , at time @xmath118 , called below the `` saturation '' time , the total energy input by injected rings balances the total energy fed into the kelvin waves . at this time
the growth of the line density stops and thereafter the tangle is in the statistically steady state with the time - averaged vortex line density @xmath119 until time @xmath120 when the injection stops and the tangle decays . the quasiclassical regime of quantum turbulence may be generated in the case where there exists a mechanism of the three - dimensional inverse energy transfer from the scales at which the energy is injected ( presumably , the scale of a single vortex ring ) to larger scales which substantially exceed the mean intervortex distance in the created vortex tangle . as we argued in our earlier papers @xcite , such a mechanism can be provided by anisotropy which favours reconnections of loops of the same polarity .
furthemore , in the cited papers we demonstrated by direct numerical simulation a generation of the kolmogorov @xmath35 energy spectrum resulting from the process of rings injection , thus mimicking the experiments @xcite .
the mechanisms of reconnections of injected quantized vortex rings , and of generation of quasiclassical , large - scale velocity fluctuations was further investigated , theoretically and experimentally , in ref .
@xcite ; the authors of cited work favoured the view that the process of rings reconnections , resulting in creation of the coarse - grained velocity field on the quasiclassical scales , is , in fact , the three - dimensional inverse cascade . however , whether this inverse energy transfer is an inverse cascade or not is irrelevant for the purpose of the current study . in the process of rings injection
the kolmogorov spectrum gradually grows through the wavenumber space to smaller wavenumbers eventually filling , at some time @xmath121 , the whole available interval of wavenumbers from @xmath122 to @xmath123 , where @xmath5 is the size of the experimental cell ( or the integral scale of the fully developed quasiclassical turbulence ) .
if the injection of rings stops at time @xmath124 , the following decay of the line density will be that typical of the quasiclassical turbulence and scale with time as @xmath125 . had the injection been stopped some time before @xmath126 , the inverse energy transfer would not have enough time to develop a full kolmogorov spectrum spanning all available lengthscales , in which case the vortex tangle would still be ultraquantum and decay , after the injection has stopped , as @xmath127 .
note that our aim is to find conditions necessary for emergence of the quasiclassical regime which is associated with motion at scales larger than the intervortex distance .
for such a motion the rate at which the energy is fed into the kelvin waves plays a rle of the dissipation rate , and the details of dissipation by phonon emission as well as the possibility of the much debated bottleneck in the energy flux between the macroscopic motion and the kelvin wave cascade are irrelevant .
the energy of an isolated quantized vortex ring of radius @xmath128 is @xcite @xmath129 where the function @xmath130 is @xmath131 @xmath132 is the vortex core radius which is of the order of coherence length .
we assume here that all injected rings are of the same radius , as practically was the case in the experiments @xcite .
( in fact , the sizes of the rings are narrowly distributed around some value of the radius .
this was taken into account in our earlier numerical studies @xcite , but is irrelevant for the purpose of the model considered below . ) assuming that the frequency of rings injection is @xmath133 , the rate of energy input , per unit mass , can be estimated as @xmath134 where @xmath135 is the volume of cubic experimental cell , and @xmath136 is the probability of the vortex ring to propagate through the vortex tangle without collisions with other vortex lines .
this probability , which depends on the current vortex line density , @xmath137 and the size of experimental cell , @xmath5 was investigated in detail in ref .
the probability @xmath44 is close to unity when the tangle is still very dilute but decreases exponentially with both the vortex line density and the size of experimental cell . for the values of parameters typical of the experiment @xcite ( @xmath138 and @xmath139 ) ,
the probability @xmath44 becomes small ( less than 0.25 ) already for @xmath140 , and negligible when the tangle reaches the saturated , statistically steady state with @xmath33 of the order @xmath141 . for the sake of simplicity
we assume below that @xmath142 so that all injected rings contribute to the growth of the line density .
( note that in the conditions of the cited experiment the probability @xmath44 is not unity even at the very early stages of tangle s formation as the rings are injected in a narrow beam so that collisions between rings of slightly different radii are not infrequent .
this can also be seen in our earlier numerical simulations @xcite of the experiment @xcite . ) the saturation time , @xmath118 required for the tangle to reach the statistically steady state will , therefore , be somewhat underestimated . however , as will be seen below , for the parameters of the experiment @xcite a formation of the large scale motion ( and , therefore , of the quasiclassical tangle ) occurs mainly after the statistically steady state has been reached , that is at times @xmath143 .
moreover , it will also be shown below that time @xmath121 required for generation of quasiclassical turbulence is independent of @xmath118 .
we will also assume that at each moment of time the vortex tangle occupies the whole experimental volume .
although there is some evidence that at the early stages of evolution the tangle occupies only part of the experimental volume , see e.g. fig . 1 in ref .
@xcite , the simplifying assumptions made above will suffice for the order - of - magnitude estimates of @xmath121 .
the rate , per unit mass , at which energy is fed into the kelvin waves , and ultimately dissipated , is given by @xcite @xmath144 where @xmath145 is the _ smoothed _ vortex line density @xcite , that is the length of line per unit volume after the excited kelvin waves have been removed , and @xmath146 is a constant whose numerical value will be discussed below . before proceeding with our model , we have to make a rather important remark .
in the zero - temperature quantum turbulence the vortex reconnections generate the kelvin waves which lead to an increase of the _ actual _ ( as opposed to the _ smoothed _ ) vortex line density , @xmath33 , for which the relation ( [ eq : kwflux ] ) still holds in the form @xmath147 , but with @xmath148 .
the product @xmath149 can be interpreted as the effective viscosity , @xmath150 which was thoroughly discussed in refs .
note also that the mean intervortex distance , @xmath6 is linked by the relation @xmath151 to the smoothed line density , not to the actual vortex line density @xmath33 . in the last two decades
significant experimental @xcite and theoretical / numerical @xcite efforts were made to determine the value of the effective kinematic viscosity in the zero - temperature turbulent @xmath0he .
in particular , it was argued @xcite that the value of @xmath150 ( or @xmath146 ) should depend on the spectrum of quantum turbulence and , therefore , its value for the quasiclassical ( kolmogorov ) regime should differ from that for the ultraquantum ( vinen ) turbulence .
however , the recent work @xcite gives convincing arguments that the value of @xmath146 is independent of the energy spectrum and , therefore , should be the same for both the quasiclassical and the ultraquantum regimes .
earlier experiments @xcite and theoretical / numerical results @xcite suggested the value of @xmath146 in the interval 0.060.10 ( although in some cases it is not clear whether @xmath146 or @xmath152 was actually measured ) .
a very recent study @xcite suggests that , in the zero - temperature limit , @xmath153 ( a more precise estimate for @xmath146 is hardly possible at present ) . as will be seen below , for our order - of - magnitude estimates of time @xmath121 required for the formation of the quasiclassical quantum turbulence a precise value of @xmath146 , and even the uncertainty
whether the value used for our estimates is that of @xmath146 or of @xmath152 are unimportant , and , following ref .
@xcite , we will assume that @xmath154 .
using the assumption that @xmath142 , we model the growth of the smoothed vortex line density during the tangle s formation by the equation @xmath155 here we introduced the empirical dimensionless constant @xmath156 to account for several phenomena .
firstly , reconnection of the vortex loop , whose line length is @xmath157 , with another vortex ring or with a vortex line within the tangle does not , in general , increase the tangle s total _ smoothed _ line length by @xmath158 as some of the length may be `` lost '' to the kelvin waves ( for example , a merger of two rings of the same radius @xmath128 results in the ring whose smoothed lengths corresponds to a radius of only about @xmath159 ) .
in fact , @xmath160 should be a function of the vortex line density , and , secondly , should also incorporate a dependence on the probability @xmath44 for the ring to propagate through the tangle without collisions . however , a somewhat simplistic model given by eq .
( [ eq : lgrowthrate ] ) with @xmath161 will suffice : as will be seen below , the formation of the full komogorov spectrum occurs after saturation of the tangle ( at least for the parameters corresponding to experiment @xcite ) .
moreover , neither the saturated line density , @xmath162 , nor the time @xmath121 of the formation of quasiclassical turbulence depend on the parameter @xmath160 . for our estimates of time @xmath121
we will assume that @xmath163 .
neglecting a small vortex line density at the beginning of the rings injection so that @xmath164 , we integrate eq .
( [ eq : lgrowthrate ] ) within the period of tangle saturation , so that @xmath165 until @xmath166 defined such that @xmath167 where @xmath168 and @xmath169 are , respectively , the total injected energy and the total energy dissipated by the kelvin wave cascade ( both quantities are per unit mass ) . the total dissipation , @xmath170 , which grows with time as @xmath171
, cannod exceed the total energy input @xmath172 whose growth with time is linear .
we , therefore , assume that at time @xmath166 following from eqs .
( [ eq : condition - tstar])-([eq : edtot ] ) in the form @xmath173 the growth of @xmath145 stops and hereafter @xmath174 where @xmath175 ( note that these quantities do not depend on the modelling parameter @xmath160 in eq .
( [ eq : lgrowthrate ] ) .
we turn now to the formation of quasiclassical turbulence .
we assume that , resulting from the inverse energy transfer ( cascade ) induced by reconnections between injected vortex loops , all injected energy gradually forms a quasiclassical tangle up to wavenumber @xmath4 which depends on time .
we also assume that during this process the quasiclassical turbulence remains quasistationary .
the energy , per unit mass , of the quasiclassical tangle is then given by @xmath176 where @xmath177 and @xmath178 is the kolmogorov constant .
quasiclassical turbulence is dissipated at the scale of the intervortex distance , @xmath179 by the kelvin wave cascade , so that the dissipation in eq .
( [ eq : ekt ] ) is given by eq .
( [ eq : kwflux ] ) with @xmath180 determined by eq .
( [ eq : lt ] ) . from the equation @xmath181 where @xmath172 grows in time as in eq .
( [ eq : ertot ] ) , we can extract @xmath182 , and the corresponding `` quasiclassical '' lengthscale @xmath183 : @xmath184^{3/2}\ , , \label{eq : lambda}\ ] ] where @xmath185 and @xmath186 is determined by eq .
( [ eq : kwflux ] ) ( for @xmath143 these quantities no longer depend on time and are determined by formulae ( [ eq : stars ] ) ) . the lengthscale @xmath187 should be compared to @xmath6 and the size @xmath5 of the experimental cell .
we expect that if @xmath188 ( one decade of the @xmath4-space ) the kolmogorov spectrum should start becoming visible . the full kolmogorov spectrum , which would yield the observed @xmath189 decay of the vortex line density after the injection of rings has stopped , requires , of course , @xmath190 to grow up to the largest possible value , @xmath191 .
we should note here that eq .
( [ eq : lambda ] ) is not applicable during a very short time period after the start of injection when the vortex configuration still consists mainly of small individual vortex loops rather than of tangled vortex lines , so that eq .
( [ eq : kwflux ] ) is not yet valid .
however , this short time period will not significantly affect our estimates of times @xmath118 and @xmath121 .
we now analyze time @xmath121 predicted by eq .
( [ eq : lambda ] ) in connection with experimental observations .
two experiments in which the quasiclassical regime of quantum turbulence was generated by ion injection were reported in refs .
@xcite and @xcite . in the first of these experiments ,
performed at temperatures ranging from @xmath192 to @xmath193 in a cube - shaped container with sides 4.5 cm ( volume @xmath194 ) , the quasiclassical regime was prominent after injections longer than 30 s , depending on the injected electron current @xmath195 which was in the range from @xmath196 to @xmath197 . in the second experiment @xcite , which was performed in the experimental cell of a more complicated shape ,
the quasiclassical regime has been observed after longer , @xmath198 injection ( a detailed description of the experimental cell can be found in ref .
note that refs .
@xcite do not provide a more detailed analysis of the relation between the duration of injection and the resulting regime of turbulence . although the second of these experiments has been carried out in the truly zero - temperature limit ( at @xmath199 ) , to illustrate our calculation of time @xmath121 we will use the parameters of the first experiment @xcite ( assuming , however , @xmath200 ) whose geometry of the experimental cell was much simpler .
in experiment @xcite , the radius of injected vortex rings was @xmath201 , which corresponds to @xmath202 . to find the frequency of rings injection
, we will follow the assumption made in ref .
@xcite that at low temperatures each injected electron dresses itself in a quantized vortex ring .
then , @xmath203 , where @xmath204 is the elementary charge . for the electron current @xmath205 the frequency @xmath206 , and ,
assuming @xmath207 , for the saturation time we have @xmath208 .
for the parameters of experiment @xcite , the evolution of the quasiclassical lengthscale , @xmath190 calculated using formula ( [ eq : lambda ] ) is shown in figure [ fig : lambda ] .
( in units of @xmath209 ) vs time ( in s ) . * here @xmath121 is the time when @xmath187 has become equal to the experimental cell s size so that the full kolmogorov spectrum is formed if the injection time @xmath124 ; @xmath118 is the saturation time .
dotted part of the curve corresponds to the short time period after the beginning of injection when the vortex configuration still consists of individual loops and @xmath6 can not yet be defined .
see text for details and values of parameters used in calculation . ] shortly after the beginning of injection the ratio @xmath210 reaches the value @xmath211^{3/2}\ , , \label{eq : lambda - ell}\ ] ] which remains constant during the period @xmath212 while the line density still grows and @xmath6 decreases .
( note that solution ( [ eq : lambda ] ) is not valid for a short period of time immediately after beginning of injection when the vortex configuration still consists of individual vortex loops and @xmath6 can not yet be defined ; for this time period solution ( [ eq : lambda ] ) is shown in figure [ fig : lambda ] by the dotted part of the curve . ) the formation of quasiclassical scales occurs mainly after the tangle has been saturated , that is for @xmath143 ( at which point @xmath213 ) when , as shown in figure [ fig : lambda ] , the quasiclassical lengthscale increases rapidly as @xmath214 : @xmath215^{3/2}\ , .
\label{eq : lambda - increase}\ ] ] in eqs .
( [ eq : lambda - ell])-([eq : lambda - increase ] ) , @xmath118 , @xmath209 and @xmath216 are determined by formulae ( [ eq : t * ] ) and ( [ eq : stars ] ) . since
, for the parameters of experiment @xcite , @xmath217 , then the time when @xmath187 becomes equal to the cell s size so that the quasiclassical turbulence is fully developed can be calculated from eq .
( [ eq : lambda ] ) , assuming @xmath218 , @xmath219 , and @xmath220 , as @xmath221\ , .
\label{eq : tk}\ ] ] note that time @xmath121 does not depend on the modelling parameter @xmath160 in eq .
( [ eq : lgrowthrate ] ) . for the parameters of experiment @xcite eq .
( [ eq : tk ] ) yields @xmath222 .
although this time compares very favourably with the results of experiment @xcite which showed that at temperatures @xmath223 the quasiclassical regime is especially prominent after more than 30 s long injection , such a good agreement might be rather coincidental .
the reason is that even at temperature as low as 0.7 k , such that the normal fluid fraction is only @xmath224 and the mutual friction coefficient @xmath53 is of the order of @xmath225 , the vortex rings of radius @xmath226 can not be treated as ballistic . indeed , at this temperature such rings decay on a distance @xmath227 which is much smaller than the size of the experimental cell .
only at temperatures @xmath228 ( @xmath229 ) the range of rings decay exceeds @xmath5 . it can be expected that , in the experiment @xcite , at the initial stage of ion injection the formation of tangle and the generation of motion on quasiclassical ( @xmath230 ) scales occur near the injector .
it can also be expected that , as time progresses , the tangle and the flow on quasiclassical scales spread through the whole experimental cell , as illustrated by the cartoon shown in fig . 1 of ref
. @xcite .
clearly , our simple model , which assumes both ballistic propagation of rings and the spatial uniformity of the tangle , does not capture these phenomena .
however , because our model captures essential features of the generation of motion on quasiclassical scales , we should probably expect an order - of - magnitude agreement between our prediction of time @xmath121 and the experimental observations . a better agreement found in this work
is somewhat surprising . obtained from our numerical simulation @xcite , typical energy spectra for ultraquantum and quasiclassical regimes of quantum turbulence , generated , respectively , by a short ( @xmath231 s ) and a long ( @xmath232 s ) injection of vortex ring , is illustrated in figure [ fig : quasi - ultra ] .
the calculation was performed in the periodic box of size @xmath233 for vortex rings , injected with initial velocities randomly confined within a small , @xmath234 angle , of radii narrowly distributed arond @xmath235 .
the frequency of rings injection was @xmath236 ( see ref .
@xcite for ) for the tangle numerically generated by injection of the vortex rings .
bottom : duration of injection @xmath237 .
top : prolonged injection for @xmath238 .
see text and ref .
@xcite for details and values of parameters used in calculation . ] numerical method and procedure ) . as seen from the bottom curve , in the case of short duration of the injection the spectrum
does indeed show the absence of the @xmath35 scaling for any interval of wave numbers but has instead a broad peak similar to that typical of the counterflow turbulence , cf .
figure [ fig : counterflow ] . on the other hand
, our simulation for a longer duration of injection clearly shows that in this case the kolmogorov spectrum is formed , see the top curve of figure [ fig : quasi - ultra ] .
based on estimates for the effective dissipation rate in quantum turbulence considered at the hydrodynamic scales , that is in the range of scales corresponding to many vortices rather than individual vortex lines , we have obtained conditions necessary for existence of the kinetic energy cascade in the superfluid component and , hence , of the @xmath35 scaling of the superfluid s kinetic energy spectrum . the phenomenological approach which we have developed has enabled us to explain why , at finite temperatures , the kolmogorov energy spectrum can not be observed in @xmath0he counterflows and @xmath1he - b flows in the presence of stationary normal fluid .
we have then extended our approach to consider generation of @xmath0he quantum turbulence at very low temperatures .
we have considered a tangle of quantized vortices generated , as in the experiments @xcite , by a beam of electrons injected , in the bubble state , into helium at temperatures significantly lower than @xmath239 . at the considered scales of many vortices the rle of the effective dissipation rate
is played by the rate at which the energy is fed into the kelvin waves .
having calculated the time required for the inverse energy transfer to form the kolmogorov energy spectrum for all available `` quasiclassical '' wavenumbers ( from that corresponding to the size of experimental volume to the wavenumber corresponding to the intervortex distance @xmath6 ) , we estimated the durations of injection required to generate the quasiclassical regime of quantum turbulence at scales larger than @xmath6 .
the calculated durations are consistent with those observed experimentally @xcite .
we conclude that the regimes of quantum turbulence which have been observed at both high and low temperatures and which are characterized by the spectral nature in which the larges eddies are weak , most of the energy is contained at the intermediate scales , and the fully developed energy cascade is absent , can be understood on the ground of simple large - scale quasiclassical ( `` hydrodynamical '' ) considerations .
data supporting this publication is openly available under an open data commons open database license . 99 barenghi , c. f. , lvov , v. s. , and roche , p .- e .
experimental , numerical , and analytical velocity spectra in turbulent quantum fluid .
sciences usa _ * 111 * ( suppl .
1 ) , 4683 - 4690 ( 2014 ) .
nore , c. , abid , m. & brachet , m. e. kolmogorov turbulence in low temperature superflows .
lett . _ * 78 * , 3896 - 3899 ( 1997 ) .
baggaley , a. w. , laurie , j. , & barenghi , c. f. vortex density fluctuations , energy spectra , and vortical regions in superfluid turbulence .
lett . _ * 109 * , 205304 - 4 ( 2012 ) .
baggaley , a. w. & barenghi , c. f. , quantum turbulent velocity statistics and quasiclassical limit .
e _ * 84 * , 067301 - 4 ( 2011 ) .
la mantia , m. & skrbek , l. quantum , or classical turbulence ?
_ europhys .
lett . _ * 105 * , 46002 - 6 ( 2014 ) .
barenghi , c. f. , skrbek , l. & sreenivasan , k. r. introduction to quantum turbulence .
* 111 * ( suppl .
1 ) 4647 - 4652 ( 2014 ) .
volovik , g. e. classical and quantum regimes of superfluid turbulence .
_ jetp lett . _ * 78 * , 533 - 537 ( 2003 ) .
baggaley , a. w. , barenghi , c. f. & sergeev , y. a. quasiclassical and ultraquantum decay of superfluid turbulence .
b _ * 85 * , 060501(r)-4 ( 2012 ) .
walmsley , p. m. & golov , a. i. quantum and quasiclassical types of superfluid turbulence .
lett . _ * 100 * , 245301 - 4 ( 2008 ) .
zmeev , d. e. _ et al .
_ dissipation of quasiclassical turbulence in superfluid @xmath0he . _
lett . _ * 115 * , 155303 - 5 ( 2015 ) .
bradley , d. i. _ et al . _ decay of pure quantum turbulence in superfluid @xmath1he - b . _
lett . _ * 96 * , 035301 - 4 ( 2006 ) .
baggaley , a. w. , sherwin , l. k. , barenghi , c. f. & sergeev , y. a. thermally and mechanically driven quantum turbulence in helium ii . _
b _ * 86 * , 104501 - 8 ( 2012 ) . frisch , u. _ turbulence : the legacy of a. n. kolmogorov_. ( cambridge university press , cambridge , england , 1995 ) .
vinen , w. f. & niemela , j. j. quantum turbulence . _ j. low temp .
phys . _ * 128 * , 167 - 231 ( 2002 ) .
vinen , w. f. theory of quantum grid turbulence in superfluid @xmath1he - b .
b _ * 71 * , 024513 - 9 ( 2005 ) .
lvov , v. s. , nazarenko , s. v. & volovik , g. e. energy spectra of developed superfluid turbulence .
_ jetp lett . _ * 80 * , 546 - 550 ( 2004 ) .
lvov , v. s. , nazarenko , s. v. & skrbek , l. energy spectra of developed turbulence in helium superfluids . _ j. low temp .
phys . _ * 145 * , 125 - 142 ( 2006 ) .
barenghi , c. f. , donnelly , r. j. & vinen , w. f. friction on quantized vortices in helium ii : a review . _ j. low temp .
phys . _ * 52 * , 189 - 247 ( 1983 ) .
roche , p .- e . , barenghi , c. f. & lvque , e. quantum turbulence at finite temperature : the two - fluids cascade .
_ europhys .
lett . _ * 87 * 54006 - 6 ( 2009 ) .
donnelly , r. j. & barenghi , c. f. the observed properties of liquid helium at the saturated vapor pressure .
_ j. phys .
chem . ref .
data _ * 27 * , 1217 - 1274 ( 1998 ) .
adachi , h. , fujiyama , s. & tsubota , m. steady - state counterflow turbulence : simulation of vortex filaments using the full biot - savart law .
b _ * 81 * , 104511 - 7 ( 2010 ) .
marakov , a. _ et al .
_ visualization of the normal - fluid turbulence in counterflowing superfluid @xmath0he .
b _ * 91 * , 094503 - 5 ( 2015 ) .
sherwin - robson , l. k. , barenghi , c. f. & baggaley , a. w. local and nonlocal dynamics in superfluid turbulence . _ phys .
b _ * 91 * , 104517 - 8 ( 2015 ) .
salort , j. _ et al .
_ turbulent velocity spectra in superfluid flows .
fluids _ * 22 * , 125102 - 9 ( 2010 ) .
dur , d. , baudet , c. , moro , j. p. , roche , p .- e . & diribarne , p. hot - wire anemometry for superfluid turbulent coflows .
instr . _ * 86 * , 025007 - 7 ( 2015 ) .
salort , j. , chabaud , b. , lvque , e. & roche , p .- e . energy cascade and the four - fifths law in superfluid turbulence .
_ europhys .
lett . _ * 97 * , 34006 - 6 ( 2012 ) .
baggaley , a. w. , barenghi , c. f. & sergeev , y. a. three - dimensional inverse energy transfer induced by vortex reconnections .
* 89 * , 013002 - 5 ( 2014 ) .
walmsley , p. m. , tompsett , p. a. , zmeev , d. e. & golov , a. i. reconnections of quantized vortex rings in superfluid @xmath0he at very low temperatures .
lett . _ * 113 * , 125302 - 5 ( 2014 ) .
donnelly , r. j. _ quantised vortices in helium ii _ , ( cambridge university press , cambridge , england , 1995 ) .
laurie , j. & baggaley , a. w. a note on the propagation of quantized vortex rings through a quantum turbulence tangle : energy transport or energy dissipation ? _ j. low temp .
phys . _ * 180 * , 95 - 108 ( 2015 ) .
walmsley , p. m. , golov , a. i. , hall , h. e. , levchenko , a. a. & vinen , w. f. dissipation of quantum turbulence in the zero temperature limit .
lett . _ * 99 * , 265302 - 4 , ( 2007 ) .
stalp , s. r. , skrbek , l. & donnelly , r. j. decay of grid turbulence in a finite channel .
lett . _ * 82 * , 4831 - 4834 ( 1999 ) .
tsubota , m. , araki , t. & nemirovskii , s. k. dynamics of vortex tangle without mutual friction in superfluid @xmath0he .
b _ * 62 * , 11751 - 11762 ( 2000 ) .
kondaurova , l. , lvov , v. , pomyalov , a. & procaccia , i. kelvin waves and decay of quantum superfluid turbulence .
b _ * 90 * , 094501 - 10 ( 2014 ) .
zmeev , d. e. a method for driving an oscillator at a quasi - uniform velocity . _ j. low temp .
phys . _ * 175 * , 480 - 485 ( 2014 ) .
we acknowledge the support by the engineering and physical sciences research council ( united kingdom ) through grant no .
ep / i019413/1 . | experiments and numerical simulations of turbulent @xmath0he and @xmath1he - b have established that , at hydrodynamic length scales larger than the average distance between quantum vortices , the energy spectrum obeys the same 5/3 kolmogorov law which is observed in the homogeneous isotropic turbulence of ordinary fluids .
the importance of the 5/3 law is that it points to the existence of a richardson energy cascade from large eddies to small eddies .
however , there is also evidence of quantum turbulent regimes without kolmogorov scaling .
this raises the important questions of why , in such regimes , the kolmogorov spectrum fails to form , what is the physical nature of turbulence without energy cascade , and whether hydrodynamical models can account for the unusual behaviour of turbulent superfluid helium . in this work
we describe simple physical mechanisms which prevent the formation of kolmogorov scaling in the thermal counterflow , and analyze the conditions necessary for emergence of quasiclassical regime in quantum turbulence generated by injection of vortex rings at low temperatures .
our models justify the hydrodynamical description of quantum turbulence and shed light into an unexpected regime of vortex dynamics . |
recent significant advances in nano - technology make it possible to investigate molecular dynamics and structures at single - molecule level .
measurement of fluorescence resonant energy transfer ( fret ) between a couple of dye molecules that are attached to complementary sites of a biomolecule like dna or protein is particularly useful , because the sharply distance - dependent dipole - dipole interaction between the dye pair can serve as a spectroscopic ruler for the biomolecule @xcite .
fret means a non - radiative quantum energy transfer from a donor that is a dye which initially absorbs light , to an acceptor which is another dye .
the fret can be considered in the framework of the theory of frster @xcite .
an input laser light excites the donor , whose one of decay channels is to migrate its excitation energy to the acceptor via dipole - dipole interaction .
the energy transfer typically finishes within nanoseconds .
the requirements for fret to occur efficiently are , at least , one of the chromophores should have a sufficient quantum yield and the donor fluorescence spectrum must overlap the acceptor absorption spectrum .
photon - photon correlations associated with the chromophores contain information on the conformational distance of the biomolecule .
such information is usually washed out in traditional ensemble measurement , but is readily available in single - molecular measurement .
recently , using a hanbury - brown
twiss @xcite time - interval apparatus , berglund et al .
@xcite have measured photon intensity correlations for individual donor - acceptor pairs on dna . to interpret the experimental data , they proposed a dual fret model .
a continuous model emphasizing overall conformational change of the biomolecule has also been studied by several authors @xcite .
however , it is important to note that there exists non - trivial interplay between biomolecular diffusion , which under physiological conditions can change drastically the molecular conformation over nanoseconds , and the quantum optical processes of a pair of fret coupled dyes , which are also of the time scale of nanoseconds . in this paper
, we show that the laser - induced fret dynamics can be modelled by the stochastic ornstein uhlenbeck ( ou ) process @xcite .
the underlying stochastic process is a gaussian random process @xcite with finite correlation time .
the density matrix equations acquire the character of stochastic differential equations which can be solved using well - established methods .
we shall demonstrate that the ou process is a good approximation to the fret dynamics as measured on biomolecules .
this paper is organized as follows . in section 2
we present our theoretical model of fret .
the idea of modelling fret dynamics as an ou stochastic process is further elucidated in the context of biomolecular dynamics in solution . in section 3 and 4
we present the results of monte - carlo simulations and compare them with the experimental work .
in order to calculate the photon - photon correlation functions , we adopt a master equation approach . to derive the master equation for the system with two molecules coupled by fret , we assume that the coherence time is much shorter than the time scale of experiments .
we further note that a fret coupled system is not a cascaded quantum system @xcite , i.e. , to obtain an equation for only donor variables , after tracing over the acceptor variables from the general master equation , is impossible .
the master equation for the density operator @xmath0 , for the donor - acceptor pair couples four possible states @xmath1 here @xmath2 ( @xmath3 ) stands for the ground and excited states of the donor ( acceptor ) @xcite : @xmath4 where @xmath5 are rate coefficients for the dominant processes : spontaneous emissions of the donor and acceptor are described by the quantum jump operators @xmath6 and @xmath7 ; laser excitations of the donor and acceptor by @xmath8 and @xmath9 ; and transfer of energy from the donor to the acceptor by @xmath10 . here
@xmath11 ( @xmath12 ) is the lowering ( raising ) operator between the excited and the ground states for the @xmath13th dye molecule ( @xmath14 for the donor and 2 for the acceptor ) . the master equation ( [ ariopeqn ] )
ignores all coherence effects as they play no important role in fret .
the conditional photon count probability of the donor - acceptor system within a time delay @xmath15 can be represented as a normally ordered correlation function @xcite @xmath16 here @xmath17 is the stationary operator solution of eq .
( [ ariopeqn ] ) , and @xmath18 is the evolution operator of the whole system satisfying @xmath19 .
fret measurements provide us information about dipole - dipole coupling , which varies as @xmath20 @xcite .
the fret rate coefficient is given by @xmath21 where @xmath22 is the frster radius , @xmath23 is the distance between two chromophores .
considering intra - chain diffusion of a biomolecule , e.g. , a protein molecule that carries the donor and the acceptor at a couple of complementary sites , the displacement of the donor - acceptor distance from its fixed ( or initial ) value , @xmath24 , can be modelled as a brownian motion in a harmonic potential .
since a biomolecule which consists of a large number of atoms and molecules , we essentially deal with the over - damped regime , where the brownian motion is described by the langevin equation @xcite @xmath25 where @xmath26 with @xmath27 being the frequency of harmonic oscillator , and @xmath28 the friction coefficient .
the random force @xmath29 is a pure gaussian characterized by @xmath30 and @xmath31 .
the solution of the langevin equation , @xmath32 , resembles a white noise : its equilibrium distribution is @xmath33 and the correlation is @xmath34 now turn to the fret rate eq .
( [ arifretrate ] ) , which can be rewritten as @xmath35 ^ 6 $ ] with @xmath36 ^ 6 $ ] corresponding to the fret rate at some fixed interdye distance .
if the displacement @xmath37 is small as compared to @xmath38 , then we have @xmath39 this is a possible interplay between the diffusion process and the fret dynamics , although neither they are completely uncorrelated nor fully correlated .
it turns out that a variance ( i.e. , ` noise ' ) of the fret rate has the statistical signature of a white noise . a sign such as
noise of the noise usually leads to colored noise in the ou stochastic dynamics .
eq.([arirelation ] ) is a linearized approximate relation .
although , some dynamic details can not be seen for large variations , analytically , the exact numerical simulation may be done .
this rather involved and so we have tried to produce reasonable results by retaining the leading term .
this captures the physics reasonably well .
note that the ou stochastic process is stationary .
thus eq.([ariopeqn ] ) becomes a stochastic differential equation as @xmath40 .
we assume that @xmath41 is a colored noise , which is described by the langevin equation @xcite @xmath42 where time averages of white noise should be @xmath43 and @xmath44 . as well known the eq .
( [ arieqncol ] ) , yields the steady state correlation function @xmath45 with @xmath46 and @xmath47 denotes the stochastic average over the initial conditions @xcite .
a parameter @xmath48 might be proportional to the diffusion coefficient @xmath49 and @xmath50 is the same in eq.([ariexpcorr0 ] ) and eq.([ariexpcorr ] ) .
the stochastic differential equation eq.([ariopeqn ] ) will be solved using monte - carlo numerical simulations . a box - mueller algorithm and the euler - maruyama method
have been used to realize the colored noise . moreover , by virtue of an integral algorithm developed in @xcite , we have verified that the monte - carlo generated correlation fits perfectly to its analytical expression eq .
( [ ariexpcorr ] ) . to achieve this , a stochastic averaging over as large as @xmath51 realizations is essential @xcite .
it must be borne in mind that the fret rate is always positive which is done by keeping a background constant value @xmath52 . however , some large negative random numbers have to be omitted . to prove that these omitted random numbers do not play important role , we have also calculated the correlation for @xmath53 which decreases exponentially as given in eq.([ariexpcorr ] ) .
namely , we assume that @xmath54 where the fret rate is finite .
@xmath55 ( @xmath56 ) corresponds to minimum ( maximum ) inter - dye distance due to continues intrachain diffusion of the protein molecule .
the quantities are determined by contour length and bending rigidity of the protein . in our case
, we have assumed that the fret does not occur between distant dyes , so that @xmath57 .
we take also @xmath58 as a maximum value of the generated random numbers .
using monte - carlo simulations we have calculated correlation functions as defined in eq .
( [ aricorr ] ) . in fig . 1 we present the results for the normalized correlations @xmath59 defined by @xmath60
since only two parameters , @xmath48 and @xmath50 , are needed to completely specify an ou process , their determination would be a desired contact between theory and experiment . as we see in fig . 1 , the normalized correlation functions are very similar to those observed in the experiment@xcite .
for the data shown in fig . 1 the correlation time is taken to be @xmath61 . in this case we take the parameter @xmath52 to be @xmath62 in order to ensure that the average fret coefficient to be around @xmath63 .
the fret coefficients fluctuate between @xmath57 and @xmath64 .
the estimated frster radius , for instance , for the tmr - cy5 dye pair that is frequently used in biomolecular measurement , is about @xmath65 @xcite .
given the calculated average fret coefficient of @xmath63 , the average distance would be approximately @xmath66 .
intensity autocorrelations show the typical quantum feature of antibunching that is characteristic of emissions from individual dye molecules . following the initial photon antibunching ,
photon bunching appears in the acceptor autocorrelation function .
a sufficient stochastic deviation from its equilibrium distribution of the fret rate is the hallmark of the generation of photon bunching , which is a tendency for clustered emissions . for large @xmath15
the normalized correlation functions go to unity indicating uncorrelated emissions .
the appearance of the photon bunching in the donor autocorrelation was first predicted theoretically by haas and steinberg @xcite .
a pronounced antibunching associated with a photon blockade effect and a photon bunching in the cross talk of the acceptor - donor pair , for short time , has also been discussed in @xcite .
we also notice in the experimental results that over longer times , photons emitted by the donor are becoming correlated with photons emitted by the acceptor and vice versa .
it is also worth noting that because of off - resonant excitation of the acceptor , the corresponding rate for the acceptor is as small as @xmath67 with @xmath68 .
otherwise , the acceptor would have already been excited by the laser , and fret may not occur . for the same reason the laser excitation rate should not be too large .
the bunching signal can be described approximately by an exponential @xmath69 , with the fitting parameter @xmath70 proportional to @xmath50 .
the tail of exponential decay , especially in the acceptor autocorrelation function is determined by the correlation time @xmath71 of the ou process .
it is obvious that a larger value of the correlation time results in a longer tail . while being supported by the simulation results ( fig . 1 )
, this point can be made clearer by assuming that the intra - chain diffusion of the biomolecule that carries the dyes , and thereby fluctuation of the donor - acceptor distance is much slower than the quantum optical process of the donor and acceptor .
the population on the excited states then adiabatically follows the slowly varying @xmath53 , so that the intensity of the acceptor can be approximated as@xcite @xmath72 according to this adiabatic approximation in longer time delay @xmath15 , the intensity correlation is an exponential @xcite @xmath73 where we have assumed that the dual fret rates result in a high value and a low value of the emission intensity , @xmath74 and @xmath75 , respectively . using @xmath64 and @xmath57 obtained in the calculation for fig . 1
, we find that @xmath76 and @xmath77 from eq.([ariintensity ] ) .
the correlation function under the adiabatic approximation given by eq.([ariadiabat ] ) has been also plotted , see the dotted curve in fig . 1 .
we have shown how the fret process as measured for biomolecules in solution can be modelled using the ornstein - uhlenbeck stochastic theory . this theory predicts the fluorescence intensity correlations from the fret coupled dye pair , which are very similar to those observed in recent experiments .
an analytic study based on the local linearization procedure also shows that it is consistent with existing theory for microscopic dynamics of the biomolecule that carries the fret coupled dye pairs .
the second - order intensity correlation functions for a fret coupled dye pair , are largely determined by only a few statistical parameters of the fret dynamics .
it is found that the stochastic ou description helps elucidate the underlying mechanism for the experimentally observed fluorescence correlations that typically exhibit exponential decay over nanosecond timescale
. * acknowledgement .
* this work was supported by onr grant n00014 - 03 - 1 - 0639 and n00014 - 02 - 1 - 0741 ; welch foundation grant a-1261 ; afrl grant f30602 - 01 - 1 - 0594 and titf .
one of us , ( g.o.a . ) also acknowledges the support of the humboldt foundation fellowship .
agarwal , _ quantum optics : quantum statistical theories of spontaneous emission and their relation to other approaches _ ( springer - verlag , berlin , 1974 ) .
coffey , yu.p .
kalmykov and j.t .
waldron , _ the langevin equation with application in physics , chemistry and electrical engineering _ ( world scientific , singapore , 1996 ) . ) are chosen as @xmath78 .
the noise parameters are taken as @xmath61 and @xmath79 .
the number of numerical realizations and a time step are taken to be @xmath80 and @xmath81 , respectively.,title="fig:",width=257 ] [ arifig1 ] | motivated by recent experiments on photon statistics from individual dye pairs planted on biomolecules and coupled by fluorescence resonance energy transfer ( fret ) , we show here that the fret dynamics can be modelled by gaussian random processes with colored noise . using monte - carlo numerical simulations ,
the photon intensity correlations from the fret pairs are calculated , and are turned out to be very close to those observed in experiment .
the proposed stochastic description of fret is consistent with existing theories for microscopic dynamics of the biomolecule that carries the fret coupled dye pairs .
keywords : flourescence resonance energy transfer , single molecule , ornstein - uhlenbeck stochastic process , nanosecond dynamics , protein folding |
microwave emission produced during solar flares is known to contain highly important information about fast electron acceleration and transport , coronal magnetic field and thermal plasma @xcite . however
, this potential of the microwave emission has not yet been converted to a routine diagnostics because of two main reasons .
the first of them is absence of well - calibrated radio imaging spectroscopy data with needed spatial , temporal , and spectral resolutions @xcite .
currently , the situation has started to change as a number of solar radio instruments ( e.g. , owens valley solar array , ovsa , and siberian solar radio telescope , ssrt ) experience significant upgrade ( upgraded ssrt , ussrt ) and expansion ( expanded ovsa , eovsa ) , while even more powerful solar radio facilities are planned to be built in near future , and a general - purpose radio facility , evla , will soon be operational in the solar observing mode .
this implies that the required radio data will soon become available .
however , even with them there is a second reason , which is apparent lack of realistic 3d modeling of the microwave emission from flares .
this modeling is highly important because the gyrosynchrotron ( gs ) emission depends in a complicated nonlinear way on many involved parameters and source geometry including spatial inhomogeneity and angular anisotropy .
the realistic modeling has to establish a clear quantitative picture and solid detailed understanding of how the involved physics ( i.e. , source properties and parameter regimes ) affects the emission produced , e.g. , what changes in the emission can be expected from variation of a given parameter .
available 3d models of the gs emission @xcite have established valuable examples of the flaring microwave emission ; however , they are neither numerous nor comprehensive and rely on isotropic pitch - angle distribution and uniform spatial distribution in most of the cases .
the observations , however , suggest that fast electrons in flares have often anisotropic and/or inhomogeneous distributions ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
analysis of the pitch - angle anisotropy effect has yielded controversial conclusions : although the anisotropy was found to have huge effect on the gs emission from a uniform ( or spatially resolved ) source @xcite , this effect can become much weaker when averaging over significant volume with a nonuniform loop magnetic field comes into play @xcite .
this calls for a more systematic study of the gs emission from anisotropic electron distributions in 3d magnetic loops .
the situation is additionally complicated by the fact that the mentioned pitch - angle anisotropy implies , as a side effect , a spatial inhomogeneity of the electron distribution due to the fast electron accumulation at the loop - top @xcite . as a first step towards addressing the whole problem of gs modeling in realistic 3d coronal magnetic loops ,
this paper presents a convenient modeling tool , `` 3d gs simulator '' , which gives any user an ability to build an analytical dipole magnetic loop , select a desired viewing angle , populate the loop with thermal plasma and nonthermal electrons , and compute microwave images and spatially resolved spectra . the vacuum - like propagation from the source to the observer
is adopted in the tool , so any propagation effect on the radiation in the coronal plasma including polarization modification , which might be present in the reality , is ignored . a reasonably quick computation of a given model is made possible by the use of recently developed fast gs codes @xcite that proved to be highly accurate for both isotropic and anisotropic electron distributions .
we intentionally consider here analytical models for the magnetic field and electron distribution to fully control the input and reliably interpret the outcome ; numerical input in form of corresponding datacubes is under development and will be presented soon elsewhere .
this paper considers the microwave ( gs + free - free ) emission produced by mildly anisotropic electron distributions from a symmetric dipole magnetic loop viewed at different angles and for different parameter combinations . we find that even in such non - extreme conditions the anisotropy makes noticeable fingerprints on the emission properties , which are discussed in detail below .
as has been said , the diversity of the microwave flaring emission is huge , so neither a paper nor a paper series can , perhaps , offer a truly comprehensive table of options fully covering all relevant parameter combinations .
thus , entirely different approach to address the whole problem of the gs modeling is called for : we need a simulation tool capable to smoothly change all the involved source parameters and quickly compute and return the datacubes describing the microwave emission produced . here , we present such a tool and give an example of using it for the microwave emission simulation .
the simulations were performed using the interactive idl tool ` gs simulator ` .
this tool allows one to change the shape and orientation of the flaring loop , choose the parameters of the magnetic field , thermal plasma , and energetic electrons , and calculate the brightness maps of gs emission . in the model
used , the magnetic field is produced by a dipole located below the solar surface .
the loop is formed by a set of field lines such that at the loop top it has a circular cross section with a given radius ; near the footpoints , the loop becomes narrower and non - circular due to conservation of the magnetic flux .
as shown in fig .
[ fig01 ] , which pictures an actual implementation of such geometry in the ` gs simulator ` tool , the user is allowed to freely rotate the dipole loop model in any direction , so that to obtain an arbitrary line of sight relative to the dipole s central plane . the magnetic model [ line - through]*definitions and * geometry and adjustable parameters are described in detail in appendix .
the source code and documentation of the ` gs simulator ` tool are provided as an online supplement .
\(a ) ( b ) to keep the range of options manageable within one paper , we restrict in the present study the flexibility of the model parameters in several ways .
we consider only a simple case of symmetric dipole magnetic loops , i.e. [ line - through]*@xmath1 * with the dipole perpendicular to a local vertical . the adopted geometry ( visualized via a few reference magnetic field lines demarcating the surface of the magnetic loop ) for two different loop orientations is shown in fig .
[ fig01 ] .
dependence of the magnetic field strength at the loop axis on the distance from the loop top along the field line is shown in fig .
[ fig02]a .
we assume that the density and temperature of the thermal plasma component within the loop are constant ( since the plasma in flaring loops is often heated up to the temperatures of @xmath2 k , the corresponding barometric scale heights far exceed the loop heights , so the density variations with height can be neglected ) .
parameters of the energetic electrons can either be constant or vary with the distance from the loop top ( see below ) . by using the above assumptions
, we construct a 3d model of the flaring loop ( which is observed at a given direction ) with all the source parameters depending on the cartesian coordinates @xmath3 , where @xmath4 and @xmath5 are the coordinates in the image plane , and @xmath6 is the coordinate along the line of sight . the source volume is divided onto a number of volume elements ( voxels ) ; each voxel is considered to be quasi - homogeneous .
the radio brightness map ( or the observed emission intensity as a function of 2d coordinates @xmath4 and @xmath5 ) at a given frequency @xmath7 is calculated by numerical integration of the radiation transfer equation @xmath8 along all selected lines of sight . in eq .
( [ rt ] ) , @xmath9 and @xmath10 are the spectral intensities of the left- and right - polarized emission components , respectively , @xmath11 and @xmath12 are the corresponding gs emissivities , and @xmath13 and @xmath14 are the absorption coefficients .
we use the `` weak coupling '' model , i.e. , the left- and right - polarized emission components propagate independently .
left - polarized emission corresponds to either ordinary or extraordinary magnetoionic mode , depending on the magnetic field direction ; respectively , right - polarized emission corresponds to the opposite mode .
the plasma emissivities @xmath15 and absorption coefficients @xmath16 for the ordinary and extraordinary modes accounting for both gs and free - free contributions at each voxel are calculated using fast gs codes developed by @xcite . outside the flaring loop , the emission propagates like in a vacuum .
the loop orientation is described in general by three euler angles ; however , since rotation of the loop around @xmath6 axis results simply in the same rotation of the brightness maps , variation of two angles only is considered .
we adopt the loop to be located at the solar equator ; in this case , the loop orientation is characterized by the angle @xmath17 between the magnetic dipole and the equatorial plane and by the longitude @xmath18 .
the energetic electrons are described by the distribution function @xmath19 in a factorized form : @xmath20 , where @xmath21 is the electron kinetic energy , @xmath22 , and @xmath23 is the electron pitch - angle ( the angle between the particle velocity and the local magnetic field vectors ) .
the tool allows using a number of different model electron distribution functions , including thermal , power - law , broken power - law , a power - law matched to a maxwellian core ( so - called , thermal / nontermal , tnt , distribution ) , kappa distribution , etc . in particular , use of tnt distribution allows one to consistently take into account both gs ( nonthermal ) and gyroresonant ( gr , thermal ) [ line - through]*components * contributions to the emission and absorption . for our modeling , however , it has a disadvantage that two distributions with different anisotropies must be matched , which eventually increases the number of model parameters needed to be varied in the modeling .
for this reason , we assume that the electrons have a power - law energy spectrum @xmath24 in the energy range @xmath25 and so neglect the gr contribution entirely .
however , we make an assessment of the gr [ line - through]*effect * contribution by considering the tnt distribution in the on - line supplement album .
although the effect of gr absorption is modest for our adopted source model and parameters , it does have [ line - through]*some * a noticeable imprint on the spectrum and polarization ( sf , * ? ? ?
* ) from the lower parts of the loop legs , see figs .
29 - 30 of the on - line album
. the pitch - angle distribution can be either isotropic or a loss - cone described by the model function @xmath26 , & \textrm{for}~|\mu|\ge\mu_{\mathrm{c } } , \end{array}\right.\ ] ] where @xmath27 , @xmath28 is the loss - cone boundary , and the parameter @xmath29 determines the sharpness of this boundary .
the loss - cone boundary is adopted to exactly follow the transverse adiabatic invariant @xmath30 where @xmath31 and @xmath32 are the magnetic fields at a given point and at the loop footpoint , respectively .
dependence of the loss - cone boundary on the coordinate along the loop is shown in fig .
[ fig02]b ; this parameter equals @xmath33 at the footpoint and decreases with height , so that the distribution becomes closer to the isotropic one .
since within the adopted model the anisotropy is only strong at and around the footpoints , while becomes much weaker across most of the magnetic loop , we call this `` a moderate anisotropy ''
. we consider both the homogeneous spatial distribution of the energetic electrons ( when their number density @xmath34 is constant ) and the case when energetic electrons are accumulated at the loop top @xcite .
the inhomogeneous distribution of the particles along the loop is described by the following model function : @xmath35,\ ] ] where @xmath36 is the magnetic latitude ( or the angle between the dipole axis and the vector drawn from the dipole center to a given point ) and the parameter @xmath37 determines the inhomogeneity degree ( @xmath38 corresponds to the homogeneous case ) .
density profiles of the energetic electrons along the loop for different values of @xmath37 are shown in fig .
[ fig02]c . in the simulations , we use the following parameters of the flaring loop : height @xmath39 km from the base of the corona , dipole depth below the base of the corona @xmath40 km ( so that the distance between the footpoints @xmath41 km ) , radius at the top @xmath42 km , and magnetic field at the top @xmath43 g ( that results in the footpoint magnetic field of @xmath44 g ) . figures [ fig01]-[fig02 ] correspond to these parameters .
two loop orientations are considered in the paper in some detail : @xmath45 , @xmath46 ( a loop near the center of the solar disk , fig .
[ fig01]a ) and @xmath45 , @xmath47 ( a loop at the limb , fig .
[ fig01]b ) ; many more examples are given in the on - line album .
the thermal plasma density and temperature are @xmath48 @xmath49 and @xmath50 k , respectively .
the energetic electrons have the power - law index @xmath51 , cutoff energies @xmath52 kev and @xmath53 mev , and the loss - cone boundary width @xmath54 .
thus , in each loop orientation , the variable parameters are : type of the pitch - angle distribution ( isotropic or loss - cone ) , the number density of the energetic electrons @xmath34 , and the inhomogeneity parameter @xmath37 .
firstly , we consider effect of electron anisotropy on the gs emission .
figure [ fig03 ] shows the brightness maps at four frequencies for the loop located near the center of the solar disk .
the displayed stokes parameters are : @xmath55 and @xmath56 .
the isotropic ( top row ) and loss - cone ( bottom row ) pitch - angle distributions are considered . in both cases ,
the number density of the energetic electrons @xmath34 is constant and equals @xmath57 @xmath49 .
figure [ fig04 ] shows the spatially resolved emission spectra for the footpoint and loop top sources ( obtained by summation over the circled pixels ) as well as the total emission ( obtained by summation over all pixels ) .
the figure also shows the degree of polarization @xmath58 and spectral index @xmath59 which are defined as @xmath60 from the spatially resolved spectra , one can see that the influence of the electron anisotropy is most important at the footpoints . in the optically thin frequency range ,
the emission from the loss - cone distribution is lower than that from the isotropic electrons by a factor of about 2 - 6 .
this reflects the fact that the gs radiation is emitted mainly in the direction of the electron velocity .
near the footpoints , the electrons with the loss - cone distribution are concentrated around the pitch - angle of @xmath33 whereas the magnetic field ( in the adopted geometry ) is nearly parallel to the line of sight .
this is why the electrons with the loss - cone distribution produce only a relatively weak radiation flux towards the observer . in the optically thick frequency range ,
the emission intensities from the isotropic and anisotropic distributions are almost the same , although the differences in the degree of polarization and spectral index can be visible . near the loop top , the loss - cone boundary @xmath28 falls to about @xmath61 and thus the loss - cone distribution does not noticeably differ from the isotropic one ; as a result , these distributions produce almost identical emission . a complementary way of thinking of the emission is via the radio images at various frequencies ( fig . [ fig03 ] ) .
at the low frequencies ( 3.75 ghz ) , the whole loop is seen on the map ( although the footpoints are brighter than the top ) ; the images for the isotropic and anisotropic distributions are very similar . at the higher frequencies , the emission is strongly concentrated at the footpoints . however , since the footpoint emission from the electrons with the loss - cone distribution is weaker than that from the isotropic distribution , the difference between the footpoints and the loop top is smaller as well .
therefore the emission from the anisotropic electrons is more evenly distributed along the loop .
figures [ fig05]-[fig06 ] show the brightness maps and emission spectra for the loop located near the limb ( all other parameters are the same as in figures [ fig03]-[fig04 ] ) .
one can see that now the electrons with the loss - cone distribution produce stronger footpoint emission in the optically thin frequency range than the isotropic distribution does ( due to the same reasons as discussed above ) . in the optically thick frequency range ,
the considered characteristics of the emission are very similar for both isotropic and anisotropic cases except the effects of the individual cyclotron harmonics that are much more pronounced for the anisotropic distribution , which is discussed in section [ harm ] in greater detail .
the loop top emissions from the isotropic and loss - cone electron distributions are almost the same . in the brightness maps ( fig .
[ fig05 ] ) , the maximum of the emission at 3.75 ghz is located between the loop top and the footpoints ; the images for the isotropic and anisotropic distributions are very similar . at the higher frequencies ,
the emission is concentrated at the footpoints .
in contrast to the previous case ( fig . [ fig03 ] ) , the footpoint sources are now more compact for the loss - cone electron distribution ; this effect becomes more pronounced with the frequency increase .
changing the pitch - angle distribution from the isotropic one to the loss cone results in a shift of the spectral peak of the footpoint emission towards lower frequencies for the loop located near the disk center , and towards higher frequencies for the loop near the limb ( which could be observationally addressed via center - to - limb variation analysis of the spatially resolved footpoint spectra ) .
this spectral peak variation affects the emission spectral index around the peak . with an increasing frequency ( in the optically thin range )
, the spectral indices of the emission from the isotropic and anisotropic electrons approach each other and gradually become the same .
we have confirmed that at @xmath62 , the emission spectral indices asymptotically approach the ultrarelativistic limit @xmath63 ( provided that the high - energy cutoff @xmath64 , * ? ? ? * and neglecting the free - free contribution ) for both the isotropic and anisotropic electron distributions . however , in the frequency range of @xmath65 ghz , the emission spectral index is strongly affected by the magnetic field inhomogeneity in the source ; as a result , the spectral index varies with frequency and can be noticeably different at the different footpoints and the loop top . at the frequencies
@xmath66 ghz , the spectrum is affected by the high - energy cutoff of the electron distribution ( note that figs .
[ fig03]-[fig06 ] are calculated for @xmath53 mev ) which makes the spectrum steeper ( the spectral index increases with frequency ) ; this effect is clearly visible for the loop - top emission where the magnetic field inhomogeneity is the lowest . on the other hand , at these and higher frequencies the free - free contribution can make the spectrum flatter ( see below examples of such flattening ) .
the loss - cone anisotropy of the fast electrons has opposite effect on the gs emission for the loops located near the solar disk center and the limb .
thus , for some intermediate longitude , the anisotropy effect should be minimal .
we have found that this occurs at the longitude of @xmath67 .
figures [ fig07]-[fig10 ] show the brightness maps and spatially resolved emission spectra for the models with the inhomogeneous distribution of the energetic electrons along the flaring loop ( the particles are accumulated at the loop top ) .
the number density of the energetic electrons at the loop top is taken to be @xmath68 @xmath49 .
the density profiles of the energetic electrons are described by equation ( [ ne ] ) with @xmath69 , 6 , and 8 , which corresponds to the ratios of the footpoint and loop - top number densities of @xmath70 , @xmath71 , and @xmath72 , respectively ( for @xmath69 , we obtain the footpoint number density of @xmath73 @xmath49 , like in the previous section ) . in all cases , the loss - cone distribution of the energetic electrons over the pitch - angle is used .
figures [ fig07]-[fig08 ] correspond to the loop located near the center of the solar disk .
since the number density of the energetic electrons at the loop top is assumed to be constant , the loop - top emission is almost independent on the parameter @xmath37 . in the footpoints , the increase of @xmath37 results in a decrease of the number density of the energetic electrons and , consequently , in a decrease of the emission intensity in the optically thin frequency range ; the spectral peak shifts towards lower frequencies . in the brightness maps ,
the intensity maximum is always located near the loop top . for @xmath69 , at the frequency of 3.75 ghz , the whole loop is seen with relatively weak brightness variations along it . at higher frequencies and/or with increasing @xmath37 , the visible emission source becomes more compact .
figures [ fig09]-[fig10 ] show the brightness maps and emission spectra for the loop located near the limb . like in the previous case , the loop - top emission is independent on the parameter @xmath37 while the footpoint emission intensity ( in the optically thin frequency range ) decreases with increasing @xmath37 .
distributions of the radio brightness along the loop can be qualitatively different for the different frequencies and inhomogeneity models : for @xmath69 , the emission maximum at 3.75 ghz is located at the loop top and gradually shifts towards the footpoints with increasing frequency . for @xmath74 and 8 ,
the intensity maximum is always located at the loop top , but the visible emission source becomes more compact with increasing frequency .
thus at high frequencies ( @xmath75 ghz ) , a brightness map will be dominated by either two footpoint sources ( for the loops with a moderate accumulation of the particles at the top , up to @xmath76 ) or one loop - top source ( for the loops with higher inhomogeneity ) .
we note an interesting behavior of the polarization map : for the loop - like optically thin sources ( at 17 or 34 ghz ) in figure [ fig09 ] most of the image area is dominated by the rcp sense , which is indicative of the same direction of the line - of - sight magnetic field component throughout most of the loop area .
similar unipolar polarization patterns are highly typical for the norh data on the limb flares .
if the concentration of the electrons at the loop top develops at the course of flare due to electron transport @xcite , this process of the electron inhomogeneity build - up can be roughly described in terms of the @xmath37 increase , which implies the apparent brightness peak motion ( at the optically thin frequencies ) from footpoints to the looptop in agreement with observations of some limb flares @xcite . in all the models ,
the emission spectral index at the loop top behaves like that for a homogeneous source : the index is nearly constant ( @xmath77 ) in the frequency range of 20 - 50 ghz ( for both loop orientations ) , and increases at higher frequencies due to the high - energy cutoff of the electron spectrum . at the footpoints , an increase of the parameter @xmath37 results in a decrease of the spectral peak frequency . with an increasing frequency ( in the optically thin range )
, the spectral indices for all values of @xmath37 become the same ( if the free - free emission is negligible , see below ) .
however , in contrast to the loop top , the spectral index of the footpoint emission is also affected by the source inhomogeneity . for the loop located near the disk center
, we can notice an asymmetry of the emission spectra produced at the different footpoints , which is caused by a difference of the viewing angles .
for the loop located near the limb , the high - frequency spectral index of the footpoint emission for the model with @xmath78 differs noticeably from the indices for @xmath69 and 6 .
this is caused by the contribution of the free - free emission of the thermal electrons which becomes dominant under these conditions : at large frequencies the spectral index of the free - free emission goes to zero .
changing the pitch - angle distribution of the energetic electrons affects the spatially resolved emission spectra in the same way as discussed in the previous section : in comparison to the isotropic distribution , the loss - cone one provides stronger footpoint emission for a loop near the limb and weaker footpoint emission for the loop near the disk center , while the loop - top emission remains almost unchanged .
since in the considered inhomogeneous models the loop - top emission source often dominates the images , the effect of the electron anisotropy on the brightness maps is only moderate .
lccccc isotropic & 0 & 6.03 & 2.40 & 2.46 & 2.05 + loss cone & 0 & 4.37 & 2.98 & 2.21 & 1.86 + isotropic & 4 & 7.59 & 1.75 & 1.82 & 1.79 + loss cone & 4 & 7.59 & 1.76 & 1.77 & 1.77 + loss cone & 6 & 6.92 & 1.66 & 1.72 & 1.76 + loss cone & 8 & 6.92 & 1.64 & 1.71 & 1.75 + isotropic & 0 & 12.02 & 0.57 & 2.10 & 2.14 + loss cone & 0 & 13.80 & 0.13 & 1.67 & 2.16 + isotropic & 4 & 10.47 & 0.94 & 1.87 & 1.92 + loss cone & 4 & 10.47 & 0.79 & 1.72 & 1.95 + loss cone & 6 & 7.24 & 1.77 & 1.78 & 1.78 + loss cone & 8 & 6.61 & 1.66 & 1.71 & 1.76 + in this section , we consider the total ( spatially integrated ) emission from the flaring loop .
although in the view of the currently operational and soon to be available radio instruments , considering the total radiation may look old - fashioned , we feel that this still makes sense for the following reasons .
most of the historically accumulated databases and corresponding statistical studies are done based on the total power observations ( e.g. , * ? ? ?
the total power data are simpler manageable as they can be easily visualized by dynamic spectra and characterized by only a few simple numbers , such as spectral indices , rise and decay times , peak flux and frequency @xcite . in particular , the corresponding high - frequency spectral index is widely used to evaluate the fast electron energy spectral index @xcite . from this perspective ,
a statement made by @xcite that such a spectral index is a good measure of the electron spectral index even for anisotropic electron distributions , if confirmed , could be of a great practical value .
the parameters characterizing the total emission are shown in the right columns in figs .
[ fig04 ] , [ fig06 ] , [ fig08 ] , and [ fig10 ] . considering the right top panel in figure [ fig04 ] as a vivid example , one can easily isolate three distinct regions of the spectra low - frequency part
( region i ) , middle - frequency part ( region ii ) , and high - frequency part ( region iii ) .
visual comparison of these total power spectra with the spatially resolved spectra from the footpoints and looptop in figure [ fig04 ] suggests that the low - frequency part is formed primarily at the looptop region with low magnetic field , the high - frequency part in the footpoints where the magnetic field is large , while the middle - frequency part by the entire loop and so related to the magnetic field non - uniformity .
the low - frequency part is known to be formed by the effect of the gs optical thickness and/or the razin - effect ( suppression of the gs emission in a dense background plasma , e.g. , * ? ? ?
* ) possibly accompanied by the free - free absorption in the dense plasma @xcite ; the slope of the spectrum can here be quantified by the index of @xmath79 or less , see the right bottom panel in figure [ fig04 ] .
the high - frequency part is mainly determined by the distribution of fast electrons including the energy spectrum @xcite and pitch - angle anisotropy @xcite ; note the anisotropy - related difference between the solid ( isotropic ) and dashed ( loss - cone ) curves in this panel . the middle - frequency part is clearly seen in fig .
[ fig04 ] as it is almost flat ( the spectral index is around zero ) .
solar microwave bursts with flat spectra have been observed for decades .
for example , @xcite proposed that the free - free absorption of gs emission can form such spectra , while @xcite recognized that the electrons trapped in a large dipole magnetic loop can produce the flat radio spectra due to the source non - uniformity in certain parameter regimes .
our modeling confirms the finding made by @xcite .
in addition , we have found that this middle - frequency part is not always flat but can have either negative or positive spectral index , see , e.g. , fig .
[ fig06 ] , which can be misinterpreted as either region i or iii in observations with a limited spectral coverage .
in fact , the observations ( see , e.g. , fig . 11 in * ? ? ?
* ) reveal that the histograms of both low- and high - frequency spectral indices extend to zero implying that both low- and high - frequency spectra can be much flatter than those determined by optical thickness effect or electron energy index , respectively . for practical application
, this means that having the spectrum falling with the frequency does not guarantee that its slope is formed by either energetic or angular properties of the electron distribution function but can instead be related to the source non - uniformity .
one of the parameters controlling the shape of the middle - frequency part is the viewing angle : figure [ fig06 ] presents the case when this part grows with frequency , which overall broadens the spectrum peak . not surprisingly
, non - uniform spatial distribution also significantly affect this part of the spectrum .
in fact , with an inhomogeneous non - thermal electron spatial distribution ( with their concentration at the looptop ) , the radio spectrum begins to resemble emission from a roughly uniform source , see figures [ fig08 ] and [ fig10 ] .
the reason for this to happen for an _ inhomogeneous _ source is very simple : with the adopted inhomogeneous electron distribution most of them reside at the looptop , where the spatial variation of the magnetic field and the viewing angle are minor , so we have a situation similar to a uniform source . finally , let us consider to what extent we could rely on the high - frequency spectral index in evaluating the fast electron energy spectrum or pitch - angle anisotropy .
the obtained spectral behavior of the local spectral index can easily be tracked in the figures , so we do not describe it here in any detail . instead ,
in order to quantify it with measures relevant to available observations , we form a few distinct spectral indices , namely , an index @xmath80 at 15 ghz relevant to the ovsa spectral range @xcite , the `` nobeyama '' index @xmath81 calculated with two well separated frequencies , 17 and 34 ghz ( e.g. , * ? ? ? * ) , and higher - frequency spectral index @xmath82 taken at 50 ghz relevant to a number of earlier separated - frequency observations ( e.g. , * ? ? ? * ) ; the results are gathered in table [ tab1 ] .
one can see , that for the same energy electron distribution , @xmath83 , the measured spectral indices are highly different depending on the pitch - angle anisotropy , the viewing angle , spatial inhomogeneity of the fast electron distribution , and the instrument used to measure the spectral index .
in fact , the numbers given in table [ tab1 ] are bounded between 0.13 and 2.98 , although as has already been said , the ultrarelativistic approximation predicts the emission spectral index of @xmath63 , or @xmath84 for the adopted value of @xmath51 , while a widely used approximation proposed by @xcite predicts the value of @xmath85 , or @xmath86 for @xmath51 ; none of this approximate values is favored by table [ tab1 ] .
stated another way , the radio indices from 0.13 to 2.98 would imply the electron energy index range from 1.26 to 6.96 if the relativistic asymptote is used and from 1.50 to 4.67 if the dulk - marsh approximation is applied . again , none of the values is truly informative in terms of recovering the original electron energy index as there is no unique way of linking the measured radio spectral index with the original electron energy spectral index .
this conclusion holds ( with narrower scatter of the recovered values , @xmath87 and @xmath88 , respectively ) if we consider only higher - frequency indices @xmath81 and @xmath82 . on the other hand ,
the ovsa index @xmath80 is more sensitive to the particle anisotropy , which , thus , can be studied through a forward fit of the radio spectra like in @xcite .
the gs emission from a homogeneous source can demonstrate an oscillatory spectral structure in the low - frequency range ( @xmath89 ) , when the emission intensity increases at a narrow spectral region near the harmonics ( small integer multiples ) of the electron cyclotron frequency ( see , e.g. , the figures in the articles of * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in an inhomogeneous source , however , this harmonic structure can be hidden because of natural smoothing : the resonance giving rise to a gyroharmonics at a given location will vary with frequency due to the spatially dependent resonant condition in the spatially non - uniform magnetic field .
thus , even if a spectrum from a single pixel contains harmonics they often disappear after integration over even a relatively small part of the source .
this is why no harmonic structure is present in either footpoint or looptop spectra in figures [ fig04 ] , [ fig06 ] , [ fig08 ] , and [ fig10 ] other than a number of extremely large peaks in figures [ fig06 ] and [ fig10 ] for the loss - cone case provided that the conditions for the electron cyclotron maser ( ecm ) instability ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) are locally fulfilled .
we do not consider the coherent ecm emission here concentrating on the harmonic structure of the incoherent gs emission . in the simplest case of a uniform gs source ,
the harmonic structure is more prominently pronounced for the source viewed at a quasitransverse direction relative to the source magnetic field .
note , that for the magnetic model adopted here , the harmonic structure is only expected from the footpoint and leg regions , while not from the looptop where the magnetic field is too weak for the gyroharmonics to be produced at the considered parameter regime ; thus , the limb location of the loop is most favorable to produce distinct gyroharmonics , which is confirmed by comparing figures [ fig11]-[fig12 ] .
these figures show the brightness maps at low frequencies for two loops locations at the disk center and limb .
as expected , there is no apparent harmonic structure in the images seen at the disk center , fig .
[ fig11 ] .
in contrast , for the loop located near the limb , the images contain a number of bright stripes highlighting the isolines of the magnetic field strength corresponding to the gyroresonance conditions at a given frequency . at 2 ghz , for example
, these stripes are clearly visible from the footpoints up to about half of the loop height .
if one were gradually increasing the frequency , for which the image has been computed , each stripe would demonstrate an apparent down - motion because same harmonic number requires proportionally larger magnetic field for a higher frequency .
accordingly , at higher frequency images , the stripes are shifted toward the footpoints , and their contrast ( or amplitude of the intensity oscillations ) decreases . apparently , having these stripes detected in real observations would offer a highly efficient way of model - independent quantitative measurement of the coronal magnetic field ( note that averaging along the line of sight plays here only a minor role since the gs optical depth is large at the gyro harmonics so the layer contributing to the emission is narrow implying almost uniform magnetic field along the essential part of the integration ) . in case of rather compact loop considered here
this task requires a radio array with the spatial resolution about 1 " at the low frequencies ( where scattering of radio waves by coronal density inhomogeneities can additionally fuze the images , @xcite , however ) , which is comparable with the anticipated capability of the frequency agile solar radiotelescope ( fasr , * ? ? ?
* ) ; although for the instruments currently under development ( e.g. , ussrt and eovsa ) detecting the harmonic stripes can only be expected from the biggest flaring loops and/or for the cases with a more uniform magnetic field .
such favorable cases , although untypical , are not unlikely : @xcite reported a total power data on a microwave burst obtained with a few single frequencies that demonstrated prominent harmonic - like enhancements at two frequencies .
furthermore , the ovsa total power data reveal clear horizontal stripes at the dynamic spectra of some events , e.g. , 2003 june 15 flare @xcite ; see also @xcite , which might indicate the gyroharmonic contribution ; however , such an interpretation remains ambiguous unless confirmed by direct imaging data .
we have considered microwave emission produced by moderately anisotropic electron distributions populating ( uniformly or non - uniformly ) a non - uniform symmetric dipole magnetic loop .
we emphasize that for the adopted loop geometry , the mirror ratio is more than 10 , so in most of the loop volume ( except the close vicinity of the footpoints ) the loss - cone angle defined by the adiabatic invariant ( figure [ fig02]b ) is small and , accordingly , the angular distribution of the fast electrons is close to the isotropic one , which we here conventionally call a `` moderate anisotropy '' .
but even this moderate anisotropy affects noticeably both images and spectral characteristics of the emission . in particular , for the uniform electron distribution along the loop
the pitch angle anisotropy enhances the optically thin emission from the footpoint vicinity for the loop located at the limb , while suppresses it for the loop located at the disk center . in the case of non - uniform electron distributions due to accumulation at the loop top , the relative contribution of the electrons in the vicinity of the footpoints decreases and , so
, the effect of the anisotropy becomes less pronounced .
nevertheless , in most of the cases ( except looptop emission , where the electron angular distribution is almost isotropic ) the effect of anisotropy on the images , spectrum , polarization , and spectral index can easily be recognized .
we note that for larger thermal density than adopted for our restricted modeling , which is indeed often the case @xcite , the looptop emission will be strongly suppressed by the razin effect ( which is specified by the @xmath90 ratio and becomes especially strong at @xmath91 ) , so the emission from the footpoints and legs of the loop , where the electron anisotropy is stronger than at the looptop , will dominate the emission ; thus the anisotropy effect will be even stronger than for the cases discussed above , see on - line supplement album .
an interesting feature of the low - frequency images is the presence of the `` gyro - stripes '' , which are indicative of distinct gyroharmonics produced at certain heights changing with frequency ( since a few small integer multiples of the local gyrofrequency produce strong enhancement of the flux density ) .
having radio imaging spectroscopy data with the spatial resolution sufficient to resolve such gyro - stripes would offer a nice model - independent diagnostics of the coronal magnetic field during flares . in the near future
the expanded ovsa will be capable of such measurements at least for relatively large flaring sources .
an important issue is anisotropy / non - uniformity effect on the total power radio spectrum .
since the total power spectrum is the result of emission integration over the entire source area , the result of this integration depends on how the distinct spatially resolved contributions are weighted , which , in turn , depends on the anisotropy , inhomogeneity , and the viewing angle .
we found that all the spectrum peak , peak flux , polarization , and the spectral index depend on both anisotropy and electron distribution inhomogeneity , although the effects are counter - directed .
in particular , the radio spectral index changes noticeably with frequency and its behavior depends noticeably on the pitch - angle anisotropy , e.g. , one can see that the maximal value of the total power spectral index ( figure [ fig04 ] , right bottom ) for the anisotropic case exceeds by roughly 1 that for the isotropic case .
we compared the spectral indices with asymptotic relativistic ( synchrotron ) and approximate dulk - marsh values as the observed radio spectral indices are often used to evaluate the fast electron energy spectral indices .
we found that neither synchrotron nor the dulk - marsh value is a good approximation to the true spectral index : although the synchrotron index does represent the true index at high frequencies ( higher than considered in our plots ) , the dulk - marsh index is less meaningful even though it sometimes coincides with the true one at a given single frequency .
therefore , we do not confirm the conclusion made by @xcite that the dulk - marsh index is a quantitatively good approximation to the spectral index of the spatially integrated spectrum , which might be an artifact specific to their adopted source model .
we have introduced a flexible simulation tool capable of fast computing 3d models of the microwave emission utilizing recently developed fast gs codes and presented an example of its use for the microwave emission modeling . considering a highly restricted parameter space ( symmetric dipole magnetic loop , moderate anisotropy , weak or no razin - effect ) we have analyzed images , spectra , and polarization of the model radio emission in the view of the available and future ( more complete ) radio data . in particular
, we note that the high - frequency spectral index does not have a unique value for a given energy spectral index of radiating electrons ; instead , it noticeably varies with the frequency , the viewing angle , the anisotropy , and the inhomogeneity and also different for various parts of the source and the entire source .
this implies that the use of the radio spectral index for constraining the electron energy spectral index is not that straightforward ; in place , the electron spectrum recovery must rely on the forward fitting of the entire radio datacube ( ideally , including polarization ) like in the example presented by @xcite .
this work was supported in part by nsf grants ags-0961867 , ast-0908344 , and nasa grants nnx10af27 g and nnx11ab49 g to new jersey institute of technology , and by the rfbr grants 09 - 02 - 00226 , 09 - 02 - 00624 , and 11 - 02 - 91175 .
thanks the leverhulme trust for financial support . 41 natexlab#1#1 , a. t. , fleishman , g. d. , huang , g .-
l . , & melnikov , v. f. 2008 , , 677 , 1367 s. k. antiochos and p. a. sturrock , , 49,359 , m. j. 1990 , , 85 , 1141 , t. s. 1994 , , 426 , 774 , t. s. , benz , a. o. , & gary , d. e. 1998 , , 36 , 131 , t. s. , fleishman , g. d. , & gary , d. e. 2007 , , 666 , 1256 , s. g. & holman , g. d. 1992 , , 391 , 854 , g. a. & marsh , k. a. 1982 , , 259 , 350 , g. & arzner , k. 2000 , , 358 , 776 , g. d. 2006 , in solar physics with the nobeyama radioheliograph , 5162 , g. d. , bastian , t. s. , & gary , d. e. 2008 , , 684 , 1433 , g. d. , gary , d. e. , & nita , g. m. 2003 , , 593 , 571 , g. d. & kuznetsov , a. a. 2010 , , 721 , 1127 , g. d. & melnikov , v. f. 1998 , uspekhi fizicheskikh nauk , 41 , 1157 . 2003 , , 587 , 823 . 2003 , , 584 , 1071 , g. d. , nita , g. m. , & gary , d. e. 2009 , , 698 , l183 , d. e. 2003 , journal of korean astronomical society , 36 , 135 , d. e. & keller , c. u. , eds . 2004 , astrophysics and space science library , vol .
314 , solar and space weather radiophysics - current status and future developments , v. l. & syrovatskii , s. i. 1965 , , 3 , 297 , d. a. & castelli , j. p. 1975
, , 44 , 155 , r. g. , melrose , d. b. , & ronnmark , k. g. 1982 , australian journal of physics , 35 , 447 , g. d. , eichler , d. , & kundu , m. r. 1980 , in iau symposium , vol .
86 , radio physics of the sun , ed .
m. r. kundu & t. e. gergely , 457459 , t. a. , dulk , g. a. , kiplinger , a. l. , winglee , r. m. , bastian , t. s. , & graeter , m. 1993 , , 412 , 853 , a. a. 2011 , , 526 , a161 , j. & treumann , r. a. 2002 , , 101 , 295 , j. w. , gary , d. e. , & zirin , h. 1994 , , 152 , 409 , v. f. & magun , a. 1998 , , 178 , 153 , v. f. , shibasaki , k. , & reznikova , v. e. 2002 , , 580 , l185 , g. m. , gary , d. e. , & lee , j. 2004 , , 605 , 528 , p. & alissandrakis , c. e. 1992 , , 257 , 307 , r. 1969 , , 158 , 753 , r. & petrosian , v. 1972 , , 178 , 241 , v. e. , melnikov , v. f. , shibasaki , k. , gorbikov , s. p. , pyatakov , n. p. , myagkova , i. n. , & ji , h. 2009 , , 697 , 735 , r. r. & vlahos , l. 1984 , , 280 , 405 , p. j. a. & costa , j. e. r. 2006 , , 453 , 729 . 2010 , , 266 , 109 , m. , magun , a. , & schanda , e. 1987 , , 111 , 181 , a. v. 1978 , pisma astronomicheskii zhurnal , 4 , 193 , r. a. 2006 , , 13 , 229 , v. , nindos , a. , & alissandrakis , c. e. 2008 , , 253 , 79 , c. s. 1985 , , 41 , 215 , c. s. & lee , l. c. 1979 , , 230 , 621
similar with other models used in the past ( e.g. * ? ? ?
* ) , in our simulation tool the magnetic field is produced by magnetic dipole with a moment @xmath92 that makes an adjustable angle @xmath93 with its corresponding solar radius , and located below the solar surface at a depth @xmath94 .
the flaring loop is constructed around a central field line that is chosen to lie in the plane defined by the magnetic dipole vector and the local vertical , referred from now on as the central plane of the loop . as pictured in figure [ dipole_geometry ] , in the dipole s cartesian system of
coordinates @xmath95 , which is defined as having the axis @xmath96 oriented along the magnetic dipole @xmath92 , and the axis @xmath97 in the central plane , the central magnetic field line may be parameterized as @xmath98 where @xmath99 $ ] represents the angle between the position vector @xmath100 and the dipole s axis , i.e. @xmath101 , and @xmath102 is the height of the central line , i.e the maximum distance between the central line and the dipole s axis , which corresponds to @xmath103 . in a solar coordinate system defined as having the axis @xmath104 along the solar radius and the axis @xmath105 identical with the axis @xmath106 , the same central line is expressed as @xmath107 equations ( [ centerline0 ] ) and ( [ centerline ] ) are a direct consequence of the fact that any magnetic dipole field line is described by the general polar equation @xmath108 and that the dipole s system of coordinates is rotated relative to the solar system of coordinates by the angle @xmath109 around the @xmath105 axis , i.e. @xmath110 the magnetic flux tube centered on the central field line is defined in terms of a circular cross section of radius @xmath111 that is normal to the central field at @xmath103 , i.e the magnetic loop - top .
hence , the cartesian coordinates of an arbitrary magnetic field line , @xmath112 , that intersects the loop - top cross section at distance @xmath113 from the central field line , may be expressed in terms of three convenient free parameters @xmath114 as @xmath115 where , as shown in figure [ dipole_geometry]b , @xmath23 represents the fixed angle between the vector @xmath116 and the dipole s vertical axis @xmath97 , and @xmath36 represents the variable angle between the vector @xmath117 and the dipole moment @xmath92 that is measured in the plane containing the field line , which makes the angle @xmath118 with the central plane of the magnetic flux tube . for @xmath119
, the set of equations ( [ tubeline ] ) defines the envelope of the magnetic flux tube , which has a circular cross section shape only for @xmath103 , i.e. at the loop - top , while the shape of its cross section is continuously changing along the central field line .
the condition for an arbitrary point @xmath117 to belong to a magnetic flux tube of height @xmath102 and top circular cross section of radius @xmath111 , i.e. the condition for the magnetic field line passing trough the point @xmath95 to intersect loop - top circular cross section , may be expressed as @xmath120 ^ 2+\zeta^2\left ( 1 + \frac{\xi^2 } { \zeta^2+\eta^2}\right)^3\\\nonumber \le\rho_0 ^ 2,\end{aligned}\ ] ] which has to be combined with the condition @xmath121 in order to determine if such flux tube point is also located above the solar surface . considering a normal cross section of the flux tube corresponding to an angle @xmath36 measured in the central plane of the flux tube , the parameter @xmath122\big\},\end{aligned}\ ] ]
may be used , instead of the angular coordinate @xmath36 , as a convenient flux tube longitudinal coordinate corresponding to the normal cross - section of interest .
the strength of the magnetic field is controlled by a unique adjustable parameter @xmath123 that defines the absolute value of the loop top magnetic field vector @xmath124 , which is perpendicular on the loop - top cross section . since the vector @xmath124 uniquely determines the dipole moment @xmath125 , the magnetic field vector in any point characterized by the position vector @xmath126 relative to the dipole origin
is given by @xmath127 | microwave emission of solar flares is formed primarily by incoherent gyrosynchrotron radiation generated by accelerated electrons in coronal magnetic loops .
the resulting emission depends on many factors , including pitch - angle distribution of the emitting electrons and the source geometry . in this work ,
we perform systematic simulations of solar microwave emission using recently developed tools ( gs simulator and fast gyrosynchrotron codes ) capable of simulating maps of radio brightness and polarization as well as spatially resolved emission spectra .
a 3d model of a symmetric dipole magnetic loop is used .
we compare the emission from isotropic and anisotropic ( of loss - cone type ) electron distributions .
we also investigate effects caused by inhomogeneous distribution of the emitting particles along the loop .
it is found that effect of the adopted moderate electron anisotropy is the most pronounced near the footpoints and it also depends strongly on the loop orientation .
concentration of the emitting particles at the loop top results in a corresponding spatial shift of the radio brightness peak , thus reducing effects of the anisotropy .
the high - frequency ( @xmath0 ghz ) emission spectral index is specified mainly by the energy spectrum of the emitting electrons ; however , at intermediate frequencies ( around 10 - 20 ghz ) , the spectrum shape is strongly dependent on the electron anisotropy , spatial distribution , and magnetic field nonuniformity .
the implications of the obtained results for the diagnostics of the energetic electrons in solar flares are discussed . |
the exterior region of the maximally extended schwarzschild space - time is described by the metric @xmath2 for any null geodesic in this exterior region the null geodesic equations give @xmath3 where @xmath4 is an affine parameter along the geodesic .
the right side here is evidently positive for @xmath5 and negative for @xmath6 .
it follows that any future endless null geodesic in the maximally extended schwarzschild space - time starting at some point with @xmath5 and initially directed outwards , in the sense that @xmath7 is initially positive , will continue outwards and escape to infinity .
any future endless null geodesic in the maximally extended schwarzschild space - time starting at some point with @xmath8 and initially directed inwards , in the sense that @xmath7 is initially negative , will continue inwards and fall into the black hole .
the hypersurface @xmath9 , known as the schwarzschild photon sphere , thus distinguishes the borderline between these two types of behaviour ; any null geodesic starting at some point of the photon sphere and initially tangent to the photon sphere will remain in the photon sphere .
( see darwin @xcite for a detailed analysis of the behaviour of null and timelike geodesics in schwarzschild space - time . )
the schwarzschild photon sphere also has physical significance for massive bodies . for any timelike geodesic in the exterior region the geodesic equations
give @xmath10 where @xmath11 is arc length along the geodesic . at any point with @xmath12 one may arrange for the two terms on the right of ( [ timegeodesic ] ) to cancel and so obtain a timelike geodesic at constant @xmath13 .
for @xmath14 the right hand side of ( [ timegeodesic ] ) is evidently negative .
thus any future endless timelike geodesic in the maximally extended schwarzschild space - time starting at some point between the event horizon at @xmath15 and the photon sphere at @xmath16 and initially directed inwards , in the sense that @xmath17 is initially negative , will continue inwards and fall into the black hole .
any observer who traverses a schwarzschild photon sphere must therefore engage some form of propulsion or else be drawn in to the black hole to meet an inevitable fate .
a photon sphere has been defined by virbhadra & ellis @xcite as a timelike hypersurface of the form @xmath18 where @xmath19 is the closest distance of approach for which the einstein bending angle of a light ray is unboundedly large .
these authors subsequently @xcite considered the einstein deflection angle for a general static spherically symmetric metric and obtained an equation for a photon sphere .
the existence of a photon sphere in a space - time has important implications for gravitational lensing . in any space - time containing a photon sphere , gravitational lensing will give rise to relativistic images @xcite . the schwarzschild photon sphere may be usefully be compared with the concept of a closed trapped surface .
any null geodesic originating from any point on a closed trapped surface in schwarzschild space - time is drawn into the singularity at @xmath20 .
by contrast , any null geodesic originating from any point on the photon sphere will be drawn into the singularity if and only if it is initially directed inwards .
the main objectives of the present paper are to give a geometric definition of a photon surface in a general space - time and of a photon sphere in a general static spherically symmetric space - time .
an evolution equation is obtained for the cross - sectional area of a photon surface in a dynamic spherically symmetric space - time .
it is shown , subject to suitable energy conditions , that in any static spherically symmetric space - time a black hole must be surrounded by a photon sphere , and a photon sphere must surround either a black hole , a naked singularity or more than a certain amount of matter .
many examples are given of photon spheres in static spherically symmetric space - times .
photon surface evolution is considered for the dynamic space - time example of vaidya null dust collapse to a naked singularity .
the hypersurface @xmath21 in schwarzschild space - time has two main properties , first that any null geodesic initially tangent @xmath22 will remain tangent to @xmath22 , and second that @xmath22 does not evolve with time .
the following general definition of a photon surface is based on only the first of these properties .
a more restrictive class of photon surfaces may be defined when the space - time admits a group of symmetries ( see definition [ photonspheredef ] ) .
[ photonsurfacedef ] a photon surface of @xmath23 is an immersed , nowhere - spacelike hypersurface @xmath22 of @xmath24 such that , for every point @xmath25 and every null vector @xmath26 , there exists a null geodesic @xmath27 of @xmath23 such that @xmath28*k*@xmath29 .
any null hypersurface is trivially a photon surface .
photon surfaces are conformally invariant structures .
if @xmath22 is a photon surface of @xmath23 then @xmath22 is a photon surface of @xmath30 for any smooth function @xmath31 .
note that definition [ photonsurfacedef ] is entirely local .
in particular , a photon surface @xmath22 need contain no endless null geodesics of @xmath23 .
moreover , a photon surface need only be immersed , rather than embedded in @xmath32 , and so may have self - intersections .
if @xmath23 is of dimension @xmath33 ( @xmath34 ) then , through each point @xmath35 of a photon surface @xmath22 in @xmath23 , there is an @xmath36-parameter family of null geodesics of @xmath23 that lie entirely in @xmath22
. the paper will be principally concerned with photon surfaces in space - times of @xmath37 dimensions .
the exceptions are examples [ min3 ] and [ desitter ] which give photon surfaces in space - times of dimension @xmath38 and @xmath39 respectively .
[ min3 ] in minkowski 3-space @xmath40 , consider the single - sheeted hyperboloid @xmath22 given by @xmath41 for some constant @xmath42 .
this surface is doubly ruled , the rulings being given by @xmath43 ( @xmath44 ) , where @xmath45 identifies the intersection points with @xmath46 and @xmath47 is the parameter along the ruling lines .
the tangents @xmath48 to the ruling lines are null with respect to the @xmath40 metric .
clearly they are geodesics in @xmath40 . at each point of @xmath22
there can be just two null directions tangent to @xmath22
. these must therefore be the directions of the two ruling lines through that point .
hence @xmath22 is a photon surface in the sense of definition [ photonsurfacedef ] ( see fig .
[ claire1 ] ) .
the two families of ruling lines are null geodesics with respect to both the minkowski @xmath49-metric and the induced @xmath50-metric . the lines may be regarded as the space - time paths of pulsed laser beams.,width=288 ] note that for any circle of the form @xmath51 and any future - directed timelike vector field @xmath52 along @xmath53 that respects the symmetry of @xmath53 , in the sense of @xmath54 for constant @xmath55 , @xmath56 such that @xmath57 , there is a unique single - sheeted hyperboloid @xmath22 through @xmath53 such that @xmath52 is tangent to @xmath22 along @xmath53 . in the case
@xmath58 , equation ( [ eq2 ] ) gives the null cone through the origin .
the complement of @xmath35 in this null cone is a null photon surface of @xmath40 .
[ min4 ] one may generalize example 1 to minkowski 4-space @xmath59 as follows .
let @xmath22 be a timelike hypersurface in @xmath59 of the form @xmath60 for some constant @xmath61 .
the two - parameter family of lines @xmath62 foliate @xmath22 and are null geodesics with respect to the @xmath59 metric . for each @xmath63 , the tangents at @xmath35 to those @xmath64 that pass through @xmath35
can be shown to generate the null cone of @xmath65 .
hence @xmath22 is a photon surface in the sense of definition [ photonsurfacedef ] . in terms of the double null coordinates @xmath66 for @xmath67 , equation ( [ ellhyp4 ] ) assumes the simple form @xmath68 in the future direction @xmath22 tends asymptotically to the null hypersurface @xmath69 , whilst in the past direction @xmath22 tends asymptotically to the null hypersurface @xmath70 .
[ desitter ] de sitter space - time may be regarded @xcite as a single - sheeted hyperboloid in minkowski @xmath71-space @xmath72 . by analogy with examples 1 and 2 ,
de sitter space - time is thus realized as a photon surface in @xmath72 . [ robwal ] since all robertson - walker models are conformally flat and therefore locally conformally transformable to minkowski space , the photon surfaces of any such model may thus be obtained , at least locally , by conformal transformations of minkowski space .
[ maintheorem ] let @xmath22 be a timelike hypersurface of @xmath73 .
let @xmath74 be a unit normal field to @xmath22 and let @xmath75 be the induced metric on @xmath22 .
let @xmath76 be the second fundamental form on @xmath22 and let @xmath77 be the trace - free part of @xmath76 .
then the following are equivalent : a. @xmath22 is a photon surface ; b. @xmath78 @xmath79 null @xmath26 @xmath80 ; c. @xmath81 ; d. every affine null geodesic of @xmath82 is an affine null geodesic of @xmath23 .
\(i ) @xmath83 ( ii ) .
suppose @xmath22 is a photon surface .
let @xmath63 and let @xmath84 be null .
there exists an affine null geodesic @xmath85 of @xmath73 such that @xmath86 .
one has @xmath87 along @xmath88 . at @xmath35
this gives @xmath89 .
\(ii ) @xmath83 ( iii ) .
let @xmath63 .
by ( ii ) one has @xmath90 @xmath79 null @xmath91 .
let @xmath92 be an orthonormal basis for @xmath65 with @xmath93 timelike and @xmath94 , @xmath95 spacelike .
any null @xmath91 , normalized such that @xmath96 , has components @xmath97 with respect to @xmath98 for some @xmath99 .
a calculation gives @xmath100 this must vanish for all @xmath101 .
one thus has @xmath102 and @xmath103 .
since @xmath77 is trace - free one must also have @xmath104 .
there follows @xmath81 .
\(iii ) @xmath83 ( iv ) . for any curve in @xmath22 with null tangent @xmath105 one
has @xmath106 where @xmath107 denotes covariant differentiation in @xmath22 with respect to @xmath108 . the second term on the right of @xmath109 vanishes by hypothesis . if @xmath105 is tangent to an affine null geodesic of @xmath82 then the term on the left of @xmath109 also vanishes and so @xmath110 is tangent to an affine null geodesic of @xmath23 .
\(iv ) @xmath83 ( i ) .
let @xmath63 and let @xmath84 be null .
let @xmath111 be an affine null geodesic of @xmath82 such that @xmath112 .
then , by ( iv ) , @xmath88 is an affine null geodesic of @xmath23 such that @xmath113 , @xmath114 . condition ( iii ) of theorem [ maintheorem ] is equivalent to a requirement that @xmath76 is pure trace in the sense of @xmath115 where @xmath116 is the expansion of the unit normal to @xmath22 .
for example [ min3 ] one has @xmath117 ; for example [ min4 ] one has @xmath118 .
( note that , by a standard abuse of notation , @xmath75 denotes both the induced metric on @xmath22 and the symmetric tensor field of rank @xmath119 along @xmath22 in @xmath32 which satisfies @xmath120 and pulls back to the induced metric on @xmath22 . )
it is clear from condition ( iii ) of theorem [ maintheorem ] that a space - time must be specialized in some respect in order to admit any timelike photon surfaces in the sense of definition [ photonsurfacedef ] .
for this reason it is helpful to restrict attention to space - times which admit groups of symmetries .
[ photonspheredef ] suppose @xmath23 admits a group @xmath121 of isometries .
a photon surface @xmath22 of @xmath23 that is invariant under @xmath121 , in the sense that each @xmath122 maps @xmath22 onto itself , will be called a @xmath121-invariant photon surface .
clearly any @xmath121-invariant null hypersurface is a @xmath121-invariant photon surface . in particular , if @xmath123 or @xmath124 , then any killing horizon @xcite is a @xmath121-invariant photon surface .
by definition , a general spherically symmetric space - time admits an @xmath1 isometry group for which the group orbits are spacelike @xmath50-spheres .
the following result describes the evolution of the cross - sectional area of an @xmath1-invariant photon surface in a spherically symmetric space - time .
[ firstphotonspheretheorem ] let @xmath23 be a spherically symmetric space - time .
let @xmath22 be an @xmath1-invariant timelike hypersurface of @xmath23 and let @xmath52 be the @xmath1-invariant unit future - directed timelike tangent vector field along @xmath22 orthogonal to the @xmath1-invariant @xmath50-spheres in @xmath22 .
let @xmath125 be one such @xmath1-invariant @xmath50-sphere in @xmath22 and let @xmath126 be the @xmath1-invariant @xmath50-sphere in @xmath22 at arc - length @xmath11 from @xmath125 along the integral curves of @xmath52 . then @xmath22 is a photon surface of @xmath23 iff the area @xmath127 of @xmath126 satisfies @xmath128 where @xmath129 is the unit normal to @xmath22 , @xmath130 is the expansion of @xmath129 and @xmath131 is the einstein tensor of @xmath23 .
let @xmath75 be the induced lorentzian @xmath49-metric on @xmath22 and , for each @xmath11 , let @xmath132 be the induced riemannian @xmath50-metric on @xmath126 .
the expansion of @xmath133 in @xmath134 is given by @xmath135 where the covariant derivative is that of @xmath136 .
since @xmath52 is both shear - free and vorticity - free in @xmath82 , the raychaudhuri equation for @xmath52 in @xmath137 assumes the form @xmath138 where @xmath139 is the ricci tensor of @xmath82 . from first principles
one has @xmath140 where @xmath141 is the second fundamental form of each @xmath126 in @xmath82 .
since @xmath142 is shear - free and vorticity free in @xmath82 one has @xmath143 .
the second fundamental form of @xmath22 admits the canonical decomposition @xmath144 .
equations ( [ 2r ] ) and ( [ 3r ] ) therefore give @xmath145 which combine to yield @xmath146 one may now substitute for the second term on the right of ( [ 3dray ] ) to obtain @xmath147 from first principles one has @xmath148 , and the gauss - bonnet theorem gives @xmath149 .
substituting for @xmath150 and @xmath151 in ( [ nearlythere ] ) one obtains @xmath152 this agrees with ( [ 2areaevolution ] ) iff @xmath153 .
construct , for the tangent bundle @xmath154 of @xmath22 , an orthonormal basis field of the form @xmath155 , with @xmath94 and @xmath95 unit spacelike .
with respect to this basis one has @xmath156 by spherical symmetry the vector field @xmath157 must be proportional to @xmath142 .
hence one has @xmath158 .
the vanishing of @xmath159 is thus equivalent to the vanishing of @xmath77 .
one has @xmath81 iff @xmath22 is a photon surface .
a spherically symmetric metric is locally expressible in the form @xmath160 with respect to coordinates @xmath161 adapted to the spherical symmetry , where @xmath162 , @xmath163 , @xmath164 and @xmath165 depend only on @xmath166 and @xmath167 .
it is often convenient to introduce a radial coordinate @xmath13 , depending only on @xmath166 and @xmath167 , such that @xmath168 is a function of @xmath13 only .
one is free to specify @xmath169 as a function of @xmath13 alone since to do so is , in effect , a definition of the coordinate @xmath13 .
this will be assumed to be done throughout this paper .
the following result is useful in the locating of @xmath1-invariant photon surfaces in dynamic spherically symmetric space - times .
[ mylemma ] let @xmath23 be a spherically symmetric space - time .
let @xmath22 be an @xmath1-invariant timelike hypersurface of @xmath23 and let @xmath52 be the @xmath1-invariant unit future - directed timelike tangent vector field along @xmath22 that is orthogonal to the @xmath1-invariant @xmath50-spheres in @xmath22 . then @xmath22 is a photon surface of @xmath23 iff @xmath170 holds along @xmath22 , where @xmath129 is the unit normal field to @xmath22 in @xmath23 . by spherical symmetry , and since @xmath52 is unit timelike , the vector field @xmath171 must be proportional to @xmath74 .
hence it suffices to show that @xmath22 is a photon surface iff along @xmath22 one has @xmath172 or equivalently @xmath173 construct for @xmath154 a local orthonormal basis field of the form @xmath174 . with respect to this basis field the components of @xmath175 form a diagonal matrix with @xmath176 equation ( [ l1 ] ) is thus equivalent to @xmath177 which is in turn equivalent to @xmath178 . in view of the trace - free property of @xmath179 , equation ( [ l1 ] ) is thus equivalent to @xmath180 . from theorem [ maintheorem ]
one has that @xmath181 holds along @xmath22 iff @xmath22 is a photon surface . let us continue to work with respect to the coordinate system @xmath182 employed in ( [ gss ] ) .
let @xmath183 be an integral curve of the vector field @xmath52 in lemma [ mylemma ] .
one has @xmath184 and equation @xmath185 becomes @xmath186 since @xmath187 is unit timelike one has @xmath188 the @xmath166 and @xmath167 components of equation @xmath189 thus combine to give @xmath190 where the components of @xmath74 are given by @xmath191 for @xmath192 where @xmath193 is the determinant of the time - space part of @xmath194 in ( [ gss ] ) .
equation ( [ x12x0 ] ) is the coordinate equivalent of ( [ gradxx ] ) and provides for the easy determination of @xmath1-invariant photon surfaces ( see example [ vaidyaex ] ) .
by definition , a spherically symmetric space - time is static if it admits an @xmath195 group of isometries such that the @xmath196 orbits are generated by a killing field @xmath197 which is both hypersurface orthogonal and orthogonal to the @xmath1 orbits .
the present section will be concerned with @xmath195-invariant photon surfaces in static spherically symmetric space - times .
such surfaces may be termed photon spheres because , as will be seen , they are a natural generalization to general static spherically symmetric space - times of the schwarzschild photon sphere concept . the term `` photon sphere '' will be regarded as applicable only in static spherically symmetric space - times . for clarity ,
the term `` @xmath195-invariant photon surface '' will usually be employed in preference to `` photon sphere '' . although the space - times of examples [ min3 ] and [ min4 ] are static and spherically symmetric , the photon surfaces in these space - times are not @xmath195-invariant and so are not photon spheres . of the robertson - walker space - times of example
[ robwal ] , only the einstein cylinder is both spherically symmetric and static .
none of the photon surfaces of the einstein cylinder are @xmath195-invariant .
thus the einstein cylinder has no photon spheres .
one may characterize an @xmath195-invariant photon surface , or photon sphere , in a static spherically symmetric space - time by means of the following special case of theorem [ firstphotonspheretheorem ] .
[ statssgeomtheorem ] let @xmath23 be a static spherically symmetric space - time with killing field @xmath197 and let @xmath22 be an @xmath195-invariant timelike hypersurface of @xmath136 .
then @xmath22 is an @xmath195-invariant photon surface of @xmath23 if there exists an @xmath1-invariant @xmath50-sphere @xmath198 satisfying @xmath199 where @xmath200 is the area of @xmath125 , @xmath129 is the unit normal to @xmath22 and @xmath130 is the trace of the second fundamental form of @xmath22 .
conversely , if @xmath22 is an @xmath195-invariant photon surface of @xmath23 then ( [ mainequation ] ) holds for every @xmath1-invariant @xmath50-sphere @xmath201 .
note that the unit future - directed timelike tangent field @xmath52 along @xmath22 in theorem [ firstphotonspheretheorem ] is proportional to the restriction to @xmath22 of the killing field @xmath197 .
suppose first that there exists an @xmath1-invariant @xmath50-sphere @xmath202 such that ( [ mainequation ] ) holds .
the quantities @xmath200 , @xmath130 and @xmath203 remain constant as @xmath125 is mapped along the flow lines of the killing field @xmath197 .
so they also remain constant as they are mapped along the flow lines of @xmath52 .
hence ( [ 2areaevolution ] ) holds with the term on the left and the first term on the right both zero .
thus , by theorem [ firstphotonspheretheorem ] , @xmath22 is a photon surface of @xmath23 . by hypothesis @xmath22
is @xmath195-invariant .
for the converse , suppose that @xmath22 is an @xmath195-invariant photon surface of @xmath23 .
then ( [ 2areaevolution ] ) holds for every @xmath1-invariant @xmath50-sphere @xmath204 .
since @xmath197 induces groups of local isometries , the area @xmath205 of @xmath126 is independent of the parameter @xmath11 .
hence the term on the left and the first term on the right of ( [ 2areaevolution ] ) both vanish and one obtains ( [ mainequation ] ) .
if one has @xmath206 @xmath79 vectors @xmath207 and @xmath22 is an @xmath195-invariant timelike photon surface of @xmath23 , then for any @xmath1-invariant @xmath50-sphere @xmath202 one has @xmath208 with equality holding iff @xmath209 along @xmath22 .
for schwarzschild space - time where ( see example [ schwarzschildex ] ) the only timelike photon sphere is at @xmath210 , one has @xmath211 , @xmath212 and @xmath213 which verifies ( [ mainequation ] ) and ( [ mainequationcor ] ) for this case . if the einstein equations hold with a zero cosmological constant then , in the corollary to theorem [ statssgeomtheorem ] , the hypothesis @xmath214 for all vectors @xmath207 is equivalent to @xmath215 for all vectors @xmath207 .
this is a physically reasonable energy condition . in particular , for a perfect fluid with density @xmath216 and pressure @xmath35 , it is equivalent to a condition that @xmath216 and @xmath35 are both non - negative . more generally , the condition holds for an energy tensor with a single timelike eigenvector ( type i in the classification of hawking & ellis @xcite ) iff each energy tensor eigenvalue is non - negative .
the characterization of timelike photon surfaces provided by theorem [ statssgeomtheorem ] involves derivatives of the metric components up to second order .
the following result ( theorem [ genphotonspheretheorem ] ) provides an entirely different characterization of @xmath217-invariant photon surfaces in terms of derivatives of the metric components up to only first order .
let a general static spherically symmetric metric @xmath218 be expressed in the form ( [ gss ] ) with @xmath169 a function of @xmath13 only .
the killing equation @xmath219 and the orthogonality of @xmath197 to @xmath220 and @xmath221 gives @xmath222 and hence @xmath223 where @xmath13 is to be regarded as a scalar field on @xmath32 .
since @xmath13 is independent of @xmath45 and @xmath224 it follows that any @xmath195-invariant hypersurface @xmath22 of @xmath23 must be of the form @xmath225 .
if @xmath22 is also a timelike hypersurface then @xmath226 is a spacelike vector field along @xmath22 and @xmath197 is a timelike vector field in a neighbourhood of @xmath22 .
if the coordinates @xmath166 and @xmath167 implicit in ( [ gss ] ) are chosen such that @xmath227 then the killing equation gives that all the metric components @xmath194 in ( [ gss ] ) are independent of @xmath166 . since @xmath13 is constant along the integral curves of @xmath197 , the coordinate @xmath167 must be a function of @xmath13 only .
a natural choice is @xmath228 .
one may redefine @xmath166 according to @xmath229 .
this diagonalizes the time - space part of @xmath194 and leaves the components of @xmath194 independent of @xmath166 .
furthermore the curves @xmath230 are unchanged except that they are re - parametrized .
the vector field @xmath231 then becomes a conformal killing field .
define the tensor field @xmath232 on @xmath32 , where the components are given with respect to the coordinate basis employed in ( [ gss ] ) and @xmath233 is the determinant of the time - space part of @xmath218 in ( [ gss ] ) , as in ( [ deltadef ] ) .
[ genphotonspheretheorem ] let @xmath23 be a static spherically symmetric space - time with @xmath218 of the form ( [ gss ] ) , with @xmath169 a function of the coordinate @xmath13 only .
let @xmath22 be an @xmath195-invariant timelike hypersurface of @xmath23 and suppose that @xmath234 is nowhere - zero along @xmath22 .
then @xmath22 is an @xmath195-invariant photon surface of @xmath23 iff @xmath235 holds along @xmath22 .
since @xmath23 is both spherically symmetric and static , the surface @xmath22 is of the form @xmath236 .
the unit spacelike normal to @xmath22 is therefore given by @xmath237 for @xmath238 the second fundamental form of @xmath22 is given by @xmath239 the vector fields @xmath240 form an orthonormal frame field along @xmath22 , with @xmath241 and @xmath242 unit spacelike and @xmath52 unit timelike .
one has @xmath243
@xmath244 @xmath245 condition ( iii ) of theorem [ maintheorem ] holds iff @xmath76 is proportional to @xmath75 and hence iff @xmath246 this is equivalent to @xmath247 which is in turn equivalent to ( [ genphotonsphere ] ) .
equation ( [ genphotonsphere ] ) can have solutions such that @xmath248 is a spacelike hypersurface and therefore not a photon surface ( see e.g. example [ rnex ] ) .
it is therefore always necessary in the use of theorem [ genphotonspheretheorem ] to check that the hypersurface @xmath248 is in fact timelike or null .
note that in a region of space - time where the killing field @xmath249 is spacelike , for example behind the event horizon of schwarzschild space - time , the hypersurfaces @xmath236 are necessarily spacelike and so can not be photon spheres .
( @xmath250 , components of @xmath194 independent of @xmath166 . ) * as discussed previously , for a static spherically symmetric metric it is possible to choose @xmath251 , with @xmath169 depending only upon @xmath13 and with all the components of @xmath194 independent of @xmath166 . in this case
( [ genphotonsphere ] ) reduces to @xmath252 this agrees with an equation obtained by virbhadra & ellis @xcite on the basis of a different definition @xcite of a photon sphere .
note that even though the components @xmath253 do not appear in ( [ photonsphere ] ) , they are not assumed to vanish .
a particular sub - case of interest is that of time - space coordinates , with @xmath254 timelike in the sense of @xmath255 .
another sub - case of interest is that of single null ( radiation ) coordinates , with @xmath256 null in the sense of @xmath257 .
( double null coordinates @xmath258 . )
* let @xmath259 , @xmath260 be double null coordinates in the sense of @xmath261 .
the radial coordinate @xmath13 is to be regarded as a function of @xmath262 and @xmath263 .
then ( [ genphotonsphere ] ) assumes the form @xmath264 the metric components @xmath265 enter here through the covariant derivatives of @xmath13 .
equations ( [ genphotonsphere ] ) , ( [ photonsphere ] ) and ( [ photonspheredoublenull ] ) may be referred to as photon sphere equations since they give the location of timelike photon spheres in static spherically symmetric space - times . in order to facilitate further progress , a general static spherically symmetric metric
will be written in such a form as to cast the einstein tensor in a particularly simple and convenient form .
one has from previous remarks that a general static spherically symmetric metric is locally expressible in the form @xmath266 where @xmath267 and @xmath268 are functions of @xmath13 only .
let @xmath269 then the metric assumes the form @xmath270 and the einstein tensor is given by @xmath271 for @xmath272 where a prime denotes differentiation with respect to @xmath13 . for the right sides of ( [ den ] ) , ( [ pr ] ) and ( [ ptheta ] ) to be defined at some radius @xmath273 one evidently needs @xmath274 and @xmath275 to be twice differentiable at @xmath276 .
equations ( [ den ] ) and ( [ pr ] ) combine to give @xmath277 whereby one may rewrite ( [ ptheta ] ) in the more convenient form @xmath278 in the perfect fluid case @xmath279 equation ( [ pressurediff ] ) reduces to the tolman - oppenheimer - volkoff equation @xmath280 by means of equations ( [ den ] ) and ( [ lambdadash ] ) , the photon sphere equation ( [ photonsphere ] ) becomes @xmath281 for @xmath13 such that @xmath282 , and hence such that the hypersurface @xmath248 is timelike , equation ( [ fluidphotonsphere ] ) gives the location of the @xmath195-invariant timelike photon surfaces for the metric ( [ metricform ] ) .
equation ( [ fluidphotonsphere ] ) is the basis for the following result which shows that , subject to a suitable energy condition , any black hole in a static spherically symmetric space - time must be surrounded by an @xmath195-invariant photon surface . for the purpose of this and subsequent results , a function @xmath283 on an interval @xmath284 will be said to be piecewise @xmath285 if @xmath284 is the disjoint union of a locally finite collection of intervals @xmath286 such that @xmath287 is @xmath285 .
each interval @xmath286 may be open , closed or half - open .
[ trappedsurfacetheorem ] suppose the metric @xmath218 has the form ( [ metricform ] ) for @xmath288 , for some @xmath289 , with @xmath274 and @xmath275 both @xmath290 , piecewise @xmath291 functions of @xmath292 .
suppose the following hold : 1 .
@xmath293 and @xmath294 are bounded functions of @xmath295 ; 2 .
@xmath282 @xmath296 ; 3 .
@xmath297 , @xmath298 @xmath296 ; 4 .
@xmath299 ; 5 . for each value of @xmath47 the @xmath50-surfaces
@xmath300 , @xmath288 , are such that @xmath301 exists as an embedded spacelike @xmath50-sphere in @xmath23 and is marginally outer trapped .
then @xmath23 admits an @xmath195-invariant timelike photon surface of the form @xmath302 for some @xmath303 .
fix @xmath47 and let @xmath105 be the outward future - directed null normal field along each @xmath304 , @xmath288 , normalized such that @xmath305 , where @xmath306 is the outward radial unit tangent to @xmath307 . since @xmath105 is parallelly propagated along each of the geodesic integral curves of @xmath74 , one has that @xmath308 is a well - defined , nowhere - zero null vector field along @xmath309 .
for @xmath310 the vector field @xmath105 has the form @xmath311 for @xmath312 the expansion of @xmath105 is given by @xmath313 the condition that @xmath314 is marginally outer trapped therefore implies @xmath315 the non - negativity of @xmath293 and @xmath294 gives , by means of ( [ den ] ) and ( [ lambdadash ] ) , that @xmath274 and @xmath275 are non - decreasing functions of @xmath316 . thus ( [ limit ] ) holds iff at least one of @xmath317 holds .
suppose the first of ( [ limitpair ] ) fails .
then the second must hold and one has @xmath318 from the boundedness of @xmath293 and @xmath294 on @xmath319 one has , by means of ( [ lambdadash ] ) and ( [ finite ] ) , that @xmath320 is finite .
this is incompatible with the second of ( [ limitpair ] ) .
hence the first of ( [ limitpair ] ) must hold .
let @xmath321 be the left side of ( [ fluidphotonsphere ] ) . by the non - negativity of @xmath294 and
the first of ( [ limitpair ] ) one has @xmath322 . by condition
( 4 ) , equation ( [ den ] ) and lhpital s rule one has @xmath323 and hence @xmath324 .
hence there exists some @xmath303 such that @xmath325 .
the hypersurface @xmath302 is an @xmath195-invariant photon surface of @xmath23 .
condition ( 3 ) of theorem [ trappedsurfacetheorem ] may be expressed more succinctly as @xmath206 @xmath79 vectors @xmath207 . with regard to condition ( 5 ) of theorem [ trappedsurfacetheorem ] , to have required @xmath326 to be contained in the hypersurface @xmath327 would have been too strong since , for schwarzschild space - time , no spacelike hypersurface of the form @xmath327 in the exterior region contains a marginally outer trapped @xmath50-surface .
theorem [ trappedsurfacetheorem ] may be interpreted to the effect that , subject to the energy conditions expressed in condition ( 3 ) , any static spherically symmetric black hole must be surrounded by an @xmath195-invariant timelike photon surface .
the following result may then be regarded as a partial converse in that it shows , subject to a suitable energy condition , that if there exists an @xmath217-invariant timelike photon surface then there must be a naked singularity or a black hole , or more than a certain amount of matter .
[ nopsprop ] suppose the metric @xmath218 has the form ( [ metricform ] ) for @xmath328 , with @xmath274 and @xmath275 both @xmath290 , piecewise @xmath291 functions of @xmath329 .
if the following all hold : 1 .
@xmath293 is a non - increasing , bounded function of @xmath330 ; 2 .
@xmath331 ; 3 .
@xmath332 @xmath333 ; 4 .
@xmath334 @xmath333 , then @xmath23 can contain no @xmath195-invariant timelike photon surfaces . by conditions ( 1 ) and ( 2 ) with equation ( [ den ] ) one has @xmath335 @xmath336 . by condition
( 4 ) one therefore has @xmath337 @xmath338 .
the left side of ( [ fluidphotonsphere ] ) is thus bounded from below by @xmath339 @xmath338 .
this is positive by condition ( 3 ) .
the left side of ( [ fluidphotonsphere ] ) is therefore non - vanishing for all @xmath340 .
note that this result is valid even for negative @xmath294 and @xmath293 .
condition ( 2 ) of proposition [ nopsprop ] prohibits any curvature singularity at @xmath20 .
condition ( 3 ) may be interpreted as a requirement that there is no black hole and less than a certain amount of matter .
the result shows that one of these two conditions must fail if there is an @xmath217-invariant timelike photon surface and conditions ( 1 ) and ( 4 ) both hold . when the matter is a perfect fluid it is possible to improve condition ( 3 ) of proposition [ nopsprop ] to condition ( 3 ) of the following result .
[ nopstheorem ] suppose the metric @xmath218 has the form ( [ metricform ] ) for @xmath328 , with @xmath274 and @xmath275 both @xmath341 , piecewise @xmath291 functions of @xmath329 .
if the following all hold : 1 .
the matter is a perfect fluid with pressure @xmath342 and density @xmath293 ; 2 .
@xmath343 ; 3 .
@xmath344 @xmath333 ; 4 .
@xmath345 @xmath333 , then @xmath23 can contain no @xmath195-invariant timelike photon surfaces .
let @xmath346 be the left side of equation ( [ fluidphotonsphere ] ) .
since @xmath274 and @xmath275 are @xmath341 , piecewise @xmath291 functions of @xmath329 , one has by equation ( [ pr ] ) that @xmath347 is a @xmath290 , piecewise @xmath341 function of @xmath348 .
the function @xmath349 is then a piecewise @xmath290 function of @xmath329 which , by means of of the tolman - oppenheimer - volkoff equation ( [ tov ] ) , is given by @xmath350 for @xmath351 such that @xmath352 equation ( [ feqn ] ) reduces to @xmath353 from condition ( 3 ) and equation ( [ fiszero ] ) one has @xmath354 , whence by condition ( 4 ) one has @xmath355 .
thus ( [ feqn1 ] ) gives @xmath356 . by condition
( 2 ) , equation ( [ den ] ) and lhpital s rule one has @xmath357 and hence @xmath358 .
since it has been established that @xmath349 is negative for all @xmath348 such that @xmath359 , one must therefore have @xmath360 for all @xmath329 . hence the space - time can contain no @xmath195-invariant timelike photon surfaces .
note that , as for proposition [ nopsprop ] , theorem [ nopstheorem ] is valid even for negative pressure and density .
one would like to remove the insufficient matter parts of condition ( 3 ) of proposition [ nopsprop ] and condition ( 3 ) of theorem [ nopstheorem ] , in other words to weaken these to a no - black - hole condition @xmath282 @xmath361 . but
no result to this effect is forthcoming . on the other hand
no counterexample is known . to conclude this section
it will be shown that the physical significance of the photon sphere in schwarzschild space - time , as discussed in the introduction , carries over to the general static spherically symmetric case .
suppose the metric has the form ( [ metricform ] ) for @xmath362 and is asymptotically flat in the limit @xmath363 .
assume @xmath298 , @xmath364 @xmath365 .
the matter need not be a perfect fluid . denote the left side of ( [ fluidphotonsphere ] ) by @xmath347 .
the condition of asymptotic flatness gives @xmath366 so , if there are any @xmath217-invariant timelike photon surfaces , there will be an outermost such surface @xmath22 . for simplicity
assume @xmath367 at @xmath22 .
let @xmath368 be the connected component of @xmath369 that has @xmath22 as its inner boundary and extends to @xmath370 .
let @xmath371 be the connected component of @xmath372 that has @xmath22 as its outer boundary .
consider first the case of a future endless affine null geodesic @xmath373 .
the null geodesic equations for the metric ( [ metricform ] ) give @xmath374 at any point @xmath375 such that @xmath376 one has @xmath377 . at any point @xmath378
such that @xmath376 one has @xmath379 .
thus if @xmath88 starts outside @xmath22 ( i.e. in @xmath380 ) and is initially directed outwards , in the sense that @xmath7 is initially positive , then @xmath88 will continue outwards . if @xmath88 starts in @xmath371 and is initially directed inwards , in the sense that @xmath7 is initially negative , then @xmath88 will continue inwards until it falls either into a singularity or through an @xmath195-invariant photon surface other than @xmath22 .
consider now a unit speed timelike geodesic @xmath381 .
the timelike geodesic equations give @xmath382 for any point @xmath383 one can arrange for the first three terms on the right of ( [ gentimelikegeodesic ] ) to cancel and so obtain a unit speed timelike geodesic @xmath384 through @xmath385 at constant @xmath13 .
for @xmath386 the first three terms on the right of ( [ gentimelikegeodesic ] ) are evidently negative .
for @xmath387 such that @xmath388 one has @xmath389 .
thus if @xmath390 starts in @xmath391 and is initially directed inwards , in the sense that @xmath17 is initially negative , then @xmath390 will continue inwards until it falls either into a singularity or through an @xmath195-invariant photon surface other than @xmath22 .
the following are some examples of @xmath195-invariant and @xmath1-invariant photon surfaces in familiar space - times .
[ schwarzschildex ] the metric of schwarzschild space - time in single null ( radiation ) coordinates has the form @xmath392 in this case equation @xmath393 reduces to @xmath394 .
the timelike hypersurface @xmath395 is thus an @xmath195-invariant photon surface , or photon sphere , as expected .
there are no other @xmath195-invariant timelike photon surfaces . for a non - zero cosmological constant @xmath396 , the schwarzschild metric ( [ schmetric ] )
generalizes to the schwarzschild - de sitter metric @xmath397 one finds that equation ( [ photonsphere ] ) reduces to @xmath210 , independent of the value of @xmath396 .
this is surprising since schwarzschild and schwarzschild - de sitter space - times are not conformally related .
the schwarzschild interior solution describes a spherically symmetric distribution of perfect fluid of radius @xmath398 , bounded pressure @xmath35 and constant density @xmath399 .
the solution is to have a metric of the form ( [ metricform ] ) and is to be matched at @xmath400 to a schwarzschild vacuum solution in such a way that the pressure is a continuous function of @xmath13 .
one thus has @xmath279 for all @xmath401 , @xmath402 for all @xmath403 , @xmath404 for all @xmath405 and @xmath406 for all @xmath407 .
the pressure @xmath342 for @xmath403 is to be obtained by an integration of the tolman - oppenheimer - volkoff equation ( [ tov ] ) subject to the boundary condition @xmath408 .
this yields @xmath409 { \rm e}^{\mu ( r ) } & = \frac{\mbox{\normalsize $ ( 3-u(r))^2$}}{\mbox{\normalsize $ 4u^2(r)$ } } \ , & 0\leq r \leq r \ , \\[2ex ]
p(r ) & = \frac{\mbox{\normalsize $ u(r)-1$}}{\mbox{\normalsize$ 3-u(r)$ } } \
\rho_0 \ , & 0\leq r \leq r \ ,
\label{pconden}\\ \intertext{for } u(r ) & : = \left(\frac{3 - 8\pi\rho_0 r^2}{3 - 8\pi\rho_0 r^2}\right)^{\frac{1}{2 } } \ , & 0\leq r \leq r \ . \label{udef}\end{aligned}\ ] ] the spherically symmetric system described by the schwarzschild interior solution can exist in a state of stable equilibrium iff @xmath410 ( see in stephani @xcite ) .
this condition is equivalent to @xmath411 , which implies @xmath412 @xmath413 , and implies that the absence of a black hole is a general feature of the schwarzschild interior solution .
the left side of equation ( [ fluidphotonsphere ] ) now assumes the form @xmath414 1-\frac{\mbox{\normalsize $ 3m(r)$}}{\mbox{\normalsize $ r$ } } & : r \geq r \ . \end{array
} \right.\ ] ] for @xmath415 one has that ( [ intschps ] ) is positive for all @xmath340 , so there are no timelike photon spheres . for @xmath416 there is a single timelike photon sphere which lies at the boundary @xmath400 of the matter . for @xmath417
there is one timelike photon sphere outside the matter at @xmath418 and one timelike photon sphere inside the matter at @xmath419 for fixed @xmath398 the radius of the outer photon sphere is a strictly increasing function of @xmath420 whilst the radius of the inner photon sphere is a strictly decreasing function of @xmath420 .
thus a schwarzschild interior solution matched to a schwarzschild vacuum exterior solution contains no black hole , and contains one timelike photon sphere iff @xmath421 and two timelike photon spheres iff @xmath420 lies in the range @xmath422 .
however for such values of @xmath420 the space - time is unphysical in that the pressure at the center is @xmath423 .
therefore under the reasonable energy condition @xmath424 @xmath425 there are no photon spheres in this example .
the energy condition @xmath426 @xmath427 is in fact satisfied iff @xmath420 lies in the range @xmath428 .
the corresponding space - times ( in view of @xmath429 ) satisfy all of conditions ( 1 ) to ( 4 ) of theorem [ nopstheorem ] , except the smoothness of @xmath293 at @xmath400 . thus considering a smoothed family of solutions approximating the solution above and having it as a strict limit , but each with smooth @xmath293 at @xmath400 , we can apply theorem [ nopstheorem ] to show no photon spheres exist in all the cases discussed in this section where this energy condition is satisfied . on the other hand proposition [ nopsprop ]
is directly applicable without smoothing , but only applies to the subset of these cases with @xmath430 .
[ rnex ] the general static spherically symmetric solution to the einstein - maxwell equations is the reissner - nordstrm solution comprising a metric @xmath218 and an electromagnetic field @xmath431 given by @xmath432 where @xmath433 is the adm mass and @xmath434 is the electric charge .
we assume @xmath435 .
there is an event horizon at @xmath436 and a cauchy horizon at @xmath437 .
the event horizon exists for @xmath438 , and the cauchy horizon exists for @xmath439 . for @xmath440
they both lie at @xmath441 . there can be no timelike photon spheres between the event horizon and the cauchy horizon because the killing field @xmath197 is spacelike there .
outside the event horizon the killing field @xmath197 is timelike , so every hypersurface of the form @xmath236 which lies outside the event horizon and satisfies the photon sphere equation ( [ photonsphere ] ) is necessarily a timelike photon sphere . equation ( [ photonsphere ] ) assumes the form @xmath442 which has solutions @xmath443 given by @xmath444 the hypersurface @xmath445 exists for @xmath446 and lies outside the event horizon and is therefore a timelike photon sphere . the hypersurface @xmath447 exists for @xmath448 but lies outside the event horizon only for @xmath449 , and so is a timelike photon sphere only then .
the hypersurfaces @xmath450 and @xmath451 coincide for @xmath452 .
the cauchy horizon and event horizon are always null photon spheres .
for @xmath453 the curvature singularity at @xmath20 is locally naked but hidden behind an event horizon which lies strictly inside the only timelike photon sphere .
for @xmath454 the singularity at @xmath20 is globally naked and is surrounded by two timelike photon spheres in the case @xmath455 , one timelike photon sphere in the case @xmath452 and by no photon spheres , either timelike or null , in the case @xmath456 ( see fig .
[ rnfigure ] ) . the most general static spherically symmetric solution to the einstein massless
scalar field equations for a scalar field @xmath457 satisfying @xmath458 was obtained by janis , newman and winicour @xcite .
the ricci tensor has the form @xmath459 .
the solution is known @xcite to be expressible in the form @xmath460 for @xmath461 where the constants @xmath462 , @xmath463 are related to the adm mass @xmath433 and scalar charge @xmath464 by @xmath465 we assume @xmath466 .
there is a curvature singularity at @xmath467 . in order to obtain
@xmath468 one must assume @xmath469 .
for @xmath470 the solution reduces to the schwarzschild solution .
since all hypersurfaces of the form @xmath248 are timelike , one has from the photon sphere equation @xmath393 that the only timelike photon sphere is at @xmath471 which exists only for @xmath472 , i.e. for @xmath473 .
for @xmath474 it is known @xcite that a photon coming from infinity is deflected through an unboundedly large angle , i.e. the photon passes increasingly many times around the singularity as the closest distance of approach tends to the right side of @xmath475 .
a static spherically symmetric space - time with a charged dilaton field was obtained by horne & horowitz @xcite .
it comprises a metric @xmath218 , a dilaton field @xmath457 and an electromagnetic field @xmath431 given by @xmath476 where @xmath477 and @xmath478 are related to the adm mass @xmath433 and electric charge @xmath434 by @xmath479 and @xmath480 is a free parameter which controls the coupling strength between the dilaton and maxwell fields , with @xmath481 defined in terms of @xmath482 by @xmath483 we assume @xmath435 .
for @xmath484 the solution reduces to the reissner - nordstrm solution considered in example [ rnex ] . for @xmath485 and @xmath486
the solution reduces to the schwarzschild solution .
the solution also reduces to the schwarzschild solution for @xmath487 and arbitrary @xmath480 . here
we shall consider the case @xmath488 . in this case one
has @xmath489 .
there is an event horizon at @xmath490 and a curvature singularity at @xmath491 .
for @xmath492 the singularity at @xmath493 lies inside a black hole whilst for @xmath494 it is globally naked .
for @xmath495 the photon sphere equation ( [ photonsphere ] ) reduces to @xmath496 for @xmath492 one has @xmath497 so there is a single timelike photon sphere . for @xmath498 one
has @xmath499 so there are no timelike photon spheres . for @xmath500 both @xmath501 and @xmath502
are complex so there are no timelike photon spheres . for @xmath503 one
has @xmath504 so there are again no timelike photon spheres .
thus in the black hole case @xmath492 there is a single timelike photon sphere , whilst in the naked singularity case @xmath494 there are no timelike photon spheres .
( see fig .
[ ghsfig ] . ) [ vaidyaex ] the vaidya null dust collapse model is a non - static , spherically symmetric space - time with a metric which , in terms of single null ( radiation ) coordinates @xmath505 , assumes the form @xmath506 where @xmath507 is a freely specifiable function of @xmath262 . setting @xmath508 , @xmath251 in ( [ x12x0 ] )
one obtains @xmath509 this is the evolution equation for a spherically symmetric photon surface of the vaidya collapse metric ( [ vmetric ] ) .
consider the special case @xmath510 \lambda u & : 0\leq u \leq
u_1 \\[1ex ] m_1:= \lambda u_1 & : u_1<u<\infty \end{array } \right.\ ] ] for given constants @xmath511 , @xmath512 .
for @xmath513 the space - time is locally minkowskian , for @xmath514 there is inward falling null dust , and for @xmath515 the space - time is locally isometric to schwarzschild space - time with adm mass @xmath516 .
it is well - known there is a curvature singularity at @xmath20 and that for @xmath517 the part of this singularity at @xmath518 is locally naked .
schwarzschild photon surface through collapsing vaidya null dust .
the space - time coordinates of the evolved surface are plotted for @xmath519 ( above ) and @xmath520 ( below ) for various values of the null time @xmath512 of the junction between the null dust and schwarzschild regions . in each case
the evolved photon surface fails to intersect the minkowskian region @xmath513.,title="fig:",width=384 ] + schwarzschild photon surface through collapsing vaidya null dust .
the space - time coordinates of the evolved surface are plotted for @xmath519 ( above ) and @xmath520 ( below ) for various values of the null time @xmath512 of the junction between the null dust and schwarzschild regions . in each case
the evolved photon surface fails to intersect the minkowskian region @xmath513.,title="fig:",width=384 ] arising from the backwards evolution of the schwarzschild photon sphere .
any partial cauchy surface @xmath521 which extends to spatial infinity is cut into two components by @xmath22.,width=384 ] fix @xmath517 . for @xmath515 equation ( [ r2uv ] ) gives , as expected , that there is a photon surface at @xmath522 .
we seek to evolve this photon surface backwards in time , though the in - falling null dust , to obtain a maximally extended photon surface @xmath22 .
the boundary conditions are @xmath523 \frac{\mbox{\normalsize $ dr$}}{\mbox{\normalsize $ du$}}=0 \end{array } \right\ } \quad \text{at $ u = u_1 $ } \ .\ ] ] the results are shown in fig .
[ vaidyagraphs ] for @xmath524 and @xmath525 for selected values of @xmath526 .
one sees that in all cases @xmath22 tends in the past direction to a null hypersurface of the form @xmath527 .
the conformal diagram must therefore be of the form sketched in fig .
[ pendiag ] .
it is evident from fig .
[ pendiag ] that the naked central singularity is enclosed within the photon surface @xmath22 in the sense that any partial cauchy surface @xmath521 extending to spatial infinity must intersect @xmath22 in a 2-sphere .
the physical significance of this may warrant further investigation .
the definition of a photon surface given in section [ photonsurfacesection ] is valid in an arbitrary space - time .
however the result that a photon surface must have a second fundamental form which is pure trace indicates that a space - time must be specialized in some respect if it is to contain any photon surfaces . for spherically symmetric space - times there are always photon surfaces that respect the spherical symmetry . for space - times that are not spherically symmetric
, the definitions of a photon surface and @xmath121-invariant photon surface may seem too restrictive .
the problem is that , in general , one may not have orbiting null geodesics at a fixed radius . in kerr space - time for example , although there are orbiting null geodesics in the equatorial plane , those null geodesics which move in the direction of rotation do so at a different radius than those which move in the opposite direction .
but it seems implausible that a concept which is physically important in the case of exact spherical symmetry should become invalid when even a small amount of angular momentum is introduced .
a non - trivial generalization of the concepts of photon surface and @xmath121-invariant photon surface , at least to axially symmetric space - times , is thus required .
this research was supported by the nfr of sweden and nrf of south africa .
ve99b c. darwin , proc .
roy . soc . * 249 * , 180 ( 1958 ) . c. darwin , proc .
soc . * 263 * , 39 ( 1961 ) .
k. s. virbhadra and g. f. r. ellis , _ schwarzschild black hole lensing _ , astro - ph/9904193 . k. s. virbhadra and g. f. r. ellis , _ gravitational lensing by naked singularities _ ,
d. to be submitted .
s. w. hawking and g. f. r. ellis , _ the large scale structure of space - time _ ( cambridge university press , cambridge , 1973 ) .
b. carter , phys .
rev . * 141 * , 1242 ( 1966 ) .
r. h. boyer , proc .
a * 311 * , 245 ( 1969 ) .
h. stephani , _ general relativity .
an introduction to the theory of the gravitational field _
( cambridge university press , cambridge 1982 ) , p.222 . a. i. janis , e. t. newman and j. winicour , phys . rev .
lett . , 878 ( 1968 ) .
k. s. virbhadra , int .
j. mod .
a * 12 * , 4831 ( 1997 ) .
k. s. virbhadra , d. narasimha , and s. m. chitre , astron . & astrophys .
* 337 * , 1 ( 1998 )
. j. h. horne and g. t. horowitz , phys . rev .
d * 46 * , 1340 ( 1992 ) . | the photon sphere concept in schwarzschild space - time is generalized to a definition of a photon surface in an arbitrary space - time .
a photon sphere is then defined as an @xmath0-invariant photon surface in a static spherically symmetric space - time .
it is proved , subject to an energy condition , that a black hole in any such space - time must be surrounded by a photon sphere .
conversely , subject to an energy condition , any photon sphere must surround a black hole , a naked singularity or more than a certain amount of matter .
a second order evolution equation is obtained for the area of an @xmath1-invariant photon surface in a general non - static spherically symmetric space - time .
many examples are provided . *
the geometry of photon surfaces * + clarissa - marie claudel + _ mathematical institute , university of ume , s-901 87 ume , sweden _
+ k. s. virbhadra and g. f. r. ellis + _ department of applied mathematics , university of cape town , + rondebosch 7701 , south africa . _ |
uu aquarii is a bright eclipsing novalike ( p@xmath1 hr ) showing long - term brightness variations of 0.3 mags @xcite .
broad - band eclipse mapping indicate that its inner disc is optically thick , with an inferred mass accretion rate of m@xmath2 @xcite . it was suggested to be an sw sex star based on its relative flat radial temperature profile in the inner disc , single peaked asymmetric emission lines showing little eclipse , large phase shift between photometric and spectroscopic conjunction and orbital phase - dependent absorption in the balmer lines @xcite .
in this paper we report on the analysis of time - resolved spectroscopy of uu aqr with eclipse mapping techniques @xcite to derive spatially - resolved spectra of the accretion flow in this binary .
time - resolved optical spectroscopy ( @xmath3 35006900 ) covering 5 eclipses of uu aqr was obtained with the 2.1-m telescope at kpno on july - august 1993 at a time resolution of 50s .
the observations were performed while uu aqr was on its high brightness state .
the spectra were divided into 226 passbands ( 15 in the continuum and fainter lines , and @xmath4 across the most prominent lines ) and light curves were extracted for each passband .
average lightcurves for each passband were obtained by combining the 5 individual lightcurves .
the eclipse mapping method was used to obtain a map of the disc brightness distribution and the flux of an additional uneclipsed component in each passband .
the monochromatic maps were combined to produce spatially - resolved spectra .
fig.1 shows spatially - resolved spectra of the disc region in a logarithmic scale .
the lower panel shows the spectrum of the uneclipsed light .
the spectrum of the inner disc is characterized by a blue and bright continuum filled with deep and narrow absorption lines .
the continuum emission becomes progressively fainter and redder for increasing disc radius while the lines transition from absorption to emission showing clear p cygni profiles on all lines mapped at higher spectral resolution .
the balmer jump appears in absorption in the inner disc and weakly in emission in the intermediate and outer disc regions , suggesting that the outer disc in uu aqr is optically thin .
the change in the slope and intensity of the continuum with increasing disc radius reflects the temperature gradient in the accretion disc , with the effective temperature decreasing outwards .
spatially resolved disc spectra are shown in fig.2 as a function of velocity for the h@xmath5 , h@xmath6 and h@xmath7 lines .
the absorption lines at disc center are perceptibly narrower and deeper than expected for emission from either the white dwarf atmosphere or from disc gas in keplerian orbits around the white dwarf , and the lines show larger widths at the outer disc .
this is in clear disagreement with the expected behaviour of line emission from gas in a keplerian disc and is an evidence that most of the line emission do not arise from the disc atmosphere . on the other hand ,
the lines at intermediate and outer disc regions ( @xmath8 ) show clear p cygni profiles indicating origin in an outflowing gas , probably the disc wind
. the spectrum of the uneclipsed light shows prominent balmer and hei emission lines ( fig.1 ) . the balmer jump ( and possibly also the paschen jump ) is clearly in emission indicating that the uneclipsed light has an important contribution from optically thin gas from outside the orbital plane .
the fractional contribution of the uneclipsed light is very significant for the emission lines , reaching 40 - 60% at the balmer lines and 20 - 40% at the hei lines , and decreases steadily along the balmer serie .
the difference in fractional contribution between the balmer and hei lines indicates the existence of a vertical temperature gradient in the material above / below the disc , with the hei lines ( which require higher excitation energies ) being produced closer to the orbital plane . in all case ,
a substantial fraction of the light at these lines does not arise from the orbital plane and is not occulted during eclipse .
fig.3 shows the ratio between the spectrum of the gas stream region and the disc region at same radius as a function of radius .
the comparison reveals that the spectrum of the gas stream is noticeably different from the disc spectrum in the outer disc regions ( where one expects a bright spot to form due to the initial shock between the inflowing stream and the outer disc rim ) but is also different in the inner disc regions .
this result suggests the existence of gas stream penetration in uu aqr , which can be seen down to @xmath0 .
the spectrum of the ratio becomes redder for decreasing disc radius , possibly a combination of the disc emission becoming bluer as one moves inwards and the gas stream emission becoming redder while its energy is continuously lost in the shock with disc material along the inward stream trajectory . | time - resolved spectroscopy of the novalike variable uu aquarii is analyzed with eclipse mapping techniques to produce spatially resolved spectra of its accretion disc and gas stream as a function of distance from disc centre in the range 3600 - 7000 .
the spectrum of inner disc shows a blue continuum filled with deep , narrow absorption lines which transition to emission with clear p cygni profiles at intermediate and large radii .
the spectrum of the uneclipsed component has strong hi and hei emission lines and balmer jump in emission and is explained as optically thin emission from a vertically extended disc wind .
most of the line emission probably arises from the wind .
the spatially - resolved spectra also suggest the existence of gas stream penetration in uu aqr , which can be seen down to @xmath0 . #
1#2#3#4#5#6#7 to#2 ' '' '' -15 mm |
a conservative estimate of the mass of the galaxy out to a distance of about 2/3 of that of the large magellanic cloud is @xmath5 ( fich and tremaine 1991 ) . with a total luminosity in the v - band of @xmath6 ( binney and tremaine 1987 )
the galactic mass to light ratio in v ( @xmath7 ) out to @xmath8 kpc is @xmath9 .
since normal stellar populations do not generally produce @xmath10 ratios higher than about 3 , this is usually taken as evidence for an important component of dark matter within an extended halo surrounding the galaxy .
a related result is that of the macho microlensing experiment in the direction of the magellanic clouds which seems to indicate that about 20% ( @xmath11% is the 90% ci ) of the dark matter in the galaxy is tied up in objects with masses near @xmath12 ( 0.3 - 0.9 is the 90% ci ) ( alcock 2000 ) .
this assertion is only true if the bulk of the machos are located in the halo of our galaxy .
moreover , recent studies suggest that objects with masses from @xmath13 to @xmath14 are now largely excluded ( so planets and/or brown dwarfs are not important dark matter candidates ) and that a 100% macho halo is no longer a viable model ( alcock 2000 ) . macho masses near @xmath12 are suggestive of white dwarfs although neutron stars or even primordial black holes remain as unlikely but still possible candidates .
a halo currently consisting of 20% old white dwarfs implies that the precursors of these @xmath12 objects would have accounted for @xmath15 of @xmath16 in the universe .
this estimate comes from taking @xmath17 as the precursor masses , @xmath18 for @xmath19 , 0.04 for @xmath20 and assuming that the galactic halo is a fair sample of the types of mass seen in the universe .
it may be that galactic halos are baryon - rich in which case the derivation of 40% would be an over - estimate .
this may also be an interesting way to constrain the masses of the precursors as clearly in this analysis they could not exceed @xmath21 without violating the observed baryonic content of the universe .
even given all the uncertainties in these estimates , it is clear in this scenario that an important fraction of all the baryons in the universe were funnelled through this star formation mode .
this will surely have implications for star formation scenarios in the early universe and perhaps also on the manner in which galaxies were assembled and how they enriched themselves in heavy elements .
the interpretation of the microlensing results is , however , not without controversy as there are indications that at least some of the lenses may reside in the magellanic clouds themselves ( mcgrath and sahu 2000 ; sahu 2002 ) .
this would have the effect of lowering both the lens masses ( to @xmath22 ) and the contribution of the machos to the galactic dark matter budget .
the lower lens masses would be in line with expectations from mass function considerations and the lower macho dark matter contribution would alleviate such problems as the appearance of high redshift galaxies ( at an epoch when the white dwarf precursors were luminous ) , the chemical evolution of galaxies ( the white dwarf precursors producing too much helium and heavier elements ) and the peculiar mass function required of the precursors ( a mass function truncated at both high and low masses to avoid too much chemical enrichment from high mass stars and too many halo white dwarfs currently evolving from the low mass ones [ chabrier 1999 ] ) .
there is also the possibility that most of the machos are located in a thick ( gates _ et al .
_ 1998 ) or flaring disk ( evans _ et al .
_ 1998 ) of the galaxy .
this would have the effect of reducing their scale height and again their total galactic mass contribution .
the definitive answer as to whether old white dwarfs are important baryonic dark matter contributors will come from searches which will ( or will not ) identify significant numbers of high velocity local examples of the very cool white dwarfs that produce the microlensing .
such searches are now underway and in the ensuing sections i will discuss early results from them .
just recently , a new era opened in the study of old white dwarfs .
hansen ( 1998 , 1999 ) first calculated emergent spectra from cool ( @xmath23k ) white dwarfs with atmospheres that included opacity from the @xmath24 molecule .
similar models were also constructed by saumon and jacobson ( 1999 ) and earlier ones by bergeron , saumon and wesemael ( 1995 ) for objects down to 4000k .
all these models exhibited very strong collisionally induced opacity in the red and near ir part of the spectrum and the effects of this opacity can be seen clearly in the model spectrum shown in figure 1 as well in several dozen real objects ( see e.g. oppenheimer _ et al_. 2001a ) .
the broad bands seen in the spectrum in figure 1 are due to @xmath24 but it is of interest to point out that no cool white dwarf has yet been observed that actually shows these spectral features . perhaps it is because no objects this cool have been discovered up until now , or that the models are still somehow deficient , or that possibly the universe is not old enough for hydrogen - rich white dwarfs to have reached such low temperatures .
the effects of collisionally induced absorption are not limited , however , to extremely cool objects .
its influence can been seen in white dwarfs at least as hot as 5400k ( see , e.g. , the spectrum of lhs 1126 in bergeron _
1994 ) .
the effect of the @xmath24 opacity is to force the radiation out in the bluer spectral regions .
this has an enormous influence on their colors , the location of the white dwarf cooling track in the color - magnitude diagram , and hence their observability . down to a temperature of about 4000k ( which corresponds to an age of about @xmath25 for @xmath12 white dwarfs )
these stars become increasingly redder as they cool , the reddest color they achieve is about @xmath26 in hansen s ( 1998 , 1999 ) models . note that other sets of cooling models give somewhat different colors and ages here ( e.g. fontaine _ et al .
older white dwarfs become progessively bluer in @xmath27 as they continue to cool .
ancient white dwarfs , of age @xmath28 , and mass @xmath12 , have @xmath29 according to these models .
as has been noted several times in the past few years , _ old hydrogen - rich white dwarfs are blue , not red , in @xmath30 colors .
_ searches for such ancient objects are now concentrating on bluish objects in these colors .
a cooling track from hansen s ( 1998 , 1999 ) models illustrating all these features for @xmath12 white dwarfs is shown in figure 2 .
note that instead of getting arbitrarily faint with age , old hydrogen - rich white dwarfs seem eventually to coast at approximately constant @xmath31 for ages between about 10 and 14 gyr , just the timescales which may bracket globular cluster ages .
the implication is clear @xmath4 very old white dwarfs may be easier to detect than was previously thought .
also potentially of interest is that the @xmath27 color now moves the bulk of the ancient white dwarfs away from the swarm of field k and m - dwarfs ( making confusion with such objects @xmath32 likely ) but at the same time overlapping the colors of distant star - forming galaxies ( making the confusion with such objects @xmath33 likely ) .
the solution here will be to search for old white dwarfs using both color and proper motion criteria , an approach taken by most current programs .
one problem with these models is that they are largely untested . of serious concern
is that none of the cool white dwarfs found to date ( see following sections ) seem to exhibit the extreme blue colors predicted by most of these current models containing @xmath24 in their atmospheres .
this may be suggesting missing or inadequate physics .
one place where these models could potentially be tested is among the cool white dwarfs in globular clusters .
such systems should contain white dwarfs almost as old as the clusters themselves ( likely in the @xmath34 gyr range ) , so the extreme effects of collisionally induced opacity ought to be clearly seen .
figure 3 shows the current situation here .
it displays the deepest color magnitude diagram yet obtained for a globular cluster ( richer _ et al .
_ 1995 , 1997 ) , an hst study of m4 .
white dwarfs perhaps as faint as @xmath35 are recorded in this study .
if these objects have masses near @xmath12 , they will have ages in excess of 7 gyr and @xmath36 in hansen s ( 1998 , 1999 ) models ; cool enough to exhibit some of the effects of collisionally induced opacity but not old enough to help constrain the appearance of truly ancient white dwarfs . in hst cycle 9 , a group of colleagues and i were awarded time on hst to attempt to locate the termination point of the white dwarf cooling sequence in m4 . with the current generation of models and a cluster age in the range of @xmath34 gyr ,
this should occur near a v magnitude of about 30 , certainly a enormous challenge .
it is not clear whether the data will actually go this deep , but even if it does not , we should be in a position to provide some stringent tests of the current generation of models . in figure 4
i attach a simulation of what we expect to see with these data .
this is an optimistic simulation in that it does not include the effects of charge transfer inefficiency now known to be plaguing the hst wfpc2 ccds and also assumes that all the data are taken under low sky brightness conditions .
nevertheless , even if the faintest magnitudes are not reached , the effects of collisionally induced opacity should be readily apparent and quantitatively testable with these data .
a number of searches for an ancient population of halo white dwarfs are currently underway or have been completed .
most of these involve color and proper motion selection with some follow - up spectroscopy .
the proper motions have generally been derived from digitized photographic plates with quite long time baselines ( typically @xmath37 years ) .
i list below in table 1 a compilation of these surveys together with some additional information .
this table must be considered a rough guide only to the current situation .
however , it is indicative of the present state of our knowledge . in table 1
the first two columns give the name and the area searched for each program .
the third column is the limiting v magnitude of each survey .
this is a fairly crude estimate as some ( eg oppenheimer _
2001b ) did not survey in the v - band and it was necessary to transform their particular magnitude to v so that a homogeneous set of statistics for each cluster could be derived .
column 4 is an estimate of the distance probed in each survey for old white dwarfs .
it is the distance out to which ancient white dwarfs could have been detected in the individual surveys .
this limiting distance for detection of old white dwarfs was set to @xmath38/5)})$ ] pc ; 17 was taken as the @xmath31 of a typical old white dwarf ( hansen 1998 , 1999 ; saumon and jacobsen 1999 ; richer _ et al .
column 5 lists the number of thick disk and spheroid white dwarfs that are expected out to the limiting distance of the survey .
this number was derived following the prescription of reid _ et al . _
( 2001 ) and is dominated by the thick disk contribution .
thin disk white dwarfs are not c c c c c c c opp & 4165 & 19.8 & 36 & 3.9 & 3.7 & 4 + monet & 1378 & 20 & 40 & 2.5 & 2.4 & @xmath39 + ibata & 790 & 20 & 40 & 1.4 & 1.4 & 2 + sdss & 400 & 20.5 & 50 & 1.4 & 3.5 & 2 + jong & 2.5 & 24 & 250 & 1.1 & 3.6 & 3 + cfht & 15 & 25 & 400 & 27 & 88 & ?
+ hdf & @xmath40 & 28 & 1600 & 0.2 & 0.5 & 0 ?
+ considered in this discussion as they would be eliminated by their low velocities .
we are only concerned here with high velocity objects which could contaminate a true dark halo sample .
column 6 contains the number of observable white dwarfs expected from the dark halo ( richer _ et al .
_ 2000 ) assuming that the halo consists of 20% by mass of machos which are old @xmath12 white dwarfs , half of which have hydrogen - rich outer atmospheres .
the helium - rich ancient white dwarfs would have cooled to very low luminosities by a hubble time and would be basically unobservable .
the seventh column lists the number of old white dwarfs actually found * out to the limiting distance only * in each survey .
it is clear from examination of table 1 that in all existing surveys thus far , there is no obvious excess in the discovered number of old white dwarfs over that expected from the thick disk and spheroid .
note , however , that the cfht program , currently acquiring its second epoch of data , should be extremely important in this context .
it is quite different from most of the other surveys as it is much deeper ( @xmath41 ) .
one caveat must be kept in mind when perusing table 1 , however . in comparing , for example
, columns 5 through 7 the assumption of 100% recovery of all moving objects is built in .
if , for example , a survey was only 50% complete in recovering moving objects , then the number in column 7 should be increased by a factor of 2 .
for some surveys , such as the sdss , where proper motions are not part of the selection criteria , the numbers in table 1 are strictly correct . however , until estimates for proper motion limits are accurately known , there remains no clear evidence of a white dwarf dark matter halo signature in any of the current surveys . of course , at the moment
, it is also true that the surveys do not rule out such a dark matter halo either .
we can examine the totality of objects in the surveys of table 1 from a different perspective .
if they are , in fact , local examples of the machos , and thus important galactic dark matter contributors , they should be quite old , likely as old as the globular clusters .
this would place them in the @xmath34 gyr range if we take the current spread quoted for globular cluster ages . from the location of the white dwarfs in color - magnitude and color - color diagrams ,
temperatures can be inferred with the use of theoretical models . from these temperatures , ages
can then be assigned using models of cooling white dwarfs after choosing an appropriate mass . for the purposes of this exercise
, i used the models of hansen ( 1998 , 1999 ) and those discussed in oppenheimer _
( 2001b ) with white dwarf masses of @xmath12 , to estimate the temperature and cooling time of each object found in the various surveys . note that we are not limiting ourselves to objects within @xmath42 here .
the cumulative distribution of temperatures is shown in figure 5 and the time for a @xmath12 white dwarf to cool to a given temperature is also indicated by the vertical lines .
the immediate conclusion from the plot is that the sample in table 1 is _ not _ that of a halo sample , the objects are simply too young .
it seems more in line with that of a thick disk component which dominates the predicted numbers in table 1 .
this result is , however , strongly model dependent .
a different set of cooling models or mass would provide quite a different result .
for example , if i had chosen to assign a mass of @xmath43 to all the stars , the coolest object would have a cooling age of @xmath44 gyr , almost @xmath45 gyr older than a @xmath12 white dwarf .
further , the main sequence lifetimes of the white dwarf precursors are not included in the cooling ages .
if the masses of the precursors are in the range of @xmath17 this will add about 1 gyr to these numbers .
nevertheless , the general conclusion to be drawn from both table 1 and figure 5 is that within the context of the present generation of white dwarf cooling models , truly ancient white dwarfs do not currently appear to have been discovered .
the last row in table 1 deserves some special comments . based on 2 epochs of exposures taken in the hdf separated by 2 years , ibata _ et al_. ( 1999 ) claimed to detect 2 extremely faint objects with significant proper motion .
one of these appeared to have colors consistent with that expected from a very cool white dwarf . with lower statistical significance
, 3 other objects also appeared to be moving , one of which possessed colors expected from an old white dwarf .
the inferred space velocities for these objects placed them squarely in the category of halo objects .
this result was potentially very important as the detection of even a few such objects in the hdf would be enough to account for the entire macho contribution to the galactic dark matter .
the second epoch data were , however , taken in the non - optimal f814w filter for another program ( supernova search ) , and were quite noisy .
based on this seemingly important result , a third epoch was secured in f606w , fully 5 years after the original hdf data .
these images were of significantly higher quality .
one of the objects suggested by ibata _
( 1999 ) as possibly moving was number 4 - 141 in the williams _ et al . _
( 1996 ) compilation . using a maximum likelihood technique , ibata _ et al . _
( 1999 ) measured a motion of 35 mas / yr with an error of 8 mas / yr . with the current superior data set we measure a proper motion of @xmath46 mas / yr , a proper motion which is not statistically significant .
none of the other 4 objects was found to possess significant motion over the 5 year baseline .
a full discussion of our reanalysis of moving objects in the hdf will appear in richer _ et al .
my personal view of the current landscape regarding the viability of old white dwarfs as a component of the galactic dark matter is that while there are some interesting hints , there is no compelling evidence in the current statistics either for against the scenario .
perhaps a listing of the critical issues will be instructive . 1 .
_ where are the lenses in the direction of the lmc and smc located ?
_ this is a fundamental issue @xmath4 perhaps the most important one .
there are suggestions that some , perhaps even most , of the lenses are in the clouds themselves .
this issue may be settled in a number of ways , some of which are possible with current technology .
one road to a solution is to observe along a different line of sight @xmath4 a microlensing campaign in the direction of m31 or m33 would be extremely useful here .
there are programs currently underway looking at m31 ( e.g. uglesich _ et al .
_ 1999 ) but no comprehensive results are available at this time
. some future satellite missions may be capable of detecting parallaxes ( and masses ) for the lenses ( gould and salim 1999 ) and this would be critical data for establishing their location . if the lenses are in the clouds and are low mass main sequence stars , then it may eventually be possible to observe them directly .
_ is there a population of cool white dwarfs with halo kinematics that are present in numbers which exceed those expected from the thick disk and spheroid ?
_ as i have discussed , based solely on number count arguments , the current searches do not as yet provide a conclusive answer either in the affirmative or negative .
deep ( @xmath47 ) proper motion and color selected surveys covering 10 degrees or more of sky are required to be definitive .
there are several such searches underway and the results are eagerly anticipated .
3 . _ do we have the last word on the models of old white dwarfs ?
_ there are 2 aspects of the present theory of cooling white dwarfs which i find disturbing .
first , there are enormous problems in fitting the spectra of these cool white dwarfs with the current generation of models .
the situation is so dire that bergeron ( 2001 ) , one of the premier practitioners of this art , was required to insert a fictitious opacity source into his models to help in the fits .
he then went on to state it now remains to be seen whether this new opacity source , even when included with the approximate treatment introduced in this paper , can help resolve the mystery surrounding several ultracool white dwarfs whose energy distributions have yet failed to be successfully explained in terms of hydrogen or mixed hydrogen / helium compositions . " the second concern is simply that no truly ultracool white dwarfs have as yet been discovered , or , perhaps better put , no white dwarfs are known whose colors and/or spectra suggest that the object s age is in the range of @xmath34 gyrs .
this may be the same problem as the first one mentioned here , that is that the models are still somehow inadequate , or it may be suggesting that such objects are extremely rare or just do not exist .
it is within this context that extremely deep observations of globular cluster white dwarfs will be important .
the author would like to thank the space telescope science institute for organizing a meeting that was all that a scientific should be ; exciting new results , participants with deep and wide interests , an excellent venue .
in particular , deepest gratitude is due to mario livio who organized the meeting and provided one of the best conference summaries that this author has ever experienced .
alcock , c. 2000 , science 287 , 74 bergeron , p. 2001 , astro - ph/0105333 , to be published in the letters ,
september 10 bergeron , p. , ruiz , maria - teresa , leggett , s.k . ,
saumon , d. & wesemael , f. 1994 , 423 , 456 bergeron , p. saumon , d. & wesemael , f. 1995 , 443 , 764 binney , j. & tremaine , s. 1987 , galactics dynamics , princeton series in astrophysics , princeton university press .
chabrier , g. 1999 , 513 , l103 de jong , j. , kuijken , k. & neeser , m. 2000 , astro - ph/0009058 evans , n.w . ,
gyuk , g. , turner , m.s .
, & binney j. 1998 , 501 , l45 fich , m. & tremaine , s. , 1991 , araa 29 , 409 fontaine , g. , brassard , p. & bergeron , p. 2001 , pasp 113 , 409 gates , e.i . ,
gyuk , g. , holder , g.p . , & turner , m.s .
500 , l145 gould , a. & salim , s. 1999 , 524 , 794 hansen , b.m.s . , 1998 , nature 394 , 860 hansen , b.m.s . , 1999 , 520 , 680 harris , h.c .
; hansen , b.m.s . ,
liebert , j. , vanden berk , d.e . ,
anderson , s.f . ,
knapp , g.r . ,
fan , x. , margon , b. , munn , j.a . ,
nichol , r.c . , pier , j.r . ,
schneider , d.p . ,
smith , j.a . ,
winget , d.e . , york , d.g . ,
anderson , j.e . , jr . , brinkmann , j. , burles , s. , chen , b. , connolly , a.j . ,
csabai , i. , frieman , j.a . , gunn , j.e . ,
hennessy , g.s . ,
hindsley , r.b . ,
ivezic , z. , kent , s. , lamb , d.q . ,
lupton , r.h . ,
newberg , h.j .
, schlegel , d.j . ,
smee , s. , strauss , m.a . ,
thakar , a.r . ,
uomoto , a. & yanny , b. 2001 , 549 , l109 ibata , r. , richer , h.b . ,
gilliland , r.l .
& scott , d. 1999 , 524 , l95 ibata , r. , irwin , m. , bienayme , olivier , scholz , r. & guibert , j. 2000 , 532 , l41 mcgrath , e.j . &
sahu , k.c . , 2000 ,
aas meeting 197 , # 105.09 monet , d.g . , fisher , m.d .
, liebert , j. , canzian , b. , harris , h.c . & reid , i.n .
2000 , 120 , 1541 oppenheimer , b. r. , saumon , d. , hodgkin , s. t. , jameson , r. f. , hambly , n. c. , chabrier , g. , filippenko , a. v. , coil , a. l. & brown , m.e .
2001a , 550 , 448 oppenheimber , b.r . , hambly , n.c . ,
digby , a.p . , & saumon , d. 2001b , science 292 , 698 reid , i.n . ,
sahu , k.c . & hawley , s.l .
2001 , astro - ph/0104110 , to be published in richer , h.b .
, fahlman , g.g . ,
ibata , r.a . ,
stetson , p.b . ,
bell , r.a .
, bolte , m. , bond , h.e . ,
harris , w.e . ,
hesser , j.e . ,
mandushev , g. , pryor , c. & vandenberg , d.a .
1995 , 451 , l17 richer , h.b . ,
fahlman , g.g . ,
ibata , r.a . , pryor , c. , bell , r.a . ,
bolte , m. , bond , h.e . ,
harris , w.e .
, hesser , j.e . , holland , s. , ivanans , n. , mandushev , g. , stetson , p.b . & wood , m.a .
1997 , 484 , 741 richer , h.b . ,
hansen , b. , limongi , m. , chieffi , a. , straniero , o. & fahlman , g.g .
2000 , 529 , 318 richer , h.b .
_ 2002 , in preparation sahu , k.c . , 2001 , see paper in this volume saumon , d. & jacobson , s. , 1999 , 511 , l107 uglesich , r. , crotts , a.p.s .
, tomaney , a.b . & gyuk , g. 1999 , aas meeting 194 , # 90.04 williams , r.e . , blacker , b. , dickinson , m. , dixon , w. , ferguson , h.c . ,
fruchter , a.s . ,
giavalisco , m. , gilliland , r.l . , heyer , i. , katsanis , r. , levay , z. , lucas , r.a . , mcelroy , d.b . ,
petro , l. , postman , m. , adorf , h - m . ,
hook , r. 1996 , 112 , 1335 | a straightforward interpretation of the macho microlensing results in the direction of the magellanic clouds suggests that an important fraction of the baryonic dark matter component of our galaxy is in the form of old white dwarfs .
if correct , this has serious implications for the early generations of stars that formed in the universe and also on the manner in which galaxies formed and enriched themselves in heavy elements .
i examine this scenario in some detail and in particular explore whether the searches currently being carried out to locate local examples of these machos can shed any light at all on this scenario .
= # 1to 0pt#1 = # 1*[@xmath0 # 1 ] * # 1*[@xmath1 # 1 ] * # 1*[@xmath2 # 1 ] * # 1*[@xmath3 # 1 ] * department of physics & astronomy , university of british columbia 6224 agricultural road , vancouver , b.c .
, v6 t 1z1 , canada electronic mail : [email protected] to appear in proceedings of the dark universe : matter , energy and gravity " held at the space telescope science institute april 2 @xmath4 5 2001 . |
carbon nanotubes are the subject of intensive research last several years @xcite . their electrical and thermal conductivity , as well as mechanical properties are of great scientific and technological interest .
depending on the diameter and the helicity carbon nanotubes can be either metallic or semiconducting , and exhibit quantum wire properties .
therefore , they are important for nanoelectronic devices .
promising applications of carbon nanotubes are also based on their extraordinary mechanical properties .
investigations have shown that carbon nanotubes are among strongest materials @xcite .
the effects of deformations and various defects are studied for single - wall and multi - wall tubes , as well as for ropes composed of carbon nanotubes @xcite .
it was found that very distorted carbon nanotubes return to their original form when loading is released . as a consequence ,
carbon nanotubes are useful in applications , such as high - strength fibers , components of composite materials and tips in scanning - probe microscopy @xcite .
the property of carbon nanotubes to deform elastically is in part the outcome of the strength and rigidity of the graphitic bond .
the cylindrical structure of nanotubes increases their elasticity and strength .
it is important to investigate mechanical properties for cylindrical nanostructures made of other materials where a different type of bonding exists .
these investigations may shed light onto the choice of optimal materials and structures for elements in nanomechanical devices .
metallic wires are important for applications .
for example , gold wires are already used as interconnections in the chips .
the miniaturization of electronic and mechanical devices requires investigations of very small wires whose diameters and lengths are in the range from @xmath2 nm to @xmath3 nm .
recently a technique for fabrication of metallic nanowires with diameters down to @xmath4 nm was invented @xcite .
this method is based on the thickness resolution of molecular - beam epitaxy and formation of extremely precise templates for metal deposition .
the mechanical properties of metallic nanowires till now were investigated because of the experiments dealing with the contact between two pieces of metal .
for example , metallic nanobridges were obtained in scanning tunneling microscopy ( stm ) experiments @xcite , mechanically controllable break - junctions @xcite , and macroscopic table top metallic contacts @xcite . in this context the stability of tip - suspended nanowires , their behavior under tensile strain , fracture properties , and conductance were investigated by molecular dynamics ( md ) simulations and density functional theory calculations @xcite .
mechanical properties of metallic nanocontacts under compressive strain are less studied @xcite .
however , compression of free single - wall carbon nanotubes was investigated by md simulations @xcite .
multi - shell structures were obtained in md simulations of finite and infinite gold nanowires @xcite .
finite nanowires with radii around a nanometer and of a length / diameter ratio between @xmath2 and @xmath5 were studied @xcite .
cylindrical multi - shell structures were found for a length / diameter ratio between @xmath6 and @xmath5 .
structures and melting of infinite gold nanowires with radii around a nanometer and an initial orientation along the ( 111 ) , ( 110 ) , and ( 100 ) directions were also investigated @xcite .
the results have shown that a formation of cylindrical shells is the most pronounced for an initial fcc ( 111 ) orientation .
similar gold nanostructures were found in experiments @xcite .
these studies are enabled by recent advances in microscopic techniques which now produce images of atomic scale resolution .
gold nanostructures were prepared by contacting a gold substrate with a stm tip , and they were simultaneously imaged using an ultrahigh - vacuum electron microscope @xcite .
the conductance of these gold nanostructures was measured during retraction of the stm tip .
it was found that the conductance is quantized in units of @xmath7 , where @xmath8 is the electron charge and @xmath9 is the planck s constant @xcite .
stable and regular strands of gold atoms that are from one nanometer to several nanometers in length were observed .
the diameters of these wires were around one nanometer . in md simulations of pb and al ultrathin infinite wires at @xmath0 k several exotic structures were found , for example icosahedral and helical forms @xcite .
similar simulation has been carried out for gold nanowires @xcite .
aluminum and copper infinite nanowires were also simulated at the room temperature and multi - shell and filled structures were obtained @xcite . recently zach and coworkers have developed a method for producing metallic wires with diameters ranging from @xmath10 nanometers to @xmath11 micrometers and lengths of up to half a millimeter @xcite . they were using the graphite substrate as a template for growing molybdenum wires .
the mechanical strength of these nanowires was tested and it was found that they were able sometimes to bend at 90 degrees without breaking . in this work a md simulation study of free multi - shell gold nanowires under axial compression at @xmath1 k is presented .
our results show that these gold nanostructures are able to sustain very large values of the strain and recover initial radii and lenghts even after severe deformations . in the following the interatomic potential , the wires preparation method , and simulation details
are described in sec .
results and discussion are presented in sec .
a summary and conclusions are given in sec .
the interaction between gold atoms was modeled by the well - tested embedded - atom potential @xcite .
it is known that embedded - atom potentials provide a satisfactory description of metallic bonding in the bulk , for surfaces , and nanoparticles @xcite . in comparison with pair interactions , such as the lennard - jones and morse potentials
, the embedded - atom model takes an account of many - body effects in metals .
the particular realization of the embedded - atom potential used here , the so - called glue model , has been shown to accurately reproduce experimental values for a wide range of physical properties of gold @xcite .
the classical md simulation method was carried out .
the equations of motion were integrated using a time step of @xmath12 s. in all simulations presented in this work the temperature control was realized through velocity rescaling of all active atoms .
it is known that bulk gold crystallizes in a face - centered cubic lattice .
initially , for nanowires at @xmath0 k , atoms were arranged in a fcc structure with the ( 111 ) direction parallel to the axis of the wire .
the prepared nanowires were approximately cylindrical in cross - sections .
all atoms further than a chosen value of the radius from the nanowire axis were removed .
several nanowires with the length @xmath13 between @xmath14 nm and @xmath15 nm , the number of atoms between @xmath16 and @xmath17 , and the radius @xmath18 between @xmath19 nm and @xmath20 nm were investigated .
all wires were first relaxed at @xmath0 k. then samples were heated to @xmath21 k. this was followed by a quench to @xmath22 k , and heating to @xmath1 k. in previous simulations , where structural and vibrational properties of finite gold nanowires were investigated , md boxes in the annealing cycles were heated to @xmath23 k @xcite . however , it was found here that even finite nanowires are robust , and preserve their cylindrical structures when heated up to @xmath24 of the bulk melting temperature .
properties of nanowires under compression were investigated after evolution of @xmath25 time steps at @xmath1 k. as in previous simulations for finite @xcite and infinite gold nanowires @xcite , after a preparation procedure stable multi - shell cylindrical structures were obtained . in comparison with previous simulations of finite gold nanowires ,
a higher annealing temperature applied here produces more regular multi - shell structures .
after equilibration , as in simulations of single - wall carbon nanotubes @xcite , the edge rings were fixed , and then compressive axial loading was applied .
these simulations under compression most often were taken out to @xmath26 time steps .
it was checked that typical deformation patterns formed during this time did not change up to @xmath27 time steps .
several structures of multi - shell wires at @xmath28 k and before compression was applied are presented in fig .
1 . figure 1(a ) shows a three - shell structure with the central core .
this type of the multi - shell configuration with a single central strand of gold atoms forms most often . a nanowire with the smallest investigated radius forms a two - wall structure with a large empty core shown in fig . 1(b ) .
the core of two strands of atoms , as in the four - wall structure shown in fig .
1(c ) , was also obtained .
short and narrow nanowires ( e.g. , @xmath29 nm , @xmath30 nm ) under compression exhibit only one morphological pattern .
they progressively shorten when compression increases , but remain straight . as an exception , long and extremely narrow nanowires ( e.g. , @xmath31 nm
, @xmath32 nm ) under compression often crush into irregular ellipsoidal morphologies .
thicker short nanowires ( e.g. , @xmath33 nm , @xmath34 nm ) show three morphological patterns .
for small compressions these nanowires deform by rippling .
however , rippling is less pronounced than in single - wall carbon nanotubes @xcite , and only isolated ripples appear . at @xmath35 gpa of the stress one end of the nanowire deforms more than the other .
such a deformation is shown in fig .
2 . for larger stress ( @xmath36 gpa )
these nanowires also deform by crushing , flatten and preserve their cylindrical shapes .
nanowires return to the configuration with approximately initial @xmath37 and @xmath38 when compression is released .
the outer cylindrical surface of those straight wires is more rough than in initial ones . the most interesting behavior exhibit long and narrow nanowires [ e.g. , @xmath39 nm , @xmath32 nm , shown in fig .
these slender nanowires under compression buckle and sometimes exhibit largely distorted configurations .
figure 3(b ) shows deformation of the nanowire in fig .
3(a ) at @xmath40 gpa of the stress .
the structure formed in this sideways deformation is almost symmetric .
figure 4(a ) shows a configuration of the same nanowire at @xmath41 gpa of the stress .
although , deformed substantially [ fig .
4(a ) ] , this structure immediately straightens [ fig .
4(b ) ] when compression is released .
this structural change already takes place within @xmath42 ps of a time evolution after loading is released . the resulting configuration ( shown in fig .
4(b ) after @xmath43 ps , i.e. , @xmath44 time steps ) is less regular than initial nanowire , but has the same average radius and length . in further simulations at the same temperature
this nanowire does not change substantially up to @xmath45 ns . the same stress of @xmath46 gpa applied on the structure shown in fig .
4(b ) produces a buckled morphology similar to one shown in fig .
when loading is released in this repeated compression , a nanowire again returns to the straight form similar to one shown in fig .
4(b ) . large buckling deformations [ as shown in figs .
3(b ) and 4(a ) ] exist for average compressions .
for the highest values of the stress slender nanowires deform by crushing and flattening , and remain straight .
these simulations show that a multi - shell structure disappears under compression .
the density plots in fig .
5 illustrate the changes in a multi - shell structure under compression . for the stress of @xmath47 gpa
the shell structure is still present [ fig .
5(b ) ] , whereas at @xmath40 gpa the shells are much less pronounced [ fig .
the shell structure is almost absent at @xmath41 gpa [ fig .
5(d ) ] . when compression is released nanowires straighten , but their internal multi - shell structure only partially recovers in the simulation of @xmath48 time steps .
disordering of a multi - shell structure is more pronounced in longer nanowires and for a higher applied stress . as in initial annealing cycles , it is possible to improve this partially ordered multi - shell structure by heating . in multi - shell
gold nanowires atomic rearrangements occur for the smallest applied stress .
in contrast , carbon nanotubes up to certain values of stress are in the elastic regime .
there virtually no defects are found after compressive loading is released @xcite . beyond the elastic regime carbon nanotubes deform plastically . in the plastic regime , as here for gold nanowires , compressed carbon nanotubes recover from severe structural deformations , but local atomic rearrangements exist . a plot of the stress vs. the strain for several nanowires is presented in fig .
all shells of multi - wall nanowires are stressed and it is assumed that the axial strain and stress are uniformly distributed over the cross section .
therefore , in calculation of the stress the whole cross - sectional area of the nanowire is used .
these results show that multi - shell gold nanowires are able to sustain a large compressive stress .
multi - shell gold nanowires at large compressions behave in similar way as ropes of single - wall carbon nanotubes under pressure @xcite .
these ropes of nanotubes act as a mechanical energy storage .
this was attributed to crushing and flattening of the tube cross section .
a flattening along the wire axis was found in this simulation for gold nanowires .
those morphological patterns correspond to the vertical portions of the curves in fig .
6 . therefore , because of the property to recover their shapes even for large values of the strain , multi - shell gold nanowires also act as a mechanical energy storage .
although the chemical bonding in carbon nanotubes and gold nanowires is different , the packing effects at nanometre length scales are also important .
regular axially symmetrical distribution of carbon nanotubes in ropes and cylindrical gold shells in nanowires enable them to store a mechanical energy .
molecular dynamics simulation was carried out to study deformations of multi - shell gold nanowires under axial compressive loading at @xmath1 k. this simulation is based on a well - tested embedded - atom potential .
it was found that multi - shell nanowires are able to sustain large values of the compressive strain .
the most interesting behavior show long and narrow nanowires where large buckling distortions without a failure are possible when axial compression is applied and then released .
the property of carbon nanotubes to recover without a damage after large deformations was explained as a consequence of the carbon bonding in the graphite layer and the cylindrical structure of the tube .
the simulation presented here shows that multi - shell gold nanowires after large deformations recover their cylindrical forms , as well as initial radii and lenghts . a similar result ( i.e , that large bending deformations without breaking are possible ) was recently obtained experimentally for molybdenum nanowires @xcite .
in contrast to carbon nanotubes , because of different type of chemical bonding in gold , atomic rearrangements occur even under small compressions . as a result ,
defects and local structural changes are always present in gold nanowires when loading is released .
the property of multi - shell gold nanowires to store a mechanical energy is useful for applications .
mechanical properties of various metallic nanowires deserve further experimental and theoretical investigations .
this work has been carried under the hr - mzt project 119206 `` dynamical properties of surfaces and nanostructures '' and the ec research action cost p3 `` simulation of physical phenomena in technological applications '' . | deformation properties of multi - wall gold nanowires under compressive loading are studied .
nanowires are simulated using a realistic many - body potential .
simulations start from cylindrical fcc ( 111 ) structures at @xmath0 k. after annealing cycles axial compression is applied on multi - shell nanowires for a number of radii and lengths at @xmath1 k. several types of deformation are found , such as large buckling distortions and progressive crushing .
compressed nanowires are found to recover their initial lengths and radii even after severe structural deformations .
however , in contrast to carbon nanotubes irreversible local atomic rearrangements occur even under small compressions . |
asymptotic giant branch ( agb ) stars are post - main sequence stars that represent the most luminous stage of evolution for low- and intermediate - mass stars .
one of the main properties of agb stars is that they loose @xmath0% of their mass ( gas and dust ) , they therefore chemically enrich the interstellar medium ( ism ) and represent the main source of dust in the universe .
agb stars are also indicators of distance , structure and metallicity .
they contribute to the integrated light of galaxies and exist in large numbers in the neighbouring magellanic clouds .
all stars with an initial mass comprised between @xmath1 and @xmath2 m@xmath3 become agb stars .
the agb phase is rather brief ( @xmath4 gyr ) compared to other stellar evolutionary phases where stars spend most of their life , e.g. the main sequence .
agb stars are characterized by a double shell burning of h and he , respectively , around their c- and o - rich nucleus . throughout the agb structure a complex chemistry
is developed : molecules form right above the photosphere and these are responsible for two classes of agb stars , c - rich ( or c - type ) where carbonaceous molecules like cn and c@xmath5 develop and o - rich ( or m - type ) where oxides like tio and vo develop , after the formation of stable co. the chemical path - way an agb star will take depends on the ism metallicity prior to their origin and on the efficiency of the dredge - up process that brings elements synthesized in the stellar interior to the surface . at a given distance from the centre of the star , where temperature and pressure are appropriate , dust forms , dust will also be predominantly of carbonaceous or silicate type and this is related to the molecular composition in the agb stars atmospheres . above the dust layer
other molecules , like oh and hcn , develop in the so - called circumstellar region .
the agb phase is a dynamical phase because stars experience surface luminosity variations with long periods and large amplitudes , and mass - loss .
these and further details on the agb stars can be found in habing & olofsson ( @xcite ) .
before infra - red ( ir ) images of agb stars were obtained the magellanic cloud stellar population was characterized by optical images .
initially the sensitivity was limited to bright super giant and carbon stars ( e.g. westerlund et al @xcite ) .
later , fainter sources were studied by blanco et al ( @xcite ) in several fields in the both the large magellanic cloud ( lmc ) and the small magellanic cloud ( smc ) .
these observations were extended to additional fields by blanco & mccarthy ( @xcite ) . at the same time
the variability of agb stars was investigated by hughes & wood ( @xcite ) who referred to agb stars as long period variables ( lpvs ) .
these authors discovered that there are several c - rich agb stars .
there were , however , not enough luminous agb stars compared to the prediction by theory and to the known large number of cepheid stars .
the latter are the precursors of agb stars of moderate mass .
it emerged that because agb stars are cool ( red ) they are potentially obscured by dust preventing their observation in the optical .
therefore , ir observations were needed to progress in this field . note
that the deficit of agb stars was several hundreds !
the first ir images of agb stars in the magellanic clouds were obtained by frogel & richer ( @xcite ) .
they used the ctio 1.5 m telescope to observe a field in the bar west region of the lmc ( fig .
[ fig1 ] ) .
the observations were sensitive to stars as faint as @xmath6 and the spatial resolution corresponded to an aperture of @xmath7 .
a few red , but not very luminous agb stars , were identified .
the study of the agb population advanced in parallel with the investigation of their variability aspect ; this is still the case at present . using the @xmath8-band magnitude
it was discovered that lpvs obey a period luminosity relation that represents an alternative distance indicator to cepheid stars ( hughes et al @xcite ) . in 1983
the ir astronomical satellite ( iras ) scanned the sky .
the instrumentation on board opened a new window into the study of agb stars , especially for those surrounded by dust .
iras also allowed for making the transition between studying individual agb stars to using them in large numbers to investigating properties of the hosting galaxy ( i.e. the milky way ) .
the exploitation of the iras catalogue is still on - going .
iras discovered @xmath9 obscured agb stars in the lmc and @xmath10 in the smc ( van loon et al @xcite ) .
the distribution of cool stars in the iras colour - colour diagram , [ 12]-[25 ] versus [ 25]-[60 ] , was presented by van der venn & habing ( @xcite ) .
regions occupied by agb stars of a different type and with a different dust shell thickness were identified , as well as the region where planetary nebulae ( pne ) , the successors of agb stars , are found .
it was suggested that ageing agb stars develop pulsation variability and thicker dust shells . in the following decade the ir space observatory ( iso ; 1995 - 1998 ) performed targeted observations of the agb stars discovered by iras .
the power of the iso instrumentation relied both in the imaging and spectroscopic capabilities that allowed for studying the type of dust , o - rich if absorption is present at @xmath11 @xmath12 m and c - rich if emission is present at @xmath13 @xmath12 m or absorption at @xmath14 @xmath12 m , and to relate this with the photometric colours and stellar models .
iso performed also a mini - survey of magellanic clouds ( loup et al @xcite ) .
this survey covers an area of @xmath1 deg@xmath15 and @xmath16 deg@xmath15 of the lmc and smc , respectively .
imaging observations were obtained in the [ @xmath17 , [ @xmath18 and [ @xmath19 @xmath12 m filters .
the considerably improved spatial resolution and sensitivity of iso with respect to iras showed that iras missed @xmath20% of the dust obscured agb stars and that these stars are as luminous as c - rich agb stars with thin dust shells ( fig .
[ fig3 ] ) .
the iso imaging data were combined with the near - ir data obtained from the denis survey ( cioni et al @xcite ) . more or less
simultaneously mid - ir space observations were obtained from the mid - course space experiment ( msx ; 1996 - 1997 ) .
this satellite observed @xmath21 deg@xmath15 of sky at better spatial resolution but worse sensitivity than iras .
the colour - colour diagram , [ @xmath8]-[@xmath2 ] versus [ @xmath24-[@xmath8 ] , showed obscured agb stars but with a non - negligible overlap with pne and hii regions ( egan et al @xcite ) .
the most comprehensive near - ir sky surveys that provided a wealth of data for studying the magellanic clouds were the 2mass ( 1997 - 2001 ; @xmath25 filters ) and denis ( 1995 - 2001 ; @xmath26 filters ) surveys . among their highlights : ( i ) the distribution of stars in the colour - magnitude space , e.g. @xmath27 versus @xmath28 ( nikolaev & weinberg @xcite ) , and ( ii ) the distinction between o - rich and c - rich agb stars of the earliest spectral types ( cioni & habing @xcite ) @xmath29 see fig .
both surveys detected the almost complete population of agb stars with absent or thin dust shells , as well as several obscured agb stars .
the view of the magellanic clouds changed dramatically from that of typical irregular galaxies with central features traced in the optical by young stars and star forming regions , to smooth and regular extended structures traced by giant stars ( cioni et al @xcite ) in the near - ir .
the lmc shows a large thick bar confined within an outer elliptical structure with hints of spiral arms .
the smc resembles a dwarf elliptical galaxy .
the ratio between c - rich and o - rich agb stars , the c / m ratio , provides also a view of the metallicity distribution within these galaxies ( cioni & habing @xcite ) .
this ratio shows a positive gradient within the lmc and a clumpy distribution consistent with a flat gradient in the smc .
the conversion and interpretation of the c / m ratio versus [ fe / h ] abundance has been investigated recently by cioni ( @xcite ) .
the smoothly declining lmc agb gradient agrees with that of old ( several gyr ) stellar clusters and rr lyrae stars , but it does differ significantly from the gradient traced by young ( a few gyr ) red giant branch ( rgb ) stars and stellar clusters that appears rather flat .
the latter suggest a flattening of the gradient with time and is perhaps influenced by the effect of the bar . in the smc nor
agb stars nor other indicators such as : rgb stars , pne , stellar clusters with a different age , show a significant gradient
. this can also be due to the bar or to the projection effect of two populations with a different mean age , a young one in the disk and an old one in an outer spheroid .
their average metallicity is consistent with that in the magellanic bridge and of the lmc at @xmath30 kpc from its centre supporting tidal stripping resulting from the dynamical interaction between the magellanic clouds .
the spitzer space telescope , launched in 2003 , has surveyed the magellanic clouds in different filters as part of two major projects : the sage and s@xmath31mc surveys of the lmc and smc , respectively .
these surveys have provided a large and homogeneous database for studying agb stars , their evolution and mass - loss properties .
sage ( meixner et al @xcite ) covered @xmath32 deg@xmath15 and acquired two epochs , separated by three months , that allowed for identifying variable stars as well as their location in colour - magnitude and colour - colour diagrams .
vijh et al ( @xcite ) discusses the distribution , variability and dust properties of the sage stars : @xmath33% of the extreme / obscured agb stars are variable , @xmath34% of the c - rich agb stars and @xmath35% of the o - rich agb stars with thin circumstellar shells are also variables in the mid - ir spitzer filters ( fig .
[ fig3 ] ) .
the spectral energy distribution ( sed ) obtained from the combination between 2mass and spitzer band - widths shows that the lack of variability data has a strong influence on the integrated flux and on the estimated mass - loss rate .
the sensitivity and spatial resolution of spitzer represent a tremendous improvement versus previous data .
in particular a spatial resolution of @xmath36 in the @xmath37 @xmath12 m range is comparable to the resolution of ground based near - ir surveys , consenting a more secure identification of counterparts .
the sensitivity in the mid - ir has surpassed that in the near - ir leaving unmatched many newly discovered objects .
the ir survey facility telescope ( irsf ; 2001 - 2006 ) observed a large area encompassing the lmc , smc and the part of the magellanic bridge close to the smc wing ( kato et al @xcite ) .
this is the most sensitive near - ir survey of the magellanic clouds to - date reaching a @xmath38 limit at @xmath39 , @xmath40 and @xmath41 .
the instrument resolution corresponds to @xmath42 @xmath43/pix while the observations were obtained with an average seeing of @xmath44 .
agb stars were not easily found in the bridge and in general no agb stars were found down to @xmath45 at @xmath46 .
the akari telescope , launched in 2006 , has completed the observation of @xmath47 deg@xmath15 in the north - east area of the lmc and is currently undertaking an all - sky survey covering the range @xmath48 @xmath12 m .
a preliminary catalogue of point sources has been published by ita et al ( @xcite ) .
the band - widths are similar to spitzer , but for the @xmath49 m one , and have been matched with irsf near - ir magnitudes .
their combination reaches sources @xmath50 mag fainter than the spitzer-2mass combination .
the colour - magnitude diagram , [ @xmath14]-[@xmath52 versus [ @xmath52 , shows interesting new features traced by agb stars : a faint plume of sources just brighter than the rgb tip and bending to red colours indicating o - rich giants with al oxide dust ( blum et al @xcite , lebzelter et al @xcite ) . in the near - ir domain
the vista wide - field telescope is currently being commissioned at the european southern observatory ( eso ) .
the core programme for the next five years at vista includes six public surveys of which two are devoted to the observation of stars , the others are extragalactic surveys , and one is focused on the magellanic system .
the vista survey of the magellanic system ( vmc ; cioni et al @xcite ) will provide the missing link to optical surveys , with a similar sensitivity , as well as counterparts for the mid - ir sources .
vmc will cover @xmath53 deg@xmath15 distributed across broad areas in the lmc and smc , the entire length of the magellanic bridge connecting the two galaxies ( for a width of @xmath54 deg ) and a couple of fields in the magellanic stream ( fig .
[ fig4 ] ) .
observations will be obtained in three filters , @xmath55 providing a @xmath38 limit at @xmath56 , @xmath57 and @xmath58
. they will be executed in service mode to guarantee homogeneous conditions and data quality .
the instrument resolution is @xmath59 @xmath43/pix and observations will be obtained with an average seeing of @xmath60 depending on crowding .
vmc is also a multi - epoch survey because it will reach its nominal sensitivity by combining @xmath61 independent epochs in @xmath28 , and @xmath14 in @xmath62 and @xmath63 , respectively .
the completion of the survey requires @xmath64 hours .
the main science goals of the vmc surveys are to derive the spatially resolved star formation history ( sfh ) and to measure the three - dimensional ( 3d ) structure of the magellanic system .
the vmc data will also be used to find stellar sub - structures ( clusters and streams ) , emission line objects ( like pne ) , to derive distances and measure proper motions , to study star formation , to re - construct the system using dynamical models and eventually find extra - galactic objects ( star - burst galaxies and dusty active galactic nuclei at high redshift ) , as well as form many other scientific applications .
the vmc survey will be complementary to an on - going optical survey of the outer regions of the magellanic clouds ( @xmath65 kpc from the galaxy centres ) .
this area will also be surveyed by vista as part of the vista hemisphere survey ( vhs ) but to a shallower depth than the vmc depth , and to the space astrometry mission gaia that will measure the metallicity and motion of evolved giant stars of the magellanic system .
simulations of the stellar population that vmc will detect show that stars just below the oldest main - sequence turn - off will be well detected .
the increased in the parameter space , e.g. near - ir photometry , at this depth will allow for determining the metallicity and age distribution with improved accuracy ( kerber et al @xcite ) .
the 3d structure will be measured using different distance indicators and in particular the period - luminosity relation for short period variable stars , rr lyrae stars and cepheids .
vmc will provide the near - ir magnitude and the period will be obtained from large optical catalogues in the literature ( e.g. ogle - iii and eros - ii ) or from observations at the vlt survey telescope ( vst ) , for the bridge and smc parts only .
although vmc is a deep survey targeting faint stellar populations it will find counterparts to faint agb stars and , via an accurate analysis of the sfh of the system , it will relate them to their progenitors .
the magellanic clouds will be targeted by different new missions . during this meeting
the herschel satellite was successfully launched .
herschel will probe the mid- to far - ir regime for agb stars in the magellanic clouds .
the next step in ground - based ir astronomy may be taken really in antarctica with the development of a telescope like pilot ( see other contributions in this proceeding ) .
this will be a 2 m class telescope that will explore the dark side of the @xmath8-band and possibly extend to the @xmath66-band , reaching an unprecedented depth , and will also be ideal to monitor agb stars in the magellanic clouds .
ir imaging of agb stars in the magellanic clouds were aimed at finding the most luminous and dusty , thus obscured , sources .
different criteria were developed to distinguish and classify agb stars into o- and c - rich .
multi - wavelength and multi - epoch ( optical only ) observations have allowed to quantify mass - loss rates and the coupling with agb pulsation and evolution .
major progress in the study of agb stars has occurred not only on the observational side but also on the theoretical side with the development of models that are able to interpret the location of agb stars in the ir ( marigo et al @xcite ) .
these authors have produced isochrones including molecular opacities in o- and c - rich agb stars , the hot bottom burning process , the effect of the pulsation mode ( first overtone or fundamental mode ) to the agb lifetime and the mass - loss with respect to the different surface chemistry . by exploring the range of ages typical for agb stars , these models , reproduce well the distribution of the agb population in the magellanic clouds in both near- and mid - ir diagrams .
the most luminous agb stars experience the strongest mass - loss rate .
this result has been greatly re - affirmed with the investigation by fraser ( @xcite ) that involves the combination between spitzer and 2mass data with optical monitoring data from the macho project .
the bolometric magnitude and the mass - loss rate were obtained from the sed of individual agb stars and the period extracted from their optical light - curve . the major uncertainty still present in the period - luminosity relations
is directly reflected in the uncertainty attributed to the bolometric magnitude . to obtain average bolometric magnitudes
it is necessary to monitor the sed across a few years .
the reduced scatter around the period - luminosity relations will also contribute to refining agb stars as a powerful tool to measure distances in the universe .
there is at present a paucity of ir monitoring ( @xmath35 epochs by spitzer , @xmath61 epochs by vmc ) and a lack of foreseen projects to improve this condition .
mean bolometric luminosities are the next step in the investigation of the mass - loss mechanism and in measuring distances from the period luminosity relation .
the magellanic clouds are the closest examples of interacting galaxies and represent an ideal laboratory for stellar evolution .
the design of pilot , a telescope for the antarctica site , is well suited to advance the study of agb stars in these neighbouring galaxies by providing better sensitivity , spatial accuracy and wide - field monitoring .
99 blanco , v.m .
, mccarthy , m.f . &
blanco , b.m .
, 1980 , aj , 242 , 983 blanco , v.m . & mccarthy , m.f . , 1983 , aj , 88 , 1442 blum , r.d . , mould , j.r . ,
olsen , k.a .
, et al 2006 , aj 132 , 2034 cioni , m .- r.l . , habing , h.j . &
israel , f.p .
2000b , a&a , 358 , l9 cioni , m .- r.l . ,
loup , c. , habing , h.j . ,
et al 2000a , a&as , 144 , 235 cioni , m .- r.l .
2009 , a&a , _ submitted _ , astro - ph/0904.3136 cioni , m .- r.l . , bekki , k. , clementini , g. , et al 2008 , pasa , 25 , 121 cioni , m .- r.l . , &
habing , h.j . , 2003 ,
a&a , 402 , 133 cioni , m .- r.l . , blommaert , j.a.d.l , groenewegen , m.a.t . , 2003b , a&a , 406 , 51 egan , m.p . , van dyk , s.d . & price , s.d .
2001 , aj , 122 , 1844 fraser , o. 2008 , _ properties of long - period variables from the macho project _ , phd thesis frogel , j.a . &
richer , h.b .
1983 , apj 275 , 84 habing , h.j .
& olofsson , h. 2003 , editors , _ asymptotic giant branch stars _ , astronomy and astrophysics library , new york , berlin : springer hughes , s.m.g . & wood , p.r . , 1990 , aj , 99 , 784 kato , d. , nagashima , c. , nagayama , t. , et al 2007 , pasj , 59 , 615 kerber , l. , girardi , l. , rubele , s. & cioni , m .- r.l . 2009 ,
a&a , 499 , 697 ita , y. , onaka , t. , kato , d. , et al 2008 , pasj , 60 , 435 lebzelter , th . ,
posh , th . , hinkle , k. , et al 2006 , apj , 653 , l145 loup , c. 1999 , _ new views of the magellanic clouds _ , iau symp . , 190 , 328 marigo , p. , girardi , l. , bressan , a. , et al 2008 , a&a , 482 , 883 meixner , m. , gordon , k.d . ,
indebetow , r. , et al 2006 , aj , 132 , 2268 nikolaev , s. & weinberg , m.d .
2000 , apj , 542 , 804 van loon , j.th . ,
groenewegen , m.a.t . , de koter , a. , et al 1999 , a&a , 351 , 559 vijh , u.p . , meixner , m. , babler , b. , et al 2009 , aj , 137 , 3139 van der veen , w.e.c.j .
& habing , h.j.h . , 1988 , a&a , 194 , 125 westerlund , b.e .
1964 , _ the galaxy and the magellanic clouds _ , iau symp . , | agb stars are ideal ir targets because they are cool and bright .
most of them escaped detection in optical or shallow ir surveys in the eighties contributing to the puzzling missing number of agb stars with respect to theoretical predictions and former stages of evolution .
observations and agb models have advanced steadily in the following decades providing us with an almost complete view of the agb stars in the magellanic clouds .
their properties are tracers of structure and chemistry across galaxies .
new surveys will be able to fill - in the gaps , in terms of sensitivity and monitoring , providing new constraints for the formation and evolution of the magellanic clouds . |
our universe is seems to undergo a period of accelerated expansion and it is assumed that a considerable part of the total density consists of dark energy component with negative pressure @xcite .
there are several candidates for the dark energy : cosmological constant ( @xmath1 ) or dynamical component such as quintessence @xcite and @xmath2-essence @xcite . in connection with the solving of the problem of fine - tuning
the models of dynamical dark energy component are seem to be more realistic as they admit to construct `` tracker '' @xcite or `` attractor '' @xcite solutions .
one of the peculiar feature of the cosmological dark energy is a possibility of the big rip @xcite : the infinite expansion of the universe during a finite time .
the big rip scenario is realized if a dark energy is in the form of the phantom energy with @xmath3 . in this case
the scenario of big rip is possible when cosmological phantom energy density grows at large times and disrupts finally all bounded objects up to subnuclear scale .
note , however that the only condition @xmath3 is not enough for the realization of big rip @xcite . in @xcite
the authors analyzed the supernova data in the model independent manner and showed that the presence of the phantom energy with @xmath4 is preferable in the present moment of time .
the analogy between phantom and qft in curved space - time has been developed in @xcite .
the entropy of the universe with phantom energy is discussed in @xcite .
usually the evolution of quintessence or @xmath2-essence are considered in a view of cosmological problems .
however in the presence of compact objects such as black holes the evolution of dark energy should be sufficiently different from that in the cosmological consideration . indeed , what would be the fate of black holes in the universe filled with the phantom energy and coming to big rip ?
recently we showed that all black holes gradually decrease their masses and very near the big rip they finally disappear @xcite . in the present work we study in details the stationary accretion of dynamical dark energy into the black hole .
as a model of de we take the perfect fluid with negative pressure .
the studying of accretion of perfect fluid on the compact objects originated from bondi @xcite .
the relativistic generalization of the perfect fluid accretion were made by michel @xcite .
below we find the solution for a stationary accretion of the relativistic perfect fluid with an arbitrary equation of state @xmath0 onto the schwarzschild black hole .
using this solution we show that the black hole mass diminishes by accretion of the phantom energy .
masses of all black holes gradually tend to zero in the phantom energy universe approaching to the big rip .
the diminishing of a black hole mass is caused by the violation of the energy domination condition @xmath5 which is a principal assumption of the classical black hole ` non - diminishing ' theorems @xcite .
the another consequence of the existence of a phantom energy is a possibility of traversable wormholes @xcite . in @xcite
authors studied the accretion of scalar quintessence field into the black hole , using the specific quintessence potentials @xmath6 for the obtaining of the analytical solution for the black hole mass evolution .
we use essentially different approach for the description of de accretion into black hole , namely , we model the de by the perfect fluid with the negative pressure .
let us consider the spherical accretion of dark energy onto black hole .
we assume that the density of the dark energy is sufficiently low so that the metric can be described by schwarzschild metric .
we model the dark energy by a perfect fluid with energy - momentum tensor : @xmath7 , where @xmath8 is the density and @xmath9 is the pressure of the dark energy and @xmath10 is the four - velocity @xmath11 .
the integration of the time component of the energy - momentum conservation law @xmath12 gives the first integral of motion @xmath13 where @xmath14 , @xmath15 and @xmath16 is a constant determined below .
given the equation of state @xmath17 , one can introduce the function @xmath18 by the relation : @xmath19 the function @xmath18 plays the role of concentration of the particles , though one can use @xmath18 for the media without introducing any particles . in this case
@xmath18 is the auxiliary function . for general equation of state @xmath17 , from ( [ n ] )
we obtain the following solution for @xmath18 : @xmath20 from the conservation of energy - momentum along the velocity @xmath21 using ( [ n1 ] ) we obtain the another first integral : @xmath22 where @xmath23 ( the concentration of the dark energy at the infinity ) was introduced for convenience . in the case of inflow @xmath24 and the constant @xmath25 . from ( [ eq1 ] ) and ( [ flux ] )
one can easily obtain : @xmath26 where @xmath27 we will see below that the constant @xmath28 which determines the flux is fixed for fluids with this can be done through finding of the critical point .
following michel @xcite we obtain the parameters of critical point : @xmath29 where @xmath30 from this by using ( [ n ] ) it follows that @xmath31 , where @xmath32 is the squared effective speed of sound in the media . combining the eqs .
( [ energy ] ) , ( [ c2 ] ) , ( [ cpoint ] ) and ( [ v1 ] ) we find the following relation : @xmath33^{1/2 } \frac{\rho_\infty+p(\rho_\infty)}{n(\rho_\infty)},\ ] ] which gives the @xmath34 for arbitrary equation of state @xmath17 . given
@xmath34 one can find @xmath35 using ( [ n1 ] ) and values @xmath36 , @xmath37 , using ( [ cpoint ] ) and ( [ v1 ] ) . then substituting the calculated values in ( [ flux ] )
one can find the constant @xmath28 .
note that there is no critical point outside the black hole horizon ( @xmath38 ) for @xmath39 or @xmath40 .
this means that for unstable perfect fluid with @xmath39 or @xmath40 a dark energy flux onto the black hole depends on the initial conditions .
this result has a simple physical interpretation : the accreting fluid has the critical point if its velocity increases from subsonic to trans - sonic values . in a fluid with a negative @xmath41 or with @xmath40 the fluid velocity never crosses such a point .
it should be stressed , however , that fluids with @xmath39 are hydrodynamically unstable ( see discussion in @xcite ) .
( [ energy ] ) together with ( [ n1 ] ) and ( [ flux ] ) describe the requested accretion flow onto the black hole .
these equations are valid for perfect fluid with an arbitrary equation of state @xmath17 , in particular , for a gas with zero - rest - mass particles ( thermal radiation ) and for a gas with nonzero - rest - mass particles . for a nonzero - rest - mass gas the couple of equations ( [ flux ] ) and ( [ energy ] ) is reduced to similar ones found by michel @xcite .
one would note that the set of equations ( [ n1 ] ) , ( [ flux ] ) and ( [ energy ] ) are also correct in the case of dark energy and phantom energy @xmath3 . in this case concentration @xmath42
is positive for any @xmath8 and constant @xmath43 in ( [ energy ] ) is negative . the black hole mass changes at a rate @xmath44 due to the fluid accretion .
with the help of ( [ flux ] ) and ( [ energy ] ) this can be expressed as @xmath45.\ ] ] for the phantom energy the relation ( [ evol ] ) leads to the diminishing of the black hole mass .
that means that in the universe filled with phantom energy the black holes should melt away .
this result is general , it does not depend on the equation of state @xmath17 , the only condition @xmath46 is important .
let us consider the model of dark energy with linear dependence of pressure from the density : @xmath47 which include , among others , the ultra - relativistic gas ( @xmath48 ) and simplest models of dark energy ( @xmath49 and @xmath50 ) .
introduced value @xmath51 is connected with usual equation of state @xmath52 by the relation @xmath53 . for @xmath50 there is no critical point for the flux of the fluid into the black hole . in the case of @xmath54 , using ( [ cpoint ] ) and ( [ v1 ] ) we find the parameters for critical point in model ( [ p1 ] ) : @xmath55 it should be noted that in the linear model ( [ p1 ] ) the parameters of critical point ( [ cpoint1 ] ) determined only by @xmath56 and do not depend on the parameter @xmath57 , which determines what physical fluid is considered : relativistic gas , dark energy or phantom energy .
note also that for @xmath58 ( that corresponds to the non - physical situation of superluminal speed of sound ) there is no critical point outside the black hole .
let us calculate the constant @xmath28 which determines the flux of the fluid into the black hole . from eq .
( [ n1 ] ) we find : @xmath59 where we defined the effective density @xmath60 . using ( [ rho_c ] ) we obtain : @xmath61 where @xmath62 is the value of effective density at the critical point and @xmath63 is the effective density at the infinity . substituting ( [ rho_eff_c ] ) in ( [ n2 ] ) and then using ( [ flux ] ) we find for linear model : @xmath64 it is easily seen that @xmath65 for @xmath66 . for @xmath67 that corresponds to @xmath68
the we have @xmath69 .
from this we may conclude that for typical sound speeds the constant @xmath28 has value around unity .
for some particular choices of parameter @xmath51 the values @xmath70 and @xmath71 can be calculated analytically .
for example , for @xmath72 the fluid density is given by : @xmath73 ^ 2,\ ] ] where @xmath74 @xmath75\ ] ] and @xmath76 the density distribution for another physically interesting case @xmath67 is given by : @xmath77 the corresponding radial fluid velocity @xmath78 can be calculated by substituting of ( [ sol1 ] ) or ( [ sol2 ] ) into ( [ eq1 ] ) . for @xmath49 the solutions ( [ sol1 ] ) and ( [ sol2 ] ) describe correspondingly a thermal radiation and a fluid with ultra - hard equation of state .
in the case of @xmath79 the solutions ( [ sol1 ] ) and ( [ sol2 ] ) describe the phantom energy falling onto the black hole . for example , a phantom energy flow with parameters @xmath67 and @xmath80 results in a black hole mass diminishing with the rate @xmath81 .
now we turn to the problem of the black hole evolution in the universe with the big rip when a scale factor @xmath82 diverges at finite time @xcite . for simplicity
we will take into account only dark energy and will disregard all others forms of energy .
the big rip solution is realized for in the linear model ( [ p1 ] ) for @xmath3 and @xmath83 . from the friedman equations for the linear equation of state model one can obtain : @xmath84 .
taking for simplicity @xmath49 we find the evolution of the density of a phantom energy in the universe : @xmath85 where @xmath86 and @xmath87 is the initial density of the cosmological phantom energy and the initial moment of time is chosen so that the ` doomsday ' comes at time @xmath88 . from ( [ sol3 ] ) and ( [ tau2 ] ) it is easy to see that the big rip solutionis realized for @xmath89 .
in general , the satisfying the condition @xmath3 is not enough for the possibility for universe to come to big rip . from ( [ evol ] ) using ( [ sol3 ] ) we find the black hole mass evolution in the universe coming to the big rip : @xmath90 where @xmath91 and @xmath92 is the initial mass of the black hole . for @xmath93 and typical value of @xmath69 ( corresponding to @xmath94 )
we have @xmath95 . in the limit @xmath96 ( i.e. near the big rip )
the dependence of black hole mass on @xmath97 becomes linear , @xmath98 . while @xmath97 approaches to @xmath88 the rate of black hole mass decrease does not depend on both an initial black hole mass and the density of the phantom energy : @xmath99 .
in other words masses of all black holes in the universe tend to be equal near the big rip .
this means that the phantom energy accretion prevails over the hawking radiation until the mass of black hole is the planck mass .
however , formally all black holes in the universe evaporate completely at planck time before the big rip due to hawking radiation .
in remaining let us confront our results with the calculations of ( not phantom ) scalar field accretion onto the black hole @xcite
. the dark energy is usually modelled by a scalar field @xmath100 with potential @xmath6 .
the perfect fluid approach is more rough because for given ` perfect fluid variables ' @xmath8 and @xmath9 one can not restore the ` scalar field variables ' @xmath100 and @xmath101 . in spite of the pointed difference between a scalar field and a perfect fluid we show below that our results are in a very good agreement with the corresponding calculations of a scalar field accretion onto the black hole .
the lagrangian of a scalar field is @xmath102 , where @xmath103 is a kinetic term of a scalar field @xmath100 and @xmath104 is a potential . for
the standard choice of a kinetic term @xmath105 the energy flux is @xmath106 .
jacobson @xcite found the scalar field solution in schwarzschild metric for the case of zero potential @xmath107 : @xmath108 $ ] , where @xmath109 is the value of the scalar field at the infinity .
in @xcite it was shown that this solution remains valid also for a rather general form of runaway potential @xmath6 .
for this solution we have @xmath110 and correspondingly @xmath111 .
the energy - momentum tensor constructed from jacobson solution completely coincides with one for perfect fluid in the case of ultra - hard equation of state @xmath112 under the replacement @xmath113 , @xmath114 .
it is not surprising because the theory of a scalar field with zero potential @xmath6 is identical to perfect fluid consideration @xcite . in a view of this coincidence
it is easily to see the agreement of our result ( [ evol ] ) for @xmath115 in the case of @xmath112 and the corresponding result of @xcite . to describe the phantom energy
the lagrangian of a scalar field must have a negative kinetic term @xcite , for example , @xmath116 ( for the more general case of the negative kinetic term see @xcite ) . in this case the phantom energy flux onto black hole has the opposite sign , @xmath117 , where @xmath100 is the solution of the same klein - gordon equation as in the case of standard scalar field , however with the replacement @xmath118 . for zero potential
this solution coincides with that obtained by jacobson @xcite for a scalar field with the positive kinetic term .
lagrangian with negative kinetic term and @xmath119 does not describe , however , the phantom energy . at the same time
, the solution for scalar field with potential @xmath119 is the same as with a positive constant potential @xmath120 , which can be chosen so that @xmath121 . in this case
the scalar field represents the required accreting phantom energy @xmath122 and @xmath123 and provides the decrease of black hole mass with the rate @xmath124 .
the simple example of phantom cosmology ( without a big rip ) is realized for a scalar field with the potential @xmath125 , where @xmath126 ev @xcite .
after short transition phase this cosmological model tends to the asymptotic state with @xmath127 and @xmath128 . in the klein - gordon equation
the @xmath129 term ( with the mentioned replacement @xmath118 ) is comparable to other terms only at the cosmological horizon distance .
this means that the jacobson solution is valid for this case also .
calculating the corresponding energy flux one can easily obtain @xmath130 . for @xmath131 and @xmath132 ev the effective time of black hole mass decrease
is @xmath133 yr .
+ this work was supported in part by the russian foundation for basic research grants 02 - 02 - 16762-a , 03 - 02 - 16436-a and 04 - 02 - 16757-a and the russian ministry of science grants 1782.2003.2 and 2063.2003.2 . | solution for a stationary spherically symmetric accretion of the relativistic perfect fluid with an equation of state @xmath0 onto the schwarzschild black hole is presented .
this solution is a generalization of michel solution and applicable to the problem of dark energy accretion .
it is shown that accretion of phantom energy is accompanied with the gradual decrease of the black hole mass .
masses of all black holes tend to zero in the phantom energy universe approaching to the big rip . |
in 1981 , dicke proposed a concept of interaction - free measurement ( ifm ) @xcite .
however , current discussion of ifm appears from the following problem stated by elitzur and vaidman : `` let us assume there is an object that absorbs a photon with strong interaction if the photon approaches the object closely enough .
can we examine whether or not the object exists without its absorption ? ''
the reason that we do not want to let the object absorb the photon is that it might lead to an explosion , for example .
elitzur and vaidman themselves present a method of the ifm that is inspired by the mach - zehnder interferometer .
then a more refined one is proposed by kwiat _
et al_. @xcite .
an experiment of their ifm is reported in ref .
the ifm finds wide application in quantum information processing ( the bell - basis measurement , quantum computation , and so on ) @xcite . according to the ifm proposed by kwiat _
et al_. , the absorbing object is put in the interferometer that consists of @xmath5 beam splitters , and we inject a photon into it to examine whether or not the object exists .
the probability that we can find the object in the interferometer arrives at unity under the limit of @xmath6 in the case where the interaction between the object and the photon is strong enough and perfect . in this paper
, we consider the ifm of kwiat _ et al_. with imperfect interaction . in ordinary ifm ,
the absorbing object is expected to absorb a photon with probability unity when the photon approaches the object closely enough . however , in this paper , we assume that the photon is absorbed with probability @xmath0 ( @xmath1 ) and it passes by the object without being absorbed with probability @xmath2 when it approaches close to the object .
we estimate the success probability @xmath3 of the ifm , namely the probability that we can find the object without the photon absorbed , under this assumption .
this problem has been investigated in ref .
@xcite already . in ref .
@xcite , although a correct approximating equation of the success probability of the ifm with the imperfect interaction is derived , its derivation is wrong .
hence , we give a right treatment of this problem in this paper .
interferometer of kwiat _
et al_. for the ifm . ] in the rest of this section , we give a short review of the ifm proposed by kwiat _
et al_. they consider an interferometer that consists of @xmath5 beam splitters as shown in fig .
[ kwinterferometer ] .
we describe the upper paths as @xmath7 and lower paths as @xmath8 , so that the beam splitters form the boundary line between the paths @xmath7 and the paths @xmath8 in the interferometer .
we write a state with one photon on the paths @xmath7 as @xmath9 and a state with no photon on the paths @xmath7 as @xmath10 .
this notation applies to the paths @xmath8 as well .
the beam splitter @xmath11 in fig .
[ kwinterferometer ] works as follows : @xmath12 [ the transmissivity of @xmath11 is given by @xmath13 , and the reflectivity of @xmath11 is given by @xmath14 in eq .
( [ definition - beam - splitter - b ] ) . ]
let us throw a photon into the lower left port of @xmath8 in fig .
[ kwinterferometer ] .
if there is no object on the paths , the wave function of the photon that comes from the @xmath15th beam splitter is given by @xmath16 if we assume @xmath17 , the photon that comes from the @xmath5th beam splitter goes to the upper right port of @xmath7 with probability unity .
next , we consider the case where there is an object that absorbs the photon on the paths @xmath7 .
we assume that the object is put on every path @xmath7 that comes from each beam splitter , and all of these @xmath5 objects are the same one .
the photon thrown into the lower left port of @xmath8 can not go to the upper right port of @xmath7 because the object absorbs it .
if the incident photon goes to the lower right port of @xmath8 , it has not passed through paths @xmath7 in the interferometer . therefore , the probability that the photon goes to the lower right port of @xmath8 is equal to the product of the reflectivities of the beam splitters .
it is given by @xmath18 . in the limit of @xmath6 ,
@xmath3 approaches @xmath19 as follows : @xmath20 \nonumber \\ & = & 1.\end{aligned}\ ] ] from the above discussion , we can conclude that the interferometer of kwiat et al . directs an incident photon from the lower left port of @xmath8 with probability @xmath3 at least as follows : ( 1 )
if there is no absorbing object in the interferometer , the photon goes to the upper right port of @xmath7 , and ( 2 ) if there is the absorbing object in the interferometer , the photon goes to the lower right port of @xmath8 . furthermore , if we take large @xmath5 , we can set @xmath3 arbitrarily close to @xmath19 .
therefore , we can examine whether or not the object exists in the interferometer .
the ifm introduced in the former section is realized by beam splitters and interaction between the absorbing object and the photon . in this section ,
we consider the case where the interaction is not perfect .
( we regard the beam splitters as accurate enough . )
we assume that the photon is absorbed with probability @xmath0 and it passes by the object without being absorbed with probability @xmath2 when it approaches close to the object .
we estimate the success probability @xmath3 of the ifm under these assumptions .
we assume the following transformation in fig .
[ kwinterferometer ] .
the photon that comes from each beam splitter to the upper path @xmath7 suffers @xmath21 where @xmath1 and @xmath22 .
@xmath23 is the state where the object absorbs the photon .
we assume that it is normalized and orthogonal to @xmath24 , where @xmath22 and @xmath25 . from now on , for simplicity
, we describe the transformations that are applied to the photon as matrices in the basis @xmath24 .
writing @xmath26 we can describe the beam splitter @xmath11 defined in eq .
( [ definition - beam - splitter - b ] ) as @xmath27 where @xmath17 and the absorption process defined in eq .
( [ imperfect - absorption - process ] ) as @xmath28 where @xmath1 .
the matrix @xmath29 is not unitary because the process defined by eq .
( [ imperfect - absorption - process ] ) causes absorption of the photon ( dissipation or decoherence ) .
results of numerical calculation of the success probability @xmath30 from eqs .
( [ definition - basis - vectors ] ) , ( [ definition - matrix - b ] ) , ( [ definition - matrix - a ] ) , and ( [ definition - fidelity - ifmgate ] ) . with fixing @xmath2 , we plot @xmath30 as a function of @xmath5 and link them together by solid lines .
@xmath2 is the rate at which the object fails to absorb the photon . @xmath3 and @xmath2 are dimensionless quantities .
@xmath5 is the number of the beam splitters .
four cases of @xmath31 , @xmath32 , @xmath33 , and @xmath34 are shown in order from top to bottom as solid curves . ]
the probability that an incident photon from the lower left port of @xmath8 passes through the @xmath5 beam splitters and is detected in the lower right port of @xmath8 in fig .
[ kwinterferometer ] is given by @xmath35 we plot results of numerical calculations of the success probability @xmath30 defined by eqs .
( [ definition - basis - vectors ] ) , ( [ definition - matrix - b ] ) , ( [ definition - matrix - a ] ) , and ( [ definition - fidelity - ifmgate ] ) in fig .
[ imperfectifm - n - exactp ] . with fixing @xmath2 , we plot @xmath30 as a function of @xmath5 and link them together by solid lines . in fig .
[ imperfectifm - n - exactp ] , the four cases of @xmath31 , @xmath32 , @xmath33 , and @xmath34 are shown in order from top to bottom . seeing fig .
[ imperfectifm - n - exactp ] , we have the following question .
if @xmath36 , at what value does @xmath30 converge in the limit of @xmath6 ? at first glance , @xmath37 seems to depend on @xmath2 .
however , fig .
[ imperfectifm - n - exactp ] lets us expect @xmath38 @xmath39 . in the next section , we show that this expectation is true .
in this section , we examine @xmath30 defined in eq .
( [ definition - fidelity - ifmgate ] ) for large @xmath5 . for this purpose , first we derive an exact formula of @xmath30 , and second we expand @xmath30 in powers of @xmath4 with fixing @xmath2 .
first we derive the exact formula of @xmath30 .
we note the following .
eigenvalues of the matrices @xmath29 and @xmath11 are given by @xmath19 and @xmath40 .
thus , @xmath37 never diverges to infinity .
we let @xmath41 to be an upper triangular matrix @xmath42 by unitary transformation @xmath43 as follows : @xmath44 where @xmath45 @xmath46 @xmath47^{2 } , \nonumber \\ t & = & 4\sin^{2}\theta+[(1-\sqrt{\eta})\cos\theta - r]^{2},\end{aligned}\ ] ] and @xmath48 , \nonumber \\ y & = & -\frac{1}{\sqrt{st}}(1-\sqrt{\eta})\sin^{2}\theta [ 2(1+\sqrt{\eta})\cos\theta \nonumber \\ & & \quad + \sqrt{2 } \sqrt{1 - 6\sqrt{\eta}+\eta+(1+\sqrt{\eta})^{2}\cos 2\theta } ] , \nonumber \\ z & = & \frac{1}{2}[(1+\sqrt{\eta})\cos\theta+r].\end{aligned}\ ] ] we obtain @xmath49 by induction as follows : @xmath50 where @xmath51 from the above calculations , we obtain the exact formula of @xmath30 as @xmath52 second we expand @xmath30 in powers of @xmath4 with fixing @xmath2 .
( we note @xmath17 . )
we can expand components of the matrices @xmath11 , @xmath43 , and @xmath42 in powers of @xmath4 as follows : @xmath53 @xmath54 \nonumber \\ & & \quad\quad + o(\frac{1}{n^{5}}),\nonumber \\ u_{10 } & = & \frac{\pi}{2n}\frac{\sqrt{\eta}}{1-\sqrt{\eta } } [ 1+\frac{\pi^{2}}{24n^{2}}\frac{2 - 3\eta+\eta^{3/2}}{(1-\sqrt{\eta})^{3 } } ] \nonumber \\ & & \quad\quad + o(\frac{1}{n^{5}}),\nonumber \\ u_{11 } & = & -1+\frac{\pi^{2}}{8n^{2}}\frac{\eta}{(1-\sqrt{\eta})^{2}}+o(\frac{1}{n^{4 } } ) , \label{components - u - expansion}\end{aligned}\ ] ] and @xmath55+o(\frac{1}{n^{4 } } ) , \nonumber \\ y & = & -\frac{\pi}{2n } [ 1-\frac{\pi^{2}}{24n^{2}}]+o(\frac{1}{n^{5 } } ) , \nonumber \\ z & = & 1-\frac{\pi^{2}}{8n^{2}}\frac{1-\eta}{(1-\sqrt{\eta})^{2}}+o(\frac{1}{n^{4 } } ) .
\label{expansion - components - d}\end{aligned}\ ] ] next , from eqs .
( [ exact - components - d - n-1 ] ) and ( [ expansion - components - d ] ) , we expand components of @xmath49 in powers of @xmath4 with fixing @xmath2 . @xmath56 and
@xmath57 can be written as follows : @xmath58 , \label{expansion - component - x } \\ z & = & 1-\frac{\pi^{2}}{8n}\frac{1-\eta}{(1-\sqrt{\eta})^{2 } } + o(\frac{1}{n^{2 } } ) .
\label{expansion - component - z}\end{aligned}\ ] ] in eq . (
[ expansion - component - x ] ) , a factor @xmath59 appears .
we never expand @xmath59 in powers of @xmath4 and regard it as a constant . because @xmath5 is a large number
, we can assume @xmath60 .
thus , we can write @xmath61 in the form , @xmath62 + o(\frac{1}{n^{3 } } ) . \label{expansion - component - y}\ ] ] substituting eqs .
( [ definition - matrix - b ] ) , ( [ definition - matrix - u ] ) , ( [ definition - matrix - d - n-1 ] ) , ( [ components - b - expansion ] ) , ( [ components - u - expansion ] ) , ( [ expansion - component - x ] ) , ( [ expansion - component - z ] ) , and ( [ expansion - component - y ] ) into eq . ( [ exact - formula - fidelity - ifmgate ] ) , we obtain an expansion of the success probability @xmath3 in powers of @xmath4 as follows : @xmath63 hence , we can let @xmath3 get close to unity arbitrarily by increasing @xmath5 . [
if we make @xmath5 larger , the reflectivity of the beam splitter @xmath64 becomes larger and gets close to unity . ]
results of numerical calculation of the success probability @xmath3 as a function of @xmath5 with @xmath65 .
@xmath2 is the rate at which the object fails to absorb the photon . @xmath3 and @xmath2 are dimensionless quantities . @xmath5 is the number of the beam splitters .
a thick solid curve represents an exact result from eqs .
( [ definition - basis - vectors ] ) , ( [ definition - matrix - b ] ) , ( [ definition - matrix - a ] ) , and ( [ definition - fidelity - ifmgate ] ) .
a thin solid curve represents an approximate result from eq .
( [ approximating - eq - p ] ) . ] from eq .
( [ expansion - success - probability ] ) , we obtain the following approximating equation of @xmath3 for large @xmath5 , @xmath66 in fig .
[ imperfectifm - n - approxp ] , we plot results of numerical calculation of the success probability @xmath3 as a function of @xmath5 with @xmath65 .
a thick solid curve represents an exact result from eqs .
( [ definition - basis - vectors ] ) , ( [ definition - matrix - b ] ) , ( [ definition - matrix - a ] ) , and ( [ definition - fidelity - ifmgate ] ) .
a thin solid curve represents an approximate result from eq .
( [ approximating - eq - p ] ) . seeing fig .
[ imperfectifm - n - approxp ] , we find that eq .
( [ approximating - eq - p ] ) is a good approximation to @xmath3 for large @xmath5 .
we show that even if the interaction between the object and the photon is imperfect , we can let the success probability @xmath3 of the ifm get close to unity arbitrarily by making the reflectivity of the beam splitter larger and increasing the number of the beam splitters .
we obtain an approximating equation of @xmath3 with the imperfect interaction . to overcome the imperfection of the interaction , we need to prepare a large number of beam splitters and let their transmission rate get smaller .
99 r. h. dicke , ` interaction - free quantum measurements : a paradox ? ' , am .
j. phys . * 49 * , 925930 ( 1981 ) .
a. c. elitzur and l. vaidman , ` quantum mechanical interaction - free measurements ' , found
* 23 * , 987997 ( 1993 ) .
p. kwiat , h. weinfurter , t. herzog , a. zeilinger , and m. a. kasevich , ` interaction - free measurement ' , phys .
lett . * 74 * , 47634766 ( 1995 ) .
p. g. kwiat , a. g. white , j. r. mitchell , o. nairz , g. weihs , h. weinfurter , and a. zeilinger , ` high - efficiency quantum interrogation measurements via the quantum zeno effect ' , phys .
. lett . * 83 * , 47254728 ( 1999 ) .
h. azuma , ` interaction - free generation of entanglement ' , phys .
a * 68 * , 022320 ( 2003 ) .
h. azuma , ` interaction - free quantum computation ' , phys .
a * 70 * , 012318 ( 2004 ) . | in this paper , we consider interaction - free measurement ( ifm ) with imperfect interaction . in the ifm proposed by kwiat _
et al_. , we assume that interaction between an absorbing object and a probe photon is imperfect , so that the photon is absorbed with probability @xmath0 ( @xmath1 ) and it passes by the object without being absorbed with probability @xmath2 when it approaches close to the object .
we derive the success probability @xmath3 that we can find the object without the photon absorbed under the imperfect interaction as a power series in @xmath4 , and show the following result : even if the interaction between the object and the photon is imperfect , we can let the success probability @xmath3 of the ifm get close to unity arbitrarily by making the reflectivity of the beam splitter larger and increasing the number of the beam splitters .
moreover , we obtain an approximating equation of @xmath3 for large @xmath5 from the derived power series in @xmath4 . |
the important ingredient for quantum computation and information processing is the presence of coherent superpositions .
a single isolated two - level system can be prepared in a coherent superposition of @xmath1 and @xmath2 states , and the manipulation of such states leads to new possibilities for storage and processing of information @xcite .
in contrast to the ideal isolated case , the interactions of real quantum systems with their environment lead to the loss of these coherent superpositions , in other words , decoherence .
however , the more realistic case would be manipulation of many qubits
. coherent superposition of such states leads to the concept of entanglement , which forms a precious resource for quantum computation and information .
the fragility of entanglement is due to the coupling between a quantum system and its environment ; such a coupling leads to decoherence , the process by which information is degraded @xcite .
in fact , decoherence is one of the main obstacles for the preparation , observation , and implementation of multi - qubit entangled states .
the intensive work on quantum information and computing in recent years has tremendously increased the interest in exploring and controlling decoherence effects @xcite . in this work
we address the problem where each of the two qubits are dissipatively coupled to a local bosonic bath ; in quantum optical sense it would mean that both the two - level systems are subject to spontaneous emission and would imply there exist relaxation between the excited state to ground state .
dissipation can assist the generation of entanglement @xcite that can be used for various quantum information processing .
for example , f. verstraete _ et al _ @xcite have shown that dissipation can be used as a resource for the universal quantum computation without any coherent dynamics needed to implement it .
contrary to other methods , entanglement generation by dissipation does not require the preparation of a system in a particular input state and exists , in principle , for an arbitrary long time .
these features make dissipative methods inherently stable against weak random perturbations , with the dissipative dynamics stabilizing the entanglement .
the effects on system due to environment can be classified into the process with memory ( non - markovian ) and without memory ( markovian ) effects @xcite . in case of markovian processes ,
the environment acts as a sink for the system information ; the system of interest loses information into the environment and this lost information plays no role in the dynamics of the system .
however , due to memory effects in case of non - markovian dynamics , the information lost by the system during the interaction with the environment will return back to the system at later time .
this makes the non - markovian dynamics complicated .
understanding the nature of non - markovian dynamics is naturally a very important topic for quantum information science , where the aim is to control a quantum system for use in technological applications @xcite . in general , three time scales in an open system
exist to characterize non - markovian dynamics : ( i ) the time scale of the system ( ii ) the time scale of the bath given by the bandwidth of bath spectral density ( iii ) the mutual time scale arising from the coupling between the system and the bath .
it is usually believed that non - markovian effects strongly rely on the relations among these different time scales @xcite . in this paper
we derive a quantum master equation for interacting qubits with local dissipation .
the equation is derived utilizing the completeness of the eigen basis of the hamiltonian representing the interacting qubits .
the time evolution of the density matrix turns out to be the sum of the time evolution corresponding to individual qubits with no cross terms .
next we solve this master equation for @xmath0-type states with the assumption that individual baths have the same properties .
the main content of this paper will remain however the same as for other kind of states and assuming different bath correlation functions for each bath .
the different bath correlation functions can give rise to different time scales in the dynamics and is treated separately .
next we identify different regimes of dynamics ( markovian and non - markovian ) and show that under non - markovian regimes of the dynamics , there exists finite entanglement in an initially unentangled state .
this entanglement decay in the same way as the pure state entanglement and we find that the decay rate of entanglement is strongly modified by the non - markovian behavior . the rest of the paper is organized as follows : in section ii , we introduce the model hamiltonian and derive the quantum master equation . in the past related works
@xcite , non - interacting qubits have been considered .
these qubits are then coupled to a common bath . however , in this paper , we consider qubits interacting through isotropic heisenberg interaction which is a general kind of interaction in condensed matter physics . in section iii , we solve the quantum master equation in the eigen - basis of the system hamiltonian for a general class of initial quantum states under the assumption that the bath correlation function decay in the same way.in section iv we give the decay of entanglement of certain @xmath0-type state .
finally we conclude in section v with the remarks of wider context of our results .
in this section we will first derive master equation for the reduced density matrix of the system which govern the dynamics of the system .
we consider two qubits represented by spin-@xmath3 particles or two level atoms coupled to each other via isotropic heisenberg interaction .
the qubits are subject to local dissipation through a coupling with a bosonic bath .
the hamiltonian of the two qubit system is @xmath4,\end{aligned}\ ] ] where @xmath5 .
@xmath6 represents the energy scale of the system .
the hamiltonian @xmath7 can be diagonalized exactly i.e. @xmath8 where @xmath9 s are the eigenstates of the hamiltonian @xmath7 with the eigen energies @xmath10 and are given below ( with notation @xmath11 and @xmath12 ): we write the total hamiltonian ( system+bath ) @xmath14 where @xmath15 is the hamiltonian for the bath @xmath16 and the dissipative interaction of the system with bath is represented by the hamiltonian @xmath17 \nonumber \\ & = & \sum_{i=1}^{2 } \sigma_i^x ( b_i + b_i^{\dagger})\end{aligned}\ ] ] where @xmath18 .
let @xmath19 represent an operator defined in interaction picture with respect to system and bath ( @xmath20 ) , we can therefore write @xmath21 in the interaction picture under rotating wave approximation as @xmath22.\end{aligned}\ ] ] the time evolution of the system operators can be evaluated using the eigen basis of the system hamiltonian @xmath7 as @xmath23 \\
\sigma_2^{+}(\tau ) & = & \sum_{i , j=1}^{4 } \!\!\ !
\langle \psi_i| i_1 \otimes \sigma_2^{+ } where @xmath24 is the projection operator @xmath25 and @xmath26 . in interaction picture , the time evolution of the total density matrix ( system and bath ) @xmath27 is given by the von neuman - liouville equation as @xmath28 .
\label{loe}\end{aligned}\ ] ] here we have used @xmath29
. we can formally integrate equation ( [ loe ] ) and write its solution as : @xmath30.\end{aligned}\ ] ] subsituting this solution back into the commutator of equation ( [ loe ] ) , we get upto second order the following equation : @xmath31 \nonumber \\ & -&\int\nolimits_{0}^{t}ds\left [ \tilde{h}% _ { i}\left ( t\right ) , \left [ \tilde{h}_{i}\left ( s\right ) , \tilde{% \rho}_{t}\left ( s \right ) \right ] \right].\end{aligned}\ ] ] the solution of the above equation depends on the initial conditions of total density operator .
we consider an initially uncorrelated situation , i.e @xmath32 , where @xmath33 and @xmath34 are respectively the density operators for the system and bath . tracing over the bath degrees of freedom and assuming that @xmath35=0 $ ] , we get the following time non - local master equation for the reduced density matrix : @xmath36 ] .\end{aligned}\ ] ] as the bath degree of freedom are infinite so that the influence of the system on the bath is small in the weak system - bath coupling case . as a cosequence , we write the total density operator @xmath37 within the second order perturbation of system - bath coupling @xcite . the replacement of total density matrix @xmath38 with an uncorrelated state @xmath39 is called as born approximation .
therefore , under born approximation we write @xmath40 ] .\end{aligned}\ ] ] the above equation is in a form of delayed integro - differential equation and is therefore a time non - local master equation . replacing @xmath41 with @xmath42 in this equation @xcite we get time - local master equation : @xmath43 ] .\end{aligned}\ ] ] assuming the bath in the vacuum state initially , i.e @xmath44 ; using the form of the @xmath45 , we arrive at the following equation : @xmath46 + \mathcal{l}_2[\tilde{\rho}_s(t)].\end{aligned}\ ] ] this forms a non - trivial result .
the master equations contains sums of @xmath47 for each qubit and no cross terms with different @xmath47 s .
this result is the same as that for the non - interacting qubits .
here we have ( @xmath48 ) and the bath correlation function is defined as @xmath50.\end{aligned}\ ] ] next we revert back to the schr@xmath51dinger picture with a change in variable @xmath52 , we write this represents the quantum master equation in the schr@xmath51dinger picture .
the solution of the above master equation depends on the type of initial states . in the next section
we find its solution for general @xmath0-type initial states .
in order to obtain the dynamics of entanglement of our two qubit system , we assume that the qubits are initially prepared in an x state @xcite : @xmath54\ ] ] where we have used the standard basis @xmath55 . since the normalization and positivity of @xmath56 i.e , @xmath57 and @xmath58 , the matrix elements @xmath59 are non - negative parameters with @xmath60 , @xmath61 , and @xmath62 .
we can use more general forms of density matrix with all elements non - zero , this makes the master equation intractable analytically .
next we express the x state @xmath56 in the eigen basis of @xmath7 as @xmath63 where the various parameters of the density operator in the eigen basis of @xmath7 are related to the parameters in the standard basis in the following way : @xmath64 , @xmath65 $ ] , @xmath66 $ ] , @xmath67 $ ] , @xmath68 , @xmath69 .
next we see that the form of the density matrix is invariant during the time evolution generated by the quantum master equation .
therefore we can the density matrix at time @xmath70 as @xmath71 in order to find the time evolution equations of the various parameters involved in equation ( [ rhoe ] ) we assume that the bath correlation functions have the same form @xmath72 where @xmath73 is the spectral width of the bath , @xmath74 is related to the microscopic system - bath coupling constant .
it defines the relaxation time scale @xmath75 over which the state of the system : @xmath76 .
it can be shown to be related to the markovian decay rate @xmath77 in markovian limit of flat spectrum .
this form of correlation function corresponds to the lorenztian spectral density of the bath @xcite . assuming that @xmath78 for simplicity , we substitute the @xmath79 as in equation ( [ rhoe ] ) in the quantum master equation ( [ qme ] ) and obtain the time dependence of the parameters as @xmath80
\\ c(t ) & = & c(0 ) e^{-is_1 t-\gamma(t ) } \\
h(t ) & = & h(0 ) e^{-is_2 t - \gamma(t)}\end{aligned}\ ] ] @xmath81 where @xmath82 ; @xmath83 ; @xmath84 ; @xmath85 ; @xmath86 and the explicit forms of these functions are given in the appendix a.
in this section we study entanglement of a two qubit system by means of concurrence @xcite . for a density matrix @xmath87 , concurrence is defined as @xmath88 , where @xmath89 , @xmath90 , @xmath91 and @xmath92 are the eigen values of matrix @xmath93 in the descending order .
the matrix @xmath93 is defined as @xmath94 and @xmath95 represents complex conjugation of @xmath87 in standard basis . for x state in the standard basis
we write concurrence as @xcite @xmath96 where we have @xmath97 , @xmath98 , @xmath99 $ ] , @xmath100 $ ] , @xmath101 $ ] , @xmath102 $ ] , @xmath103 .
next we use these results to investigate the decay of entanglement in some specific cases .
first we consider the decay of the pure entangled state @xmath104 .
this state has initial entanglement @xmath105 and at time @xmath70 we write with the help of above results @xmath106 or @xmath107 this forms an important result .
it shows that even though we have initially an unentangled state @xmath108 , we still have entanglement at later time @xmath70 . this can be attributed to the dissipative interaction between the system and the bath .
let us suppose @xmath109 , which corresponds to @xmath110 state ; the effect of dissipative interaction ( @xmath111 ) results in an entangled state . with time at different values of the parameter @xmath112 and @xmath93 .
here we have used @xmath113 .
plots for different @xmath112 values at ( a ) @xmath114 ( b ) @xmath115 ( c ) @xmath116 and ( d ) @xmath117 . plots ( a ) and ( b ) are in units of markovian decay rate @xmath77 , i.e @xmath118 , while plots ( c ) and ( d ) are in units of rescaled decay rate @xmath119 i.e @xmath120 , title="fig:",width=144,height=120 ] with time at different values of the parameter @xmath112 and @xmath93 .
here we have used @xmath113 .
plots for different @xmath112 values at ( a ) @xmath114 ( b ) @xmath115 ( c ) @xmath116 and ( d ) @xmath117 . plots ( a ) and ( b ) are in units of markovian decay rate @xmath77 , i.e @xmath118 , while plots ( c ) and ( d ) are in units of rescaled decay rate @xmath119 i.e @xmath120 , title="fig:",width=144,height=120 ] + with time at different values of the parameter @xmath112 and @xmath93 .
here we have used @xmath113 .
plots for different @xmath112 values at ( a ) @xmath114 ( b ) @xmath115 ( c ) @xmath116 and ( d ) @xmath117 .
plots ( a ) and ( b ) are in units of markovian decay rate @xmath77 , i.e @xmath118 , while plots ( c ) and ( d ) are in units of rescaled decay rate @xmath119 i.e @xmath120 , title="fig:",width=144,height=120 ] with time at different values of the parameter @xmath112 and @xmath93 .
here we have used @xmath113 .
plots for different @xmath112 values at ( a ) @xmath114 ( b ) @xmath115 ( c ) @xmath116 and ( d ) @xmath117 . plots ( a ) and ( b ) are in units of markovian decay rate @xmath77 , i.e @xmath118 , while plots ( c ) and ( d ) are in units of rescaled decay rate @xmath119 i.e @xmath120 , title="fig:",width=144,height=120 ] .
[ plot2 ] next we analyze the markovian and non - markovian regimes of the dynamics , for that we define the following parameters : @xmath121 . therefore
, using this parametrization we have @xmath122 } - \frac{[1 - 9q^2][1-e^{-r\tau } \cos(3qr\tau)]}{2r[1 + 9q^2]^2 } \nonumber\\ & & -\frac{3q}{r[1 + 9q^2]^2}e^{-r\tau}\sin(3qr\tau ) \\
\gamma_{-}(t)&= & \frac{\tau}{2[1 + q^2 ] } - \frac{[1-q^2][1-e^{-r\tau } \cos(qr\tau)]}{2r[1+q^2]^2 } \nonumber\\ & & -\frac{q}{r[1+q^2]^2}e^{-r\tau}\sin(qr\tau).\end{aligned}\ ] ] in order to understand how the markovian limit is obtained from the above expressions , we plot in fig .
[ plot2](a)-(b ) @xmath123 for @xmath113 with respect to dimensionless parameter @xmath124 for @xmath114 and @xmath125 at different values of @xmath112 .
we observe that the markovian curve is recovered for @xmath126 with @xmath127
. we can understand this behaviour of @xmath128 by looking at the different paramters involved .
the typical time scale over which the system of two qubits changes is @xmath129 and the time scale over which the bath changes is @xmath130 while relaxation time scale for each qubit would be given by @xmath131 .
it means @xmath126 implies @xmath132 and @xmath133 implies @xmath134 .
thus physically @xmath126 and @xmath133 would imply that the system evolves over a large time compared to very fast bath dynamics .
therefore markovian regime corresponds to @xmath126 and @xmath133 and we have @xmath135 and therefore we get standard markovian limit : @xmath136 we observe that under the markovian limit an initially unentangled state remains unentangled always . in situations where the spectral width @xmath137 of the bath is narrower than the energy scale @xmath6 involved for the system implying @xmath138 .
this would mean @xmath139 . in fig .
[ plot2](b ) , we observe for @xmath115 that as @xmath112 increases from 0 , there is larger devaition from the markovian dynamics of the concurrence @xmath128 .
the general trend is similar for all values of @xmath140 as can be seen on comparison to fig .
[ plot2](a ) .
these observations suggest markovian regime is in fact opposite to the regime @xmath141 and @xmath138 , which we call as _ non - markovian _ regime .
the term that is responsible for this larger deviation can be attributed to first terms of @xmath142 and @xmath143 containing @xmath144 in the denominator . for @xmath145
suggest defining another time scale @xmath146 the decay of entanglement defined by @xmath128 at various values of @xmath112 for non - markovian regime @xmath141 in terms of rescaled time @xmath147 is shown in the fig.[plot2](c)-(d ) .
we see that for large @xmath148 the concurrence @xmath149 coincides with exponential decay in units of the rescaled time .
next we see that before reaching the limiting behavior of exponential decay in rescaled time ( [ rescaled ] ) , we observe some oscillatory behavior figure [ plot2](c)-(d ) .
the deviation from an exponential decay can be attributed to the memory effects developed in the two qubit system .
this occurs clearly due to the second terms in @xmath142 and @xmath143 .
for @xmath150 we may approximate this as @xmath151 \label{approxgamma1}\\ \gamma_{+}(\tau ' ) & \approx & \frac{1}{18}\left [ \tau ' + \frac{1- e^{-r q^2 \tau ' } \cos ( 3r q^3 \tau')}{r q^2}\right ] .
\label{approxgamma2}\end{aligned}\ ] ] in order for the oscillatory term to be visible , we require the exponential decay term in ( [ approxgamma1])-([approxgamma2 ] ) to be not too fast giving @xmath152 , but simultaneously the oscillation frequency should be faster than the overall decay envelope @xmath153 .
the strongest oscillations therefore occur when @xmath154 , which agree with the numerical plots in figure [ plot2](c)-(d ) .
the deviation from an exponential decay can be attributed to the memory effects developed initially , typical of non - markovian behavior .
the criterion for the strongest oscillatory behavior are satisfied when all the characteristic time scales are all approximately the same i.e @xmath155 .
in conclusion , we have derived a quantum master equation for system of two interacting qubits under the influence of local dissipation . using the assumption that the correlation functions have the same form for each of the baths , the solution of master equation is found for the general @xmath0-type state .
the time dependence of concurrence , a measure of entanglement is studied for a pure entangled state @xmath104 ( a special case of @xmath0-type state ) under both markovian and non - markovian regimes of the dynamics .
it is found that for finite time evolution an unentangled state can go to an entangled state in contrast to the markovian case where the unentangled state remains unentangled always . by identifying the parameter space ,
we have found that our results reduce to standard markovian decay rate which in general is not a physically relevant regime @xcite . in the physically relevant regime with narrower spectral width as compared to @xmath6 ,
the decay rate is better approximated by @xmath156 , which is the standard markovian decay rate divided by the @xmath157 , which can be quite large in practice .
next we compare our work with the several other works that studied non - markovian dynamics of entanglement .
taking the example of ref .
@xcite , the authors derive the non - markovian decay of the entanglement for the pure state @xmath104 .
the time dependence of concurrence is given by : where @xmath159\end{aligned}\ ] ] and @xmath160 and @xmath77 is the markovian decay rate .
our result eqn .
( [ concu ] ) is more general than above result .
the results in these works @xcite do not tell about the amount of entanglement and its decay that would be present in an entangled state generated from dissipation . to see more clearly the behavior ,
let us examine this in two limiting cases _ weak coupling limit _
@xmath161 : this regime corresponds to a weak coupling regime or a very broad coupling to many frequency modes , which gives markovian behavior . here
@xmath162 and the decay function @xmath163 gives purely exponential behavior . to first order the decay function may be approximated as @xmath164 which is nothing but standard markovian spontaneous decay .
_ strong coupling limit _
@xmath165 : the reverse regime is when the linewidth of the bath is extremely narrow , which gives rise to strongly non - markovian behavior . here we may approximate @xmath166 and @xmath167 \nonumber \\\end{aligned}\ ] ] which corresponds to damped oscillations at frequency @xmath168 and a decay envelope with rate @xmath73 .
thus we see that in both the cases the previous results do not yield the scaling factor @xmath144 as derived in eq .
( 35 ) .
the current result would be important for applications where spontaneous emission is a serious drawback of using excited states , such as for quantum information processors , quantum simulators , and quantum metrological applications .
@xmath170}\left [ \lambda t -\frac{\lambda^2-j^2}{\lambda^2+j^2}[1-e^{-\lambda t}\cos jt ) ] \right.\nonumber \\ & & ~~~~~~~~~~~~~~~~~~\left .
- \frac{2\lambda j}{\lambda^2+j^2}e^{-\lambda t } \sin jt \right ] \\
\gamma_{+}(t ) & = & \frac{1}{2 } \int_0^t \!\!dz~\sigma(z ) \\ & = & \frac{\gamma_m \lambda}{2[\lambda^2 + 9j^2]}\left [ \lambda t -\frac{\lambda^2 - 9j^2}{\lambda^2 + 9j^2}[1-e^{-\lambda t}\cos 3jt ) ] \right.\nonumber \\ & & ~~~~~~~~~~~~~~~~~~\left .
- \frac{6\lambda j}{\lambda^2 + 9j^2}e^{-\lambda t } \sin 3jt \right ] \end{aligned}\ ] ] @xmath171}\left [ j t + \frac{2j}{\lambda^2+j^2}[1-e^{-\lambda t}\cos jt ) ] \right.\nonumber \\ & & ~~~~~~~~~~~~~~~~~~\left .
+ \frac{\lambda^2-j^2}{\lambda^2+j^2}e^{-\lambda t } \sin jt \right ] \\
s_{+}(t ) & = & \frac{\gamma_m \lambda}{2[\lambda^2 + 9j^2]}\left [ 3j t + \frac{6j}{\lambda^2 + 9j^2}[1-e^{-\lambda t}\cos3 jt ) ] \right.\nonumber \\ & & ~~~~~~~~~~~~~~~~~~\left .
+ \frac{\lambda^2 - 9j^2}{\lambda^2 + 9j^2}e^{-\lambda t } \sin3 jt \right ] \end{aligned}\ ] ] 47 m. a. nielsen and i. chuang , _ quantum computation and quantum communication _ ( cambridge university press ) @xmath172 .
m. schlosshauer , rev .
phys . * 76 * , 1267 @xmath173 w. h. zurek , rev .
mod . phys . * 75 * , 715(2003 ) .
j. t. barreiro , p. schindler , o. ghne , t. monz , m. chwalla , c. f. roos , m. hennrich and r. blatt , nature phys .
* 6 * , 943 ( 2010 ) .
s. schneider and g. j. milburn , phys .
a * 57 * , 3748 ( 1998 ) .
q. a. turchette , c. j. myatt , b. e. king , c. a. sackett , d. kielpinski , w. m. itano , c. monroe and d. j. wineland , phys .
a , * 62 * , 053807 ( 2000 ) . c. j. myatt , b. e. king , q. a. turchette , c. a. sackett , d. kielpinski , w. m. itano , c. monroe and d. j. wineland nature , * 403 * , 269 ( 2000 ) .
e. knill , r. laflamme and w. h. zurek , science * 279 * , 342 ( 1998 ) .
s. diehl , a. micheli , a. kantian , b. kraus , h. buechler , p. zoller , nature phys .
* 4 * , 878 ( 2008 ) .
f.verstraete , m. m.wolf , and j. i.cirac , nat .
phys.*5 * , 633 ( 2009 ) .
h. weimer , m. mller , i. lesanovsky , p. zoller and h.p .
bchler , nature phys .
* 6 * , 382 ( 2010 ) .
m. b. plenio et al . , phys . rev .
a 59 , 2468 ( 1999 ) .
beige et al . , j. mod .
47 , 2583 ( 2000 ) .
p. horodecki , phys .
a 63 , 022108 ( 2001 ) .
h. p. beuer and f. petruccione , _ the theory of open quantum systems _
( oxford , new york : oxford university press)(2005 ) .
r. lo franco , b. bellomo , s. maniscalco , and g. compagno , int . j. mod
b * 27 * , 1345053 ( 2013 ) .
b. bellomo , r. lo franco , and g. compagno , phys .
99 * , 160502 ( 2007 ) . b. bellomo , r. lo franco , and g. compagno , phys .
a * 77 * , 032342 ( 2008 ) .
k. m. fonseca romero and r. lo franco , phys .
* 86 * , 065004 ( 2012 ) .
m. q. lone , and t. byrnes , phys .
a * 92 * , 011401(*r * ) , ( 2015 ) .
b. m. terhal and g.burkard , phys .
a 71 , 012336 ( 2005 ) . m. ban , j. phys . a : math .
39 , 1927 ( 2006 ) .
m. m. wolf , j. eisert , t. s. cubitt , and j. i. cirac , phys .
lett . 101 , 150402 ( 2008 ) .
f. pastawski , l. clemente and j. i. cirac , phys .
a 83 , 012304 ( 2011 ) .
p. haikka and s. maniscalco , phys . rev .
a 81 , 052103 ( 2010 ) .
w. m. zhang , p. y. lo , h. n. xiong , m. w. y. tu , and f. nori , phys .
109 , 170402 ( 2012 ) .
m. a. cirone , g. de .
chiara , g. m. palma and a. recati , new j. phys . 11 103055 ( 2009 ) .
h. j. carmichael , _ statistical methods in quantum optics i _
( berlin : springer - verlag ) ( 2008 ) . h. p. breuer , b. kappler and f. petruccione , phys .
a * 59 * , 1633 ( 1999 ) .
h. p. breuer , b. kappler and f. petruccione , ann .
( ny ) * 291 * , 36 ( 2001 ) .
m. schrder , u. kleinekathfer and m. schreiber , j. chem . phys . * 124 * 084903 ( 2006 ) .
e. ferraro , m. scala , r. migliore and a. napoli , phys .
a * 80 * , 042112 ( 2009 ) .
t. yu and j. h. eberly , quantum inf .
comput . 7 , 459 ( 2007 ) w.k.wooters , phys . rev.lett . *
80 * , 2245 ( 1998 ) . | we investigate the dynamics of entanglement given by the concurrence of a two - qubit system in the non - markovian setting .
a quantum master equation is derived which is solved in the eigen basis of the system hamiltonian for @xmath0-type initial states .
a closed formula for time evolution of concurrence is presented for a pure state .
it is shown that under the influence of dissipation that non - zero entanglement is created in unentangled two qubit states which decay in the same way as pure entangled states .
we also show that under real circumstances , the decay rate of concurrence is strongly modified by the non - markovianity of the evolution . |
in random flows , in particular in chaotic or turbulent ones , the increase of the local , instantaneous gradient of a scalar such as temperature or concentration results in the enhancement of mixing through acceleration of molecular diffusion .
this `` gradient production '' is caused by the mechanical action of the velocity gradient , more precisely by its straining part , either in the pure straining motion of hyperbolic regions or in the vicinity of vortices where strain stems from differential rotation .
quite equivalently , at least as far as only convective mechanisms are considered , hastening of the mixing process in random flows finds expression in the stretching of material surface areas . in turbulent
flows the growth of the scalar gradient is closely connected to the production of small scales of a scalar field and to cascade phenomena . actually , the mean dissipation rate of the energy of fluctuations of a scalar is proportional to the variance of its fluctuating gradient . in the view
where mixing properties of flows are explained through the evolution of the gradient of a scalar the key mechanisms rest on the alignment of the gradient with respect to the local strain principal axes @xcite .
the rise of the scalar gradient is indeed promoted by alignment with the most compressive direction of strain .
but the general problem of scalar gradient alignment is quite complicated . in turbulent
flows the question amounts to understanding how the gradient behaves under the combined actions of molecular diffusion as well as of fluctuating strain and effective rotation ( i.e. vorticity plus strain basis rotation ) . in three - dimensional turbulence random alignment of vorticity and vorticity / strain interaction
make the problem even more complex . from pure kinematical considerations it is obvious that in incompressible flows the scalar gradient is drawn by the local compressional direction .
yet it is also quite understandable that the latter is generally not the equilibrium orientation of the scalar gradient . in other words , when not only strain but also rotation and , possibly , molecular diffusion are present , the fixed point of the gradient orientation equations is certainly not the strain compressional axis . in three - dimensional flows ,
the bare existence of this fixed point is even not proved in the general case @xcite . in two - dimensional flows
@xcite as well as in a special , three - dimensional situation @xcite , however , the equilibrium orientation and conditions for its existence have been derived analytically at least when the scalar is non - diffusive .
it has also been shown that in two - dimensional turbulence the scalar gradient statistically aligns better with the local equilibrium direction defined by the balance between strain and rotation than with the compressional direction @xcite .
even so , the statistical alignment of the scalar gradient is determined by the gradient dynamics . if the velocity field is time varying , alignment with a possible equilibrium orientation requires that the response time scale of the scalar gradient is short enough compared to the time scale of velocity gradient fluctuations .
alignment dynamics is thus an essential feature of the scalar gradient behaviour and has already been addressed in some studies @xcite .
the influence on mixing that follows from this nonstationary aspect is still not clear .
trying to understand how the mixing process is affected by the dynamics of the scalar gradient is precisely one of the main goals of the present study .
this is done in section [ sec3 ] by analyzing numerically the growth rate of the gradient norm for different unsteady regimes in the two - dimensional case .
we first devote section [ sec2 ] to the numerical analysis of alignment statistics when the scalar gradient undergoes a fluctuating strain persistence .
this investigation aims at bearing out the generality of the partial findings of ref . @xcite .
in particular , the orientation equation of the scalar gradient in a two - dimensional flow is solved for simulating the strain - dominated as well as the effective - rotation - dominated regimes .
different strain persistence signals are used to show the existence of a gradient alignment that is neither the compressional nor the equilibrium direction defined by the instantaneous strain persistence .
section [ sec4 ] reports on experimental results supporting the latter numerical study .
conclusion is drawn in section [ sec5 ] .
we restrict the analysis to two - dimensional , incompressible flow and non - diffusive passive scalar . writing the scalar gradient in the fixed frame of reference as @xmath0 ,
the equation for the gradient orientation is @xcite @xmath1 where @xmath2 gives the gradient orientation in the local strain basis ; angle @xmath3 defines the orientation of the strain principal axes through @xmath4 with @xmath5 and @xmath6 denoting the normal and shear components of strain , respectively .
time @xmath7 is a strain - normalized time @xmath8 with @xmath9 standing for the lagrangian time and @xmath10 for strain intensity , @xmath11 .
note that eq .
( [ eq1 ] ) just proceeds from the usual equation for the passive scalar ( with molecular diffusion neglected ) through the equation for the scalar gradient @xcite which is handled as explained in ref .
@xcite .
parameter @xmath12 represents the persistence of strain @xcite and is defined as @xmath13 where @xmath14 is vorticity .
strain persistence , @xmath12 , defines an objective criterion for partitioning the flow ; in strain - dominated regions @xmath15 , while in regions where effective rotation prevails @xmath16 . in this article these regions
are occasionally termed hyperbolic and elliptic respectively .
it is to be noticed that this partition does not necessarily coincide with the one derived from the okubo - weiss criterion ; because @xmath12 includes the pressure hessian through strain basis rotation , @xmath17 , the corresponding criterion is more general @xcite .
lapeyre et al .
@xcite analyzed eq .
( [ eq1 ] ) assuming slow variations of @xmath12 along lagrangian trajectories and showed the way in which the scalar gradient orientation is determined by the local flow topology . for prevailing strain the orientation equation has a stable fixed point , @xmath18 which corresponds to an equilibrium orientation . in the special case @xmath19 ( i.e. in the pure hyperbolic regime ) the equilibrium orientation , @xmath20 , coincides with the local compressional direction , @xmath21 if effective rotation dominates , eq . ( [ eq1 ] ) has no fixed point ; there is no equilibrium orientation for the scalar gradient , but a most probable one given by @xmath22 \pi/2.\ ] ] exact balance between strain and effective rotation , namely @xmath23 , drives the scalar gradient to align with the bisector of strain principal axes . the general solution , @xmath24 , of eq . ( [ eq1 ]
) can be derived for any of the latter three different regimes @xcite . the study of garcia et al .
@xcite reveals that the above approach remains valid as long as the response time scale of the scalar gradient is short enough compared to the time scale of the lagrangian fluctuations of @xmath12 .
they also put forward that in the opposite case , namely when the gradient does not keep up with @xmath12 fluctuations and its response is poor , the alignment of the scalar gradient is determined by the mean value of @xmath12 , @xmath25 . in the following we try to generalize and support these results by the study of regimes that have not been originally addressed .
indeed we extend the previous analysis by examining the effective - rotation - dominated regime in addition to the strain - dominated one . considering different values of @xmath25 we also confirm the existence of the orientation defined by the mean value of strain persistence .
we simulate the fluctuations of strain persistence , @xmath12 , with a stochastic differential equation , @xmath26 in eq .
( [ eq2 ] ) @xmath27 is a standardized gaussian random variable .
this equation depends on three parameters , namely @xmath28 , @xmath25 and @xmath29 .
giving @xmath28 a value is equivalent to prescribe the nondimensional integral time scale of @xmath12 , @xmath30 , through @xmath31 .
parameter @xmath29 stands for the root mean square of @xmath12 .
the few data on strain persistence statistics @xcite show it is not a gaussian variable .
however , the actual statistics are immaterial here , for the concern is mainly in the gradient response to time - varying strain persistence . the evolution of the scalar gradient orientation , then , is derived by solving numerically @xmath32 \ , d \tau.\ ] ] since @xmath33 , it is clear from eq .
( [ eq3 ] ) that the response time scale of the gradient orientation to the forcing mechanism represented by @xmath12 is of the order of @xmath34 .
parameter @xmath28 gives a measure of the gradient response compared to forcing stimulation .
small values of @xmath28 ( resp .
slow variations of @xmath12 ) mean that the scalar gradient responds well to @xmath12 fluctuations , while large values ( resp .
fast variations of @xmath12 ) correspond to a poor response of the scalar gradient .
in the special case for which strain intensity , @xmath10 , is constant @xmath35 where @xmath36 is the integral time scale of @xmath12 signal and thus @xmath28 is the ratio of gradient response time scale to @xmath12 fluctuations time scale .
garcia et al . @xcite
have shown that the approach of lapeyre et al .
@xcite ( section [ sec2.1 ] ) does not apply to scalar gradient orientation if strain persistence fluctuates too rapidly for the scalar gradient to respond .
they argued that in this case the most probable orientation of the gradient should be determined by the mean value of @xmath12 , @xmath25 , rather than by the instantaneous one .
in particular , it was put forward that if @xmath37 ( in other words , if the regime is hyperbolic _ on an average _ ) and the time scale of @xmath12 fluctuations is shorter than the gradient response time scale , the preferential orientation of the gradient should be defined by @xmath38 in reality , the most probable orientation can not be determined analytically , but is close to @xmath39 .
this results from the fact that the weaker the gradient response , the closer @xmath40 to @xmath39 . a simplified proof is given in appendix [ appa ] .
the behaviour of the scalar gradient when its response to @xmath12 fluctuations is poor is confirmed by the present simulations for two different values of @xmath25 as well as for small and large values of @xmath28 .
the p.d.fs of scalar gradient alignment have been derived from the numerical evolution of @xmath40 computed using eq .
( [ eq3 ] ) and eq .
( [ eq2 ] ) with @xmath41 and @xmath42 .
the fluctuations of @xmath12 thus describe the hyperbolic regime ( @xmath15 ) , but also make significant inroads into the elliptic one ( @xmath16 ) .
the p.d.fs of @xmath43 and @xmath44 conditioned on @xmath15 are shown in fig .
they indicate which of the orientations @xmath20 ( determined by the instantaneous value of @xmath12 ; section [ sec2.1 ] ) and @xmath39 ( defined by @xmath25 ) is statistically the best one in the hyperbolic regime .
clearly , for @xmath45 the scalar gradient preferentially aligns with the instantaneous , equilibrium direction , @xmath20 , which is generally different from the compression one , in agreement with the approach by lapeyre et al .
for @xmath46 it is the alignment with the direction defined by @xmath39 which is the most probable one .
note that for @xmath41 @xmath47 and thus coincides with the compressional direction .
interestingly , in the present case the scalar gradient thus statistically aligns with the compressional direction because the gradient response to @xmath12 fluctuations is poor ( @xmath48 ) and the mean value of @xmath12 corresponds to a pure hyperbolic regime ( @xmath41 ) .
this is reminiscent of the experimental situation analyzed by garcia et al .
@xcite .
the results in the elliptic regime ( p.d.f s conditioned on @xmath49 ; fig .
[ fig2 ] ) are similar . for @xmath45
the most probable gradient orientation is given by @xmath50 ( section [ sec2.1 ] ) which agrees with the analysis of lapeyre et al .
@xcite . in the case
@xmath46 the scalar gradient again aligns preferentially with the direction defined by @xmath39 , that is , the compressional one .
yet this is not paradoxical .
when the scalar gradient does not respond to the fluctuations of @xmath12 its orientation is governed by the mean value , @xmath25 , rather than by the instantaneous one and remains close to @xmath39 .
this may drive the gradient to align with the compressional direction , even though @xmath12 takes elliptic values , provided that @xmath25 is close to zero .
note that the secondary maxima in fig .
[ fig2 ] are explained by @xmath51 .
alignment p.d.fs derived with @xmath52 are plotted in figs .
[ fig3 ] and [ fig4 ] .
if @xmath28 is small the value of @xmath25 is immaterial ; the scalar gradient preferentially aligns with the equilibrium direction , @xmath20 , and the same orientation p.d.fs as those plotted in figs .
[ fig1 ] and [ fig2 ] for @xmath45 are derived .
figures [ fig3 ] and [ fig4 ] thus only display the orientation p.d.fs in the case where @xmath28 is given a large value , @xmath53 .
results for the hyperbolic regime are presented in fig .
[ fig3 ] . plainly , @xmath39 ( which is computed as @xmath54 ) is the most probable orientation of the scalar gradient .
alignment with the equilibrium direction , @xmath20 , is weak and as shown by the p.d.f of @xmath55 there is no trend of the gradient to align with the compressional direction . in the elliptic regime ( fig .
[ fig4 ] ) , too , the orientation defined by @xmath25 is statistically the best one . as a matter of course
, the gradient does not align with the compressional direction , but there is also no trend at all toward the direction defined by @xmath50 .
with the same assumptions as those stated in section [ sec2.1 ] the equation for the norm of the scalar gradient is @xcite @xmath56 from which it is clear that the mean growth rate of the gradient norm , @xmath57 , is given by @xmath58 .
note that alignment with the compressional direction , @xmath59 , corresponds to the maximum growth rate .
the mean growth rate has been derived from the statistics of @xmath40 computed by solving eqs .
( [ eq2 ] ) and ( [ eq3 ] ) .
the simulations have been run for three different @xmath12 signals , namely @xmath60 , @xmath42 and @xmath61 with @xmath28 ranging from 0.1 to 100 .
for all cases @xmath41 .
figure [ fig5 ] shows the evolution of @xmath57 _ vs. _ @xmath62 .
interestingly , @xmath57 decreases with @xmath62 whatever the value of @xmath63 ; in other words , the better the response of gradient orientation to @xmath12 fluctuations ( large values of @xmath62 ) , the lower the mean growth rate of the norm . when it responds to @xmath12 fluctuations the scalar gradient preferentially aligns with either @xmath20 or @xmath50 which are determined by the instantaneous value of @xmath12 and are both mostly different from the compressional direction ( section [ sec2.2 ] ) . hence the weaker growth rate .
this feature gets more pronounced as the variance @xmath63 is increased . for the lowest variance ( @xmath64 )
@xmath12 mostly fluctuates within the bounds of the hyperbolic regime assuming low values around @xmath19 and makes @xmath20 not much different from the compressional direction .
wider fluctuations ( @xmath42 or @xmath61 ) , by contrast , bring about instantaneous orientations @xmath20 corresponding to larger values of @xmath65 and thus lying further from the compressional direction ; in addition , more frequent , deep inroads into the elliptic regime ( @xmath16 ) even lead the gradient to align with @xmath50 , that is , @xmath66 away from the compressional direction , during longer time intervals . as a result , for moderate and large values of @xmath62 the mean growth rate of the scalar gradient norm
is lowered as the amplitude of @xmath12 fluctuations is increased . for the smallest values of @xmath62 , though , the growth rate is insensitive to @xmath12 variance and is close to its maximum value . indeed a sluggish response to @xmath12 fluctuations compels the gradient to remain aligned with the direction defined by @xmath25 ( section [ sec2.2 ] ) which in the present case ( for which @xmath41 ) coincides with the compressional direction .
there is an interesting consequence for mixing properties . if @xmath67 , then favourable conditions for mixing are not only achieved in the hyperbolic regime ; provided that the gradient response is weak , they are also fulfiled in those cases where large fluctuations of @xmath12 span both the hyperbolic and elliptic regimes .
more generally , when @xmath12 fluctuates with @xmath25 remaining close to 0 it is the poor response of the scalar gradient to @xmath12 fluctuations that promotes the best conditions for fast mixing .
as far as we know , simultaneous measurements of velocity and scalar gradients are scarce . apart from experiments in turbulent flows @xcite
, joint statistics of velocity gradient and temperature gradient have been measured in a low - reynolds number , two - dimensional , bnard - von krmn street @xcite .
the latter data confirm that for fast fluctuations of strain persistence the preferential orientation of the scalar gradient is determined by the mean strain persistence rather than by its instantaneous value .
a detailed description of the experiment and measurement techniques are given in refs .
@xcite . in brief , the experimental set - up consists of a 2 mm - diameter ( @xmath68 ) circular cylinder used for generating a two - dimensional bnard - von krmn street ; the reynolds number based on the cylinder diameter is @xmath69 .
temperature is passively injected through a 20 @xmath70m - diameter heated line source located in the near wake of the cylinder ( fig .
[ fig6 ] ) . the line source can be set either in or off the centre of the street .
velocity gradients and temperature gradients are derived from simultaneous measurements of temperature and velocity components .
these data have been used to obtain lagrangian statistics of the scalar gradient orientation conditioned on the strain persistence parameter , @xmath12 , as explained in refs .
@xcite .
part of the latter results have been already analyzed @xcite .
in particular , it has been found that in this experiment the temperature gradient does not respond to @xmath12 fluctuations .
more precisely , @xmath71 , where the mean value of strain , @xmath72 , and @xmath36 , the autocorrelation time scale of @xmath12 , have been computed by averaging over lagrangian trajectories .
this previous statistical study was mainly focused on the hyperbolic zones of the flow , far enough downstream ( i.e. @xmath73 with @xmath74 the distance of the source to the cylinder ) , when the heated line source is in the centre of the cylinder wake .
the analysis showed ( fig .
[ fig7 ] ) that the alignment of the temperature gradient with the compressional direction is statistically better than with the equilibrium direction derived from the approach of lapeyre et al .
@xcite ( section [ sec2.1 ] ) .
this result lends support to the findings of the numerical study presented in section [ sec2.2 ] .
most likely , preferential alignment of the temperature gradient with the compressional direction does not result from any kinematic attraction , but from the gradient dynamics .
because of the poor response of the temperature gradient to @xmath12 fluctuations , the preferential alignment is given by @xmath25 computed along lagrangian trajectories which , in the above conditions , is found to be close to 0 and thus defines an orientation that almost coincides with the compressional direction .
previous studies @xcite suggest that molecular diffusion may influence scalar gradient orientation , but that large gradients are weakly affected . as to the present case , garcia et al . @xcite
have shown that the departure of the experimental results from the lapeyre et al .
approach is not to be ascribed to molecular diffusion .
the latter certainly plays on the shape of the orientation p.d.fs , but is not the cause for the compressional direction being more probable than the equilibrium one .
further experimental data for the elliptic regions of the flow and in the case where the line source is located off the centre of the cylinder wake confirm the above picture .
results for elliptic regions when the line source is in the centre of the wake are displayed in fig .
they concern the p.d.fs of orientation of the temperature gradient with respect to the compressional direction , @xmath75 , and to direction @xmath50 imposed by dominating effective rotation @xcite ( section [ sec2.1 ] ) .
the p.d.f of orientation with respect to direction @xmath76 is also shown .
the latter direction corresponds to the lowest eigenvalue of tensor @xmath77 which is defined as : @xmath78 and is related to the pressure hessian @xcite .
according to klein et al .
@xcite , when strain intensity varies significantly along lagrangian trajectories @xmath76 is the preferential orientation of the scalar gradient in effective - rotation - dominated regions . from fig .
[ fig8 ] it is clear that the temperature gradient does not tend to align with @xmath50 . there is a trend of the gradient to align with @xmath76 , but the best alignment is with the compressional direction .
this striking behaviour strongly pleads for the above - described scenario and for the conclusions of the numerical analysis of section [ sec2.2 ] : when strain persistence fluctuations are too fast for the scalar gradient to respond ( which in the experiment is measured by the low value of @xmath79 ) the latter is blind to the local topology and its orientation is entirely governed by the mean value of the strain persistence parameter .
when the heated line source is set off the centre of the bnard - von krmn street ( more precisely , at transversal distance @xmath80 from the previous position ) the mean value @xmath25 of strain persistence computed for lagrangian trajectories along which the temperature gradient keeps significant values ( @xmath81 ) is found to be close to -0.8 .
figure [ fig9 ] shows the p.d.fs of temperature gradient orientation with respect to either the compressional direction or the equilibrium direction , @xmath20 , derived from the approach of lapeyre et al .
@xcite , in strain - dominated regions .
there is a trend of the gradient to align with @xmath20 .
however , a much better statistical alignment is found with the direction corresponding to @xmath82 , that is , with @xmath83 .
p.d.fs of fig .
[ fig9 ] are more distributed and thus display lower peaks than those corresponding to the source being in the centre of the cylinder wake ( fig .
[ fig7 ] ) .
this may result from a bigger influence of rotation on the temperature gradient when the line source is off the centre of the wake .
orientation statistics in effective - rotation - dominated regions have similar trends .
figure [ fig10 ] shows that the temperature gradient aligns slightly better with @xmath76 than with @xmath50 .
but in this case , too , the best statistical alignment
broader though the p.d.f is is found around @xmath84 . these latter experimental data with @xmath85
thus give a more general support to the previous findings on the dynamics of scalar gradient orientation .
analysis of the behaviour of the passive scalar gradient undergoing the influence of varying velocity derivatives through fluctuating strain persistence uncovers some interesting results regarding both the statistical gradient orientation and mixing properties . generalizing the analysis of garcia et al .
@xcite to the effective - rotation - dominated regime , we confirm that the statistics of scalar gradient alignment depend on the gradient response to strain persistence fluctuations .
more precisely , the numerical study based on the equation for the scalar gradient orientation in two - dimensional flow shows that perfect response drives the gradient to preferentially align with a direction determined by the instantaneous value of strain persistence as predicted by the analysis of lapeyre et al . @xcite .
when the response of the scalar gradient is poor , though , the preferential alignment is given by the mean value of strain persistence which indicates whether the flow regime is , on an average , either strain- or effective - rotation - dominated .
it follows interestingly that the scalar gradient preferentially aligns with the compressional direction provided that it is almost insensitive to strain persistence fluctuations and the mean strain persistence is close to the pure hyperbolic value .
this result opposes the usual statement that it is a perfect response to the fluctuating rotation of the strain basis ( which in the two - dimensional case is explicit in the strain persistence parameter ) that leads the gradient to align with the compressional direction . the general picture derived from the numerical study is firmly supported by experimental , lagrangian joint statistics of velocity gradient and scalar gradient derived from simultaneous measurements of velocity and temperature in a two - dimensional , low - reynolds number bnard - von krman street .
it is worth noting that its low reynolds number marks this flow from the simulated two - dimensional turbulent flows in which scalar gradient alignment was previously studied @xcite . in particular , it may be that alignment with the equilibrium direction found by lapeyre et al .
@xcite results from a good response of the scalar gradient to @xmath12 fluctuations in turbulent flows ( moderate or large @xmath86 ) . checking this surmise
would require lagrangian data on strain persistence in two - dimensional turbulence .
regarding the lagrangian properties of strain persistence and scalar gradient dynamics , another interesting and unanswered issue is whether the present flow is a special one or belongs to a more general class of flows .
finally , unsteady behaviour of the scalar gradient may unexpectably affect mixing properties through alignment statistics .
an interesting result is that poor response of the gradient to strain persistence fluctuations does not inevitably oppose efficient mixing .
in particular , the study of the growth rate of the gradient norm clearly shows that when the mean strain persistence is close to the pure hyperbolic value ( in other words , strain is persistent on an average ) it is the weak response of the gradient to strain persistence fluctuations which promotes the highest growth rate and thus the best conditions for mixing .
we restrict to the case where @xmath37 and @xmath87 is defined .
we get rid of the nonlinearity of eq .
( [ eqa1 ] ) by assuming small values of @xmath92 . to first order , then : @xmath93 with @xmath94 .
the proof is thus not general , but is given as an additional argument for the behaviour of the scalar gradient when its response to @xmath12 fluctuations is poor .
we assume a sine signal for @xmath29 , @xmath95 , in which frequency is normalized as @xmath96 with @xmath97 .
the nondimensional frequency , @xmath98 , thus gives a measure of the response of the scalar gradient to @xmath12 fluctuations .
in particular , poor response is expected for large values of @xmath98 . the solution of eq .
( [ eqa2 ] ) is : @xmath99 \exp(-\gamma \tau ) + \frac{a}{\gamma(1 + { \omega^{\star}}^2/\gamma^2 ) } \left(\sin \omega^ * \tau - \frac{\omega^{\star}}{\gamma } \cos \omega^{\star } \tau \right).\ ] ] note that eq .
( [ eqa3 ] ) can be recast in the more standard form : @xmath100 \exp(-\gamma \tau ) + \frac{a}{\gamma{(1+{\omega^{\star}}^2/\gamma^2)}^{1/2 } } \sin(\omega^{\star } \tau - \phi)\ ] ] with @xmath101 and phase , @xmath102 , is present when the response to the stimulation is not perfect .
it follows that when @xmath12 fluctuates on a much shorter time scale than the response time scale of the scalar gradient so much so that the latter does not keep up with @xmath12 fluctuations the gradient orientation , @xmath40 , remains close to @xmath39 .
increasing the amplitude of @xmath12 fluctuations , however , leads to the opposite behaviour .
00 a. pumir , phys .
fluids * 6 * , 2118 ( 1994 ) .
w. d. smyth , j. fluid mech .
* 401 * , 209 ( 1999 ) .
p. vedula , p. k. yeung , r. o. fox , j. fluid mech .
* 433 * , 29 ( 2001 ) .
g. brethouwer , j. c. r. hunt , f. t. m. nieuwstadt , j. fluid mech .
* 474 * , 193 ( 2003 ) .
m. tabor , i. klapper , chaos , solitons fractals * 4 * , 1031 ( 1994 ) .
g. lapeyre , p. klein , b. l. hua , phys .
fluids * 11 * , 3729 ( 1999 ) .
a. garcia , ph .
d. thesis , university of rouen , 2006 .
a. garcia , m. gonzalez , phys .
fluids * 18 * , 058101 ( 2006 ) . g. lapeyre , b. l. hua , p. klein , phys .
fluids * 13 * , 251 ( 2001 ) .
p. klein , b. l. hua , g. lapeyre , physica d * 146 * , 246 ( 2000 ) .
a. garcia , m. gonzalez , p. paranthon , phys .
fluids * 17 * , 117102 ( 2005 ) .
e. dresselhaus , m. tabor , j. fluid mech . * 236 * , 415 ( 1991 ) .
b. l. hua , p. klein , physica d * 113 * , 98 ( 1998 ) .
b. galanti , g. gulitsky , m. kholmyansky , a. tsinober , s. yorish , in _ advances in turbulence x _ , proceedings of the 10th european turbulence conference , edited by h. i. andersson and p .- .
krogstad , trondheim , norway , 2004 .
g. godard , ph.d .
thesis , university of rouen , 2001 .
g. godard , m. gonzalez , p. paranthon , in _ advances in turbulence ix _ , proceedings of the 9th european turbulence conference , edited by i. p. castro , p. e. hancock and t. g. thomas , southampton , united kingdom , 2002 .
p. paranthon , g. godard , f. weiss , m. gonzalez , int .
j. heat mass transfer * 47 * , 819 ( 2004 ) .
p. constantin , i. procaccia , d. segel , phys .
e * 51 * , 3207 ( 1995 ) .
. 1 p.d.fs of scalar gradient orientation for @xmath41 and @xmath42 conditioned on @xmath15 ( dominating strain ) .
p.d.f of @xmath108 : @xmath109 @xmath46 , @xmath110 @xmath45 ; p.d.f of @xmath111 : @xmath112 @xmath46 , @xmath113 @xmath45 .
2 p.d.fs of scalar gradient orientation for @xmath41 and @xmath42 conditioned on @xmath16 ( dominating effective rotation ) .
p.d.f of @xmath108 : @xmath109 @xmath46 , @xmath110 @xmath45 ; p.d.f of @xmath114 : @xmath112 @xmath46 , @xmath113 @xmath45 .
. 3 p.d.fs of scalar gradient orientation for @xmath52 , @xmath42 and @xmath53 conditioned on @xmath15 ( dominating strain ) .
@xmath109 p.d.f of @xmath44 ; @xmath112 p.d.f of @xmath43 ; @xmath115 p.d.f of @xmath55 .
4 p.d.fs of scalar gradient orientation for @xmath52 , @xmath42 and @xmath53 conditioned on @xmath16 ( dominating effective rotation ) .
@xmath109 p.d.f of @xmath44 ; @xmath112 p.d.f of @xmath116 ; @xmath115 p.d.f of @xmath55 .
experimental p.d.fs of temperature gradient orientation conditioned on @xmath15 ( dominating strain ) and @xmath118 in the far field of the heated line source ( @xmath119 ) .
the source is in the centre of the cylinder wake ; experimental @xmath67 .
@xmath110 p.d.f of @xmath55 ; @xmath113 p.d.f of @xmath43 . fig .
8 experimental p.d.fs of temperature gradient orientation conditioned on @xmath16 ( dominating effective rotation ) and @xmath118 in the far field of the heated line source ( @xmath119 ) .
the source is in the centre of the cylinder wake ; experimental @xmath67 .
@xmath110 p.d.f of @xmath55 ; @xmath113 p.d.f of @xmath116 ; @xmath115 p.d.f of @xmath120 . fig .
9 experimental p.d.fs of temperature gradient orientation conditioned on @xmath15 ( dominating strain ) and @xmath118 in the far field of the heated line source ( @xmath119 ) .
the source is off the centre of the cylinder wake ( @xmath121 ) ; experimental @xmath122 .
@xmath110 p.d.f of @xmath55 ; @xmath113 p.d.f of @xmath43 .
10 experimental p.d.fs of temperature gradient orientation conditioned on @xmath16 ( dominating effective rotation ) and @xmath118 in the far field of the heated line source ( @xmath119 ) .
the source is off the centre of the cylinder wake ( @xmath121 ) ; experimental @xmath122 .
@xmath110 p.d.f of @xmath55 ; @xmath113 p.d.f of @xmath116 ; @xmath115 p.d.f of @xmath120 . | the dynamics of a passive scalar gradient experiencing fluctuating velocity gradient through the lagrangian variations of strain persistence is studied .
to this end , a systematic , numerical analysis based on the equation for the orientation of the gradient of a non - diffusive scalar in two - dimensional flow is performed .
when the gradient responds weakly its orientation properties are determined by the mean value of strain persistence .
statistical alignment of the scalar gradient with the direction defined by the opposed actions of strain and rotation , by contrast , requires the gradient to keep up with strain persistence fluctuations .
these results have been obtained for both strain- and effective - rotation - dominated regimes and are supported by relevant experimental data .
consequences of the unsteady behaviour of the scalar gradient on mixing properties are also analyzed .
, , passive scalar gradient , alignment properties , mixing properties 47.51.+a , 47.61.ne , 47.27.-i |
electricity demand in the residential sector can be decomposed into a combination of individual appliances aggregated by individual households .
these appliances are tied together through different activities performed by users throughout a day and each of these activities may involve one or more of these power consuming devices .
these appliances are conventionally managed by each user according to his / her preferences , e.g. one may decide to wash clothes early in the morning before he leaves for work , and washing clothes is an activity or task which involves the use of washing machine , dryer etc .
different users can perform this task at different hours of the day according to their convenience . and many of such acvitives / task are flexible and can be performed at any time during a day . on the other hand , there may be certain activities which can be regarded as essential and which needs to be performed daily at exactly specified time slots e.g. after sunset from 7 pm till mid night one has to turn on the lights . such activities and the devices involved in these activities then contribute towards electricity load which is essential and which has strict scheduling requirements . in a traditional grid , the dominant setup has been to serve the preferences of the users as the priority need and match electricity supply to the instantaneous demand .
this however requires constant manipulation of electricity production levels . as a consequence power generating plants
suffer large deviations from their steady operating points which impose additional costs to the overall system .
all this is changing as the grid is becoming smart @xcite-@xcite .
a smart grid can help the operator in shaping the demand ( e.g. schedule the washing machine at a later time slot when there is less demand ) so as to reduce the overall societal cost for them , this can be done through the flattening of the demand curve @xcite . to achieve a flatter demand curve
, it can propose incentives ( e.g. discount ) to users to change their preference levels for different activities .
users can then allow the grid to manage and schedule certain appliances to enjoy these benefits at the expense of suffering some level of inconvenience . in this paper , we attempt to quantify the inconvenience levels , by varying the number of appliances that participate through deviation from their preferred scheduling time slots and also by varying the number of time slots each activity deviates .
we can thus identify a compromise between the grid operator objectives and user convenience levels .
we believe such understanding is beneficial for the grid to design effective incentive to achieve load balancing in a smart grid .
there are some recent studies on this problem . in @xcite authors design incentives and propose scheduling algorithms considering strictly convex functions of costs .
users are given incentives to move to off peak hours and these incentives are proposed using game theoretic analysis .
however they do not consider or quantify the inconvenience levels of the users . in
@xcite authors propose pricing scheme for users in order to achieve a perfectly flat demand curve .
they show that finding an optimum schedule is np - hard problem .
they propose centralized and distributed algorithms depending on the degree of knowledge of the state of the network .
the authors in @xcite propose a strategy to achieve a uniform power consumption over time .
their algorithm schedules the devices in such a way that a target power level is not exceeded in each time slot .
however again the authors do not take into account the inconvenience level of users while designing these algorithms .
in @xcite the authors use convex optimization tools and solve a cooperative scheduling problem in a smart grid .
the authors in @xcite use a water - filling based scheduling algorithm to obtain a flat demand curve .
the proposed algorithm does not require any communication between scheduling nodes .
the authors also study the possible errors in demand forecast and incentives for customer participations .
it should be noted that the objective of all these studies is to achieve flat demand curve for the grid .
however in this paper we study the compromise between the grid objective of flat demand vis - a - vis the user inconvenience levels , as the acceptance from the users is the key to have smart grid to be succeed .
the rest of the paper is organized as follows . in section
ii we describe the load model , our approach and problem formulation .
proposed solution , algorithms and metric for comparing various schedules are described in section iii .
simulation results are presented in section iv while the paper is concluded in section v.
in this paper we consider two types of loads in the grid i.e. essential and flexible .
essential load is due to essential activities and the devices involved in these activities have fixed scheduling needs .
flexible load is due to flexible activities and the devices involved in these activities can have flexible scheduling requirements .
there is a preferred scheduling time slot for these flexible activities and user feels most convenient if these activities are performed according to their preferences .
however we assume a generalized framework that if some activity or task is declared as flexible then it can be scheduled in time slots either before or after the preferred time slot for this activity .
for example , pre - cooling a room is an activity that can be scheduled before the preferred time slot , while cloth washing is an activity that can be scheduled after the preferred time slot .
we understand that there is no activity that can be scheduled both before and after the preferred time slot , but in this study , we just assume a generic load with such flexibility to facilitate the problem formulation .
the level of inconvenience is measured by the deviation of an activity from its specified time slot .
the more an activity is scheduled beyond its specified preferred time slot ( either to the left or to the right of it ) the more inconvenience a user faces . in the rest of the paper ,
the terms , devices , activities and tasks are used interchangeably .
similarly the terms , flexible and shiftable are also used interchangeably . given a set of tasks and their energy consumptions , we propose two extreme schedules to serve as bounds .
the first schedule is optimal for the grid in terms of load balancing and the second schedule is best for the user in terms of its preference for non - essential tasks : * * grid convenient ( gc ) schedule * : for the given set of essential and shiftable loads , this represents the best schedule from the perspective of the grid .
this schedule does not care about the user preferences in scheduling essential as well as non - essential tasks .
instead the objective of this schedule is to achieve maximum load balancing across various time slots .
we can obtain this schedule by equally dividing all the load in each time slot . * * user convenient ( uc ) schedule * : this schedule is the best schedule from the customer s perspective .
this is another extreme schedule which does not take into account the load balancing preferences of the grid .
instead it schedules all the non - essential tasks at the most preferred time slots specified by the users .
this schedule is most convenient for the users . any other schedule for the given set of loads will lie between these two extremes .
for a given set of essential and shiftable loads , the gc schedule is practically impossible to achieve because there is in reality , not much flexibility in shifting the essential loads . since we assume that we can only shift the non - essential loads , we study the region between these two extreme schedules through the following parameters : * we change the allowable time slot deviation of non - essential devices from their preferred time slots , serving as a proxy to changing the convenience levels of users .
it allows for us to schedule a device within a flexible number of time slots either to the left or to the right of its preferred time slot .
* we vary the number of non - essential devices willing to be flexible . all the devices which declare themselves as non - flexible
will then be treated as essential loads and will start exactly at their preferred time slots . through this study , results can influence the stakeholders involved in this system .
the grid can define incentives by measuring the deviation of a given schedule from the perfectly flat demand profile while also keeping in view the gc schedule for given load conditions .
similarly a customer can through feedback from its deviation of a given schedule from the uc schedule , readjust its preference conditions .
let @xmath0 denote the set of all essential tasks .
we assume that the electricity consumption data of these essential tasks on an hourly basis are known .
let @xmath1 denote the consumption of electricity by all the essential tasks to be performed during the @xmath2 time slot ( maybe hour or half hour etc ) .
let @xmath3 denote the set of all @xmath4 non - essential tasks .
the electricity consumption of these non - essential tasks is also assumed to be known .
for a non - essential task @xmath5 , let @xmath6 denote its total energy consumption .
let @xmath7 denote the total time required to complete non - essential task @xmath8 .
we allow for non - essential tasks to require several time slots to complete , and once the we decide to carry out this task at time @xmath9 then we can not stop it until it is completed .
let @xmath10 denotes the best operating time for task @xmath8 . since we have to finish all the non - essential tasks within @xmath11 time slots
, therefore we assume that @xmath12 ( to allow task @xmath8 to finish by time @xmath11 ) .
let @xmath13 denote the portion of non - essential load @xmath8 scheduled at time @xmath9 .
similarly , let @xmath14 contain the per time slot load of non - essential device @xmath8 .
it should be noted that if device @xmath8 is schedule in time slot @xmath15 then , @xmath16 the objective of this schedule is to achieve perfect load balancing for the grid .
this schedule re - distributes the essential as well as flexible load equally in all time slots .
let us denote the perfectly flat schedule by @xmath17 .
it can be obtained as follows : @xmath18 once again note that this schedule is not a practical schedule for the given set of essential and shiftable loads .
however this schedule represents the ideal situation for the grid , and merely serve as benchmark purposes .
the objective of this schedule is to carry out all the essential and non - essential tasks at their specified best time slots .
this schedule can be determined by treating the non - essential tasks like essential load at the specified time slots .
e.g. if for task @xmath8 the best time slot is @xmath19 and @xmath20 then , @xmath21 let us denote this schedule by @xmath22 .
we determine @xmath23 and then the total scheduled load during time slot @xmath9 is given as , @xmath24 this is a practical schedule , representing the current status quo and the most convenience for the users .
we can obtain a range of schedules between the above two extreme schedules by changing the number of devices declaring themself as flexible and also by defining the number of time slot deviations they are willing to tolerate .
if all the devices declare them as non - flexible then we will obtain schedule @xmath22 ( uc schedule ) . on the other hand if all the all non essential devices declare them as flexible and are willing to tolerate maximum possible time slot deviation then such a schedule , though not perfectly flat ( due to the presence of essential loads in each time slot ) will be the best schedule for the grid for a given set of loads .
let @xmath25 denote the set of devices which declare themselves as flexible .
similarly let @xmath26 denote the time slot deviation that device @xmath27 is willing to tolerate .
it means that we aim to schedule non - essential task @xmath8 within @xmath26 time slots of its preferred start time @xmath10 .
this deviation can either be to the left or to the right of the preferred time slot .
we assume here that in terms of inconvenience , the scheduling of a device @xmath26 time slots before its preferred time slot is equivalent to the inconvenience caused by scheduling the same device @xmath26 time slots after its preferred time slot .
since @xmath28 $ ] , therefore if e.g. @xmath29 then we can only perform task @xmath8 ahead of @xmath10 and schedule it in interval @xmath30 $ ] . similarly if @xmath31 then we can only perform task @xmath8 before @xmath10 and
schedule it in interval @xmath32 $ ] .
thus any non - essential task @xmath27 willing to tolerate @xmath33-time slot deviation can be scheduled in the interval @xmath34 $ ] where @xmath35 and @xmath36 .
all the non - essential devices @xmath37 have @xmath38 and they have to be scheduled exactly at time slot @xmath10 and completed after @xmath39 time slots .
we then treat all such devices @xmath37 as essential load , determine @xmath40 ( as explained in the description of the uc schedule ) and then update the essential load accordingly i.e. @xmath41 we can now formulate the scheduling problem as follows ( we refer this problem as @xmath42 ) , @xmath43 subject to , @xmath44 @xmath45 \label{sch1_const2}\ ] ] eq ( [ sch1_const1 ] ) indicates that the total energy consumed by all the non - essential tasks should be equal to their total required energy . eq ( [ sch1_const2 ] ) says that if non - essential task @xmath27 starts at time @xmath9 then it should be finished at time @xmath46 without interruption .
the start time of flexible devices can lie in the interval @xmath47 $ ] .
we can then discuss some special cases of the above general problem .
if all the devices are flexible then @xmath48 and if all the devices are 100% flexible then @xmath49 and @xmath50 ( this value of @xmath51 will allow non - essential task to finish within @xmath11 time slots ) .
the solution of this * 100% flexible problem * is the best possible practical schedule for the grid and achieves maximum flatness for given set of essential and non - essential tasks .
similarly , if all the devices declare themselves as flexible but allow only @xmath33-time slot deviation ( we assume the same @xmath33 for all the devices ) then we call this special problem as * @xmath33-time slot deviation problem*. if @xmath52 , devices declare them as 100% flexible then we call this special problem as * @xmath53-device flexible problem*. a * 100% flexibile device * can be scheduled at any time @xmath54 $ ] .
in this section we discuss the solution of the above scheduling problems and design practical scheduling algorithms .
the optimal solution of the above problem ( including all the special cases ) in general depends on the sequence or order in which we consider non - essential loads .
we illustrate this fact by following simple example .
+ * example : * consider @xmath55 time slots .
the essential load is given as @xmath56 .
there are two 100% shiftable loads with demands per time slot given as @xmath57 and @xmath58 .
there are two possible permutations , load 1 followed by load 2 or load 2 followed by load 1 . in the first case , the final load per time slot is , \{4,3,5 } with a peak load of 5 in third time slot . for the second case when load 2 is scheduled before load 1
we obtain two possible schedules , \{7,3,2 } or \{2,3,7 } both of which are optimal for this order and give a peak load of 7 in both schedules .
thus , in order to reduce the peak , we should schedule load 1 before scheduling load 2 . therefore the sequence in which we consider non - essential loads can not be ignored .
we now prove that the above problem @xmath42 is np hard problem .
[ np ] the defined problem @xmath42 is np - hard .
we consider the special case of the defined problem , where we restrict that @xmath59 , and @xmath60 .
we then prove that the special case is np - hard by a induction from the multi - processor scheduling problem , which is a well - know np - hard problem in the strong sense .
multi - processor scheduling problem : we are given @xmath61 identical machines in @xmath62 and @xmath63 jobs in @xmath64 .
job @xmath65 has a processing time @xmath66 .
the objective of multi - processor scheduling problem is to assign jobs to the machines so as to minimize the maximum load of the machines .
given an instance of multi - processor scheduling problem , we can construct an instance of the decision version of the above special case of the defined problem in polynomial time as follows .
let there be @xmath67 time slots that can be scheduled for tasks , there be @xmath68 shiftable tasks , and @xmath69 be the power consumed for the @xmath8-th task .
then , the load of the tasks scheduled at time @xmath9 is equal to the load of the jobs assigned at the @xmath9-th machine . in other words ,
the objective of minimizing the maximum load at each time is is to minimize the maximum working load assigned at each machine .
thus , the instance of multi - processor scheduling problem is equivalent to an instance of the special case of the defined load balancing problem .
thus , by induction , the defined load balancing problem is np - hard .
despite the fact that the problems are np hard we can still design an algorithm to find the optimal schedules .
however the complexity of the optimal algorithm is exponential which makes it infeasible when the number of flexible devices is large .
we give the optimal algorithm for problem @xmath42 below .
let @xmath70 hold the total power consumption , required completion time and lower and upper limits of scheduling interval ( calculated based on specified @xmath26-time slot deviations ) for non - essential tasks willing to be flexible .
note that @xmath71 and @xmath72 .
let @xmath73 denote all possible permutations of this set @xmath74 i.e. all possible ways of arranging the shiftable tasks in this set .
the total number of permutations is @xmath75 .
let @xmath76 denote the total load scheduled in time slot @xmath9 .
update essential load according to eq ( [ ess_update ] ) for all @xmath37 .
+ 2 . for all @xmath77 + 3 .
initialize : @xmath78 .
+ 4 . for all @xmath79 do , + 5
for all @xmath47 $ ] + 6 . for @xmath80 + 7 .
@xmath81 + 8 .
end for + 9 .
end for + 10 .
@xmath82 + 11 .
@xmath83 + 12 .
end for + 13 .
end for + 14 .
@xmath84 + 15 .
@xmath85 + + in the first line we update the essential load if there are some non - essential devices which declare them as non - flexible . for each permutation
@xmath61 task @xmath8 can be scheduled any time between @xmath34 $ ] . if task @xmath8 starts at time
@xmath9 then it will be complete at @xmath86 .
we obtain all the schedules with all possible start times in lines 5 - 9 for task @xmath8 . from all possible schedules we select the one which gives the minimum peak in line 10 and select this best schedule . in line 11
we update the total load and repeat for the next task @xmath8 . finally in line 14
we select the best order @xmath87 in which we should consider the shiftable tasks .
the final schedule is @xmath76 given in line 15 .
this is the optimal algorithm .
however , the complexity of this algorithm is exponential which may not be feasible when @xmath88 .
we discuss a special case of the above problem when all non - essential tasks have the same power consumption @xmath89
i.e. @xmath90 the required number of time slots to complete each task however are different i.e. @xmath91 . in this case
the sequence in which we pick the tasks for scheduling becomes irrelevant .
based on this observation we now develop a low complexity sub - optimal algorithm
. + + 1 .
initialize : @xmath92 and @xmath93 .
+ 2 . for all @xmath94 do , + 3
for all @xmath47 $ ] + 4 . for @xmath80 + 5
@xmath95 + 6 .
end for + 7 .
end for + 8 .
@xmath82 + 9 .
@xmath96 + 10 .
end for + 11 . for all @xmath8 + 12 . for @xmath97 + 13 .
@xmath98 + 14 .
end for + 15 .
end for in this algorithm we initially assume that all the shiftable loads have same power consumption @xmath99 per time slot where @xmath99 is taken as the maximum power consumption across all the non - shiftable devices willing to tolerate @xmath26-time slot deviation . in lines 2 - 10 , we arbitrarily pick the tasks one after another and find the best scheduling time @xmath100 for each shiftable task @xmath8 in their scheduling interval @xmath47 $ ] .
once we obtain the schedule then in lines 11 - 15 we restore the loads to their actual power consumption levels .
we can measure the difference between any two schedules @xmath101 and @xmath102 where @xmath103 denotes the load at time slot @xmath9 by measuring their mean square error i.e. @xmath104 let @xmath105 denote any arbitrary schedule . as defined before
let @xmath17 denote the gc schedule while @xmath22 denote the uc schedule . then we define , @xmath106 @xmath107 where @xmath108 measures the deviation of any arbitrary schedule @xmath109 for the given set of load conditions from the gc schedule @xmath17 ; while @xmath110 measures the deviation of any arbitrary schedule @xmath109 from the uc schedule @xmath22 .
the smaller the value of @xmath108 means that schedule is more flat ; while a small value of @xmath110 means that schedule is more close to the uc schedule .
we consider a generalised simulation setup of residential household appliances where electricity consumption is assumed to be constant over the consumption duration and represented in kwh .
we generate essential loads in each time slot as discrete uniform integer random variables , taking values between 1kwh and 5kwh .
each time slot represents one hour . in addtion
, we assume that there are 100 generalized devices which can be shifted .
the total power consumption of each shiftable device is generated as a discrete uniformly distributed random variable taking values between 1kwh and 5kwh .
the total duration of each shiftable task is generated as a discrete uniform random integer variable taking values between 1 and 5 time slots .
we also assume that each shiftable device has a preferred time slot . again
this preferred time slot is generated as a discrete uniform random integer variable . in fig .
[ fig0:fig ] we compare the optimal algorithm with the sub - optimal algorithm .
we assume that all the non - essential devices are 100% flexible i.e. they can be scheduled at any time @xmath111 $ ] .
for comparison we measure the mean square difference of both the schedules from a perfectly flat schedule ( gc schedule ) .
since the complexity of the optimal algorithm is exponential , we restrict ourself to only 7 non - essential devices .
we can see that although there is a small difference between the performance of both the schedules , the complexity reduction between the two algorithms is significant , and we will use only the sub - optimal algorithm in the following simulations . )
vs number of 100% shiftable devices ] in fig .
[ fig1:fig ] and fig .
[ fig2:fig ] we vary the number of devices which are willing to be 100% flexible .
all other devices which are not willing to be flexible are then treated as essential load and their power consumption is added to the essential load at their preferred time slots .
we plot @xmath108 , the deviation of our proposed sub - optimal schedule from the gc schedule in fig .
[ fig1:fig ] .
it is obvious that as more devices become flexible this deviation decreases .
however , we can observe that after 40 devices the value of @xmath108 does not decrease much which means that there is not much gain for the grid if more devices become flexible .
the flatness level achieved by 40 devices is comparable to the flatness level achieved by 100 devices . in fig .
[ fig2:fig ] we plot @xmath110 , the deviation of our proposed sub - optimal algorithm from the uc schedule . as more devices become flexible their scheduling is not performed at their most convenient time slots and thus users suffer more inconvenience .
the level of inconvenience keeps on increasing as more devices become flexible .
when 40 devices are 100% flexible the value of @xmath110 is 517 and the corresponding value of @xmath108 is 206 .
similarly when all the devices are 100% flexible the value of @xmath110 is 1375 and that of @xmath108 is 154 . if we define relative inconvenience level as , @xmath112 and relative flatness level as @xmath113 .
then for 40 devices @xmath114% and @xmath115% while for 100 devices we have @xmath116% and @xmath117% .
thus if a user only allows 40 devices to become 100% flexible he can reduce relative inconviniec level by @xmath118% while the non - flatness will only increase by @xmath119% .
* this results shows that there is a minimum level of customer participation in the smart grid that the grid should aim for that would maximize the gain to the operator while at the same time imposing minimal inconvenience .
based on this observation , the inconvenience to the customer will not be too significant .
although these results may not be representative of the system , but it does indicate a great research opportunity to reduce system wide costs at relatively small individual inconvenience . * )
vs number of 100% shiftable devices , width=170,height=170 ] ) vs number of 100% shiftable devices , width=170,height=158 ] in fig .
[ fig5:fig ] and fig .
[ fig6:fig ] , we obtain various curves by varying the number of flexible devices . in these simulations
we assume that the total number of available flexible devices can be up to 50 .
when the devices declare themselves as flexible we can schedule them according to their described x - time slot deviation levels .
the load of devices declaring them as non - flexible is then added to the essential load .
e.g. if 10 devices declare themselves as flexible then the load of remaining 40 devices is added to the essential load .
therefore the total load in all these curves is same .
as more devices become flexible , we can achieve more flatness as evident in fig .
[ fig5:fig ] .
however the gains in flatness diminish and are not very significant as the number of flexible devices are increased from 30 to 50 . *
again , this could represent large system savings at minimal individual costs .
* similarly the gains in flatness also does not increase much as the x - time slot deviation increases beyond 10 time slots . on the other hand in fig .
[ fig6:fig ] we can see that increasing the number of flexible devices significantly increases the inconvenience levels of users .
when 50 devices are flexible users experience much more inconvenience compared to 30 flexible devices . also increasing x - time slot deviation also increases user inconvenience . * hence , much further research is required to quantify the trade - off between benefit versus the users participation and inconvenience caused . * from these observations , in our test case , we can conclude that significant gains in flatness can be achieved by declaring a small number of devices as flexible and keeping x - time slot deviation up to 10 time slots .
how this translates to larger more representative systems need further examination . ) vs x - time slot deviation ] ) vs x - time slot deviation ]
in this paper we study the problem of load balancing in smart grids . we proposed algorithms to obtain schedules by varying the number of flexible devices and inconvenience levels
. then we identify the level of compromise between the grid objective of load balancing and user convenience levels .
we show that by allowing only a small portion of the activities to become flexible , users can contribute significantly towards load balancing due to aggregating effect .
similarly by letting the scheduling of activities deviate just a few hours from their preferred time slots can also significantly impact load balancing for the grid . more practical system and load models
will be used in the future work to quantify these results .
it is also interesting to investigate what kind of incentives that can be provided by the grid to encourage the users to have their load be flexible .
this research is partly supported by sutd - zju / res/02/2011 , international design center , eirp , sse - lums via faculty research startup grant , and fundamental research funds ( no .
2012hgbz0640 ) .
mohsenian - rad , v. w.s .
wong , j. jatskevich , r. schober , optimal and autonomous incentive - based energy consumption scheduling algorithm for smart grid , " in proc .
_ ieee pes conference on innovative smart grid technologies _ , jan
. 2010 .
s. caron and g. kesidis , incentive - based energy consumption scheduling algorithms for the smart grid , " in proc . _ 1st ieee international conference on smart grid communications ( smartgridcomm ) _ , pp .
391396 2010 .
s. huang , j. xiao , j. f. pekny , g. v. reklaitis , and a. l. liu , `` quantifying system - level benefits from distributed solar and energy storage , '' _ journal of energy engineering _ , vol .
138 , issue .
2 , jun 2012 . | the purpose of this paper is to study conflicting objectives between the grid operator and consumers in a future smart grid .
traditionally , customers in electricity grids have different demand profiles and it is generally assumed that the grid has to match and satisfy the demand profiles of all its users
. however , for system operators and electricity producers , it is usually most desirable , convenient and cost effective to keep electricity production at a constant rate . the temporal variability of electricity demand forces power generators , especially load following and peaking plants to constantly manipulate electricity production away from a steady operating point .
these deviations from the steady operating point usually impose additional costs to the system . in this work , we assume that the grid may propose certain incentives to customers who are willing to be flexible with their demand profiles which can aid in the allowance of generating plant to operate at a steady state . in this paper
we aim to compare the tradeoffs that may occur between these two stakeholders . from the customers perspectives , adhering to the proposed scheduling scheme
might lead to some inconvenience .
we thus quantify the customers inconvenience versus the deviations from an optimal set by the grid .
finally we try to investigate the trade - off between a grid load balancing objective and the customers preferences . |
lattice qcd allows a first principles determination of hadronic properties , including the matrix elements of operators between hadronic states . among the more interesting matrix elements are those of the @xmath5 , four - fermion operators which are responsible for the decay of a kaon into two pions .
these matrix elements are needed to relate the values of the cabbibo - kobayashi - maskawa matrix in the standard model to physical measurements .
direct measurements of @xmath6 on the lattice are not possible @xcite , due to the two pions in the final state , although a recent suggestion involving tuning the lattice volume @xcite looks promising .
alternatively , lowest order chiral perturbation theory can be used to relate @xmath7 to @xmath8 and @xmath2 matrix elements @xcite , which can be measured on the lattice .
this approach has been extensively studied and , while quite straight - forward in principle , requires careful control over a wide range of phenomena . of central importance to a calculation relying on chiral perturbation theory
is precise control over chiral symmetry on the lattice .
a major advance in this long - standing problem is the idea of four - dimensional chiral fermions arising from mass defects in a higher - dimensional fermion formulation @xcite . with improved chiral symmetry at finite lattice spacing , the lattice ward - takahashi identities , without complicated renormalization ,
are much closer to the continuum ones . here
we seek to understand how the residual chiral symmetry breaking effects for domain wall fermions enter the removal of unphysical , power divergent contributions , which is needed to get @xmath9 from @xmath1 and @xmath10 matrix elements .
the riken - bnl research center / bnl/ columbia ( rbc ) collaboration has recently finished extensive quenched simulations of the hadron spectrum and low - energy qcd physics , using domain wall fermions @xcite , with a particular focus on the residual chiral symmetry breaking effects due to finite values for the fifth dimension , @xmath11 @xcite .
( similar work was also reported in @xcite . )
this work produced quantitative values for the residual quark mass @xmath12 , which enters low energy qcd physics through the effective quark mass , @xmath13 , where @xmath14 is the input bare quark mass .
the residual quark mass was shown to be small , @xmath15 the strange quark mass for @xmath16 and @xmath17 .
it was also found that topological near - zero modes of the domain wall fermion operator , which are not suppressed by a fermionic determinant in quenched simulations , have pronounced effects on quark propagators for small @xmath14 .
in addition , for large volumes where topological near - zero mode effects are suppressed , a non - linear dependence of @xmath18 on @xmath14 was found , consistent with a quenched chiral logarithm .
thus , even though the chiral properties of domain wall fermions are quite improved over other formulations , these quenched pathologies can appear in simulation results .
we have also finished measuring @xmath1 and @xmath19 matrix elements for the 10 @xmath3 effective four - fermion operators @xcite relevant to determinations of @xmath20 and the real parts of the @xmath21 and 3/2 amplitudes for @xmath22 decays .
we have results for 200 quenched lattices of size @xmath23 with @xmath17 and @xmath24 and a variety of light and heavy quark masses .
( for more details about the simulations see @xcite . )
the analysis of these results requires control over the residual chiral symmetry breaking effects and an understanding of the role of quenched pathologies in our data .
this report is a summary of our progress to date .
a final step in determining physical matrix elements is the matching of continuum and lattice operators .
this has been done using non - perturbative renormalization @xcite and details can be found in @xcite .
furman and shamir @xcite defined axial transformations for domain wall fermions , which lead to the ward - takahashi identity @xmath25 here @xmath26 at the four - dimensional point @xmath27 includes spinors from all values of the the fifth coordinate . the pseudoscalar density @xmath28 is constructed from four - dimensional fields in the standard way and @xmath29 is constructed from `` mid - point '' spinors at @xmath30 and @xmath31 .
( the notation is as in @xcite . ) for physics that involves momentum scales much less than the cutoff , the extra term in eq .
[ eq : ward_tak_id ] involving @xmath32 ( the `` mid - point '' term ) , gives a contribution of the form @xmath33 .
this defines @xmath12 to @xmath34 , since the low - energy qcd physics of domain wall fermions should by describable by an effective lagrangian .
however , in matrix elements involving divergent quantities one only expects that @xmath35 . since divergent expressions enter in @xmath36 amplitudes and we must manipulate the ward - takahashi identities to remove them , these effects need to be understood . to test our understanding of @xmath12 effects and our data ,
we first consider @xmath37 in eq .
[ eq : ward_tak_id ] . summing over
@xmath27 gives ( no sum on @xmath38 ) @xmath39 j^a_5(y ) \rangle - \langle \overline{q } q \rangle = 0 \label{eq : int_wi}\ ] ] the term in the sum without @xmath40 gives @xmath41 pion pole terms plus @xmath42 contact terms along with constant terms and terms of higher order in @xmath14 .
similarly the mid - point term gives @xmath43 pion pole terms , plus @xmath44 contact terms as well as constant contributions .
the pion pole terms enter as @xmath45 in eq .
[ eq : int_wi ] while the contact contributions enter as @xmath46 .
this form for the pion pole term is consistent with @xmath47 at @xmath48 .
thus we are lead to expect a gell - mann oakes renner relation for domain wall fermions with the form @xmath49 this relation holds at non - zero values for @xmath14 , with the @xmath50 contact term contribution being cancelled by the similar term in @xmath51 .
of course , there may be other quenched pathologies entering this expression , such as effects from topological near - zero modes or quenched chiral logarithms .
we have measured @xmath52 , @xmath53 , @xmath12 and @xmath54 for @xmath55 lattices at @xmath16 @xcite and plot our results in figure [ fig : gmor ] .
one sees that for @xmath17 , the larger mass points are approximately linear , but do not seem to extrapolate towards the value of @xmath54 at @xmath56 . for @xmath57 ,
the extrapolation is much closer , as is expected since @xmath12 has decreased by almost a factor of 2 .
also shown are solid lines representing the left - hand side of eq .
[ eq : gmor ] when a quenched chiral logarithm fit is used for @xmath58 .
one sees that the effect of chiral logarithms is to exacerbate the disagreement in the extrapolation .
thus we have evidence for residual mass contact terms in this simple ward - takahashi identity . in conclusion , we note that for @xmath17 , the quantity @xmath59 , which is 1 in chiral perturbation theory , differs from 1 by around 20% for our data .
thus one should avoid introducing extraneous powers of this quantity while manipulating the data .
we now turn to the ward - takahashi identities useful for subtracting the divergent contributions to @xmath1 matrix elements .
here we discuss our tests of a particular method for removing the divergent effects in @xmath61 matrix elements .
following @xcite , we define a subtracted operator @xmath62 \label{eq : osub}\ ] ] where @xmath63 is a @xmath3 four - fermion operator and @xmath64 is chosen such that @xmath65 then following @xcite , but noting that contact terms in the ward - takahashi identities coming from @xmath12 effects will enter , we find @xmath66 where the derivative removes any contact terms proportional to @xmath12 . the successful removal of the effects of mixing with the dimension three operators implemented in eq .
[ eq : osub ] requires chiral symmetry .
we first check this with a simpler ward - takahashi identity , obtained by inserting an operator @xmath67 in eq .
[ eq : ward_tak_id ] .
the pion pole terms in the ward - takahashi identity then lead to @xmath68 there are no power divergent operators in this ward - takahashi identity , so the coefficient of the @xmath14 and @xmath12 contact terms are finite , but there may still be quenched chiral logarithm effects present . in figure
[ fig : k2pi_sbard ] we plot the left - hand side of eq .
[ eq : k2pi_sbard ] versus @xmath14 .
we see that for larger @xmath14 , the points appear to extrapolate reasonably well to 1 , but there is noticeable curvature for small quark masses .
the figure includes results for axial vector sources , which agree with those for pseudoscalar sources , indicating that topological near - zero modes are not the origin of this non - linearity .
we are investigating whether this can be explained as a quenched chiral logarithm .
= we determine @xmath64 from measurements of @xmath69 for non - degenerate masses for the strange and down quarks .
the upper panel of figure [ fig : o2 ] shows a plot of @xmath70 versus @xmath71 .
we see good linearity which allows for a reliable determination of @xmath72 , with statistically errors at the few percent level for this particular operator .
= with the value of @xmath72 determined , @xmath73 is known and figure [ fig : o2 ] shows @xmath74 where @xmath75 is the @xmath76 part of @xmath4 . in lowest order chiral perturbation theory ,
the denominator is a constant , so the desired @xmath0 matrix element is given by the slope and the intercept is related to @xmath12 .
we are looking carefully for non - linearities in our data , paying particular attention to the quenched chiral logarithm effects recently calculated in @xcite .
we have seen that the ward - takahashi identities for domain wall fermions include @xmath44 contact terms which can be removed by taking derivatives with respect to @xmath14 .
our data also show evidence for non - linear dependence on @xmath14 , some of which may be due to quenched chiral logarithms .
we have shown that the subtracted operator @xmath73 can be determined and are now doing the subtractions for all the operators .
9 l. maiani and m. testa , phys .
* b245 * ( 1990 ) 585 .
m. luscher and l. lellouch , hep - lat/0003023 . c. bernard , _ et .
rev . * d32 * ( 1985 ) 2343 .
d. kaplan , phys .
b288 * ( 1992 ) 342 . y. shamir , nucl .
* b406 * ( 1993 ) 90 , v. furman and y. shamir , nucl .
* b439 * ( 1995 ) 54 .
t. blum , _ et .
al . _ , hep - lat/0007038 , l. wu , m. wingate , these proceedings .
a. ali kahn , _ et .
al . _ , hep - lat/0007014 .
m. ciuchini , _ et .
z. phys .
* c68 * ( 1995 ) 239 .
t. blum , these proceedings .
g. martinelli , _ et .
al . _ , nucl .
* b445 * ( 1995 ) 81 . c. dawson , these proceedings .
s. sharpe , _ et .
* b286 * ( 1987 ) 253 .
m. golterman and e. pallante , hep - lat/0006029 . | we are using domain wall fermions to study @xmath0 matrix elements by measuring @xmath1 and @xmath2 matrix elements on the lattice and employing chiral perturbation theory to relate these to the desired physical result .
the residual chiral symmetry breaking of domain wall fermions with a finite extent in the fifth dimension impacts these measurements . using the ward - takahashi identities , we investigate residual chiral symmetry breaking effects for divergent quantities and study pathologies of the quenched approximation for small quark mass .
we then discuss the @xmath3 operator @xmath4 , where chiral symmetry is vital for the subtraction of unphysical effects . |
one of the extensions of electrostatics beyond coulomb interactions is the interaction of charged conductors with multiple modes of polarization .
the relatively simple problem of two charged metal spheres has been studied from a theoretical and pedagogical standpoint using image charges @xcite @xcite .
in addition , more complex methods using image charges @xcite and boundary element methods @xcite have been developed for multiple - particle configurations .
here we derive a new method to calculate the electrostatic interaction energy of charged conducting spheres that utilizes multipole expansions of the potential and charge density .
this method can be expanded to any number of conductors or to an arbitrary precision by including higher order multipoles in the calculation .
this approach is motivated by the assembly configurations of binary nanoparticle super lattices ( bnsls ) .
guided by a complex combination of entropic and electromagnetic interactions , these lattices can assemble in a vast array of different shapes and configurations . in order to better understand the assembly of these lattices ,
a number of experimental studies have tried to isolate and analyze the different interactions driving the assembly@xcite . here
we focus on analyzing the electrostatic interactions that guide the assembly of the metal nanoparticles at the center of these lattices . normally in electrostatic problems ,
most of the higher order interactions can be ignored since the monopole and dipole terms dominate the energy equation for large separations .
once the scale of the configuration approaches the scale of the nanoparticles , the quadrupole or octupole interactions can be on the same order as some monopole interactions @xcite . since these assemblies can also include nanoparticles in contact , there is no assumption that can justify ignoring multipole terms beyond the dipole .
this multipole expansion method allows the interaction between any number of conducting spheres to be calculated using an arbitrary number of multipoles
. this will provide insight into why bnsls only assemble with certain configurations of metal nanoparticles by allowing a direct energy comparison across different configurations .
moreover , it will allow us to analyze how the geometry of a configuration influences the relative strength of various multipoles in a configuration . + we proceed in deriving the equations for the energy of the configuration by first looking at the energy between two charged spheres . to make this calculation easier ,
we begin by assuming that the two spheres lie on the z - axis of our arbitrarily selected coordinate system .
once we find the energy equations in this simplified aligned case , we derive the transformation to express the energy equations in any arbitrary frame .
the accuracy of this method is then checked against existing literature calculations , and then we generalize the energy equations to include an arbitrary number of spheres . finally , we present calculations for a number configurations and comment on the accuracy and insights of the multipole expansion method .
to calculate the multipole expansion of the energy , we start with the energy equation : @xmath0 this can be broken up into the pairwise sphere interactions for n spheres : @xmath1 where @xmath2 is the charge density on sphere i , @xmath3 is the potential on sphere j , and the integration is carried over the volume of sphere i. the terms where @xmath4 represent the energy cost of holding a charge distribution on the sphere , further referred to as the self - energy .
the other terms represent the interaction energy , or the energy associated with charge distributions and potentials on different spheres .
we now find the multipole expansion of the charge distribution and potential for a general sphere @xmath5 .
the charge distribution on sphere @xmath5 ( with radius @xmath6 ) is given by : @xmath7 where @xmath8 , etc .
represent the complex conjugates of @xmath9 etc .
the @xmath10 s are related to the multipole moments , @xmath11 s , as defined by jackson ( gaussian or cgs units ) @xcite where : @xmath12 to calculate the potential , we expand the formula for the potential in terms of the sphere s multipoles : @xmath13 to calculate the energy explicitly , we start by just considering two spheres , a and b , with their charge distributions and potentials given by replacing the label s with labels a and b in equations [ eq : dist ] and [ eq : potent ] . thus the total interaction energy is given by : @xmath14 we will continue by deriving the formula for @xmath15 .
the other half of the interaction energy , @xmath16 , can be derived by using @xmath17 and @xmath18 in place of @xmath19 and @xmath20 , all odd order multipoles need to be multiplied by a @xmath21 since the alignment of a positive multipole this adds a @xmath22 factor to the odd ordered b multipoles in the derivation of @xmath16 . ] . for the calculation of two aligned spheres , the radius term in the potential ( @xmath23 in eq .
[ eq : potent ] ) is the distance between a point on sphere a and the center of sphere b. this gives the @xmath15 interaction energy as : @xmath24 we can find the form for @xmath25 using the law of cosines : @xmath26 plugging this in and integrating over the radius variable , gives the energy as : @xmath27 the expression contained inside the radical is the generating function for the gegenbauer polynomials@xcite @xmath28 with @xmath29 , @xmath30 , and @xmath31 .
this allows us to express the denominator as a power series in @xmath32 with the coefficients being the gegenbauer polynomials@xmath33 . through a transformation matrix of coefficients ,
the denominator is changed from an expansion of gegenbauer polynomials into an expansion of spherical harmonics .
this is done through the following transformation : @xmath34 where we defined @xmath35 as : @xmath36 and utilized the following identity : @xmath37 where @xmath38 is the p - th legendre polynomial @xcite .
+ the formula for @xmath39 is derived from the law of sines : @xmath40 ( see figure [ fig : diag1 ] ) . as before , we can expand @xmath25 using the law of cosines and the gegenbauer series expansion : @xmath41\right)}\end{aligned}\ ] ] where @xmath42 is the t - th gegenbauer polynomial .
because of the azimuthal symmetry of the set - up we see that @xmath43 . plugging this all into the energy equation ( leaving @xmath44 in its place for simplicity ) : @xmath45\mathrm{d}\omega\ ] ] @xmath46 where @xmath47 is defined as : @xmath48 and @xmath31
. we can also simplify the integral to be performed by utilizing the wigner 3-j symbols to decompose the product of spherical harmonics into a sum .
this formula is given by @xcite : @xmath49 where @xmath50 is the appropriate wigner 3-j symbol with the convention used in @xcite and @xmath51 .
this gives us a final expression for @xmath15 : @xmath52 this is the interaction energy that we will use to calculate the energy for a configuration of spheres .
to give an example of what the individual energy terms are , we integrate this formula explicitly for the first few multipoles . for simplicity in reporting the expansion below
, we assume that the spheres are of equal size ( @xmath53 ) .
in addition , we will assume the two spheres are separated by a distance much larger than their individual radii ( @xmath54 ) .
this allows us to take the following simplification : multipoles that occur in the general case interaction energy . ]
@xmath55\right ) } \rightarrow \pi\ ] ] using this simplification we expand @xmath15 and report the terms of order @xmath56 or lower : @xmath57 these are the terms of the multipole expansion of the partial interaction energy , @xmath15 , up to the dipole - dipole interaction in the special case where the spheres are far apart and aligned along the @xmath58 axis .
to generalize for spheres not aligned along the @xmath59 axis , the expression for @xmath25 becomes : @xmath60\end{aligned}\ ] ] where @xmath61 and @xmath62 are the polar and azimuthal angles between @xmath63 and the vector @xmath25 .
when the spheres were aligned along the @xmath64 axis , we were able to simplify the equations since @xmath65 and @xmath66 .
+ we start with same gegenbauer expansion method presented above , but we also need to rotate the spherical harmonics in @xmath67 to spherical harmonics in @xmath68 by using the euler angles @xmath69 that separate the two frames . in our specific case , we are only rotating the center point of a sphere to the z - axis without changing the orientation .
this means that one of the euler angles , @xmath70 , will not have any effect on the orientation of the spheres .
thus we can take @xmath71 and just use two euler angles to describe the rotation .
the following transformation converts the spherical harmonics with @xmath67 arguments to spherical harmonics with @xmath68 arguments : @xmath72 where @xmath73 is the standard finite rotations operator that can be found in landau and lifshitz @xcite . + this transformation allows us to easily calculate the interaction energy equation for a configuration of multiple spheres . in a configuration with more than two spheres ,
each pairwise interaction energy has the same form as equation [ eq : f_aligned ] when the coordinate systems are aligned along each line of action ( i.e. along @xmath63 as @xmath61 is in figure 2b . ) . instead of recalculating the entire energy series for every orientation
, we can utilize the fact that the rotation of the spherical harmonics is a linear transformation ( see appendix [ app : linear ] ) .
the transformation allows us to take each multipole referred in these local pairwise frames and transform them all to the same configuration - wide @xmath68 coordinate system .
this expands the @xmath74 and @xmath75 multipoles in @xmath76 in terms of new multipoles @xmath77 and @xmath78 in the @xmath68 frame .
this gives us the transformation : @xmath79 the rotation coefficients can be further decomposed into their three constituent parts : @xmath80 @xmath81 applying this transformation to the multipoles changes the expression for @xmath82 to : @xmath83 this new form for @xmath82 is then plugged into equation [ eq : f_aligned ] .
this gives us a completely general expression for the interaction energy between two arbitrarily oriented spheres .
to see what the first few energy terms are , we integrate equation [ eq : f_aligned ] explicitly and obtain the formula for the interaction energy between the two spheres . due to limited space
we show only the monopole - monopole and dipole - monopole terms in the @xmath84 limit below : @xmath85 the self - energy formulas are much easier to derive .
starting with the expression for the energy in equation [ eq : ijenergy ] , we set @xmath4 and use the charge density and potential definitions from equations [ eq : dist ] and [ eq : potent ] .
this gives the self - energy for sphere a is : @xmath86 simplifying and utilizing the orthogonality of the spherical harmonics to perform the integration yields : @xmath87 applying the same process to sphere b yields the same exact result in terms of the @xmath88 s instead of the @xmath89 s . by combining the interaction energies and the self energy from both spheres we obtain a energy formula that can be expanded to any order interaction .
due to the many summations in the interaction energy formula , the computation of relatively small order interactions ( like the monopole - dipole interaction ) can be extremely complicated and asymmetric .
for example , the computation time needed to calculate @xmath90 ( a monopole on sphere a and a dipole on sphere b ) is the interaction energy between a dipole on sphere a and a monopole on sphere b calculated with the @xmath15 formula .
@xmath91 is the interaction energy between a monopole on sphere b and a quadrupole on sphere a calculated from the @xmath16 formula ] is much longer than the computation time needed to calculate @xmath92 even though both represent the same electrostatic interaction .
looking at eq .
[ eq : integral ] , it is clear that the @xmath93 exponent makes @xmath15 more difficult to calculate as the multipole order increases .
this calculation can be simplified by making use of green s reciprocation theorem to reduce the order of the exponent in eq .
[ eq : integral ] @xcite .
when constructing the interaction energy @xmath15 , we can split the calculation up into each of the different components : @xmath94 from green s reciprocation theorem , we know that : @xmath95 in order to calculate @xmath90 , we would need to expand @xmath96 in another series .
thus , it is advantageous to use @xmath92 instead since it avoids having to use any further expansions or unnecessary computation time
. this renders the interaction energy as : @xmath97 this same idea is then applied to further terms in the expansion ( e.g. replacing @xmath98 with @xmath99 ) .
+ in addition , green s reciprocal theorem also states that @xmath100 . with proper attention to the conjugation of multipoles
, there is no need to calculate @xmath16 . since substitutions
can rotate our aligned energy into any rotated frame ( see section [ sec : nonaligned ] ) , no calculations beyond @xmath15 in the aligned case are needed to obtain the full interaction energy ( @xmath101 ) , for any orientation . here
we will check specific terms of the energy equation against other known equations .
we will use the simpler aligned version of the energy equation for these calculations .
the first and easiest check is that the monopole - monopole interaction gives the correct coulomb energy . for two charged spheres
, we use the definitions of the multipole moments in jackson @xcite , @xmath102 to find : @xmath103 which agrees with the coulomb energy of two point charges .
the dipole - monopole and the dipole - dipole energies derived from our equations also agree with the jackson formulas .
to check the dipole - monopole energy , we use the simplification of section [ sec : green ] and plug in the values of the multipoles to find : @xmath104 where @xmath105 is the dipole on sphere a , and @xmath106 is the electric field due to a monopole on sphere b. this agrees with the formula for the energy of a perfect dipole in an electric field . for the dipole - dipole interaction ,
we start with the formula for the interaction between two dipoles given by jackson @xcite : @xmath107 since this formula is for the energy between two perfect dipoles , we recalculate the dipole - dipole term of our interaction energy assuming @xmath108 in order to compare it to the jackson formula . using the green s theorem simplification from section [ sec : green ] again and taking the dipole - dipole term from the @xmath15 expansion yields : @xmath109 now looking at the jackson formula , we note that the distance between the two dipoles will be @xmath63 . since the dipoles could either be aligned in the same direction or in opposing directions , there is a plus or minus factor in front of the formula : @xmath110 which agrees with our lowest order term from the dipole - dipole interaction energy expansion .
+ we can also compare a numerical result from our calculations with an analytical solution done by maxwell@xcite .
maxwell finds that two touching spheres of radius @xmath111 held at potential @xmath112 statvolt each have a charge of @xmath113 .
taking @xmath114 cm and using the equations for the energy stored in a capacitor we find : @xmath115 @xmath116 where @xmath117 is the energy stored in the capacitor .
this is extremely close to our energy values for the @xmath118 limiting case .
in addition , our formula converges rather quickly to the correct value .
.[tab : table1 ] comparison of calculated energies with analytic maxwell energies [ cols="^,^,^ " , ] to calculate the energies reported above , we explicitly performed the integrations from our final energy equation using mathematica .
we then minimized the energy equation utilizing the matrix formalism detailed in the next section .
the gegenbauer expansion of the denominator , @xmath25 , was taken out to the sixth gegenbauer polynomial .
this was the point at which we started getting marginal returns in the percent difference between our calculated energy and maxwell s figures .
the second gegenbauer expansion used in the expansion of @xmath39 was taken out to the first term .
this provided improvement over the small angle approximation and left the integrals still analytically computable by mathematica .
these formulas are generalized for any number of spheres by writing the multipoles as a vector and the individual terms of the energy as the entries of a matrix .
this gives us the following matrix equation for the energy between any number of spheres : @xmath119 where @xmath120 is a vector containing all the multipoles for all the spheres to a particular order and @xmath121 is the energy matrix that contains all the coefficients for the different multipole interactions .
the diagonal terms of @xmath122 are the self - energy terms , while generally , @xmath123 is the interaction between the @xmath124 and the @xmath125 multipoles .
the left section of @xmath122 for a three sphere calculation up to dipoles is produced in the appendix .
the minimum of equation [ eq : matrix ] occurs in the trivial case when all the multipoles are equal to zero . to examine the non - trivial cases we need to find the constrained minimum by minimizing the function : @xmath126 where @xmath127 are the undetermined multipliers for the constraints .
the term @xmath128 is a dummy variable that uses the dot product @xmath129 to select out a specific multipole to constrain .
the variable @xmath130 is the constrained value for that specific multipole .
for example , to constrain the monopole term on sphere a to be one , the second term of the equation would take the following form : @xmath131= \lambda ( a_{0}^{0 } -1)\ ] ] while we can arbitrarily constrain any multipole , our calculations only constrain the monopole terms since they correspond to the total charge on each sphere . for a configuration with three spheres ,
there are three different constraint equations ; one for the monopole term on each sphere . in order to minimize @xmath132
, we take the derivative of equation [ eq : langmatrix ] and set it equal to zero .
the derivative is easy to calculate since @xmath122 is a hermitian matrix , and this yields the multipoles , @xmath120 , that minimize the constrained equation : @xmath133 by using this matrix formalism and solving for the unconstrained multipliers we can find the multipoles that minimize the electrostatic energy for any given number of charged conducting spheres .
here we present the configuration energy and the work required to remove a bounded sphere for various configurations .
the configuration energy includes both the interaction and self energies of all the spheres except the monopole self energy .
since we constrain the monopole terms to constant values , we omit them from the energy data . the work required to remove a sphere is calculated by finding the difference in configuration energy between the original configuration and the configuration with one sphere removed . in the case of two spheres ,
the work needed to remove a sphere is the same as the configuration energy . in the case of three spheres ,
we subtract the total configuration energy from the energy of the two - sphere system that results after we remove a sphere . for asymmetrical configurations ,
we compare the work required to remove each of the different spheres and report the minimum .
the spheres in the figures are shaded with their charge density .
the coloring goes from blue to red as the charge density goes from negative to positive ( color online , dark to light in grayscale ) .
in addition there are lines drawn around the z - axis of each sphere to help qualitatively show the orientation of the charge on the spheres .
the figures are meant to show alignment of the various higher - order multipoles relative to the other spheres in its own figure .
the shading scale was varied between different figures to provide enough contrast within a figure .
in constructing these configurations , we label spheres with letters ( a , b , c ... ) usually placing sphere a at the origin as a convention .
since the equations are rotationally invariant usually only one specific configuration is present ( i.e. no more than one equilateral triangle with one charged sphere will be looked at since all three possibilities are equivalent ) . for the calculations
, we used the same parameters discussed in section [ sec : check ] when we compared our energy values with maxwell s calculations .
charged spheres had their monopole constrained so that the sphere has one statcoulomb of charge .
all spheres have radius of 1 cm , and all energies are reported in ergs ( gaussian or cgs units ) . in the case of two spheres ,
there are only two different configurations of interest .
the first is when both spheres have equal charges .
this induces multipoles equal in magnitude for both spheres and equal binding energies for each sphere . when only one sphere is charged , the higher order multipole terms have a much stronger role . looking back at the explicit interaction energy expansion , the first three terms of the energy expansion are the monopole - monopole energy and the two monopole - dipole interactions . however , only one of those three terms is non - zero when one sphere is left uncharged .
this results in the binding energy for the more weakly bound sphere ( sphere b ) to increase 43% when the expansion is extended from the dipole - dipole interaction up to the quadrupole - quadrupole interaction .
if we take the calculation up to the octupole - octupole interaction , the binding energy of that sphere increases by another 15% .
this is compared to only about a 0.2% difference between the quadrupole - quadrupole and octupole - octupole calculations in the configuration with both spheres charged .
if we had arbitrarily cut off the calculation at the dipole - dipole term of the configuration with only one sphere charged , we would have seen a binding energy for sphere b that is about one - third of the strength it actually is .
this shows evidence for why higher ordered terms can not be arbitrarily ignored in these energy calculations .
also , since the configuration energy relies heavily on higher order multipoles , the energy values do not converge quickly when only one sphere is charged .
for the three sphere case , the minimal energy configuration is the equilateral triangle ( configuration a ) .
this is followed by the isosceles triangle ( config .
b ) and the middle - charged line of spheres ( config .
c ) . as more
multipoles are added , the minimal energy configuration changes . when only dipoles are taken into account ,
both the isosceles triangle and the middle - charged line of spheres are favored over the equilateral triangle .
this is due to minimal dipole - dipole interaction between the two uncharged spheres in each configuration .
the isosceles triangle configuration induces mutually orthogonal dipoles on the two uncharged spheres , reducing the energetic cost of the configuration . in the middle - charged line of spheres ,
the monopole on the middle sphere dominates the energy equation through the monopole - dipole terms .
the antiparallel dipoles that are induced on the two uncharged spheres have only a slight effect on the configuration energy . however , this effect is strong enough to give the configuration a slightly higher energy than the isosceles triangle .
once we add quadrupoles , the dipole advantages discussed above are much less influential , causing the equilateral triangle becomes the minimal energy configuration .
since all the dipoles in the equilateral triangle line up to the same point in space , there is much more repulsive interaction between the dipoles than in the isosceles triangle configuration .
this caused it to have a misleadingly higher configuration energy when only dipoles are considered . with
the quadrupoles added , the equilateral triangle experiences the largest increase in binding energy , making it the lowest energy configuration .
the symmetry of this configuration causes the energy to be more dependent on quadrupole terms than the isosceles triangle or the middle - charged line of spheres .
this shows that some configurations show non - regular convergence when different multipoles are added to the calculation .
in particular , the work to remove a sphere from configuration d is not even monotonically decreasing as we add higher order terms , demonstrating the significant effect that the geometry of the configuration has on the binding energies .
in addition , configurations with uncharged spheres equidistant from the charged sphere have stronger binding energies . the line with a charge on the end ( config .
d ) and the isosceles triangle with a charge on one of the legs ( config e ) are the most weakly bound configurations .
both of these configurations are asymmetrical in the sense that the two uncharged spheres are not equidistant from the charged sphere .
judging from the charge distribution diagrams of these configurations , it seems that the asymmetric configurations restrict the polarization of one of the uncharged spheres .
this prevents the configuration energy from lowering as far as the other three configurations that all feature the uncharged spheres equidistant from the charged sphere .
we also see that the energy convergence of the equilateral triangle s configuration energy is similar to the the two sphere case .
the binding energy increases by 86% when the calculation order is taken from dipoles up to quadrupoles and a further 16% when octupole interactions are added in , comparable to the convergence seen in the two sphere case .
however , there is variability in energy convergence for different configurations .
for example , the binding energy for configurations c and d only increase by 1.5% and 8% respectively when the calculation is extended from quadrupoles to octupoles . for configurations with four spheres ,
the minimal energy configuration is the regular tetrahedron ( config .
a ) . much like in the three - sphere case , the minimal energy configuration changes as additional multipole terms are added to the calculation . when only dipoles are considered , the minimal energy configuration is the non - regular tetrahedron ( config .
since most of the faces of this tetrahedron are isosceles triangles , the behavior of this configuration is similar to the isosceles triangle that we analyzed in the three - sphere case .
as before , the configurations that induce dipoles in mutually orthogonal direction will appear to have an energetic advantage when the calculation only includes dipole terms .
the regular tetrahedron ( config .
a ) mimics the behavior of the equilateral triangle , with all of the spheres induced dipoles focusing on a common point .
this causes strong interactions between the induced dipoles and makes configuration a look less favorable when only including dipole interactions in the calculation .
thus , when only dipoles are included in the calculations , there is an intrinsic bias for configurations that feature mutually orthogonal dipoles .
this bias , caused by arbitrarily cutting off the calculation at the dipole - dipole interaction , results in incorrect conclusions for the minimal energy configuration .
once quadrupoles are added , the configurations take the same order as when the calculation is carried out to octupoles . the regular tetrahedron ( config .
a ) is the preferred configuration with the charged blunt - end rhombus as the next most favored .
note the significant energy difference between the two rhombus configurations ( configs .
b and d ) .
when the blunt - end particle is charged in a rhombus , the charged sphere is in contact with all of the other three spheres in the configuration causing significant polarizations in the three uncharged sphere .
however , in the sharp - end charged rhombus ( config .
d ) , the charged sphere is only in contact with two other spheres .
this causes a much smaller polarizing effect on the third sphere and makes the configuration less energetically favorable when compared with the blunt - end charged rhombus ( config b ) .
here we presented a new method utilizing multipole expansions to calculate the energy of charged conducting sphere configurations .
this method can be used for any number of spheres in an arbitrary configuration , and it is easily expanded to any multipole order . as seen from the configurations we considered , cutting off the energy expansion at arbitrary multipole orders ( like at the dipole - dipole or quadrupole - quadrupole level ) gives an incorrect picture of the energetically preferred configurations .
in addition , there are intrinsic biases towards particular configurations .
configurations that induce mutually orthogonal dipoles , were favored when we cut off the energy expansion at the dipole - dipole interaction . in this study , we truncated the energy expansion at the octupole - octupole level . in the case of two charged spheres ,
this was sufficient since the energy converged quickly ; the energy only increased 0.2% when the calculation was taken from quadrupoles up to octupoles .
however , this is not the case when we move on to more general configurations with multiple spheres . in general
, the more spheres in a configuration , the more the energy calculations rely on higher order multipoles , and the less quickly the energy converges .
the calculations here were also subject to an arbitrary cut - off for interactions at the octupole - octupole level . given more computing power or patience ,
the calculations could be extended even further .
we can extend this study by analyzing further how different multipoles interact in various configurations .
for example , a family of configurations could be found where the configuration energy only slightly changes with the addition of quadrupoles .
this implies that the specific geometry of the configuration reduces or neutralizes the quadrupole - quadrupole interactions . by identifying similar configurations
, more complex tailor - made materials can be created that utilize configuration geometry to reduce individual inter - particle interactions .
this paper was the product of an independent research project done by alex moore under the supervision of thomas witten while at the university of chicago .
in addition nathan krapf , also at the university of chicago , helped to extensively revise this paper .
alex would like to thank both prof .
witten and nathan krapf for their patience and guidance while working on the project , as well as dmitri talapin for his support and for providing the initial motivation for this study .
the charge distribution on a sphere and the energy between two spheres are invariant with respect to any choice of coordinates . therefore , given two systems of coordinates , @xmath134 and @xmath135 , a multipole in the @xmath136 frame can be expanded in terms of the multipoles in the @xmath137 frame for any point @xmath138 : @xmath139 now we can rotate the @xmath140 into the @xmath136 frame by utilizing the identity @xcite : @xmath141\ ] ] where @xmath142 are the euler angles needed to rotate the @xmath137 coordinate system to the @xmath136 coordinate system . plugging this in : @xmath143 @xmath144 by the linear independence of the @xmath145 s , we must have @xmath146 .
this removes the summation over n and presents us with our final indentity : @xmath147 if we are trying to rotate a conjugated multipole , we just use the rules of complex conjugation of sums and products to get the identity : @xmath148
this section of the matrix @xmath122 ( defined in equation [ eq : matrix ] ) covers the projection of @xmath149 onto the monopole and dipole terms from spheres a , b , and the monopole from sphere c ( given by vector @xmath150 ) . for simplicity , we assume the spheres are all the same size ( @xmath151 ) , and we take the limit @xmath152 .
this gives the first few rows and columns of @xmath122 as : @xmath153}{r_{\text{ab}}^2 } & \frac{4 \pi \text{cos}\left[\beta _ { \text{ab}}\right]}{\sqrt{3 } r_{\text{ab}}^2 } & -\frac{2 \sqrt{\frac{2}{3 } } e^{i \alpha _ { \text{\tiny{ab } } } } \thinspace \pi \text{sin}\left[\beta _ { \text{ab}}\right]}{r_{\text{ab}}^2 } \hdots \\ -\frac{2 \sqrt{\frac{2}{3 } } e^{i \alpha _ { \text{ab } } } \thinspace \pi \text{sin}\left[\beta _ { \text{ab}}\right]}{r_{\text{ab}}^2 } & -\frac{4 \pi \text{sin}\left[\beta _ { \text{ab}}\right]{}^2}{3 r_{\text{ab}}^3 } & -\frac{2 \sqrt{2 } e^{i \alpha _ { \text{ab } } } \thinspace \pi \text{sin}\left[2 \beta _ { \text{ab}}\right]}{3 r_{\text{ab}}^3 } & \frac{4 e^{2 i \alpha _ { \text{ab } } } \thinspace \pi \text{sin}\left[\beta _ { \text{ab}}\right]{}^2}{3 r_{\text{ab}}^3 } \hdots \\
-\frac{4 \pi \text{cos}\left[\beta _ { \text{ab}}\right]}{\sqrt{3 } r_{\text{ab}}^2 } & -\frac{2 \sqrt{2 } e^{-i \alpha _ { \text{ab } } } \thinspace \pi \text{sin}\left[2 \beta _ { \text{ab}}\right]}{3 r_{\text{ab}}^3 } & -\frac{8 \pi \text{cos}\left[\beta _ { \text{ab}}\right]{}^2}{3 r_{\text{ab}}^3 } & \frac{2 \sqrt{2 } e^{i \alpha _ { \text{ab } } } \thinspace \pi \text{sin}\left[2 \beta _ { \text{ab}}\right]}{3 r_{\text{ab}}^3 } \hdots \\ \frac{2 \sqrt{\frac{2}{3 } } e^{-i \alpha _ { \text{ab } } } \thinspace \pi \text{sin}\left[\beta _ { \text{ab}}\right]}{r_{\text{ab}}^2 } & \frac{4 e^{-2 i \alpha _ { \text{ab } } } \thinspace \pi \text{sin}\left[\beta _ { \text{ab}}\right]{}^2}{3 r_{\text{ab}}^3 } & \frac{2 \sqrt{2 } e^{-i \alpha _ { \text{ab } } } \thinspace \pi \text{sin}\left[2 \beta _ { \text{ab}}\right]}{3 r_{\text{ab}}^3 } & -\frac{4 \pi \text{sin}\left[\beta _ { \text{ab}}\right]{}^2}{3 r_{\text{ab}}^3 } \hdots \\
\frac{4 \pi } { r_{\text{ac } } } & \frac{2 \sqrt{\frac{2}{3 } } e^{-i \alpha _ { \text{ac } } } \thinspace \pi \text{sin}\left[\beta _ { \text{ac}}\right]}{r_{\text{ac}}^2 } & \frac{4 \pi \text{cos}\left[\beta _ { \text{ac}}\right]}{\sqrt{3 } r_{\text{ac}}^2 } & -\frac{2 \sqrt{\frac{2}{3 } } e^{i \alpha _ { \text{ac } } } \thinspace \pi \text{sin}\left[\beta _ { \text{ac}}\right]}{r_{\text{ac}}^2 } \hdots \\ \vdots & \vdots & \vdots & \vdots \end{array}\right)\ ] ] as an example , the interaction between the @xmath154 dipole on sphere a and the @xmath154 dipole on sphere b is given by @xmath155{}^2}{3 r_{\text{ab}}^3}$ ] since the seventh term in the vector @xmath120 is @xmath156 and the third term in @xmath120 is @xmath157 .
this term will also appear as @xmath158 due to symmetry of the interaction . | the electrostatic interaction between metal spheres is an influential component in the assembly of many nanoscale materials in chemistry . here
we derive a method to calculate the energy and polarizations of metal spheres in arbitrary configurations to an arbitrary multipole order .
this helps provide insight into the preferred configurations of charged metal particles and demonstrates the sensitivity of electrostatic interactions to configuration geometry . |
the `` hydrogen atom '' problem of meson spectroscopy is the study of the pion in terms of its fundamental constituents .
typically , this difficult problem is finessed by using an effective lagrangian treatment of the composite field which includes the important feature of ( almost ) spontaneously broken chiral symmetry .
then one explains the presumed next highest mass meson the rho as a @xmath2 bound state and continues up the spectrum .
but nowadays there is increasing support for the existence of the old `` sigma '' resonance which may be lighter than the rho .
if this is true it certainly seems worthwhile to pause and examine the issue in detail .
it is also a difficult problem because the sigma is in an energy range just above where one expects chiral perturbation theory to be practical but well below where asymptotic freedom permits a systematic perturbative qcd expansion . in this talk
a recent paper@xcite on the subject will be discussed .
other work is referenced in that paper and in other contributions @xcite to this conference .
first , a brief review of our previous results based on the non linear chiral lagrangian will be given .
then we try to check the form of these results by using the linear sigma model .
this model , while less general , provides the usual physical intuition about the problem as it contains a scalar nonet linked to the pseudoscalars .
* pi pi scattering*@xcite .
it was noted that the @xmath3 partial wave amplitude up to about 1 gev could be simply explained as a sum of four pieces : i. the current algebra `` contact '' term , ii .
the @xmath4 exchange diagram iii . a non breit wigner @xmath5 pole diagram and exchange , iv .
an @xmath6 pole in the background produced by the other three .
this is illustrated in a step by step manner for the real part @xmath7 in figs . 1 , 2 and 3 .
we see in fig .
1 that the current algebra " piece starts violating the unitarity bound , @xmath8 at about 0.5 gev and then runs away .
however the inclusion of the @xmath4 meson exchange diagrams turns the curve in the right direction and improves , but does not completely cure , the unitarity violation .
these pieces , which do not involve any unknown parameters , give encouragment to our hope that the cooperative interplay of various pieces can explain the low energy scattering . in order to fix up fig .
1 we note that the real part of a resonance contribution vanishes at the pole , is positive before the pole and _ negative _ above the pole .
thus a scalar resonance with a pole roughly about 0.5 gev ( above which @xmath7 in fig .
1 needs a negative contribution to stay below @xmath9 ) should do the job .
the result of including such a @xmath5 pole , with three parameters , is shown in fig .
2 . now note that the predicted @xmath10 in fig .
2 vanishes around 1 gev .
thus the phase @xmath11 at 1 gev ( assumed to keep rising ) is about @xmath12 . considering this as a background phase for the known @xmath6 ,
the real part of the @xmath6 contribution will get reversed in sign ( ramsauer townsend effect ) . as fig .
3 shows this is the missing piece needed to give a simple explanation of the @xmath13 @xmath14 scattering up to about 1 gev .
* pi k scattering*@xcite . in this case
the low energy amplitude is taken to correspond to the sum of a current algebra contact diagram , vector @xmath4 and @xmath15 exchange diagrams and scalar @xmath16 , @xmath6 and @xmath17 exchange diagrams .
the situation in the interesting @xmath18 s - wave channel turns out to be very analogous to the @xmath19 channel of s - wave @xmath14 scattering .
now a non breit wigner @xmath20 is required to restore unitarity ; it plays the role of the @xmath16 in the @xmath14 case .
we found that a satisfactory description of the 1 - 1.5 gev s - wave region is also obtained by including the well known @xmath21 scalar resonance , which plays the role of the @xmath6 in the @xmath14 calculation . *
putative light scalar nonet*@xcite .
the nine states associated with the @xmath16 , @xmath17 , @xmath6 and @xmath22 are required in order to fit experiment in our model .
what do their masses and coupling constants suggest about their quark substructure ?
clearly the mass ordering of the various states is inverted compared to the `` ideal mixing''@xcite scenario which approximately holds for most meson nonets .
this means that a quark structure for the putative scalar nonet @xmath23 is unlikely since the mass ordering just corresponds to counting the number of heavier strange quarks .
then the degenerate @xmath6 and @xmath22 which must have the structure @xmath24 would be lightest rather than heaviest .
however the inverted ordering will agree with this counting if we assume that the scalar mesons are schematically constructed as @xmath25 where @xmath26 is a `` dual '' quark ( or anti diquark ) .
this interpretation is strengthened by consideration @xcite of the scalars coupling constants to two pseudoscalars .
that shows @xmath27 `` small '' , so it is a predominantly non - strange particle in this picture .
furthermore the states @xmath24 now would each have two strange quarks and would be expected to be heaviest .
the four quark picture was first suggested a long time ago@xcite on dynamical grounds . * mechanism for next heavier scalar nonet*@xcite .
of course , the success of the phenomenological quark model suggests that there exists , in addition , a nonet of conventional `` p - wave @xmath2 scalars in the 1 + gev range .
the experimental candidates for these states are @xmath28 , @xmath29 and for @xmath19 , @xmath30 , @xmath31 and @xmath32 .
these are enough for a full nonet plus a glueball .
however it is puzzling that the strange @xmath21 is nt noticeably heavier than the non strange @xmath33 and that they are not lighter than the corresponding spin 2 states .
these and another puzzle may be solved in a natural way@xcite if the heavier p - wave scalar nonet mixes with a lighter @xmath34 nonet of the type mentioned above .
the mixing mechanism makes essential use of the ' ' bare `` lighter nonet having an inverted mass ordering while the heavier ' ' bare `` nonet has the normal ordering .
a rather rich structure involving the light scalars seems to be emerging . at lower energies one may consider as a first approximation , ' ' integrating out " the heavier nonet and retaining just the lighter one .
in @xcite we employed the conventional chiral field @xmath35 , where @xmath36 is a hermitian matrix of nine scalars and @xmath37 is a hermitian matrix of nine pseudoscalars .
the lagrangian density is , @xmath38 the three @xmath39 s are numbers proportional to the ( current ) quark masses .
@xmath40 may be considered to be an arbitrary function of the chiral @xmath41 invariants constructed from @xmath42 and @xmath43 .
note that most consequences at tree level follow just from chiral symmetry , irrespective of the form of @xmath40 . * pi pi scattering amplitude*. the computed @xmath3 partial wave amplitude at tree level has the form , @xmath44 + sin^2\psi[{\tilde \alpha}(s)+\frac{{\tilde
\beta}(s ) } { m^2_{bare}(\sigma^\prime)-s } ] , \label{pipiamp}\ ] ] where @xmath45,@xmath46 etc .
are given in connection with eq ( 3.2 ) of @xcite .
@xmath47 is a mixing angle between the two @xmath19 scalars , denoted as @xmath5 and @xmath48 .
the subscript `` bare '' on their masses means the value at tree level . if @xmath40 is general , the three quantities @xmath47 , @xmath49 and @xmath50 may be chosen at will .
however if @xmath40 is taken to be renormalizable there is only one arbitrary parameter ( say @xmath49 ) in the theory when the input set ( say @xmath51 ) is fixed .
it is instructive to first go back to the widely treated two flavor case .
this corresponds to choosing @xmath52 in ( [ pipiamp ] ) .
then @xmath49 is the only unknown parameter . near threshold ,
if @xmath49 is not too low , the amplitude is the `` current algebra '' result which agrees fairly well with experiment .
it is a small quantity which emerges from an almost complete cancellation of the pole and non pole terms in ( [ pipiamp ] ) .
one would like to keep this result and utilize the effect of the sigma at higher energies . since there is a true pole in ( [ pipiamp ] ) it seems reasonable to regulate this in the usual way by adding a term @xmath53 in the pole denominator .
however , as achasov and shestakov @xcite pointed out , this regulation completely destroys the good current algebra result .
they instead adopt the k matrix approach ( whereas the usual solution is to adopt the non linear model instead since the derivative coupling of the @xmath5 there suppresses the pole contribution near threshold ) . in this way
the tree amplitude is not only regularized but made exactly unitary .
one calculates the amplitude in terms of its tree value as : @xmath54 when @xmath55 is small , @xmath56 so the behavior near threshold will now not be spoiled . at the other extreme , when @xmath55 gets very large @xmath57 .
note that the pole position , @xmath58 of the unitarized @xmath59 will typically correspond to mass and width ( via @xmath60 ) which differ from @xmath61 and the starting perturbative width .
which one should be chosen ?
since @xmath59 in ( [ kmatrix ] ) evidently has the structure of a `` bubble sum '' in field theory it seems reasonable to regard the k matrix unitarization as an approximation to including the `` radiative corrections '' .
then , as in usual field theory , the pole found is interpreted as giving the physical mass and width while the values of @xmath61 and @xmath62 would have no special significance . for the two flavor model we verified the result of @xcite that a choice of @xmath49 around 0.8 to 1.0 gev would result in a physical @xmath63 around 0.45 gev and fit the first bump in fig .
the physical mass is not very sensitive to the exact choice of bare mass and also the physical width is very greatly reduced .
the predicted mass in this model is a bit less than the one we found in the non linear model reviewed in the previous section , but this is readily understandable as being due @xcite to the neglect of vector mesons in the present model .
* three flavor linear model amplitude*. the procedure was simply to use the full two pole tree amplitude ( [ pipiamp ] ) in the unitarization formula ( [ kmatrix ] ) .
we were not able to fit the entire @xmath64 amplitude shown in fig .
3 up to about 1.2 gev in the renormalizable model ( which contains only the single unknown parameter @xmath49 ) .
however it is easy to find a fit in the chiral model with general @xmath40 , in which we were able to choose the three parameters @xmath65 gev , @xmath66 gev and @xmath67 .
the physical isoscalar masses ( after unitarization ) turned out to be 0.457 gev and 0.993 gev associated with respective widths 0.632 gev and 0.05 gev .
again these represent large shifts from the bare values . for illustrative purposes the unitarized amplitude
is reasonably approximated as @xmath68 neither the first ( @xmath5 ) pole nor the second ( @xmath6 ) pole is precisely of breit wigner type .
however the @xmath6 pole approximates a breit wigner except for an overall minus sign , which corresponds to the well known `` flipping '' of this resonance .
we similarly studied the @xmath69 scattering amplitude to find the properties of the @xmath20 resonance in the linear model .
the bare mass of the @xmath20 is fixed once the input parameters are given . to allow us to vary this quantity ( in a range where the @xmath70 scattering is not much affected ) we chose the alternative input set @xmath71 and varied @xmath72 . in this case , because the @xmath21 is not included in the model we can only fit the data up to about 1 gev .
it was found that the best fit corresponded to the bare @xmath20 mass about 1.3 gev .
after unitarization the physical kappa mass turned out to be about 0.800 gev and this did nt change much as the bare value was varied from 0.9 to 1.3 gev .
unitarization also substantially narrowed the physical kappa width .
furthermore , as for the case of the @xmath5 the @xmath20 pole is not of breit wigner type .
an analogous calculation was carried out to study the properties of the @xmath22 as observed in @xmath73 scattering . a summary shown in table 1
compares the physical widths obtained in this linear model with those obtained in the non linear model . in the cases of the @xmath6 and the @xmath22
the entries were taken from the particle data group @xcite , with which the non linear model calculations agree . clearly , the complex pole positions and nature of the poles ( non breit wigner for @xmath5 and @xmath20 and `` ramsauer townsend '' for @xmath6 ) of the scalar nonet in the linear sigma model are similar to those obtained previously ( putative scalar nonet ) using a non linear chiral lagrangian with a different `` local '' regulation .
this statement makes heavy use of the unitarization of the three flavor linear model ; otherwise the @xmath6 and @xmath20 might be considered too high and wide to belong to a light scalar nonet . in particular ,
the @xmath20 clearly can not be identified with the @xmath21 in this unitarized linear sigma model . * speculation on scalar quark structure * ( section v of @xcite ) . at an intuitive level one might expect the scalar nonet , being the `` chiral partner '' of the light pseudoscalar nonet , to have a quark- anti quark structure .
it was stressed @xcite however that in the more general non linear lagrangian approach ( e.g. @xcite ) the scalar and pseudoscalar transformation properties are decoupled .
only the flavor su(3 ) transformation property , not the chiral one , of the scalars is fixed in the effective non - linear lagrangian treatment .
features , mentioned above , like isoscalar mixing angle and mass ordering suggest in fact the @xmath74 structure for the light scalars as an initial approximation .
how might this kind of scenario play out in the linear model where the chiral properties of the scalars and the pseudoscalars are clearly linked ? even there , the quark substructure implied by the @xmath75 transformation properties of the chiral matrix @xmath42 in ( [ lagrangian ] ) is not unique @xcite ( however the @xmath76 transformations do distinguish between @xmath2 and @xmath77 ) .
there are three different `` four quark '' structures with the same transformation properties .
physically , they correspond to making the chiral mesons as a ) meson meson `` molecule '' b ) spin 0 diquark -spin 0 anti diquark and c ) spin 1 diquark - spin 1 anti diquark . actually these three are not linearly independent . thus the molecule @xcite and diquark- anti diquark @xcite pictures are not clearly distinguished at the effective lagrangian level .
presumably , large changes in the properties of the scalars due to unitarization in the effective theory must be counted as `` four quark '' admixtures at the underlying level . in detail
, the schematic structure for the matrix @xmath78 realizing a @xmath79 composite in terms of quark fields @xmath80 can be written @xmath81 where @xmath82 and @xmath83 are respectively flavor and color indices .
for the `` molecule '' model a ) the schematic quark structure with the same @xmath84 transformation property is , @xmath85 in the spin 0 diquark - spin 0 anti diquark case the same transformation property is realized with , @xmath86 where @xmath87 in which @xmath88 is the charge conjugation matrix of the dirac theory .
finally the spin 1 diquark - spin 1 anti diquark case c ) has the schematic structure , @xmath89 where @xmath90 and @xmath91 $ ] .
now , as discussed before , the realistic situation is likely to contain substantial mixing between scalar @xmath2 and @xmath77 nonets . to explore this
we formulated a linear sigma model containing both a @xmath2 chiral matrix @xmath92 and a @xmath77 chiral matrix @xmath93 .
this is a very complicated system so we started with a `` toy model '' in which all current quark masses are neglected and only a minimum number of non derivative terms are included .
in addition to two minimal kinetic terms as in ( [ lagrangian ] ) , we took the simplified potential , @xmath94 here @xmath95 , @xmath96 and @xmath97 are positive real constants .
the @xmath42 matrix field is chosen to have a wrong sign mass term so that there will be spontaneous breakdown of chiral symmetry .
a pseudoscalar octet will thus be massless .
the mixing is controlled by the parameter @xmath98 .
it is amusing to note that there will then be an induced @xmath77 condensate @xmath99 in addition to the usual @xmath2 condensate @xmath100 .
we found that it is easy to obtain a situation where the the next highest state above the predominantly @xmath2 nambu goldstone pseudoscalar octet is a predominantly @xmath101 scalar octet .
still heavier is the predominantly @xmath101 pseudoscalar while heaviest of all is the @xmath2 scalar octet . of course , su(3 ) symmetry breaking and unitarization would be expected to modify this picture .
it seems very interesting to further pursue a model of this type .
there is evidently a possibility of learning a lot about non perturbative qcd from the light scalar system .
we would like to thank francesco sannino and masayasu harada for fruitful collaboration .
one of us ( j.s . ) would like to thank the organizers for arranging a stimulating and enjoyable conference .
the work has been supported in part by the us doe under contract de - fg-02 - 85er40231 .
see the write - ups of n. achasov , c. gobel , s. ishida , m. ishida , t. kunihiro , j. l. lucio - martinez , g. moreno , w. ochs , j. pelaez , yu .
surovtsev , k. teshihiko , t. teshima , s. f. tuan and e. van beveren . | we first briefly review a treatment of the scalars in meson meson scattering based on a non - linear chiral lagrangian , with unitarity implemented by a `` local '' modification of the scalar propagators .
it is shown that the main results are confirmed by a treatment in the su(3 ) linear sigma model in which unitarity is implemented `` globally '' .
some remarks are made on the speculative subject of the scalars quark structure .
.1cm_@xmath0 department of physics , syracuse university , syracuse , new york 13244 - 1130 , usa._.2cm_@xmath1 department of mathematics / science , state university of new york institute of technology , utica , new york 13504 - 3050 , usa . _ |
some quotes in the many worlds literature suggest a belief that one can derive the canonical structure from the hamilton operator taken alone , given as an abstract linear operator in some hilbert space , without any additional structure .
for example , tegmark describes the construction of a `` preferred basis '' in many worlds : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` this elegant mechanism is now well - understood and rather uncontroversial [ ] . essentially , the position basis gets singled out by the dynamics because the field equations of physics are local in this basis , not in any other basis . ''
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ this is ( as indicated by the `` essentially '' ) an oversimplification : the decoherence - based construction considered there depends not only on the dynamics ( the hamilton operator ) , but also on some `` subdivision into systems '' a tensor product structure as can easily be seen in the quoted papers by zurek : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` one more axiom should [ be ] added to postulates ( i ) - ( v ) : ( o ) the universe consists of systems . ''
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ but some comments made by zurek suggest that he shares the belief that physics is completely defined by the hamilton operator as well : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` both the formulation of the measurement problem and its resolution through the appeal to decoherence require a universe split into systems .
yet , it is far from clear how one can define systems given an overall hilbert space of everything and the total hamiltonian . ''
@xcite `` [ a ] compelling explanation of what are the systems how to define them given , say , the overall hamiltonian in some suitably large hilbert space would be undoubtedly most useful . ''
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ indeed , the problem `` how '' to define these systems seems to assume , at least implicitly , that these systems can be defined given the hamiltonian .
then , based on the decoherence technique , the preferred basis can be defined as well .
the following quote suggests that vaidman shares this belief too : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ i believe that the decomposition of the universe into sensible worlds is , essentially , unique .
the decomposition , clearly , might differ due to coarse or fine graining , but to have essentially different decompositions would mean having a multi - meaning escher - type picture of the whole universe continuously evolving in time .
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if we interpret the `` decomposition into sensible worlds '' as something based on the decoherence - constructed `` preferred basis '' , the uniqueness of this decomposition implies the uniqueness of the `` preferred basis '' as well .
schlosshauer talks about _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the physical definition of the preferred basis derived from the structure of the unmodified hamiltonian as suggested by environment - induced selection @xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ which also suggests that he shares the belief .
last but not least , let s quote brown and wallace @xcite .
they discuss the possible non - uniqueness of the `` preferred basis '' picked out by decoherence : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` granted that decoherence picks out a quasi - classical basis as preferred , what is to say that it does not also pick out a multitude of other bases very alien with respect to the bases with which we ordinarily work , perhaps , but just as preferred from the decoherence viewpoint . such a discovery would seem to undermine the objectivity of everettian branching , leaving room for the bohmian corpuscle to restore that objectivity . ''
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ and present the following response as preferable : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` granted that we can not rule out the possibility that there might be alternative decompositions , and that this would radically affect the viability of the everett interpretation
well , right now we have no reason at all to suppose that there actually are such decompositions .
analogously , logically we ca nt absolutely rule out the possibility that there s a completely different way of construing the meaning of all english words , such that they mostly mean completely different things but such that speakers of english still ( mostly ) make true and relevant utterances . such a discovery would radically transform linguistics and philosophy , but we do nt have any reason to think it will actually happen , and we have much reason to suppose that it will not . to discover one sort of higher - level structure in microphysics (
be it the microphysics of sound - waves or the micro - physics of the wave - function ) is pretty remarkable ; to discover several incompatible structures in the same bit of microphysics would verge on the miraculous . ''
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the aim of this paper is to show that this miracle happens the theory of the korteweg - de vries ( kdv ) equation gives nice counterexamples for this thesis : if a potential @xmath0 is a solution of the kdv equation , then the operators @xmath1 for different @xmath2 appear to be unitarily equivalent , despite defining different physics .
thus , the physics of canonical quantum theories is not completely defined by the hamilton operator alone .
this fact seems fatal for the idea to derive a preferred basis , using decoherence techniques , from the hamilton operator taken alone .
one needs an additional structure be it the tensor product structure related with the `` decomposition into systems '' or whatever else which has to be postulated .
we consider the question if this construction could be , nonetheless , used as a foundation of quantum theory , as a replacement for simply postulating the configuration space .
we argue that this construction of the preferred basis combines the disadvantages of postulated structures ( lack of explanatory power ) and emergent structures ( uncertainty , dependence from other structures , especially dynamics ) , and , therefore , should be rejected in favour of the canonical way to postulate the configuration space as a non - dynamical structure , as done in canonical quantum theories as well as pilot wave theories .
thus , the `` derivation '' of the `` preferred basis '' based on decoherence techniques seems useless in the domain of fundamental physics .
it has it s useful applications in situations where we already have ( as in the copenhagen interpretation or in pilot wave theories ) a classical part , which defines the decomposition into systems which one needs to derive a decoherence - preferred basis .
some clarification of how we interpret this belief seems useful .
canonical quantum theories are defined in a more or less standard way : first , one defines the kinematics by defining some hilbert space with some set of canonical operators @xmath3 , @xmath4 with commutation relation @xmath5 = -i\hbar\delta^j_i$ ] on , or , equivalently , to postulate some configuration space @xmath6 with coordinates @xmath7 so that @xmath8 and @xmath9
. then , in a second step , one defines the dynamics by postulating the schrdinger equation @xmath10 for some hamilton operator , usually of the form @xmath11 implicit part of the definition of the canonical operators @xmath3 , @xmath4 is their identification with classical observables .
it is hard to press this part of the definition into some formal property it depends on the particular theory .
the point is that the definition is only complete if we know what it means to measure the configuration @xmath6 it is some procedure , usually known from the corresponding classical theory .
the details of this procedure do not matter in a general discussion .
but in order to apply the theory to make concrete predictions , one has to know how to measure the @xmath12 and @xmath13 . without this additional information
the physical definition of the theory is not complete .
the quantum theory predicts the result of this measurement for a state @xmath14 as @xmath15 . but
this information would be useless if we do nt know how to measure @xmath6 . given this physical meaning of @xmath6 , it seems , at a first look , completely unreasonable to think that the physics is completely defined by the hamilton operator alone .
nobody would apply , for example , some unitary transformation @xmath16 to @xmath17 and @xmath18 , but not to @xmath3 , and @xmath4 , but nonetheless claim that the physics remain unchanged .
everybody knows that one has to apply the same unitary transformation to @xmath19 and @xmath20 as well to preserve physics .
now , with our interpretation of the quotes , we are in no way suggesting that the authors do not recognize this .
the idea that the hamilton operator alone is sufficient is a different one . given the very special form of the hamilton operator , one can imagine that it may be possible to _ reconstruct _ the configuration space @xmath6 more or less uniquely given as an abstract operator on some abstract hilbert space .
last but not least , the straightforward ideas to construct counterexamples fail : one can apply coordinate transformations @xmath21 , but these do not change the physics , and leave the abstract configuration space @xmath6 unchanged .
one can consider a canonical rotation @xmath22 , @xmath23 , but this gives a very strange nonlocal operator of type @xmath24 , which forbids a physical interpretation of the new @xmath25 in terms of a configuration .
therefore , one can conclude that there are only few sufficiently nice choices of @xmath4 which allow a meaningful interpretation of the @xmath4 as configuration space coordinates , and one could hope or expect that these few choices lead , at least approximately , to physically equivalent theories . in this case
, it would be indeed the hamilton operator taken alone , as an abstract operator on a hilbert space , which would be sufficient to define the physics completely .
but this expectation is false , as we will show below .
a really beautiful one - dimensional example of different representations follows from the theory of solutions of the korteweg - de vries ( kdv ) equation : if the function @xmath0 is a solution of the korteweg - de vries equation @xmath26 then the operators @xmath27 for different @xmath2 are unitarily equivalent . indeed , as has been found by lax ( @xcite ) , the kdv equation is equivalent to the operator equation @xmath28.\ ] ] for the self - adjoint operator @xmath29 as one can easily check . but this type of operator equations defines , for a self - adjoint operator @xmath30 , a unitary evolution of @xmath31 .
indeed , this is simply the analogon of the heisenberg equation for the `` hamilton operator '' @xmath30 , applied to @xmath31 .
the unitary transformation we need is defined by the equation @xmath32 for results about the existence of solutions of the kdv equation see , for example , @xcite .
this representation leaves and invariant but changes @xmath31 and @xmath33 by @xmath34 resp .
this is not yet exactly our point .
but , given the unitary operators @xmath36 , we can also consider another equivalent representation : [ hfixed ] for a given hamilton operator @xmath37 , with @xmath38 as defined by , there exist canonical operators @xmath39 , @xmath40 , so that the representation of in terms of @xmath41 is given by .
indeed , it is sufficient to define the operators @xmath39 , @xmath40 by @xmath42 and @xmath43 . in this case , the unitary transformation @xmath44 , applied to the operators @xmath45 , gives @xmath46 , thus , the representation of in terms of @xmath41 is @xmath31 .
it is worth to note that the hamiltonian in this theorem does not depend on @xmath2 .
only the operators @xmath41 depend on @xmath2 .
this has the consequence that the representation @xmath47 in terms of these @xmath2-depending operators depends on @xmath2 too .
but is , nonetheless , the same for all @xmath2 .
thus , given the operator alone , we can not reconstruct the operators , uniquely .
the operators @xmath40 , @xmath39 give equally nice candidates for canonical variables for : the representation of in these operators has the same canonical form @xmath48 , and the potential functions @xmath49 in this standard form are as nice and well - behaved as the original @xmath50 .
two - soliton - solution @xmath51 of the kdv equation for different values of the evolution parameter @xmath2 .
picture taken from @xcite , scaledwidth=80.0% ] let s clarify if the different potentials @xmath52 really define different physics .
this seems obvious , if one looks at particular examples , like in fig .
[ fig : solitons ] .
the two sharply localized part of the solutions in the first and last picture of figure [ fig : solitons ] are so - called solitons , special solutions , which , taken alone , have the exact form @xmath53 and move with velocity @xmath54 .
their spectrum is defined by a single eigenvalue @xmath55 , with an eigenfunction @xmath56 localized in the same domain .
some superpositions of the two eigenstates would be , in one case ( first and last picture ) , clearly delocalized , in another one ( upper right picture ) they are all localized in the same region . an experiment which would allow us to measure a phase difference between the reflection coefficients @xmath57 of the horizontal and vertical mirrors . for reflection angles close enough to @xmath58 , only the one - dimensional reflection coefficient in orthogonal direction matters .
thus , we can put the different one - dimensional potentials for different @xmath2 into the mirrors , extended trivially in the other direction , say , a localized one at the bottom , and one with two solitons on the right.,scaledwidth=60.0% ] we can also consider the scattering matrix .
the inverse scattering method ( developed by @xcite , see @xcite ) for solving the kdv equation gives the following explicit result for the one - dimensional scattering matrix : one of the two coefficients of the scattering matrix , namely @xmath59 , appears to be an integral of motion of the kdv equation .
instead , the reflection coefficient @xmath57 depends explicitly on @xmath2 . to construct an experiment which allows
the measurement of such differences is not difficult ( see fig . [
fig : reflection ] ) . if not a different scattering matrix , what else defines different physics ?
the example we have given is one - dimensional . one can consider a straightforward generalization to higher dimension by considering hamilton operators of the form @xmath60 with different one - dimensional potentials @xmath61 . but all we obtain in this way are only non - interacting degrees of freedom
this seems to leave some hope for the case of non - trivially interacting hamiltonians in higher dimension .
the interactions between different degrees of freedom could , possibly , allow the choice of a preferred basis .
this is , essentially , the way used in the decoherence - based approach for the construction of a preferred basis .
one starts with a `` decomposition into systems '' : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \(o ) the universe consists of systems.@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ that means , a tensor product structure @xmath62 on the hilbert space of everything .
in general , each factor @xmath63 interacts with it s environment @xmath64
. in particular , if has the standard form @xmath65 the interaction hamiltonian is @xmath66 .
the preferred basis is , in a simplified version,[multiblock footnote omitted ] the one which is measured by the interaction hamiltonian . in case of @xmath66 as the interaction hamiltonian between @xmath63 and its environment @xmath67 , it is @xmath4 which is measured by the environment .
thus , we can recover the operator @xmath4 on @xmath63 by taking into account the interaction with the environment .
the one - dimensional counterexample , as well as the straightforward non - interacting examples , could be interpreted as irrelevant exceptions , which play no role in the multi - dimensional interacting case .
unfortunately , this construction depends on the predefined tensor product structure .
and this tensor product structure is not unique : [ th : tensorproduct ] there exists a hamilton operator in a hilbert space such that there exist different tensor product structures @xmath68 with canonical variables @xmath69 and @xmath70 on the factor spaces @xmath71 resp .
@xmath72 so that has in all of them the standard canonical form @xmath73 with a non - trivial interaction potential @xmath74 , which depends non - trivially on @xmath2 .
this can be easily seen in a minor variant of the straightforward multidimensional extension of the kdv example .
let s start with a simple degenerated two - dimensional hamilton operator @xmath75 similar to theorem [ hfixed ] , we choose as well as @xmath76 as fixed , but @xmath77 as depending on @xmath2 .
now , we define the tensor product structure we need by @xmath78 in these variables , the interaction potential is already nontrivial .
but , it has yet the same nice standard canonical form , and as in theorem [ hfixed ] , the resulting potential @xmath79 is of comparable nice quality for different @xmath2 .
the tensor product structure depends on @xmath2 . a nice implicit way to see
this is to use the fact that the simplified version of decoherence already allows the unique derivation of the positions @xmath80 as the decoherence - preferred observables , for the given tensor product structure @xmath81 .
but , on the other hand , the result is obviously not unique .
this contradiction disappears once we recognize that the tensor product structure is not unique .
but the @xmath2-dependence of the tensor product structure can be seen directly as well : if the tensor product structure would be the same for different @xmath2 , we would be able to express @xmath82 as a function @xmath83 .
but an attempt to express @xmath82 in this way fails for a general @xmath36 there is no chance to get rid of the dependence on @xmath84 : @xmath85 let s emphasize again that the operator does not depend on @xmath2 . only the operators @xmath86 depend on @xmath2 .
therefore , also the representation of in terms of these operators depends on @xmath2 .
but the operator itself has not only the same spectrum , but is simply the same for different @xmath2 . as a consequence
, one should give up the hope expressed by zurek : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` [ a ] compelling explanation of what are the systems how to define them given , say , the overall hamiltonian in some suitably large hilbert space would be undoubtedly most useful . ''
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ already in the two - dimensional case there is no unique way to reconstruct a physically reasonable , nice tensor product structure , or `` decomposition into systems '' , from a given hamilton operator taken alone .
but , maybe the decoherence - based reconstruction of the configuration space basis is , nonetheless , worth something ?
last but not least , even if it depends on some other structure does nt it , nonetheless , explain something important about the fundamental nature of the configuration space ?
we do nt think so , and the aim of this section is to explain why .
note that we do not want to question at all that there are lot s of useful non - fundamental applications of this construction applications where a subdivision into systems is defined by the application .
the systems in these applications will be various measurement instruments and state preparation devices , various parts of the environment , and the quantum system which is interesting in the particular application .
the preferred basis , constructed in this way , is also application - dependent . while it may be very important to find such a basis for a particular application , these constructions seem irrelevant in considerations of the foundations of physics .
a variant is to start from a tensor product decomposition given by the fundamental , postulated configuration space @xmath87 .
the resulting decoherence - preferred basis may be different from the position basis and better suited for the consideration of the classical limit .
but this would be a non - fundamental application as well , with no relevance for the foundations of quantum theory , which is , in this variant , postulated in the usual way .
the only variant which seems relevant for fundamental physics is to replace the a - priori definition of the configuration space in canonical quantum theories as well as pilot wave theories by the decoherence - based construction of the preferred basis , based on some fundamental tensor product structure . to evaluate this replacement
, we have to compare it with the standard alternative to postulate the configuration space @xmath6 as a predefined , non - dynamical structure .
first , the above competitors depend on predefined , non - dynamical structures , which are introduced into the theory in an axiomatic way .
the standard approach postulates the configuration space @xmath6 itself , and the decoherence - based construction postulates a tensor product structure . in this sense , they are on equal footing .
but the resulting structure the configuration space @xmath6 is , in the standard approach , a predefined , non - dynamical object . instead , in the decoherence - based construction
, it depends on dynamics .
this dependence of @xmath6 on dynamics has at least one obvious disadvantage : we can no longer define the dynamics in the canonical way as @xmath88 which uses simple and natural structures on @xmath6 the laplace operator @xmath89 on @xmath6 , as well as the special subclass of multiplication operators in @xmath90 , and , in addition , some special function @xmath91 like @xmath92 .
indeed , such a definition of in terms of @xmath6 would become circular .
thus , we loose a nice and simple way to define the dynamics of the theory
. we would have to define the dynamics in some other way .
we can expect that this other way is more complex , and less beautiful .
thus , for the emergent , derived character of @xmath6 we have to pay . in itself , this is not untypical for emergent objects : we have to pay some costs for the possibility to explain them . in a previous theory , they have been postulated as fundamental , simple , independent objects , with some nice , well - defined properties . in the new theory , which derives them as emergent objects , they become more complex , often with uncertain boundaries , and they depend on other objects and structures .
nonetheless , the special character of the loss in our case seems untypical even for emergent objects .
usually , the structures which depend on the objects which now become emergent , are or of some higher , emergent level already in the previous theory , or they become emergent together with these objects in the new theory .
i do nt know of an example of an object which depends on another object in the old theory , and then the other object becomes emergent , but the object itself remains fundamental , as it would be the case for .
( of course , this could happen , but it would indicate that the dependencies were wrong already in the old theory . )
thus , i would characterize the costs related with the dependence of @xmath6 on the dynamics as higher than usual for emergent objects .
let s consider now what we gain .
usually , if an object becomes emergent which was previously fundamental , our gain is explanatory power .
do we have such a gain in our case ? here , we have to take into account that we need another , postulated structure the tensor product structure to construct @xmath6 .
thus , we can `` explain '' the configuration space @xmath6 only in a relative sense , in comparison with the unexplained , postulated tensor product structure .
now , explaning one structure in terms of another may also give large explanatory power .
last but not least , in some sense all our more fundamental theories are of this type they explain the previously postulated objects and structures in terms of some more fundamental , but also postulated , objects and structures .
but we would talk about explanatory power only if the new , more fundamental structure has some advantages in simplicity , beauty , generality , or whatever else .
what is the situation in our case ?
at least i can not see any advantage .
instead , i see a lot of disadvantages : first , we have simple examples of quantum theories which have natural configuration spaces but do not have a ( similarly natural ) tensor product structure : for example , finite - dimensional quantum theories with prime dimension do not have nontrivial tensor product structures .
indeed , from follows @xmath93 , thus the only possible tensor product structure for a space with prime dimensnion is the trivial one , which is worthless because it does not allow the start of the decoherence procedure .
then we have spaces of identical particles , which are factor spaces of a tensor product , but do not have their own natural to an arbitrary basis @xmath94 of . ]
tensor product structure .
last but not least , for topologically sufficiently non - trivial manifolds @xmath6 , starting with @xmath95 , there exists no decomposition @xmath96 into factor - manifolds @xmath97 , and therefore no natural tensor product structure .
but for all these hilbert spaces we have natural configuration spaces . then , the tensor product structure is often less symmetric in comparison with the configuration space .
the simplest example is the tensor product structure of @xmath98 of one - particle theory , which destroys rotational symmetry .
in addition , the tensor product structure based on points in field theory requires an identification of points in different time slices , if considered as fixed over time , destroying galilean or relativistic symmetry .
these arguments seem sufficient to argue that a tensor product structure , as a fundamental object , is worse than a configuration space .
thus , an explanation of @xmath6 in terms of a more fundamental tensor product structure has no explanatory power at all . the gain which we usually want to reach by deriving structures previously considered to be fundamental , namely explanatory power ,
can not be reached in this approach .
moreover , this can not be hailed in some way , say , by deriving the tensor product structure from something else .
the hamilton operator alone is not sufficient , as shown by our counterexample .
thus , one needs some additional structure anyway .
therefore , the main line of our argumentation remains intact .
the construction combines only disadvantages : the lack of explanatory power of postulated objects , with the uncertainty and dependency of emergent objects .
what remains intact is also the dependence of @xmath6 on the dynamics , and therefore the very special loss of the possibility to define the dynamics as @xmath99 on @xmath6 .
what also remains is the simplicity and the general and very natural character of postulating a configuration space @xmath6 .
therefore , whatever the new additional structure , we can not expect a large gain in explanatory power .
given this situation , there seem to be no gains but only losses , in a construction which constructs the configuration space @xmath6 postulated in canonical quantum theories and pilot wave theories using decoherence techniques , starting from a fundamental tensor product structure ( or some replacement ) .
we have shown that it is not the hamilton operator alone which defines the physics of quantum theories .
in addition , one needs the canonical configuration space @xmath6 , or some similar structure , which connects this hamilton operator with observable configurations . as a consequence , hopes to derive the configuration space basis using decoherence techniques from a hamilton operator taken alone have to be given up .
such constructions necessarily depend on some additional structure , like a `` subdivision into systems '' , which have to be postulated .
the derivation of the configuration space basis from such additional structures combines the disadvantages of predefined structures ( lack of explanatory power ) and emergent structures ( dependence on dynamics ) without giving any advantages thus loses in comparison with a simply postulated configuration space .
if these arguments against a decoherence - based construction of the preferred basis are of any relevance for everettians is a completely different question .
they may also completely ignore the non - uniqueness and follow some reasoning like this : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ suppose that there were several such decompositions , each supporting information - processing systems . then the fact that we observe one rather than another is a fact of purely local signicance : we happen to be information - processing systems in one set of decoherent histories rather than another .
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ indeed , once one introduces many worlds anyway , some more of them do not matter anymore .
all what i can say about this is to recommend a further `` improvement '' in this direction to assign reality not only to the state vector of our multiverse , but to all other states as well , and to be consequent , to all hamilton operators as well .
another possibility would be to throw away the `` solution of the preferred basis problem '' and instead to use the same predefined configuration space as used in canonical quantization and pilot wave theories . given the arguments in section [ sec : comparison ] , this would be an improvement . for pilot wave interpretations @xcite ,
the clarification that the configuration space is necessary to fix the physics of a canonical quantum theory is clearly helpful .
it weakens a quite common argument that the choice of the configuration space in pilot wave interpretations is artificial , like _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the artificial asymmetry introduced in the treatment of the two variables of a canonically conjugated pair characterizes this form of theory as artificial metaphysics .
( @xcite , as quoted by @xcite ) , `` the bohmian corpuscle picks out by fiat a preferred basis ( position ) ''
@xcite _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ instead , recognizing that the configuration space is part of the definition of the physics gives more power to an old argument in favour of the pilot wave approach , made already by de broglie at the solvay conference 1927 : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` it seems a little paradoxical to construct a configuration space with the coordinates of points which do not exist . ''
_ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
thanks ch . roth for correcting my poor english .
de broglie , l. , la nouvelle dynamique des quanta , in `` electrons et photons : rapports et discussions du cinquieme conseil de physique '' , ed .
j. bordet , gauthier - villars , paris , 105 - 132 ( 1928 ) , english translation in : bacciagaluppi , g. , valentini , a. : `` quantum theory at the crossroads : reconsidering the 1927 solvay conference '' , cambridge university press , and http://arxiv.org/abs/arxiv:quant-ph/0609184[arxiv:quant-ph/0609184 ] ( 2006 ) freire jr .
, o. : science and exile : david bohm , the hot times of the cold war , and his struggle for a new interpretation of quantum mechanics , historical studies on the physical and biological sciences 36(1 ) , 1 - 34 , http://arxiv.org/abs/arxiv:quant-ph/0508184[arxiv:quant-ph/0508184 ] ( 2005 ) wolfgang pauli , remarques sur le problme des paramtres cachs dans la mcanique quantique et sur la thorie de londe pilote , in andr george , ed . , louis de broglie physicien et penseur ( paris , 1953 ) , 33 - 42 schlosshauer , m. : decoherence , the measurement problem , and interpretations of quantum mechanics , rev .
76 , 1267 - 1305 ( 2004 ) http://arxiv.org/abs/arxiv:quant-ph/0312059[arxiv:quant-ph/0312059 ] vaidman , l. : on schizophrenic experiences of the neutron or why we should believe in the many - worlds interpretation of quantum theory , international studies in philosophy of science 12 , 245 - 261 http://arxiv.org/abs/arxiv:quant-ph/9609006[arxiv:quant-ph/9609006 ] ( 1998 ) brown , h.r .
, wallace , d. : solving the measurement problem : de broglie - bohm loses out to everett , foundations of physics , vol .
35 , no . 4 , 517 ( 2005 ) http://arxiv.org/abs/arxiv:quant-ph/0403094[arxiv:quant-ph/0403094 ]
zurek , w.h .
: decoherence , einselection , and the existential interpretation , philos .
london , ser . a 356 , 1793 - 1821 , http://arxiv.org/abs/arxiv:quant-ph/9805065[arxiv:quant-ph/9805065 ] ( 1998 ) zurek , w.h . : relative states and the environment : einselection , envariance , quantum darwinism , and the existential interpretation , http://arxiv.org/abs/arxiv:0707.2832[arxiv:0707.2832 ] and los alamos preprint laur 07 - 4568 ( 2007 ) | in the many worlds community there seems to exist a belief that the physics of quantum theory is completely defined by it s hamilton operator given in an abstract hilbert space , especially that the position basis may be derived from it as preferred using decoherence techniques .
we show , by an explicit example of non - uniqueness , taken from the theory of the kdv equation , that the hamilton operator alone is not sufficient to fix the physics . we need the canonical operators , as well . as a consequence ,
it is not possible to derive a `` preferred basis '' from the hamilton operator alone , without postulating some additional structure like a `` decomposition into systems '' .
we argue that this makes such a derivation useless for fundamental physics . |
the question of how the proton spin is carried by its constituents is still very actual @xcite .
the determination of the gluon contribution , @xmath7 , to the nucleon spin is the object of different experiments with polarized beams and targets @xcite . in particular
, the production of charmed particles in collisions of longitudinally polarized muons with polarized proton target will be investigated by the compass collaboration @xcite . in the framework of the photon - gluon fusion model @xcite , @xmath8 , ( @xmath9 is a virtual photon ) , the corresponding asymmetry ( for polarized @xmath10-collisions ) ,
is related to the polarized gluon content in the polarized protons @xcite . due to the future impact of such result
, it seems necessary to understand all the other possible mechanisms which can contribute to inclusive charm photoproduction , such as @xmath11 , for example .
one possible and @xmath12 important background , which can be investigated in detail , is the process of exclusive associative charm photoproduction , with pseudoscalar and vector charmed mesons in the final state , @xmath13 , with @xmath14 or @xmath15 .
the mechanism of photon - gluon fusion , which successfully describes the inclusive spectra of @xmath16 and @xmath4 mesons at high photon energies , can not be easily applied to exclusive processes , at any energy .
threshold considerations of such processes in terms of standard perturbative qcd have been applied to the energy dependence of the cross section of @xmath17 @xcite . however , in such approach , polarization phenomena can not be calculated without additional assumptions . in this respect , the formalism of the effective lagrangian approach ( ela ) seems very convenient for the calculation of exclusive associative photoproduction of charmed particles , such as @xmath18 @xcite at least in the near threshold region .
such approach is also widely used for the analysis of various processes involving charmed particles @xcite , as , for example , @xmath19-suppression in high energy heavy ion collisions , in connection with quark - gluon plasma transition @xcite .
we analyzed earlier different exclusive processes of associative charm particles photoproduction , @xmath18 and indicated that polarization phenomena can be naturally predicted for those reactions . in this paper
we apply the ela to the collision of circularly polarized photons with a polarized proton target , starting from the reaction threshold .
the asymmetry @xmath21 , for the collision of circularly polarized photons with a polarized proton target , @xmath22 is defined as : @xmath23 where @xmath24 @xmath25 is the cross section of @xmath26-collisions with parallel ( antiparallel ) spins ( along the direction of the photon beam , taken as the @xmath2direction ) .
simple kinematical considerations show that this asymmetry for exclusive photoproduction processes , such as @xmath27 is sizeable .
for example , in the @xmath28-wave approximation , which applies to the near threshold region , the @xmath29-spin structure of the corresponding amplitudes , ( where @xmath30 is the photon polarization vector ) induces a maximal asymmetry , @xmath31 , independently on the @xmath32 production angle in the reaction cms .
this result is correct for any value of the @xmath33wave threshold amplitude and can be directly obtained from the conservation of spin projection in threshold conditions . in the general case ,
any reaction @xmath34 is described by two different asymmetries and the dependence of the differential cross section on the polarization states of the colliding particles can be parametrized as follows ( taking into account the p - invariance of the electromagnetic interaction of charmed particles ) : @xmath35 where @xmath36 is the photon helicity , @xmath37 and @xmath38 are the possible components of the proton polarization @xmath39 , and @xmath40 , @xmath41 are the two independent asymmetries . due to the t - even nature of these asymmetries , they are non vanishing in ela , where the photoproduction amplitudes are real .
we will use the standard parametrization @xcite of the spin structure for the amplitude of pseudoscalar meson photoproduction on the nucleon : @xmath42\chi_1,\ ] ] @xmath43 and @xmath44 are the unit vectors along the three - momentum of @xmath45 and @xmath20 ; @xmath46 , @xmath47=1 - 4 , are the scalar amplitudes , which are functions of two independent kinematical variables , @xmath48 and @xmath49 , where @xmath50 is the @xmath51meson production angle in the reaction cms , @xmath52 and @xmath53 are the two - component spinors of the initial nucleon and the produced @xmath54-baryon . after summing over the @xmath54-polarizations ,
one finds the following expressions for the asymmetries @xmath40 and @xmath41 in terms of the scalar amplitudes @xmath46 : @xmath55 , \label{asymx}\ ] ] @xmath56 where @xmath57 \right\ } .
\label{eq : asymd}\end{aligned}\ ] ] note that @xmath40 vanishes at @xmath58 and @xmath59 .
moreover @xmath60 , for @xmath58 and @xmath59 , for any photon energy .
this is a model independent result , which follows from the conservation of helicity in collinear kinematics .
it is independent from the dynamics of the process , its physical meaning is that the collision of @xmath45 and @xmath61 with parallel spins can not take place for collinear regime .
this result holds for any process of pseudoscalar meson photo - production on a nucleon target .
but for @xmath62 and @xmath63 the results for @xmath40 and @xmath41 are model dependent . to estimate the @xmath64 and the @xmath65-dependences of these asymmetries we use the ela model .
the explicit expressions for the amplitudes @xmath66 , in terms of the @xmath67 coupling constant and the magnetic moments of the charmed baryons , in framework of ela , can be found in @xcite .
the starting point of this model is the @xmath68-quark exchange ( through the @xmath69-channel @xmath54-exchange ) .
other diagrams ( d - exchange in @xmath70channel and @xmath71-exchange in @xmath48-channel ) have to be included because the matrix element of this process has to satisfy the gauge invariance .
it is this symmetry of the electromagnetic interaction which drives the relative role of different possible mechanisms in @xmath72 , and therefore the angular distributions and the @xmath73-asymmetry induced by linearly polarized photons colliding with unpolarized target . in the near threshold region ,
the main contribution is represented by the @xmath69-channel mechanism @xcite .
this mechanism occurs for charm photoproduction at short distances , @xmath74 , where @xmath75 is the @xmath68-quark mass , in agreement with a previous study of threshold exclusive @xmath19 photoproduction @xcite .
this is the typical radius for exclusive charmed particle photoproduction , and a larger radius ( up to 1 fm ) can be achieved only for more complicated processes , where the light mesons exchanges are possible , such as @xmath76 .
applying this model in @xcite , we found the angular and the energy dependence of the differential cross section for the processes @xmath72 and compared the results with the existing experimental data on open charm photoproduction .
the absolute values of the corresponding cross sections are determined by the @xmath67 coupling constants , which are presently not known , and have been considered as adjustable parameters .
the deduced values for these coupling constants are too low in comparison with the predictions of @xmath77-symmetry .
such conclusion has been confirmed by our analysis @xcite of vector meson @xmath78photoproduction .
as the constant @xmath67 can not be independently determined , the relative contribution of individual channels to the inclusive open charm photoproduction can not be assessed .
however , a crude estimation of this contribution can be done from the measured charm - anticharm asymmetry in open charm photoproduction , because the exclusive processes generate @xmath20-mesons and charmed baryons . from the observed @xmath79-asymmetry @xcite
, one can estimate a contribution of exclusive photoproduction of open charm of the order of 20% , which may seriously affect the experimental @xmath51asymmetry in polarized @xmath80-collisions .
the asymmetries @xmath40 and @xmath41 do not depend on the coupling constant @xmath67 , in the framework of the considered model , but they depend on @xmath81 , the anomalous magnetic moment of the charmed baryons . the energy and the angular dependences of @xmath40 and @xmath41 , for the process @xmath82 are shown in figs . [
fig : fig1 ] and [ fig : fig2 ] .
the asymmetry @xmath41 is large , @xmath83 , in all the considered kinematical region with small sensitivity to the @xmath84-magnetic moment .
the asymmetry @xmath40 is negative in most part of the considered kinematical region and smaller in absolute value .
note also that the asymmetry @xmath40 is more sensitive to the @xmath84-magnetic moment ( see fig . [
fig : fig3 ] ) .
the vector meson @xmath4-photoproduction can be considered in a similar way . as in case of photoproduction of pseudoscalar mesons ,
the asymmetry for polarized @xmath26- collisions takes the maximal value near threshold .
let us write the most general parameterization of the spin structure of the amplitudes for @xmath85 , with @xmath28-wave production of final particles : @xmath86\chi_1 , \label{mdst}\end{aligned}\ ] ] where @xmath87 is the three - vector of @xmath88-polarization , @xmath89 and @xmath90 are the multipole amplitudes corresponding to the absorption of electric dipole photons with production of final particles in states with spin @xmath91 and parity @xmath92 , @xmath93 and @xmath94 respectively , and @xmath95 is the multipole amplitude for absorption of magnetic quadrupole photon , so @xmath96 . only the asymmetry @xmath41 does not vanish , at the reaction threshold , and it can be written as a function of the threshold amplitudes @xmath89 , @xmath90 and @xmath95 : @xmath97 for pure multipole transitions , with @xmath89 , @xmath90 or @xmath95 amplitude
, we have , respectively @xmath98 , -1/2 and 1/2 . a model for @xmath4-photoproduction , based on the pseudoscalar @xmath16-exchange , has been considered in @xcite .
such mechanism gives @xmath99 and @xmath41=0 , for any photon energy and any production angle , because no spin correlation between the polarizations of @xmath45 and @xmath61 can be induced by the exchange of a zero spin particle .
therefore the discussed asymmetries can be considered a powerful signature of the possible exchange mechanism .
note that these results do not depend on the exact values of the corresponding coupling constants and the form of the phenomenological hadronic form factors , which are typical ingredients of such calculations . on the basis of our previous consideration @xcite of possible baryon exchange contribution to the threshold multipole amplitudes @xmath89 , @xmath90 or @xmath95
, one can derive the following formula for the asymmetry @xmath100 : @xmath101\right \ } , \label{eq : asym1}\ ] ] with @xmath102 where @xmath103 and @xmath104 , @xmath105 , @xmath106 are the masses of nucleon , @xmath107-hyperon and @xmath4-meson , respectively .
the real parameter @xmath108 characterizes the ratio of the coupling constants : determines the relative role of @xmath16-exchange mechanism and baryon - exchange mechanisms for @xmath109 @xcite . ]
@xmath110 where @xmath111 is the vector coupling constant for the vertex @xmath112 and @xmath113 is the magnetic moment for the @xmath114 transition .
this moment determines the width of the radiative decay @xmath115 , with experimental branching ratio @xmath116 .
only the upper limit is experimentally known for the total width of the neutral @xmath117 : @xmath1182.1 mev @xcite .
the total width of the charged @xmath119-meson is determined from the recent measurements of the cleo collaboration @xcite : @xmath120 kev . using the known value for
the branching ratio of the radiative decay @xmath121 @xcite : @xmath122 one can find a value of the magnetic moment for the transition @xmath123 , which agrees with the predictions from different theoretical approaches .
therefore it seems reasonable to use these models for the calculation of the coupling constant @xmath124 .
but the main problem , in our case , is how to determine the coupling constants for the vertices @xmath125 and @xmath126 .
the @xmath127-symmetry , which connects these coupling constants with the corresponding coupling constants for the vertices @xmath128 and @xmath129 is essentially violated @xcite .
we showed earlier @xcite , that the parameter @xmath108 can be determined , in principle , in the reaction @xmath130 through the angular dependence of the pseudoscalar @xmath16-meson produced in the subsequent decay @xmath131 .
[ fig : fig4 ] shows the dependence of the asymmetry @xmath41 , eq .
( [ eq : asym1 ] ) , on the ratio @xmath108 .
one can see that only for @xmath132 the asymmetry @xmath133 is positive , being generally negative , @xmath134 , with very small sensitivity to the ratio @xmath108 of the fundamental coupling constants . from this brief discussion
it appears that the prediction of the asymmetry @xmath41 for charmed vector meson photoproduction is more complicated and more model dependent in comparison with processes involving pseudoscalar mesons , as this asymmetry depends on the value of @xmath108 .
we can suggest the study of the asymmetry in @xmath135 as a possible way to determine the fundamental coupling constants for the vertices involving charmed particles .
the model of photon - gluon fusion predicts a value of @xmath136 , for the @xmath41-asymmetry in the inclusive @xmath137 process , at @xmath138 50 gev , depending on the assumptions on the polarized gluon distribution , @xmath7 @xcite .
therefore , even a @xmath139 contribution of the exclusive process @xmath140 to the inclusive @xmath16-cross section can induce a @xmath139 correction to the @xmath141-fusion asymmetry , at forward angles .
it is a very large effect , which should be taken into account in the extraction of the @xmath7-effect from the asymmetry in the process @xmath142 .
evidently this effect does not contribute to the @xmath143 production in the process @xmath144 .
the observation of the @xmath145 or @xmath146 asymmetry in @xmath147-collisions will be important in order to test the validity of the photon - gluon fusion mechanism .
note that the exclusive photoproduction of open charm will result in @xmath145 and @xmath146 asymmetry in @xmath147-collisions ( with unpolarized particles ) .
such asymmetries have been experimentally observed .
for example , the focus experiment found a ratio for @xmath146 production @xmath148 , demonstrating that @xmath84-production is more probable than @xmath149 @xcite .
this is in contradiction with the model of @xmath141-fusion , which predicts symmetric @xmath150-yields and can be considered as an indication of the presence of other mechanisms .
the exclusive processes , discussed in this paper , explain naturally such asymmetry .
similar charm asymmetries have been observed also at @xmath65=20 gev , at slac @xcite .
we calculate the t- and p - even asymmetries for the processes @xmath0 and @xmath1 , induced by circularly polarized photons on a polarized proton target . using an effective lagrangian approach
, we found that these asymmetries are large in absolute value for all considered photon energies .
for the pseudoscalar meson production the @xmath41 asymmetry is essentially positive , large in magnitude and of the same sign of the predictions based on the photon - gluon fusion mechanism .
this means that exclusive processes have to be viewed as a non negligeable contribution to charm photoproduction at high energies and they have to be carefully treated in order to extract the polarization properties of the gluon content of the nucleon from the future compass data . for vector charmed meson photoproduction
the predictions for the discussed asymmetries are more model dependent .
vanishing for the @xmath51meson exchange mechanism , the threshold asymmetry @xmath41 shows large sensitivity to the ratio of the coupling constants for the different vertices involving charmed particles .
d. l. adams _ et al . _
[ fnal e581/704 collaboration ] , phys .
b * 336 * , 269 ( 1994 ) .
g. bunce _
et al . _ ,
world * 3 * , 1 ( 1992 ) .
g. baum _ et al . _ [ compass collaboration ] , cern - spslc-96 - 14 .
f. halzen and d. m. scott , phys .
b * 72 * , 404 ( 1978 ) ; + h. fritzsch and k. h. streng , phys .
b * 72 * , 385 ( 1978 ) ; + v. a. novikov , m. a. shifman , a. i. vainshtein and v. i. zakharov , nucl .
b * 136 * , 125 ( 1978 ) [ yad .
* 27 * , 771 ( 1978 ) ] ; + j. babcock , d. w. sivers and s. wolfram , phys .
d * 18 * , 162 ( 1978 ) .
m. p. rekalo and e. tomasi - gustafsson , phys .
b * 500 * , 53 ( 2001 ) .
m. p. rekalo and e. tomasi - gustafsson , phys .
d * 65 * , 014010 ( 2002 ) .
m. p. rekalo and e. tomasi - gustafsson , phys . rev .
d * 65 * , 074023 ( 2002 ) .
z. w. lin and c. m. ko , phys .
034903 ( 2000 ) and refs herein . t. matsui and h. satz , phys .
b * 178 * , 416 ( 1986 ) .
k. wijesooriya _ et al .
_ [ jefferson lab hall a collaboration ] , phys .
* 86 * ( 2001 ) 2975
. g. f. chew , m. l. goldberger , f. e. low and y. nambu , phys .
106 * , 1345 ( 1957 ) .
d. e. groom _ et al .
_ , eur . phys .
j. * c 15 * , 1 ( 2000 ) .
a. anastassov _ et al . _
[ cleo collaboration ] , hep - ex/0108043 .
j. bartelt _ et al . _
[ cleo collaboration ] , phys .
lett . * 80 * , 3919 ( 1998 ) .
m. j. savage , phys
b 326 , 303 ( 1994 ) .
s. bianco , arxiv : hep - ex/9911034 . | we calculate the asymmetries for the processes @xmath0 and @xmath1 , induced by circularly polarized photons on a polarized proton target in the framework of an effective lagrangian model .
we found large and positive values for the asymmetry induced by the @xmath2component of the proton polarization , in a wide energy region , from threshold up to @xmath3 gev . in the case of pseudoscalar meson photoproduction ,
the predictions are qualitatively model - independent . for @xmath4-photoproduction , there is a strong dependence of the asymmetries on @xmath5-coupling constants .
# 1@xmath6#1 |
the casimir force , arising from quantum fluctuations of the electromagnetic field in vacuum , is a striking manifestation of quantum field theory at the mesoscopic scale .
casimir s computation of the force between two parallel metallic plates @xcite gives the classic demonstration of this phenomenon .
following its experimental confirmation in the past decade @xcite , however , the casimir force is now important to the design of microelectromechanical systems @xcite .
potential practical applications have motivated the development of large - scale numerical methods to compute casimir forces for objects of any shape @xcite .
in contrast , the simplest and most commonly used analytic methods for dealing with complex shapes , such as the proximity force approximation ( pfa ) , rely on pairwise summations , limiting their applicability .
recently we have developed a formalism @xcite that relates the casimir interaction among several objects to the scattering of the electromagnetic field from the objects individually .
this method decomposes the path integral representation of the casimir energy @xcite as a log - determinant @xcite in terms of a multiple scattering expansion , as was done for asymptotic separations in ref .
it can also be regarded as a concrete implementation of the perspective emphasized by schwinger @xcite that the fluctuations of the electromagnetic field can be traced back to charge and current fluctuations on the objects .
( for additional perspectives on the scattering formalism , see also references in @xcite . )
this approach allows us to take advantage of the well - developed machinery of scattering theory .
in particular , the availability of exact scattering amplitudes for simple objects , such as spheres and cylinders , has made it possible to compute the casimir force for two spheres @xcite , a sphere and a plate @xcite , multiple cylinders @xcite , and cases with more than two objects @xcite .
this formalism has also been applied and extended in a number of other situations @xcite . here
we expand on recent work @xcite that showed how to apply these techniques to _ parabolic _ cylinders , another example where the scattering amplitudes can be computed exactly .
the limiting case when the radius of curvature at the tip vanishes , so that the parabolic cylinder becomes a semi - infinite plate ( a `` knife - edge '' ) , provides a particularly interesting application of this approach .
one can also model the knife - edge as the limit of a wedge of zero opening angle @xcite ; the two approaches are rather complementary as the former is most amenable to _ numerical _ computation , while the latter yields approximate _ analytic _ formulae via a multiple reflection expansion ( which is useful for other sharp geometries , such as the cone @xcite ) .
edge geometries have also been considered in refs .
@xcite .
the remainder of the manuscript is organized as follows : the helmholtz equation in the parabolic cylinder coordinate system is reviewed in sec .
[ scatter ] , and exact formulae are derived for the scattering of the electromagnetic field from a perfectly conducting parabolic cylinder . the techniques of refs .
@xcite , are then employed to find the electromagnetic casimir interaction energy for a variety of situations : we let the parabolic cylinder interact with a plane ( sec .
[ pcyl - plane ] ) , a second parabolic cylinder ( sec .
[ 2pcyls ] ) , or an ordinary cylinder ( sec .
[ pcyl - cyl ] ) . in these calculations
we consider arbitrary rotations of the parabolic cylinder around its focal axis and arbitrary translations perpendicular to that axis , transformations that are particularly useful when considering the `` knife - edge '' limit .
we also position an ordinary or parabolic cylinder inside a parabolic cylinder ( sec . [ interior ] ) , and incorporate the effects of thermal corrections in all of these calculations ( sec .
[ finitet ] ) .
we begin with a review of scattering theory in parabolic cylinder coordinates @xcite . because the system is translationally invariant in the @xmath0 direction and perfectly reflecting , we can decompose the electromagnetic scattering problem into two scalar problems , one with dirichlet and the other with neumann boundary conditions .
parabolic cylinder coordinates are defined by @xmath1 such that for @xmath2 , @xmath3 where we have chosen a convention where @xmath4 but @xmath5 can have either sign , which is explained in more detail below .
this restricted domain is sufficient to include all points in space .
note that here we must take @xmath6 , not @xmath7 . at fixed @xmath0 , the surfaces of constant @xmath8 are confocal parabolas opening upward ( toward positive @xmath9 ) and the surfaces of constant @xmath5 are confocal parabolas opening downward ( toward negative @xmath9 ) , as shown in fig .
[ fig : coords ] .
the scale factors ( metric coefficients ) are @xmath10 , @xmath11 .
we would like to solve the helmholtz equation , which in these coordinates takes the form @xmath12 where eventually we will set @xmath13 . for @xmath14 real , we expect oscillating traveling wave solutions , while for @xmath15 real , we expect exponentially growing and decaying solutions .
this equation is amenable to separation of variables : @xmath16 separation of the @xmath0 variable is trivial , @xmath17 , leaving @xmath18 which gives the separated equations @xmath19 where @xmath20 is a separation constant .
the solutions are @xmath21 in terms of the parabolic cylinder function @xmath22 . here
@xmath23 and @xmath24 the second solution in each case is obtained by making the replacements @xmath25 , @xmath26 , and @xmath27 , the combination of which leaves the differential equations invariant . for @xmath28 , @xmath29 as @xmath30 .
so for both @xmath5 and @xmath8 , we have solutions that both grow and decay exponentially as their argument approaches positive infinity .
since the cartesian radial distance is @xmath31 , these are ordinary exponentials ( not gaussians ) when expressed in terms of cartesian coordinates . if we send @xmath5 to @xmath32 and @xmath33 , we return to the same point in space , so a wavefunction defined everywhere in the plane must be invariant under this transformation . from the identity
@xmath34 we can conclude that @xmath35 must be a non - negative integer to satisfy this requirement .
we begin with the case @xmath36 for @xmath37 . for these values of @xmath35 ,
the parabolic cylinder functions with real arguments are simple rescalings of the solutions to the quantum harmonic oscillator , and thus are given by a gaussian times a hermite polynomial . the combined solution in eq .
( [ eqn : regular ] ) then takes the form of a polynomial in @xmath8 and @xmath5 times @xmath38 , and thus represents a traveling parabolic wave in the @xmath39 direction .
these solutions represent `` regular '' waves , the analogs of solutions in spherical coordinates involving spherical bessel functions and spherical harmonics , @xmath40 .
since we have chosen to restrict @xmath8 to positive values , it will represent the analog of the radial coordinate @xmath41 .
we will also require `` outgoing '' solutions to the same differential equations , the analogs of solutions in spherical coordinates involving spherical hankel functions and spherical harmonics , @xmath42 .
as in the spherical case , in the irregular solution the function of the `` angular '' variable @xmath5 is the same , but the function of the `` radial '' variable is an independent solution to the same differential equation , @xmath43 again for @xmath37 . even though they do not blow up at @xmath44 ( as the outgoing spherical wavefunctions do at @xmath45 ) , these solutions are not permissible for @xmath46 because they are not invariant under the combined substitution @xmath33 and @xmath47 .
the solution in eq .
( [ eqn : irregular ] ) asymptotically approaches a polynomial in @xmath8 and @xmath5 times @xmath48 , and thus represents an outgoing radial parabolic wave .
we define the full regular and outgoing solutions @xmath49 using which the free green s function becomes , for @xmath50 @xcite , is incorrectly omitted in ref .
@xmath51 here @xmath52 ( @xmath53 ) is the point with the smaller ( larger ) of @xmath54 and @xmath55 , and we have made use of the wronskian of the two independent solutions for each @xmath35 , @xmath56 = i^{\nu-1 } \,.\ ] ] the decomposition of a plane wave in regular parabolic cylinder functions is @xcite @xmath57 where @xmath58 and @xmath59 . here the logarithm defines the arctangent of @xmath60 in the appropriate quadrant .
note that the expansion in eq .
( [ eqn : plane1 ] ) converges only for @xmath61 , since it is built out of parabolic waves that propagate upward . to determine the t - matrix we consider dirichlet or neumann boundary conditions at @xmath62 . in the region
@xmath63 we have the scattering solution @xmath64 for dirichlet boundary conditions and @xmath65 for neumann boundary conditions , where prime denotes the derivative of the parabolic cylinder function with respect to its argument and @xmath37 .
these wavefunctions correspond to the scattering @xmath66-matrix elements @xmath67 , with @xmath68 for the process where an incoming parabolic wave propagating in the @xmath39 direction is scattered into an outgoing parabolic wave propagating radially .
the solutions we have obtained allowed us to construct the complete free green s function , the decomposition of a plane wave , and the scattering @xmath66-matrices , which contain all the information we will need to carry out our calculations .
however , we note that there also exists a second set of solutions , representing the time - reversed scattering process , @xmath69 for @xmath70 , which also are unchanged for @xmath71 and @xmath72 . for real @xmath14 ,
these solutions go like @xmath73 and so propagate in the @xmath74 direction , in contrast to the solutions in eq .
( [ eqn : regular ] ) , which go like @xmath75 and propagate in the @xmath39 direction .
we also have the the corresponding irregular solutions , @xmath76 where again @xmath70 .
these solutions go like @xmath77 and thus correspond to incoming radial parabolic waves . since a decomposition of the green s function like eq .
( [ eqn : green1 ] ) typically consists of a sum over all scattering solutions , one might wonder why this second set of solutions does not appear there .
one can formally extend the sum to include all values of @xmath35 , but for @xmath78 the @xmath79 in the denominator becomes a divergent gamma function , so that these terms all gives zero contribution .
however , we can also write the green s function solely in terms of these solutions as @xmath80 and the decomposition of the plane wave as @xmath81 which now converges only for @xmath82 , since it consists only of waves propagating downward . using these solutions , we could construct the analogous scattering solutions for neumann and dirichlet boundaries , which represent the process where an incoming radial parabolic wave is scattered into an outgoing parabolic wave propagating in the @xmath74 direction .
to calculate the casimir force for a parabolic cylinder opposite a plane , we will need an appropriate expression for the free green s function in terms of plane waves , and expansions translating between these two bases . for @xmath83
, the free green s function can be written in cartesian coordinates as @xmath84 where @xmath59 .
we equate this expression to the green s function in eq .
( [ eqn : green1 ] ) , expand the plane wave @xmath85 in eq .
( [ eqn : greenplane ] ) using eq .
( [ eqn : plane1 ] ) , make the substitution @xmath86 , and then use the orthogonality of the regular parabolic solutions to equate both sides term by term in the sum over @xmath35 .
the result is an expansion for the irregular parabolic solutions in terms of plane waves : @xmath87 e^{-i k_y y + i k_x x } e^{ik_z z } \ , , \label{eqn : expandout}\ ] ] which is valid for @xmath88 and @xmath70 .
we have not found this result in the previous literature , though it is hinted at in @xcite .
the quantity in brackets then defines the conversion matrix between outgoing parabolic cylinder functions and plane waves propagating in the @xmath74 direction .
it allows us to propagate the outgoing waves from the parabolic cylinder downward to the plane .
we displace the origin of the cartesian coordinates for the plane from the origin of the parabolic cylinder coordinates by a distance @xmath89 in the @xmath9 direction , which simply introduces a factor of @xmath90 . due to invariance along the time and @xmath0 directions
, we can make independent computations for each @xmath15 and @xmath91 , and then integrate over both quantities in the final result for the casimir energy . in the scattering theory approach , the calculation can be formulated in terms of scattering amplitudes by considering fluctuating multipoles @xcite , or equivalently by using a generalized @xmath66-operator formalism @xcite . in the latter approach , which we adopt here , the ingredients we will need are the @xmath66-matrix elements , the expansion of the outgoing wave in terms of plane waves , and the normalization factors appearing in the green s functions in eqs .
( [ eqn : green1 ] ) and ( [ eqn : greenplane ] ) @xcite .
the @xmath66-matrix elements for the parabolic cylinder are given in eq .
( [ eqn : tmat ] ) , and the @xmath66-matrix elements for the plane are simply @xmath92 for neumann and dirichlet boundary conditions respectively .
finally , we must include the appropriate normalization factor @xcite @xmath93 , where we can read off @xmath94 and @xmath95 from the expressions for the free green s function in eqs .
( [ eqn : green1 ] ) and ( [ eqn : greenplane ] ) . we can then write the energy per unit length as @xmath96 where the matrix determinant runs over @xmath97 . here
we have defined the translation matrix @xmath98 and , for convenience in later expressions in which we consider different orientations of the parabolic cylinder , we have written the reverse translation matrix as @xmath99 , with @xmath100 .
the complete energy per unit length is then @xmath101 where we have dropped a factor of @xmath102 since it does not change the determinant .
we sum this result over dirichlet and neumann boundary conditions to obtain the full electromagnetic result .
this can be compared to the proximity force approximation , @xmath103 which is the sum of equal contributions from the dirichlet and neumann cases .
we can make the following simplifications in eq .
( [ eqn : cpenergy ] ) : * the integral over @xmath104 is zero if @xmath105 is odd , it is symmetric in @xmath35 , @xmath106 , and the integrand is even in @xmath104 .
* we can replace @xmath107 by @xmath108 , where @xmath109 .
* we can further simplify the integral @xmath110 which appears in eq .
( [ eqn : cpenergy ] ) with @xmath111 .
here @xmath112 is always an integer , since the translation matrix element vanishes if @xmath113 is odd . setting @xmath114 , we have @xmath115 which is given in terms of the confluent hypergeometric function of the second kind @xmath116 as @xmath117 where @xmath118 is the bateman @xmath119-function @xcite .
we now review some results that were reported previously using this formalism @xcite . to connect back to the physical configuration ,
it is convenient to represent the final casimir energy in terms of the radius of curvature at the tip @xmath120 and the separation @xmath121 . at small separations ( @xmath122 )
the proximity force approximation , given by @xmath123 ^ 3 } = -\frac{\pi^3}{960\sqrt{2 } } \sqrt{\frac{r}{h^5}}\ , , \label{eq : pfa}\ ] ] should be valid .
the numerical results in fig .
[ fig : evsh ] confirm this expectation with a ratio of actual to pfa energy of @xmath124 at @xmath125 ( with @xmath126 ) .
we note that since the main contribution to pfa is from the proximal parts of the two surfaces , the pfa result in eq .
( [ eq : pfa ] ) also applies to a circular cylinder with the same radius @xmath127 . in the opposite limit , @xmath128 , the parabolic cylinder becomes a half - plane , and we can express the @xmath66-matrix in closed form as well : @xmath129 the nonzero elements of which can be summarized compactly as @xmath130 where even @xmath35 corresponds to dirichlet boundary conditions and odd @xmath35 corresponds to neumann boundary conditions .
we can thus write the full electromagnetic energy for the half - plane perpendicular to the plane as @xmath131 where the bateman @xmath119-function is nonzero only for @xmath105 even , and we have dropped factors that cancel in the determinant .
the factor of @xmath132 in this expression arises from the @xmath66-matrix element for the plane .
numerically , we find @xmath133 , which is shown by a dashed line in fig . [
fig : evsh ] .
this geometry was studied using the world - line method for a scalar field with _ dirichlet _ boundary conditions in ref .
( the world - line approach requires a large - scale numerical computation , and it is not known how to extend this method beyond the case of a scalar with dirichlet boundary conditions ) . in our calculation , the dirichlet component of the electromagnetic field makes a contribution @xmath134 to our result , in reasonable agreement with the value of @xmath135 in ref .
these results are also in agreement with the calculation in ref .
@xcite . ,
@xmath136 , plotted versus @xmath137 for @xmath138 and @xmath126 on a log - linear scale .
the dashed line gives the @xmath128 limit and the solid curve gives the pfa result . ] it is straightforward to extend this calculation to the case where the parabolic cylinder ( of any radius ) is rotated by an angle @xmath139 around its focal axis , as shown in fig .
[ fig : tilt ] . in place of eq .
( [ eqn : cptrans ] ) , we have @xmath140 however , now the integral over @xmath104 is not symmetric , and the matrix elements with @xmath113 odd need not vanish . , which is oriented perpendicular to the plane for @xmath138 and parallel to the plane for @xmath141 .
the left panel shows the coefficient @xmath142 ( see text ) as a function of @xmath139 , with the exact parallel plate result at @xmath141 marked with a cross .
the inset shows the dirichlet ( circles ) and neumann ( squares ) contributions to the full electromagnetic result .
the right panel again shows @xmath142 , but now in comparison to the proximity force approximation ( solid line ) .
note the large discrepancy between the pfa and the exact results as @xmath143 .
, title="fig : " ] , which is oriented perpendicular to the plane for @xmath138 and parallel to the plane for @xmath141 .
the left panel shows the coefficient @xmath142 ( see text ) as a function of @xmath139 , with the exact parallel plate result at @xmath141 marked with a cross .
the inset shows the dirichlet ( circles ) and neumann ( squares ) contributions to the full electromagnetic result .
the right panel again shows @xmath142 , but now in comparison to the proximity force approximation ( solid line ) .
note the large discrepancy between the pfa and the exact results as @xmath143 .
, title="fig : " ] we again consider the @xmath144 limit in analyzing this result . from dimensional analysis ,
the electromagnetic casimir energy at @xmath128 takes the now @xmath139-dependent form @xmath145 where @xmath146 for @xmath128 .
following ref .
@xcite , which considers the casimir energy for a scalar field with dirichlet boundary conditions in this geometry , we plot @xmath147 in fig .
[ fig : camp ] .
a particularly interesting limit is @xmath148 , as the two plates become parallel . in this case
, the leading contribution to the casimir energy should be proportional to the area of the half - plane according to the parallel plate formula , @xmath149 with @xmath150 , plus a subleading correction due to the edge .
multiplying by @xmath151 has removed the divergence in @xmath152 as @xmath148 . as in ref .
@xcite , we assume @xmath153 , although we can not rule out the possibility of additional non - analytic forms , such as logarithmic or other singularities . with this assumption , we can estimate the edge correction @xmath154 from the data in fig .
[ fig : camp ] .
from the inset in fig .
[ fig : camp ] , we estimate the dirichlet and neumann contributions to this result to be @xmath155 ( in agreement with @xcite within our error estimates ) and @xmath156 respectively . because higher partial waves become more important as @xmath157 , reflecting the divergence in @xmath152 in this limit , we have used larger values of @xmath158 for @xmath139 near @xmath159 . in fig .
[ fig : camp ] we also show a comparison to the proximity force approximation .
the pfa is clearly of no use at @xmath138 , since it simply gives zero , while at @xmath141 the pfa gives the correct energy but incorrectly has zero slope , since it misses the edge correction .
we have found that the edge correction is small in the electromagnetic case , as a result of the near - cancellation between the dirichlet and neumann contributions . by using babinet s principle ,
it is possible to show that this suppression of edge effects is a general feature of any thin conductor , arising because the leading term in the multiple reflection expansion is identically zero @xcite .
we next consider the force between two parabolic cylinders opening in opposite directions , as shown in the left panel of fig .
[ fig : parageomout ] .
we will consider the generalization to arbitrary orientation below .
we need to express the outgoing waves from one parabolic cylinder in terms of the regular waves for the other .
we let @xmath160 represent the coordinates of the second parabolic cylinder , @xmath161 , @xmath162 , and @xmath163 . using eq .
( [ eqn : expandout ] ) for @xmath160 , we have @xmath164 now we use the expansion of the plane wave , eq .
( [ eqn : plane1 ] ) , to obtain @xmath165 \psi_{\nu'}^{\hbox{\tiny reg}}(\bm{r } ) \,,\end{aligned}\ ] ] where @xmath89 is the interfocal separation .
we can then obtain the translation coefficient from the quantity in brackets .
we thus obtain the casimir interaction energy @xmath166 where @xmath167 and @xmath168 are the scattering @xmath66-matrix elements for the two parabolic cylinders ( which can have different radii ) , the translation matrix elements are given by @xmath169 and the determinant runs over @xmath170 . here
again we have defined @xmath171 , where @xmath172 is the reverse translation matrix .
we sum the results for dirichlet and neumann boundary conditions to obtain the result for electromagnetism .
the analogous numerical simplifications apply here as in the case of the plane , and we can use eq .
( [ eqn : bateman ] ) , now with @xmath89 instead of @xmath173 , to express the translation matrix elements in eq .
( [ eqn : cctrans ] ) in terms of the bateman @xmath119-function .
the extension to the tilted case is also analogous ; now the angle of rotation can be different for the two translation matrices , corresponding to different angles of rotation for the two parabolic cylinders .
we can also introduce a translation in the @xmath174-direction @xmath175 , in addition to the existing translation @xmath89 in the @xmath9-direction . for rotations @xmath139 and @xmath176 of the two parabolic cylinders and @xmath174-translation @xmath175 , we have @xmath177 where we must have @xmath178 , but @xmath175 can have either sign , representing a translation in either horizontal direction .
by considering two parabolic cylinders of zero radius , we can study the casimir interactions of two half - planes , as illustrated in fig .
[ fig : planegeom ] . these techniques ,
together with a multiple reflection expansion , were used in ref .
@xcite to obtain a variety of results in half - plane geometries . and @xmath179 , and displaced by @xmath175 and @xmath89 .
] we take @xmath180 , so that we are considering parallel half - planes , where positive @xmath175 gives the width of the region over which they overlap , while negative @xmath175 gives a horizontal displacement of the edges away from each other . in this case , eq .
( [ eqn : trans2par ] ) simplifies to @xmath181 results are shown in fig .
[ fig : overlap ] , along with approximations valid in two limiting cases : first , for @xmath175 very negative , we can ignore the vertical displacement .
the configuration is then equivalent to the case of @xmath182 , which gives @xmath183 at separation @xmath89 , which is shown as a dashed line in fig .
[ fig : overlap ] .
second , for @xmath175 large and positive , we can take the standard result for parallel plates @xmath184 plus twice the edge correction @xmath185 found above for a half - plane parallel to a plane , which is shown as a solid line in fig .
[ fig : overlap ] .
as these examples illustrate , our description of the two half - planes is redundant : different parameter choices lead to the same physical configuration , a property we have used to check our calculations .
the numerical convergence of physically equivalent configurations can be quite different , however .
for example , in the case of @xmath186 , when both @xmath175 and @xmath89 increase , the casimir interaction energy decreases , since the half - planes are becoming further apart . in the scattering bases we have chosen , however , in @xmath89
this effect appears directly through a decaying exponential , while in @xmath175 it appears through the cancellation of an oscillating integrand . as a result
, we need to maintain @xmath178 , but can consider either sign of @xmath175 . .
solid points are obtained from the exact calculation described in the text .
the solid line connecting them is a rational function fit to guide the eye .
the dashed line gives the energy for the limit where the planes are edge - to - edge , while the solid straight line gives the standard parallel plate result for the overlap area , plus edge corrections . ]
we next consider the case of an ordinary cylinder outside a parabolic cylinder , as shown in the right panel of fig .
[ fig : parageomout ] . in ordinary cylindrical coordinates
, we have regular solutions given in terms of bessel functions , @xmath187 , and outgoing solutions given in terms of hankel functions of the first kind , @xmath188 , both indexed by angular momentum @xmath189 .
we use the expansion of a plane wave in regular ordinary cylindrical wavefunctions , @xmath190 where @xmath191 is defined as in eq .
( [ eq : phi ] ) and @xmath192 and @xmath41 are the ordinary cylindrical coordinates for @xmath193 . in these coordinates ,
the free green s function is given by @xmath194 where @xmath195 ( @xmath196 ) is the smaller ( larger ) of @xmath197 and @xmath198 .
the @xmath66-matrix elements for an ordinary cylinder of radius @xmath199 are given in terms of bessel and hankel functions and their modified counterparts by @xmath200 , with @xmath201 where prime indicates a derivative with respect to the function s argument .
for an ordinary cylinder outside a parabolic cylinder with separation @xmath89 between the center of the ordinary cylinder and the focus of the parabolic cylinder , we substitute eq .
( [ eq : cylplane ] ) into eq .
( [ eq : outreg ] ) to obtain @xmath202 e^{ik_z z } e^{i\ell \theta } j_\ell(\sqrt{k^2-k_z^2 } r ) \,,\end{aligned}\ ] ] where again the quantity in brackets is the coefficient we need to compute the translation matrix .
we thus obtain the casimir interaction energy @xmath203 where @xmath204 and the determinant again runs over @xmath170 .
as before , for convenience we have defined @xmath205 , where @xmath206 is the reverse translation matrix .
we can again simplify this expression for numerical computation by combining the @xmath15 and @xmath91 integrals , and by exploiting symmetries in @xmath207 and @xmath86 .
generalizations to include horizontal translation and tilt of the parabolic cylinder also work in the same way as before , and taking the limit in which the radius of the parabolic cylinder goes to zero gives the casimir energy for a cylindrical wire opposite a half - plane .
up to now , we have considered `` exterior '' geometries in which the objects are outside one another . however , simple modifications of these techniques enable us to also consider `` interior '' geometries using the formalism of ref . @xcite .
( interior geometries were also considered in @xcite , using large - scale computation . )
there are two changes required for this case : the @xmath66-matrix of the outside object must be inverted , and we require the translation matrix connecting the regular solutions for the different objects , rather than the one connecting outgoing solutions for one object to regular solutions for the other .
we first consider two parabolic cylinders inside one another .
we parameterize the displacement between their foci in parabolic cylinder coordinates by @xmath208 and @xmath209 , as shown in the left panel of fig .
[ fig : parageom ] . following ref .
@xcite , we can derive the translation matrix for regular solutions appropriate to the inside problem .
let @xmath210 and consider the equation @xmath211 .
using eq .
( [ eqn : plane1 ] ) , we have @xmath212 now we let @xmath213 , so @xmath214 , and consider the case where @xmath215 to obtain @xmath216 next we write @xmath217 substituting this result into eq .
( [ eqn : addition ] ) and equating powers of @xmath218 results in @xmath219 \psi_{\nu}^{\hbox{\tiny reg } } ( \lambda , \mu ) \,,\ ] ] which yields the coefficient we need from the quantity in brackets .
the casimir interaction energy is then @xmath220 where @xmath221 ( @xmath222 ) is the @xmath66-matrix for the outer ( inner ) parabolic cylinder , @xmath223 and we have dropped normalization factors that cancel in the determinant . for an ordinary cylinder inside a parabolic cylinder , as shown in the right panel of fig .
[ fig : parageom ] , we again let @xmath210 and consider the equation @xmath211 , but now we expand the left - hand side in parabolic cylinder coordinates and the right - hand side in ordinary cylindrical coordinates , to obtain @xmath224 as before , setting @xmath225 , so that @xmath226 and @xmath227 , yields @xmath228 we take @xmath35 derivatives with respect to @xmath218 and then set @xmath229 to obtain @xmath230 e^{i k_z z } e^{i\ell \theta } j_{\ell}(\sqrt{k^2-k_z^2 } r ) \,,\ ] ] where again the quantity in brackets will give the coefficient we need . using the generalized binomial expansion , we obtain @xmath231 where the @xmath232 is @xmath233 for @xmath234 and @xmath235 for @xmath236 , and the sum starts at @xmath237 for @xmath35 even and @xmath238 for @xmath35 odd , and then in both cases goes in steps of @xmath239 . for a configuration where the displacement from the focus of the parabolic cylinder to the center of the ordinary cylinder is parameterized in ordinary cylindrical coordinates by distance @xmath240 and angle @xmath241 from the @xmath174-axis ,
the casimir interaction energy is then @xmath242 where @xmath221 is the @xmath66-matrix for the ( outer ) parabolic cylinder , @xmath243 is the @xmath66-matrix for the ( inner ) ordinary cylinder , and @xmath244 \,.\ ] ] as an example , we consider a thin wire near the focal axis of a parabolic cylinder with parabolic radius @xmath245 . to leading order in the needle radius @xmath199
, we only need the @xmath246 dirichlet @xmath66-matrix element , which goes like @xmath247 for @xmath199 small . keeping only the leading term in @xmath248 and also expanding in the displacement from the focus @xmath240
, we obtain @xmath249 here energy is calculated in comparison to the configuration where the ordinary cylinder is placed at @xmath250 , @xmath251 . in deriving this result ,
we have assumed that @xmath252 is small enough that we can drop terms proportional to @xmath253 in comparison to terms proportional to @xmath254 , since for @xmath255 the integrand in eq .
( [ eqn : eparordint ] ) is exponentially suppressed .
the first term gives the ( negative ) casimir interaction energy per unit length when the wire is at the focus , while the second and third terms give the correction as it is moved a small distance away .
the angular dependence is exactly as we would expect : as the wire moves closer to the vertex axis of the parabolic cylinder ( @xmath256 ) , the energy gets more negative ; as it moves away from the vertex axis ( @xmath257 ) the energy is less negative , and if it moves in a direction perpendicular to the plane of symmetry of the parabolic cylinder ( @xmath258 or @xmath259 ) , the energy is unchanged to first order . as a result , unlike the geometric optics calculation considered in ref .
@xcite , here we do not see any unusual behavior of the casimir energy at the focus , which is in agreement with the results in ref .
@xcite . in fig .
[ fig : cylinterior ] we illustrate this result for the case where the radius of the ordinary cylinder and its displacement are not small .
we choose the same radii for the parabolic and ordinary cylinders as in ref .
@xcite , and the results we obtain are approximately in agreement with what was found there .
we can not make a precise comparison , however , because in that work the parabolic cylinder is of finite size and closed at the far end . and
@xmath260 , so that the vertex line of the parabolic cylinder lies at @xmath250 , @xmath261 .
the right panel shows the casimir interaction energy per unit length @xmath262 as a function of @xmath174 and @xmath9 , the displacement of the center of the ordinary cylinder from the focus of the parabolic cylinder , where the center of the ordinary cylinder lies within the shaded region of the left panel.,title="fig : " ] and @xmath260 , so that the vertex line of the parabolic cylinder lies at @xmath250 , @xmath261 .
the right panel shows the casimir interaction energy per unit length @xmath262 as a function of @xmath174 and @xmath9 , the displacement of the center of the ordinary cylinder from the focus of the parabolic cylinder , where the center of the ordinary cylinder lies within the shaded region of the left panel.,title="fig : " ]
it is straightforward to extend all of these results to temperature @xmath263 , a subject that has been of significant recent interest @xcite . in each calculation
, we simply replace the integral @xmath264 by the sum @xmath265 over matsubara frequencies @xmath266 , where @xmath267 is boltzmann s constant and the prime indicates that the @xmath237 mode is counted with a weight of @xmath268 @xcite . in the classical limit of asymptotically large temperature , only the @xmath237 term contributes .
the numerical calculation is more cumbersome for @xmath263 , because for @xmath269 we could always make the substitution @xmath270 , where @xmath271 , since the quantity we integrate depends only on @xmath20 .
we find it convenient to continue to use the integration variable @xmath20 , since the quantity we now sum and integrate still depends only on this quantity .
we therefore carry out the sum and integral via the replacement @xmath272\,,\ ] ] where @xmath273 denotes the greatest integer less than or equal to @xmath174 . as an example , we consider thermal corrections for a conducting half - plane ( a parabolic cylinder with @xmath128 ) oriented perpendicular to a conducting plane , at separation @xmath274 . in the classical limit , only the @xmath237 mode contributes and we obtain the energy @xmath275 , with @xmath276 .
the dirichlet contribution to this result is @xmath277 , in agreement with ref . @xcite . in fig .
[ fig : evst ] the energy for this geometry is shown as a function of temperature .
for typical separation distances at room temperature , the thermal corrections are small . , @xmath136 , plotted versus @xmath278 for @xmath138 and @xmath128 .
the solid line gives the @xmath279 limit determined from the lowest matsubara frequency .
for reference , at a separation of @xmath280 , a temperature @xmath281 corresponds to @xmath282 . ]
there are only a limited set of coordinate systems in which the vector helmholtz equation for electromagnetism can be solved exactly .
taking advantage of one of these few cases , we have obtained complete scattering amplitudes for a perfectly conducting parabolic cylinder and employed these results to compute casimir forces . in principle , casimir forces can be computed in configurations involving parabolic cylinders and other shapes for which scattering amplitudes are known , as long as we can obtain the _ translation matrices _ , which convert expressions of electromagnetic waves between different coordinate basis , appropriate to the individual shapes . following this procedure ,
we have computed casimir forces between a parabolic cylinder , a plane , an ordinary cylinder , and a second parabolic cylinder .
the formalism is versatile enough to treat situations in which one object is enclosed in the interior of a parabolic cylinder , and is also easily extended to finite temperatures .
we focus special attention to the limit when the radius of the parabolic cylinder goes to zero , and it evolves into a semi - infinite plate a _ knife edge_. in this limit we can quantify the contribution of edges to the casimir force . by examining tilted plates , we can consider a broad range of cases involving interacting edges which should be useful to the design of microelectromechanical devices . until
recently , the state of art computation of casimir forces relied upon the pfa , which is demonstrably unreliable for a knife edge : a thin metal disk perpendicular to a nearby metal surface experiences a casimir force described by an extension of eq .
( [ eq : zeror ] ) , while as indicated in fig .
[ fig : evsh ] , the pfa approximation to the energy vanishes as the thickness goes to zero .
based on the full result for perpendicular planes , however , we can formulate an `` edge pfa , '' which yields the energy by integrating @xmath283 from eq .
( [ eq : zeror ] ) along the edge of the disk .
letting @xmath41 be the disk radius , in this approximation we obtain @xmath284 which is valid if the thickness of the disk is small compared to its separation from the plane .
( for comparison , note that the ordinary pfa for a metal sphere of radius @xmath41 and a plate is proportional to @xmath285 . )
a disk may be more experimentally tractable than a plane , since its edge does not need to be maintained parallel to the plate .
one possibility is a metal film , evaporated onto a substrate that either has low permittivity or can be etched away beneath the edge of the deposited film .
micromechanical torsion oscillators , which have already been used for casimir experiments @xcite , seem readily adaptable for testing eq .
( [ eq : etheta ] ) . because the overall strength of the casimir effect is weaker for a disk than for a sphere , observing casimir forces in this geometry will require greater sensitivities or shorter separation distances than the sphere - plane case .
as the separation gets smaller , however , the dominant contributions arise from higher - frequency fluctuations , and deviations from the perfect conductor limit can become important .
while the effects of finite conductivity could be captured by an extension of our method , the calculation becomes significantly more difficult in this case because the matrix of scattering amplitudes is no longer diagonal . to estimate the range of important frequencies , we consider @xmath286 and @xmath138 . in this case , the integrand in eq .
( [ eq : zeror ] ) is strongly peaked around @xmath287 . as a result , by including only values of @xmath20 up to @xmath288 , we still capture @xmath289 of the full result ( and by going up to @xmath290 we include 99% ) .
this truncation corresponds to a minimum `` fluctuation wavelength '' @xmath291 . for the perfect conductor approximation to hold
, @xmath292 must be large compared to the metal s plasma wavelength @xmath293 , so that these fluctuations are well described by assuming perfect reflectivity .
we also need the thickness of the disk to be small enough compared to @xmath274 that the deviation from the proximity force calculation is evident ( see fig .
[ fig : evsh ] ) , but large enough compared to the metal s skin depth @xmath294 that the perfect conductor approximation is valid . for a typical metal film , @xmath295 and @xmath296 at the relevant wavelengths . for a disk of radius @xmath297
, the present experimental frontier of @xmath298 sensitivity corresponds to a separation distance @xmath299 , which then falls within the expected range of validity of our calculation according to these criteria .
the force could also be enhanced by connecting several identical but well - separated disks . in that case
, the same force could be measured at a larger separation distance , where our calculation is more accurate . in the case of overlapping planes ,
the correction to the traditional pfa energy is of a similar magnitude to the total force for perpendicular planes in the above example , and thus should also be measurable at these separations .
we have shown that thermal corrections are generally small at room temperature for typical separations , and furthermore our methods allow these corrections to be computed precisely .
we thank u. mohideen for helpful discussions and f. khoshnoud for correspondence regarding the example of overlapping planes .
this work was supported in part by the national science foundation ( nsf ) through grants phy08 - 55426 ( ng ) , dmr-08 - 03315 ( sjr and mk ) , defense advanced research projects agency ( darpa ) contract no .
s-000354 ( sjr , mk , and te ) , and by the u. s. department of energy ( doe ) under cooperative research agreement # df - fc02 - 94er40818 ( rlj ) .
s. k. lamoreaux , phys .
lett . * 78 * , 5 ( 1997 ) ; u. mohideen and a. roy , phys .
lett . * 81 * , 4549 ( 1998 ) ; g. bressi , g. carugno , r. onofrio , and g. ruoso , phys .
lett . * 88 * , 041804 ( 2002 ) ; h. b. chan , v. a. aksyuk , r. n. kleiman , d. j. bishop , and f. capasso , science * 291 * , 1941 ( 2001 ) . | an exact calculation of electromagnetic scattering from a perfectly conducting parabolic cylinder is employed to compute casimir forces in several configurations .
these include interactions between a parabolic cylinder and a plane , two parabolic cylinders , and a parabolic cylinder and an ordinary cylinder . to elucidate the effect of boundaries
, special attention is focused on the `` knife - edge '' limit in which the parabolic cylinder becomes a half - plane .
geometrical effects are illustrated by considering arbitrary rotations of a parabolic cylinder around its focal axis , and arbitrary translations perpendicular to this axis .
a quite different geometrical arrangement is explored for the case of an ordinary cylinder placed _ in the interior _ of a parabolic cylinder .
all of these results extend simply to nonzero temperatures . |
numerical methods to study electronic structure in condensed matter use mainly techniques developed for periodic systems.in many cases , however , methods developed for finite systems are also used .
they are especially adequate for ionic solids , liquids , molecular crystals , clusters of molecules , for instance , to study features of the electronic structure which are local in character .
in such a case , the electronic structure is modelled only in some well - defined region in space of direct relevance .
the effect of the atoms outside of this selected region ( referred to as environment in this work ) is taken into account by some _ embedding
potential_. different strategies are applied in practice to represent the embedding potential .
they differ in the choice of descriptor of the environment .
the roughest approximation is to neglect the environment entirely .
such simplification is commonly used to study chemical bonding and reactivity in condensed phase if the solvent in which the reaction takes place is known to play a secondary role . representing the environment ( discrete or continuous , polarisable or not ) by the electric field it generates ,
makes it possible to take into account the effect of the environment @xcite .
such classical treatment of the effect of the environment on the electronic structure is commonly used both in chemistry and in materials science ( for review see ref .
the embedding potential in such methods is obviously orbital - free .
it is , however , not exact because the quantum statistics nature of electrons is completely neglected . taking into account the fermion nature of electrons might proceed by following a similar strategy as the one applied by phillips and kleinman in the construction of pseudopotentials in order to eliminate explicit treatment of core electrons @xcite . for recent developments along these lines ,
see ref .
@xcite .
using the following elements of the hohenberg - kohn - sham formulation of density functional theory : hohenberg - kohn theorems @xcite , a reference system of non - interacting electrons @xcite , and the corresponding density functional of the kinetic energy ( @xmath4 $ ] ) @xcite in particular , leads to the embedding potential which is _ exact _ in the limit of exact functionals and _ orbital - free _ i.e. does not involve other descriptors of the environment than its electron density @xcite .
the pure - state non - interacting @xmath5-representable electron density @xmath6 , such that added to some arbitrarily chosen density associated with the environment ( @xmath7 ) minimises the hohenberg - kohn energy functional for the whole system , can be obtained from the following one - electron equations ( eqs .
20 - 21 in ref .
@xcite ) : @xmath8\right]\phi^a_{i}=\epsilon^a_{i}\phi^a_{i } \;\;\;\;\ ; i=1,n^{a } \label{ksceda } \\ \nonumber\end{aligned}\ ] ] where @xmath9 and and @xmath10 $ ] denotes a local potential which depends only on electron densities @xmath11 and @xmath7 .
the label ksced ( kohn - sham equations with constrained electron density ) is used here to indicate that the local potential differs from that in kohn - sham equations @xcite for either the total system ( @xmath12)$ ] ) or the isolated subsystem @xmath13 ( @xmath14)$ ] ) . also the one - electron functions ( @xmath15 ) obtained from eq .
[ ksceda ] are not optimal orbitals in neither kohn - sham systems .
atomic units are applied in all formulas which are given for spin unpolarised systems .
the total effective potential in eq . [ ksceda ] is the sum of the conventional kohn - sham effective potential @xmath16 $ ] for the whole system evaluated for the electron density @xmath17 and another local potential ( @xmath0(\vec{r})$ ] ) : @xmath18= v_{eff}^{ks}\left[\rho_a+\rho_b;\vec{r}\right]+ v_t[\rho_a,\rho_b](\vec{r } ) % % \label{kscedpot0 } \\
\nonumber \end{aligned}\ ] ] where @xmath0(\vec{r})$ ] involves functional derivatives of the functional @xmath4 $ ] : @xmath19(\vec{r})= \left.\frac{\delta t_{s}[\rho]}{\delta \rho}\right|_{\rho=\rho_a+\rho_b } - \left.\frac{\delta t_{s}[\rho]}{\delta \rho}\right|_{\rho=\rho_a } \label{definition_of_vt } \end{aligned}\ ] ] note that no restriction is made concerning the overlap between @xmath11 and @xmath7 in real space .
the potential @xmath0(\vec{r})$ ] can be alternatively expressed as : @xmath20(\vec{r})= \left.\frac{\delta t_{s}^{nad}[\rho,\rho_{b}]}{\delta \rho}\right|_{\rho=\rho_a } \label{definition_of_vt_alt } \end{aligned}\ ] ] where @xmath21 $ ] denotes the following difference : @xmath22=t_s[\rho_a+\rho_b]-t_s[\rho_a]-t_s[\rho_b]\label{tsnad_def } \\ \nonumber\end{aligned}\ ] ] for the sake of the subsequent discussions , it is convenient to split the total effective potential @xmath10 $ ] into two components : the kohn - sham effective potential for the isolated subsystem @xmath13 ( @xmath23 $ ] ) , which is @xmath7 independent , and the remaining part representing the environment : + @xmath18= v_{eff}^{ks}\left[\rho_a;\vec{r}\right]+ v_{emb}^{ksced}\left[\rho_a,\rho_b;\vec{r}\right ] \label{kscedpot1 } \\
\nonumber \end{aligned}\ ] ] where @xmath24&= & v_{ext}^b(\vec{r } ) + \int \frac{\rho_{b}(\vec{r}')}{|\vec{r}'-\vec{r}| } d\vec{r } ' \label{kscedembpot } \\
\nonumber \\ & + & \left.\frac{\delta e_{xc}\left[\rho\right]}{\delta\rho}\right\vert_{\rho=\rho_a+\rho_b}- \left.\frac{\delta e_{xc}\left[\rho\right]}{\delta\rho}\right\vert_{\rho=\rho_a}+ v_t[\rho_a,\rho_b](\vec{r } ) % \nonumber \\ \nonumber\end{aligned}\ ] ] where @xmath25 $ ] denotes the kohn - sham functional of the exchange - correlation energy @xcite .
orbital - free effective embedding potential given in eq .
[ kscedembpot ] , and its @xmath0(\vec{r})$ ] component in particula , are used in various types of multi - level numerical simulations ( for a review , see ref .
@xcite or refs .
@xcite for representative recent reports ) .
such simulations deal with condensed matter systems , for which the electronic features of a selected subsystem ( subsystem @xmath13 ) are subject to detailed investigation whereas @xmath7 is subject to additional simplifications .
other formal frameworks use also @xmath0(\vec{r})$ ] such as : cortona s formulation of density functional theory @xcite , where @xmath7 is not an assumed quantity but a result of fully variational calculations @xcite or linear - response time - dependent density - functional - theory description of electronic excitations localised in embedded systems @xcite . finally , the orbital - free effective embedding potential given in eq
. [ kscedembpot ] , and its @xmath0(\vec{r})$ ] component in particular , are used in combination with traditional wave - function based methods by carter and collaborators ( see for instance ref .
@xcite ) . for
the formal analysis of applicability of such a combination , see ref .
@xcite which shows that the exact embedding potential in such a case always comprises the @xmath0(\vec{r})$ ] component . in practical applications , @xmath0(\vec{r})$ ] is not used but some analytic expressions approximating this quantity ( @xmath26(\vec{r})$ ] ) for the obvious sake of practical advantages . this replacement results in errors in all derived quantities which will be referred to as _
@xmath27-induced errors_. in each case , @xmath26(\vec{r})$ ] is obtained by using analytic form of an approximated functional @xmath28 $ ] into eqs .
[ definition_of_vt]-[tsnad_def ] .
we will refer to such @xmath26(\vec{r})$ ] as _ decomposable _ because the analytic form of all relevant quantities : @xmath28 $ ] , @xmath29 $ ] , and @xmath26(\vec{r})$ ] , is available .
if the form of the used @xmath28 $ ] comprises only low - level gradient - expansion @xcite contributions , the corresponding decomposable @xmath26(\vec{r})$ ] violates the exact limit for @xmath30(\vec{r})$ ] at @xmath2 and @xmath3 ( see appendix a ) .
our interest in the local behaviour of @xmath30(\vec{r})$ ] at this limit is motivated by the fact that the corresponding conditionsoccur if @xmath7 comprises two electrons tightly bound to a distant nucleus in the environment such as in the case of the helium atom , li@xmath31 cation , be@xmath32 , etc .
we expect that they are relevant also for heavier nuclei if in a volume element centred on the nucleus @xmath7 is dominated by a doubly - occupied orbital .
the present work focuses on the investigation whether the considered exact limit is of any practical relevance . to this end
, we apply the following strategy : _ i _ ) we use a model system ( appendix b ) , for which the conditions @xmath2 and @xmath3 apply rigorously , to analyse the importance of enforcing the correct local behaviour of @xmath26(\vec{r})$ ] .
_ ii _ ) we construct a simple approximation to @xmath30(\vec{r})$ ] obeying the considered exact limit in the vicinity of nuclei and analyse the numerical significance of imposing the considered condition in real systems where the conditions @xmath2 and @xmath3 do not apply rigorously .
our ultimate goal is a new approximation to @xmath0(\vec{r})$ ] which can be inexpensively evaluated in practice and obeys as much as possible of the relevant exact properties .
it should pointed out in this context that the position - dependency of @xmath0(\vec{r})$ ] is the result of non - homogeneity of @xmath11 and/or @xmath7 .
therefore , the symbol @xmath0(\vec{r})$ ] ( or @xmath26(\vec{r})$ ] if approximated ) is used throughout this work to indicate that this local quantity is a functional of @xmath11 and @xmath7 .
explicit position dependence is strongly undesired in density - functional - theory based methods because it is not straightforward to obtain _ i _ ) such potential as a functional derivative of some density functional , and _ ii _ ) functional derivatives of such potential needed in some formal frameworks @xcite .
general symbols such @xmath1 or @xmath33 are used in some discussions where the issue of explicit position - dependence is not relevant .
before proceeding to the construction of the desired approximation to @xmath0(\vec{r})$ ] obeying the considered limit , we overview the conventional construction of approximation to @xmath0(\vec{r})$ ] and the local behaviour near a nucleus of the obtained potential .
the conventional strategy , which is applied in our own works and the works by others so far , is to start from some explicit density functional @xmath28 $ ] and to use its analytic form to derive the corresponding approximate expression for @xmath21 $ ] : @xmath22\approx \tilde{t}_s^{nad}[\rho_a,\rho_b]= \tilde{t}_s[\rho_a+\rho_b]- \tilde{t}_s[\rho_a ] - \tilde{t}_s[\rho_b]\label{decomposablege}\end{aligned}\ ] ] and to use the obtained analytic expression to obtain @xmath26(\vec{r})$ ] by means of functional differentiation .
\tilde{v}_t[\rho_a,\rho_b](\vec{r } ) = \left.\frac{\delta \tilde{t}_{s}^{nad}[\rho,\rho_{b}]}{\delta \rho}\right|_{\rho=\rho_a } \label{decomposablev } \end{aligned}\ ] ] this strategy can be applied for any approximated functional @xmath28 $ ] provided its form makes it possible to obtain the analytic expression for @xmath26(\vec{r})$ ] .
simple functionals @xmath28 $ ] , which depend explicitly on densities and their gradients , are of particular practical interest .
they lead to @xmath26(\vec{r})$ ] which depends explicitly only on @xmath11 and @xmath7 and their first- and second derivatives . in the original work by cortona
@xcite , where the subsystem formulation of density functional theory was introduced , a decomposable @xmath26(\vec{r})$ ] derived from the thomas - fermi @xcite kinetic energy functional was used to study ionic solids . in our own works ,
only decomposable @xmath26(\vec{r})$ ] derived from gradient - dependent approximations to @xmath4 $ ] were considered so far ( see for instance the analyses of their accuracy in ref .
@xcite or their recent applications in multi - level computer simulations of condensed matter @xcite ) .
thomas - fermi kinetic energy functional @xcite , which is exact for the uniform electron gas , leads to the following approximate expression for @xmath34 $ ] : @xmath35= c_{tf}\int \left ( \left(\rho_a+\rho_b\right)^{5/3 } -\rho_a^{5/3 }
-\rho_b^{5/3 } \right)d\vec{r } \label{tsnad_gea0}\end{aligned}\ ] ] where @xmath36 .
+ the associated expression for @xmath37(\vec{r})$ ] reads : @xmath38(\vec{r})= \frac{5}{3}c_{tf}\left ( \left(\rho_a+\rho_b\right)^{2/3 } -\rho_a^{2/3 } \right ) \label{vtsnad_gea0}\end{aligned}\ ] ] approximating @xmath4 $ ] by means of the gradient expansion of the kinetic energy @xcite truncated to the second order leads to the following approximate expression for @xmath34 $ ] @xcite : @xmath39=t_{s}^{nad(tf)}[\rho_{a},\rho_{b}]- \frac{1}{72 } \int\frac{|\rho_a\nabla\rho_b-\rho_b\nabla\rho_a|^2}{\rho_a\rho_{b}(\rho_a+\rho_b)}d\vec{r } \label{tsnad_gea2 } \\ \nonumber\end{aligned}\ ] ] the associated expression for @xmath40(\vec{r})$ ] is given in ref .
@xcite . for the the group of gradient - dependent approximations to @xmath4 $ ] of the _ generalised gradient approximation _ form @xcite
the analytic expression for @xmath41 $ ] reads : @xmath42=c_{tf}\int \left [ \left(\rho_a+\rho_b\right)^{5/3 } f(s_{ab } ) -\rho_a^{5/3 } f(s_{a } ) -\rho_b^{5/3 } f(s_{b } ) \right ] d\vec{r } \label{tsnad_gga } \\
\nonumber\end{aligned}\ ] ] where @xmath43 ( _ enhancement factor _ ) depends on a dimensionless quantity @xmath44 ( _ reduced density gradient _ ) .
various analytic forms of @xmath43 were proposed in the literature @xcite . the associated analytic expression for @xmath45(\vec{r})$ ]
is given in ref .
it is worthwhile to notice that the gga form is flexible and includes @xmath46 $ ] and @xmath47 $ ] as special cases .
numerical values of @xmath48 provide useful information about shell structure and the distance from the nucleus in atoms @xcite .
for an atom , @xmath48 is known to be small near the nucleus , reach the values of about 3 in the valence region , and diverge exponentially to @xmath49 at large distances . in molecules
, it behaves similarly with a noticeable exception of stationary points of electron density ( bond midpoints for instance ) where @xmath50 .
each approximated functional given in eqs .
[ tsnad_gea0]-[tsnad_gga ] comprises a dominant thomas - fermi component and satisfies two exact conditions : * @xmath51-t_s[\rho_a]-t_s[\rho_b]=0 $ ] for non - overlapping @xmath11 and @xmath7 . * for uniform @xmath11 and @xmath7
, they recover the exact analytical expression for @xmath51-t_s[\rho_a]-t_s[\rho_b]$ ] .
the common feature of each among the above approximations for @xmath4 $ ] is that none of them yields the exact analytic form of @xmath30(\vec{r})$ ] at @xmath2 and @xmath3 ( see appendix a ) : @xmath52(\vec{r } ) % { \longrightarrow } % % % v_t^{limit}[\rho_b](\vec{r})= \frac{1}{8}\frac{\vert \nabla \rho_b\vert^2}{\rho_b^2 } - \frac{1}{4}\frac{\nabla^2\rho_b}{\rho_b } \label{vsnadw } \\ \nonumber\end{aligned}\ ] ] the expression given in eq .
[ vtsnad_gea0 ] , which provides the dominant contribution to gradient - expansion based approximations to @xmath0(\vec{r})$ ] does not comprise the relevant term at all whereas the second - order term provides only 1/9 of the exact expression .
figure [ lda_embedding_potential ] shows @xmath53(\vec{r})$ ] for a spherically symmetric case : @xmath54 , @xmath55 , and @xmath2 , which represents a helium atom far from subsystem @xmath13 . in the figure as well as in the following discussion , @xmath56 denotes the distance from the considered nucleus .
the potential in the figure shows the features which are common also for heavier atoms if gradient expansion based approximations to @xmath4 $ ] are used to derive @xmath0(\vec{r})$ ] : a very narrow and deep well ( reaching @xmath57 ) centred on the nucleus and surrounding it repulsive shell .
the @xmath30(\vec{r})$ ] component of the shown embedding potential is finite at the nucleus instead of behaving as @xmath58 .
note that the exact term has the same form as the potential due to coulomb attraction by the nucleus of the charge @xmath59 .
therefore , the exact @xmath1 partially compensates this attraction to some extend because @xmath60 is smaller than @xmath59 @xcite . for the particular case considered in the figure , the missing @xmath58 component does not lead to any bound states ( appendix b ) . in general , however , the wrong asymptotic of the singularity at the nucleus , can lead to an unphysical transfer of electron density from the investigated system to its environment ( charge - leak @xcite ) .
this can occur if the artificially attractive ( not sufficiently repulsive ) approximation to the orbital - free effective embedding potential generates a bound state in the environment of the energy which is lower than the eigenvalue of the highest occupied embedded orbital associated with embedded subsystem @xmath13 .
numerical cases confirming such scenario are known @xcite . moreover
, the numerical solution of the schrdinger equation with @xmath53(\vec{r})$ ] for @xmath61 and the external potential of the @xmath62 form , shows a deeply lying node - less bound state of the energy -0.209665 hartree with the maximum of the radial electron density at @xmath63 bohr ( appendix b ) .
the above observations indicate that decomposable @xmath26(\vec{r})$ ] obtained from low - order gradient - expansion based approximations to @xmath4 $ ] might not be adequate for , at least , li@xmath31 cations in the environment .
the fact that the bound state associated with an atom in the environment is too tightly bound to the nucleus indicate , however , that the problem might be also present in atoms comprising more electrons . in a subsequent section , a simple approximation to @xmath30(\vec{r})$ ] is constructed based on these observations . in principle
, the exact limit should be applied in any volume element in which @xmath11 vanishes and @xmath7 is obtained from a doubly occupied orbital . in the proposed construction ,
the exact limit is imposed only at volume elements near heavier - than - hydrogen nuclei expecting that the considered condition is most relevant there .
the aforementioned flaws of decomposable strategy to construct gradient- and laplacian dependent approximations to @xmath30(\vec{r})$ ] suggest a bottom - up approach in which @xmath30(\vec{r})$ ] is directly a target .
a given approximated potential @xmath26(\vec{r})$ ] will be referred to as _ non - decomposable _ if the analytic form of its two individual components @xmath65}{\delta\rho}\right|_{\rho=\rho_a+\rho_b}$ ] and @xmath65}{\delta\rho}\right|_{\rho=\rho_a}$ ] can not be reconstructed .
the non - decomposable strategy is motivated by the fact that there are exact properties of @xmath30(\vec{r})$ ] , which can be taken into account quite easily in @xmath26(\vec{r})$ ] , whereas building - in them into some approximate functional @xmath28 $ ] is less straightforward .
abandoning the decomposable strategy is motivated also by the results of our recent dedicated studies of the accuracy of various gradient - dependent approximations to @xmath21 $ ] , which revealed that there is no correlation between the accuracy of @xmath29 $ ] , @xmath26(\vec{r})$ ] and the errors in the parent gradient - dependent approximation to @xmath28 $ ] @xcite .
it should be also pointed out that , that the individual contributions @xmath29 $ ] are not needed in practice .
the non - decomposable approximation to @xmath30(\vec{r})$ ] is constructed by enforcing the following exact conditions into its analytic form : * @xmath29\longrightarrow \tilde{t}_s^{nad(lda)}[\rho_a,\rho_b]$ ] for uniform @xmath11 and @xmath7 .
* @xmath29\longrightarrow 0 $ ] for non - overlapping @xmath11 and @xmath7 . * @xmath26\longrightarrow v_t^{limit}[\rho_b]$ ] at @xmath2 and @xmath64 .
the first two conditionsare automatically satisfied by the decomposable gradient - expansion based approximations discussed in the previous section .
since such approximations proved to be sufficiently accurate for many systems the same conditions are retained in the new construction .
the last condition is the key element of the present construction . before proceeding to the construction of the approximation obeying the considered exact condition we note that @xmath30(\vec{r})$ ]
can be alternatively expressed as : @xmath52= \tilde{v}_t^{decomposable}[\rho_a,\rho_b]+f[\rho_a,\rho_b]\cdot v_t^{limit}[\rho_b ] \label{vtnad_alt}\end{aligned}\ ] ] all functionals in the above equation are determined locally and the argument @xmath66 is not written explicitly for simplicity and the functionals @xmath30 $ ] and @xmath67 $ ] are simply related ( @xmath67=\frac{v_t[\rho_a,\rho_b]-\tilde{v}_t^{decomposable}[\rho_a,\rho_b]}{v_t^{limit}[\rho_a,\rho_b]}$ ] if @xmath68 $ ] is non - zero ) .
the above form of @xmath30 $ ] provides a convenient for construction of approximation .
it can be used for any decomposable approximation to @xmath30 $ ] , which violaties the considered condition , and the functional @xmath67 $ ] has a clear physical meaning as a _ switching factor _ determining whether it is needed to add locally the missing component of the embedding potential . as far as the choice for the decomposable component , both the gradient - free potential given in eq .
[ vtsnad_gea0 ] and the decomposable potential derived from the lembarki - chermette @xcite approximation to @xmath4 $ ] were shown in dedicated studies @xcite to be reasonably accurate if @xmath11 and @xmath7 do not overlap strongly .
both these approximations comprise the zeroth order contribution . if @xmath37 $ ] is used as decomposable component in eq .
[ vtnad_alt ] , the terms @xmath69 $ ] and @xmath70 $ ] are local functions depending explicitely on @xmath11 , @xmath7 , @xmath71 , and @xmath72 . approximating @xmath67
$ ] by a local function depending explicitely on these quantities applied in eq .
[ vtnad_alt ] leads to an approximated potential requiring a similar computational effort as conventional low - order gradient - expansion based decomposable approximations to @xmath30 $ ] .
the first approximation made here is replacing the switching factor defined in eq .
[ vtnad_alt ] by a _ switching function _ : @xmath73(\vec{r})\approx\tilde{f}(\rho_a,\rho_b,\nabla\rho_b,\nabla^2\rho_b)\end{aligned}\ ] ] the above considerations lead to the following general form of the approximation to @xmath0(\vec{r})$ ] : @xmath74= % \tilde{v}_t^{tf}[\rho_a,\rho_b]+\tilde{f}(\rho_a,\rho_b,\nabla\rho_b,\nabla^2\rho_b ) \cdot v_t^{limit}[\rho_b ] \label{ndsd}\end{aligned}\ ] ] the above general form provides a clear interpretation for the switching function , which can be used as guideline in construction of approximations - it `` detects '' such volume elements for which the conditions @xmath2 and @xmath64 are most relevant . in constructing @xmath76 the following additional requirements ( simplifications )
are made : * @xmath76 is one in the vicinity of a nucleus to account fully for the missing @xmath58 component . *
the criterion for determining the range at which @xmath69 $ ] is nuclear number independent . *
@xmath76 does not depend on @xmath11 ( to obtain the analytic form of @xmath29 $ ] : ( @xmath67\approx \tilde{f}[\rho_b]$ ] ) the above criteria are very restrictive and leave us with not many choices .
the last one leads to the following form of the switching factor : @xmath73\approx \tilde{f}(\rho_b,\nabla\rho_b,\nabla^{2}\rho_b)\end{aligned}\ ] ] approximating @xmath67 $ ] by some function @xmath77 is one of possible further simplifications .
it is , however , very unlikely that a @xmath7-based switching function could be universal . the electron density near the nucleus depends on the effective nuclear charge @xmath60 @xcite and varies strongly from atom to atom .
it is possible , however , to design an @xmath60-independent criterion . to this end , we consider the reduced density gradient ( @xmath78 ) defined as : @xmath79 for @xmath7 obtained from hydrogenic orbital @xmath80 defined by some effective nuclear charge @xmath60
, the following @xmath60-independent observations can be made : _ i _ ) for @xmath81 , @xmath82 , and _ ii _ ) @xmath83 $ ] changes sign from positive to negative at @xmath84 .
these observations suggest that the switching function can take a very simple and @xmath60-independent form @xmath85 where @xmath86 for @xmath87 and @xmath88 for @xmath89 and @xmath90 and @xmath91 . since eq .
[ ksceda ] can be used also to obtain forces @xcite , it is preferable to use a smooth switching from @xmath92 to @xmath93 instead of @xmath94 in the above definition .
the simplest form of such a switching functional has the fermi - dirac statistics form : @xmath95 where the parameter @xmath96 determines the smoothness of the switch .
@xmath97 in eq .
[ smoothswitch1 ] leads to equivalent results to that obtained with the step function @xmath98 ( differences in dipole in the range of 10@xmath99 debye and orbital energies in the range of 10@xmath100 hartree ) .
all results discussed in this work are obtained with @xmath97 .
smaller values corresponding to even `` softer '' switching can be also used in our numerical implementation but is not considered here in order to minimise the use of adjustable parameters . a switching factor
@xmath75 constructed following the above restrictions can be used to investigate the importance of enforcing the exact limit for @xmath0(\vec{r})$ ] at @xmath2 and @xmath64 but only in cases where the environment comprises one nucleus and two electrons . not only the condition @xmath3 apply rigorously but the considerations leading to the @xmath60-independent values of @xmath101 and @xmath102 apply as well . such model systems as li@xmath31-h@xmath103o and be@xmath32-h@xmath103o complexes if the water molecule is considered as subsystem @xmath13 and the cation as subsystem @xmath104 ( environment ) fall into this category . at equilibrium geometry addition of @xmath105 $ ] results in a desired effect on calculated properties for these complexes .
the lowest unoccupied embedded orbital associated with subsystem @xmath13 is indeed localised on the cation and its energy is shifted by 0.262 ev in the case of li@xmath31 and by 0.830 ev in the case of be@xmath32 .
addition of @xmath105 $ ] reduces also the dipole moment of a water molecule in the vicinity of the cation by 0.071 debye and 0.409 debye for li@xmath31 and be@xmath32 , respectively .
such noticeable numerical effects obtained for systems , for which the condition @xmath3 applies rigorously , indicate clearly that the exact limit considered might be relevant for practical calculations where the environment is larger . using only universal parameters @xmath102 and @xmath101 is very appealing but the reasoning leading to their numerical values does not apply in real systems .
it allows one to study the importance of the considered exact conditions but only in particular cases . the approximation to @xmath0(\vec{r})$ ] defined in eq .
[ ndsd ] and using the switching function given in eq .
[ smoothswitch1 ] has , therefore , little practical value .
the construction of the switching function described in the previous section and choice of the thresholds @xmath102 and @xmath101 in particular can be expected not to be adequate for other systems for the following reasons : * the criterion @xmath106 applies at hydrogen nucleus where addition of @xmath69 $ ] is not expected to be needed .
we recall here that it is the danger of a collapse of electron density on a doubly occupied hydrogenic @xmath80 orbital provides the physical motivation for introducing the @xmath69 $ ] .
moreover , numerical studies on molecular electron densities indicate clearly that enforcing the local behaviour of the density of the kinetic energy near nucleus corresponding to the von weizscker expression leads indeed to significant improvements of the approximation to @xmath4 $ ] for all nuclei except that of hydrogen @xcite . * for heavier atoms , electron density at the nucleus comprises contributions from other orbitals than the hydrogenic @xmath80
this can lead to the possibility that is @xmath69 $ ] is negative although @xmath107 . *
the criterion @xmath106 might also be satisfied near stationary points of the electron density such as bond midpoints .
it is very unlikely that the condition @xmath3 can be relevant to any volume element centred on a stationary points .
therefore addition of @xmath69 $ ] lacks formal justification there .
although the condition @xmath108 assures that @xmath69 $ ] is not added at the stationary point , where @xmath109 , or in the close proximity to it , the numerical value of this threshold requires verification in real systems.it should be added at this point also that , if linear combination of atomic orbitals is used to construct embedded orbitals , the quality of description of the density at the nuclear cusp depends on the used basis set
. a weaker criterion should be used in practice to assure that the @xmath69 $ ] is indeed added in the vicinity of the nucleus .
we start with the choice made for @xmath101 .
numerical analyses in the systems discussed in the next section show that @xmath69 $ ] is negative locally even if @xmath110 is smaller than 1.426 for heavier nuclei .
this suggest that this threshold should be reduced .
the smaller is the value for this threshold the less probably is inclusion of negative @xmath69 $ ] , which is desired from the point of view of universality of this threshold , but it comes at the expense of loosing a part of the desired effect at two - electron nuclei by reducing the range at which addition of @xmath69 $ ] applies .
we use the model system considered in appendix b to estimate the effect associated with the reduction of this range .
the underlying assumption leading to the value of 1.426 value is rigorously true in the model system .
the desired effect of reducing the charge distribution on top of the nucleus is achieved mainly by adding @xmath69(\vec{r})$ ] very close to the nucleus i.e. where @xmath111 .
increasing further the value of the @xmath110 threshold leads to smaller effect .
moreover , adding locally @xmath69 $ ] to the potential near the nucleus leads to negligible effect on orbital energies in the system considered in appendix b ( it reaches a peak of about @xmath112 ev at @xmath113 ) .
these results for the model system indicate that any choice for @xmath114 is acceptable . in real systems discussed in the next section , lowering the threshold from @xmath115 to to @xmath116
assures that @xmath69 $ ] is added only if it is positive .
the fine - tuning of @xmath102 follows other considerations .
we note that , in the model system considered in appendix b , @xmath102 can be reduced even to zero without affecting the results because lower values of @xmath110 than @xmath117 do not occur near the nucleus . to make sure that no nuclei is overlooked
even if the chosen atomic basis set in practical calculations is such that the exact relation @xmath118 for @xmath119 can not be rigorously satisfied , @xmath102 is reduced from 0.376 to 0.3 .
this change leads to negligible numerical effects if @xmath7 corresponds to atomic electron densities . for molecular @xmath7 , retaining the criterion based on the @xmath102
is necessary to avoid unjustified additions of @xmath69 $ ] near stationary points .
the criteria based only on @xmath102 and @xmath101 are not sufficient if the environment comprises hydrogen atoms because they are satisfied also at hydrogen nucleus . to avoid adding the @xmath69 $ ] near hydrogens ,
the proposed switching function includes additionally the criterion based on smallness of @xmath7 .
it is required that @xmath7 is larger than the square of the @xmath80 wave function of the hydrogen atom ( z=1 ) at @xmath120 which equals to @xmath121 . concerning @xmath122 , increasing the idealised value of 0.318 to even 1
does not affect the results for hydrogen - free systems because the density on top of any nucleus , which is heavier than hydrogen , is at least one order of magnitude larger .
increasing the value @xmath122 is desired for the same reasons as the ones motivating the decrease of @xmath102 .
the value of @xmath123 was arbitrary chosen for practical calculations .
the final form of the `` fine - tuned '' switching function of more general applicability and used in the subsequent section for studying the importance of imposing the considered exact limit in real systems take the following form : @xmath124 where @xmath125 , @xmath126 , @xmath123 .
[ vtsnad_gea0 ] , [ vsnadw ] , [ ndsd ] , and [ smoothswitch ] define the potential which will be referred to as @xmath127 $ ] , ( non - decomposable approximation using first- and second derivatives of @xmath128 ) .
the notion of non - decomposability is brought up here because the the second term in eq .
[ ndsd ] does not have the form of a difference between two functional derivatives of some common explicit density functional @xmath28 $ ] .
the analytic expression for @xmath129 $ ] , which yields @xmath127 $ ] after functional differentiation with respect to @xmath11 can be easily constructed .
its decomposable component is given in eq .
[ tsnad_gea0 ] and the non - decomposable @xmath69 $ ] component is @xmath11 independent .
the functional generating @xmath130 $ ] reads therefore : @xmath131&= & c_{tf}\int \left ( \left(\rho_a+\rho_b\right)^{5/3 } -\rho_a^{5/3 } -\rho_b^{5/3 } \right)d\vec{r } \\ & + & \int f(\rho_b,\nabla\rho_b)\cdot \rho_a(\vec{r } ) \tilde{v}_t^{limit}[\rho_b](\vec{r } ) d\vec{r } + c[\rho_b ] \nonumber\end{aligned}\ ] ] where @xmath132 $ ] is @xmath11-independent . to assure the proper dissociation limit @xmath132 $ ]
must vanish .
in practical applications of eq .
[ ksceda ] , the results depend on @xmath7 as well as on the used approximation to @xmath0(\vec{r})$ ] .
as far as the quality of the used approximation to @xmath0(\vec{r})$ ] is concerned , a general procedure was proposed in one of our earlier works @xcite .
its principal element is the comparison between numerical values of the calculated property ( energy components , dipole moments , total electron density , etc . )
obtained from two fully variational formal frameworks : that of cortona @xcite and that of kohn and sham @xcite .
results obtained from both frameworks are not exact but the difference between them can be attributed only to the approximation used for @xmath0(\vec{r})$ ] if all technical parameters ( approximation to the exchange - correlation functional , basis sets for expanding orbitals , algorithms to calculate used matrix elements ) are the same .
we point out here , however , that direct comparisons between the total electron densities derived from kohn - sham- and cortona s calculationsare cumbersome because these quantities are local . in practice
, it is more convenient to use global quantities ( norm of the difference between these densities , or selected observables ) in such analyses ( see the next section ) . to obtain the pair of electron densities @xmath11 and @xmath7 which minimises the total energy in cortona s type of calculations , a self - consistent super cycle of embedding calculations ( _ freeze - and - thaw _ cycle ) is performed . at each iteration , eq .
[ ksceda ] are solved . in the subsequent iteration , @xmath11 and
@xmath7 exchange their role in eqs .
[ ksceda].the _ freeze - and - thaw _ iterations continue until self - consistency . in the end , a pair of electron densities ( @xmath133 and @xmath134 ) and the corresponding two sets of embedded orbitals is obtained .
obviously , the notion of _ embedded system _ and its _ environment _ becomes meaningless because both subsystems are treated on the equal footing .
_ freeze - and - thaw _ calculations are conducted in practice for small model systems to validate the used @xmath7 in large scale multi - level numerical simulations @xcite or , as it is made in the present work , to asses the used approximation to @xmath0(\vec{r})$ ] .
@xmath27-generated errors in complexation induced dipole moments can be expected to be strongly affected by the local behaviour of the used @xmath26(\vec{r})$ ] near nuclei .
lack of sufficient repulsion near the nucleus might lead to an artificial transfer of electron density between subsystems reflected in the numerical values of the dipole moment . therefore , this quantity was chosen for the analysis of the @xmath27-generated errors in a representative set of intermolecular complexes including charged , polar and non - polar ones at their equilibrium geometries . for key details of the numerical implementation of the relevant equations ,
see ref .
@xcite .
tables [ dipolemoments.tab ] and [ dipolemoments_charged.tab ] collect the complexation induced dipole moments in neutral or charged complexes , respectively .
first of all , switching on the @xmath135(\vec{r})$ ] term decreases the @xmath27-generated errors in each of the considered cases . for systems comprising neutral subsystems the effect on the errors are negligible .
this indicated that the origin of the errors lies not in the violation the condition considered in this work . for systems comprising charged components ,
the effect of imposing the considered limit is evident .
the errors are invariably reduced .
the reduction of the relative errors depends on the system from such a case as li@xmath31-h@xmath103o ( from 9.8% to 7.9% ) to a reduction by factor 2 or 3 in na@xmath31-br@xmath136 and ( from 1.4% to 0.7% ) and na@xmath31-h@xmath103o ( from 0.23% to 0.07% ) .
similarly as in the group of complexes formed by neutral molecules , the origin for the remaining errors lies somewhere else .
the above numerical examples lead to the following principal conclusions : * violation of the limit for @xmath0(\vec{r})$ ] at @xmath2 and @xmath64 contributes to the overall error in the calculated quantities but this contribution varies from one system to another .
it is rather negligible for complexes formed by neutral components .
it is numerically significant for complexes comprising charged components . * our simple strategy to impose the considered limit locally in the vicinity of nuclear cusps leads invariably to reduction of errors .
therefore it can be used generally as correction to any approximation violating the above limit . *
the construction of the approximation obeying the considered limit , and the used switching criteria in particular , corresponds to a real case where a distant nucleus is surrounded by a frozen - density shell comprising two electrons ( he , li@xmath31 , be@xmath32 , etc . ) .
the fact that the errors are reduced also for systems , where these idealised conditions do not apply , indicates that the considered condition is important and should be taken into account in construction of approximations to @xmath30(\vec{r})$ ] . * for systems , where @xmath27-generated errors were not reduced by imposing the considered exact limit , their origin
must be looked for somewhere else . in this section , we analyse the complexation induced shifts of orbital energies derived using the approximated potential @xmath127(\vec{r})$ ] considered in the previous section .
opposite to the dipole moments discussed previously , direct comparisons between the calculated shifts and the corresponding reference data are less straightforward .
however , we investigate the numerical effect associated with imposing the exact limit for @xmath0(\vec{r})$ ] at @xmath2 and @xmath64 in view of the numerical practice which indicates that shifting the levels of unoccupied orbitals localised in the environment would be desirable . conventional decomposable approximations to @xmath30(\vec{r})$ ]
lead to artificially low levels of unoccupied orbitals in the environment which might cause unphysical effects such as charge - transfer between subsystems @xcite or erroneous other observables @xcite .
the numerical results for a model system considered in appendix b show that inclusion of the @xmath69 $ ] term everywhere where @xmath137 into the effective embedding potential leads to a positive shift of the energy level of the unoccupied orbital . in real intermolecular systems ,
the conditions considered in appendix b ( @xmath2 and @xmath138 ) are not satisfied .
the subsystems are in finite separation and the use of atom - centred basis sets for each subsystem , which include all atoms ( supermolecular expansion labelled as ksced(s ) in ref .
@xcite ) , results in the fact that @xmath11 can be significant at a nucleus associated with subsystem @xmath104 . for the same reasons and the fact that the considered nuclei include also atoms with occupied @xmath139 shell ,
also the second assumption is not satisfied rigorously .
therefore , it is useful to verify in practice to which extend inclusion of the @xmath69 $ ] term affects the orbital levels if these asymptotic conditions do not apply . tables [ lueo_shifts_1 ] and [ lueo_shifts_3 ] collect the values of energy levels corresponding to the lowest lying orbital localised mainly in the environment for the previously considered complexes .
the calculations are made not in the end of the _ freeze - and - thaw _ cycle but for the same @xmath7 obtained from kohn - sham calculations for the isolated subsystem b. including @xmath69 $ ] leads to the shifts of the energy levels of unoccupied orbitals of the magnitude which significantly larger than that in the model system .
it is a very desired effect of the new approximation . in the li@xmath31-h@xmath103o case discussed in ref .
@xcite , the energy of the lowest unoccupied embedded orbital localised on li@xmath31 crosses that of the highest occupied embedded orbital localised on h@xmath103o at the intermolecular distance of 13 if eq . [ vtsnad_gea0 ] is used for @xmath0(\vec{r})$ ] . at larger separations , the self - consistent procedure to solve eqs .
[ ksceda ] does not converge due to localisation of the highest occupied embedded orbital which jumps between subsystems in subsequent iterations .
addition of the @xmath69 $ ] term shifts the energy of the unoccupied embedded orbital localised at li@xmath31 . as the result , no crossing of levels occurs even at intermolecular separations as large as 18 .
the occupied levels are affected less strongly ( see tables [ hoeo_shifts_1 ] and [ hoeo_shifts_3 ] ) .
in principle , any decomposable approximation can be used as the first term of eq .
[ ndsd ] instead of the term derived from thomas - fermi functional . in the approximation introduced in this work ,
the decomposable component of @xmath127 $ ] is gradient - free and does not contribute to the asymptotic local behaviour of @xmath0(\vec{r})$ ] at the nuclei .
only the second term enforces the desired behaviour .
should a gradient - dependent alternative for the first term be considered , a proper care should be taken to avoid double - counting of @xmath69 $ ] at a nucleus .
for instance , @xmath40 $ ] comprises already 1/9 of @xmath69 $ ] . to verify whether further improvements are possible following this lines , two approximations were considered by replacing @xmath37 $ ] in eq .
[ ndsd ] by either @xmath40 $ ] @xcite or @xmath140 $ ] @xcite . including the @xmath69 $ ] term into @xmath141(\vec{r})$ ] brings improvements in both cases ( see table [ other.decomposable ] ) .
this indicates @xmath69 $ ] should be enforced universally on any approximation to @xmath142(\vec{r})$ ] .
compared to the results obtained using @xmath143(\vec{r})$ ] discussed earlier in this work , the two alternative non - decomposable approximations @xmath142(\vec{r})$ ] are not better .
we recall here the main reasons for singling out @xmath140 $ ] ) among other decomposable ones which depend explicitly on densities @xmath11 and @xmath7 as well as their first- and second derivatives .
* @xmath40 $ ] obtained from the second - order gradient - expansion approximation leads typically to worse results than that obtained from zeroth order @xcite indicating that the contribution to @xmath0 $ ] due to the second term in eq .
[ tsnad_gea2 ] is erroneous .
the deterioration of the results is pronounced the most at very small overlaps between @xmath11 and @xmath7 .
note that in the present work this flaw of @xmath40 $ ] manifests itself in the absence of convergent solutions of eq .
[ ksceda ] in na@xmath31-cl@xmath136 and li@xmath31-h@xmath103o cases ( see table [ other.decomposable ] ) . * @xmath140 $ ] was introduced as a pragmatic solution replacing @xmath40 $ ] .
due to its analytic form , the gradient - dependent contribution disappears at small overlaps between @xmath11 and @xmath7 . *
the functional @xmath144 $ ] generating the decomposable approximated potential @xmath145 $ ] is known to be a very good approximation to @xmath4 $ ] .
the above reasons for singling out @xmath140 $ ] are not applicable for the non - decomposable construction presented in this work . in @xmath127
$ ] the problematic second - order term lying at the origin of flaws of @xmath40 $ ] is either present where it is needed to assure the correct asymptotic limit ( i.e. in the vicinity of nuclei ) or absent .
numerical results collected in tables [ dipolemoments.tab ] , [ dipolemoments_charged.tab ] , and [ other.decomposable ] support fully the above formal reasons to consider @xmath127 $ ] as the successor of @xmath140 $ ] .
enforcing the considered exact limit for @xmath30(\vec{r})$ ] , as it is made in the proposed approximation , leads to a significant reduction of @xmath27-generated errors for charged systems .
errors in the complexation indiced dipole moments are reduced by more than 50% in some cases . for neutral systems ,
reduction of error takes also place but its magnitude is typically negligible .
this indicates that one of important sources of inaccuracies in the conventional ( i.e. decomposable and gradient - expansion based ) approximations to @xmath30(\vec{r})$ ] was identified .
the origin of the remaining contributions @xmath27-generated errors lies probably somewhere else .
the analytic form of the component of the local embedding potential enforcing the correct considered limit is , indeed , an approximation to @xmath30(\vec{r})$ ] because its position dependency is indirect - through the density and its gradient .
the function which was used to switch on the exact limit was designed based on the analysis of model system of relevance for elements of the first- , second- and the third period . using the local potential introduced in this work ( @xmath127(\vec{r})$ ] given in eq .
[ ndsd ] ) as an alternative to potentials derived using conventional strategy of deriving it from analytic form of functionals based on low - order terms in the gradient expansion of @xmath4 $ ] @xcite is recommended for the following reasons : + _ i _ ) * formal : * we believe that a proper strategy to improve approximations to functionals in density functional theory should proceed by imposing the most relevant exact conditions and @xmath127(\vec{r})$ ] was constructed in this way .
+ _ ii _ ) * practical : * the numerical results reported in this work show indeed that imposing this condition improves the obtained electron density and that the improvement varies from negligible to significant depending on the system .
+ _ iii _ ) * numerical : * evaluating @xmath127(\vec{r})$ ] involves the same quantities as evaluating its counterparts derived from gradient - expansion- ( up to second order ) and so called generalised gradient approximations to @xmath4 $ ] . concerning the area of applicability of @xmath127(\vec{r})$ ]
, it should be underlined that some arbitrary choices were made concerning the criteria for `` detecting '' the vicinity of a nucleus based only on electron density @xmath7 and its derivatives in the construction of this approximation.the chosen criteria are most adequate for such nuclei , where the total electron density is dominated by the hydrogenic @xmath80 orbital .
the local potential @xmath127(\vec{r})$ ] introduced in this work is intrinsically non - decomposable .
although the analytic form of the functional @xmath129 $ ] and the potential @xmath127(\vec{r})$ ] are given in this work , the analytic form of neither @xmath146 $ ] nor @xmath147}{\delta\rho}$ ] is available .
therefore , the introduced here non - decomposable strategy can be seen as the first attempt to decouple the search for @xmath28 $ ] and its functional derivative , which are needed in orbital - free calculations , from the search for an adequate approximation for the kinetic - kinetic - energy dependent component of the effective orbital - free embedding potential .
opposite to orbital - free strategy @xcite , neither @xmath28 $ ] nor its functional derivative are needed in methods applying orbital - free effective embedding potential given in eq .
[ kscedembpot ] . finally , imposing the exact limit for @xmath0(\vec{r})$ ] at @xmath2 and @xmath64 leads to shifts of the level of the unoccupied orbitals localised in the environment .
such shift is strongly desired in view of the earlier reports on possible practical inconveniences resulted from artificially low position of such levels when approximations not taking into account the cusp condition are used @xcite .
_ acknowledgement : _ this work was supported by swiss national science foundation
. 99 e.g. mcrae , _ j. phys
_ , * 61 * ( 1957 ) 562
. o. tapia and o. goscinski , _ mol .
phys._,*29 * ( 1975 ) 1653 .
j. quist and a. warshel , _ chem .
_ , * 93 * ( 1993 ) 2523 ; j. gao , in : _ reviews in computational chemistry _ vol .
k.b . lipkowitz and d.b .
boyd ( vch publishers , new york , 1996 ) 119 ; j. sauer , p. ugliengo , e. garrone , v.r .
sounders , _ chem . rev . _ ,
* 94 * ( 1994 ) 2095 .
phillips , l. kleinman , _
phys.rev._ , * 116 * ( 1959 ) 287 .
henderson , _
_ , * 125 * , 2006 , 014105 .
p. hohenberg , w. kohn , _ phys .
b _ , * 136 * ( 1964 ) 864 .
w. kohn , l.j .
sham , _ phys .
_ * 140 * ( 1965 ) a1133 .
m. levy , _ proc .
* 76 * ( 1979 ) 6062 .
wesoowski , a. warshel , _ j. phys . chem .
_ , * 97 * ( 1993 ) 8050
wesolowski , _ one - electron equations for embedded electron density : challenge for theory and practical payoffs in multi - level modeling of complex polyatomic systems _ in : computational chemistry : reviews of current trends - vol .
10 , j. leszczynski , ed .
, world scientific , 2006 , pp . 1 - 82 . e.v .
stefanovitch , t.n .
_ , * 104 * ( 1996 ) 2946
mei , l.l .
boyer , m.j .
mehl , m.m .
ossowski , h.t .
stokes , _ phys .
b _ , * 61 * , ( 2000 ) 11425 . j.r .
trail , d.m .
bird , _ phys .
b _ , * 62 * , ( 2000 ) 16402 .
m. zbiri , m. atanasov , c. daul , j .-
garcia lastra , t.a .
wesolowski , _ chem .
_ , * 397 * ( 2004 ) 441 - 446 .
f. shimojo , r.k .
kalia , a. nakano , p. vashishta , _ computer phys .
comm . _ , * 167 * 151 .
j. neugebauer , c.r .
jacob , t.a .
wesolowski , e.j .
baerends _ j. phys .
* 109 * ( 2005 ) 7805 .
olsson , g.y .
hong , a. warshel , _
_ , * 125 * , ( 2003 ) 5025 .
n. choly , g. lu , e. weinan , e. kaxiras , _ phys .
b _ , * 71 * ( 2005 ) 094101 . c.r .
jacob , l. visscher , _ j. chem .
_ , * 125 * ( 2006 ) 194104 .
j. neugebauer , e.j .
baerends , _ j. phys
a _ , * 110 * ( 2006 ) 8786 .
p. cortona , _
b _ , * 44 * ( 1991 ) 8454 . t.a .
wesolowski , j. weber , _ chem .
_ , * 248 * ( 1996 ) 71 .
wesolowski , f. tran , _ j. chem
_ , * 118 * ( 2003 ) 2072 .
m. iannuzzi , b. kirchner , j. hutter , _ chem .
_ , * 421 * ( 2006 ) 16 .
m. dulak , j. kaminski , t.a .
wesolowski , _
theor . & comput .
_ , ( 2007 ) * 3 * 735 .
casida , t.a .
wesolowski , _ intl . j. quant .
_ , * 96 * ( 2004 ) 577 . t.a .
wesolowski , _ j. am .
_ , * 126 * , ( 2004 ) 11444 .
t. klner , n. govind , y.a .
wang , e.a .
carter , _ j. chem .
_ , * 116 * ( 2002 ) 42 .
wesolowski , _ phys .
* 77 * 012504 .
kirzhnits , _ sov
jetp _ , * 5 * , ( 1957 ) 64 .
\(a ) l.h .
thomas , _ proc .
_ , * 23 * ( 1927 ) 542 ; ( b ) e. fermi , _ z. physik _ , * 48 * ( 1928 ) 73 .
wesolowski , _ j. chem .
_ , * 106 * ( 1997 ) 8516
bernard , m. dulak , j.w .
kaminski , t.a .
wesolowski , _ j. phys .
, * 41 * ( 2008 ) 055302 .
wesolowski , h. chermette , j. weber , _ j. chem
_ , * 105 * ( 1996 ) 9182 .
h. lee , c. lee , r. g. parr , _ phys . rev . a. _ ,
* 44 * ( 1991 ) 768 .
a. lembarki , h. chermette , _ phys .
a _ , * 50 * ( 1994 ) 5328 p. fuentealba , o. reyes , _ chem .
_ , * 232 * ( 1995 ) 31 . f. tran , t.a .
wesolowski , _ int .
j. quantum chem .
_ , * 89 * ( 2002 ) 441 .
a. zupan , j.p .
perdew , k. burke , m. causa , _ int .
j. quantum chem .
_ , * 61 * ( 1997 ) 835 .
e. clementi , d.l .
raimondi , _ j. chem
_ , * 38 * ( 1963 ) 2686
. m. dulak , t.a .
wesolowski , _ j. chem
_ , * 124 * ( 2006 ) 16410 .
jacob , s.m .
beyhan , l. visscher , _
_ , * 126 * ( 2007 ) 234116 .
m. dulak and t.a .
wesolowski , _ unpublished results _ c.f .
von weizsaecker , _
z. physik _ , * 96 * ( 1935 ) 431 .
wesolowski , j. weber , _ int .
j. quantum chem .
_ , * 61 * ( 1997 ) 303
. y. tal , r. f. w. bader , int .
j. quantum chem . : quantum chem .
12 ( 1978 ) 153 .
f. tran , t.a .
wesolowski , _ chem .
_ , * 360 * ( 2002 ) 209 .
dirac , _ proc .
cambridge philos .
_ , * 26 * ( 1930 ) 376 . )
lda approximation for exchange - correlation functional : exchange @xcite and vosko et al .
parametrisation @xcite of the ceperley - alder data @xcite , _ ii _ ) tzp atomic basis sets of the sto type @xcite ,
_ iii _ ) supermolecular expansion of electron density of each subsystem ( ksced(s ) label in ref .
@xcite , _ iv _ ) equilibrium geometries obtained from kohn - sham lda calculations with tzp basis sets ( available from the authors upon request ) , _ v _ ) numerical implementation of the @xmath148 in adf code @xcite version adf2007 downloaded from svn repository on 2007 - 05 - 23 .
d.m ceperley , b.j .
alder , _ phys .
_ , * 45 * ( 1980 ) 566 . s.h .
vosko , l. wilk , m. nusair , _ can .
_ , * 58 * ( 1980 ) 1200 . g. te velde , f.m .
bickelhaupt , e.j .
baerends , c. fonseca guerra , s.j.a .
van gisbergen , j.g .
snijders , t. ziegler , _ j. comput . chem .
_ , * 22 * ( 2001 ) 931 .
b. fornberg , _ siam .
_ , * 40 * ( 1998 ) 685
. v. weber , c. daul , r. baltensperger , _ comp .
_ , _ 163 _ ( 2004 ) 133 .
matlab 2006 http://www.mathworks.com/
@xmath150 + for small @xmath151 such that @xmath152 , + @xmath153&=&t_s^{nad}[\rho_a+\delta \rho,\rho_b]-t_s^{nad}[\rho_a,\rho_b ] \\
\nonumber \\ & = & t_s[\rho_a+\delta\rho+\rho_b]-t_s[\rho_a+\rho_b]-t_s[\rho_a+\delta\rho ] + t_s[\rho_a ] \nonumber \\
\nonumber \\ & = & \int \left.\frac{\delta t_s[\rho]}{\delta \rho}(\vec{r})\right|_{\rho=\rho_a+\rho_b } \delta{\rho}(\vec{r})d\vec{r } % % % - \int \left.\frac{\delta t_s[\rho]}{\delta \rho}(\vec{r})\right|_{\rho=\rho_a } \delta{\rho}(\vec{r})d\vec{r } + o(\delta^2\rho ) \nonumber \end{aligned}\ ] ] if also @xmath11 is small i.e. @xmath2 , + @xmath154-t_s^{nad}[\rho_a,\rho_b ] = % \int \left.\frac{\delta t_s[\rho]}{\delta \rho}(\vec{r})\right|_{\rho=\rho_b } \delta{\rho}(\vec{r})d\vec{r } + o(\delta^2\rho ) \nonumber \\ \end{aligned}\ ] ] therefore , @xmath155}{\delta \rho}(\vec{r})\right|_{\rho\longrightarrow 0 } \approx \left.\frac{\delta t_{s}[\rho]}{\delta \rho}(\vec{r})\right|_{\rho\longrightarrow \rho_b } \label{approx1}\\ \nonumber\end{aligned}\ ] ] the above result that the kinetic - energy component of @xmath156 is just the functional derivative of the @xmath4 $ ] calculated for @xmath157 makes it possible to express it analytically for any @xmath7 which comprises just two electrons . for one - electron and two - electron- spin - compensated systems the exact expression reads @xcite : @xmath158=t^w_{s}[\rho]=\int \frac{1}{8}\frac{\nabla \vert\rho\vert^2}{\rho}d\vec{r}\;\;\mathrm{for}\;\;\int\rho d\vec{r}=2\end{aligned}\ ] ]
therefore , @xmath159}{\delta
\rho}\right|_{\rho=\rho_b}=\frac{1}{8}\frac{\vert \nabla \rho_b\vert^2}{\rho_b^2 } - \frac{1}{4}\frac{\nabla^2\rho_b}{\rho_b}\;\;\mathrm{if}\;\;\int\rho_b d\vec{r}=2 \label{derivativew } \\
\nonumber\end{aligned}\ ] ] using eq . [ derivativew ] in eq .
[ approx1 ] leads to the asymptotic form of the kinetic energy component of @xmath156 in the case where @xmath11 and @xmath7 do not overlap significantly and @xmath7 is a two - electron spin - less electron density reads : @xmath155}{\delta \rho}\right|_{\rho=\rho_a\longrightarrow 0,\int\rho_b d\vec{r}=2 } = \frac{1}{8}\frac{\vert \nabla \rho_b\vert^2}{\rho_b^2 } - \frac{1}{4}\frac{\nabla^2\rho_b}{\rho_b } \label{assymptotic_vtnad}\end{aligned}\ ] ] @xmath7 representing a doubly occupied hydrogenic @xmath80 function ( @xmath160 ) reads : @xmath161 for @xmath162 , eq . [ assymptotic_vtnad ] leads to the following potential : @xmath155}{\delta \rho}\left(\vec{r}\right)\right|_{\rho=\rho_a\longrightarrow 0,\rho_b = \rho_b^{1s}}=\frac{\zeta}{r } - \frac{\zeta^2}{2}\label{exactpotential } \\
\nonumber\end{aligned}\ ] ] the potential given in eq .
[ exactpotential ] is repulsive for @xmath163 . for hydrogenic densities @xmath162
, the reduced density gradient equals to a @xmath60-independent value of 1.426 .
near the nuclear cusp , therefore , a pair of electrons on the @xmath80 shell provides a local repulsive potential which compensates the coulomb attraction due to the nuclear charge .
note that the effective nuclear charge @xmath60 for the most tightly bound orbital ( @xmath80 ) in a multi - electron atom is smaller than the charge of the corresponding nucleus ( @xmath59 ) @xcite . as a consequence ,
the compensation is perfect only for one - electron hydrogenic systems .
we consider numerical solutions of the schrdinger equation for one electron in the spherically symmetric potential which takes the general from given in eqs . [ kscedpot1]-[kscedembpot ] .
the analysed potential corresponds to @xmath164=0 $ ] , @xmath165 , and @xmath166 $ ] defined in eq .
[ kscedembpot ] for @xmath2 and @xmath167 .
such a case represents a local potential around a nucleus with the charge @xmath59 localised in the environment far from from the investigated subsystem @xmath13 and @xmath7 comprising entire contribution from doubly occupied @xmath80 shell centred on this nucleus .
dirac s exchange energy expression @xcite is used to derive the exchange - correlation component of @xmath166 $ ] .
the solutions of one - electron schrdinger equation for such potential is obtained by radial quadrature described in refs .
@xcite implemented numerically using matlab2006 environment @xcite .
finite difference approximation and the matrix representation with 127 radial points is used .
the above model system is used to investigate the effect of approximations to @xmath0(r)$ ] on the lowest energy level .
using such approximate potential @xmath26(r)$ ] , which is finite at nuclear cusp , might lead to appearance of artificially stabilised bound states due to improper balance between the nuclear attraction , classical electron - electron repulsion , which are both descried exactly and improperly behaving @xmath26(r)$ ] .
this imbalance does not cause qualitative problems in the z=2 ( helium atom ) case .
no bound states occur if @xmath0(r)$ ] is approximated by : @xmath168}{\delta\rho}\right|_{\rho=\rho_a\longrightarrow 0 } \label{vt_model1}\\ & = & \frac{5}{3}c_{tf } \rho_b^{2/3}\nonumber % % % % \end{aligned}\ ] ] derived from the uniform - electron gas expression for @xmath1 given in eq .
[ vtsnad_gea0 ] . for @xmath169 ( i.e. li@xmath31 )
, however , @xmath170 leads to a bound state of the energy -0.2096654 hartree . this value will be used as a reference for analysis of other approximations to @xmath0(\vec{r})$ ] . +
the expression derived from the second - order gradient approximation to @xmath171}{\delta\rho_a}$ ] reads : + @xmath172 } { \delta \rho}\right|_{\rho=\rho_a\longrightarrow 0}\label{vtnad_gea2 } \\ & = & \frac{5}{3 } c_{tf } \rho_b^{2/3 } -\frac{1}{72 } \frac{\nabla^2\rho_b}{\rho_b } + \frac{1}{144 } \frac { \vert \nabla \rho_b \vert^2 } { \rho_b^2 } \nonumber \\ \nonumber\end{aligned}\ ] ] leads to a further lowering of the energy to -0.611935 hartree . in the following part , other potentials will be considered , for which a local non - decomposable contribution is added to that given in eq .
[ vt_model1 ] near the nucleus .
@xmath173 figure [ levels_sb ] shows that , compared to the decomposable result , the addition of a non - decomposable component destabilises the energy but only for such small values of @xmath56 at which @xmath174 .
the maximal destabilisation occurs at @xmath113 in line with the change of sign of the added term .
a similar picture emerges from the analysis of the electron density at @xmath120 .
comparisons of figures [ levels_sb ] and [ ontopdensity_sb ] reveals that addition of the non - decomposable term to the effective potential affects the orbital energies and electron density in a different manner .
even very close to the nucleus this addition reduced electron density without affecting the orbital energy noticeably . from the point of view of choosing @xmath101 determining the range of this additional potential it is worthwhile to notice that the limit @xmath175 should not be exceeded .
the energy level starts a rapid descend and the on - top density starts to rise again at larger values of @xmath176 . as far as the lower limit for @xmath110
is concerned , it should not be smaller than about 0.5 because points at which @xmath177 influence significantly the charge density at the nucleus . .
total dipole moment ( @xmath178 in debye ) obtained from the _ freeze - and - thaw _ calculations using two approximations @xmath0(\vec{r})$ ] . the target
kohn - sham results are given for comparison .
only dipole moment components along the principal axis connecting the subsystems are given .
relative percentage errors of the dipole moments derived from subsystem based calculations ( @xmath179 ) are given in parentheses . [ cols="^,^,>,>,>",options="header " , ] | a new approach to approximate the kinetic - energy - functional dependent component ( @xmath0(\vec{r})$ ] ) of the effective potential in one - electron equations for orbitals embedded in a frozen density environment ( eqs .
20 - 21 in [ wesolowski and warshel ,
_ j. phys .
chem . _ * 97 * , ( 1993 ) 8050 ] ) is proposed .
the exact limit for @xmath1 at @xmath2 and @xmath3 is enforced .
the significance of this limit is analysed formally and numerically for model systems including a numerically solvable model and real cases where @xmath3 . a simple approximation to @xmath0(\vec{r})$ ]
is constructed which enforces the considered limit near nuclei in the environment .
numerical examples are provided to illustrate the numerical significance of the considered limit for real systems - intermolecular complexes comprising , non - polar , polar , charged constituents .
imposing the limit improves significantly the quality of the approximation to @xmath0(\vec{r})$ ] for systems comprising charged components . for complexes comprising neutral molecules or atoms
the improvement occurs as well but it is numerically insignificant . |
the quantum ising chain is the basic model for quantum phase transitions at zero temperature @xcite : the pure chain is exactly soluble @xcite and is the same universality class as the two - dimensional classical ising model via the classical - quantum correspondence @xcite ; the random chain is also exactly soluble via the ma - dasgupta - fisher strong disorder renormalization approach @xcite and is the prototype of infinite disorder fixed points with unconventional scaling laws with respect to the pure case ( see @xcite for a review ) .
remarkably , for an open ising chain containing @xmath1 quantum spins with arbitrary transverse fields @xmath12 and arbitrary couplings @xmath13 @xmath14 there exists an explicit exact formula for the surface magnetization at one end @xmath15 when the spin at the other end @xmath16 is fixed @xmath17 @xcite @xmath18^{-\frac{1}{2 } } \nonumber \\ & & = \left [ 1 + \frac{h^2(n)}{j^{2}(n-1 ) } + \frac{h^2(n)}{j^{2}(n-1 ) } \frac{h^2(n-1)}{j^{2}(n-2 ) } + ... + \frac{h^2(n)}{j^{2}(n-1 ) } \frac{h^2(n-1)}{j^{2}(n-2 ) } ... \frac{h^2(2)}{j^{2}(1 ) } \right]^{-\frac{1}{2 } } \label{msurfzexact}\end{aligned}\ ] ] for the disordered chain where the transverse fields @xmath19 and the couplings @xmath13 are random , the surface magnetization thus involves the specific structure of kesten random variables satisfying simple multiplicative recurrence @xcite .
kesten random variables also appear either directly in many other discrete disordered systems @xcite , or via their continuous analogs given by exponential functionals of brownian motion @xcite .
the probability distribution of the surface magnetization over disordered samples has thus been analyzed in detail @xcite .
unfortunately , other interesting observables are not given by explicit exact formulas in terms of the initial disorder variables , but various proposals have been made to relate them to the known surface magnetizations : \(i ) for the chain of eq .
[ h1chain ] with free boundary conditions , the end - to - end correlation @xmath20 whose statistics has been computed via the appropriate extension @xcite of the strong disorder renormalization @xcite , was found numerically to be very close to the product of the two surface magnetizations @xcite @xmath21 \(ii ) the energy gap between the ground state and the first excited state of the chain of eq .
[ h1chain ] with free boundary conditions @xmath22 whose statistics has also been studied via strong disorder renormalization @xcite , was found to be related to the end - to - end correlation of eq .
[ corre ] via @xmath23 as a consequence of the structure of the strong disorder renormalization rules @xcite . on the other hand , using the free - fermion approach , the gap was found to be related to the product of the two surface magnetizations @xmath24 provided @xmath25 decays faster than @xmath26 @xcite .
eq [ gapcorre ] and eq .
[ gapsurf ] are equivalent if one takes into account eq .
[ corremsurf ] . in this paper
, we obtain via real - space renormalization some generalizations of these formula for star junctions of @xmath11 quantum ising chains , and we use them to analyze the properties of the energy gap .
more precisely , we consider @xmath27 quantum ising chains containing @xmath1 spins @xmath28 each with @xmath29 and @xmath30 @xmath31 \label{hmchains}\end{aligned}\ ] ] when they are coupled via a star junction involving an additional central spin @xmath32 @xmath33 or when they are coupled via two star junctions involving two additional spins @xmath32 and @xmath34 leading to the watermelon geometry @xmath35 the star junction of eq . [ hstar ] with random transverse fields and random couplings has been studied recently via strong disorder renormalization @xcite to obtain the disorder - averaged magnetization at the center of the junction @xmath36 in the thermodynamic limit of semi - infinite chains @xmath37 .
pure quantum ising chains with uniform transverse field @xmath38 and uniform couplings @xmath39 when they are coupled via a delta - junction @xmath40 have also been studied in the limits of weak coupling @xmath41 @xcite and of strong coupling @xmath42 @xcite . the delta junction of eq .
[ hdelta ] is expected to have the same physical properties as the star junction of eq .
[ hstar ] . in @xcite ,
the case of two delta junctions at the two ends has been also considered and is expected to be similar to the watermelon geometry of eq . [ hwatermelon ] involving instead two star - junctions .
let us finish this introduction by mentioning that other quantum models have been studied on the star - junction geometry @xcite .
the paper is organized as follows . in section [ sec_rg ]
, we introduce a real - space renormalization procedure for the star junction that can be explicitly solved in terms of the surface magnetizations of eq .
[ msurfzexact ] and other related kesten variables , and we obtain generalizations of eq . [ corremsurf ] for the correlation between the end spins , and of eq .
[ gapsurf ] for the energy gap . in section [ sec_pure ] , we describe the results for the pure star junction , and in particular the dynamical exponent @xmath43 at criticality . in section [ sec_random ]
, we discuss the results for the disordered star junction governed by an infinite disorder fixed point .
section [ sec_watermelon ] is devoted to the watermelon geometry .
our conclusions are summarized in section [ sec_conclusion ] .
the fernandez - pacheco self - dual real - space renormalization introduced initially for the pure quantum ising chain @xcite has been recently extended to the disordered chain @xcite . as in any block renormalization procedure
, the renormalization rules are applied _ in parallel _ to all blocks of size @xmath44 . here
we propose to apply the fernandez - pacheco self - dual renormalization rules _ sequentially _ around the central site of the junction of eq .
[ hstar ] .
chains of length @xmath1 : + top : initial hamiltonian of eq .
[ hstar ] with the transverse fields @xmath45 and the couplings @xmath46 + bottom : renormalized hamiltonian of eq .
[ hstarrg ] after @xmath47 elementary rg steps , where the renormalized center is characterized by the renormalized transverse field @xmath48 and the renormalized couplings @xmath49 .
, title="fig:",height=355 ] chains of length @xmath1 : + top : initial hamiltonian of eq .
[ hstar ] with the transverse fields @xmath45 and the couplings @xmath46 + bottom : renormalized hamiltonian of eq .
[ hstarrg ] after @xmath47 elementary rg steps , where the renormalized center is characterized by the renormalized transverse field @xmath48 and the renormalized couplings @xmath49 .
, title="fig:",height=355 ] more precisely , after @xmath47 rg steps , we consider that the initial hamiltonian of eq . [ hstar ]
has been renormalized into ( see fig .
[ figstar ] ) @xmath50 \label{hstarrg}\end{aligned}\ ] ] where all the spins @xmath51 of the @xmath11 chains have kept their initial transverse fields @xmath52 and initial couplings @xmath53 of eq .
[ hmchains ] , and where the central renormalized spin @xmath54 with its renormalized transverse field @xmath48 and its renormalized couplings @xmath55 replaces the central @xmath56 initial spins , namely the initial central spin and the initial first @xmath47 spins of the @xmath11 chains .
the initial hamiltonian of eq .
[ hstar ] corresponds to @xmath57 so that the initial condition of the rg flow reads @xmath58 to go from @xmath59 to @xmath60 , we decompose @xmath59 into the sum of two contributions @xmath61 \nonumber \\ v & & = - h^{r^{n-1}}(0)\sigma^x(r^{n-1}(0 ) ) + \sum_{p=1}^m \left[- \sum_{i = n+1}^n h_p(i ) \sigma^x_p(i ) - \sum_{i = n}^{n-1 } j_p(i ) \sigma_p^z(i ) \sigma^z_p(i+1 ) \right ] \label{hstarrgsum}\end{aligned}\ ] ] the hamiltonian @xmath62 is diagonal in @xmath63 , and for each value @xmath64 , one has to diagonalize independently the single - spin effective hamiltonian for @xmath65 .
we refer to @xcite for more details and simply state the final output : the projection of the complementary hamiltonian @xmath66 onto the two degenerate ground states of @xmath67 leads to the following rg rules @xcite for the central transverse field @xmath68 and for its couplings @xmath69 it is convenient to rewrite eq [ rgj ] as @xmath70 \label{rgjsrb}\end{aligned}\ ] ] in order to obtain the simple iteration @xmath71= ... \nonumber \\ & & = \frac{1 } { j_p^2(n ) } \left[1 + \frac{h^2_{p}(n)}{j_p^2(n-1 ) } + \frac{h^2_{p}(n)}{j_p^2(n-1 ) } \frac{h^2_{p}(n-1)}{j_p^2(n-2 ) } + ... + \frac{h^2_{p}(n)}{j_p^2(n-1 ) } \frac{h^2_{p}(n-1)}{j_p^2(n-2 ) } ...
\frac{h^2_{p}(2)}{j_p^2(1 ) } \frac{h^2_{p}(1)}{j_p^2(0 ) } \right ] \nonumber \\ & & = \frac{1 } { j_p^2(n ) } \left[1 + \sum_{i=1}^n \prod_{k = i}^n \frac{h_p^2(k)}{j_p^{2}(k-1 ) } \right ] \label{jrsol}\end{aligned}\ ] ] so that the final result reads @xmath72^{-\frac{1}{2 } } \label{jrsolfinal}\end{aligned}\ ] ] the parenthesis on the right hand side coincides with the surface magnetization @xmath73 at @xmath74 of the @xmath75 chain when the magnetization at the center @xmath76 is fixed ( eq . [ msurfzexact ] ) @xmath77^{-\frac{1}{2 } } \label{mz0p}\end{aligned}\ ] ] so the solution obtained in eq .
[ jrsolfinal ] reads @xmath78 note that this formula also suggests that the sequential rg that we have described is actually somewhat equivalent after @xmath47 steps to the large - block self - dual procedure involving a block a size @xmath47 studied in ref @xcite .
it is convenient to consider the ratio obtained from eqs [ rgh ] and [ rgj ] @xmath79 \frac{h^{r^{n-1}}(0 ) } { \prod_{p=1}^m j_p^{r^{n-1}}(0 ) } \label{rgratiosr}\end{aligned}\ ] ] in order to obtain the simple iteration @xmath79 \left [ \prod_{p=1}^m \frac{h_p(n-1 ) } { j_p(n-1 ) } \right ] \frac{h^{r^{n-2}}(0 ) } { \prod_{p=1}^m j_p^{r^{n-2}}(0 ) } = ... \nonumber \\ & & = \left [ \prod_{p=1}^m \frac{h_p(n ) } { j_p(n ) } \right ] \left [ \prod_{p=1}^m \frac{h_p(n-1 ) } { j_p(n-1 ) } \right ] ...
\left [ \prod_{p=1}^m \frac{h_p(1 ) } { j_p(1 ) } \right ] \frac{h(0 ) } { \prod_{p=1}^m j_p(0 ) } \label{rghsr}\end{aligned}\ ] ] the solution obtained in eq [ jmz ] for the renormalized couplings in terms of the surface magnetizations of eq . [ mz0p ] yields for the renormalized transverse field @xmath80 \left ( \frac{h(0 )
\prod_{p=1}^m \prod_{i=1}^n h_p(i)}{\prod_{p=1}^m \prod_{i=0}^{n-1 } j_p(i ) } \right ) \label{hmz}\end{aligned}\ ] ] using the explicit form of the surface magnetization of eq
. [ mz0p ] , this can be also rewritten as @xmath81^{-\frac{1}{2 } } \right ) \nonumber \\ & & = h(0 ) \prod_{p=1}^m \left[\frac { j_p^2(0 )
j_p^2(1 ) ... j_p^2(n-1)}{h_p^2(1 ) h_p^2(2 ) ...
h_p^2(n ) } + \frac { j_p^2(0 ) j_p^2(1 ) ...
j_p^2(n-2)}{h_p^2(1 ) h_p^2(2 ) ... h_p^2(n-1 ) } + ... + \frac{j_p^2(0 ) } { h^2_{p}(1 ) } + 1 \right]^{-\frac{1}{2 } } \nonumber \\ & & = h(0 ) \prod_{p=1}^m \left[1 + \sum_{i=1}^{n } \prod_{k=1}^i \frac{j_p^2(k-1 ) } { h^2_{p}(k ) } \right]^{-\frac{1}{2 } } \label{rghsrfinal}\end{aligned}\ ] ] for the finite star junction of eq .
[ hstar ] , the renormalization procedure described above ends after @xmath1 rg steps where the renormalized hamiltonian of eq .
[ hstarrg ] contains a single renormalized spin @xmath82 representing the whole star junction @xmath83 the remaining renormalized transverse field @xmath84 corresponds to the energy scale associated to the global flip of the junction and thus represents the energy gap ( up to an unimportant factor of @xmath85 ) @xmath86 eq . [ hmz ] yields @xmath87 \left ( \frac{h(0 ) \prod_{p=1}^m
\prod_{i=1}^n h_p(i)}{\prod_{p=1}^m \prod_{i=0}^{n-1 } j_p(i ) } \right ) \label{gapmz}\end{aligned}\ ] ] this formula is thus the analog of eq . [ gapsurf ] : the numerator contains the product of the @xmath11 surface magnetizations and the product of all the transverse fields , whereas the denominator involves the product of all the couplings . the alternative form of eq .
[ rghsrfinal ] reads @xmath88^{-\frac{1}{2 } } \label{gapexpli}\end{aligned}\ ] ] before the last hamiltonian of eq .
[ hstarrgfinal ] , the renormalized hamiltonian at step @xmath89 ( eq . [ hstarrg ] ) @xmath90 involves the initial end - spins @xmath91 of the @xmath11 chains , and the renormalized central spin @xmath92 representing the rest of the sample .
the eigenfunction corresponding to these two last rg steps reads ( see @xcite for more details on the eigenfunctions of @xmath62 of eq .
[ hstarrgsum ] ) @xmath93 \label{psilast}\end{aligned}\ ] ] the joint distribution of the @xmath11 end - spins @xmath94 is thus obtained after the integration of the renormalized central spin @xmath95 \nonumber \end{aligned}\ ] ] in particular , the correlation between the @xmath11 end - spins reduces to @xmath96 and can be thus rewritten using eqs [ mz0p ] and [ jmz ] as the product of the @xmath11 surface magnetizations when the center is fixed @xmath97 this formula is thus somewhat the analog of eq .
[ corremsurf ] .
for the pure case , the control parameter of a single chain is the ratio between the ferromagnetic coupling @xmath98 and the transverse field @xmath99 @xmath100 the exact surface magnetization of eq .
[ msurfzexact ] reads @xmath101^{\frac{1}{2 } } \label{msurfpure}\end{aligned}\ ] ] in the ferromagnetic phase @xmath102 , the surface - magnetization of eq .
[ msurfpure ] converges towards the finite limit @xmath103^{\frac{1}{2 } } \propto_{k \to 1^+ } ( k-1)^{\beta_s } \ \ \ { \rm with } \ \
\beta_s=1/2 \label{msurfpureferro}\end{aligned}\ ] ] as a consequence the renormalized coupling of eq .
[ jmz ] also converges towards a finite limit @xmath104^{\frac{1}{2 } } \label{jmzpureferro}\end{aligned}\ ] ] whereas the renormalized transverse field of eq .
[ rghsrfinal ] converges exponentially towards zero as @xmath105^{-\frac{m}{2 } } \oppropto_{n \to + \infty } k^{- n m } \label{gapexplipureferro}\end{aligned}\ ] ] in the paramagnetic phase @xmath106 , the surface magnetization of eq .
[ msurfpure ] decays exponentially with the distance @xmath107^{\frac{1}{2 } } \oppropto_{n \to + \infty } k^n = e^{n \ln ( k ) } \label{msurfpurepara}\end{aligned}\ ] ] and involves the correlation length exponent @xmath108 . as a consequence , the renormalized coupling of eq .
[ jmz ] behaves similarly @xmath109 whereas the renormalized transverse field of eq .
[ rghsrfinal ] converges to the finite limit @xmath110^{-\frac{m}{2 } } \oppropto_{n \to + \infty } h(1-k^2)^{\frac{m}{2 } } \label{hmxpara}\end{aligned}\ ] ] at the critical point @xmath111 , the surface magnetization of eq .
[ msurfpure ] decays as the power - law @xmath112 as a consequence , the renormalized coupling of eq .
[ jmz ] decays with the same exponent @xmath113 whereas the renormalized transverse field of eq .
[ rghsrfinal ] representing the gap decays as the power - law @xmath114 with the dynamical exponent @xmath115 for @xmath4 corresponding to a single open chain with free boundary conditions , this result reduces to @xmath116 as it should . for @xmath117 , eq . [ zm ]
is in agreement with the strong - coupling perturbative expansion of eq .
( 12 ) of reference @xcite .
for the disordered junction , the usual control parameter @xcite @xmath118 valid for gaussian distributions can be more precisely defined as the root of the equation @xcite @xmath119 that makes the link with the literature on kesten variables @xcite .
it is also usual in the field to redefine the length scale to absorb the variance @xcite , in order to have the single parameter @xmath120 .
let us briefly summarize the known results on the statistics of the surface magnetization of random chains @xcite in the universal critical region around @xmath121 : \(i ) in the paramagnetic phase @xmath122 , the surface magnetization reads @xcite @xmath123 where @xmath124 is a zero - mean random variable distributed with the gaussian distribution @xmath125 in particular , the typical value @xmath126 involves the typical correlation length exponent @xmath127 , whereas the finite - size scaling properties determined by the balance between the two terms of eq .
[ mspara ] are governed by @xmath128 .
\(ii ) at criticality @xmath121 , the surface magnetization reads @xcite @xmath129 where @xmath130 is a positive random variable distributed with the half - gaussian distribution @xmath131 \(iii ) in the ferromagnetic phase @xmath132 , the surface magnetization reads @xcite @xmath133 where @xmath134 is a positive random variable distributed with the exponential distribution @xmath135 from these known results concerning the surface magnetization of a single random chain , we may now derive the statistical properties of the renormalized parameters for the random star junction . at the critical point @xmath121
, the renormalized coupling @xmath136 behaves as the surface magnetization of eq .
[ mscriti ] @xmath137 the gap @xmath138 given by the renormalized transverse field will be also of the form @xmath139 and thus involves the same activated exponent @xmath140 as the chain corresponding corresponding to the special case @xmath4 .
the variable @xmath141 in eq .
[ gapcritirandom ] is distributed with the convolution of @xmath11 half - gaussian distributions ( eq . [ wcriti ] ) @xmath142 \delta\left(g- \sum_{p=1}^m w_p \right ) \label{halfgaussconvol}\end{aligned}\ ] ] however this expression is expected to be valid only for sufficiently large @xmath9 , whereas the region of small @xmath9 requires a more precise analysis , as shown by the strong disorder calculations of @xcite concerning the case @xmath4 of the chain . in the ferromagnetic phase @xmath143 ,
the renormalized coupling @xmath136 remains finite as the surface magnetization of eq .
[ msferro ] @xmath144 whereas the renormalized transverse field will behave as @xmath145 where @xmath124 is a gaussian random variable distributed with eq .
[ vpara ] . in the paramagnetic phase @xmath122 , the renormalized coupling inherits from the behavior of eq .
[ mspara ] of the surface magnetization @xmath146 whereas the renormalized transverse field remains finite @xmath147 where @xmath148 is distributed with the convolution of @xmath11 exponential distributions of eq .
[ bferro ] @xmath149
\delta\left(b- \sum_{p=1}^m b_p \right ) = \frac { \mu^m b^{m-1 } } { ( m-1 ) ! } e^{- \mu b } \label{expconvol}\end{aligned}\ ] ]
in this section , we use the results of the previous sections concerning the single star junction with free boundary conditions ( eq . [ hstar ] ) to analyze the watermelon geometry of eq .
[ hwatermelon ] where the @xmath11 chains are coupled by two star junctions at the two ends .
we apply the same renormalization rules as in section [ sec_rg ] around each of the two junctions , so that after @xmath150 rg steps , the renormalized hamiltonian reads @xmath151 the renormalized transverse fields of the two centers of the junctions are obtained from eq .
[ rghsrfinal ] @xmath152^{-\frac{1}{2 } } \nonumber \\ & & = h(0 ) \prod_{p=1}^m \left[1 + \sum_{i=1}^{{\frac{n}{2 } } } \prod_{k=1}^i \frac{j_p^2(k-1 ) } { h^2_{p}(k ) } \right]^{-\frac{1}{2 } } \label{rgh0w}\end{aligned}\ ] ] and @xmath153^{-\frac{1}{2 } } \nonumber \\ & & = h(n+1 ) \prod_{p=1}^m \left[1 + \sum_{i=\frac{n}{2}+1}^n \prod_{k = i}^n \frac{j_p^2(k ) } { h^2_{p}(k ) } \right]^{-\frac{1}{2 } } \label{rgh0np1}\end{aligned}\ ] ] whereas their coupling is obtained by summing the contribution of the @xmath11 chains @xmath154 with @xmath155^{-\frac{1}{2 } } \nonumber \\ & & = \left [ 1 + \frac{h_p^2(\frac{n}{2})}{j_p^{2}(\frac{n}{2}-1 ) } + \frac{h_p^2(\frac{n}{2})}{j_p^{2}(\frac{n}{2}-1 ) } \frac{h_p^2(\frac{n}{2}-1)}{j_p^{2}(\frac{n}{2}-2 ) } + ... + \frac{h_p^2(\frac{n}{2})}{j_p^{2}(\frac{n}{2}-1 ) } \frac{h_p^2(\frac{n}{2}-1)}{j_p^{2}(\frac{n}{2}-2 ) } ... \frac{h_p^2(1)}{j_p^{2}(0 ) } \right]^{-\frac{1}{2 } } \label{msurf1}\end{aligned}\ ] ] and @xmath156^{-\frac{1}{2 } } \nonumber \\ & & = \left [ 1 + \frac{h_p^2(\frac{n}{2}+1)}{j_p^{2}(\frac{n}{2}+1 ) } + \frac{h_p^2(\frac{n}{2}+1)}{j_p^{2}(\frac{n}{2}+1 ) } \frac{h_p^2(\frac{n}{2}+2)}{j_p^{2}(\frac{n}{2}+2 ) } + ... + \frac{h_p^2(\frac{n}{2})+1}{j_p^{2}(\frac{n}{2}+1 ) } \frac{h_p^2(\frac{n}{2}+2)}{j_p^{2}(\frac{n}{2}+2 ) } ... \frac{h_p^2(n)}{j_p^{2}(n ) } \right]^{-\frac{1}{2 } } \label{msurf2}\end{aligned}\ ] ] for the pure case , where the surface magnetization is given by eq .
[ msurfpure ] in terms of the control parameter @xmath157 of eq .
[ kpur ] , the renormalized coupling of eq .
[ jwatermelonr ] reads @xmath158 whereas the transverse fields of eqs [ rgh0w ] and [ rgh0np1 ] have the same value @xmath159 in the ferromagnetic phase @xmath102 , the renormalized coupling of eq .
[ jwatermelonrpure ] converges towards the finite limit @xmath160 and the renormalized transverse field of eq .
[ rgh0wpure ] converges exponentially towards zero as @xmath161 in the paramagnetic phase @xmath106 , the renormalized coupling of eq .
[ jwatermelonrpure ] decays exponentially with the distance @xmath162 and the renormalized transverse field of eq .
[ rghsrfinal ] converges to the finite limit @xmath163 at the critical point @xmath111 , one obtains the following power - laws for the renormalized coupling of eq .
[ jwatermelonrpure ] @xmath164 and for the renormalized transverse field of eq .
[ rghsrfinal ] @xmath165 for @xmath4 corresponding to a single chain with periodic boundary conditions , both involve the same exponent , that represents the dynamical exponent @xmath116 as it should .
however for @xmath117 , the coupling of eq .
[ jwpurecriti ] dominates over the transverse fields of eq .
[ hwpurecriti ] , so that the last decimation corresponds to the projection onto the two ferromagnetic states @xmath166 and @xmath167 defining the final renormalized spin representing the whole watermelon , and its renormalized transverse field @xmath168 represents the gap @xmath169 with the dynamical exponent @xmath170 for @xmath117 , the dynamical exponent of eq .
[ zwm ] is bigger than the dynamical exponent of a single star junction of eq .
[ zm ] .
the statistical properties of the renormalized transverse fields of eqs [ rgh0w ] and [ rgh0np1 ] are the same as in the star junction case given in eqs [ gapcritirandom ] , [ hrferrorandom ] and [ hrpararandom ] with the replacement @xmath171 .
the statistical properties of the renormalized coupling of eq .
[ jwatermelonr ] can be obtained from the properties of the surface magnetization recalled in section [ subsec_surf ] . in the paramagnetic phase @xmath122
, it decays as @xmath172 where @xmath124 is the minimum value among @xmath11 variables drawn with the gaussian distribution of eq .
[ vpara ] . in the ferromagnetic phase @xmath132 , it remains finite @xmath173 where @xmath134 is minimum among @xmath11 positives variables drawn with the convolution @xmath174 of the exponential distribution of eq .
[ bferro ] . at criticality @xmath121 , it decays as @xmath175 where @xmath130 is the minimum value among @xmath11 positives variables drawn with the half - gaussian distribution of eq .
[ wcriti ] .
in particular , the gap of the watermelon involves the same activated exponent @xmath176 and it is only the scaling distribution of the gap that depends upon @xmath11 .
for @xmath0 pure or random quantum ising chains of @xmath1 spins coupled via a single star junction or via two star junctions at the two ends , we have introduced a sequential self - dual real - space renormalization procedure that can be explicitly solved in terms of kesten variables containing the initial couplings and and the initial transverse fields .
we have described the properties of the renormalized couplings and renormalized transverse fields in the ferromagnetic phase , in the paramagnetic phase and at the critical point . in particular , we have discussed the finite - size properties of the energy gap @xmath177 at criticality : \(i ) in the pure case , the dynamical exponent @xmath178 governing the power - law decay of the gap @xmath2 was found to be @xmath3 for the single star junction and @xmath6 for the watermelon geometry ( for the case @xmath4 corresponding to a single chain with open or periodic boundary conditions , both dynamical exponents reduce to @xmath116 as it should ) .
\(ii ) in the random case , the gap was found to follow the infinite disorder fixed point scaling @xmath7 with the same activated exponent @xmath8 as the single chain corresponding to @xmath4 , so that the number @xmath11 of chains and the geometry ( star or watermelon ) only enter in the probability distribution of the @xmath10 random positive variable @xmath9 . as a final remark ,
let us mention some link with the recent studies concerning the shannon and rnyi entropies of the ground state wavefunction of the pure quantum ising chain @xcite .
the rnyi entropy of index @xmath179 of the quantum chain amounts to study the binding of the ising book ( see fig . 4 of @xcite ) where @xmath180 half - planes of classical two - dimensional ising models are glued together . taking the anisotropic limit in the other direction leads to the pure quantum star junction discussed in section [ sec_pure ] .
although the present study does not deal with observables that are directly relevant for the rnyi entropy , we feel that the expansion of the dynamical exponent @xmath181 for @xmath182 around the ising plane @xmath183 @xmath184 suggests that the appropriate finite - size scaling variable could be @xmath185 with @xmath186 for the periodic quantum chain , grgoire misguich and jean - marie stphan have replotted their numerical data showned on fig .
5 of their paper @xcite in terms of the variable @xmath187 and have concluded that this finite size scaling could be also compatible with their data @xcite . for the open quantum chain ,
the double logarithm @xmath188 found numerically at criticality for @xmath183 by jean - marie stphan @xcite and the singular logarithmic terms in @xmath189 found at criticality for @xmath190 via conformal field theory in ref @xcite also points towards the finite size scaling variable of eq .
it is a pleasure to thank grgoire misguich and jean - marie stphan for very useful discussions and for replotting their numerical data @xcite in terms of the variable of eq . | we consider @xmath0 pure or random quantum ising chains of @xmath1 spins when they are coupled via a single star junction at their origins or when they are coupled via two star junctions at the their two ends leading to the watermelon geometry .
the energy gap is studied via a sequential self - dual real - space renormalization procedure that can be explicitly solved in terms of kesten variables containing the initial couplings and and the initial transverse fields . in the pure case at criticality
, the gap is found to decay as a power - law @xmath2 with the dynamical exponent @xmath3 for the single star junction ( the case @xmath4 corresponds to @xmath5 for a single chain with free boundary conditions ) and @xmath6 for the watermelon ( the case @xmath4 corresponds to @xmath5 for a single chain with periodic boundary conditions ) . in the random case at criticality
, the gap follows the infinite disorder fixed point scaling @xmath7 with the same activated exponent @xmath8 as the single chain corresponding to @xmath4 , and where @xmath9 is an @xmath10 random positive variable , whose distribution depends upon the number @xmath11 of chains and upon the geometry ( star or watermelon ) .
# 1#2 # 1#2 |
one of the most important tasks for a quantum computer would be to efficiently obtain eigenvalues and eigenvectors of high - dimensional matrices . it has been suggested @xcite that the quantum phase estimation algorithm ( pea ) @xcite can be used to obtain eigenvalues of a hermitian matrix or hamiltonian . for a quantum system with a hamiltonian @xmath0 , a phase factor , which encodes the information of eigenvalues of @xmath0 ,
is generated via unitary evolution @xmath1 . by evaluating the phase , we can obtain the eigenvalues of @xmath0 .
the conventional pea consists of four steps : preparing an initial approximated eigenstate of the hamiltonian @xmath0 , implementing unitary evolution operation , performing the inverse quantum fourier transform ( qft ) , and measuring binary digits of the index qubits .
the pea is at the heart of a variety of quantum algorithms , including shor s factoring algorithm @xcite .
a number of applications of pea have been developed , including generating eigenstates associated with an operator tm , evaluating eigenvalues of differential operators @xcite , and it has been generalized using adaptive measurement theory to achieve a quantum - enhanced measurement precision at the heisenberg limit @xcite .
the pea with delays considering the effects of dynamical phases has also been discussed @xcite . the implementation of an iterative quantum phase estimation algorithm with a single ancillary qubit is suggested as a benchmark for multi - qubit implementations @xcite .
the pea has also been applied in quantum chemistry to obtain eigenenergies of molecular systems @xcite .
this application has been demonstrated in a recent experiment @xcite .
moreover , several proposals have been made to estimate the phase of a quantum circuit @xcite , and the use of phase estimations for various algorithms @xcite , including factoring and searching .
the conventional pea is only designed for finding eigenvalues of either a hermitian or a unitary matrix . in this paper
, we propose a measurement - based phase estimation algorithm ( mpea ) to evaluate eigenvalues of _
non_-hermitian matrices .
this provides a potentially useful generalization of the conventional pea .
our proposal uses ideas from conventional pea , frequent measurement , and techniques in one - qubit state tomography .
this proposal can be used to design quantum algorithms apart from those based on the standard unitary circuit model .
the proposed quantum algorithm is designed for systems with large dimension , when the corresponding classical algorithms for obtaining the eigenvalues of the non - unitary matrices become so expensive that it is impossible to implement on a classical computer .
the structure of this work is as follows : in sec .
[ const ] we introduce how to construct a non - hermitian evolution matrix for a quantum system . in sec .
[ mpea ] , we present the measurement - based phase estimation algorithm , introducing two methods for obtaining the complex eigenvalues of the non - hermitian evolution matrix .
we give two examples for the application of mpea and discuss how to construct hamiltonian for performing the controlled unitary operation in sec .
[ example ] . in sec .
[ diss ] , we discuss the success probability of the algorithm and the efficiency of constructing the non - hermitian matrix .
we close with a conclusion section .
now we describe how to construct non - unitary matrices on a quantum system .
a bipartite system , composed of subsystems _ a _ and _ b _ , evolves under hamiltonian @xmath2where @xmath3 is the hamiltonian of subsystem _ a_(_b _ ) and @xmath4 is their interaction .
we prepare the initial state of subsystem _ a _ in its pure state @xmath5 and the initial state of subsystem _ b _ in an arbitrary state @xmath6 .
then at time @xmath7 , the state of the system is @xmath8 .
let the system evolve under the hamiltonian @xmath0 for a time interval @xmath9 , if subsystem _ a _ is subject to a projective measurement @xmath10 applied at time interval @xmath9 , this is equivalent to driving subsystem _
b _ with an evolution matrix @xmath11this evolution matrix is in general neither unitary nor hermitian . the hamiltonian @xmath0 of the whole quantum system can be spanned as : @xmath12with eigenenergies @xmath13 , and the corresponding eigenvectors @xmath14 } and \{@xmath15 } of hilbert spaces of subsystems _ a _ and _ b _ , which are of dimensions @xmath16 and @xmath17 , respectively , and @xmath18 . using the bases for _
a _ and _ b _ , we have @xmath19and the evolution matrix on subsystem _ b _ , after the measurement @xmath20 performed on subsystem _ a _ at time interval @xmath9 , becomes @xmath21where@xmath22and @xmath23 more generally , we can construct different evolution matrices by performing measurements on subsystem _ a _ with different time intervals and/or different measurement bases .
for example , by sequentially performing projective measurements with time intervals @xmath24 , @xmath25 , @xmath26 , an evolution matrix @xmath27is constructed .
we can also combine unitary evolution matrices with the non - unitary transformations on subsystem _ b _ to construct some desired evolution matrices .
for the bipartite system , set the initial state of the system in a separable state @xmath28and let the system evolve under the hamiltonian in eq .
( @xmath29 ) .
then after performing @xmath30 successful projective measurements on subsystem _ a _ with time intervals @xmath9 , the evolution on the hilbert space of subsystem _ b _ is driven by @xmath31 ^{m}$ ] , and the state of subsystem _ b _ evolves to @xcite@xmath32^{m}\rho _ { b}\bigl[% v_{b}^{\dagger } ( \tau ) \bigr]^{m}}{p^{\left ( \tau \right ) } \left ( m\right ) } , \]]where @xmath33^{m}\rho _ { b}[v_{b}^{\dagger } ( \tau ) ] ^{m}\bigr\}\]]is the probability of finding subsystem _ a _ still in state @xmath34 after each of the @xmath30 measurements .
the evolution matrix @xmath35 can be spanned as @xmath36where @xmath37 and @xmath38 are the right- and left - eigenvectors of @xmath35 and @xmath39 is the corresponding eigenvalue @xcite satisfying @xmath40 . in the large @xmath30 limit ,
the operator @xmath41^{m}$ ] is dominated by a single term @xmath42 , provided the largest eigenvalue @xmath43 is unique , discrete , and non - degenerate . in the limit of large @xmath30 and finite @xmath9
, @xmath44 tends to a pure state , independent of the initial ( mixed ) state of subsystem _
b_. the final state of @xmath45is dominated by @xmath46 , and this outcome is found with probability @xcite @xmath47the state of subsystem _ b _ evolves to @xmath46 , after performing a number of operations of @xmath35
. then we can evaluate @xmath48 by resolving the phase of the state .
if we prepare the initial state of subsystem _ b _ in a pure initial state that is close to an eigenstate of the matrix @xmath35 , the state of the subsystem _ b _ can evolve to other eigenstates of @xmath49
. then we can also obtain the corresponding eigenvalues of @xmath35 .
based on the above analysis , we suggest a measurement - based phase estimation algorithm for evaluation of the eigenvalues of the matrix @xmath35 . as in the circuit shown in fig .
@xmath29(@xmath50 ) , three quantum registers are prepared . from top to bottom : an index register , a target register and an interacting register .
the index register is a single qubit , which is used as control qubit and to readout the final results for the eigenvalues ; the target register is used to represent the state @xmath51 of subsystem _ b _ ; and the interacting register represents the state @xmath52of subsystem _
the initial state of the circuit is prepared in the state@xmath53with subsystem _ a _ in a pure state @xmath54 , and subsystem _ b _ in state @xmath51 .
the construction of the controlled evolution matrix @xmath35 on the target register is achieved by implementing the controlled unitary ( @xmath55-@xmath56 ) transformation for the whole quantum system and successfully performing the projective measurement @xmath57 on the interacting register with time interval @xmath9 .
note here for the unitary transformation @xmath58 , we set @xmath59 such that @xmath30 projective measurements are performed successfully on subsystem @xmath60 at the time interval @xmath9 , while the unitary transformation of the whole system evolves for time period @xmath59 . after performing @xmath30 successful periodic measurements on the interacting register with time intervals @xmath9 , as shown in fig .
@xmath29(@xmath61 ) , the state of the system is transformed to @xmath62^{m}\rho _ { b}\bigl[v_{b}^{\dagger } ( \tau ) % \bigr]^{m}\biggr\ } \nonumber \\ & & \otimes |\varphi _
{ a}\rangle \langle \varphi _ { a}|.\end{aligned}\]]the dominant term of this is @xmath63% qubit is dominated by @xmath64.\]]in general , @xmath43 is a complex number and can be written as @xmath65 .\]]we can obtain @xmath43 by resolving the phase factor @xmath66 two approaches can be used to resolve @xmath43 : @xmath67 using single - qubit quantum state tomography ( qst ) @xcite , and @xmath68 using the measured quantum fourier transform ( mqft ) combined with projective measurements on a single qubit .
the details of these two approaches are given below .
quantum state tomography can fully characterize the quantum state of a particle or particles through a series of measurements in different bases @xcite . in the approach using qst to resolve the eigenvalue of the matrix @xmath35 , we prepare a large number of identical copies of the state on the index qubit @xmath69 , as shown in eq .
( @xmath70 ) , by running the mpea circuit a number of times .
then the value of @xmath48 can be obtained by determining the index qubit state . the state of the index qubit in eq .
( @xmath70 ) can be written as @xmath71 |1\rangle \bigr ) .\]]in the qst approach , we perform a projective measurement on the index qubit in the basis @xmath72 to obtain the probability of finding the index qubit in state @xmath73 , thus obtaining the value of @xmath61 . with the knowledge of @xmath61 , then perform a @xmath74 rotation around the @xmath75-axis and a measurement in the basis of the pauli matrix @xmath76 on the index qubit , we can obtain the observable @xmath77and thus obtain the value of @xmath50 .
the measurement errors of qst , from counting statistics , obey the central limit theorem . to obtain more accurate results
, we have to prepare a larger ensemble of the single qubit states . )
shows the circuit for the mpea using qst . from top to bottom , an index register , a target register and an interacting register are prepared in the states @xmath78 , @xmath79 , and @xmath80 , respectively .
the index register is a single qubit used as control qubit ; the target register is used to represent the state of subsystem _ b _ ; and the interacting register is used to represent the state of subsystem _ a , _ which interacts with _ b_. part ( @xmath61 ) shows the circuit for performing @xmath30 projective measurements with period @xmath81 . in the circuit , the unitary transformation @xmath82 , we set @xmath59 such that @xmath83 projective measurements are performed successfully on subsystem _ a _ separated by a time interval @xmath81 , while the unitary transformation of the whole system evolves for time @xmath59 . ] in the second approach , we use the techniques of measured quantum fourier transform and projective measurements to resolve the eigenvalue of the matrix @xmath35 .
the phases which encode the eigenvalues of @xmath84 are in general complex numbers ; the inverse qft can be used to resolve the real part @xmath50 of the phase @xmath85 .
the imaginary part @xmath61 of the phase factor , @xmath86 , can be obtained by performing single - qubit projective measurements .
the details of this method are discussed below . in order to resolve @xmath50 up to @xmath87 binary digits using the inverse qft
, one has to construct a series of controlled evolution matrices , @xmath55-@xmath88 , in successive binary powers , from @xmath89 to @xmath90 . in the mpea , this is done by implementing the @xmath55-@xmath56 operation on the whole system and performing a series of @xmath91 periodic measurements separated by time intervals @xmath9 for @xmath92 , respectively .
the @xmath55-@xmath56 operation evolves for a time @xmath59 , during which all the measurements are performed successfully on the interacting register .
then we can obtain a series of controlled transformation matrices in binary powers , @xmath55-@xmath93^{2^{k}},$ ] @xmath94 @xmath95 @xmath90 . in fig .
@xmath96(@xmath97 ) , we show the circuit for the @xmath98th projective measurement with period @xmath9 , @xmath99 , where the measurement @xmath100 is performed @xmath91 times with period @xmath9 on the interacting register , while the whole system evolves under the controlled unitary operation @xmath56 . the measurements on the interacting register are sequentially performed @xmath101 .
then , correspondingly , on the index qubit , we obtain single qubit states as @xmath102,\frac{1}{\sqrt{2}}\biggl[|0\rangle + ( \lambda _ { \max } ) ^{2^{(n-2)}}|1\rangle \biggr ] , \nonumber \\ & & \cdots , \frac{1}{\sqrt{2}}\bigl(|0\rangle + \lambda _ { \max } |1\rangle \bigr)% .\end{aligned}\]]afterward , we can retrieve @xmath87 binary digits of the real part @xmath50 of the phase factor @xmath86 , of @xmath43 , by performing a mqft .
the mqft technique implements a qft using only a single qubit @xcite .
it uses the fact that gates within the fourier transform are applied sequentially on qubits .
this modification of the qft algorithm preserves the probabilities of all measurements @xcite . the procedure for obtaining the real part of the phase factor of @xmath43 by using mqft
is shown in fig .
@xmath96(@xmath50 ) , where the circuit in the dotted square is replaced by circuits in the dotted squares shown in fig .
@xmath96(@xmath61 ) , sequentially obtaining @xmath87 binary digits of @xmath50 .
the details of this procedure are shown below .
the initial state of the mpea circuit is prepared as in eq .
( @xmath103 ) .
after performing the @xmath98th periodic measurements @xmath99 on the interacting register , the dominant term of the state of the system becomes @xmath104% can be written as @xmath105 \nonumber \\ & = & \frac{1}{\sqrt{2}}\biggl[|0\rangle + \exp \left ( -b2^{k}\right ) \exp \left ( ia2^{k}\right ) |1\rangle \biggr].\end{aligned}\]]in order to resolve the real part @xmath50 of the phase factor , @xmath86 , we first need to obtain the value of @xmath61 , the imaginary part of the phase factor .
this can be achieved by using a single - qubit projective measurement .
one can prepare an ensemble of @xmath69 and perform projective measurements @xmath72 .
the value of @xmath61 can be obtained through the probability for observing the index qubit in state @xmath106^{1/2},\]]and let us run mpea again and perform a single qubit operation @xmath107 on the index qubit such that the index qubit is rotated to state latexmath:[\[|\psi _
{ \text{ind}}^{\prime } \rangle = r_{k}\frac{1}{\sqrt{2}}\biggl[% single - qubit operation @xmath107 is defined as @xmath109where @xmath110 } $ ] .
then we apply the mqft technique to resolve the real part @xmath50 of the phase factor @xmath85 .
we therefore obtain the eigenvalue @xmath111 $ ] of the matrix @xmath35 . in the mpea ,
the @xmath87th binary digit of the phase factor is retrieved first , and the partial measurement on the interacting register is performed in sequence of @xmath112 to @xmath113 times .
this procedure provides high fidelity for the state of the target register since each measurement drives the state of the target register closer to @xmath46 , the eigenstate of @xmath35 . , @xmath114 , and @xmath115 , respectively .
the index register is a single qubit used as control qubit and to perform mqft ; the target register is used to represent the state of subsystem _ b _ ; and the interacting register is used to represent the state of subsystem _ a , _ which interacts with _
b_. the circuit in the dotted square in ( a ) is used for obtaining the @xmath87th binary digit of the real part of the phase factor . in part
( b ) , a sequence of circuits inside the dashed squares is used to replace the circuit in the dashed square in part ( @xmath50 ) , in order to resolve from the ( @xmath116)th to the first binary digits of the real part of the phase factor . here
, @xmath0 is the hadamard gate , @xmath107 is a single - qubit rotation as defined in eq .
( @xmath117 ) and @xmath118 is a single - qubit rotation in quantum fourier transform .
part ( c ) shows the circuit for the @xmath98th partial measurement @xmath99 with period @xmath81 . in the circuit , for the unitary transformation @xmath82
, we set @xmath59 such that all projective measurements are performed successfully on subsystem _ a _ while the unitary transformation of the whole system evolves for time @xmath59 . ]
now we use a simple model to show how mpea works .
we consider here a quantum system consisting of two subsystems _ a _ and _ b _ , where _ b _ contains two non - interacting spin qubits , _ _ b__@xmath119 and _ _ b__@xmath120 ; and subsystem _ a _ is a photon .
the whole system is described by the jaynes - cummings hamiltonian @xcite @xmath121,\]]where @xmath61 ( @xmath122 ) is a bosonic annihilation ( creation ) operator of the photons .
consider the case @xmath123 , and perform projective measurements in the basis of a single photon state @xmath124 .
then we have @xmath125% /5\times e^{-2iw_{0}\tau } , \nonumber \\ & & \cos ( \sqrt{6}\tau j)e^{-iw_{0}\tau } , \cos ( \sqrt{2}\tau j)\biggr\},\end{aligned}\]]in the ordered basis @xmath126 , where @xmath127 , then we construct an evolution matrix @xmath35 as @xmath128% /5\times e^{-i } , \nonumber \\ & & \cos ( \sqrt{6}/2)e^{-i/2},\cos ( \sqrt{2}/2)\biggr\}.\end{aligned}\ ] ] on the mpea circuit , now let us prepare the target register in a mixed state : @xmath129and the interacting register in state @xmath130 . then implement the controlled hamiltonian of eq .
( @xmath131 ) for a time @xmath59during which the projective measurements on the interacting register are performed successfully . after a series of projective measurements , in the basis of @xmath130 , on the interacting register
, the state of the target register evolves to a singlet state @xmath132 , which corresponds to the largest eigenvalue of @xmath35 .
we resolve the corresponding phase as zero , thus its eigenvalue is one .
the survival probability , @xmath133 , of the state latexmath:[$%
@xmath30 successful measurements , and the fidelity , @xmath135 , for the target register to be in state @xmath46 , are shown in fig .
@xmath136 . the fidelity @xmath135 is defined as@xmath137since @xmath138 is close to one , the success probability is determined by @xmath139 .
( @xmath140 ) and fidelity @xmath141 ( @xmath142 ) for @xmath143 versus @xmath30 , the number of successful measurement , for the jaynes - cummings model . ]
if we prepare the target register in a pure initial state that is close to an eigenstate of the matrix @xmath49 , by applying mpea , the state of the target register can evolve to other eigenstates of @xmath49 . then we can also obtain the corresponding eigenvalues of @xmath49 .
for example , applying mpea to the above system and preparing the target register in state @xmath144 , by performing projective measurements with @xmath130 on the interacting register , the state of the target register would remain in the @xmath144 state .
we can retrieve the real part of the phase factor of the corresponding eigenvalue up to an accuracy of @xmath96 , @xmath145 and @xmath146 binary digits , respectively , and obtain the eigenvalues of the matrices @xmath49 as @xmath147 $ ] , @xmath148 $ ] , and @xmath149 $ ] , assuming we have already obtained the imaginary part of the eigenvalue of @xmath49 through projective measurements .
the true eigenvalue is @xmath150 , which is quite close . to implement a controlled unitary evolution on the mpea circuit , we set the control qubit as a single spin and label it as subsystem _
c_. thus , the controlled hamiltonian of the whole system becomes @xmath151\biggr\}.\end{aligned}\]]this hamiltonian contains three - body interactions and can not be implemented directly .
one could decompose the three - body interaction into two - body interactions @xcite and then implement the two - body interaction . in general , an arbitrary unitary matrix @xmath152 can be decomposed @xcite into tensor products of unitary matrices of @xmath153 and @xmath154 , which correspond to two- and single - qubit operations respectively , and can be implemented on a universal quantum computer . for another example , we consider the axial symmetry model @xcite .
this is relevant for quantum information processing in solid state @xcite and atomic @xcite systems .
the quantum system is composed of two subsystems _ a _ and _ b _ , where _ b _ contains two non - interacting spins , and subsystem _ a _ contains _
_ _ _ a single spin interacting with subsystem _ b_. the hamiltonian for the whole system is @xcite @xmath155,\]]where @xmath156 and @xmath157 are the pauli operators . by performing projective measurements on subsystem
_ a _ in the basis of the @xmath76-eigenvector , then in the basis @xmath158 , we obtain @xmath159operating on subsystem _
b_. if we prepare the initial state of the target register in state @xmath160 , then the fidelity of the target register to be in state @xmath160 is @xmath29 after performing a number of successful measurements on the interacting register .
for the case @xmath161 and @xmath162 , the corresponding eigenvalue is @xmath163 .
the success probability of the successful measurement on the interacting register versus the number of measurements on the interacting register is shown in fig .
@xmath164 . from that figure
, we can see that even for @xmath165 successful measurements , we can still have a success probability of @xmath166 . for @xmath167 versus @xmath30 , the number of successful measurements , by using the @xmath168-eigenstate as the measurement basis for the axial symmetry model . ]
on a quantum computer , a unitary matrix can be efficiently represented , i.e. , for a unitary matrix of dimension @xmath169 , only @xmath170 qubits are needed to represent it on a quantum computer . in this paper , we have tried to represent a non - unitary matrix on a quantum system by performing periodic projective measurements . whether an arbitrary matrix can be constructed
using this technique still remains an open problem , and this would be a subject for future study . in the conventional pea ,
the phase factor is resolved through a quantum fourier transform . to resolve the binary expansion of the phase , up to @xmath87 binary digits
, one has to implement @xmath87 controlled unitary transformations in successive binary powers : @xmath55-@xmath171 , @xmath55-@xmath172-@xmath173 . in the mpea approach of using mqft combined with projective measurements to obtain the eigenvalues of @xmath49
, we need to implement the controlled transformations in successive binary powers : @xmath55-@xmath174 , @xmath55-@xmath175-@xmath176 , and followed by the corresponding mqft circuit as shown in fig .
@xmath96 . the controlled transformations @xmath55-@xmath177 ,
@xmath55-@xmath178 , @xmath179 , @xmath55-@xmath180 , are achieved by implementing the controlled hamiltonian on the whole system only once and during a time @xmath59 until the successful measurements on the interacting register finish , and performing measurements @xmath181 , i.e. , a series of periodic measurements ( each one separated by the time interval @xmath9 ) for @xmath182 @xmath183 times on the interacting register , respectively .
the success probability of the mpea is @xmath184 , where @xmath185 is the fidelity of the state on the target register to be in the eigenstate of @xmath186 after performing @xmath30 successful measurements , and @xmath187 is the probability of performing @xmath30 successful measurements on the interacting register . note that @xmath187 depends on @xmath188 , and also on the initial guess of the state on the target register as shown in eq .
( @xmath189 ) .
since @xmath190 is close to one as the number of successful measurements @xmath30 increases , the success probability is determined by @xmath191 .
it must be emphasized that the present quantum algorithm is designed for systems with large @xmath17 ( the dimension of subsystem @xmath192 ) , when the corresponding classical algorithms for obtaining the eigenvalues of @xmath193 become so expensive that it is impossible to implement on a classical computer .
the efficiency of our algorithm _ does _ depend on @xmath191 .
note that the success probability of projective measurements on the interacting register decreases exponentially in terms of @xmath30 when @xmath194 .
this is not an essential obstacle because this exponential decrease can be overcome by running the algorithm for a number of times to prepare a large but _ fixed _ number of copies of the index qubit state as shown in eq .
( @xmath195 ) . in the qst approach
to obtain @xmath43 , the measurement errors of qst obey the central limit theorem .
accurate results can be obtained by preparing a larger ensemble of the single qubit states .
the tomographic estimation converges with statistical error that decreases as @xmath196 , where @xmath197 is the number of copies prepared in the qst and is not relevant to @xmath17 .
also , in the approach of using single - qubit qst to obtain the eigenvalues of @xmath198 , we prepare a number of copies of the index qubit state as shown in eq .
if we have a good initial guess of the eigenstate of @xmath200 , then , as shown in the second example , we can still obtain a high success probability ( @xmath201 ) for the algorithm and this does not require a large @xmath30 .
the other eigenvalues of @xmath49 can be obtained by setting the initial state of the target register in a pure state .
if the overlap of the initial guess of the eigenstate with the real eigenstate is not exponentially small and @xmath30 is a fixed number , the success probability , @xmath202 , for preparing a index qubit state as shown in eq .
( @xmath195 ) is not exponentially small .
then each copy of the index qubit state can be prepared in a polynomial number of trials .
another issue that needs to be addressed is the efficiency for implementing the projective measurement @xmath203 , which is linked to the efficiency of constructing the non - unitary matrix @xmath204 , therefore connected to the efficiency of the algorithm .
since the measurement @xmath20 is a non - unitary process , it can not be implemented deterministically .
also , a number of projective measurements are required in mpea , and thus the overall efficiency of the algorithm might be affected . to deal with this problem , we can design a scheme such that subsystem @xmath60 can have a simple structure , containing either a single qubit or a few qubits , by controlling the interaction between the subsystems .
then the implementation of the measurement on subsystem @xmath60 will be simple .
the measurement performed on @xmath60 does not depend on the qubit number @xmath205 of the subsystem @xmath192 , on which the matrix @xmath49 is constructed .
therefore , the measurement on @xmath60 can avoid the exponential scaling with respect to the size of subsystem @xmath192 .
note that the corresponding classical algorithms scale as @xmath206 .
we have presented a measurement - based quantum phase estimation algorithm to obtain the eigenvalues and the corresponding eigenvectors of non - unitary matrices . in mpea
, we implement the unitary transformation of the whole system only once ; the non - unitary matrix is constructed as the evolution matrix on the target register . by performing periodic projective measurements on the interacting register ,
the state of the target register is driven automatically to a pure state of the transformation matrix . using single - qubit state tomography and mqft combined with single - qubit projective measurements ,
we can obtain the complex eigenvalues of the non - unitary matrix .
the success probability of the algorithm and the efficiency of constructing the matrix @xmath207 have been discussed .
this algorithm can be used to study open quantum system and in developing other new quantum algorithms .
fn acknowledges partial support from darpa , air force office for scientific research , the laboratory of physical sciences , national security agency , army research office , national science foundation grant no .
0726909 , jsps - rfbr contract no .
09 - 02 - 92114 , grant - in - aid for scientific research ( s ) , mext kakenhi on quantum cybernetics , and funding program for innovative r&d on s&t ( first ) .
law thanks the support of the ikerbasque foundation . | we propose a quantum algorithm for finding eigenvalues of non - unitary matrices .
we show how to construct , through interactions in a quantum system and projective measurements , a non - hermitian or non - unitary matrix and obtain its eigenvalues and eigenvectors .
this proposal combines ideas of frequent measurement , measured quantum fourier transform , and quantum state tomography .
it provides a generalization of the conventional phase estimation algorithm , which is limited to hermitian or unitary matrices . |
quantum algorithm shows us a tempting future to conquer some problems that are hard to solve using classical computers@xcite , such as , factorizing large numbers@xcite , searching in a large database@xcite , and so on .
however , the physical realization of a quantum computer turns out to be a really difficult work .
the greatest obstacle is the decoherence of system induced by unavoidable coupling with the surrounding environment@xcite .
since quantum devices work in a microscopic scale , their interactions with the environment become fatal to the quantum coherence of the system .
someting has to be done to increase the decoherence time so as to complete enough operations within so short a peroid of time .
therefore , error correcting code ( ecc ) is needed@xcite . with the help of ecc
a small error rate can be tolerated .
howerver , since the necessary error - rate threshold still lies far beyond the present available technology , other methods to fight against error should be considered .
one promising method is topological quantum computation ( tqc ) @xcite@xcite@xcite .
topological order@xcite@xcite@xcite@xcite is a new kind of quantum order .
although the mechanism of topological order is not quite clear , we can find the following features in a system in topological order : 1 .
the degeneracy of the ground state closely relates to the topology of the physical system ( usually 2-dimensional ) .
2 . the system is immune to local perturbation and this property can be used to protect qubits .
3 . excited quasiparticles may obey anyon statistics ( abelian or non - abelian)@xcite . with its great advantages , non
- abelian tqc@xcite attracts us the most . in a non - abelian topological phase , not only
can the encoded qubits be protected , but also the operation process can be protected intrinsically through the braiding of non - abelian anyons .
much has been discussed about the fibonacci anyons that can be used for tqc directly in quantum hall effect .
however it is hard to control anyons in real quantum hall systems . besides , kitaev proposed a spin-1/2 honeycomb model@xcite , which seems achievable , followed by some different possible physical implementations@xcite@xcite@xcite@xcite@xcite .
the work in ref.@xcite proposed a more sophisticated method for the anyon detection process .
unforturnately , the non - abelian anyons that emerged in this model do not act well enough to complete the full braiding operation of tqc . attractively , b. doucot , _ et al .
_ gave a complete description of the rhombus josephson junction array ( jja ) with abelian and non - abelian gauge theories @xcite .
however , this is too complex for implementation .
because of all these difficulties we are facing , we will turn to the abelian topological phase in which qubits can be protected as well .
recently , kou showed how to do computations in the toric code model by using the tunneling effect theoretically @xcite@xcite@xcite .
however , the original toric code model calls for four - body interaction , which is hard to implement in real systems .
the josephson junction seems to one of the candidates to implement a realistic quantum computation @xcite@xcite@xcite@xcite .
ioffe and feigelman proposed a hexagonal josephson junction lattice that showed some topological protecting properties which can be used as quantum memory @xcite . based on their work
, we propose a complete tqc scheme in this article .
we focus on encoding logical qubits in the array and the construction of quantum gates .
similar to the method of kou , we also show how to implement fundamental quantum gates , especially the cnot gate .
the article is organized as follows . in sec .
ii , we introduce ioffe and feigelman s model in detail . in sec .
iii , we study the ground - state behavior .
topological degeneracy is used to encode qubits in the array . in sec .
iv we show how to manipulate the qubits to complete basic quantum gate operation .
finally , we draw summary in sec . v.
to describe the computation process we first introduce the model proposed by ioffe and feigelman @xcite in this section . in this work , they showed that the array system exhibits some topological protecting properties and thus can be used as quantum memory .
we discuss how to do computations based on their work in the next section .
now we recount some basic properties of their model in detail first .
these properties are essential for the construction of computation gates . for the convenience of the computation we talk about
, we make a small modification with the boundary and discuss further the quasiparticle excitation , especially the vortex , which agrees , in essence , with the original model .
first , we give the basic building element and explain how it works .
then we show how the array is arranged and give the dynamics of the system . finally , we talk about the topological degeneracy of the ground space and the low - lying excitations .
the classical potential is shown on the right .
, width=96 ] ( b ) .
the classical potential is shown on the right .
, width=115,height=86 ] as shown in fig.[potential ] , the basic building block is a rhombus made of a four - junction superconducting loop placed in a magnetic field . the flux through the rhombus is @xmath0 , where @xmath1 is the flux quanta .
we require that all the josephson junctions are the same and work in the `` phase regime '' , that is , @xmath2 , where @xmath3 and @xmath4 .
this system , as is much discussed in the literature , forms a simple two - level `` flux qubit '' ( or persistent - current qubit)@xcite@xcite . in the following ,
we briefly explain how the qubit system emerges .
we denote the phase of the superconductor islands with @xmath5 .
the josephson system is described with the lagrangian @xmath6 where @xmath7 .
the quantity in the cosine term is the the gauge - invariant phase difference across each junction . here
, we make an approximation that three of the junctions have the same phase difference .
as all phase differences should sum up to @xmath8 , we get @xmath9 where @xmath10 indicates the phase difference between the two far ends of the rhombus . neglecting the kinetic part , this energy has two equivalent classical minima at @xmath11 ( as shown in fig.[potential ] ) , that is , the phase difference across each junction is @xmath12 .
so the two classical ground states correspond to states in which there are both clockwise and counterclockwise supercurrents in the loop .
this is the reason why we call it a persistent - current qubit .
we denote the two states as @xmath13 and @xmath14 .
quantum fluctuation makes the electron tunnel between the two classical states in the wells and that splits the degenerate ground states .
the low - energy configuration can be regarded as a two - level qubit system .
the effective hamiltonian is @xmath15 where @xmath16 , and @xmath17 is the classical action taking the path between the two minima .
the value of tunneling energy is estimated regarding the potential as quadratic .
let us see what happens when the condition is slightly changed from the ideal .
notice that the dynamics require that the flux through the rhombus is _ exactly _ half a flux quanta .
if the external flux is a bit different from the ideal value , @xmath18 , we add a first order change to the potential , @xmath19 as a result of the slight change of the flux , the two minima of the double well are no longer degenerate but separated by a gap @xmath20 .
then the effective hamiltonian is @xmath21 .
this is important for operations .
we can control the evolution of the system by adjusting the flux through the rhombus .
below , in the josephson array building from the basic rhombi , logical qubits are encoded for computation .
some gate operations are done just by the method of tuning the flux of the rhombi along specific paths .
the array is punctured with several holes .
the number of the holes relates to the degeneracy of the ground space .
the boundary , as shown in the lower figure and indicated by a ( blue ) solid line , is a whole superconductor wire connecting all the rhombi along the boundary .
the flux through the polygon loops on the boundary are @xmath22 and @xmath23 .
, title="fig:",width=288 ] \(b ) .
the array is punctured with several holes .
the number of the holes relates to the degeneracy of the ground space .
the boundary , as shown in the lower figure and indicated by a ( blue ) solid line , is a whole superconductor wire connecting all the rhombi along the boundary .
the flux through the polygon loops on the boundary are @xmath22 and @xmath23 .
, title="fig:",width=288 ] now we have got the bricks we need . in this part
we show how to build the josephson array system .
the basic rhombi are arranged as shown in fig.[array ] . for simplicity ,
a rhombus is indicated by a short line in the figure .
six rhombi form a hexagon .
the array is placed in a uniform magnetic field .
we should carefully tune the shape of the rhombi , so that the flux through the hexagram in the center is a half - integer multiple of @xmath24 , @xmath25 .
the centers of the hexagons are denoted with @xmath26 and so on and individual rhombi between the corresponding two centers with @xmath27 .... the array contains @xmath28 holes .
soon we will show that , in fact , each hole can be treated as a single logical qubit . in the array , the rhombi form loops ,
so we should check the gauge - invariant condition over again . to satisfy the constraint condition for all the closed loops in the array ,
the sum of the phase differences around each hexagon should equal @xmath29 [ i.e. , @xmath30 .
this is well consistent with the configuration that all the @xmath31 of the six rhombi on the edges of a hexagon take the ground - state value @xmath32 ( or all take @xmath33 ) .
the flip of a single rhombus is not allowed because it changes the sum of the phase differences around the loop by @xmath34 and violates the constraint .
however , we can flip rhombi of an even number in a hexagon at the same time ( e.g. , two , four or six ) . choosing the two ground states as a basis , we can write the constraint in the operator form , @xmath35 the product takes the rhombi on the hexagon loops . because of the existence of the constraint , the rhombi embedded in the array can no longer behave as in eq.([rhoh ] ) , which does not commute with the constraint .
the simplest dynamics compatible with constraint contain flips of three adjacent rhombi , @xmath36 , such that two rhombi flip at the same time in the corresponding three hexagons .
this gives the main contribution to the hamiltonian of the system @xmath37 here , @xmath38 is estimated to be analogous to @xmath39 in the last section .
the action in the exponential term is three times of that of a single rhombus because of the simultaneous flip of the three rhombi .
now we talk about the boundary .
we should treat it carefully .
the boundary here is arranged a bit differently from ioffe s original model .
different boundary conditions of the toric code have been discussed in detail@xcite . in the system considered here , similarly , the boundary is arranged more like the `` rough edge '' . here
, we arrange the boundary rhombi in such a way that they behave like those in the bulk , as described by eq.([aryh ] ) .
each rhombus on the boundary still stays close to the other two .
a superconductor line connects the unattached rhombi at the outer and inner boundary ( like the blue line in fig.[array ] ) .
the geometry of the polygon loops on the boundary should be carefully treated .
a boundary loop has @xmath40 rhombi , usually three or four .
we want to maintain the dynamics and constraint of the form as before , that is , two rhombi must flip at one time .
so the flux embedded should be consistent with the sum of the phase differences of the @xmath40 rhombi , i.e. , @xmath41 , where @xmath42 is an integer . then the hamiltonian eq.([aryh ] ) is still kept for the rhombi on the boundary .
we complete the construction of the josephson array and give the low - energy dynamics . in the following section
, we will discuss the properties of the ground - state space and excitations in detail . in this section ,
we discuss the properties of the ground space and excitations .
the topological degeneracy of the ground space can be used to encode logical qubits for computation .
we show two kinds of quasiparticles in the system , charge and vortex .
we can operate qubits by controlling their motion in the array .
first , let us find the ground state of the system . ignoring the constraint
, we can easily get the ground state of hamiltonian ( [ aryh ] ) as @xmath43 .
of course , it is illegal and incompatible with the constraint .
we can project it into the physical space with the help of an operator @xmath44 @xmath45 is a ground state of the system , but not the only one . to find all the ground states
, we can turn to the help of a path operator @xmath46 where @xmath47 takes the big ( red ) loop in fig.[array ] . notice that all the @xmath48 with the paths homotopic with @xmath47 are equivalent .
also it is easy to check that @xmath49 commutes with both @xmath50 and @xmath51 .
so obviously @xmath52 is another ground state of @xmath50 . noting that @xmath53
, we can construct all the ground states @xmath54 here , the product contains all the @xmath28 `` basic loops '' ( i.e. , loops around each single hole ) .
note that in the representation above , @xmath55 .
so we can interpret @xmath49 as the operator measuring the `` spin '' direction of the @xmath56th hole .
we will utilize this physical meaning of @xmath49 for encoding in the next section .
now we find the whole ground space of the array system with @xmath28 holes .
the dimension of the ground space is just @xmath57 .
it is interesting that the degeneracy correlates with the topology of the system .
this is why it is called topological degeneracy .
we can encode @xmath28 qubit in this space and do computations in it .
based on the knowledge of the ground space , let us study the low - lying excitations of the system .
an open string product of @xmath58 , which is compatible with the constraint , excites two `` charges '' at the two ends of the string . with the help of the anti - commutation relation with the hamiltonian , after a simple algebra computation
, we can deduce that the energy of each single `` charge '' is @xmath59 above the ground .
unfortunately , a string product of @xmath60 is not allowed by the constraint , which is different from the case in toric code model@xcite . to get another type of excitation
, we should break the constraint eq.([constraint ] ) and contemplate the problem from the original lagrangian of the array .
approximately , we only keep the degree of freedom of the superconductor islands at the two ends of the rhombi ( i.e. , at the vertices of the hexagons ) .
correspondingly , we use the modified potential eq.([rhou ] ) .
then the lagrangian of the array is @xmath61 where @xmath62 . taking the continuum limit ,
the josephson array system can be mapped to a type - ii superconductor@xcite .
a vortex solution exists in the system , taking one flux quanta .
the energy of a single vortex is @xmath63 here , @xmath64 , and @xmath65 is the length between the two centers of the hexagons .
vortices stay at the hexagons where the constraint is broken .
the dynamics there are changed correspondingly .
there is one kind of dynamic that interests us most , the single flip of a rhombus . to get this kind of dynamics
, we carefully change the flux of two adjacent hexagrams to integer multiples of @xmath24 .
two vortices appear and stay here .
single flip happens only at the rhombus between the two hexagons .
since the other rhombi stay close to the hexagons that satisfy the constraint , the single flip is still not allowed .
this is also the reason why we must change the flux of two adjacent hexagons , but not one or two separated hexagons .
if we change the flux of only one hexagon , we still can not get the single flip because it is forbidden by the adjacent hexagon .
now we can add a @xmath60 term to @xmath66 with the amplitude @xmath39 at the rhombus between the two adjacent vortices . or rather
, we can interpret it in such a way that two vortices with each energy @xmath67 are excited by a @xmath60 operation .
this term is important for operations that we will discuss later .
so we have got two types of quasi - particle excitations in the array .
after all the discussions about the properties of the array we focus , in this section , on the computation process . it was mentioned in ref .
@xcite that motion of quasi - particles along topologically non - trivial paths changes the state .
we care about how to achieve it . in this section ,
we show how to encode logical qubits in the array , and propose an exact method to operate qubits in lab . however , first we give a short review of stabilizer code formalism .
in fact , this system can be regarded as stabilizer code .
then we show how the topological degenerated ground space is protected from local noise .
logical qubits are encoded in the `` holes '' .
the spin directions of holes can be used to denote the state of the @xmath28 logical qubits .
finally , we show how to implement the single qubit and multi - qubits operations .
stabilizer code is developed as a general theory of quantum error - correction code@xcite . in this formalism
, @xmath68 physical qubits are used to encode @xmath28 logical qubits ( @xmath69 ) .
certain types of errors can be detected and corrected so that the reliability of quantum computation is greatly improved .
first , we define the pauli group @xmath70 on @xmath68 qubits as @xmath71 where @xmath72 are integers .
@xmath70 contains all the @xmath68-body pauli operators with coefficients @xmath73 .
note that any two elements in the pauli group either commute or anti - commute with each other .
a subgroup @xmath74 of @xmath70 is called a _ stabilizer group _ if there exists a nontrivial space @xmath75 stabilized by @xmath74 , @xmath76 it can be proved that @xmath75 is @xmath57-dimensional if the group @xmath74 has @xmath77 independent generators .
we denote the stabilizer group as @xmath78 , where @xmath79 is one of the @xmath77 generators of the group .
the elements in @xmath74 are called _ stabilizer operators _ , or just _
stabilizers_. @xmath28 logical qubits are encoded in the stabilized space @xmath75 . to correct errors
, we should first notice that the effects of errors can be represented by the one - body elements in the pauli group @xmath70 , because errors are usually local .
when no error happens , after measuring all the stabilizers in @xmath74 ( just measuring all the @xmath77 generators is ok . ) , all results are 1 .
assume one error , denoted by @xmath80 , happens , so that the original computation state @xmath81 is changed to @xmath82 .
if @xmath83 anti - commutes with at least one stabilizer in @xmath74 , e.g. , @xmath84 , the measurement of @xmath84 gives the result @xmath85 , then we know the error type and we can correct it by corresponding operation . however , there may exist some elements in @xmath86 that also commute with all the elements in @xmath70 .
this type of errors can not be detected or corrected .
the basis of logical qubits should be written out for computation .
we can always find @xmath28 operators @xmath87 from @xmath86 such that @xmath88 form an independent operator set .
@xmath87 plays the role of a logical operator @xmath89 .
the logical computational basis @xmath90 is actually the state stabilized by the group @xmath91 the logical @xmath92 can be also found from @xmath86 satisfying : 1 .
@xmath93=0 $ ] and 2 .
@xmath94 .
we emphasize here that the logical operator set @xmath95 is not unique .
also , we did not limit @xmath87 to be of some certain form such as a product of several @xmath58 .
it can be a mixed product of @xmath58 and @xmath60 , or even a product that contains only @xmath60 .
the only condition required is that @xmath88 form an independent operator set .
for example , suppose we find a set @xmath96 that satisfies the condition and @xmath97 is the set of corresponding logical flipping operators , it is easy to check that @xmath98 is also an independent set .
so the status of logical sets @xmath96 and @xmath97 are the same , and their roles can be exchanged . in this section ,
we show how to encode logical qubit in the topological protected array . as an example
, the toric code is a kind of stabilizer code .
the computation is done in the ground space of @xmath99 , where @xmath100 and @xmath101 are the star and plaquette operators .
the ground space is stabilized by the operators @xmath102 .
the loop products of @xmath60 and @xmath58 along two directions give rise to the logical gate @xmath103 , @xmath104 , @xmath105 and @xmath106 .
but toric code model only encodes two logical qubits .
if we want to encode more qubits , we have to implement a much too complex boundary condition for more genus . in our system with puncture topology
, we can see that encoding is much easier .
we can write down the stabilizers in our considering model , @xmath107 and @xmath108 .
the physical ground space satisfying @xmath109 constitutes the protected code space .
but this is not enough .
now we have to find the logical operators @xmath96 and @xmath97 . in an array with @xmath28 holes
, we can see that the product of @xmath60 around the @xmath56th hole ( i.e. , @xmath49 mentioned earlier ) , actually plays the role of logical operator @xmath110 .
we have mentioned that eigenvalues @xmath111 of @xmath110 can be regarded as the spin values of the @xmath56th hole .
we can also find the logical operator @xmath112 as the product of @xmath58 along the path from the outer boundary to the inner of the @xmath56th hole ( the yellow lines connecting the right hole to the outer boundary in fig.[array ] ) .
@xmath112 flips the spin of the @xmath56th hole to the opposite direction
. it may feel strange that we take the loop product of @xmath60 as logical @xmath113 while taking the string product of @xmath58 as logical @xmath114 .
we should notice that the string product of @xmath60 is not allowed by constraints at the two ends of the string . on the other hand
, although we can exchange the roles of @xmath114 and @xmath113 as mentioned earlier , the representation we chose here has a better physical meaning as `` spin '' .
therefore , we denote the logical computation basis by the spin directions of the holes , like @xmath115 . the corresponding single qubit operations are @xmath110 and @xmath112 , formally .
we emphasize again that all the homotopic paths are equivalent . around the @xmath56th hole measures the direction of its flux .
the state shown is @xmath116 .
, width=272 ] now we show why the code space is protected against local noise .
we assume that the noise can be treated as local perturbations on the rhombus , @xmath117 . to consider the perturbation effect of the first order , we compute the matrix element @xmath118 .
@xmath119 is a set of orthogonal basis of the ground space @xmath120 , as discussed in the last section .
there must exist at least one @xmath121 that anti - commutes with @xmath117 , so @xmath122 is driven outside @xmath120 and @xmath118 is zero .
only perturbations of an order higher than @xmath68 take effect , such that @xmath68 pauli operators form a topologically nontrivial path across different boundaries , or form a loop embedding at least one hole .
when @xmath68 is large enough , the system can be well protected and this is the reason why it is called topologically protected . when it is cooled to a low enough temperature ,
the system arrives at the protected ground space automatically .
the computation is well protected from local noises . however
, proper perturbations can drive quasi - particles tunnel through nontrivial paths , and this can be used to construct quantum gates .
we will do computations by this method in the following part .
we showed how the computation basis for logical qubits are represented .
now let us see how to implement quantum gate operations in this system .
from the discussion of the system previously , we know that quasi - particles are excited by pauli operators .
the product operators of @xmath60 and @xmath58 along topologically non - trivial paths change the state of the system .
so we get a more physical picture that the transmission of both charge and vortex changes the state .
this can be used for operation in this section . to complete the qubit gates operation ,
we use a method similar to kous@xcite .
we change the external magnetic field .
effectively , the ground space behaves as a pseudo - spin system .
we can see that it just drives the quasi - particles to tunnel through different sectors . as an example
, we slightly tune the flux of the rhombi along a path from the @xmath56th hole to the outer boundary . according to the previous discussion ,
perturbations of the @xmath58 form appear in the single - rhombus hamiltonian , and this is allowed by the constraint .
so we get the hamiltonian @xmath123 confined in the degenerate ground space .
the main part gives identity with an energy coefficient .
the effect of perturbation appears in the @xmath124th order .
@xmath68 is the number of rhombi along the path .
one can get the energy splitting@xcite
@xmath125 where @xmath126 is the ground energy of @xmath66 .
each @xmath127 contributes one single @xmath58 .
only the term that contains a complete string product of @xmath58 connecting both the inner and outer boundaries takes effect as @xmath112 ( remember that in the representation we use here , products of @xmath58 behave as @xmath114 not @xmath113 ! ) , while others vanish .
so we get the effective hamiltonian of the array as @xmath128 . in this process ,
a charge tunnels across the array from the @xmath56th hole to the outer boundary .
degenerated states split just as in the case in the double - well of a single rhombus .
similarly , we want to get an effective hamiltonian like @xmath129 by adding perturbations of the form such as @xmath130
. however , as mentioned before , this is not allowed by the constraint under ideal condition .
something more should be done .
we have talked about the vortex excitation in the last section . to achieve this purpose
, we carefully change the flux of the hexagons along the loop around the @xmath56th hole ( as shown in fig.[array ] ) to be integer multiples of @xmath24 .
the constraint is broken , and the dynamics of the corresponding hexagon changes .
the single flip of a rhombus is allowed .
so we get the desired perturbation as @xmath131 where @xmath39 takes the value in eq.([rhoh ] ) .
two vortices emerge .
one is transmitted around the @xmath56th hole , and then annihilates the other one .
this process gives the @xmath113 term .
similar to eq.([split ] ) , we get the energy splitting @xmath132 so we have got @xmath133 .
as we known , any single - qubit rotation can be written as @xmath134 we have got both @xmath135 and @xmath136 , then we can get any single qubit gate , at least in principle .
we can tune the corresponding flux of rhombi or hexagons and get the effective hamiltonian we need , carefully control the evolution time of the system ( that is to tune the parameters @xmath137 above ) , then we can construct any single qubit gate we want . .
this gives the effective hamiltonian @xmath138 .
the red dashed loop represents the hexagons with flux changed to integer multiples of @xmath24 . that gives @xmath139 .
, width=272 ] in the last part , we saw that any single qubit gate can be constructed . in the this section
, we show that in this array multi - qubit operations are also easy to realize on any two qubits .
when we change the flux of the rhombi on the line connecting two holes ( see the yellow line in fig.[multi ] ) , similarly , the states of the corresponding two qubits are flipped at the same time , and this gives @xmath140 term in the hamiltonian . by using the perturbation approach , the amplitude is @xmath141 here , @xmath142 , and the summation contains all the rhombi on yellow line .
we get a effective hamiltonian @xmath138 .
analogously , we can get another type of operation .
we can tune the flux of the hexagons along the loop around two holes to integer multiples of @xmath24 .
this gives a perturbation @xmath143 .
the amplitude is @xmath144 then we get the effective hamiltonian @xmath139 .
the most important thing we are concerned with is to construct cnot operation , because cnot gate together with arbitrary single qubit rotation gate gives rise to a universal quantum computation .
this is also realizable in this array .
first , we should construct another effective hamiltonian .
we tune the flux in such a way that a vortex is transmitted around the @xmath40th hole , and a charge is transmitted from the @xmath145th hole to the outer boundary at the same time .
the effective hamiltonian contains contributions of three parts , @xmath146 .
@xmath147 and @xmath148 are evaluated in the same way with @xmath149 and @xmath150 in the last section .
a cnot gate can be constructed with the help of @xmath151 , @xmath152 and @xmath153 .
we can let @xmath151 evolve for @xmath154 .
then we should change the external magnetic field to construct @xmath152 and let the system evolve for @xmath155 . finally , we construct @xmath153 and let the system evolve for @xmath156 .
now we get the cnot operation with an external global phase @xmath157 , which is negligible .
in summary , we propose a tqc scheme with protected qubits in a jja system , and we indicate how to perform qubit encoding , single qubit operation , and multi - qubit operation . especially , we show the way to implement the cnot quantum gate operation which is a key element in quantum computation .
the scheme we propose here can be realized in real jja system since the technology of the josephson junction is developing rapidly .
considering the convenience of realizability and scalability of our system , the dimension of ground space is related to the number of punctures on the array .
the logic qubit encoding here is in a protected space so that our system is topologically protected from noise . besides
, specific quasiparticles are driven by proper perturbation transmitted along topologically nontrivial path .
this makes small splittings happen .
we give the effective hamiltonian under different condition corresponding to various splitting situations and this can be used to do quantum computation . | we propose a complete quantum computation process on a topologically protected josephson junction array system , originally proposed by ioffe and feigelman [ phys .
rev .
b 66 , 224503 ( 2002 ) ] .
logical qubits for computation are encoded in the punctured array .
the number of qubits is determined by the number of holes .
the topological degeneracy is lightly shifted by tuning the flux along specific paths .
we show how to perform both single - qubit and basic quantum - gate operations in this system , especially the controlled - not ( cnot ) gate . |
since the realization of the bose - einstein condensed ( bec ) atomic gases @xcite , there has been much interest in ultracold trapped atomic systems to study quantum many - body phenomena . besides the bose - einstein condensates ( bec ) @xcite , one can now study degenerate atomic fermi gases @xcite and bose - fermi mixtures @xcite .
these systems offer great promise to exhibit new and interesting phenomena of quantum many - particle physics .
sogo and yabu @xcite studied the ground state properties of fermi gas composed by two spin components with various repulsive interaction and showed the phase transition between the paramagnetic and ferromagnetic states .
furthermore we extended this study on the spin phases of the ground state to that for the asymmetric gas with a nets spin in the presence of an external magnetic field and gave the phase diagram about these spin phases @xcite .
an important diagnostic signal for many - body systems is the spectrum of collective excitations .
such oscillations are common to a variety of many - particle systems and are often sensitive to the interaction and the structure of the ground state . in ref.@xcite
we have also studied the spin excitation on the dipole and monopole oscillations in the two - component fermion condensed system .
the frequencies of the collective oscillations with the out - of - phase show the dependence on the fermion - fermion interaction while those with the in - phase are not sensitive to the interaction .
particularly the frequencies of the out - of - phase modes show rapid changes in the phase transition between the paramagnetic and ferromagnetic spin - phases .
the monopole oscillations can be defined only in the spherical symmetric system . in usual experiments traps
have elliptically deformed shapes , where the monopole oscillations are coupled with the quadrupole oscillations .
if the trap is largely deformed with the axial symmetry the oscillations must be decoupled into the modes of oscillations along the transverse and longitudinal directions ; we shall call them the transverse breathing ( tb ) and longitudinal breathing ( lb ) modes .
in this paper we study the coupled oscillation states of the two component fermi gases between the tb and lb modes in deformed traps .
the oscillation behaviors are expected to show various features as the deformation of the trap is changed .
the underlying theoretical tool to treat dynamic problems of dilute quantum gases is the time - dependent mean field theory .
this is reduced to the random - phase approximation ( rpa ) for small amplitudes , and the theory in this form has been applied to density oscillations of these systems @xcite . when the single - particle spectrum is regular , the long - wavelength excitations are collective and simpler methods can be used to calculate the frequencies , in particular with sum rules @xcite or the scaling approximation . in this work
we calculate the frequencies of these collective oscillations by using the scaling method @xcite .
this scaling method is physically quite transparent , but it is in fact equivalent to the theory based on energy - weighted sum rules @xcite , as used for example in ref .
@xcite .
we will only consider here the case of positive scattering lengths which correspond to a repulsive interaction .
when the interaction is attractive , the system is superfluid in its ground state and the excitation properties are controlled by the energy gap .
in addition we will focus only on the collective oscillations in the paramagnetic spin phases , and do not deeply discuss those in the ferromagnetic regime . in the next section
we will explain our formulation , and show calculational results in sec .
3 . in sec.4
we will summarize our work .
we begin with the expression for the energy of a dilute trapped condensate in the mean - field approximation : @xmath0 \label{etot}\end{aligned}\ ] ] with @xmath1 where @xmath2 are orbital wave functions indexed by orbit @xmath3 and component @xmath4 , @xmath5 are the longitudinal and transverse frequencies of the trapping field and @xmath6 is the coupling strength of a contact interaction .
here we assume that all fermions have the same mass , @xmath7 , but we do not need to treat these components as the spin state of the same fermion . in order to study the collective oscillation
, we introduce the following scaling for the fermion wave - function : @xmath8 with @xmath9\ ] ] where @xmath10 is the wave - function in the ground state , @xmath11 are the time - dependent collective coordinates for the oscillation , and @xmath12 are the time - derivatives of @xmath11 .
the factor @xmath13 is the gallilei transformation factor which is determined to satisfy the continuum equation .
let us obtain the total energy under the above scaling by substituting the wave - function ( [ scwf ] ) into the total energy functional ( [ etot ] ) . the total energy becomes @xmath14 where the superscript @xmath15 indicates @xmath16 and @xmath17 at the ground state . in this work we use the thomas - fermi ( tf ) approximation to evaluate the above total energies .
the tf approximation makes the spherical symmetric in the momentum distribution at each position and gives the following relation : @xmath18 in addition the ground state densities @xmath19 become functions of @xmath20 .
here we can then make a change of variables to simplify the appearance of the thomas - fermi equations as follows .
@xmath21 where @xmath22 is a certain coupling constant which is chosen later . with the scaled variables and the thomas - fermi approximation , the expression for the total energy at the ground state is @xmath23 where @xmath24 .
the tf equations for the densities @xmath25 are derived by variation of the energy with a constrain on the fermion numbers .
this yields @xmath26 in this equation @xmath27 and @xmath28 are the lagrange multiplier to constraint the numbers of the fermion-1 , @xmath29 and -2 , @xmath30 , respectively , and they are rewritten with @xmath31 and @xmath32 ; these @xmath31 and @xmath32 have the meanings of the fermi energy and the external magnetic field in fermion gas when the two components of fermions are described with two different spin states of one fermion . in this expression the ground state
is determined by the three variables , @xmath33 , @xmath31 and @xmath32 , but the two variables , @xmath33 and @xmath31 , are not independent because of the scale symmetry in eq.([tfeq ] ) @xcite . in this work
we choose @xmath22 as a critical coupling constant of the paramagnetic and ferromagnetic phase transition at a fixed chemical potential in the symmetric system @xmath34 . with this choice
the fermi energy is fixed to be @xmath35 .
the solution of these equations is discussed in detail in ref .
@xcite . with fixed values of @xmath36 and @xmath37 ,
the spin phase of the two component fermi gas turns to be paramagnetic ( @xmath38 ) , ferromagnetic(@xmath39 ) in order as the coupling constant , @xmath33 , increases .
furthermore the total energy in collective oscillations is given as @xmath40 with @xmath41 where @xmath24 , @xmath42 and @xmath43 ( @xmath44 ) .
the interaction energy @xmath45 appears @xmath46 in order to describe the collective oscillations , we obtain the variation of the total energy up to the order of @xmath47 ) as @xmath48 with @xmath49 @xmath50 with @xmath51 where the vector @xmath52 is defined as @xmath53 . the terms propritional to @xmath52 disappears because of the virial theorem .
then the classical equation of motion for @xmath52 is harmonic , giving rise to the following eigenvalue equation for the oscillation frequencies , @xmath54 { \mbox{\boldmath $ \lambda$}}= 0 . \label{eigeq}\ ] ] before performing actual calculations , we should discuss the collective oscillations in some extreme conditions . in the largely deformed system , @xmath55 or @xmath56 ,
first , the four oscillation modes are decoupled into two groups , two oscillations modes in the longitudinal directions , and the other two modes in the transverse directions .
we can obtain these decoupled modes by taking @xmath57 . in the symmetric system ,
@xmath58 , furthermore , the two kinds of modes are further decoupled into the in - phase ( @xmath59 ) and the out - of - phase ( @xmath60 ) modes ; namely the four excited oscillations are decoupled into the in - phase transverse breathing ( itb ) , in - phase longitudinal breathing ( ilb ) , out - of - phase transverse breathing ( otb ) and out - of - phase longitudinal breathing ( olb ) modes .
the oscillation frequencies in the symmetric system become @xmath61 where the superscripts @xmath62 and @xmath63 indicate the in - phase and out - of - phase modes , respectively , and the subscripts t and l denote the oscillations in the transverse and longitudinal directions , respectively .
note that @xmath64 , @xmath65 and so on .
next we discuss the oscillations in the near - spherical symmetric system , @xmath66 . in this system
the oscillation modes are decoupled into the monopole and quadrupole oscillation modes . by taking the collective
coordinates to be @xmath67 , we can obtain the frequencies of the monopole oscillations with the following equation : @xmath68 \left ( \begin{array}{c } \lambda_1 \\ \lambda_2 \end{array } \right ) = 0\ ] ] with @xmath69 by taking the collective coordinates to be @xmath70 , similarly , we can also calculate the frequencies of the quadrupole oscillations with @xmath71 \left ( \begin{array}{c } \lambda_1 \\ \lambda_2 \end{array } \right ) = 0\ ] ] with @xmath72 in the symmetric system @xmath34 , these modes are also decoupled to the in - phase and the out - of - phase modes .
then the oscillation frequencies in the symmetric system become @xmath73 where the subscript m and q indicate the monopole and quadrupole frequencies , respectively .
in this section we discuss the collective oscillations coupled with the transverse and longitudinal breathing modes . as mentioned in the previous section we fix the fermi energy to be @xmath35 , which makes the parametric and ferromagnetic spin phase transition at @xmath74 when @xmath37 .
density profiles illustrating these three regimes are shown in fig .
the panels show the densities as a function of distance @xmath75 in a weak magnetic field ( @xmath76 ) and at three different interaction strengths : @xmath77(a ) , 0.95(b ) and 1.05(c ) .
the solid and dashed lines represent the scaled density distribution of major and minor components of fermions . from top to bottom , the panels show paramagnetic ( a , b ) and ferromagnetic ( c ) .
next we calculate the collective oscillations by solving eigenvalue equation ( [ eigeq ] ) . in this equation
we obtain four eigenvalues corresponding to collective oscillation frequencies . for convenience
we define these four frequencies as @xmath78 , @xmath79 , @xmath80 and @xmath81 in order of the frequencies from the larger one to the lower one .
in addition we refer to the four collective modes with the frequencies @xmath78 , @xmath79 , @xmath80 and @xmath81 as modes-1 , -2 , -3 and -4 , respectively . in fig .
[ mdlt1 ] we show the frequencies of oscillations as functions of the deformation parameter @xmath82 at the coupling constant @xmath83 ( a ) , @xmath84 ( b ) and @xmath85 ( c ) .
the solid lines represent the first and third frequencies @xmath78 and @xmath80 , and the chain - dotted lines indicate @xmath79 and @xmath81 . in these calculation
we take the magnetic field to be @xmath86 , where the system is almost symmetric between the two component , @xmath87 . in the same figures ,
furthermore , we also plot the frequencies of the transverse and longitudinal breathing modes with the dashed and dotted lines , respectively .
first we see the strong level mixing between modes-2 and -3 at two deformation parameters , @xmath88 and @xmath89 ( @xmath90 ) . when @xmath91 and @xmath92 , the frequencies in the full calculation ( solid lines ) are almost the same as those in the calculations decoupled with the transverse and longitudinal modes ( dashed and dotted lines ) , though they are quite different when @xmath93 . in fig
[ mdlt2 ] we also give the same quantities but with @xmath94 .
the system is not symmetric @xmath95 , but the results are quite similar to those in fig .
[ mdlt1 ] though the difference in the frequency at the level mixing points between the modes-2 and -3 is a little larger than those in fig .
[ mdlt1 ] . for future discussions we refer to the three deformation parameter regions , @xmath96 , @xmath97 and @xmath92 as the prolate deformed , the semi - spherical and the prolate deformed regions , respectively .
the transverse and longitudinal breathing oscillation modes are almost decoupled in both the oblate and prolate deformed regions , while these modes are coupled in the spherical region . in order to know the level mixing behaviors more ,
we calculate the four following components : @xmath98 where @xmath99 , @xmath100 , @xmath101 and @xmath102 are components of the itb , ilb , otb and olb modes , respectively . in fig .
[ evc1 ] we show the components of the eigenvectors at the mode-1 ( a ) , the mode-2 ( b ) , the mode-3 ( c ) and the mode-4 ( d ) with @xmath103 and @xmath86 .
the solid , chain - dotted , dashed and dotted lines represent @xmath99 , @xmath100 , @xmath101 and @xmath102 , respectively .
the results show that the phase of the modes-1 and -4 are in - phase in all @xmath104 regime . when the system is oblate deformed ( @xmath105 ) .
the mode-1 is almost the itb mode .
as @xmath104 increase , the system approaches the spherical , the ilb component in the mode-1 mode gradually becomes larger .
when the system becomes spherical ( @xmath106 ) , the mode-1 becomes a mixed mode with the itb and ilb in the same ratio : namely the mode-1 is the monopole oscillation .
as the system is further deformed to the prolate shape , the transverse component decreases and the longitudinal one increases and the oscillation becomes the in - phase longitudinal oscillation in @xmath107 limit .
the mode-4 has the same behavior , but its phase is out - of - phase ; the otb mode when @xmath96 , the out - of - phase monopole mode at @xmath108 and the olb mode when @xmath109 .
the variations of the modes-2 and -3 show more complicated behaviors .
the mode-2 is mainly the otb mode in the oblate deformed region ( @xmath96 ) , and the olb mode in the prolate deformation region ( @xmath109 ) .
the mode-3 is the ilb mode in the oblate deformation , and the itb mode in the prolate deformation . at @xmath88 and @xmath89 , the modes-2 and
-3 make the level mixing and exchange their roles .
namely the modes-2 and -3 are in - phase out - of - phase oscillations in the semi - spherical region , respectively .
as @xmath104 approaches zero , the tb and lb oscillation modes are mixed . at @xmath106 ,
the mode-2 becomes the in - phase quadrupole oscillations , ( @xmath110 ) and the mode-3 becomes the out - of - phase ones , respectively .
thus we know that the four modes agree with the frequencies of the tb and lb modes , when @xmath111 ( @xmath112 ) , and @xmath113 ( @xmath105 ) . on the other hand these modes show different behaviors in the nearly spherical symmetric system ( @xmath114 ) , where the collective oscillations are described as the coupled modes of the monopole and quadrupole oscillations . in fig .
[ evc2 ] , furthermore , we show the same quantities in fig .
[ evc1 ] but with @xmath94 .
the feature discussed above is less clear , but also seen in the case of the strong magnetic field .
when @xmath94 , the number asymmetry is not so small , @xmath115 .
the mixing between the in - phase and out - of - phase modes is very small in modes-1 and -4 .
the similar result was shown for the monopole oscillation in the spherical trap . even in modes-2 and -3 ,
this mixing is seen only in the region near @xmath116 and @xmath117 .
hence we can disregard the mixing between the in - phase and out - of - phase modes in qualitative discussions , and consider that modes-2 and -3 exchange their roles around @xmath116 and @xmath117 . in order to know more details of the deformation dependence
, we should examine relations between the frequencies of the collective oscillations and the coupling constant @xmath33 .
first we show the frequencies of the monopole and quadrupole oscillations versus the coupling constant @xmath33 in the spherical symmetric system ( @xmath118 ) with the @xmath86 in fig .
the solid and dashed lines represent the results of the monopole and quadrupole oscillations . in the completely spherical symmetric system the monopole and quadrupole states are decoupled , and there is no level mixing between these two modes . in the paramagnetic phase ( @xmath119 ) ,
larger frequencies correspond to the in - phase oscillation modes both in the monopole and quadrupole modes .
the monopole modes make the level mixing between the in - phase and out - of - phase modes in the ferromagnetic phase ( @xmath119 ) , while the quadrupole modes do not show the level mixing .
as the coupling becomes larger , the frequency of the oscillation modes except the in - phase monopole oscillation mode decreases monotonically until @xmath120 , and increases above that , while the frequencies of the in - phase monopole oscillation are almost invalid . when the coupling constant increases further , the level mixing between the two modes of the monopole oscillation occurs .
next we study the non - symmetric system , where all the four modes can be mixed . in fig .
[ spclmq ] we show the frequencies of the mode-1 , -2 , -3 and -4 as the functions of the coupling constant with @xmath121 ( @xmath122 ) ( a ) , @xmath123 ( @xmath124 ) ( b ) , @xmath125 ( @xmath126 ) ( c ) and @xmath127 ( @xmath128 ) ( d ) . in these figures
we see again that the levels between modes-2 and -3 are mixed at a certain @xmath33 . when @xmath129 ( fig .
[ spclmq ] ) , for example , the frequencies of modes-1 and -2 are @xmath130 , and those of modes-3 and -4 are @xmath131 at @xmath132 . as the coupling becomes larger , the frequencies of modes-2 and -4 monotonously decreases while those of modes-1 and -3 is almost invalid , and then the the level mixing between modes-2 and -3 occurs at @xmath133 .
as the deformation parameter @xmath104 increases , the difference between @xmath134 and @xmath135 becomes larger , and the level mixing occurs at larger coupling constant .
as the deformation parameter further increases , the level mixing disappears when @xmath96 because the minimum value of @xmath79 is larger than @xmath80 , and the level mixing can not occur .
the @xmath79 becomes minimum at the border between the paramagnetic and ferromagnetic phases @xmath136 , where the level mixing occurs when @xmath88 . in fig .
[ spvcg ] we show the components of the eigenvectors at the mode-1 ( a ) , the mode-2 ( b ) , the mode-3 ( c ) and the mode-4 ( d ) with @xmath86 and @xmath137 .
the mode-1 and the mode-4 are almost itb and olb modes in the weak coupling region , respectively . as the coupling , @xmath33 , increases ,
the ilb and otb components are gradually mixed in both the modes . in the case of the mode-1 , this mixing is small , while in the mode-4 the olb and otb components are more largely mixed , particularly around @xmath136 . in this figure , furthermore , we can also see the level mixing between modes-2 and -3 . at @xmath132 the mode-2 and
the mode-3 are the ilb and otb modes . as the coupling increases , the itb and olb components are more mixed in the mode-2 and the mode-3 , respectively . at @xmath133 the level mixing occurs . the strength of the components is exchanged between the modes-2 and -3 .
in this paper we study the collective breathing oscillations in the deformed traps .
the oscillations contain the four kinds of the oscillation , itb , otb , ilb and olb modes .
when the shape of the traps are largely deformed , either oblate or prolate , these modes are decoupled . as this shape becomes close to being spherical ,
the transverse and longitudinal modes are coupled , and make the monopole and quadrupole oscillations .
the borders between these coupling and decoupling are at @xmath138 and @xmath139 , where the otb and ilb modes make level crossing . in the deformed traps the frequencies of the itb and ilb modes agree with twice of the trap frequencies and are not so different from those of the non - interacting system .
the frequencies of the otb and olb modes show some coupling dependence and give information of the interaction , their qualitative behaviors are however similar to those in the dipole oscillation @xcite .
in the semi - spherical traps , on the other hand , the collective oscillations exhibit more complicated behavior because of the level mixing .
hence we can get significant information on many body system of fermi gas by investigate breathing oscillations with the various fermion - fermion couplings and deformations of the trap near being spherical ( @xmath140 ) . in this work we do not discuss extremely deformed system ( @xmath141 or @xmath142 ) .
in such extreme conditions the system becomes a quasi - low dimension one @xcite , and the collective oscillation can be expected to show quite different behaviors from those in our work @xcite . for reviews ,
see : a.s .
parkins and h.d.f .
walls , phys .
rep . * 303 * , 1 ( 1998 ) ; f. dalfovo , s. giorgini , l.p .
pitaevskii and s. stringari , rev .
mod . phys . * 71 * , 463 ( 1999 ) ; + w. ketterle , d.s .
durfee , and d.m stamper - kum , in _ bose - einstein condensation in atomic gases _ proc . of internationa school of physic
enrico fermi , edited by m. ingusico , s. stringari and c. wieman ( ios press , amsterdam , 1999 ) m. h. anderson , et al .
science * 269 * , 198 ( 1995 ) ; + k. b. davis , et al .
lett . * 75 * , 3969 ( 1995 ) .
b. demarco and d.s .
jin , science * 285 * , 1703 ( 1999 ) ; + s.r .
granade , et al .
lett . * 88 * , 120405 ( 2002 ) .
tuscott , k.e .
strecker , w.i .
mcalexander , g.b .
parridge , and r.g .
hullet , science * 291 * , 2570 ( 2001 ) ; + f. schreck , et al .
, phys . rev .
lett . * 87 * , 080403 ( 2001 ) ; + z. hadzibabic , et al .
lett . * 88 * , 160401 ( 2002 ) ; * 91 * , 160401 ( 2003 ) ; + t. maruyama , h. yabu and t. suzuki , phys . rev .
* a72 * , 013609 ( 2005 ) .
bertsch , nucl . phys . *
a249 * , 253 ( 1975 ) ; + g.f .
bertsch and k. stricker , phys . rev .
* c13 * , 1312 ( 1976 ) ; + d.m .
brink and leobardi , nucl . phys . *
a258 * , 285 ( 1976 ) ; + t. suzuki , prog .
64 * , 1627 ( 1980 ) . at @xmath145 ( a ) , @xmath146 ( b ) and @xmath103 ( c ) with @xmath36 and @xmath86 .
the solid lines represent the results with the full calculations , and the dashed and dotted lines indicate the frequencies of the transverse and longitudinal breathing modes , respectively . ] with @xmath86 .
the solid and chain - dotted lines represent the in - phase transverse and longitudinal components , respectively , and the dashed and dotted lines denote the out - of - phase transverse and longitudinal components , respectively . ] with the external magnetic fields with @xmath148 @xmath147 for @xmath121 ( a ) , 0.7 ( b ) , 0.5 ( c ) and 0.3 ( b ) . the solid lines represent the results of the mode-1 and the mode-3 , and the chain - dotted lines indicate those of the mode-2 and the mode-4 . ] and
the solid and chain - dotted lines represent the in - phase transverse and longitudinal components , respectively , and the dashed and dotted lines denote the out - of - phase transverse and longitudinal components , respectively . ] | we investigate collective excitations coupled with monopole and quadrupole oscillations in two - component fermion condensates in deformed traps .
the frequencies of monopole and dipole modes are calculated using thomas - fermi theory and the scaling approximation .
when the trap is largely deformed , these collective motions are decoupled to the transverse and longitudinal breathing oscillation modes . as the trap approaches becoming spherical , however ,
they are coupled and show complicated behaviors . |
an important aim of the study of young stellar objects ( ysos ) has been to classify them in an evolutionary sequence and identify when in this sequence protostellar collapse occurs .
the most widely used classification scheme is that of lada and wilking ( 1984 ) and lada ( 1987 ) which divides ysos into 3 categories based on the slope of their infrared spectral energy distribution ( sed ) .
the oldest population , class iii , have a blackbody - like sed and are interpreted as pre - main - sequence stars without a disk .
class ii sources have a flat or negatively - sloped sed in the infrared and are understood as classical t tauri stars with an optically thick disk of circumstellar dust .
class i sources have an sed that rises with wavelength up to 100 @xmath7 m and are believed to be collapsing protostars ( adams et al .
1987 ; wilking et al.1989a ) .
recently , andr , ward - thompson , and barsony ( 1993 ) ( hereafter awb ) noted that several sources have not been detected in the near - infrared and have blackbody - like seds that peak in the submillimeter , suggesting extremely low dust temperatures ( @xmath8 20 k ) .
they called these sources class 0 ( which they defined as a submillimeter source with l@xmath9/l@xmath10 @xmath11 2 @xmath12 10@xmath13 ) because their cold temperature suggested an object even younger than typical class i objects .
barsony ( 1994 ) listed the following as characteristics of these sources in addition to the low l@xmath9/l@xmath10 : a sed like a 30 k blackbody , undetected at @xmath14 10 @xmath7 m , and a molecular outflow .
chen et al . ( 1995 ) found that class 0 sources have bolometric temperatures ( t@xmath9 ) less than 70 k. t@xmath9 is the temperature of a blackbody with the same mean frequency as the sed of the source ( myers & ladd 1993 ) .
andr and montmerle ( 1994 ) argued that class 0 sources have more circumstellar mass than stellar mass .
they suggested that class 0 sources are still in their active accretion phase , while class i sources have already accreted most of their mass .
if this suggestion is correct , class 0 sources should be good candidates to show kinematic evidence of collapse .
the goal of this project was to survey all of the class 0 objects that had been cited in the astronomical literature for signs of protostellar collapse . since we were unable to determine if every claimed class 0 object meets the criterion of awb or barsony , the selected sample of 23 sources ( see table 1 ) includes any source that has been identified as either a definite or a probable class 0 source .
although this criterion is rather loose , we have checked the credentials of our sources by computing t@xmath9 , using the procedure of myers & ladd ( 1993 ) , for all sources with sufficient data . since many of these sources have few spectral data points , we relaxed one criterion of myers & ladd .
they required six spectral data points but we used as few as four , which seemed to be adequate as long as the photometric data lay on either side of the peak of the sed .
the values of t@xmath9 are given in table 2 , along with information on l@xmath9 , distance , and references .
all 16 of the sources for which t@xmath9 can be calculated lie in the class 0 category ( t@xmath15 k ) within error bars .
the mean t@xmath9 is 44 k and the mean l@xmath9 is 5.7 l@xmath3 , excluding the extremely luminous sources iras 20050 and smm1 .
this sample is primarily one of very embedded , low luminosity objects . for two other sources ,
there are enough photometric data to calculate upper limits ( if data are available on the blue side of the peak ) or lower limits ( if data are available on the red side of the peak ) for t@xmath9 .
some of the sources without sufficient data to compute t@xmath9 or l@xmath9 lie in regions of considerably higher luminosity , where confusion prevents separation of the contributions of different sources in the far - infrared .
the method of spectral line asymmetry ( leung & brown 1977 ; zhou 1992 ) was used to identify collapse candidates . an optically thick line in a collapsing core shows a blue asymmetry in the line profile toward the collapse center , while an optically thin line appears symmetric .
the optically thick lines we used were the @xmath1 and @xmath2 transitions of hco@xmath0 , and the optically thin lines were those of rarer isotopomers of hco@xmath0 , like h@xmath5co@xmath0 and hc@xmath16o@xmath0 .
clumpy outflows can also produce peaks that may be either red or blue ; we address the possible confusion between infall and outflow in 4.3 .
observations were made of twenty - three class 0 sources at the 10.4-m telescope of the caltech submillimeter observatory ( cso ) at mauna kea , hawaii in 1994 october , 1994 november , 1994 december , 1995 march , 1995 june , 1995 december , and 1996 june .
the sources are listed in table 1 with their celestial coordinates and the off position used for position switching .
the off positions were checked for absence of detectable hco@xmath0 emission .
we used sis receivers operating at 230 and 345 ghz , and the backend was a 1024-channel , acousto - optic spectrometer with a bandwidth of 49.5 mhz . the frequency resolution varied over the five runs from about 2.5 to 3 channels , which corresponds to a velocity resolution of 0.14 to 0.17 at 260 ghz and of 0.10 to 0.12 at 357 ghz .
chopper - wheel calibration was used to obtain the antenna temperature , @xmath17 .
the lines we observed and their frequencies are listed in table 3 along with the main beam efficiency , the velocity resolution , and the beamsize at each frequency .
the efficiencies were calculated from numerous observations of planets .
data from separate runs were resampled to the resolution of the run with the worst frequency resolution before averaging .
a first order baseline was removed before scans were averaged .
seven sources which had noticeable asymmetry ( either to the blue or red side ) at the central position were mapped ( map sizes ranged from 30 by 30 to 60 by 90 ) and observed in at least one transition of h@xmath5co@xmath0 .
twenty - three sources were observed in this survey .
all twenty - three sources were observed in the hco@xmath0 @xmath4 line .
eighteen sources were observed in the hco@xmath0 @xmath1 line .
nineteen and four sources were observed in the h@xmath5co@xmath6 @xmath4 and @xmath18 lines , respectively .
seven sources were observed in the @xmath4 line of hc@xmath16o@xmath6 .
nine sources showed blue asymmetry in both the the hco@xmath0 @xmath1 and @xmath2 lines .
three sources had a red asymmetry in both lines and one source , l1157 , had opposite asymmetry in each of the two lines .
five sources showed no asymmetry in both lines .
five other sources were observed only in the @xmath4 line .
of the five sources observed only in the @xmath4 line , three showed no asymmetry , one showed a red asymmetry , and one showed a blue asymmetry .
the line properties are listed in table 4 .
@xmath17 is the highest temperature in the line profile . for singly - peaked lines ,
the velocity and @xmath19v are the centroid and width of the line , found by fitting a single gaussian to the line profile . for lines that have two peaks , we list two values of @xmath17 and velocity , one for each peak , and we give one value for the line width , which is the full width across the spectrum at the temperature where the weaker peak falls to half power .
this core in perseus ( @xmath20 = 300 pc ) is coincident with irs 3 and lies in the outflow lobe of l1448-c .
an h@xmath21o maser coincides with the peak of the ammonia emission ( anglada et al.1989 ) .
bachiller et al . (
1990 ) mapped the outflow in co and found it to be quite weak and not well collimated .
a clear red asymmetry is seen in hco@xmath0 @xmath4 ( figure 5 ) .
bachiller et al . (
1990 ) discovered this highly collimated outflow ( @xmath20 = 300 pc ) which had no known exciting source .
curiel et al .
( 1990 ) found a radio continuum source at the center of the outflow .
line asymmetries consistent with infall have been seen in the 85 ghz line of c@xmath22h@xmath21 ( mardones et al .
the hco@xmath0 @xmath1 and @xmath2 lines display an asymmetry with the red peak stronger than the blue peak .
the h@xmath5co@xmath0 @xmath4 line peaks on the red side of the self - absorption dip , but the h@xmath5co@xmath0 @xmath1 line peaks on the red peak of the hco@xmath0 @xmath1 line ( figures 3 and 4 ) , suggesting that the blue peak is an optically thin clump of outflowing gas . in support of this interpretation
, we note that the velocity of the blue peak in the hco@xmath6 spectrum is well outside the line profile seen in c@xmath22h@xmath21 by mardones et al .
consequently , our data do not provide evidence for or against the suggestion by mardones et al . that l1448-c may be collapsing .
jennings et al .
( 1987 ) discovered ngc 1333 iras 4 ( @xmath20 = 350 pc ) in iras cpc observations .
submillimeter and millimeter photometry by sandell et al .
( 1991 ) revealed a double source ( iras 4a and 4b ) with a 31 separation and a sed that peaks at about 500 @xmath7 m .
recent interferometer observations have shown that iras 4a is itself a binary with a separation of 18 ( lay et al .
our hco@xmath0 @xmath1 and @xmath2 data show a blue asymmetry , with the h@xmath5co@xmath0 @xmath4 line slightly offset to the blue from the self - absorption dip ( figure 1 ) .
this source ( @xmath20 = 350 pc ) appears to be a system of more than two stars ( lay et al .
a strong blue asymmetry is seen in the hco@xmath0 @xmath4 line ( figure 1 ) .
the red peak is very weak in the hco@xmath0 @xmath1 line ( figure 2 ) .
the slight blue offset of the h@xmath5co@xmath0 @xmath4 line from the absorption dip that was seen in iras 4a is also present here .
this source ( @xmath20 = 140 pc ) lies in a lobe of the famous l1551 outflow .
moriarty - schieven et al .
( 1995a ) identified this as an object between class i and class 0 ; we find a t@xmath9 of 75@xmath2318 k. cs @xmath4 and @xmath24 observations show a blue asymmetric line profile ( moriarty - schieven et al .
our hco@xmath0 @xmath4 observations show a multi - peaked line profile ( figure 5 ) .
the blue peak in our hco@xmath0 @xmath4 spectrum has the same velocity as that observed in the double - peaked cs spectra of moriarty - schieven et al .
( 1995b ) .
the h@xmath5co@xmath0 @xmath4 line has two peaks and the hco@xmath0 @xmath4 has three .
the spectrum is too complex to decide whether collapse is occurring .
the spectral energy distribution of this source ( @xmath20 = 140 pc ) is quite similar to those of b335 and l483 ( ladd et al .
zhou et al . ( 1994 ) , mardones et al.(1994 ) , and myers et al.(1995 ) have presented observations that show infall asymmetry in the h@xmath21co @xmath25 line , the 85 ghz line of c@xmath22h@xmath21 , and the h@xmath21co @xmath26 line , respectively .
our hco@xmath0 @xmath1 and @xmath2 data also show a prominent infall signature . the h@xmath5co@xmath0 @xmath1 and @xmath4 lines peak in the self - absorption dips of the hco@xmath0 @xmath1 and @xmath2 line profiles ( figures 1 and 2 ) .
no emission has been detected from this source ( @xmath20 = 400 pc ) at 12 or 25 @xmath7 m ( iras psc ) and its spectral energy distribution peaks at about 100 @xmath7 m ( zinnecker et al.1992 ) .
the derived @xmath27 of 36 k confirms that it is a class 0 source .
our hco@xmath0 @xmath4 data show a symmetric line ( figure 5 ) .
these two sources ( fir 5 and fir 6 ) ( @xmath20 = 415 pc ) were found by mezger et al .
( 1988 ) who suggested that they were isothermal protostars .
a unipolar jet has been seen near this source but the identification of fir 5 as the driving source is uncertain ( richer et al .
the identification of fir 5 and 6 as isothermal protostellar condensations is controversial because warm gas is present in these cores , suggesting that the objects may be more evolved than mezger et al .
( 1988 ) believed ( lis et al .
cs observations suggest that these two sources are massive enough to form stars ( chandler and carlstrom 1996 ) .
no collapse asymmetry is seen in the hco@xmath0 @xmath4 line but a low velocity shoulder was detected ( figure 5 ) .
a bipolar outflow with a very young dynamical time scale ( @xmath28400 yr ) is present ( richer 1990 ) .
an h@xmath21o maser is associated with this object ( @xmath20 = 415 pc ) ( johnston et al .
barnes and crutcher ( 1990 ) observed a torus around this source , perpendicular to the the outflow ( see also lis et al .
1991 ) . like ngc 2024
fir 5 , a low velocity shoulder was detected in the hco@xmath0 @xmath4 line ( figure 5 ) .
this source ( @xmath20 = 400 pc ) is within the l1630 cloud .
gibb and heaton ( 1993 ) detected a co outflow centered on this source .
bontemps et al .
( 1995 ) identified this source as either a class 0 or a class i source .
the hco@xmath0 @xmath1 line at the central position has a clear blue asymmetry but the hco@xmath0 @xmath4 line is only slightly asymmetric ( figures 1 and 2 ) .
this source ( @xmath20 = 400 pc ) is located in the gum nebula and is the exciting star of hh120 .
it has a steep spectral index in the infrared ( persi et al .
persi et al .
( 1994 ) noted that the 1.3 mm continuum flux was similar to those of other class 0 sources .
our calculated t@xmath9 , 74@xmath2315 k , puts it on the boundary of the class 0 category .
our observations show a symmetric line in hco@xmath0 @xmath1 and and a slight blue asymmetry in @xmath4 ( figures 5 and 6 ) .
this source ( @xmath20 = 125 pc ) , the prototype for the class 0 category , has a highly collimated co outflow that extends over 15 in length ( andr et al . 1990a ; dent et al .
awb ( 1993 ) argued that the low temperature , high extinction , and massive envelope of vla 1623 were evidence of an extremely young object .
our observations are ambiguous regarding collapse .
the hco@xmath0 @xmath4 and hco@xmath0 @xmath29 lines both have a three - peaked spectrum with red asymmetry ( figures 3 and 4 ) .
the h@xmath5co@xmath0 @xmath4 line peaks between the middle and the red peak of the hco@xmath0 @xmath4 line ( figure 3 )
. these line profiles probably result from confusion with clumps in the outflow .
walker et al .
( 1986 ) reported the first observations of spectral line signatures of infall in this source ( @xmath20 = 125 pc ) .
menten ( 1987 ) disputed this interpretation and claimed that the asymmetry was a rotational effect .
however , zhou ( 1995 ) has modeled this core as collapse with rotation .
there is an elongated structure in the central region which was originally attributed to a disk - like structure ( mundy et al.1986 ) but is now known to be a binary system with a separation of 5 ( wootten 1989 ) .
the hco@xmath0 @xmath1 and @xmath4 lines are quite strong ( 15 k and 18 k , respectively ) and have a clear infall signature ( figures 1 and 2 ) .
even the h@xmath5co@xmath0 @xmath4 line shows central self - absorption ( figure 1 ) suggesting it is optically thick , as van dishoeck et al .
( 1995 ) found for the @xmath1 line .
our detection of the hc@xmath16o@xmath0 @xmath4 line ( figure 7 ) would agree with the interpretation of van dishoeck et al .
the hc@xmath16o@xmath0 @xmath4 line profile may have two peaks , suggesting self - absorption .
the co outflow ( @xmath20 = 200 pc ) was observed by parker et al .
( 1991 ) who noted its similarity to that of b335 in that it has some blue emission in the red lobe and vice versa .
fuller and myers ( 1993 ) traced the dense gas in the cloud with the hc@xmath22n @xmath1 line .
an h@xmath21o maser has been detected by wilking et al .
fuller et al.(1995 ) calculated a t@xmath9 of 46 k , similar to the 48@xmath235 k we calculated .
myers et al.(1995 ) observed a double - peaked line profile with a stronger blue peak in the h@xmath21co @xmath25 and @xmath26 lines .
the hco@xmath0 @xmath4 line profile has a red - peaked spectral line profile with h@xmath5co@xmath0 @xmath4 peaking to the red side of the dip , while the hco@xmath0 @xmath1 line displays a slight red asymmetry ( figures 3 and 4 ) . unlike the case of l1448-c , the velocities of the red and blue peaks in hco@xmath6 agree with those in the h@xmath30co spectra of myers et al . the fact that hco@xmath0 has a red asymmetry weakens the case for collapse .
this source in serpens ( @xmath20 = 310 pc ) is also known as firs1 ( harvey et al.1984 ) .
casali et al .
( 1993 ) detected this source , discovered the next three sources in submillimeter and millimeter continuum photometry of the serpens molecular cloud core , and noted the lack of near - ir counterparts , a trait that many other class 0 objects share .
however , they also pointed out that this could be a more evolved object because a nearby jet shows reflected 2 @xmath7 m emission from a possible hot source at the center . hurt et al .
( 1996b ) observed a double - peaked and symmetric line profile with a dip at 9 in the @xmath31 line of h@xmath21co .
the hco@xmath0 @xmath1 and @xmath4 lines reveal quite strong self - absorption of a symmetric line suggesting a large optical depth ( figures 5 and 6 ) , but there is no collapse signature , in agreement with hurt et al .
( 1996b ) .
this source ( @xmath20 = 310 pc ) and the following two sources were discovered in the submillimeter continuum maps of casali et al .
hurt et al .
( 1996b ) saw a slight scallop at 8.75 in the @xmath31 line of h@xmath21co .
the hco@xmath0 @xmath1 and @xmath4 lines display a strong blue asymmetry ( figures 1 and 2 ) .
the h@xmath5co@xmath0 @xmath4 line peaks in the self - absorption dip of the hco@xmath0 @xmath4 line which agrees in velocity with the dip in the 218 ghz h@xmath21co line .
the 218 ghz h@xmath21co line shows two peaks ( hurt et al .
a slight blue asymmetry can be seen in the hco@xmath0 @xmath4 line at the central position ( @xmath20 = 310 pc ) .
the blue asymmetry is also present in the hco@xmath0 @xmath1 line , but it is not as prominent ( figure 2 ) .
the h@xmath5co@xmath0 @xmath4 line has two velocity components at the same velocity as the two peaks of the hco@xmath0 @xmath4 line ( figure 1 ) , suggesting that two cloud components exist in this region , rather than collapse .
the hc@xmath16o@xmath0 @xmath4 line ( figure 7 ) has one component , but higher signal - to - noise observations are needed .
a blue outflow wing can be seen in the hco@xmath0 @xmath4 line , but the central line core ( @xmath20 = 310 pc ) displays no asymmetry in either the hco@xmath0 @xmath1 or @xmath4 line ( figures 5 and 6 ) .
davidson ( 1987 ) identified this far - infrared and submillimeter source ( @xmath20 = 300 pc ) .
the outflow of this cloud has a quadrupolar structure ( goldsmith et al . 1984 ) ,
which avery et al .
( 1990 ) interpreted as the limb - brightened walls of a cavity .
anglada et al.(1991 ) found two radio continuum sources within l723 , suggesting two outflows , but cabrit and andr ( 1991 ) reported a single source .
the class 0 identification ( cabrit and andr ) is supported by our calculation of t@xmath32 k. our hco@xmath0 observations show a symmetric line ( figures 5 and 6 ) .
b335 is a well - studied class 0 source ( @xmath20 = 250 pc ) .
zhou et al.(1990 ) mapped the 6 cm h@xmath21co line in this source and found a density gradient consistent with the shu ( 1977 ) model of inside - out collapse .
more direct evidence of collapse was found by zhou et al .
( 1993 ) with the observation of an asymmetric spectral line signature in the cs @xmath33 and @xmath4 lines and the 140 and 225 ghz lines of h@xmath21co .
choi et al .
( 1995 ) have modeled the line profiles of zhou et al .
( 1993 ) as inside - out collapse .
velusamy et al . (
1996 ) have seen clumpy structure in the infalling envelope .
our central position is taken from the 2.7 mm continuum observations of chandler and sargent ( 1993 ) .
our hco@xmath0 observations also show an infall signature ( figure 1 and 2 ) .
this core ( @xmath20 = 700 pc ) was included in the sample of cold iras sources of wilking et al .
( 1989b ) .
the value of 69@xmath34 for t@xmath9 places it barely within the category of class 0 objects .
however , it is considerably more luminous than the other sources .
recently , bachiller et al . ( 1995 ) discovered an outflow with three pairs of lobes suggesting more than one central star .
one outflow has an extremely high velocity jet with molecular bullets like l1448-c .
the cs @xmath4 and c@xmath35s @xmath4 line observations of bachiller et al .
( 1995 ) show the optically thin c@xmath35s peaking at 6.1 , quite close to the cs self - absorption dip at 6 .
chen et al . (
1996 ) have detected a cluster of 200 near - infrared sources .
the hco@xmath0 @xmath4 data show a broad double - peaked line toward the center .
the h@xmath5co@xmath0 @xmath4 line peaks in the absorption dip of the hco@xmath0 @xmath4 line but is rather broad ( figure 1 ) .
the hco@xmath0 @xmath1 line shows a slight blue asymmetry toward the central position ( figure 2 ) .
this source ( @xmath20 = 600 pc ) lies in the bipolar nebula s106 .
s106 is bisected by a dark lane that has possibly contains a disk ( bally et al .
however , more recent observations have shown that was what believed to be a disk is actually several clumps ( barsony et al .
1989 , richer et al .
this source was first detected by richer et al . in continuum maps at 450 , 850 , and 1100 @xmath7 m at a position 15 west of s106ir .
its position coincides with an h@xmath30o maser ( stutzki et al . 1982 ) .
no asymmetry was detected , but broad line wings , possibly from an outflow , were seen in the hco@xmath0 @xmath1 and @xmath4 lines ( figures 5 and 6 ) .
this source ( @xmath20 = 440 pc ) has a well - collimated co outflow in which the blue lobe of the outflow seems to be shock - heating the ambient gas ( umemoto et al .
tafalla and bachiller ( 1995 ) pointed out that the sed is quite similar to that of a class 0 object ; we find t@xmath36 k. our hco@xmath0 @xmath4 data shows the required blue asymmetry for collapse , but the hco@xmath0 @xmath1 line shows the opposite asymmetry ( figure 3 and 4 ) .
the blue outflow wing is quite clear at the center position in hco@xmath0 @xmath4 .
the h@xmath5co@xmath0 @xmath4 line peaks in the self - absorption dip of the hco@xmath0 @xmath4 line .
the line profiles in nine sources are consistent with predictions of line profiles from collapsing clouds .
caution is suggested by the fact that three sources show the opposite asymmetry and one shows different asymmetries in different lines . in this section , we will present some predictions for the expected shape of hco@xmath6 lines in collapsing clouds .
we will then use these predicted line profiles to get quantitative measures of the line asymmetry which can be compared with the same measures for the observations .
we will then discuss the effects of outflows on the spectra and present some detailed modeling of one source , l1527 . based on these considerations
, we will construct a truth table ( table 6 ) to help decide which sources are good collapse candidates ( 5 ) .
zhou ( 1992 ) has modeled the behavior of cs lines as a function of time in a collapsing cloud .
he found that the cs lines get broader , weaker , and more asymmetric as the collapse proceeds , as a result of the rising velocities and dropping densities inside the infall radius .
because hco@xmath6 has different optical depth and excitation properties , the detailed behavior of its line profiles may differ from those of cs .
consequently , we have modeled the evolution of the hco@xmath0 @xmath4 and @xmath1 line profiles , as well as the h@xmath5co@xmath0 @xmath4 and @xmath1 lines , with time .
we used the density and velocity fields of the inside - out collapse model ( shu 1977 ) and a temperature distribution derived from scaling that used for b335 ( zhou et al .
1990 ) to a luminosity of 6.5 l@xmath3 , the average luminosity of the sources with well - determined bolometric temperatures .
the sound speed , @xmath37 , was set at 0.21 .
altering the sound speed would slow down or speed up the evolution of the line profile .
models were run for infall radii , @xmath38 , of 0.005 , 0.01 , 0.02 , 0.03 , 0.04 , 0.05 , 0.06 , and 0.08 pc , corresponding to infall times of 2.3@xmath1210@xmath13 yr for the earliest model and 3.8@xmath1210@xmath39 yr for the last .
the models had 30 shells , 15 inside the infall radius and 15 outside the infall radius .
the outer radius , @xmath40 , was 0.2 pc .
changing the outer radius affects the line profile very little .
we used a constant abundance of x(hco@xmath0 ) = 6@xmath1210@xmath41 .
the terrestrial abundance ratio of x(hco@xmath0)/x(h@xmath5co@xmath0 ) = 90 was assumed .
the h@xmath30-hco@xmath0 collisional rate coefficients are taken from monteiro ( 1985 ) for levels up to @xmath42 ; the rate coefficients for @xmath43 and 6 are scaled appropriately from those calculated by green ( 1975 ) for n@xmath21h@xmath0 .
these parameters were used as input into a code which simulates the shu infall model and includes the temperature and abundance distribution .
the resulting output , the density , velocity , kinetic temperature , turbulent width , and x(hco@xmath0 ) at each shell , was then fed into the monte carlo code described by choi et al .
the output of this code , populations of each level of hco@xmath6 in each shell , then became the input to a virtual telescope program that simulated observations with the cso , convolving the emission from each shell with the spatial and velocity resolution of the actual observations ( figure 8) .
the distance to the model cloud was 140 pc , which is the distance to the taurus molecular cloud and only slightly farther than sources in the ophiuchus cloud core ( @xmath20 = 125@xmath2325 pc ) ( de geus et al .
1989 ) .
the line profiles which emerge from these calculations are qualitatively similar to those of cs calculated by zhou ( 1992 ) , but some important differences emerge because of the difference between cs and hco@xmath0 .
the asymmetry between blue and red peaks increases with age , and it is stronger than in the cs profiles . unlike the cs models ,
the intensity of the hco@xmath6 @xmath4 line grows with time until @xmath44 yr because the dropping opacity outweighs the effects of dropping density .
the @xmath1 line also grows initially , but it decreases after @xmath45 yr since it is more sensitive to density .
the h@xmath5co@xmath0 @xmath4 and @xmath1 lines decrease with age and become clearly weaker than the observations after a very short time and before the blue - red asymmetry matches the observations . to get stronger h@xmath5co@xmath6 lines while maintaining an isotope ratio of 90
, we increased the abundance to x(hco@xmath0 ) = 6@xmath1210@xmath46 inside the infall radius , while retaining x(hco@xmath0 ) = 6@xmath1210@xmath41 in the static envelope , based on some models of rawlings et al .
( 1992 ) , which have lower hco@xmath0 abundances in the outer envelope than in the interior .
these models ( figure 9 ) can reproduce the observed h@xmath5co@xmath6 lines , but the hco@xmath6 lines are now very opaque .
one consequence is that the intensities of both @xmath4 and @xmath1 lines increase with time , as the opacity drops , until even later times .
models were also run for a range of distances from 70 pc to 700 pc to check the effect of distance on the observed line profiles , as observed with a beam of fixed angular extent .
the blue / red ratio decreases with distance from as high as 3.52 to as low as 1.59 for our constant abundance model . at 700 pc ,
the line profiles become noticeably asymmetric only at the oldest times .
the linewidth and line strength also decrease with distance .
all of these effects will make collapse harder to detect in more distant clouds of low luminosity .
the more distant sources are often more luminous , and the temperature distribution we used ( appropriate for the average luminosity of our sample ) will underestimate the line strength . because of the dependence on distance , temperature , and abundance , we consider the models presented here only as rough sketches of the evolution of a collapsing cloud .
we can use the models of line profiles to predict the behavior of various quantities with time .
for example , the asymmetry of the line profile can be characterized in various ways .
one is the ratio of the blue peak intensity and the red peak intensity .
the blue / red ratio is plotted versus time in figure 10 for both lines and both models of the abundance .
the blue / red ratio generally increases with time for both models of abundance and both lines , but the values are clearly larger for the model with constant abundance .
thus absolute values of the blue / red ratio can not be used to assign an age without further information on the abundance . for lines which are not clearly double - peaked
, the blue / red ratio will be undefined , but we can still characterize the asymmetry of the line by a measure like the skewness . the skewness is a non - dimensional quantity defined as the ratio of the third moment of a distribution and the three - halves power of the second moment , both normalized by the first moment .
we calculated the skewness as follows : @xmath47 where @xmath48 is the central velocity of the source and @xmath49 is the channel width .
line profiles with a negative skewness have a blue asymmetry , and sources with a positive skewness have a red asymmetry .
the skewness of the model line profiles is also plotted in figure 10 .
the skewness rapidly becomes more negative with time , but eventually reaches a maximum negative value .
we have measured the blue / red ratio and skewness of the observed hco@xmath0 @xmath4 and @xmath1 lines to see how these quantities compare to the model values ( see table 5 ) .
lines with a single peak , even if skewed , or three peaks were assigned a blue / red ratio of one .
skewness was measured in the interval of @xmath48 @xmath23 1.25 . the @xmath48 of the source was calculated from the velocity centroid of the optically thin h@xmath5co@xmath0 line .
this velocity interval was chosen because it covered the central line core for all the sources while excluding as much of the outflow wings as possible .
( changing the velocity interval over which the skewness is measured to 0.95 or 2 causes , at most , a 25% change in the skewness .
the sign is unchanged . )
as can be seen from table 5 and figure 11 , all of the double - peaked blue asymmetric sources , save the hco@xmath0 @xmath1 line in serpens smm2 , have a negative skewness .
one of the two red asymmetric double - peaked sources , l483 , has a positive skewness in @xmath1 and zero skewness in @xmath4 ; and the other , l1448-c , has a negative skewness .
it is possible to have a line with red asymmetry and negative or zero skewness if the line profile displays strong blue wings . in figure 11 , we plot the skewness versus the blue / red ratio .
the observed sources are represented by pentagons and the models as triangles ( @xmath4 line ) or boxes ( @xmath1 line ) connected by solid or dashed lines , according to the same scheme as in fig .
the good collapse candidates ( 4.3 ) are shown as filled pentagons .
the sources with a blue / red ratio greater than unity have negative skewness , as expected .
the @xmath1 lines lie nicely in the region of the models , but many of the @xmath4 lines have larger blue / red ratios than are achieved in the models . in this plot ,
the difference between the two models is small , except that the model with constant abundance extends to much higher blue / red ratios and covers the region with quite a few sources in the @xmath1 line . in fig
12 , we show the blue / red ratio versus the peak line temperature .
this plot reveals another difference between the data and the models : a number of sources have considerably stronger lines than are seen in either model .
the arrow at the right indicates iras 16293 , which lies off the plot .
some of these sources have higher luminosity and consequently higher temperatures than we have modeled ; higher temperatures would produce stronger lines . for the @xmath1 line
, we can imagine that intermediate values of abundance may match the sources with blue / red ratios above two and line strengths greater than 2 k , but the high blue / red ratios in the @xmath4 line may be harder to explain .
this figure also reveals that high peak temperatures and high blue / red ratios are hard to achieve with the models we have run probably because the assumed luminosity is lower than that of many of the sources .
many of the spectra in figures 16 show substantial high velocity wings , which are not produced in the infall models of the last section .
models with faster infall or enhanced abundances of hco@xmath6 in the innermost parts of the infall might produce stronger line wings , but these would be concentrated toward the center of the infall .
maps of the distribution in space of the emission in different velocity ranges can distinguish between infall and bipolar outflow .
the top two panels of figure 13 are integrated intensity maps of one of our infall models . the simulated map , like the actual observations , was made with 15 spacing .
the integrated intensity of the two peaks are centered on the infrared source and the red peak is more compact than the blue peak ( fig . 13 ) .
the emission in the line wings should be very compact .
( the dashed contours are the integrated intensity of the red - shifted peak and the red line wing . ) figure 13 also shows maps of the intensity integrated over different velocity intervals in l1527 .
the integrated blue and red line wing intensities are peaked on either side of the source , indicating that the line wings are due to outflow lobes and are clearly not related to collapse .
the map of the velocity interval corresponding to the blue peak is fairly circular and peaks near the center position , but the red peak is somewhat displaced , indicating some contamination by the outflow .
several other sources ( notably ngc 1333 iras 4a , serpens smm4 , l483 , and l1448-c ) have bipolar outflow emission in hco@xmath6 , based on the maps of the line wings ( figures 13 to 15 ) .
if bipolar outflows dominate the line wings , they might also dominate the peaks . in this case
a high blue / red ratio would not reflect collapse , but rather the relative strength of the two bipolar lobes in the beam centered on the source . in that case
, we might expect an equal number of sources to have stronger red peaks .
considering only the sources with two clearly distinct peaks , there are 7 with stronger blue peaks and 2 with stronger red peaks ; for a sample of nine objects with intrinsically equal probability for either peak to be stronger , the probability is only 0.07 that 7 will have stronger blue peaks .
thus it is unlikely , but not impossible , that the peaks are all caused by outflows . despite the statistical evidence that not all the double - peaked spectra are caused by outflows ,
the chance of contamination by outflows is serious enough that we do not consider the hco@xmath6 data alone to be sufficient evidence for collapse .
instead we take a blue asymmetry in both hco@xmath6 @xmath1 and @xmath4 as an indication of possible collapse worthy of further investigation . of the nine sources with blue asymmetry , three ( b335 , iras 16293 , and l1527 )
have been previously suggested as collapse candidates , based on other lines . of the other six ,
serpens smm2 can probably be eliminated from consideration because the h@xmath5co@xmath0 @xmath4 line has two velocity components , which have nearly the same velocities as the two peaks of the hco@xmath0 @xmath4 line . for the others
, we can compare the integrated intensity maps of the peaks to those produced by an infall model .
in ngc 1333 iras 4a and 4b , the blue peak is centered on iras 4a , but the red peak increases in intensity to the north . in serpens
smm4 , as in l1527 , the two peaks are separated by about 10 ( fig .
14 ) , in the same direction that the peaks are displaced in the maps of the line wings , suggesting contamination from outflow emission . in hh25mms and
iras 20050 , the maximum of one peak stretches over a larger area than the other peak ( figure 15 ) .
since rotation will also cause a shift of peaks , these sources can not be ruled out as collapse candidates , but more evidence would be needed .
it is noteworthy that both l483 and l1448-c , which have blue / red ratios less than one , have strong bipolarity in the integrated line wing emission ( fig .
15 ) . in l1448-c ,
the blue peak is displaced from the center by about 15 , supporting our earlier suggestion that it is a feature in the outflow . in l483 , the red peak and the blue peak
are displaced about 20 in an east - west direction , as are the peaks in the maps of the line wing emission .
also , in l483 , the peaks are separated in the direction of the outflow .
table 6 is a truth table , which includes information on whether the lines have the correct asymmetry for collapse and whether the h@xmath5co@xmath6 line peaks in the self - absorption dip .
table 6 also includes information on whether the integrated intensity maps of the line peaks are compatible with collapse , with a n " indicating that the maps of the line peaks are displaced in a way that suggest they are dominated by outflow .
given the uncertainties described above , we did not make detailed models of all the sources with blue asymmetries .
instead we focused on l1527 , a source with relatively good evidence for collapse ( zhou et al .
1994 , myers et al .
1995 , zhou et al.1996 ) , to see if the hco@xmath6 was consistent with previous models . using the procedure described in 4.1 ,
we have modeled the spectra toward l1527 .
all models had an outer radius of 0.2 pc , a temperature distribution appropriate for the source luminosity ( see 4.1 ) , and constant abundance . by comparing the model profiles to the observed h@xmath5co@xmath0 @xmath1 and @xmath4 line profiles
, we constrained the abundance for a range of infall radii . after scaling down the abundance in the h@xmath5co@xmath0 models by a factor of 5 ,
the predicted profiles for the hc@xmath16o@xmath0 @xmath1 line are consistent with our upper limit of 0.07 k. scaling the abundances that fit the h@xmath5co@xmath6 profiles up by factors of 50 or 90 to model hco@xmath0 , we found the model which best matches the observed line profiles , focusing on the inner part of the profiles to avoid the outflow ( figure 16 ) .
this model has a constant abundance of x(hco@xmath0 ) = 2.5@xmath1210@xmath46 , 50 times the abundance of h@xmath5co@xmath6 .
models using 90 for the isotope ratio typically produced worse fits .
the absorption dip is deeper in the model than in the observations , suggesting that a lower abundance in the static envelope would fit better .
we did not pursue this point , to avoid introducing too many parameters .
the best - fit model has an infall radius of 0.026 pc , in agreement with modeling of single - dish h@xmath30co spectra and interferometer data on @xmath5co by zhou et al .
( 1996 ) , but larger than that found by myers et al .
( 1995 ) , r@xmath50 = 0.015 pc , using different h@xmath30co data .
the wings of the observed lines are clearly not produced by an infall model , and we attribute those to outflow as discussed earlier . also , we have noted earlier that the integrated intensity maps of the line peaks show a @xmath5110 displacement between the two peaks in an east - west direction , as do the peaks in the maps of the line wing emission .
although the peak regions of the l1527 spectrum were used to determine the best - fitting model , there is still much excess emission in the peak regions .
this emission is probably contamination from the outflow and contributes to the displacement of the two peaks seen in figure 13 .
note that the large blue / red ratio in the observations is not well - matched by the models , as was also noted by myers et al .
( 1995 ) in their modeling of the h@xmath21co @xmath26 line .
we have surveyed 23 class 0 sources in hco@xmath0 @xmath4 and @xmath1 .
nine sources ( 39% of the sample ) have a blue asymmetry , making them candidates for protostellar collapse .
three sources have a red asymmetry , and one source has different asymmetries in the two lines . the evolutionary progress of a collapsing core was modeled to provide predictions of what line profiles and statistical quantities that characterize them would look like . in general , the model line profiles are similar to the observed line profiles , but they have difficulty reproducing the large blue / red ratios seen in some sources . hco@xmath0 shows the collapse signature quite well in sources previously identified as collapse candidates . according to chemical models , it may be relatively abundant in collapsing envelopes where other molecules may deplete ( rawlings et al .
1992 ) . however , this molecule has some definite disadvantages .
first , it emits strongly in outflowing gas , confusing collapse signatures .
second , it is a difficult molecule to model because its lines can be quite optically thick . we conclude that it is a good molecule for surveys of possible collapse candidates , because its lines are strong , but confirmation with other molecular lines is needed before collapse can be definitely claimed . referring to table 6 , we find that b335 , iras 16293 , l1527 ( all previously suggested ) , as well as hh25mms , serpens smm4 , and iras 20050 , are good candidates on all the criteria .
ngc 1333 iras 4a and 4b meet the criteria other than the outflow maps .
serpens smm2 fails the criterion that h@xmath5co@xmath6 peaks in the dip .
we reiterate that these six sources are only collapse _
candidates_. before we can definitely say that protostellar collapse is happening in these sources , further work needs to be done . to strengthen the case for collapse , multi - transition observations are needed in other optically thick molecules like cs and h@xmath21co
also , two - dimensional modeling should be done to distinguish between the effects of collapse and rotation .
lllll l1448-n & 03:22:31.5 & 30:34:49.0 & ( 1200,0 ) & 1 l1448-c & 03:22:34.4 & 30:33:35.0 & ( 1200,0 ) & 1 ngc 1333 iras 4a & 03:26:04.78 & 31:03:13.6 & ( 150,120 ) & 2 ngc 1333 iras 4b & 03:26:06.48 & 31:02:50.8 & ( 150,120 ) & 2 l1551-ne & 04:28:50.5 & 18:02:10 & ( 600,0 ) & 3 l1527 & 04:36:49.3 & 25:57:16 & ( 600,0 ) & 4 rno43 mm & 05:29:30.6 & 12:47:35.0 & ( 900,0 ) & 5 ngc 2024 fir 5 & 05:39:13.0 & 01:57:08.0 & ( 1800,0 ) & 6 ngc 2024 fir 6 & 05:39:13.7 & 01:57:30.0 & ( 1800,0 ) & 6 hh25mms & 05:43:34.0 & 00:14:42 & ( 900,0 ) & 7 iras 08076 & 08:07:40.2 & 35:56:07.0 & ( 900,0 ) & 8 vla 1623 & 16:23:24.9 & 24:17:46.3 & ( 0,900 ) & 9 iras 16293 & 16:29:20.9 & 24:22:13.0 & ( 0,900 ) & 10 l483 & 18:14:50.6 & 04:40:49.0 & ( 600,0 ) & 11 serpens smm1 & 18:27:17.3 & 01:13:23.0 & ( 900,0 ) & 12 serpens smm4 & 18:27:24.7 & 01:11:10.0 & ( 900,0 ) & 12 serpens smm3 & 18:27:27.3 & 01:11:55.0 & ( 900,0 ) & 12 serpens smm2 & 18:27:28.0 & 01:10:45.0 & ( 900,0 ) & 12 l723 & 19:15:42.0 & 19:06:49.0 & ( 900,0 ) & 13 b335 & 19:34:35.4 & 07:27:24.0 & ( 600,0 ) & 14 iras 20050 & 20:05:02.5 & 27:20:09.0 & ( 900,0 ) & 15 s106fir & 20:25:32.44 & 37:12:48.0 & ( 900,0 ) & 16 l1157 & 20:38:39.6 & 67:51:33.0 & ( 900,0 ) & 17 llllll l1448-c & 12 - 3500 & 55@xmath239 & 300 & [email protected] & 1 ngc 1333 iras 4a & 50 - 2000 & 34@xmath232 & 350 & [email protected] & 2,3 ngc 1333 iras 4b & 50 - 2000 & 36@xmath232 & 350 & [email protected] & 2,3 l1551-ne & 12 - 1260 & 75@xmath2318 & 140 & [email protected] & 4,5 l1527 & 25 - 800 & 41@xmath236 & 140 & [email protected] & 6,7 rno43 mm & 60 - 1300 & 36@xmath234 & 400 & [email protected] & 6,8 iras 08076 & 12 - 1300 & 74@xmath2315 & 400 & [email protected] & 6,9 iras 16293 & 25 - 3000 & [email protected] & 125 & [email protected] & 6,10,11 l483 & 25 - 450 & 48@xmath235 & 200 & [email protected] & 6,12 serpens smm1 & 20 - 2000 & 45@xmath235 & 310 & 45@xmath233 & 13,14,15serpens smm4 & 800 - 2000 & 35@xmath235 & 310 & [email protected] & 13,14,15serpens smm3 & 60 - 1300 & 38@xmath235 & 310 & [email protected] & 13,14,15 serpens smm2 & 60 - 1100 & 38@xmath235 & 310 & [email protected] & 13,14,15 l723 & 25 - 1300 & 39@xmath234 & 300 & [email protected] & 6,16,17 b335 & 60 - 1300 & 29@xmath233 & 250 & [email protected] & 6,16,18,19,20,21 iras 20050 & 12 - 2700 & 69@xmath237 & 700 & 206@xmath2313 & 6,22 vla 1623 & 350 - 2000 & @[email protected] & 125 & @[email protected] & 11,23l1157 & 25 - 100 & @xmath1144@xmath239 & 440 & @[email protected] & 6l1448-n & 12 - 100 & & 300 & & 6 ngc 2024 fir 5 & 350 - 2700 & & 415 & & 24,25,26 ngc 2024 fir 6 & 870 - 2700 & & 415 & & 24,25,26 hh25mms & 52 - 1300 & & 400 & & 27,28 s106fir & 450 - 1100 & & 600 & & 29 llllll h@xmath16co@xmath0 & @xmath4 & 27 & 0.66 & 0.17 & 255.479389 h@xmath5co@xmath0 & @xmath4 & 26 & 0.66 & 0.17 & 260.255478 hco@xmath0 & @xmath4 & 26 & 0.66 & 0.16 & 267.557620 h@xmath5co@xmath0 & @xmath1 & 20 & 0.62 & 0.13 & 346.998540 hco@xmath0 & @xmath1 & 20 & 0.62 & 0.12 & 356.734288 llllll l1448-n & hco@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] l1448-c & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] &
& h@xmath5co@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & @xmath4 & [email protected] & [email protected] & [email protected] ngc 1333 iras 4a & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & &
h@xmath5co@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & @xmath4 & [email protected] & [email protected] & [email protected] 1333 iras 4b & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & h@xmath5co@xmath0 & @xmath1 & @xmath53 & & & & @xmath4 & [email protected] & [email protected] & [email protected] l1551-ne & hco@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & & [email protected] & [email protected] & & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] l1527 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & h@xmath5co@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & @xmath4 & [email protected] & [email protected] & [email protected] & h@xmath16co@xmath0 & @xmath1 & @xmath54 & & rno43 mm & hco@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] ngc 2024 fir 5 & hco@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] ngc 2024 fir 6 & hco@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] hh25mms & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] iras 08076 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & @xmath4 & [email protected] & [email protected] & [email protected] & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] vla 1623 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & & [email protected] & [email protected] & &
h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] iras 16293 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & &
h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & hc@xmath16o@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & hc@xmath16o@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] serpens smm1 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & hc@xmath16o@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] serpens smm4 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & [email protected] & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected]
& hc@xmath16o@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] serpens smm3 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & @xmath4 & [email protected] & [email protected] & [email protected] serpens smm2 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & hc@xmath16o@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] l723 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & @xmath4 & [email protected] & [email protected] & [email protected] b335 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] & hc@xmath16o@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] iras 20050 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & &
h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] s106fir & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & @xmath4 & [email protected] & [email protected] & [email protected] l1157 & hco@xmath0 & @xmath1 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & & @xmath4 & [email protected] & [email protected] & [email protected] & & & [email protected] & [email protected] & & h@xmath5co@xmath0 & @xmath4 & [email protected] & [email protected] & [email protected] llll l1448-n & @xmath4 & [email protected] & 1.00 l1448-c & @xmath1 & [email protected] & 0.47 & @xmath4 & [email protected] & 0.60 ngc 1333 iras 4a & @xmath1 & [email protected] & 4.09 & @xmath4 & [email protected] & 4.17 ngc 1333 iras 4b & @xmath1 & [email protected] & 3.37 & @xmath4 & [email protected] & 4.22 l1551-ne & @xmath4 & [email protected] & 1.00 l1527 & @xmath1 & [email protected] & 2.39 & @xmath4 & [email protected] & 2.02 rno43 mm & @xmath4 & [email protected] & 1.00 ngc 2024 fir 5 & @xmath4 & [email protected] & 1.00 ngc 2024 fir 6 & @xmath4 & [email protected] & 1.00 hh25mms & @xmath1 & [email protected] & 1.30 & @xmath4 & [email protected] & 1.00 iras 08076 & @xmath1 & [email protected] & 1.00 & @xmath4 & [email protected] & 1.00 vla 1623 & @xmath1 & [email protected] & 1.00 & @xmath4 & [email protected] & 1.00 iras 16293 & @xmath1 & [email protected] & 3.17 & @xmath4 & [email protected] & 2.64 l483 & @xmath1 & [email protected] & 0.32 & @xmath4 & [email protected] & 0.47 serpens smm1 & @xmath1 & [email protected] & 0.99 & @xmath4 & [email protected] & 0.96 serpens smm4 & @xmath1 & [email protected] & 2.64 & @xmath4 & [email protected] & 3.05 serpens smm3 & @xmath1 & [email protected] & 1.00 & @xmath4 & [email protected] & 1.00 serpens smm2 & @xmath1 & [email protected] & 1.63 & @xmath4 & [email protected] & 1.16 l723 & @xmath1 & [email protected] & 1.00 & @xmath4 & [email protected] & 1.00 b335 & @xmath1 & [email protected] & 2.25 & @xmath4 & [email protected] & 2.08 iras 20050 & @xmath1 & [email protected] & 1.00 & @xmath4 & [email protected] & 1.41 s106fir & @xmath1 & [email protected] & 1.00 & @xmath4 & [email protected] & 1.00 l1157 & @xmath1 & [email protected] & 0.77 & @xmath4 & [email protected] & 1.14 lllll l1448-n & r & & y & l1448-c & r & r & close & n ngc 1333 iras 4a & b & b & close & n ngc 1333 iras 4b & b & b & y & n l1551-ne & b & & n & l1527 & b & b & y & y rno43 mm & n & & & ngc 2024 fir 5 & n & & n & ngc 2024 fir 6 & n & & n & hh25mms & b & b & y & ?
iras 08076 & n & n & n & vla 1623 & r & r & y & ?
iras 16293 & b & b & y & y l483 & r & r & y & n serpens smm1 & n & n & y & serpens smm4 & b & b & close & y serpens smm2 & b & b & n & ?
serpens smm3 & n & n & & l723 & n & n & & b335 & b & b & y & iras 20050 & b & b & y & ?
s106fir & n & n & & l1157 & b & r & y & n adams , f. c. , lada , c. j. , & shu , f. h. 1987 , , 312 , 788 andr , p. , martin - pintado , j. , despois , d. , & montmerle , t. 1990a , , 236 , 180 andr , p. , montmerle , t. , feigelson , e. d. & steppe , h. 1990b , , 240 , 321 andr , p. & montmerle , t. 1994 , , 420 , 837 andr , p. , ward - thompson , d. , & barsony , m. 1993 , , 406 , 122 anglada , g. , estalella , r. , rodriguez , l. f. , torrelles , j. m. , lopez , r. , & canto , j. 1991 , , 376 , 615 anglada , g. , rodriguez , l. f. , torrelles , j. m. , estalella , r. , ho , p. t. p. , canto , j. , lopez , r. , & verdes - montenegro , l. 1989 , , 341 , 208 avery , l. w. , hayashi , s. s. , & white , g. j. 1990 , , 357 , 524 bachiller , r. , andr , p. , & cabrit , s. 1991 , , 241 , l43 bachiller , r. , cernicharo , j. , martin - pintado , j. , tafalla , m. , & lazareff , b. 1990 , , 231 , 174 bachiller , r. , fuente , a. , & tafalla , m. 1995 , , 445 , l51 bally , j. , snell , r. l. , & predmore , r. 1983 , , 272 , 154 barnes , p. j. & crutcher , r. m. 1990 , , 351 , 176 barsony , m. 1994 in clouds , cores , and low mass stars , ed .
d. p. clemens & r. barvainis ( provo : astronomical society of the pacific ) , 197 barsony , m. & chandler , c. j. 1993 , , 406 , l71 barsony , m. , scoville , n. s. , bally , j. & claussen , m. j. 1989 , , 343 , 212 bontemps , s. , andr , p. , & ward - thompson , d. 1995 , , 297 , 98 cabrit , s. & andr , p. 1991
, , 379 , l25 casali , m. m. , eiroa , c. , & duncan , w. d. 1993 , , 275 , 195 chandler , c. j. & carlstrom , j.e .
1996 , , 466 , 388 chandler , c. j. , gear , w. k. , sandell , g. , hayashi , s. , duncan , w. d. , griffin , m. j. , & hazell , a. s. 1990 , , 243 , 330 chandler , c. j. & sargent , a. i. 1993 , , 414 , l29 chen , h. , myers , p. c. , ladd , e. f. , & wood , d. o. s. 1995a , , 445 , 377 chen , h. , tafalla , m. , greene , t. p. , myers , p. c. , & wilner , d. j. 1997 , , 475 , l163 choi , m. , evans , n. j. , ii , gregersen , e. m. , & wang , y. 1995 , , 448 , 742 cohen , m. , harvey , p. m. , schwartz , r. d. , & wilking , b. a. 1984 , , 278 , 671 curiel , s. , raymond , j. c. , rodriguez , l. f. , canto , j. , & moran , j. m. 1990 , , 365 , l85 davidson , j. a. 1987 , , 315 , 602 de geus , e. j. , de zeeuw , p. t. , & lub , j. 1989 , , 216 , 44 dent , w. r. f. , matthews , h. e. , & walther , d. m. 1995 , , 277 , 193 emerson , j. p. , harris , s. , jennings , r. e. , beichman , c. a. , baud , b. , beintema , d. a. , marsden , p. l. , & wesselius , p. r. 1984 , , 278 , l49 fuller , g. a. & myers , p. c. 1993 , , 418 , 273 fuller , g. a. , lada , e. a. , masson , c. r. , & myers , p. c. , 1995 , 453 , 754 gee , g. , griffin , m. j. , cunningham , c. t. , emerson , j. p. , ade , p. a. r. , & caroff , l. j. 1985 , , 215 , 15p gibb , a. g. & heaton , b. d. 1993 , , 276 , 511 goldsmith , p. f. , snell , r. l. , hemeon - heyer , m. , & langer , w. d. 1984 , , 286 , 599 green , s. 1975 , , 201 , 366 harvey , p. m. , wilking , b. a. , & joy , m. 1984 , , 278 , 156 hurt , r. l. & barsony , m. 1996a , , 460 , l45 hurt , r. l. , barsony , m. , & wootten , a. 1996b , , 456 , 686 jennings , r. e. , cameron , d. h. m. , cudlip , w. , & hirst , c. j. 1987 , , 226 , 461 johnston , k. j. , sloanaker , r. m. , & bologna , j. m. 1973 , , 182 , 67 keene , j. , davidson , j. a. , harper , d. a. , hildebrand , r. h. , jaffe , d. t. , loewenstein , r. f. , low , f. j. , & pernic , r. 1983 , , 274 , l43 lada , c. j. 1987 , in star forming regions , ed . m. peimbert & j. jugaku ( dordrecht : reidel ) , 1 lada , c. j. , & wilking , b. a. 1984 , , 287 , 610 ladd , e. f. , adams , f. c. , casey , s. , davidson , j. a. , fuller , g. a. , harper , d. a. , myers , p. c. & padman , r. 1991 , , 366 , 203 lay , o. p. , carlstrom , j. e. , & hills , r. e. 1995 , , 452 , l73 leung , c. m. & brown , r. l. 1977 , , 214 , l73 lis , d. c. , carlstron , j. e. , & phillips , t. g. 1991 , , 370 , 583 mardones , d. , myers , p. , caselli , p. , & fuller , g. 1994 in clouds , cores , and low mass stars , ed .
d. p. clemens & r. barvainis ( provo : astronomical society of the pacific ) , 192 mcmullin , j. p. , mundy , l. g. , wilking , b. a. , hezel , t. , & blake , g. a. 1994 , , 424 , 222 menten , k. m. , serabyn , e. , gusten , r. , & wilson , t. l. 1987 , , 177 , l57 mezger , p. g. , chini , r. , kreysa , e. , wink , j. e. , & salter , c. j. 1988 , , 191 , 44 mezger , p. g. , sievers , a. w. , haslam , c. g. t. , kreysa , e. , lemke , r. , mauersberger , r. , & wilson , t. l. 1992 , , 256 , 631 monteiro , t. s. 1985 , , 214 , 419 moriarty - schieven , g. h. , butner , h. m. , & wannier , p. g. 1995a , , 445 , l55 moriarty - schieven , g. h. , wannier , p. g. , magnum , j. g. , tamura , m. , & olmsted , v. k. 1995b , , 455 , 190 mundy , l. g. , wilking , b. a. , & myers , s. t. 1986 , , 311 , l75 myers , p. c. , bachiller , r. , caselli , p. , fuller , g. a. , mardones , d. , tafalla , m. , & wilner , d. j. 1995 , , 449 , l65 myers , p. c. & ladd , e. f. 1993 , , 413 , l47 parker , n. d. , padman , r. , & scott , p. f. 1991 , , 252 , 442 persi , p. , ferrari - toniolo , m. , busso , m. , origlia , l. , robberto , m. , scaltriti , f. , & silvestro , g. 1990 , , 99 , 303 persi , p. , ferrari - toniolo , m. , marenzi , a. r. , anglada , g. , chini , r. , krugel , e. , & sepulveda , i. 1994 , , 282 , 233 rawlings , j. m. c. , harquist , t. w. , menten , k. m. , & williams , d. a. 1992 , 255 , 471 richer , j. s. 1990 , , 245 , 24p richer , j. s. , hills , r. e. , & padman , r. 1992 , , 254 , 525 richer , j. s. , padman , r. , ward - thompson , d. , hills , r. e. , & harris , a. i. 1993 , , 262 , 839 sandell , g. , aspin , c. , duncan , w. d. , russell , a. p. g. , & robson , e. i. 1991 , , 376 , l17 shu , f. h. 1977 , , 214 , 488 stutzki , j. , ungerechts , h. , & winnewisser , g. 1982 , , 111 , 201 tafalla , m. & bachiller , r. 1995 , , 443 , l37 umemoto , t. , iwata , t. , fukui , y. , mikami , h. , yamamoto , s. , kameya , o. , & hirano , n. 1992 , , 392 , l83 van dishoeck , e. f. , blake , g. a. , jansen , d. j. , & groesbeck , t. d. 1995 , , 447 , 760 velusamy , t. , kuiper , t. b. h. , & langer , w. d. 1996 , , 451 , l75 walker , c. k. , adams , f. c. , & lada , c. j. 1990 , , 349 , 515 walker , c. k. , lada , c. j. , young , e. t. , maloney , p. r. , & wilking , b. a. 1986 , , 309 , l47 wilking , b. a. , claussen , m. j. , benson , p. j. , myers , p. c. , terebey , s. & wootten , a. 1994 , , 431 , l119 wilking , b. a. , lada , c. j. , & young , e. t. 1989a , , 340 , 823 wilking , b. a. , mundy , l. g. , blackwell , j. h. , & howe , j. e. 1989b , , 345 , 257 wilson , t. l. , mehringer , d. m. , & dickel , h. r. 1995 , , 303 , 840 wootten , a. 1989 , , 337 , 858 zhou , s. 1992 , , 394 , 204 zhou , s. 1995 , , 442 , 685 zhou , s. , evans , n. j. , ii , butner , h. m. , kutner , m. l. , leung , c. m. , & mundy , l. g. 1990 , , 363 , 168 zhou , s. , evans , n. j. , ii , kompe , c. , & walmsley , c. m. 1993 , , 404 , 232 zhou , s. , evans , n. j. , ii , & wang , y. 1996 , , 466 , 296 zhou , s. , evans , n. j. , ii , wang , y. , peng , r. , & lo , k. y. 1994 , , 433 , 131 zinnecker , h. , bastien , p. , arcoragi , j .-
p . , & yorke , h. w. 1992 , , 265 , 726 | we have observed 23 class 0 sources in the hco@xmath0 @xmath1 and @xmath2 lines .
the mean bolometric temperature of the 16 sources with well - determined values is 44 k and the mean luminosity is 5.7 l@xmath3 , excluding two sources of considerably higher luminosity .
nine sources , including three sources previously suggested to be collapsing , have the correct ( blue ) spectral line asymmetry for infall in both lines .
three sources have the opposite ( red ) asymmetry in both lines , and one source , l1157 , has a red asymmetry in hco@xmath0 @xmath1 and a blue asymmetry in @xmath4 .
the rest have no significant or consistent asymmetry .
the h@xmath5co@xmath0 @xmath1 and @xmath2 lines were also observed to find the velocity of the ambient gas , and sources with an interesting line asymmetry were mapped .
a monte carlo code was used to produce an evolutionary sequence of collapsing cloud models of the hco@xmath0 @xmath1 and @xmath2 lines and to compare various diagnostics of the resulting line profiles .
the same code was used to compare infall models to the observations in one source , l1527 .
the results were consistent with previous collapse models . based on integrated intensity maps of the line peaks and wings , as well as the velocity of the h@xmath5co@xmath0 line , we select six of the nine sources with a blue line asymmetry as good _ candidates _ for protostellar collapse .
further evidence is needed to establish that infall is taking place .
the hco@xmath6 spectra are not conclusive because bipolar outflows produce strong emission , which can confuse the issue in any individual source
. however , the predominance of blue asymmetries over red asymmetries is not naturally explained in outflow models , whereas it is expected in collapse models .
= 2 ed at min . , |
the purpose of this paper is to propose and analyze a method based on the riccati transformation for solving a time dependent hamilton - jacobi - bellman equation arising from a stochastic dynamic optimal allocation problem on a finite time horizon , in which our aim is to maximize the expected value of the terminal utility subject to constraints on portfolio composition .
investment problems with state constraints were considered and analyzed by zariphopoulou @xcite , where the purpose was to maximize the total expected discounted utility of consumption for the optimal portfolio investment consisting of a risky and a risk - free asset , over an infinite and finite time horizon .
it was shown that the value function of the underlying stochastic control problem is the unique smooth solution to the corresponding hjb equation and the optimal consumption and portfolio are presented in a feedback form .
she furthermore showed that the value function is a constrained viscosity solution of the associated hjb equation .
classical methods for solving hjb equations are discussed by benton in @xcite . in @xcite , musiela and
zariphopoulou applied the power - like transformation in order to linearize the non - linear pde for the value function in the case of an exponential utility function . in the seminal paper @xcite karatzas _
investigated a similar problem of consumption - investment optimization where the problem is to maximize total expected discounted utility of consumption over time horizon @xmath0 $ ] . for a class of utility functions , they derived explicit solutions to the hjb equation .
however , in our case the aim is to maximize the expected value of the terminal utility from portfolio for a general utility function under constraints imposed on the control function and for the case of nontrivial defined contributions to the portfolio . as consequence
, we have to solve the dynamic hjb equation and , in general , explicit solutions to such nonlinear pde are no longer available . regarding numerical approaches for solving hjb equations associated with portfolio optimization
, we can refer to finite difference methods for approximating its viscosity solution developed and analyzed by tourin and zariphopoulou @xcite , crandall , ishii and lions @xcite , nayak and papanicolaou @xcite . in @xcite , muthamaran and sunil solved a multi - dimensional portfolio optimization problem with transaction costs .
they used finite element method and iterative procedure that converts a free - boundary problem into a sequence of fixed boundary problems . in @xcite , peyrl _ et al .
_ applied a successive approximation algorithm for solving the corresponding hjb equation .
the fixed point - policy iteration scheme for solving discretized hjb equations is discussed in huang _
_ @xcite . in @xcite , witte and reisinger presented a penalty approach for the numerical solution of discrete continuously controlled hjb equations . in our approach
we follow a different approach . rather than solving the fully nonlinear hjb equation directly ,
we first transform it into a quasi - linear parabolic equation by means of the riccati transformation .
we prove existence and uniqueness of a solution to the transformed quasi - linear parabolic equation .
moreover , we derive useful bounds on the solution . these bounds can be interpreted as estimates for the coefficient of risk aversion .
a special attention is put on a solution of an auxiliary parametric quadratic programming problem .
it is shown that the derivative of the value function of such a convex program plays the role of a diffusion coefficient of the quasi - linear equation . although the diffusion function need not be sufficiently smooth , we are able to prove existence , uniqueness and derive useful bounds of classical hlder smooth solutions
. the resulting equation can be solved numerically by an iterative method based on finite volume approximation .
there is an analogy to a solution of fully nonlinear generalizations of the black - scholes equation for pricing derivative securities ( cf .
evovi , stehlkov and mikula @xcite ) and the fully nonlinear hjb equation investigated in this paper . in @xcite jandaka and evovi suggested a numerical method for solving a fully nonlinear generalization of the black scholes equation by means of its transformation to the so - called gamma equation stated for the second derivative of the option price .
in fact , the riccati transformation is the logarithmic derivative of the derivative of the value function . here
we apply the riccati transformation proposed and analyzed in a series of papers by ishimura _
_ @xcite . in the context of a class of hjb equations with range constraints , such
a transformation has been analyzed recently by ishimura and evovi in @xcite where also a traveling wave solution to the hjb equation has been constructed .
concerning numerical methods for solving the transformed quasi - linear parabolic pde there are recent papers by ishimura , koleva and vulkov @xcite where they considered a simplified problem without inequality constraints on the optimal control function .
the paper is organized as follows . in section 2
we formulate the problem of our interest and the motivation behind it .
section 3 is devoted to analysis of the riccati transformation of the hjb equation into a quasi - linear parabolic equation .
the transformed function can be interpreted in terms of the coefficient of relative risk aversion of an investor . in section 4
we analyze a class of parametric quadratic optimization problems .
the goal of this section is to show that the value function is a sufficiently smooth and increasing function .
lipschitz continuity of the derivative of the value function is a crucial requirement for the proof of existence and uniqueness of a classical solution to the transformed quasi - linear parabolic equation presented in section 5 .
we also derive useful bounds of a solution to the cauchy problem for the corresponding quasi - linear parabolic equation . using these bounds and smoothness properties of the value function of the auxiliary parametric quadratic optimization problem ,
we prove existence of a classical hlder smooth solution .
a special semi - explicit solution having the form of a traveling wave is analyzed in section 6 .
such a special solution is then utilized as a benchmark solution in section 7 , where we construct an iterative fully implicit numerical approximation scheme for solving a quasi - linear parabolic equation .
section 8 is devoted to application of the method to construction of an optimal response strategy for the german dax 30 index .
our motivation arises from a dynamic stochastic optimization problem in which the purpose is to maximize the conditional expected value of the terminal utility of a portfolio : @xmath1 , \label{maxproblem}\ ] ] where @xmath2 is the it s stochastic process on the finite time horizon @xmath0 $ ] , @xmath3 is a given terminal utility function and @xmath4 a given initial state condition of @xmath2 at @xmath5 .
the function @xmath6 mapping @xmath7 represents an unknown control function governing the underlying stochastic process @xmath8 . here
@xmath9 for @xmath10 denotes the restriction of the control function @xmath11 to the time interval @xmath12 .
we assume that @xmath13 is driven by the stochastic differential equation @xmath14 where @xmath15 denotes the standard brownian motion and the functions @xmath16 and @xmath17 are the drift and volatility functions depending on the control function @xmath11 .
the parameter @xmath18 represents a constant inflow rate of property to the system whereas @xmath19 is the interest rate .
many european pension systems use @xmath20 , representing regular contribution rate to the saver s pension account as a prescribed percentage of their salary .
for example , @xmath21 in slovakia , @xmath22 in bulgaria , @xmath23 in sweden ( c.f .
@xcite ) . throughout the paper
we shall assume that the control parameter @xmath24 belongs to the compact simplex @xmath25 where @xmath26 .
it should be noted that the process @xmath27 is a logarithmic transformation of a stochastic process @xmath28 driven by the sde : @xmath29 where @xmath30 with @xmath31 .
it is known from the theory of stochastic dynamic programming that the so - called value function @xmath32\ ] ] subject to the terminal condition @xmath33 can be used for solving the stochastic dynamic optimization problem ( [ maxproblem ] ) ( cf .
bertsekas @xcite , fleming and soner @xcite or bardi and dolcetta @xcite ) .
if the process @xmath34 is driven by ( [ processx ] ) , then the value function @xmath35 satisfies the hamilton - jacobi - bellman ( hjb ) equation @xmath36 for all @xmath37 subject to the terminal condition @xmath33 ( see e.g. macov and evovi @xcite or ishimura and evovi @xcite ) . as a typical example leading to the stochastic dynamic optimization problem ( [ maxproblem ] ) in which the underlying stochastic process satisfies sde ( [ processx ] ) one can consider a problem of dynamic portfolio optimization in which the assets are labeled as @xmath38 and associated with price processes @xmath39 , each of them following a geometric brownian motion @xmath40 ( cf .
merton @xcite , browne @xcite , bielecki and pliska @xcite or songzhe @xcite ) .
the value of a portfolio with weights @xmath41 is denoted by @xmath42 .
it can be shown that @xmath43 satisfies ( [ processyeps ] ) .
the assumption @xmath24 corresponds to the situation in which borrowing of assets is not allowed ( @xmath44 ) and @xmath45 .
we have @xmath46 and @xmath47 with @xmath48 and @xmath49 where @xmath50 .
the terminal function @xmath51 represents the predetermined terminal utility function of the investor .
[ rem : merton ] in the case of zero inflow @xmath52 , assumption ( [ processyeps ] ) made on the stochastic process @xmath28 is related to the well - known merton s model for optimal consumption and portfolio selection ( cf .
merton @xcite ) .
however , for merton s model , one has to consider a larger set of constraints for control function @xmath11 .
namely , the simplex @xmath53 has to be replaced by a larger set @xmath54 .
it is worth to note that all results concerning @xmath55 smoothness of the value function @xmath56 ( see theorem [ smootheness ] ) as well as those regarding existence and uniqueness of classical solutions ( see theorem [ existence ] ) and numerical discretization scheme remain true when @xmath53 is replaced by @xmath57 .
following the methodology of the riccati transformation first proposed by abe and ishimura in @xcite and later studied by ishimura _
@xcite , xia @xcite , or macov and evovi @xcite for problems without inequality constraints , and further analyzed by ishimura and evovi @xcite , we introduce the following transformation : @xmath58 [ rem : ara ] the function @xmath59 can be viewed as the coefficient of absolute risk aversion for the value function @xmath60 , representing the intermediate utility function of an investor at a time @xmath61 $ ] ( cf .
pratt @xcite ) . in the original variable @xmath62 ,
denoting @xmath63 , we can deduce that the function @xmath64 is the coefficient of relative risk aversion of the intermediate utility function @xmath65 , which is defined as the ratio : @xmath66 .
[ rem : transactioncosts ] it is worth to note that the pension saving s model model based on the sde ( [ processx ] ) takes into account neither transaction costs nor consumption .
it follows from recent papers by dai _ _ e__t al .
@xcite that a model incorporating these effects leads to a hjb equation in two spatial dimensions . in such a case , transformation based on a simple one dimensional riccati transformation ( [ eq_varphi ] ) is not possible .
suppose for a moment that @xmath67 for all @xmath68 and @xmath61 $ ] .
this assumption is clearly satisfied for @xmath69 if we consider a function @xmath70 which is an increasing and concave function in the @xmath71 variable .
we discuss more on this assumption in section [ sec : existence ] .
now , problem ( [ eq_hjb ] ) can be rewritten as follows : @xmath72 where @xmath73 is the value function of the following parametric optimization problem : @xmath74 if the variance function @xmath75 is strictly convex and @xmath76 linear ( as discussed in section [ sec : motivation ] ) , problem ( [ eq_alpha_def ] ) belongs to a class of parametric convex optimization problems ( cf .
@xcite ) .
suppose that the value function @xmath77 satisfies ( [ eq_hjbtransf ] ) and the function @xmath78 is defined as in ( [ eq_varphi ] ) .
then @xmath78 is a solution to the cauchy problem for the quasi - linear parabolic equation : @xmath79 = 0 , \quad x\in{\mathbb{r } } , t\in[0,t ) , \label{eq_pdephi_1 } \\ & & \varphi(x , t ) = 1 - u^{\prime\prime}(x)/u^\prime(x ) , \quad x\in{\mathbb{r}}. \nonumber\end{aligned}\ ] ] the statement can be easily shown by differentiating ( [ eq_varphi ] ) with respect to @xmath80 and calculating derivatives @xmath81 , @xmath82 , @xmath83 from ( [ eq_hjbtransf ] ) .
indeed , as @xmath84 , we have @xmath85 let us denote @xmath86 then @xmath87 and therefore @xmath88 \partial_x v , \\ \partial^2_x\partial_t v & = & [ \partial^2_x g + \partial_x(g ( 1-\varphi ) ) + ( \partial_x g + g ( 1-\varphi ) ) ( 1-\varphi ) ] \partial_x v .\end{aligned}\ ] ] hence @xmath89 and @xmath90 $ ] , as claimed .
finally , we notice that @xmath91 .
moreover , if @xmath56 is strictly increasing then indeed is a quasi - linear parabolic pde with terminal condition at @xmath69 ( see ladyzhenskaya _
* chapter 1 , ( 2.4 ) ) ) .
conversely , one can construct a solution @xmath60 to the hjb equation ( [ eq_hjbtransf ] ) using a solution @xmath78 satisfying equation ( [ eq_pdephi_1 ] ) . indeed , suppose that the function @xmath78 satisfies ( [ eq_pdephi_1 ] ) .
we can define a function @xmath35 as the unique solution to the first order linear pde satisfying the terminal condition : @xmath92 where the function @xmath93 is given by ( [ ag ] ) .
let us introduce @xmath94 as follows : @xmath95 then , following derivation of ( [ bg ] ) we end up with an equation for the function @xmath96 : @xmath97 hence the difference @xmath98 satisfies a linear pde : @xmath99 . since @xmath100 we deduce @xmath101 for all @xmath68 and @xmath61 $ ] .
but it means that @xmath77 fulfills the fully nonlinear equation : @xmath102 \partial_x v = 0\ , , \quad v(x , t)=u(x ) .
\label{hjb - nonlin}\ ] ] in other words , @xmath35 satisfies hjb equation ( [ eq_hjbtransf ] ) .
consequently , it is a solution to hjb equation ( [ eq_hjb ] ) .
moreover , equation ( [ hjb - nonlin ] ) is a fully nonlinear parabolic equation which is monototone in its principal part @xmath103 .
this way one can deduce that the solution @xmath77 to ( [ hjb - nonlin ] ) is unique . in summary , we have shown that we can replace solving hjb equation ( [ eq_hjb ] ) by solving the auxiliary quasi - linear equation ( [ eq_pdephi_1 ] ) .
[ equivalence ] let @xmath104 be a solution to the cauchy problem ( [ eq_pdephi_1 ] ) .
then the function @xmath60 given by ( [ transform ] ) is a solution to hjb equation ( [ eq_hjb ] ) .
moreover , @xmath105 .
the advantage of transforming ( [ eq_hjb ] ) to ( [ eq_hjbtransf])([eq_alpha_def ] ) is that we can define and compute the function @xmath73 in advance as a result of the underlying parametric optimization problem ( either analytically or numerically ) .
this can be then plugged into the quasi - linear equation ( [ eq_pdephi_1 ] ) which can be solved for @xmath78 , instead of solving the original fully nonlinear hjb equation ( [ eq_hjbtransf ] ) as well as ( [ eq_hjb ] ) . in this way we do not calculate the value function @xmath77 itself . on the other hand , it is only the optimal feedback strategy @xmath11 which is of investor s interest and therefore @xmath77 is not important , in fact .
the optimal strategy @xmath106 can be computed as the unique optimal solution to the quadratic optimization problem ( [ eq_alpha_def ] ) for the parameter values @xmath107 .
in the case of the example of a portfolio consisting of @xmath108 assets , we denote @xmath109 the vector of expected asset returns and @xmath110 the covariance matrix of returns which we assume to be symmetric and positive definite . for the portfolio return and variance
we have @xmath111 and @xmath112 . for @xmath113 , ( [ eq_alpha_def ] )
becomes a problem of parametric quadratic convex programming @xmath114 over the compact convex simplex @xmath115 . in this section
, we shall discuss qualitative properties of the value function @xmath116 for this case . by @xmath117
we denote the space of all functions defined on @xmath118 whose @xmath119-th derivative is lipschitz continuous . by @xmath120
we denote the derivative of @xmath73 w.r . to @xmath78 .
[ smootheness ] let @xmath121 be positive definite and @xmath122 .
then the optimal value function @xmath73 defined as in ( [ eq_alpha_def_quad ] ) is a @xmath55 continuous function .
moreover , @xmath123 is a strictly increasing function and @xmath124 where @xmath125 is the unique minimizer of ( [ eq_alpha_def_quad ] ) for @xmath126 .
the function @xmath127 is locally lipschitz continuous .
first , we notice that the mapping @xmath128 is continuous , which can be deduced directly from basic properties of strictly convex functions minimized over the compact convex set @xmath53 .
let us denote @xmath129 the objective function in problem ( [ eq_alpha_def_quad ] ) . since @xmath130 is a continuous function on the compact set @xmath53 , we have @xmath131 in @xmath11 implies the existence of a unique minimizer @xmath132 to ( [ eq_alpha_def_quad ] ) .
moreover , @xmath133 is continuous in @xmath78 due to continuity of @xmath134 . applying the general envelope theorem due to milgrom and segal ( * ? ? ?
* theorem 2 ) the function @xmath73 is differentiable on the set @xmath118 .
next , we prove that @xmath135 .
the function @xmath136 is linear in @xmath78 for any @xmath137 .
therefore it is absolutely continuous in @xmath78 for any @xmath11 .
again , applying ( * ? ? ?
* theorem 2 ) , we obtain @xmath138 therefore @xmath139 , which is strictly positive on @xmath53 .
hence @xmath140 is a @xmath141 continuous and increasing function for @xmath113 .
local lipschitz continuity of @xmath120 now follows from the general result proved by klatte in @xcite ( see also aubin @xcite ) . indeed , according to ( * ? ? ?
* theorem 2 ) the minimizer function @xmath142 is locally lipschitz continuous in @xmath78 .
hence the derivative @xmath143 is locally lipschitz , as well .
equation ( [ eq_pdephi_1 ] ) is a strictly parabolic pde , i.e. there exist positive real numbers @xmath144 , such that for the diffusion coefficient @xmath120 of equation ( [ eq_pdephi_1 ] ) the following inequalities hold : @xmath145 these inequalities follow directly from ( [ eq_alphader_vzorec ] ) , which is a quadratic positive definite form on a compact set @xmath53 . with regard to ( [ eq_alphader_vzorec ] ) the function @xmath120 attains its maximum @xmath146 and minimum @xmath147 . [ example - dax ] an illustrative example of the value function @xmath56 having discontinuous second derivative @xmath148 is based on real market data and it is depicted in fig . [
fig : alphader_dax ] . in this example
we consider the german dax index consisting of 30 stocks . based on historical data from august 2010 to april 2012
we have computed the covariance matrix @xmath110 and the vector of mean returns @xmath109 .
one can observe that there are at least two points of discontinuity of the second derivative @xmath149 . and its second derivative @xmath148 for the portfolio of the german dax 30 index , computed from historical data , august 2010april 2012 .
source : finance.yahoo.com , title="fig:",scaledwidth=45.0% ] and its second derivative @xmath148 for the portfolio of the german dax 30 index , computed from historical data , august 2010april 2012 .
source : finance.yahoo.com , title="fig:",scaledwidth=45.0% ] in this section we discuss further smoothness properties of the value function @xmath116 in the @xmath78 variable , for the case specified at the beginning of section [ sec : multiportf ] .
we furthermore show that the function @xmath56 is locally a rational function which is concave on an open set .
let us denote @xmath150 the set @xmath151 then @xmath152 and @xmath153 varies over all subsets of active indices , @xmath154 . here
@xmath155 denotes the number of elements of the set @xmath153 .
since @xmath156 is continuous , the set @xmath157 is open .
first , let us consider the case @xmath158 .
if we introduce the lagrange function @xmath159 then the optimal solution @xmath160 and the lagrange multiplier @xmath161 are given by : @xmath162 hence @xmath163 where @xmath164 can be expressed as follows : @xmath165 after straightforward calculations we conclude @xmath166 the inequality @xmath167 follows from the cauchy - schwartz inequality .
notice that @xmath168 unless the vectors @xmath109 and @xmath169 are linearly dependent .
now , if @xmath170 for some subset @xmath171 of active indices , then the quadratic minimization problem ( [ eq_alpha_def ] ) can be reduced to a lower dimensional simplex @xmath172 . hence the function @xmath73 is smooth on @xmath173 and therefore @xmath142 and @xmath73 are given by : @xmath174 for any @xmath175 where @xmath176 and @xmath177 and @xmath178 are constants calculated using the same formulas as in ( [ a1b ] ) and ( [ eq : abc ] ) , where data ( columns and rows ) from @xmath110 and @xmath109 corresponding to the active indices in the particular set @xmath153 are removed .
[ pocastiachder ] the function @xmath179 defined in ( [ eq_alpha_def ] ) is a @xmath180 smooth function on the open set @xmath181 .
it is given by ( [ a1 ] ) for @xmath182 and by ( [ a2 ] ) for @xmath183 where @xmath184 , respectively .
there is a useful information that can be extracted from the shape of @xmath149 . for illustration ,
let us observe points of discontinuity of @xmath149 depicted in fig .
[ fig : alphader_dax ] for the example of the german dax 30 index .
the intervals between the points of discontinuities correspond to the sets @xmath185 . for the portfolio of the german dax 30 index
we obtain the sets of active indices , corresponding to the continuity intervals as summarized in tab .
[ tab : activeindices ] .
high values of @xmath78 represent high risk - aversion of the investor .
there is only one single asset present with a nonzero weight ( equal to one ) in the first interval .
this asset is the most risky one and with highest expected return .
indeed , for lowest values of @xmath78 , investor s risk aversion is low and therefore they do not hesitate to undergo high risk for the sake of gaining high return .
.sets of active indices for the german dax 30 index .
the assets are labeled by 1 - adidas , 15 - fresenius , 16 - fres medical , 21 - linde , 23 - merck , 27 - sap , 30 - volkswagen . [ cols="<,<",options="header " , ] in section [ ex : alpha_discontin ] we showed that the sets of active indices can be identified directly from the function @xmath149 .
moreover , based on proposition [ th : comppsi ] , there is an upper bound on investor s coefficient of absolute risk aversion @xmath186 given by @xmath187 . when the utility function is given as in ( [ eq : ara2 ] ) , we have @xmath188 and so @xmath189 for all @xmath71 and @xmath80 .
hence , only the interval @xmath190
$ ] gives relevant information for the investor .
knowing the sets of active indices computed for @xmath191 $ ] , the investor knows the set @xmath192 } \{i\ |\ \hat{\theta}_i(\varphi ) > 0\}$ ] , i.e. the set of assets which will be entering the optimal portfolio with a nonzero weight . to identify the set @xmath193 on a particular interval ,
it is enough to calculate the optimal @xmath194 in one single point from the given interval .
we proposed and analyzed a method of the riccati transformation for solving a class of hamilton - jacobi - bellman equations arising from a problem of optimal portfolio construction .
we derived a quasi - linear backward parabolic equation for the coefficient of relative risk aversion corresponding to the value function - a solution to the original hjb equation . using schauder s theory we showed existence and uniqueness of classical hlder smooth solutions .
we also derived useful qualitative properties of the value function of the auxiliary parametric quadratic programming problem after the transformation .
a fully implicit iterative numerical scheme based on finite volume approximation has been proposed and numerically tested .
we also provided a practical example of the german dax 30 index portfolio optimization .
we are thankful to professor milan hamala for stimulating discussions on parametric quadratic programming .
this research was supported by the vega project 1/2429/12 ( s.k . ) and eu grant program fp7-people-2012-itn strike - novel methods in computational finance , no .
304617 ( d. . ) .
bielecki , t.r . ,
pliska , s.r . and sheu , s.j .
: risk sensitive portfolio management with cox ingersoll
ross interest rates : the hjb equation .
_ siam j. control and optimization _ , * 44*(5 ) ( 2005 ) , 18111843 .
ishimura , n. , koleva , m. n. and vulkov , l. g. : numerical solution of a nonlinear evolution equation for the risk preference , _ lecture notes in computer science _
6046 , springer - verlag new york , heidelberg , 2011 , 445 - 452 .
ishimura , n. and evovi , d. : on traveling wave solutions to a hamilton - jacobi - bellman equation with inequality constraints , _ japan journal of industrial and applied mathematics _ * 30*(1 ) ( 2013 ) , 5167 .
ishimura , n. and nakamura , m. : risk preference under stochastic environment . in : bmei
2011 - proceedings 2011 international conference on business management and electronic information , vol . 1 , 2011 , article number 5917024 , 668670 .
klatte , d. : on the lipschitz behavior of optimal solutions in parametric problems of quadratic optimization and linear complementarity , _ optimization : a journal of mathematical programming and operations research _ * 16*(6 ) ( 1985 ) , 819831 .
koleva , m. and vulkov , l. : quasilinearization numerical scheme for fully nonlinear parabolic problems with applications in models of mathematical finance , _ mathematical and computer modelling _ * 57 * 2013 , 2564 - 2575 .
ktik , p. and mikula , k. : finite volume schemes for solving nonlinear partial differential equations in financial mathematics . in : finite volumes for complex applications vi problems & perspectives , springer proceedings in mathematics , 2011 , vol .
4(1 ) , 643651 .
ladyzhenskaya , o.a . ,
solonnikov , v.a . and
uraltseva , n.n . :
linear and quasi - linear equations of parabolic type .
translations of mathematical monographs 23 , providence , ri : american mathematical society ( ams ) xi , 1968 , 648 pp .
macov , z. and evovi , d. : weakly nonlinear analysis of the hamilton - jacobi - bellman equation arising from pension saving management , _ international journal of numerical analysis and modeling _ * 7*(4 ) ( 2010 ) , 619638 .
melicherk , i. and ungvarsk , c. : _ pension reform in slovakia : perspectives of the fiscal debt and pension level . _ finance a vr - czech journal of economics and finance , * 54 * ( 910 ) ( 2004 ) , 391404 .
peyrl , h. , herzog , f. and geering , h.p . :
numerical solution of the hamilton - jacobi - bellman equation for stochastic optimal control problems . in : _
wseas int .
conf . on dynamical systems and control _ , venice , italy , november 2 - 4 , 2005 , 489497 . | in this paper we propose and analyze a method based on the riccati transformation for solving the evolutionary hamilton - jacobi - bellman equation arising from the stochastic dynamic optimal allocation problem .
we show how the fully nonlinear hamilton - jacobi - bellman equation can be transformed into a quasi - linear parabolic equation whose diffusion function is obtained as the value function of certain parametric convex optimization problem . although the diffusion function need not be sufficiently smooth , we are able to prove existence , uniqueness and derive useful bounds of classical hlder smooth solutions .
we furthermore construct a fully implicit iterative numerical scheme based on finite volume approximation of the governing equation .
a numerical solution is compared to a semi - explicit traveling wave solution by means of the convergence ratio of the method .
we compute optimal strategies for a portfolio investment problem motivated by the german dax 30 index as an example of application of the method . |
over the past three decades the pomeron model has been very successful in describing a wealth of data on hadronic collisions : the total cross section , elastic scattering , and single diffractive scattering @xcite . as viewed from perturbative qcd the pomeron
may be thought of as the @xmath0-channel exchange of two gluons in a color singlet state @xcite .
this simple picture , to be called the low - nussinov pomeron in the following , has been refined in subsequent years in terms of regeizzed gluon ladders leading to the lipatov hard pomeron @xcite .
one of the defining features of diffractive scattering is the pattern of final state hadrons . in a collider configuration hadronization
occurs in narrow cones around the beam directions while the central rapidity region is free of produced hadrons ( rapidity gap )
. the same phenomenon has recently been observed in _
hard _ scattering events at the tevatron @xcite .
the d0 collaboration has studied a sample of dijet events with jet transverse energies in excess of 30 gev . in about @xmath2% of all events with widely separated jets
no sign of hadronic activity is observed between the two jets .
this observation is consistent with predictions for rapidity gap event rates within the pomeron model @xcite .
bjorken estimated the size of the low - nussinov pomeron exchange contribution and found that a fraction @xmath3% of all dijet events should be due to this mechanism .
the other factor in his estimate is the `` survival probability '' @xmath4 , the probability that the underlying event does not fill the rapidity gap which is present at the parton level .
the survival probability is expected to be of order 10% at the tevatron , with large uncertainties @xcite , leading to a rapidity gap fraction @xmath5 , as compared to the observation at the 0.005 level .
both pomeron and @xmath0-channel photon exchange are believed to lead to rapidity gaps since both scattering processes proceed via the exchange of a @xmath0-channel color singlet object . in the case of @xmath6 exchange in forward @xmath7 scattering
, color coherence @xcite between initial and final state gluon radiation is known to lead to an exponentially suppressed gluon emission probability into the rapidity region between the two final state quarks @xcite .
after hadronization this translates into suppressed hadronic activity between the two quark jets , thus allowing the formation of a rapidity gap .
due to its color singlet structure one expects that the same would occur for the @xmath0-channel exchange of a pomeron .
however , since the low - nussinov pomeron is an extended object with colored constituents , the validity of this analogy is not obvious : gluon radiation may resolve the internal color structure of the pomeron . in this paper
we investigate the conditions ( if any ) under which gluon emission between the two final state quarks is suppressed in the pomeron exchange process .
this radiation pattern is necessary for the interpretation of the sizable rate of rapidity gap events observed by d0 since electroweak processes alone , _
i.e. _ @xmath0-channel @xmath6 , @xmath8 or @xmath9 exchange , have too small a cross section to account for the observed rate @xcite .
we calculate the gluon radiation pattern for the imaginary part of the @xmath0-channel exchange of two gluons in quark - quark scattering .
there are practical reasons why we do not attempt a full calculation .
the lowest order cross section for gluon radiation in two gluon color singlet exchange is of order @xmath10 and hence the full calculation corresponds to a two - loop evaluation of 3-jet production at hadron colliders , which is clearly beyond the scope of the present work .
instead , we aim at a qualitative understanding of the radiation pattern only .
since the imaginary part of the amplitude is known to dominate over the real part in forward , elastic quark - quark scattering via pomeron exchange @xcite , the imaginary part of the emission process should be sufficient for our purposes . the gluon radiation pattern for color singlet two - gluon exchange
is then compared to the ones obtained for ( color singlet ) @xmath6 exchange and for ( color octet ) gluon exchange . for the exchange of a low - nussinov pomeron
we find that the angular distribution of a single radiated gluon can be described in terms of three components : a qed - like contribution which is consistent with the formation of rapidity gaps , a qcd like term with gluon emission in the backwards direction , and a collinear term which describes @xmath11 splitting and subsequent pomeron exchange between the gluon and the second quark .
for the emission of gluons at low transverse momentum the qed - like component is much larger than the qcd - like one .
since these soft gluons dominate the overall gluon radiation , hadronization will produce a pattern similar to the one expected for @xmath0-channel photon exchange .
therefore pomeron exchange should indeed lead to the formation of rapidity gaps in hard scattering events .
for those willing to extrapolate to low @xmath12 this finding gives further credence to a qcd explanation also of the soft pomeron .
the main part of this paper is organized as follows . in section
ii we discuss the most general color structure of the @xmath13 process .
two of the terms in the color decomposition define what we mean by `` color singlet exchange in the @xmath0-channel '' .
we then apply this decomposition to the tree level photon and single gluon exchange processes , which are later needed for comparison , and discuss their radiation patterns .
section iii is technical and describes our calculation of the imaginary part of @xmath13 scattering via two gluon exchange . the resulting radiation pattern is then discussed in section iv and section v gives our conclusions .
we start by analyzing the color structure and the resulting gluon radiation patterns for the scattering of two quarks of different flavor at tree level .
the amplitudes for single gluon and single photon exchange already exhibit the full color structure of the process @xmath14 and these results will be needed later for comparison with pomeron exchange .
we thus study the process @xmath15 where @xmath16 and @xmath17 denote the quark and gluon momenta and their colors are labeled by @xmath18 and @xmath19 . scattering involving anti - quarks
is related by crossing and we do not consider processes with initial state gluons . in general , the amplitude of the @xmath14 process can be written in terms of four orthogonal color tensors , which we denote by @xmath20 , @xmath21 , @xmath22 , and @xmath23 , @xmath24 in an @xmath25 gauge theory they are given explicitly by @xmath26 @xmath22 marks the process where the quark @xmath27 keeps its color .
similarly , @xmath23 multiplies the amplitude for @xmath0-channel color singlet exchange as viewed from quark @xmath28 . within qcd , @xmath20 and @xmath21
correspond to @xmath0-channel color octet exchange as viewed from either of the two scattering quarks .
let us first apply this color decomposition to the tree level qed and qcd amplitudes .
the five feynman graphs for the qcd process are shown in fig .
[ figone ] . lumping the momentum and helicity dependence of the individual feynman diagrams into reduced amplitudes @xmath29 , one obtains for the qcd amplitudes at tree level @xmath30 the non - abelian three - gluon - vertex only contributes to the color octet exchange amplitude @xmath31 . both @xmath31 and @xmath32 vanish for @xmath0-channel photon exchange , while the color singlet exchange amplitudes are given by @xmath33 where @xmath34 and @xmath35 denote the electric charges of the two quarks .
the resulting color and polarization summed squared amplitudes , @xmath36 are shown in fig .
[ qedqcdtree](a ) for both the full qcd and the qed case and for the corresponding color singlet contributions .
shown is the dependence of @xmath37 on the rapidity of the emitted gluon when all other phase space parameters are kept fixed , namely the two quarks are held at @xmath38 gev and pseudorapidities @xmath39 and @xmath40 and the gluon transverse momentum is chosen to be @xmath41 gev .
the details of this choice are irrelevant : forward scattering of the two quarks is a sufficient condition to obtain the qualitative radiation pattern of fig .
[ qedqcdtree ] . in the qcd case
the color octet contributions , via the non - abelian amplitude @xmath42 , lead to enhanced gluon emission in the angular region between the two jets . for @xmath0-channel photon exchange
this region is essentially free of gluons due to color coherence between initial and final state gluon radiation @xcite : gluon emission into the central region is exponentially suppressed as the rapidity distance from the quarks increases .
it is exactly this difference between the qed and the qcd case which lets us expect the formation of rapidity gaps for color singlet exchange in the @xmath0-channel .
single gluon exchange also produces a color singlet exchange piece which is defined by the @xmath43 term in eq .
( [ summ2 ] ) . since its shape is identical to the one for tree level photon exchange ( dashed and dash - dotted lines in fig . [ qedqcdtree](a ) ) , one might be misled to conclude that the color singlet contribution to single gluon exchange also produces rapidity gaps
. however , there is an important difference between the two : the role of the amplitudes @xmath44 and @xmath45 is interchanged in the qcd _ vs. _
qed color singlet exchange amplitudes .
the appearance of @xmath46 in the qed amplitude @xmath47 corresponds to the first two feynman graphs in fig .
[ figone ] , _ i.e. _ emission of the final state gluon from the upper quark line .
color coherence between these two graphs strongly suppresses gluon emission except in the angular region between the initial and final direction of the upper quark @xmath27 ( dashed line in fig . [ qedqcdtree](b ) ) . in
forward quark scattering via photon exchange the color @xmath48 of the initial quark @xmath27 is thus transferred to a low mass color triplet object which emerges close to the original @xmath27 direction . at lowest order
this is the final state @xmath27 , at @xmath49 it is the @xmath50 system .
the situation is thus stable against gluon emission at even higher order for the qed case and gluon radiation is suppressed in the rapidity range between the two final state quarks .
formally , the qcd amplitude @xmath51 also describes @xmath0-channel color singlet exchange as seen from the lower quark line .
proportionality to @xmath45 implies that the gluon is emitted from the lower quark , @xmath28 , and hence preferentially between the initial @xmath28-beam and the final @xmath28 directions ( dash - dotted line in fig . [ qedqcdtree](b ) ) .
thus the color triplet @xmath50 system , into which the initial quark @xmath27 evolves , consists of a widely separated quark and gluon .
higher order corrections will lead to strong gluon radiation into the angular region between the two and thus also into the rapidity range between the two final state quarks .
one thus finds that a color singlet exchange structure as identified by the color tensors is not sufficient to infer the production of a rapidity gap in the gluon radiation pattern .
stability of the pattern against multiple gluon emission is another necessary requirement .
the shapes produced by @xmath52 and @xmath53 may now be taken as a gauge for the radiation pattern produced by @xmath0-channel two - gluon exchange when the color index on the @xmath28-quark line is preserved .
only a qed like pattern can be expected to lead to the formation of a rapidity gap in dijet events .
a complete calculation of gluon emission in @xmath55 scattering via two - gluon exchange involves squaring box - diagrams like the ones in fig .
[ figthree ] and thus corresponds to a two - loop calculation of the 3-jet cross section at hadron colliders , which is clearly beyond the scope of the present work .
since @xmath56 scattering via color singlet two gluon exchange is dominated by its imaginary part at @xmath57 @xcite , we restrict our attention to the imaginary part of @xmath54 , as defined by unitarity , @xmath58 here @xmath59 denotes the number of on - shell intermediate particles and @xmath60 and @xmath61 symbolize the initial and final state .
to lowest order , @xmath62 , only @xmath63 and @xmath64 particle intermediate states contribute to @xmath65 .
representative feynman graphs are depicted in fig .
[ figthree ] .
row a ) shows graphs corresponding to @xmath63 particle intermediate states @xmath66 .
factoring out color factors and coupling constants , the contributions to @xmath67 are given by the products of the tree level amplitudes listed below the individual feynman diagrams . here
@xmath68 corresponds to the 1-gluon exchange graph in @xmath55 elastic scattering and the @xmath69 are the tree level @xmath14 amplitudes of fig .
[ figone ] .
when considering @xmath70 3-particle intermediate states at order @xmath71 the @xmath72 amplitude decomposes into one disconnected particle and a tree level @xmath73 amplitude .
the three possibilities are shown in rows b ) , ( disconnected gluon ) , c ) ( disconnected quark @xmath28 ) , and d ) ( disconnected quark @xmath27 ) of fig . [ figthree ]
. @xmath74 refers to the tree level @xmath56 1-gluon exchange graph and @xmath75 and @xmath76 ( @xmath77 ) correspond to the three tree level graphs contributing to @xmath78 and @xmath79 scattering , respectively . because of the disconnected particle in the @xmath72 amplitudes the phase space integral in the unitarity relation of eq .
( [ unitarity ] ) reduces to a solid angle integral over the direction of one of the two intermediate state particles participating in the scattering , taken in the two particle rest frame . using the abbreviations @xmath44 , @xmath45 , and @xmath80 ( see eq .
( [ ampqcd])(10 ) ) for the gauge invariant combinations of tree level @xmath14 amplitudes , one finds @xmath81 \nonumber \\ & -&\left [ n^2a_i^{(na)}+(n^2 - 2)a_i^{(12)}+(n^2 - 4)a_i^{(34 ) } \right]\ , b_f \nonumber \\ & + & 2a_i^{(34)}\,\left [ ( n^2 - 1)c_f^{(1)}-c_f^{(2)}+n^2c_f^{(3 ) } \right ] \nonumber \\ & + & \left [ ( n^2 - 4 ) a_i^{(12)}+(n^2 - 2)a_i^{(34 ) } \right ] \left [ d_f^{(1)}+d_f^{(2 ) } \right ] \nonumber \\ & + & n^2a_i^{(na)}\left [ d_f^{(1)}-d_f^{(2)}+2d_f^{(3)}\right ] \biggr ) \ ; . \label{m34tot}\end{aligned}\ ] ] because of the massless @xmath0-channel gluon propagators in the feynman graphs of fig .
[ figone ] the phase space integral in eq .
( [ m34tot ] ) actually diverges .
the same problem is already encountered for low - nussinov pomeron exchange in @xmath55 elastic scattering and can be circumvented by replacing the massless gluon propagator by a regularized version which avoids gluon propagation over long distances @xcite .
qcd pomeron models of this kind , with a dynamically generated gluon mass @xcite , have been found to give a good description of available data @xcite .
we approximate these refinements by multiplying the four products @xmath67 in the unitarity relation ( corresponding to the four groups of intermediate states in fig .
[ figthree ] ) by factors @xmath82 for each internal gluon propagator of momentum @xmath27 .
effectively this corresponds to a replacement of a massless propagator by a massive one . since any complete gauge invariant set of amplitudes
is multiplied by a common factor , gauge invariance is preserved by this procedure . finally , the phase space integral of eq .
( [ m34tot ] ) was performed numerically by evaluating individual tree level amplitudes with the aid of the helicity amplitude techniques of ref .
the numerical evaluation of this `` loop''-integral is quite slow ( of order ten minutes for one individual phase space point on our alpha - station ) and does not allow detailed investigations via a monte carlo program .
however , the code is quite adequate for a first qualitative study .
numerical results for @xmath83 are shown in fig .
[ figfour ] .
the square of this color singlet exchange amplitude was obtained for particular kinematic configurations ( the same ones as used in section ii ) where the two quarks are held fixed at @xmath38 gev and pseudorapidities @xmath39 and @xmath40 as indicated by the arrows
. the gluon then is taken at fixed @xmath84 gev and the gluon rapidity is varied between -8 and + 8 .
the solid lines depict the gluon angular distribution for the imaginary part of the two gluon exchange amplitude , @xmath65 .
the distributions follow neither the ones for @xmath0-channel photon exchange ( dashed lines ) nor the ones for color octet single gluon exchange ( dash - dotted lines ) .
rather one finds a strong enhancement in the central region , @xmath85 , which is particularly pronounced for the emission of high transverse momentum gluons .
this enhancement can be traced to feynman graphs like the first one in fig .
[ figfive ] , the @xmath86 term .
since the upper quark , @xmath27 , is fixed at small scattering angle , this diagram corresponds to @xmath11 splitting and subsequent elastic @xmath87 scattering by @xmath0-channel color singlet two gluon exchange , _
i.e. _ by the exchange of a low - nussinov pomeron .
there is no reason why this process should be suppressed except for @xmath88 , which corresponds to large scattering angles in the @xmath87 rest frame .
indeed the fall - off observed in fig .
[ figfour ] at @xmath89 confirms this interpretation . in order to isolate gluon
radiation off `` pomeron exchange '' between the two quark lines one needs to eliminate the collinear contributions involving pomeron exchange between the gluon and the quarks .
a full treatment of the collinear region requires convolution with quark and gluon structure functions . because we do not yet have analytic formulas for the imaginary part or at least a fast monte carlo program ,
we can not at present perform the full calculation
. however , for our qualitative discussion a crude approximation will suffice .
from the full amplitude @xmath65 we subtract the gluon box diagram @xmath86 of fig .
[ figfive ] and the minimal set of additional graphs required for gauge invariance .
we thus define the pomeron amplitude in @xmath13 scattering as @xmath90 \left [ d_f^{(1)}+d_f^{(3 ) } \right ] \biggr ) \ ; . \label{m34pom}\end{aligned}\ ] ] the above choice is fixed not by gauge invariance alone , but rather by the necessity to avoid over - subtraction in all regions of phase space . while the combination @xmath91 for the last factor in eq .
( [ m34pom ] ) is gauge invariant also , the concomitant contribution @xmath92 ( see second graph in fig . [ figfive ] ) is strongly peaked for back - scattered gluons and thus would not correspond to pomeron exchange in @xmath87 scattering .
similarly , the @xmath93-channel quark exchange graphs @xmath94 and @xmath95 ( see fig .
[ figone ] ) are chosen to form a gauge invariant combination with @xmath96 .
the additional pieces are numerically small , unlike @xmath97 and @xmath98 terms which receive strong enhancements from the @xmath99-channel quark propagators and lead to over - subtraction in the @xmath100 and @xmath88 regions , respectively . the resulting radiation pattern in `` pomeron exchange '' between the two quark lines is given in fig . [ figsix ] for three values of the regularizing gluon mass parameter , @xmath101 gev ( solid lines ) , @xmath102 gev ( dotted lines ) , and @xmath103 gev ( dash - double dotted lines ) .
the @xmath41 gev case shows that soft gluon emission strongly resembles the pattern found for @xmath0-channel photon exchange : emission into the gap region between the two quarks and backwards emission are strongly suppressed .
soft gluon radiation in pomeron exchange therefore produces the pattern expected for @xmath0-channel color singlet exchange , and as argued earlier for photon exchange this pattern should be preserved for multiple soft gluon emission .
this strongly suggests that quark scattering by exchange of a hard pomeron indeed leads to the formation of rapidity gaps in dijet events @xcite . for hard gluon radiation ( @xmath104 gev in fig .
[ figsix](b ) ) backward emission , like in the qcd case , dominates and no suppressed radiation into the gap region can be expected at higher orders .
qualitatively , the radiation pattern in pomeron exchange , @xmath105 , can be described as a superposition of a qed - like component , which dominates at small gluon transverse momenta , and a qcd - like component , which dominates in the hard region .
the transition between the soft and hard regions is displayed in fig .
[ figseven ] by showing @xmath83 as a function of @xmath106 for the same quark configuration as in fig .
[ figsix ] . in the back - scattering region
( see fig .
[ figseven](a ) the emission of soft gluons is severely suppressed compared to the tree level qcd case . in the forward region
( see fig .
[ figseven](b ) soft gluon emission in `` pomeron exchange '' essentially follows the shape expected for photon exchange . for large gluon transverse momenta
, however , the qed - like component of @xmath105 is severely suppressed .
the more qcd - like behavior at large @xmath106 may be understood by noting that the phase space integral for color singlet two - gluon exchange in eq .
( [ m34tot ] ) is dominated by the region of small @xmath107 for one of the two gluons .
a hard radiated gluon is insensitive to the color screening by this soft gluon and effectively only the color charge of the second , hard @xmath0-channel gluon is seen , leading to a more qcd - like radiation pattern .
this interpretation is confirmed in figs .
[ figsix ] and [ figseven ] by the curves for different values of the regularizing gluon mass .
larger values of @xmath108 eliminate the soft gluon exchange region where the emitted gluon can resolve the color charges of the two @xmath0-channel gluons .
thus the color singlet structure of @xmath0-channel two gluon exchange gains importance and the radiation pattern becomes more qed - like .
[ figsix ] and [ figseven ] also exhibit the dependence of our results on the regularization of the gluon propagator . in the regions of strong destructive interference in @xmath65 , which are visible in fig .
[ figsix ] , the quantitative results strongly depend on the details of soft gluon exchange and are thus non - perturbative in nature .
the qualitative features , however , suppression of backwards soft gluon emission in color singlet two gluon exchange events and a more qcd - like behavior in the hard region , are insensitive to these uncertainties
. in particular , the variation with @xmath108 is modest in the most important region , forward gluon emission at small @xmath106 .
hence , our main finding , the qed like radiation pattern of soft gluons in pomeron exchange , is unaffected .
we have analyzed the different properties of @xmath0-channel color singlet @xmath109 color octet exchange processes in @xmath55 scattering at next to leading order , namely including gluon bremsstrahlung .
the radiation pattern of emitted gluons is intimately connected , after hadronization , to the angular distribution of produced hadrons and hence directly bears on the question whether rapidity gaps are to be expected in @xmath0-channel color singlet exchange processes .
essentially , we arrive at an affirmative answer , also for the case of color singlet two gluon exchange .
while this answer was expected by many in the field , our calculations reveal a rich internal structure of the pomeron .
in addition , our demonstration had to deal with a number of problems which do not yet appear at the level of @xmath73 processes .
while the definition of @xmath0-channel color singlet exchange is obvious in the case of @xmath56 scattering , the emission of an extra gluon complicates the picture .
the color tensors of the @xmath14 process suggest identification of color singlet exchange as those terms where the color of at least one of the incident quarks remains unchanged . in the qcd case ( single gluon exchange )
this also occurs since the color of the exchanged gluon can be compensated for by emission of a gluon of the same color .
the crucial difference emerges when considering the angular distribution of emitted gluons : @xmath0-channel photon exchange leads to forward gluon emission and hence preserves the spatial distribution of the color charges which are present at lowest order .
by contrast gluons are radiated in the backwards direction in the qcd color singlet term , and multiple emission will wash out the pattern produced by the first emitted gluon .
color singlet and color octet exchange terms do not interfere because they are orthogonal in color . because of their different behaviour under multiple gluon emission any interference between a qed- and a qcd - like `` color singlet pattern '' will be severely suppressed at higher order .
the square of the qed - like component in the color singlet exchange amplitude can hence be treated as the probability density for producing a hard scattering event which may evolve into a rapidity gap signature . for technical reasons ( numerical speed )
we can not yet calculate rapidity gap cross sections , rather we are limited to an analysis at the amplitude square level when considering the @xmath0-channel exchange of two gluons . for @xmath0-channel two gluon exchange
we have calculated the imaginary part of the color singlet exchange amplitude , @xmath65 , and we have analyzed the emission probability for the extra gluon when the quarks are scattered by small angles only . at first sight gluon emission is most likely in the angular region between the two quarks . however , this pattern is solely due to @xmath11 splitting and subsequent @xmath87 scattering via the exchange of a low - nussinov pomeron . only after subtracting this contribution
does the final pattern emerge for gluon emission in @xmath55 scattering via pomeron exchange . for high transverse momentum of the emitted gluon ( of order of the quark momenta )
the radiation pattern is quite similar to the one obtained for single gluon exchange .
hard emitted gluons have too short a wavelength to see the screening of the color charge of the harder exchanged gluon by the second , typically very soft , exchanged gluon .
the low - nussinov pomeron thus reveals itself as an extended object .
hard gluon emission is able to resolve the internal color structure of the pomeron . as the transverse momentum of the emitted gluon
is decreased , a qualitative transition occurs .
the gluon radiation can no longer resolve this internal color structure and hence the pomeron appears as a color singlet object . as a result the emission of a soft gluon ( @xmath110 ) follows a pattern very similar to the one observed for @xmath0-channel photon exchange .
this pattern is expected to lead to the formation of rapidity gaps .
since the overall gluon emission rate is dominated by the soft region , we conclude that two gluon color singlet exchange in dijet events may indeed lead to the formation of rapidity gap events as observed at the tevatron @xcite .
we thank j. d. bjorken , s. ellis , f. halzen , and g. sterman for helpful discussions .
this research was supported in part by the university of wisconsin research committee with funds granted by the wisconsin alumni research foundation , by the u. s. department of energy under contract no .
de - ac02 - 76er00881 , and by the texas national research laboratory commission under grants no .
rgfy9273 and fcfy9212 . | 14.5pt the @xmath0-channel exchange of two gluons in a color singlet state represents the lowest order approximation to the pomeron .
this exchange mechanism is thought to also explain the formation of rapidity gaps in dijet events at the tevatron . at the perturbative level
this requires suppressed gluon emission in the rapidity interval between widely separated jets , analogous to color coherence effects in @xmath0-channel photon exchange . by calculating the imaginary part of the two gluon , color singlet exchange amplitude
we show how this pattern does emerge for gluon emission at small transverse momenta . at large @xmath1 the radiation pattern characteristic for color octet single gluon exchange
is reproduced . |
recently , various noise effects for nonlinear systems have been studied . in stochastic resonance ,
the response of a bistable system or an excitable system to a periodic force is enhanced with the addition of noise .
the stochastic resonance improves signal detection by the superposed noise @xcite .
noise - enhanced frequency locking is observed in stochastic bistable systems driven by a relatively strong periodic force or a chaotic signal and phase diagrams similar to the arnold tongues are obtained @xcite .
noise induced entrainment among coupled oscillators is found experimentally in belousov - zhabotinsky reactions and brain waves @xcite .
stochastic resonance has been studied also in globally or locally coupled systems of many bistable elements and it was shown that the coupling can lead to the enhancement of the response @xcite .
various types of fokker - planck equations have been used to study the stochastic resonance theoretically . on the other hand ,
various types of collective dynamics in globally coupled oscillators have been studied @xcite .
globally coupled phase oscillators under external noises can be studied with a nonlinear diffusion equation for the number density .
the nonlinear diffusion equation is obtained from the mean field approximation of the fokker - planck equation for the @xmath0 oscillators , which is considered to be correct in the globally coupled system in the limit @xmath1 @xcite .
the nonlinear diffusion equation for the globally coupled phase oscillators can be transformed into coupled nonlinear equations for the fourier amplitudes of the number density .
the numerical simulation of the coupled nonlinear equations for the fourier modes is relatively easy .
various nonequilibrium phase transitions were found in the globally coupled phase oscillators using the numerical simulation of the coupled equations for the fourier modes and their bifurcation analysis @xcite . in this paper , we study an extended model of the coupled phase oscillators , in which an external periodic force is added to the model studied in [ 15 ] . each element may exhibit stochastic resonance and stochastic synchronization , however , we study dynamical behaviors of the order parameter . the nonlinear diffusion equation is a deterministic equation , and the stochastic behaviors for each element are averaged out in the description of the order parameter . the resonant response and the synchronization to a periodic force
are more clearly shown in this deterministic system .
at first , we consider a globally coupled bistable system .
each phase oscillator evolves according to an equation @xmath2 where @xmath3 is the phase of the oscillator , @xmath4 and @xmath5 are positive constants , and @xmath6 is the frequency of the periodic force .
there are two stationary solutions : @xmath7 and @xmath8 .
if the coefficient of the second term on the right - hand side is not temporally periodic as @xmath9 but constant @xmath10 , the two solutions are bistable for @xmath11 . in that case , the dynamical system has a potential function @xmath12 , and the solution @xmath7 has a lower potential for @xmath13 and a higher potential for @xmath14 than the solution @xmath15 . a model of a globally coupled system with a noise term
is written as @xmath16 where @xmath0 is the total number of elements , and the second term on the right - hand side represents the periodically forcing term and the third term represents mutual coupling , and the last term represents gaussian white noise characterized by @xmath17 where @xmath18 represents noise strength .
the mean field treatment holds exactly in the limit of @xmath1 .
the normalized number density defined by @xmath19 obeys a nonlinear diffusion equation @xmath20 the @xmath21-periodic function @xmath22 can be expanded as @xmath23 the nonlinear equation ( 3 ) is rewritten with coupled ordinary differential equations , @xmath24 we have performed numerical simulation of eq .
( 4 ) retaining the first 50 modes with the runge - kutta method of timestep @xmath25 .
the order parameter is expressed as @xmath26 by eq .
( 4 ) at @xmath27 ( b ) @xmath28 and ( c ) @xmath29 for @xmath30 and @xmath31 . , width=453 ] in this model , the order parameter takes a real value , that is , @xmath32 .
figure 1 displays three time evolutions of @xmath33 at @xmath34 and 1 for @xmath30 and @xmath31 .
the order parameter is regularly oscillating with the frequency @xmath6 of the external force .
the average value of the order parameter takes a nonzero value at @xmath27 and zero at @xmath28 and 1 .
the amplitude of the periodic oscillation is maximum at @xmath28 .
we have calculated the average values @xmath35 of the order parameter and the temporal fluctuations around the average value , @xmath36 , where @xmath37 . the results are shown in fig . 2 .
the average values of order parameters take nonzero values for @xmath38 .
this is a kind of symmetry breaking phase transition induced by noises .
the amplitude of the periodic oscillation of the order parameter takes a maximum near the phase transition point . the response function to the periodic modulation becomes large in an intermediate range of noise strength .
this is interpreted as a kind of stochastic resonance in this globally coupled bistable system . a phase transition in a globally coupled bistable system without a periodic force and the enhancement of the susceptibility to a weak periodic force
was discussed in [ 9],[10 ] . of the order parameter and ( b ) the temporal fluctuations @xmath39 as a function of @xmath18 for @xmath30 and @xmath31 .
, width=415 ] we have found that a kind of phase transition occurs even for nonzero amplitude of periodic modulation .
next , we consider a globally coupled oscillator system . each phase oscillator is assumed to obey an equation @xmath40 where @xmath3 is the phase of the oscillator , @xmath41 and @xmath5 are positive constants , and @xmath6 is the frequency of the periodic force . if @xmath42 , this equation describes a simple phase rotator , since the solution is @xmath43 , where @xmath44 is the natural frequency .
if @xmath45 , and @xmath46 but @xmath47 , @xmath48 is not constant but always takes a positive value , and this equation describes still a phase rotator .
the natural frequency of the oscillator is given by @xmath49 .
when @xmath45 and @xmath47 but @xmath4 is close to @xmath44 , the oscillator behaves like the relaxation oscillation ( it may be suitable to call it relaxation oscillation of phase ) , since the phase motion is clearly separated into fast motion and very slow motion near @xmath50 ( @xmath51 is an integer ) . on the other hand , when @xmath52 and @xmath45 , the equation has a stable stationary solution given by @xmath53 and all trajectories are attracted to the stable point .
if @xmath45 , and @xmath54 but @xmath4 is close to @xmath44 , the system behaves like an excitable system , since one - cycle rotation is excited by some small perturbation to the stationary state .
the parameter @xmath4 is a parameter which determines a transition from an excitable system to an oscillatory system .
the threshold value is @xmath55 .
the parametric perturbation @xmath56 implies that the threshold value is periodically modulated .
it is known in an experiment of light - sensitive belousov - zhabotinsky reaction that illumination can control the threshold .
the noise entrainment was observed by controlling the illumination @xcite .
the periodic modulation in our model may be related to such periodic modulation of the threshold value in experimental systems . a model of a globally coupled system with a noise term
is written as @xmath57 where @xmath0 is the total number of elements , and the third term represents mutual coupling , and the last term represents gaussian white noise characterized by @xmath58 a coupled phase oscillator model similar to this equation was studied by kim et al .
@xcite . in their model ,
time delay was further assumed . in their system with time delay , there are multistable states , i.e. , fast and slow synchronized states and a desynchronized state .
they studied noise induced transitions among the multistable states in one oscillator system and a coupled system of 10 oscillators .
we do not consider such time delay effect and study the stochastic synchronization using a nonlinear diffusion equation for a coupled system in the limit @xmath59 .
the normalized number density obeys @xmath60 the corresponding coupled ordinary differential equations are written as @xmath61 we have performed numerical simulation of eq .
( 8) retaining the first 50 modes with the runge - kutta method of timestep @xmath62 . in this model ,
the order parameter @xmath63 is a complex variable . by eq .
( 8) at @xmath64 ( b ) @xmath65 and ( c ) @xmath66 for @xmath67 and @xmath31 .
, width=491 ] figure 3 displays three time evolutions of @xmath68 at @xmath69 and 0.12 for @xmath70 and @xmath71 .
the order parameter is entrained to the external periodic force at @xmath65
. however , the motion of the order parameter is not completely entrained to the periodic force at @xmath64 and 0.12 , and quasi - periodic motions appear . of the order parameter oscillation vs. @xmath18 for @xmath67 and @xmath31 .
( b ) fourier amplitude @xmath72 vs. @xmath18 .
( c ) phase diagram for the entrainment in @xmath73 plane .
the entrainment is observed in the parameter range surrounded by two curves .
, width=453 ] of the order parameter oscillation vs. @xmath18 and ( b ) fourier amplitude @xmath72 vs. @xmath18 for @xmath74 and @xmath75 . ,
width=415 ] figure 4(a ) displays the frequency @xmath76 of the motion of the order parameter as a function of @xmath18 for @xmath77 and @xmath71 .
the frequency was calculated from the rotation number of @xmath78 around the origin @xmath79 .
the frequency of the order parameter increases with the noise strength @xmath18 .
it can be understood from the dynamical behavior of each oscillator .
each oscillator behaves like the relaxation oscillation and the phase motion becomes slow near @xmath50 when @xmath80 , since @xmath81 .
an effect of noises is to make the regular slow motion randomly faster , and then the average frequency increases with the noise strength .
the periodic modulation is further applied .
the frequency entrainment is observed between @xmath82 .
this is a kind of the stochastic synchronization , although the motion of the order parameter is not stochastic .
figure 4(b ) displays the fourier amplitude of the order parameter motion with frequency @xmath6 , that is , @xmath83 .
the fourier amplitude takes a maximum in the parameter range where the stochastic synchronization is observed .
figure 4(c ) displays a parameter region in the @xmath73 parameter space , in which the synchronization is observed .
this phase diagram seems to be the arnold tongues discussed in [ 4 - 6 ] , however , the ordinate is not the amplitude @xmath5 of the periodic forcing but the coupling constant @xmath84 , so the meaning is different . for a fixed value of noise strength , the coupling constant also needs to take an intermediate value for the stochastic synchronization to appear , as shown in this figure .
too large coupling constant makes the distribution of the number density too narrow , and noise effects are effectively reduced . even for a parameter range @xmath85 ,
in which each elemental system is purely excitable without noises , noises can induce phase slips and the oscillation is observed on an average .
when the periodic force is further applied , the stochastic synchronization can occur .
we have performed a numerical simulation for @xmath86 and @xmath71 .
figure 5(a ) displays the frequency of the motion of the order parameter as a function of @xmath18 .
the frequency entrainment is observed for @xmath87 .
figure 5(b ) displays the fourier amplitude of the order parameter motion with frequency @xmath88 .
the fourier amplitude also has a maximum in an intermediate value of noise strength .
we have proposed globally coupled phase oscillator models and performed numerical simulations of the corresponding nonlinear diffusion equations for the number density .
the response to the periodic force becomes large in an intermediate range of noise strength .
it can be interpreted as a stochastic resonance or a stochastic synchronization , although the motion of the order parameter is regular .
we have shown numerical results for parametrically perturbed system , but we have also studied an additively perturbed model : @xmath89 and another type of forced model : @xmath90 we have observed similar stochastic synchronization also in these systems . it does not seem to depend on the detailed form of periodic forcing . the stochastic synchronization may be found in many systems driven by noises and periodic forcing , if the parameter for each element is near a transition point between an excitable state and an oscillatory state .
99 b. mcnamara and k. wiesenfeld , phys .
a * 39 * , 4854 ( 1989 ) .
l. gammaitoni , p. hanggi , p jung and f. marchensoni , rev .
mod . phys . * 70 * , 223 ( 1998 ) .
b. j. gluckman , t. i. netoff , e. j. neel , w. l. ditto , m. l.spano and s. j. schiff , phys .
* 77 * , 4098 ( 1996 ) .
b. shulgin , a. neiman and v. anishchenko , phys .
lett . * 75 * , 4157 ( 1995 ) .
a. silchenko , t. kapitaniak and v. anishchenko , phys .
e * 59 * , 1593 ( 1999 ) .
r. rozenfeld , j. a. freund , a. neiman and l. schimansky - geier , phys .
e * 64 * , 051107 ( 2001 ) .
k. fujii , d. hayashi , o. inomoto and s. kai , forma * 15 * , 219 ( 2000 ) .
t. mori and s. kai , phys .
* 88 * , 218101 ( 2002 ) .
p. jung , u. behn , e. pantazelou and f. moss , phys .
rev . a * 46 * , r1709 ( 1992 ) . m. morillo , j. gmez - ordoez and j. m. casado , phys .
e * 52 * , 316 ( 1995 ) .
m. e. inchiosa and a. r. bulsara , phys .
e * 52 * , 327 ( 1995 ) .
j. f. lindner , b. k. meadows , w. l. ditto , m. e. inchiosa and a. r. bulsara , phys .
lett . * 75 * , 3 ( 1995 ) .
y. kuramoto , _ chemical oscillations , waves and turbulence _ ( springer - verlag , berlin , 1984 ) .
s. h. strogatz , physica d * 143 * , 1 ( 2000 ) . s. shinomoto and y. kuramoto , prog .
* 75 * , 1105 ( 1986 ) , h. sakaguchi , s. shinomoto and y. kuramoto , prog .
. phys . * 79 * , 600 ( 1988 ) .
s. kim , s. h. park and h .- b .
pyo , phys .
* 82 * , 1620 ( 1999 ) . | cooperative effects of periodic force and noise in globally coupled systems are studied using a nonlinear diffusion equation for the number density .
the amplitude of the order parameter oscillation is enhanced in an intermediate range of noise strength for a globally coupled bistable system , and the order parameter oscillation is entrained to the external periodic force in an intermediate range of noise strength .
these enhancement phenomena of the response of the order parameter in the deterministic equations are interpreted as stochastic resonance and stochastic synchronization in globally coupled systems .
+ + pacs : 05.40.+j , 05.70.ln , 02.50-r + |
a hallmark of mesoscopic systems is the presence of permanent current upon application of an external flux @xcite .
although a noninteracting electron theory is quite successful in describing qualitatively the features of such an effect , it strongly underestimates the magnitude of the observed permanent currents compared to the observed experimental values @xcite .
one possible way out of this discrepancy would be to include the electron - electron interactions .
in general the combined study of disorder and interactions proves difficult to tackle , so that one has to resort to various approximations . in three dimensions hartree - fock - like calculations @xcite tend to suggest that permanent currents are enhanced by the interactions .
arguments based on level statistics reach the same conclusions @xcite . in one dimension ,
on the other hand , numerical and analytical techniques alike allow to treat the interactions exactly enough so that firm results can be obtained .
surprisingly , however , various independent calculations performed on a spinless hubbard model with nearest neighbor interactions @xcite or long range coulomb interactions @xcite have reached the opposite conclusion , namely that repulsive interactions are detrimental to permanent currents , which , therefore , casts doubts on the validity of the approximate calculations in higher dimensions .
we here consider the full problem of interacting electrons with spin , in the presence of disorder , and examine the permanent currents in such a system .
we will mainly concentrate for the sake of simplicity on a purely local interaction ( hubbard model ) , but will also consider briefly finite range interactions .
the interactions are treated exactly and we use a renormalization group calculation @xcite to take care of the disorder .
we show analytically that the permanent currents are * enhanced * by the presence of repulsive interactions , and that the result of the spinless model , although correct , was an artifact . as a byproduct of the study we also give a very simple formulation of the bethe - ansatz equations for the pure attractive hubbard model .
the plan of the paper is as follows .
section [ meissner ] relates the so called charge stiffness to the permanent currents , and discusses various peculiarities that can occur for finite temperature .
section [ shubbard ] introduces the model and treats the interactions using the bosonization procedure .
this section is merely to fix the notations .
section [ currents ] computes the stiffness using a renormalization group calculation .
section [ bansatz ] focuses on the case of attractive interactions , both with a large @xmath1 expansion and using bethe ansatz , and section [ longrange ] examines an extended ( nearest neighbors ) hubbard model .
finally the conclusions of our study can be found in section [ conclusions ] .
instead of computing the permanent currents it is simpler to focus on the so - called conductivity stiffness @xmath2@xcite , which measures the strength of the drude peak in a macroscopic system @xmath3 .
the stiffness @xmath2 can be related to the change of the energy of the ground state of the system in presence of an external flux by @xmath4 @xmath5 being the ground state energy of a ring in a field . here
@xmath6 denotes the boundary angle @xmath7 where @xmath8 is the flux threading the ring and @xmath9 is the flux quantum . on the other hand , for a mesoscopic system , the permanent current measures the response to a finite flux by @xmath10 therefore the stiffness @xmath2 provides a measure of the permanent currents for small ( or close to a multiple of @xmath11 ) flux since @xmath12 . although the complete calculation of the permanent currents at finite flux is also possible for a one dimensional interacting system , the calculation is more complicated in the presence of disorder , and the stiffness carries enough information for our present purposes . in order to compute the permanent currents , one should in principle distinguish between an odd and even number of electrons , assumed spinless for the present argument . for an odd number of electrons the energy is minimum in zero external flux , and the number of right and left moving electrons
is equal . for an even number of electrons , due to the @xmath13 state , the number of right and left moving electrons can not be equal , and the energy is not minimum at zero flux . if one has non - disordered noninteracting electrons , it is obvious that in that case the minimum of the energy corresponds to half a flux quantum .
an external flux can be incorporated in the wave function by making the usual gauge transformation @xmath14 where @xmath15 is the vector potential .
this transformation modifies the kinetic energy term and absorbs an external flux at the cost of a twist in boundary conditions .
since this transformation is purely local and the interactions and disorder do depend on density only , it will not affect the interaction and disorder terms .
therefore the permanent currents for a system with an even number of electrons , even in the presence of disorder and interactions , will be identical to those of a system with an odd number of electrons , with a shift corresponding to half a flux quantum . in the calculation of the stiffness one
implicitly assumes that the equilibrium state is a minimum of energy with zero current ( equal number of right and left carriers ) .
therefore the stiffness for the case of an even number of electrons , measures the permanent current produced as a function of @xmath16 .
the above connection between the permanent currents and the stiffness is valid only at zero temperature . at finite temperature
the permanent current is given by the derivative of the free energy .
one could think that the stiffness would be given by the second derivative of the free energy , within the canonical ensemble .
but such a quantity is the so called meissner fraction @xmath17 although both of @xmath18 and @xmath2 are related to the current correlation function , they correspond to different limits .
@xmath18 , being a thermodynamic quantity corresponds to the limit @xmath19 first and then to @xmath20 , whereas @xmath2 which is related to transport corresponds to the limits taken in the reverse order . for finite @xmath21
the two quantities are distinct . for a macroscopic system
@xmath18 measures the superfluid density and will be zero for a non - superfluid system , whereas @xmath2 can be non zero if the system is perfectly conducting but not superconducting ( e.g. free electrons in the absence of impurities ) . if one has a finite system then @xmath18 needs not to be zero even if the system is not superconducting .
in general the meissner fraction is given by @xmath22 where @xmath23 denotes a state of the system , @xmath24 is the hamiltonian and @xmath25 the energy of the eigenstates .
we have separated out the contribution from the manifold of degenerate states , which are of especial importance for charge transport . in ( [ stiff ] ) the limits
@xmath26 and @xmath27 do not commute in general .
if one takes the limit @xmath26 first , then @xmath18 tends to zero unless the system is really superconducting , as can be checked by computing explicitly @xmath18 for e.g. free electrons . for a mesoscopic system , where @xmath0 is finite
, @xmath18 will be finite and gives the slope of the permanent currents with the flux @xmath28 .
one now has to distinguish whether we have an odd or even number of electrons , or more generally whether the ground state is not degenerate ( odd number of electrons ) or whether it has a twofold degeneracy ( even number of electrons ) . in the first case , for small enough temperatures , the sums in ( [ stiff ] ) will be dominated by the ground state , other terms being suppressed by factors like @xmath29 , where @xmath30 would be the gap between the ground state and the first excited states .
such a gap remains finite for a systems of finite size .
if the ground state is not degenerate then the average value of the current in zero external flux is zero in the ground state .
one has @xmath31 the last term in ( [ stiff ] ) vanishes and one finds a positive @xmath18 , roughly temperature independent ( dominated by the first term in ( [ stiff ] ) ) .
there is , therefore , a paramagnetic permanent current for small flux . on the other hand , if the ground state has a twofold degeneracy , which occurs e.g. for the case of an even number of electrons , then in each of the degenerate ground states @xmath32 the current can ( and will in general ) be nonzero @xmath33 in that case ( [ stiff ] ) is dominated by the last term , which gives rise to a curie type behavior , @xmath34 .
there will therefore be a diamagnetic permanent current , with a slope proportional to @xmath35 .
many of these properties are well known for free electrons ( see e.g. @xcite and references therein ) but the arguments presented here show that they are much more general and rest only on the degeneracy of the ground state and are valid for interacting electrons as well .
only a short derivation will be given here in order to fix the notations . more details can be found in @xcite .
let us consider for example the discrete 1d hubbard model with @xmath0 sites @xmath36 where @xmath37 stands for nearest neighbors .
using the well known boson representation of fermion operators in one dimension the complete hamiltonian ( [ hubbard ] ) becomes , away from half - filling @xcite @xmath38 where @xmath39 and @xmath40 are defined by @xmath41\ ] ] @xmath42 and @xmath6 are canonically conjugate variables and @xmath43 is a short distance cutoff that can be identified with the lattice spacing .
the @xmath44 and @xmath45 parts of the hamiltonian ( [ complet ] ) respectively describe the charge and spin degrees of freedom of the system .
the @xmath46 term is the scattering between electrons of opposite spins with an exchange of momentum of @xmath47 .
the hamiltonian ( [ complet ] ) describes in fact the most general one dimensional hamiltonian with spin conserving interactions , provided that the proper values of the @xmath48 and @xmath49 parameters are used .
@xmath50 controls the charge excitations .
@xmath51 corresponds to dominant superconducting fluctuations , whereas @xmath52 corresponds to charge- or spin - density wave ( depending on the spin part of the hamiltonian ) dominant fluctuations . for the hubbard hamiltonian ( [ hubbard ] ) the various coefficients in ( [ complet ] ) and ( [ free ] )
are given , if @xmath53 is small compared to @xmath54 , by @xmath55 for a macroscopic system @xmath46 renormalizes to zero and @xmath56 to one for repulsive interactions . in the asymptotic limit
@xmath57 are the only parameters needed to describe the long range properties and have been computed exactly using bethe - ansatz @xcite . for attractive interactions
there is a gap in the spin sector , and only @xmath58 and @xmath50 are needed to describe the low energy properties of the model .
they can also be computed from bethe ansatz @xcite as will be seen in more details in section [ bansatz ] .
the charge stiffness @xmath2 can be obtained @xcite from the bosonized hamiltonian ( [ complet ] ) , and is simply given by @xmath59 .
disorder can be added to ( [ complet ] ) by @xmath61 where @xmath62 is a random potential . as is well known the forward scattering due to the potential
does not affect the conductivity in one dimension ( see e.g. ) and one can retain only the @xmath63 fourier components of the random potential .
higher fourier components are less effective since they do not scatter electrons on the fermi surface , and do not correspond to low energy processes . a notable exception
are @xmath64 components that will be discussed later .
when expressed in term of the bosons variables ( [ dis ] ) becomes @xmath65 where @xmath66 is a complex random potential corresponding to the part of the random potential having fourier components close to @xmath63 .
for simplicity we will take it gaussian @xmath67 the effect of ( [ disbos ] ) can be computed using a renormalization group calculation @xcite , where one changes the cutoff ( lattice spacing ) @xmath43 into @xmath68 .
we will just quote the results here @xmath69 k_\sigma ^2 \\
\frac{d u_\rho(l)}{dl } & = & -\frac{u_\rho^2k_\rho}{2u_\sigma } \delta(l)\\ \frac{d u_\sigma ( l)}{dl } & = & -\frac{u_\sigma k_\sigma } { 2 } \delta(l)\\ \frac{d y(l)}{dl } & = & [ 2 - 2k_\sigma ( l)]y(l ) - \delta ( l ) \\ \frac{d \delta ( l)}{dl } & = & [ 3-k_\rho ( l)-k_\sigma ( l)-y(l ) ] \delta(l ) \label{renormfin}\end{aligned}\ ] ] with the dimensionless quantities : @xmath70 the renormalizations of @xmath71 and @xmath72 are of first order in @xmath30 and @xmath73 , and consequently can be neglected on the right sides of the first three equations .
contrary to the pure case , charge and spin degrees of freedom are now no more decoupled . for a macroscopic system
, the physics implied by the equations ( [ renorm]-[renormfin ] ) has been studied at length @xcite . as can be seen from ( [ renormfin ] )
, there are two regimes depending on the initial values of @xmath74 .
@xmath30 can scale to zero , the system is in that case delocalized , and as shown in @xcite is dominated by divergent superconducting fluctuations . in the other regime
@xmath30 scales to large values , and the corresponding phase can be identified with a localized phase . in that case
the rg equations break down below a certain length scale that can be identified with the localization length . in the limit where @xmath75 , as can be seen from ( [ renormfin ] ) the localized - delocalized transition occurs when @xmath76 if @xmath77 ( since @xmath78 renormalises to zero and @xmath56 to one ) or @xmath79 if @xmath80 . for a mesoscopic system of size @xmath0
, one can expect the size to play the role of an infra - red cutoff in the rg equations .
when the renormalized cutoff @xmath81 , i.e. @xmath82 , one can treat the disorder term in perturbation , and the stiffness is therefore simply given by the quadratic part of the hamiltonian , with the renormalized coefficients @xmath83 .
provided the size @xmath0 is smaller than the length at which the equations cease to be valid one can use them to compute the various values of @xmath84 as function of the size @xmath0 of the system , and from that to obtain the stiffness @xmath2 .
such a calculation is similar to the one performed for a macroscopic system to get the exponent @xmath50 @xcite at finite temperatures . in that case
the cutoff is provided by the dephasing length @xmath85 .
the full dependence of the stiffness on the size of the sample needs a numerical integration of the rg equations , but the qualitative features can be understood by looking at the very small disorder limit . in that case
one can neglect the renormalization of @xmath86 in the equation for @xmath30 , which gives @xmath87 where @xmath88 is the initial value of the disorder . here
we focus on the case of the hubbard model . for small
@xmath53 one can use the values of the parameters ( [ param ] ) and one gets @xmath89 therefore the disorder grows more slowly for the repulsive hubbard model than in the attractive one . in the same limit of a very small initial @xmath30 ,
the stiffness is roughly given by @xmath90 the bare stiffness @xmath91 can be considered roughly independent of the interactions if the system is far enough from half filling as can be seen from ( [ param ] ) .
the dependence in @xmath53 comes only from lattice effects @xcite that breaks galilean invariance and are sensitive , for repulsive interactions , only for fillings close to a commensurate filling ( mainly half filling where the pure system would be a mott insulator ) . for attractive interactions the renormalization of the stiffness of the pure system due to interactions will become much more important and
will be discussed in section [ bansatz ] .
we will ignore in the following the change of the bare stiffness due to the interactions and will only focus on the effects due to the disorder .
as can be seen from ( [ reduc ] ) , the disorder term tends to drastically reduce the stiffness @xmath2 , and this effect will be smaller for the repulsive model than the attractive one and the stiffness ( the permanent currents ) will be * enhanced * by repulsive interactions for a given size and a given disorder
. for finite disorder one has to numerically integrate the rg equations .
the result is shown in figure [ figure1 ] and is in agreement with the simplified analysis above .
this rather counter - intuitive result can be simply understood with the following argument : with the repulsive hubbard model , the ground state is almost a spin density wave ( with a power - law decay of the correlation functions ) whose density is uniform . such a ground state couples very weakly to non - magnetic impurities as is obvious from ( [ dis ] ) . to couple to disorder ,
one has to distort the spin density wave and make a fluctuation of the density , a process that will cost an energy increasing with @xmath53 .
the disorder effect is therefore very weak , at least if the size of the system is not too big . on the other hand
the attractive hubbard model has a ground state that contains charge density wave fluctuations ( although superconductive fluctuations are the dominant ones ) which can get very easily pinned by impurities .
on such a ground state the disorder will act very efficiently and drastically reduce the stiffness compared to the pure value therefore making the permanent currents smaller .
such an argument is in agreement with higher dimensions @xcite .
this is to be contrasted with a previously studied spinless model . in that case
both the attractive and repulsive ground state have density fluctuations , and both can be equally well pinned by disorder . since in the attractive case
the superconducting fluctuations tend to screen the disorder , the stiffness increases with attractive interactions . for the interactions to have a beneficent effect on the permanent currents one must necessary
take a realistic model in which the main effect of the interactions will be to homogenize the density as is the case for the hubbard model .
if the interactions are infinitely repulsive , the system becomes equivalent to a model of spinless fermion with a fermi momentum of @xmath47 . in that case , although the @xmath47 component of the disorder is inefficient ( as is also obvious from the fact that it will no more correspond to a process on the new fermi surface ) , one should worry about the @xmath92 component of the disorder .
such a fourier component acts on the free spinless fermion , so that one recovers the stiffness of free electrons in the presence of disorder .
the crossover between the two regimes would need a detailed analysis of the coupling of the @xmath64 component of the charge density to disorder which is way beyond the scope of this paper , but one could naively expect a maximum of the permanent currents for an intermediate value of the interactions .
if the size of the system becomes large enough the disorder will renormalise to large values and the system will be localized .
this is always the case for repulsive interactions . for attractive interactions ,
a localized - delocalized transition is in principle possible ( for @xmath93 ) and the stiffness could saturate to a finite value .
we will show in the following section that for the particular case of the attractive hubbard model , where one has only an on - site attraction , this transition does not occur and the system remains always localized .
let us consider the case of a large negative @xmath53 .
in that case one would naively imagine that the system should delocalize .
in fact a very large on - site attraction can not delocalize , and * increases * the localization ( and therefore decreases the permanent currents ) . in the @xmath94 limit , one can perform a large @xmath95 expansion . only pairs of particles can hop and if one introduces the operators @xmath96 , the attractive hubbard model then becomes a model of hard core bosons with a hopping @xmath97 and a disorder @xmath98 .
the residual interaction between the bosons is also on the scale of @xmath99 .
in fact using the superexchange formulation or degenerate perturbation theory , the model maps on precisely to the 1-d heisenberg antiferromagnet at a fixed magnetization ( related to the density of particles ) , with an exchange energy @xmath100 . by a jordan wigner transformation
this model corresponds to spinless fermion with a narrow bandwidth and with nearest neighbors interaction , in the presence of the old disorder .
such a system is obviously localized , and since the kinetic energy reduces with @xmath95 one expects the localization length to diminish when the attraction is increased .
one can make the statements more quantitative for finite @xmath53 by studying the rg equations in the attractive regime . in that case
it is well known that there is a gap in the spin excitation spectrum and that only the charge sector remains ungapped .
keeping only the charge excitations into account the rg equations become @xcite @xmath101\delta ( l ) \label{renormattf}\end{aligned}\ ] ] with @xmath102 and @xmath103 is a constant of order unity coming from the @xmath104 correlations in the perturbation expansion .
the equations ( [ renorm ] ) can be used at scales above the size @xmath105 of a cooper pair .
this approach will therefore be adapted for reasonably large @xmath53 . for small @xmath53
it will be better to use the equations ( [ renorm]-[renormfin ] ) , the crossover between the two regimes occurring when @xmath106 . in order to get the stiffness of the disordered system one needs the initial values of @xmath50 and @xmath58 in the absence of disorder as a function of the attraction @xmath53 .
as for the repulsive hubbard model , they can be deduced from the bethe - ansatz solution @xcite .
we will here give a derivation based on an appealing formulation introduced by sutherland ( in the absence of @xmath6 ) involving the formation of cooper pairs which scatter without diffraction .
let the number of particles be @xmath107 , for which case we note the bethe equations for the attractive @xmath53 hubbard model , with energy @xmath108 , where @xmath109 , and @xmath110 satisfy the bethe equations @xcite @xmath111\ ] ] where @xmath112 are integers ( half odd integers ) for m odd ( even ) .
the flux @xmath6 comes in with a factor of @xmath113 due to the charge of the cooper pair .
these equations can also be obtained from the repulsive case @xcite by using a particle hole transformation on the up electrons of a half filled model , the spin excitations then map on to the above equations .
this can be checked explicitly , using the idea of complementary solutions due to woynarovich @xcite , which essentially rests on the recognition that the equations for real @xmath114 in the bethe equations of the repulsive hubbard model are the @xmath0 real zeroes of a polynomial of degree @xmath115 .
hence the residue theorem of cauchy helps in transforming equations involving the real @xmath114 to those over complex @xmath114 .
the complex @xmath114 come in pairs , and are essentially pinned to be @xmath116 , in order to satisfy the growth conditions .
the error involved in writing down the above cooper pair representation is of @xmath117 , with @xmath118 the cooper pair radius . the parameters @xmath58 and @xmath50 can be obtained by computing the compressibility @xmath119 and the charge stiffness @xmath120 of the pure system from the bethe ansatz ground state energy .
various values of @xmath58 and @xmath50 are plotted in figure [ figure2 ] together with the stiffness @xmath2 ( for the pure system ) . as can be guessed from the large @xmath53 expansion @xmath121 at large @xmath53 .
the fact that the parameter @xmath50 remains finite shows that the system remains interacting .
one can check that the limiting value of @xmath50 is in agreement with the one obtained for the @xmath122 chain @xcite on which this system maps in the large @xmath53 limit . due to the reduction of the velocity
, the stiffness of the clean system itself goes to zero at large @xmath53 .
conversely to what happened for repulsive interactions where the stiffness of the pure system was nearly interactions independent , there is here a drastic * reduction * of the stiffness when the attraction is increased @xcite . using equations ( [ renorm ] )
one computes the stiffness in the presence of disorder . here
the main in the decrease of the velocity @xmath58 , which increases the relative strength of the disorder given by the dimensionless parameter ( [ dimensionless ] ) .
some results are shown in figure [ figure3 ] , where we have normalized the stiffness to its value in the absence of disorder to avoid the trivial effect of renormalization of the bare stiffness by attractive interactions . in agreement with the previous section
the reduction of the stiffness due to disorder becomes more and more important as the attraction @xmath53 is increased .
in order to check the validity of the arguments presented here for a slightly more general model than the hubbard model , we also look at an extended hubbard model with a nearest neighbor interaction @xmath123 defined by @xmath124 in that case , and for small @xmath53 and @xmath123 the various parameters entering the equations are @xmath125 for small @xmath53 and @xmath123 the renormalization of velocities is of second order in @xmath126 and can be neglected . when replaced in equations ( [ renormfin ] ) one
gets @xmath127 if @xmath128 the results are unchanged compared to the case of the pure hubbard model . in order to check whether the physical ideas introduced here on the increase of the permanent current due to repulsive interactions are correct , or whether they are an artifact of the purely local hubbard model , one can consider the artificial limit where @xmath129 and @xmath123 remains finite . note that the model does not boil down in that case to the spinless fermion model since the @xmath123 term still introduces interactions among opposite spins . in that case the effect of the interaction depends on the filling . for low filling , a repulsive @xmath123 will tend to favor a spin density wave ground state again , whereas an attractive one would tend to pair particles on neighboring sites , giving a modulation of the density .
for large fillings the situation changes : a positive @xmath123 will now tend to favor two particles on the same site , to avoid paying the repulsion , and therefore to give a charge density wave .
an attractive @xmath123 favoring two particles on neighboring site will this time give a spin density wave .
the change between a sdw towards cdw ground state occurs when @xmath130 and as can be seen from ( [ ksigma ] ) this will occur when @xmath131 .
added to this is the competing effect that the more attractive @xmath123 we have , the more there are superconducting fluctuations in the system which tend to reduce the disorder . above quarter filling the two effects go hand in hand and repulsive interactions are detrimental to the stiffness , whereas below quarter filling the two effects will compete . from ( [ renorv ] )
one can see that the point where a repulsive @xmath123 again becomes favorable to the stiffness is @xmath132 .
below this filling , the fluctuations of the density generated by a repulsive @xmath123 are too strong to be balanced by the superconducting fluctuations and a positive @xmath123 will increase the permanent currents .
we have looked in this paper at the stiffness constant of a hubbard model as a function of the size of the system .
the stiffness constant is directly related to the permanent currents in the presence of an external flux by @xmath133 for small flux .
we have shown that both the attractive and repulsive hubbard model are always localized for a macroscopic system regardless of the strength of the interactions .
in fact , using bethe ansatz solution or for large @xmath1 simple perturbation theory , one shows that the localization length decreases for attractive interactions due to the reduction of charge velocity . for a mesoscopic system ,
the stiffness in the repulsive hubbard model is much less sensitive to disorder than for the attractive one .
therefore the permanent currents are * enhanced * by repulsive interactions .
this surprising result is related to the fact that for the attractive hubbard model the ground state contains strong charge density wave fluctuations that pins easily on the impurities , whereas repulsion favors a uniform density and makes the pinning harder .
this property remains valid for a model with longer range interactions .
in general the effects of the interactions on the permanent currents is controlled by two competing effects .
one is the presence of density fluctuations in the ground state .
the more there will be , the more easily the system will be pinned by disorder and the more the permanent currents will be reduced compared to the pure value . in general repulsive interactions will tend to favor a homogeneous density ( local fluctuations in density will cost an energy increasing with the repulsion ) , and therefore will tend to increase the permanent currents . on the other hand , attractive interactions promote superconducting fluctuations in the system that tend to screen the disorder and therefore tend to increase the permanent currents .
previous studies of one dimensional systems , leading to the conclusion that repulsive interactions reduced permanent currents , were performed on a spinless model .
in such a rather artificial model , the first effect does not occur , since density fluctuations are always present both for attractive and repulsive interactions and therefore repulsive interactions are detrimental to permanent currents . in a more realistic model where the local interactions are
the dominant ones ( interactions in a real system do decrease with distance ! ) , the density effect will dominate and the permanent current are increased .
this study is , strictly speaking , restricted to one - dimensional systems , and a direct comparison of our results with experimental , three - dimensional , rings is not feasible .
it nevertheless suggests that in mesoscopic systems the presence of repulsive interactions can considerably enhance the permanent currents , and confirms in the exactly solvable one - dimensional case , that the increase of the permanent currents is linked to a reduction of the local density fluctuations by the repulsive interactions .
it is therefore tempting to ascribe the discrepancies observed between the measured and the computed ( with a free electron theory ) values of the permanent currents to such an interaction effect , an interpretation compatible with recent perturbative calculations @xcite .
it is a pleasure to thank g. montambaux , h. bouchiat and b. reulet for many illuminating discussions . | we consider a one - dimensional hubbard model in the presence of disorder . we compute the charge stiffness for a mesoscopic ring , as a function of the size @xmath0 , which is a measure of the permanent currents .
we find that for finite disorder the permanent currents of the system with repulsive interactions are larger than those of the system with attractive interactions .
this counter intuitive result is due to the fact that local density fluctuations are reduced in the presence of repulsive interactions . |
in this paper , we generalize the construction of schemes to other algebraic categories .
there are already many attempts of extending the category of schemes , for various reasons .
one of it is from the geometry over @xmath1 : the rough idea is to construct a scheme - like topological object from monoids .
there are numerous attempts towards this goal , and we can not list them all . here , we just mention that there is a survey on this topic @xcite . in this article , we focus on the relation between the construction of ours and those of ton - vaqui @xcite , and deitmar @xcite .
( the latter two constructions are actually equivalent , according to @xcite . )
however , not many have mentioned its universal property of the proposed new objects and categories . to claim that the new objects are good "
, we must make it explicit how it is similar to schemes .
in other words , we should ask to ourselves , we know how schemes are useful , but we do nt know why . despite its heavy and complicated machinary , the category of schemes behaves surprisingly good .
this implies that the construction of schemes , or almost equivalently , the spectrum functor , should be realized as an adjoint of a simpler functor . in this paper
, we go back to this fundamental question : we give the universal characterization of the category of coherent schemes .
the universality can not be obtained , only by specifying the algebraic object : even if we restrict ourselves to rings , there are several ways of defining a scheme - like objects .
therefore , we must designate its grothendieck topology .
also , we must be aware that the category of schemes is not complete , although the category of rings is co - complete .
therefore , in order to give a universal property , we should be able to apply the spectrum functor to schemes : if we call the scheme - like @xmath2-valued spaces as @xmath2-schemes , then we must be able to define @xmath3-schemes .
this implies that the usual definition of schemes is not appropriate : it uses the notion of local rings and local homomorphisms where infinite operations occur , which is not available for schemes in general .
therefore , we give another way of describing this local property , only using finite operations on the value category .
this idea is already mentioned in the previous preprint of the author @xcite .
the advantage of this definition is that we can define the spectrum functor as an adjoint of the global section functor , which is by nature , much simpler .
another advantage is that this gives a way to define a larger category containing all coherent schemes , which is complete : see @xcite . here , we will list up the main results .
let @xmath0 be a category with pull backs and finite coproducts , equipped with a grothendieck topology @xmath4 ( the precise definition is given in 3 ) .
we call the triple @xmath5 a coherent site
. this is _ schematic _
, if it admits finite open patchings ( again , the definition is given in 3 ) .
let @xmath6 , @xmath7 be the ( 2-)categories of coherent sites , schematic coherent sites , respectively . 1 .
the underlying functor @xmath8 admits a left adjoint @xmath9 .
the unit is the spectrum functor @xmath10 .
( theorem [ thm : univ : coh : sch ] ) 2 .
when @xmath0 is the opposite category of commutative rings , with the topology induced from ideals , then @xmath0-schemes are coherent schemes .
( theorem [ thm : sch : local : space ] ) the universal property of the zariski topology is summarized as follows : let @xmath2 be a self enhancing algebraic type with a constant operator , and @xmath11 be the type of commutative monoid objects in the category of @xmath2-algebras . if the morphism class @xmath12 consists of localizations of finite type , then the zariski topology is the coarsest topology satisfying the following condition : for any @xmath11-algebra @xmath13 and @xmath13-module @xmath14 , the induced @xmath15-module @xmath16 is zero if and only if @xmath17 .
note that this paper is only a stepping stone to various generalizations .
however , we decided to rewrite it from the first step : since although the scheme theory is already a classical topic , its universal property is scarcely discussed , and we need it to be clarified to apply its machinary to other workfields .
this paper is organized as follows : in section 2 , we give a brief summary of algebraic types and lattice theories .
the reader may skip this section , if he or she is familiar with the terminology . in 3 , we discuss the universal property of @xmath0-schemes : namely it is given by the universal property of schematic topological category under @xmath0 . in the process , we introduce the notion of weak @xmath0-schemes , which is the generalization of @xmath18-schemes . using this
, we obtain the spectrum functor as an adjoint of the global section functor . in 4 ,
we compare our notion of @xmath0-schemes with the conventional coherent schemes .
this is somewhat time - consuming , since the definitions of the two are very different . here
, we also show the correspondence between the ideals and the grothendieck topology .
finally in 5 , we discuss the universal property of ideal topology .
note that the ground of choosing the ideal topology is only given by looking at modules .
when we talk of algebraic objects other than rings , we know that congruences take place of ideals .
however , the result obtained here shows that this is a groundless fear , since the ideal topology detects modules similarly as in the case of rings .
* notation and conventions : * in the sequel , we assume the reader to be familiar with category - theoretic languages ; the textbook @xcite will be sufficient for this purpose .
we fix a universe , and do not mention it when unnecessary .
any ring is unital , and @xmath19 is the set of non - negative rational integers .
* acknowledgements : * we would like to thank h. minamoto for valuable discussions .
i was partially supported by the grant - in aid for young scientists ( b ) # 23740017 .
here , we begin with the definition of algebraic types . conventionally , an algebraic type is a pair @xmath20 of the set of operators and equational classes .
however , to define equational classes ( and therefore , derived operators ) explicitly , it requires some terminology of _ trees _ , or _ languages_.
to emphasize the relation between algebras and category theories , we decided to describe these in categorical terms , as below . 1 .
let @xmath21 be the category of small categories with finite products ( here , we include null products , namely we also assume the existence of the terminal object ) and product - preserving functors .
we have a left adjoint @xmath22 of the underlying functor @xmath23 which sends a small category @xmath0 to its object set @xmath24 .
2 . we set @xmath25 .
this is actually equivalent to the opposite category of finite sets and maps .
we denote by @xmath26 \in \obj(\scr{x})$ ] the set @xmath27 .
an object in the overcategory @xmath28 is called a _ graded set_. + for an object @xmath29 in @xmath30 ,
we denote by @xmath31 for the inverse image @xmath32 for a non - negative integer @xmath33 .
let @xmath34 be a functor defined by @xmath35 \mapsto [ \amalg_{n \in \nnn}\scr{c}(j[n],j[1 ] ) \to \nnn],\ ] ] where the latter map is determined by sending elements of @xmath36,j[1])$ ] to @xmath33 . a morphism @xmath37 in @xmath38
canonically gives a map @xmath39,j[1 ] ) \to \amalg_{n \in \nnn}\scr{d}(j[n],j[1])$ ] over @xmath19 .
later , we omit @xmath40 if there seems to be no confusion .
then , @xmath41 admits a left adjoint .
it is straightforward to see that @xmath21 is small complete .
therefore , it remains to show that the solution set condition is satisfied to apply freyd s adjoint functor theorem .
suppose @xmath42 is an object of @xmath43 and @xmath44 an object of @xmath38 with a map @xmath45,[1])$ ] over @xmath19
. then there exists the smallest subcategory @xmath46 of @xmath0 which is an object of @xmath38 , and @xmath47 factors through @xmath48,[1])$ ] .
it is clear that the set of these @xmath46 s @xmath49 modulo equivalence is small .
hence , the solution set condition is satisfied and we have the left adjoint of @xmath41 . 1 .
we denote again by @xmath22 the left adjoint functor @xmath50 obtained in the above proposition . for a graded set @xmath51 , @xmath52 is the _ free algebraic type _ generated by @xmath51 .
the graded set @xmath53,[1 ] ) \to \nnn\ ] ] is the set of _ derived operators _ generated by @xmath51 .
2 . a _ finitary algebraic type _ is a pair @xmath54 , where 1 .
@xmath51 is a graded set , and 2 .
@xmath55 is a sequence of sets indexed by @xmath19 , and @xmath56 is an equivalence relation on @xmath57 .
+ @xmath51 is called the set of _ operators _ of @xmath2 and the elements in @xmath31 are called _
@xmath33-ary operators_. @xmath58 is called the _
equational class _ of @xmath2 .
3 . for a finitary algebraic type @xmath59 ,
let @xmath60 be the full subcategory of @xmath61 , consisting of objects @xmath62 such that @xmath63 for any @xmath64 .
the existence of the initial object in @xmath60 can be shown by applying freyd s adjoint theorem again , and is denoted by @xmath65 .
4 . a morphism @xmath66 of algebraic types is a morphism @xmath67 in @xmath68 .
5 . for any algebraic type @xmath2 and a symmetric monoidal category @xmath69 ,
a _ @xmath2-algebra _ in @xmath0 is a morphism @xmath70 of monoidal categories .
we often just indicate @xmath71 ) \in \scr{c}$ ] for @xmath72 , if there seems to be no ambiguity .
+ a _ morphism of @xmath2-algebras _ in @xmath0 is a @xmath73-preserving natural transformation .
we denote by @xmath74 the category of @xmath2-algebras in @xmath0 .
when @xmath0 is the category @xmath75 of small sets , regarded as a cartesian closed category , then we omit the indication of @xmath0 . 1 . to present a fixed type of algebra as a functor is already realized by lawvere , known as the _
lawvere theory_. moreover , it is true that this definition of algebras is quite complicated , and category theorists would rather treat finitary algebraic category instead ; see definition [ def : finitary : alg ] below .
it is also true that the following construction of scheme theories is available in this generalization @xcite .
however , here we reviewed the classical definition of algebras ( lawvere theory and finitary algebraic categories are generalizations of algebras , but are not equivalent ) , since anyway we need to clarify that these generalizations are indeed a generalization of algebras .
2 . the reader should be aware that , when the monoidal structure @xmath73 of a symmetric monoidal category @xmath0 is not cartesian , the notion of @xmath2-algebras in @xmath0 introduced above may be different from the conventional one .
for example , if @xmath0 is the category of abelian groups with the tensor structure and @xmath2 is the algebraic type of monoids , then @xmath74 does _ not _ coincide with the category of unital associative rings , since a @xmath2-algebra @xmath72 in @xmath0 in the above sense is also required to have first and second projections @xmath76 , the diagonal mapping @xmath77 , and so on .
+ this also shows that some of the algebraic type can not be realized , using the modern fashion of operads@xcite : for example , the algebraic type of groups can not be described as an operad , since the inverse law @xmath78 requires multiple times of references to one variable @xmath79 .
[ def : finitary : alg ] 1 .
let @xmath80 be an adjunction between two categories , and set @xmath81 .
let @xmath82 be the category of @xmath83-algebras , namely objects @xmath84 in @xmath85 with a @xmath83-action @xmath86 and @xmath83-equivariant morphisms .
we have a natural functor @xmath87 , called the _
comparison functor_. the above adjoint is _ monadic _ , if @xmath88 is an equivalence of categories .
a _ finitary algebraic category _ is a category @xmath18 equipped with a adjunction @xmath89 which is monadic .
let @xmath90 be a morphism of algebraic types .
then , @xmath47 induces an underlying functor @xmath91 which has a left adjoint @xmath22 .
in particular , we have an adjoint @xmath92 for any algebraic type @xmath2 , and this is monadic .
consequently , @xmath93 is small complete and small co - complete .
we we refer to the funtor @xmath22 as the _
free generator_. in this section , we give a brief summary of lattice theories .
this also emphasizes the fact that restricting to coherent schemes is somewhat better than considering all schemes . a topological space @xmath84 is _ sober _
, if any irreducible closed subset has a unique generic point .
@xmath84 is _ quasi - separated _
( @xcite , proposition 1.2.7 ) , if the intersection of any two quasi - compact open subset of @xmath84 is again quasi - compact .
@xmath84 is _ coherent _ , if it is sober , quasi - compact , quasi - separated , and admits a quasi - compact open basis .
a poset @xmath94 is a _ distributive lattice _ , if a. any two elements of @xmath95 of @xmath96 admit a _ join _
@xmath97 , namely the supremum of @xmath98 , and a _ meet _
@xmath99 , namely the infimum of @xmath98 , b. @xmath96 has a unique maximal ( resp .
minimal ) element @xmath100 ( resp .
@xmath101 ) , and c. distribution law holds : @xmath102 .
a distributive lattice @xmath96 is _ complete _ , if the supremum is defined for any subset of @xmath96 , and infinite distribution law holds .
in fact , a distributive lattice can be regarded as a semiring : recall that , an element @xmath79 of a monoid @xmath14 is _ absorbing _ , if @xmath103 for any @xmath104 .
a 5-uple @xmath105 is an _ idempotent semiring _ if a. @xmath13 is a set , @xmath106 are two binary operators on @xmath13 , and @xmath107 are two elements of @xmath13 , b. @xmath108 is a commutative idempotent monoid , and @xmath109 is a commutative monoid , c. @xmath101 ( resp .
@xmath100 ) is an absorbing element with respect to the multiplication ( resp .
the addition ) , and d. distribution law holds .
let @xmath13 be a semiring .
then , @xmath13 can be regarded as a poset by setting @xmath110 then @xmath111 gives the supremum of @xmath95 with respect to this order , and @xmath107 become the maximum and the minimum of @xmath13 , respectively .
in fact , to give a distributive lattice is equivalent to giving a semiring with idempotent multiplication .
indeed , if we are given a semiring @xmath13 with idempotent multiplication , then we may replace @xmath112 by @xmath113 and @xmath114 by @xmath115 .
in particular , a distributive lattice can be regarded as an algebra .
let @xmath116 ( resp .
@xmath117 ) be the category of distributive lattices ( resp .
complete distributive lattices ) and their homomorphisms .
these categories are finitary algebraic categories . as a consequence ,
@xmath116 is small complete and small co - complete .
the underlying functor @xmath118 admits a left adjoint @xmath119 , defined as follows : for a distributive lattice @xmath96 , @xmath120 is the set of all non - empty _ lower sets _ of @xmath96 : @xmath121 @xmath122 ( resp .
@xmath123 ) is defined as the lower set generated by @xmath124 ( resp .
@xmath125 ) for @xmath126 , @xmath127 respectively . let @xmath128 be the initial object in @xmath116 .
this is the simplest boolean lattice @xmath129 . 1 . for a topological space @xmath84
, @xmath130 is the set of open subsets of @xmath84 .
this becomes a complete distributive lattice via setting @xmath131 and @xmath132 , and a continuous map @xmath133 between two topological spaces induces a lattice homomorphism @xmath134 .
therefore , we have a contravariant functor @xmath135 .
conversely , for a complete distributive lattice @xmath96 , let @xmath136 be the set of homomorphisms @xmath137 .
this has a natural topology , the open set of which is of the form @xmath138 where @xmath139 .
then @xmath136 becomes a sober space .
if we denote the category of sober spaces by @xmath140 , then we have a contravariant functor @xmath141 .
the functor @xmath51 is not essentially surjective .
the objects in the image category of @xmath51 is called _ spatial _ , and we denote by @xmath142 the full subcategory of @xmath117 consisting of spatial complete distributive lattices .
it turns out that @xmath51 and @xmath143 gives an equivalence of categories @xmath144 .
this also gives the right adjoint of the underlying functor @xmath145 , namely the _
@xmath146 , and the left adjoint of @xmath147 . summarizing
, we have a commutative diagram of adjunctions : @xmath148 \ar[d]_{\omega } & \cat{sob}^{\op } \ar@<.5ex>[l]^{u } \ar@{<->}[d]_{\omega}^{\simeq } \\ \cat{cdlat } \ar@<.5ex>[r ] & \cat{scdlat } \ar@<.5ex>[l]^{u } } \ ] ] let us see for distributive lattices .
the correspondence between the spaces and lattices becomes more clear in this case .
for a distributive lattice @xmath96 , @xmath120 turns out to _ have enough points _
, namely , @xmath120 is spatial .
this is a consequence of a more general statement : let @xmath13 be a semiring , and @xmath149 an ideal of @xmath13 which is not the unit ideal .
then , there exists a maximal ideal containing @xmath149 .
actually , we use this proposition when we prove the existence of a maximal ideal of a given ring . when @xmath13 is a distributive lattice , the ideals of @xmath13 correspond to the lower sets of @xmath13 . 1 . for a distributive lattice @xmath96
, @xmath150 becomes a coherent space , and a homomorphism @xmath151 induces a quasi - compact morphism @xmath152 .
hence , we have a functor @xmath153 .
conversely , for a coherent space @xmath84 , let @xmath154 be the set of quasi - compact open subsets of @xmath84 .
then , @xmath154 becomes a distributive lattice and @xmath155 induces a functor @xmath156 .
this gives the inverse of @xmath157 , and hence an equivalence of categories .
the functor @xmath119 corresponds to the underlying functor @xmath158 : note that this is _ not _ fully faithful , as we only consider quasi - compact morphisms between coherent spaces .
summarizing , we have a commutative square : @xmath159^{u } \ar@{<->}[d]_{\simeq}^{\omega_{c } } & \cat{sob}^{\op } \ar@<.5ex>[l ] \ar@{<->}[d]_{\simeq}^{\omega } \\
\cat{dlat } \ar@<.5ex>[r ] & \cat{scdlat } \ar@<.5ex>[l]^{u } } \ ] ] since @xmath116 is small complete and co - complete , so is @xmath160 .
this justifies our standing point that we stick on to coherent spaces . here
, we will explain some other aspects , which give reasonings of sticking to coherent spaces . 1 . let @xmath161 be a direct product of the copies of a field @xmath162 , with an infinite index set @xmath84
. note that @xmath163 , since @xmath164 is quasi - compact , while @xmath165 is not : @xmath84 is discrete , and @xmath166 , the stone - ech compactification of @xmath84 .
@xmath167 is a hausdorff coherent space , in other words , a totally disconnected compact hausdorff space .
a point @xmath168 on @xmath169 is a prime ideal corresponding to a non - principal ultrafilter on @xmath84 .
the stalk @xmath170 is a non - standard extension field of @xmath162 , if the cardinality of @xmath162 is not less than @xmath171 .
( see for example , @xcite or other textbooks on set theory or model theory . )
+ this is saying that considering coproducts of the underlying space in @xmath172 might not be the best choice , since product of rings and coproduct of schemes does not coincide .
in contrast , when we consider coproducts in @xmath160 , then @xmath173 coincides with @xmath164 .
this implies that we are ignoring some data of the algebra , when we regard the underlying space merely as a topological space , especially when we consider infinite operations .
one of the motivation of extending the notion of schemes is from arithmetics : we want to handle the infinite places of a given number field equally with the finite places .
however , the archimedean - complete field @xmath174 appears only after some transcendental operations .
let @xmath13 be a subring of @xmath175 whose element @xmath176 is uniformly bounded : namely , there is an upper bound @xmath177 such that @xmath178 for any @xmath33
. then , @xmath174 appears as the residue field @xmath179 of the maximal ideal @xmath180 of @xmath13 , corresponding to a non - principal ultrafilter @xmath181 on @xmath19 : @xmath182 this hints us that , these transcendental points may bridge the gap between the archimedean world and the non - archimedean world . before we go on to the next section , we mention a simple , but important fact on sheaves on coherent spaces .
let @xmath84 be a topological space .
since @xmath130 and @xmath154 are posets , they can be regarded as a category : the object set is @xmath130 ( resp .
@xmath154 ) and there is a unique morphism @xmath183 if and only if @xmath184 .
let @xmath84 be a coherent space , and @xmath0 be a small complete category . then , to give a @xmath185-valued sheaf on @xmath84 is equivalent to give a finite continuous functor @xmath186 .
let @xmath187 be a finite continuous functor .
we will define a continuous functor @xmath188 as follows : any @xmath189 is a union of quasi - compact subsets @xmath190 s of @xmath84 .
then , @xmath191 is defined as the equalizer of @xmath192 .
this does not depend on the choice of the refinement @xmath193 .
in this section , we characterize the category of coherent schemes by a universal property .
note that this can not be achieved only by specifying the algebraic type ; we must also fix the grothendieck topology
. the essential idea of the construction of general schemes is already given in @xcite . in the sequel , we fix the following data : @xmath0 is a category which admits pull backs and finite coproducts .
let @xmath12 be a lluf subcategory of @xmath0 ( namely , @xmath194 is faithful and essentially surjective ) such that a. @xmath12 is closed under isomorphisms , and base change : namely , if @xmath183 is a morphism in @xmath12 , then @xmath195 is also in @xmath12 .
b. @xmath12 admits pushouts , and @xmath194 is pushout preserving .
in particular , the initial object of @xmath0 coincides with that of @xmath12 . c. any morphism @xmath196 in @xmath12 is monic and _ flat _
, namely @xmath197 is coproduct preserving .
d. if @xmath198 is a finite set of @xmath12-morphisms , then the diagram @xmath199 has the coequalizer in @xmath12 , where @xmath200 and @xmath201 are canonical morphisms .
[ def : groth : top ] @xmath202 is a family indexed by the objects in @xmath0 , and @xmath203 is a family of finite sets @xmath204 of morphisms in @xmath12 , which satisfies the following condition : 1 .
the canonical fork @xmath205 is exact .
this is what we call _ descent datum_. 2 .
@xmath206 gives a coherent grothendieck topology on @xmath0 , namely : 1 .
@xmath207 is in @xmath208 .
2 . if @xmath209 is an inclusion of finite sets of @xmath12-morphisms over @xmath210 and @xmath211 , then @xmath212 .
3 . if @xmath213 is a morphism in @xmath0 , and @xmath214 , then @xmath215 , where @xmath216 \in s\}.\ ] ] 4 . if @xmath217 and @xmath218 are finite sets of @xmath12-morphisms such that @xmath219 for any @xmath220 \in s$ ] , then @xmath221 . 1 . we will refer to the triple @xmath222 as a _
coherent site_. 2 . let @xmath223 ( @xmath224 ) be two coherent sites .
a morphism of coherent sites @xmath225 is a functor such that @xmath226 and @xmath227 .
we denote by @xmath6 the 2-category of coherent sites . here , we will give some examples of coherent sites . 1 . the category @xmath160 of coherent spaces , or equivalently , the opposite category @xmath228 of distributive lattices : @xmath12 is the subcategory of open immersions , and @xmath229 if and only if @xmath230 covers @xmath84 , in the usual sense .
the opposite category @xmath231 of commutative ring with units : @xmath12 is the subcategory of localizations @xmath232 of finite type , namely the multiplicative system @xmath233 is finitely generated ( hence @xmath234 for some @xmath235 ) @xmath236 if and only if @xmath237 generates the unit ideal .
we will discuss the generalization of this example in section 5 .
the opposite category @xmath238 of commutative monoids with an absorbing element .
@xmath12 is the subcategory of localizations of finite type , similar to the case of rings .
@xmath239 is an element of @xmath240 if and only if one of the @xmath241 s is a unit . the crucial difference between the category of rings and that of schemes is that we can patch " objects along open subobjects . here
, we will axiomatize what patching " means .
let @xmath222 be a coherent site .
@xmath0 is _ schematic _ , if any cocartesian diagram @xmath242 \ar[d ] & b \ar[d ] \\
c \ar[r ] & b \amalg_{a } c } \ ] ] in @xmath12 is bicartesian , namely @xmath79 is the pullback of @xmath243 $ ] .
we will denote by @xmath7 the full subcategory of @xmath6 consisting of schematic coherent sites .
the main theorem of this paper is as follows : [ thm : univ : coh : sch ] the underlying functor @xmath8 admits a left adjoint @xmath9 .
the next section is devoted to the proof .
[ rmk : sch : cat : incomp ] note that until now , we havent assumed completeness of the category @xmath0 .
this enables us to apply @xmath9 to the category of schemes , and we can say that the above adjoint is idempotent . note that the category of ordinary schemes is _ not _ complete , for example , we can not define an infinite product of projective spaces @xmath244 .
the reason of this incompleteness relies on the fact that we assume schemes to be locally isomorphic to the spectrum of a ring : @xmath244 has the product topology , which means that there are no open affine subsets .
this is another motivation of considering weak @xmath0-schemes later on . to begin with
, we must construct the spectrum @xmath245 for each object @xmath13 of a coherent site @xmath0 .
let @xmath0 be a coherent site , and @xmath246 . 1 .
@xmath247 is the family of finite sets @xmath248 of @xmath12-morphisms .
we define a relation @xmath249 on @xmath247 by @xmath250 this relation satisfies the reflexivity and transitivity .
3 . let @xmath251 be the equivalence relation generated by @xmath249 , namely @xmath252 set @xmath253 .
4 . we can define @xmath113 and @xmath115 on @xmath254 by @xmath255 this does not depend on the representation , and @xmath254 becomes a distributive lattice : @xmath101 is an empty set , and @xmath100 is @xmath256 .
let @xmath257 be a morphism in @xmath0 .
then , we have a morphism @xmath258 defined by @xmath259 .
this gives a contravariant functor @xmath260 .
we will denote the functor @xmath261 by @xmath262 . here , we will only give a brief proof of ( 2 ) and ( 4 ) .
in the sequel , we omit the subscript of @xmath263 and simply write @xmath114 . 1 .
we first show the reflexivity : @xmath264 . for any @xmath40 ,
@xmath265 is isomorphic to @xmath266 , since @xmath267 is monic .
hence , @xmath268 . from the condition ( b ) of definition [ def : groth : top ] , we have that @xmath269 . + next
, we show the transitivity : suppose @xmath270 and @xmath271 .
then , @xmath272 for any @xmath40 , and ( c ) shows that @xmath273 for any @xmath274 . combining with the fact that @xmath275 for any @xmath276 , ( d ) tells that @xmath277 .
2 . we will only prove for the join .
suppose @xmath278 and @xmath279 .
then , @xmath280 tells that @xmath281 also , @xmath282 implies @xmath283 combining these two , we obtain @xmath284 .
[ def : weak : sch ] 1 . on a coherent space @xmath84
, we have a canonical @xmath116-valued sheaf @xmath285 defined by @xmath286 .
a morphism @xmath133 of coherent spaces induces a morphism @xmath287 of @xmath116-valued sheaves on @xmath84 , defined by @xmath288 2 . on a @xmath185-valued coherent space @xmath289
, we have a canonical @xmath116-valued sheaf @xmath290 defined by the sheafification of @xmath291 .
weak @xmath0-scheme _ is a triple @xmath292 , where @xmath289 is a @xmath185-coherent space , and @xmath293 is a morphism of @xmath228-valued sheaves on @xmath84 , which satisfies the following condition : + for any inclusion @xmath294 of quasi - compact open subsets of @xmath84 , the restriction functor @xmath295 factors through the @xmath12-morphism @xmath296 whenever @xmath297 .
+ we will refer to @xmath298 as the support morphism " on @xmath84 : we will explain the reason of this name later in [ exam : asch : nagata ] .
+ also , we will denote the underlying coherent space by @xmath299 to avoid confusion .
4 . a morphism @xmath133 of weak @xmath0-schemes is a pair @xmath300 , where @xmath301 is a quasi - compact morphism , and @xmath302 is a morphism of @xmath185-valued sheaves on @xmath303 which makes the following diagram commutative : @xmath304^{f^{\ # } } \ar[d]_{\beta_{y } } & f_{*}\sigma_{x } \ar[d]^{f_{*}\beta_{x } } \\ \tau_{y } \ar[r]_{f^{-1 } } & f_{*}\tau_{x } } , \ ] ] where the top arrow is defined by @xmath305 5 .
we denote by @xmath306 the category of weak @xmath0-schemes and their morphisms .
one advantage of this new definition is that , we can obtain the spectrum functor as the adjoint of the global section functor , as below .
the global section functor @xmath307 admits a right adjoint , namely the spectrum functor @xmath308 .
moreover , @xmath308 is fully faithful . for an object @xmath246
, we define a weak @xmath0-scheme @xmath309 as follows : a. @xmath310 , b. @xmath208 is a @xmath185-valued sheaf on @xmath84 , in other words , a finite continuous functor @xmath311 .
this is defined by @xmath312 this does not depend on the representation of the open set @xmath41 .
for an inclusion @xmath313 of quasi - compact open subsets of @xmath299 , we have a commutative diagram @xmath314 \ar@<-.5ex>[r ] \ar[d ] & \amalg_{k}(v_{k } \times u_{i } ) \ar[r ] \ar[d ] & u_{i } \ar@{.>}[d ] \\
\amalg_{k , l}(v_{k } \times v_{l } ) \ar@<.5ex>[r ] \ar@<-.5ex>[r ] & \amalg_{k } v_{k } \ar[r ] & \scr{o}_{x}(v ) } , \ ] ] which patches up to give @xmath315 .
+ this is the _ structure sheaf of @xmath84_. c. the support morphism @xmath293 is defined canonically : any quasi - compact open subset @xmath41 of @xmath299 is represented by a family of @xmath12-morphisms @xmath316 .
set @xmath317 .
then , for any @xmath12-morphism @xmath318 , the composition @xmath319 is an @xmath12-morphism .
the family of @xmath12-morphisms @xmath320 defines a quasi - compact open subset @xmath321 of @xmath41 , and is independent of the choice of the representation of @xmath41 .
let @xmath322 be a morphism in @xmath0 .
we can define a morphism @xmath323 as follows : we already have the morphism between the underlying spaces @xmath301 .
the morphism between the structure sheaves @xmath302 is given as follows : for a quasi - compact open set @xmath324 of @xmath303 , we have a commutative diagram @xmath325 \ar@<.5ex>[r ] \ar@<-.5ex>[r ] & \amalg_{i}(a \times_{b}u_{i } ) \ar[d ] \ar[r ] & \scr{o}_{x}(f^{-1}u ) \ar@{.>}[d ] \\
\amalg_{i , j } ( u_{i } \times_{b } u_{j } ) \ar@<.5ex>[r ] \ar@<-.5ex>[r ] & \amalg_{i}u_{i}\ar[r ] & \scr{o}_{y}(u ) \\ } \ ] ] which gives a morphism @xmath326 .
we can easily check that @xmath300 is indeed a morphism of weak @xmath0-schemes .
hence , we have a functor @xmath327 .
the unit functor @xmath328 is given as follows : the support morphism @xmath329 gives a lattice homomorphism @xmath330 , which induces the morphism @xmath331 between the underlying spaces .
the morphism between the structure sheaves @xmath332 is given as follows : for each @xmath12-morphism @xmath333 , there is a morphism @xmath334 .
these patch up to give a morphism @xmath335 .
the counit functor is the canonical morphism @xmath336 .
since @xmath337 is a natural isomorphism , @xmath308 is fully faithful . 1 .
a weak @xmath0-scheme which is isomorphic to a spectrum of some object in @xmath0 is called _ affine_. 2 .
a _ @xmath0-scheme _ is a weak @xmath0-scheme which is locally affine .
let @xmath338 be the full subcategory of @xmath306 consisting of @xmath0-schemes .
we have a pullback in @xmath338 constructed as follows : let @xmath339 be a @xmath0-schemes over a @xmath0-scheme @xmath233 .
a. case @xmath340 are all affine , say @xmath341 , @xmath342 , @xmath343 respectively .
then @xmath344 is defined as @xmath345 .
b. case @xmath233 is affine : let @xmath346 and @xmath347 be open affine coverings of @xmath84 and @xmath303 . then , @xmath348 patches up to define @xmath349 . c. case @xmath233 is arbitrary :
let @xmath350 be an open affine covering of @xmath233 , and set @xmath351 , @xmath352 where @xmath353 and @xmath354 are structure maps .
then , @xmath355 patches up to give @xmath349 .
+ these definitions does not depend on the choice of affine open covers of @xmath339 and @xmath233 .
the lluf subcategory @xmath356 of @xmath338 are open immersions , and @xmath357 is the family of finite quasi - compact open covers .
then , @xmath358 becomes a schematic coherent site , and the spectrum functor @xmath359 becomes a morphism of coherent sites .
let @xmath360 be a morphism of coherent sites .
then , we can extend @xmath361 to a functor @xmath362 as follows : a. if @xmath341 is an affine @xmath0-scheme , then @xmath363 .
b. for a general @xmath0-scheme @xmath84 , let @xmath364 be an affine open covering of @xmath84
. then @xmath365 is defined by the patching of @xmath366 .
this definition does not depend on the choice of the affine covering @xmath230 . c. let @xmath133 be a morphism of @xmath0-schemes , and @xmath367 be an affine open covering of @xmath303 .
then , @xmath368 is determined by a morphism @xmath369 .
this induces a morphism @xmath370 , and these patch up to give @xmath371 .
d. in other words , we have a functor @xmath372 , sending @xmath0 to @xmath338 . 1 .
this definition of schemes is seemingly different from the conventional one : we usually use the notion of locally ringed spaces and local homomorphisms , and in fact , the above definition is almost equivalent to this conventional one .
however , the usual definition of schemes bothers us since it uses the limit process to look at the stalks .
note that we can not take arbitrary limits in the categories of usual schemes .
2 . there is another fancier way of defining @xmath0-schemes .
@xmath0 becomes a site by the fixed grothendieck topology , and @xmath0-scheme is a set - valued sheaf on @xmath0 which locally isomorphic to @xmath373 for some @xmath246 .
when we consider groupoid - valued sheaves , then we can immediately define @xmath0-stacks ( @xcite ) .
3 . the notion of weak @xmath0-schemes coincides with @xmath18-schemes introduced in @xcite , when @xmath0 is the opposite category of rings , monoids , etc .
we will summarize the advantage of considering @xmath18-schemes and weak @xmath0-schemes later on .
[ prop : sch : idem ] if @xmath0 is schematic , then the functor @xmath359 is an equivalence of categories .
it suffices to show that any @xmath0-scheme @xmath84 is affine .
let @xmath374 be an affine open covering of @xmath84 .
it suffices to show for @xmath375 , since the rest follows from the induction .
let @xmath376 be an affine open covering of @xmath377 .
again , it suffices to show for @xmath378 .
therefore , the proof is reduced to the following lemma : suppose the following diagram is a bicartesian square in @xmath12 of @xmath0 : @xmath379^{u } \ar[d]_{v } & b \ar[d ] \\
c \ar[r ] & d } \ ] ] then , applying the spectrum functor to this square gives a bicartesian square in @xmath338 .
suppose we are given a commutative diagram @xmath380 \ar[d ] & \spec b \ar[d]^{f } \\
\spec c \ar[r]_{g } & x } \ ] ] in @xmath338 .
we can define the unique morphism @xmath381 by an explicit construction : @xmath382 is given by @xmath383 .
the morphism @xmath384 is given by @xmath385 we have come to the stage of proving theorem [ thm : univ : coh : sch ] .
we only have to show that the functor @xmath9 is the left adjoint of @xmath8 .
the unit @xmath386 is the spectrum functor @xmath387 .
the counit @xmath388 is defined by the inverse of @xmath308 , since @xmath389 is an equivalence for schematic coherent site @xmath390 by proposition [ prop : sch : idem ] .
in the previous sections , we obtained the distributive lattice from open immersions .
however , in the usual definition we approach from the ideal description . in this section
, we will explain how to obtain the grothendieck topology from ideals .
let @xmath391 be an algebraic type and @xmath0 a category of @xmath2-algebras .
@xmath2 has _ self enhancing property _ , if : a. @xmath392 has a natural structure of @xmath2-algebras for any two objects @xmath393 , b. the composition is bilinear , namely the natural morphisms @xmath394 and @xmath395 induced from any homomorphism @xmath396 of @xmath2-algebras are also homomorphisms .
@xmath2 has self enhancing property if any two operators @xmath397 are _ commutative _ , namely if @xmath398 is @xmath399-ary and @xmath400 is @xmath33-ary , then @xmath401 then , for any @xmath393 , the action of the operators on @xmath392 can be defined , by the action on the value of homomorphisms : @xmath402 for any homomorphisms @xmath403 and any @xmath33-ary operator @xmath398 .
let @xmath2 be an algebraic type with self enhancing property , @xmath0 the category of @xmath2-algebras , and @xmath14 a @xmath2-algebra .
then , the functor @xmath404 admits a left adjoint @xmath405 . for any @xmath406 ,
let @xmath233 be the set of quotient @xmath2-algebras of @xmath407 , where @xmath408 is the free generator .
suppose we are given a homomorphism @xmath409 .
then , we have a homomorphism @xmath410 induced by the map @xmath411 , @xmath412 .
let @xmath413 be the image of @xmath414 .
then we have a commutative diagram @xmath415^{\varphi } \ar[d ] & \scr{c}(m , x ) \\
\scr{c}(m , a ) \ar[ru]_{\tilde{\varphi } _ { * } } } \ ] ] since @xmath233 is a small set , we can apply freyd s adjoint functor theorem @xcite to claim the existence of the left adjoint of @xmath416 . as in the case of abelian categories , @xmath417 becomes a closed symmetric monoidal category : the unit @xmath128 with respect to @xmath73 is @xmath418 , and the composition @xmath419 naturally extends to @xmath420 . in the sequel
, we fix an algebraic type @xmath2 with self enhancing property , and @xmath0 the category of @xmath2-algebras . 1 .
an object @xmath246 is a _
monoid object _
. _ commutative monoid object _ ) if there are morphisms @xmath421 and @xmath422 satisfying the monoid ( resp .
commutative monoid ) axioms .
we denote by @xmath423 ( resp .
@xmath424 ) the category of monoid ( resp .
commutative monoid ) objects in @xmath0 .
note that this notion is different from what we have defined in the preliminary ( 1 ) ; therefore , we use different notations . 2 .
let @xmath246 be a monoid object .
an object @xmath425 is an _
@xmath13-module _ if there is an @xmath13-action @xmath426 with the unital law .
when @xmath13 is commutative , the category @xmath427 of @xmath13-modules shares similar properties with @xmath0 , namely it has self enhancing property , so that the symmetric monoidal structure @xmath428 is well defined and is closed .
also , there is a free generator @xmath429 .
an _ ideal _ of a commutative monoid object @xmath13 is an @xmath13-submodule of @xmath13 .
4 . for two
finitely generated ideals @xmath430 of a commutative monoid object @xmath13 , the multiplication @xmath431 is defined by the image of the morphism @xmath432 .
@xmath433 is defined by the image of the morphism @xmath434 .
these two operations make the set @xmath435 of finitely generated ideals of @xmath13 into a semiring .
we have a set - theoretic description of monoid objects : @xmath13 is a monoid object in @xmath0 if @xmath13 has a multiplicative monoid structure which satisfies the distribution laws : @xmath436 for any @xmath33-ary operator @xmath398 .
note that for some @xmath2 , the emptyset may become an ideal of a commutative monoid object : for example when @xmath2 is null .
the followings are algebraic types @xmath2 with self enhancing property : @xmath11 is the algebraic type of commutative monoid object of @xmath2-algebras 1 .
@xmath2 is the null algebraic type , namely @xmath2-algebras are sets . in this case
, @xmath73 is the cartesian product , and @xmath11 is the type of commutative monoids , in the usual sense .
@xmath2 is the type of pointed sets .
then @xmath73 is the smash product , and @xmath11 is the type of commutative monoids with an absorbing element .
@xmath2 is the type of abelian groups .
then @xmath73 is the usual tensor product , and @xmath11 is the type of commutative rings .
@xmath2 is the type of commutative idempotent monoids with absorbing elements .
@xmath11 is the type of semirings , except that we do not assume that @xmath100 is absorbing with respect to @xmath112 .
[ def : congruence ] let @xmath13 be a @xmath2-algebra . 1
. a _ congruence _ @xmath251 of @xmath13 is an equivalence relation induced by a morphism @xmath437 of @xmath2-algebras : @xmath438 2 .
we have a left adjoint @xmath439 of the underlying functor @xmath440 .
the unit morphism @xmath441 induces a congruence @xmath58 on @xmath435 .
the _ radical ideal _ of @xmath13 is an ideal @xmath442 of @xmath13 such that @xmath443 [ lem : equiv : dlat ] let @xmath13 be a semiring , and @xmath444 .
then , @xmath445 in @xmath446 if and only if there is an integer @xmath33 such that @xmath447 and @xmath448 .
let @xmath251 be an equivalence relation on @xmath13 defined by @xmath449 set @xmath450 .
then , the semiring structure of @xmath13 descends to @xmath72 , and @xmath72 becomes a distributive lattice .
conversely , it is obvious that the canonical map @xmath451 factors through @xmath72 , hence @xmath452 .
let @xmath453 be a commutative monoid object , @xmath14 an @xmath13-module and @xmath454 a multiplicative subset of @xmath13 . the _ localization _ @xmath455 is defined by the universal property of making the actions of the elements of @xmath233 invertible .
we can give an explicit construction @xmath455 in the usual way : @xmath455 is the quotient set of @xmath456 divided by the equivalence @xmath457.\ ] ] as usual , @xmath458 also becomes a commutative monoid object , and @xmath459 .
any @xmath13-multilinear operator @xmath398 on @xmath14 can be extended to a @xmath458-multilinear operator on @xmath455 : @xmath460 if @xmath233 is generated by a single element @xmath47 , then @xmath458 is denoted by @xmath461 .
[ lem : ideal : derive : op ] let @xmath13 be a commutative monoid object in @xmath0 , and @xmath462 be a finitely generated ideal of @xmath13 .
then , the element of @xmath463 can be written in a form @xmath464 , where @xmath398 is a derived operator in @xmath0 and @xmath465 .
in particular , for any element @xmath466 , there exists an @xmath13-equivariant map @xmath467 such that @xmath468 . here , @xmath469 is a product of @xmath33-copies of @xmath13 , regarded as an @xmath13-module .
this can be easily proven by using the induction on the length of derived operators and the distribution law .
let @xmath13 be a commutative monoid object in @xmath0 , and suppose @xmath470 be a finitely generated ideal of @xmath13 such that @xmath471 in @xmath472 .
then @xmath473 covers @xmath13 , namely @xmath474 is exact .
let @xmath475 be an element of the equalizer , namely @xmath476 .
this is equivalent to saying that @xmath477 in @xmath478 .
we may assume that @xmath479 by replacing @xmath241 s by sufficiently large powers ; this is a standard method , and can be seen in hartshorne ( @xcite , ii 2.2 ) .
since @xmath480 in @xmath435 for sufficiently large @xmath481 by lemma [ lem : equiv : dlat ] , we have @xmath482 for some @xmath13-equivariant map @xmath467 by lemma [ lem : ideal : derive : op ] . set @xmath483 .
then , @xmath484 this implies that @xmath485 in @xmath486 so that @xmath487 is in the image of @xmath488 .
[ lem : sieve : cond : local ] let @xmath13 be a commutative monoid object in @xmath0 , and @xmath489 elements of @xmath13 .
if @xmath490 in @xmath435 and @xmath491 in @xmath492 for any @xmath276 , then @xmath493 in @xmath435 . for every @xmath276 , we have a @xmath13-equivariant map @xmath494 such that @xmath495 in @xmath486 .
multiplying by some power of @xmath241 if necessary , we obtain a derived operator @xmath496 such that @xmath497 for some @xmath481 and some @xmath498 .
since @xmath490 in @xmath435 , we also have @xmath499 .
hence , there is a derived operator @xmath500 such that @xmath501 for some @xmath502 . composing @xmath503 s and @xmath398 gives an @xmath13-equivariant map @xmath504 such that @xmath505 , hence the result .
[ cor : zar : top ] let @xmath12 be the lluf subcategory of @xmath185 consisting of localizations of finite type .
for each @xmath246 , let @xmath240 be the family of sets @xmath239 of @xmath12-morphisms such that the ideal @xmath237 is unital .
then , @xmath506 satisfies the axioms of coherent sites , and the induced spectrum functor @xmath507 is the desired one : @xmath508 in the sequel , we will refer to this grothendieck topology as the _
zariski topology_. first , we show that @xmath506 is indeed a coherent site .
it is easy to see that @xmath509 is indeed monic and flat in @xmath185 .
the descent condition is satisfied by definition , hence we only need to verify that @xmath206 is a grothendieck topology .
the only non - trivial statement is condition ( d ) in definition [ def : groth : top ] , but this is proved in lemma [ lem : sieve : cond : local ] .
next , we show that the topology coincides with that obtained from ideals .
let @xmath13 be a commutative monoid object of @xmath0 , and @xmath510 a quasi - compact open set of @xmath511 .
we define the corresponding radical ideal @xmath512 to be the ideal generated by @xmath237 .
this does not depend on the representation of @xmath41 , hence @xmath513 is well defined .
conversely , let @xmath514 be a radical ideal .
the quasi - compact open set @xmath515 is defined by @xmath516 .
this again does not depend on the choice of the generators of @xmath463 , hence we have a bijection between @xmath254 and @xmath472 . also , it is straightforward to see that this is a lattice isomorphism , and that it is functorial with respect to @xmath13 . for some algebraic category @xmath0 with its zariski topology , @xmath0 is already schematic , namely it is vacuous to consider @xmath0-schemes .
for the followings , @xmath0 is the opposite category of @xmath11-algebras ( with the zariski topology ; see corollary [ cor : zar : top ] ) .
the category @xmath0 becomes schematic , if any localization is surjective .
the following cases are such : 1 .
@xmath11 is the type of distributive lattices , 2 .
@xmath11 is the type of idempotent semirings , 3 .
@xmath11 is the algebraic type of von neumann regular commutative rings : a commutative ring @xmath13 is _ von neumann regular _ , if any element @xmath517 has a _ weak inverse _
@xmath518 , namely the unique element @xmath518 which satisfies @xmath519 and @xmath520 .
the krull dimension of a von neumann regular ring is @xmath101 , hence any localization becomes surjective .
this is related to serre s cohomological criterion of affineness : a scheme @xmath84 is affine if and only if @xmath521 for any quasi - coherent sheaf @xmath522 and @xmath523 .
+ hence , any von neumann regular scheme is affine , since its krull dimension is @xmath101 . in this section ,
we discuss local objects of a complete coherent site ( a coherent site @xmath5 is _ complete _ if @xmath0 is complete ) , and show that the notion of @xmath0-scheme coincides with 1 . the notion of coherent schemes , when @xmath0 is the opposite category of rings , and 2 . the notion of monoid schemes , in the sense of ton - vaqui , or deitmar , when @xmath0 is the opposite category of commutative monoids .
let @xmath0 be a complete coherent site . 1 .
an object @xmath524 is _ local _ , if any covering of @xmath72 is principal , namely : @xmath525 for some @xmath276 if @xmath526 is a covering of @xmath72 .
this is equivalent to saying that @xmath527 has a unique closed point .
2 . a morphism @xmath396 between two local objects
@xmath528 is _ local _ , if the induced morphism @xmath529 sends the unique closed point of @xmath530 to that of @xmath531 . in the language of distributive lattices ,
this is equivalent to @xmath532 for any quasi - compact open @xmath533 .
a @xmath0-coherent space @xmath534 is _ local _ , if the stalk @xmath535 is local for any @xmath536 .
a morphism @xmath133 of @xmath185-spaces is _ local _ , if the induced morphism @xmath537 on the stalks is local .
we denote by @xmath538 the category of locally @xmath185-spaces and local morphisms .
the following proposition shows that the category of @xmath0-schemes coincides with the notion of coherent schemes , when @xmath185 is the opposite category of commutative rings : [ thm : sch : local : space ] we have a natural fully faithful functor @xmath539 .
let @xmath292 be a weak @xmath0-scheme .
first , we will show that @xmath84 is a local @xmath185-space .
fix a point @xmath536 , and let @xmath540 be a quasi - compact open cover of @xmath303 .
we have a natural isomorphism @xmath541 since @xmath308 is a right adjoint , and @xmath298 induces a morphism @xmath542 the left - hand side is a local object in @xmath116 .
since @xmath193 covers @xmath303 , one of @xmath543 must coincide with @xmath544 . since @xmath545 is obtained by localizations of open neighborhoods of @xmath168 , @xmath546 for some quasi - compact open neighborhood @xmath2 of @xmath168 .
this shows that @xmath547 in @xmath548 , and hence also in @xmath549 .
therefore , @xmath84 is a local @xmath185-space .
suppose we have a morphism @xmath133 of weak @xmath0-schemes .
we need to show that the induced morphism @xmath550 is local , for every @xmath536 and @xmath551 .
we have a commutative diagram @xmath552^{f^{-1 } } \ar[d]_{\beta_{y } } & \omega_{1}(\scr{o}_{x , x } ) \ar[d]^{\beta_{x } } \\ \omega_{c}(y_{y } ) \ar[r ] & \omega_{c}(x_{x } ) } \ ] ] of distributive lattices .
suppose @xmath553 maps to @xmath554 .
then , @xmath555 : this implies that @xmath556 , and hence @xmath557 .
this shows that @xmath558 , which means that @xmath550 is indeed a local morphism . as a consequence
, we have a functor @xmath559 .
it remains to prove that @xmath560 is fully faithful .
let @xmath339 be two weak @xmath0-schemes , and @xmath133 a morphism of locally @xmath185-spaces .
for any inclusion of two quasi - compact open subsets @xmath294 of @xmath303 , it suffices to show that @xmath561 .
let @xmath168 be any point in @xmath562 , and set @xmath551 .
we have a sequence of local morphisms @xmath563 which shows that @xmath564 in @xmath565 .
we have another sequence of local morphisms @xmath566 induced by @xmath567 , and this sends @xmath100 to @xmath100 .
this is equivalent to saying that @xmath568 .
therefore , @xmath569 .
the converse can be proved similarly .
hence , we see that the notion of @xmath0-scheme coincides with that of conventional schemes , when we set @xmath0 and grothendieck topology properly . 1 .
let @xmath2 be the algebraic type of abelian groups .
then , the commutative monoid objects in the category of @xmath2-algebras are exactly commutative rings . for a ring @xmath13 , @xmath435 is the set of _ finitely generated _ ideals of @xmath13 .
note that @xmath570 is the set of radical ideals , and @xmath571 is the usual zariski spectrum of @xmath13 .
+ let @xmath572 be the opposite category of commutative unital rings , and @xmath12 the lluf subcategory of localizations of finite type .
then we see that @xmath0-schemes are exactly coherent schemes .
2 . when @xmath2 is the null algebraic type , the the commutative monoid objects in the category of @xmath2-algebras are exactly commutative monoids .
then , we obtain the coherent monoid schemes , in the sense of deitmar , of equivalently , that of ton - vaqui @xcite .
the category of coherent schemes include all noetherian schemes and affine schemes , which are practically all the schemes we treat . moreover , the morphisms between noetherian ( or affine ) schemes are all quasi - compact .
this tells that the category of coherent schemes is sufficiently large to work on .
also , remark [ rmk : sch : cat : incomp ] implies that we had better work within this category . in this section ,
we compare weak @xmath0-schemes with @xmath18-schemes introduced in @xcite .
we recall the definition of @xmath18-schemes , in the most general type .
a quadruple @xmath573 is a _
schematizable algebraic type _ , if a. @xmath11 is an algebraic type with commutative multiplicative monoid structure .
we denote by @xmath574 the category of @xmath11-algebras and their homomorphisms .
b. @xmath575 is a functor @xmath576 . c. @xmath577 is a natural transformation @xmath578 such that @xmath579 is multiplication - preserving .
d. for each @xmath11-algebra @xmath13 and a multiplicative system @xmath233 , @xmath580 is a natural isomorphism @xmath581 .
when @xmath522 is a @xmath582-valued sheaf on a coherent space @xmath84 , then the functor @xmath575 induces a @xmath116-valued sheaf @xmath583 on @xmath84 , defined by the sheafification of @xmath584 .
let @xmath585 be a schematizable algebraic type .
@xmath18-scheme _ is a triple @xmath586 such that 1 .
@xmath84 is a coherent space , @xmath208 is a @xmath574-valued sheaf on @xmath84 , @xmath587 is a morphism of @xmath116-valued sheaves on @xmath84 which we refer to as the support morphism " , and 2 .
the restriction maps reflect localizations : let @xmath294 be an inclusion of quasi - compact open subsets of @xmath84 , and we denote by @xmath588 the localization of @xmath589 by the multiplicative system @xmath590 a _ morphism _ @xmath133 is a pair @xmath300 , where a. @xmath301 is a morphism of coherent spaces , b. @xmath591 is a morphism of @xmath574-valued sheaves on @xmath303 , such that the following diagram commutes : @xmath592^{\alpha_{1}f^{\ # } } \ar[d]_{\beta_{y } } & f_{*}\alpha_{1}\scr{o}_{x } \ar[d]^{f_{*}\beta_{x } } \\ \tau_{y } \ar[r]_{f^{-1 } } & f_{*}\tau_{x } } \ ] ] here , we list up the properties of @xmath18-schemes : let @xmath593 be the category of @xmath18-schemes and their morphisms .
then : 1 .
@xmath593 is a full subcategory of locally @xmath574-spaces .
the category of @xmath18-schemes is small complete and small co - complete .
3 . we can consider the image @xmath18-scheme for any morphism of @xmath18-schemes .
practically , we only consider the following case : [ exam : asch : alg : type ] @xmath2 is a self enhancing algebraic type , and @xmath11 is the algebraic type of commutative monoid objects in @xmath2-algebras . for each @xmath11-algebra @xmath13
, @xmath594 is the set of finitely generated ideals of @xmath13 , divided by the congruence defined by @xmath595 ( cf .
lemma [ lem : equiv : dlat ] ) .
@xmath596 sends @xmath597 to the principal ideal generated by @xmath79 .
this preserves multiplication . for each multiplicative system @xmath233 of @xmath13
, we have a natural isomorphism @xmath598 .
the set @xmath594 can be regarded as the distributive lattive corresponding to @xmath164 , and we identify a finitely generated ideal @xmath599 with the quasi - compact open subset which is the complement of the support of @xmath463 .
note that @xmath600 corresponds to @xmath601 appeared in the definition of weak @xmath0-schemes ( cf .
definition [ def : weak : sch ] ) . then , it is straightforward to see that let @xmath585 be as above . then , the category of @xmath18-schemes is equivalent to weak @xmath0-schemes defined in section 5 .
[ exam : asch : nagata ] 1 .
suppose @xmath18 is the schematizable algebraic type induced from rings : namely , @xmath2 is the algebraic type of abelian groups in example [ exam : asch : alg : type ] .
then for each ring @xmath13 , @xmath594 is the idempotent semiring of finitely generated ideals of @xmath13 modulo the congruence generated by @xmath602 . for each element @xmath599 ,
@xmath298 simply gives the complement of the support of @xmath463 which is a quasi - compact open subset of @xmath164 .
this is why we call @xmath298 ` the support morphism ' .
+ actually , the data @xmath298 implies that @xmath289 is a locally ringed space by theorem [ thm : sch : local : space ] .
2 . let @xmath18 be as in ( 1 ) .
let @xmath41 be a non - empty open subset of a coherent scheme @xmath84 .
the zariski - riemann space @xmath603 is the limit of @xmath41-admissible blowups , hence an object of @xmath18-schemes , but in general not a scheme .
however , this object happens to be useful to prove pure algebro - geometric theorems , such as nagata embeddings ( @xcite , @xcite ) .
3 . there is a notion of zariski - riemann spaces for monoids .
this is used to prove the existence of equivariant compactification of toric varieties in @xcite .
let @xmath2 be a self enhancing algebraic type , @xmath11 the type of commutative monoid objects in @xmath2-algebras , and @xmath13 a @xmath11-algebra .
so far , we have introduced the zariski topology on the category of @xmath11-algebras , which is defined by the ideals of @xmath13 , without the reason why .
in fact , there is no natural reason for this choice , if we only aim to construct a natural @xmath582-valued space .
in addition , we have at least two natural topologies , different from the zariski topology .
let @xmath2 be a self enhancing algebraic type , and @xmath13 be a monoid object in @xmath93 .
then , we have a natural isomorphism @xmath604 of monoid objects in @xmath93 , where @xmath605 is the center of @xmath13 , and @xmath606 is the category of right @xmath13-modules . in particular , @xmath13 can be recovered from the category of @xmath13-modules when @xmath13 is commutative .
let @xmath607 be a natural endomorphism of @xmath608 .
then @xmath609 is given by a left multiplication by an element @xmath610 , since @xmath13 is a free @xmath13-module generated by one element .
since @xmath361 is natural , @xmath611 must commute with any left multiplication @xmath612 by an element @xmath597 .
this shows that @xmath613 is in the center of @xmath13 .
the universal property of coproducts tells that @xmath614 is also a left multiplication by @xmath613 , for any free @xmath13-module @xmath615 .
let @xmath14 be an arbitrary @xmath13-module .
then we have a surjective homomorphism @xmath616 from a free @xmath615-module . the commutative diagram @xmath617 \ar[d]_{z \times ( - ) } & m \ar[d]^{f_{m } } \\
p \ar[r ] & m } \ ] ] tells that @xmath618 is also a left multiplication by @xmath613 .
suppose there is a surjective homomorphism @xmath619 of @xmath13-modules when @xmath13 is a ring
. then this map is essentially determined by the kernel , which is an ideal of @xmath13 .
in contrast , the surjective map can not be parametrized by the ideal of @xmath13 , if we consider algebraic types other than that of rings ; we need congruences ( see definition [ def : congruence ] ) .
the typical example if when @xmath2 is the null algebraic type , namely when an @xmath13-module is a set with an action of a monoid @xmath13 .
then , the set of congruences of @xmath13 may not have a maximal element , since the set of all congruences is not finitely generated . 1 .
let @xmath623 be the quotient @xmath13-module .
then for any @xmath104 , @xmath624 2 .
this follows from the fact that the set @xmath625 of non - unital congruences of @xmath13 including @xmath463 is an inductively ordered set , since the unit congruence is principally generated by @xmath626 . in the sequel
, @xmath0 is the opposite category of @xmath11-algebras , @xmath12 is the subcategory of @xmath0 consisting of localizations of finite type ; this restriction on @xmath12 is reasonable , since 1 .
it is the simplest and the most important case that occurs , and 2 . in general , to describe flat epimorphisms ( the condition which we require for @xmath12 ) of @xmath11-algebras is not at all trivial , even for the classical case of rings ; see @xcite , @xcite . 1
let @xmath13 be a @xmath11-algebra and @xmath628 .
let @xmath14 be an @xmath13-module .
the @xmath208-module @xmath629 is defined by the sheafification of @xmath630 , where @xmath257 is a morphism in @xmath12 .
2 . let @xmath84 be a @xmath0-scheme .
an @xmath208-module is _ quasi - coherent _ , if it is isomorphic to @xmath629 on each affine open subset @xmath631 of @xmath84 , where @xmath14 is an @xmath13-module . 1 .
for any @xmath11-algebra @xmath13 , let @xmath635 be the set of all finite sets @xmath636 \to r\}_{i}$ ] of localizations which satisfies the descent datum .
the family @xmath637 gives a grothendieck topology .
2 . for any @xmath11-algebra @xmath13 ,
let @xmath638 be the set of all finite sets @xmath636 \to r\}_{i}$ ] of localizations such that either one of the @xmath241 s is a unit .
the family @xmath639 gives a grothendieck topology .
the topology @xmath640 ( resp .
@xmath641 ) gives the coarsest ( resp .
finest ) topology : when we denote by @xmath642 ( resp .
@xmath643 ) the corresponding spectrum functors , then we have immersions @xmath644 these immersions are not isomorphisms in general . for example , when @xmath13 is a noetherian commutative ring , then the points of @xmath645 correspond to the points of @xmath164 of height less than @xmath646 via krull s hauptidealsatz ( @xcite , theorem 13.5 ) .
first , we will show that @xmath648 implies @xmath17 .
let @xmath104 be a non - zero element .
we have an isomorphism @xmath649 of @xmath13-modules , where @xmath463 is a congruence on @xmath13 as an @xmath13-module .
since @xmath650 , there is a maximal congruence @xmath180 containing @xmath463 by proposition [ prop : congruence : max ] .
let @xmath651 be the subset of @xmath13 consisting of elements @xmath79 such that @xmath652 .
we see that @xmath651 is a prime ideal , and the stalk @xmath653 is non - zero .
next , we will show that the zariski topology is indeed the coarsest topology satisfying ( * ) .
let @xmath206 be a topology satisfying ( * ) .
\to r\}_{i } \in \mathcal{o}_{r}$ ] is not in @xmath654 , namely @xmath237 does not generate the unit ideal in @xmath13 .
let @xmath655 be the congruence on @xmath13 generated by @xmath656 .
then @xmath657 is not a zero @xmath13-module .
however , @xmath658 , where @xmath659 $ ] is an open subset of @xmath628 for each @xmath276 . by definition , @xmath190 s
cover @xmath84 .
therefore , @xmath647 , a contradiction .
comfort , w. w. , negerpontis , s. : _ the theory of ultrafilters _ , grunlehren der mathematischen wissenschaften ( 1974 ) springer - verlag maclane , s. : _ categories for the working mathematician .
2nd edtion _ , graduate texts in math . * 5 * ( 1998 ) springer - verlag | in this paper , we generalize the construction method of schemes to other algebraic categories , and show that the category of coherent schemes can be characterized by a universal property , if we fix the class of grothendieck topology .
also , we introduce the notion of @xmath0-schemes , which is a further generalization of coherent schemes and still shares common properties with ordinary schemes . |
the study of entanglement measures in the eigenstates of condensed matter hamiltonians is a rapidly growing field of interdisciplinary research @xcite .
while a great many results have been reported , the typical situation is that entanglement studies of condensed matter systems provide alternative signatures of phenomena already known from more traditional condensed - matter techniques . in this article , we use measures from quantum information to characterize a spin system which is poorly understood in the condensed matter literature , namely , spin rings with ferromagnetic nearest - neighbor interactions frustrated by antiferromagnetic next - nearest - neighbor interactions .
we show how the unusual properties of this system are reflected in quantum information quantities .
the systems we present are closed rings of @xmath0 spins .
the study of finite - size quantum systems have enjoyed a huge resurgence in the past decade because of cold - atom and nanoscience developments , which provide the context for designing and studying finite - size hamiltonians for their own fundamental properties rather than as mere tools for investigating the thermodynamic limit @xmath1 .
we will focus on a sequence of ground - state level crossings in such systems , which reflect lattice - incommensurate correlations .
we show that certain measures of entanglement display remarkably strong parameter dependences near these level crossings .
our calculations of the _ concurrence _ show that the entanglement structure of the states between successive level crossings is ideal for tuning the distance between spins which are mutually entangled .
we also present some results on entanglement _ dynamics _ , again focusing on the effect of the level crossings .
it is natural to think of ground state level crossings as finite - size analogs of first - order phase transitions in the thermodynamic limit . as a corollary ,
one expects that ground state properties change suddenly at the crossing ( ` transition ' ) point and that away from the exact crossing point the system possesses no knowledge of the proximity of the crossing . in this
regard , our quantum information quantities reveal a peculiarity of the crossings in our spin rings , because they seem to evolve in response to the proximity of the level crossings , already at some distance away from the parameter value where the crossing occurs , _
i.e. _ , the ground state is `` aware '' of the nearby crossing .
this counter - intuitive phenomenon is clearly seen in multi - spin ( but bipartite ) entanglements as well as fidelities , but not in traditional condensed - matter quantities such as correlation functions .
note that , since the level crossings are not expected in the thermodynamic limit , this issue is meaningful only in our mesoscopic context . to calculate a bipartite entanglement
, one has to first decide on a partition . for a condensed - matter model or even its finite - size versions
, there are a large number of ways in which the system can be bi - partitioned .
a variety of different partitionings , and correspondingly different entanglements , have been used for entanglement studies in condensed matter models .
many studies have focused on the entanglement between two sites as measured by the concurrence or negativity ( _ e.g. _ , @xcite ) , or the entanglement between one spin / site with the rest ( _ e.g. _ , @xcite ) .
such measures generically are not expected to provide information different from traditional condensed matter quantities like 2-point correlation functions , since the entanglement measure can be related easily to correlation functions ( _ e.g. _ , @xcite ) . in this sense , many - site measures are more innovative .
examples are block entanglement in 1d @xcite and in 2d @xcite , and more recently studied spatially disconnected partitions such as sublattice entanglement @xcite and entanglement between labeled itinerant particles @xcite .
such entanglement measures are not easily related to usual correlation functions , and therefore hold the possibility of providing genuinely new tools for probing collective phenomena in many - particle systems . for our finite - sized chains , we will present entanglement results for various partitions , both two - site and many - site .
the dynamics of entanglement measures in many - particle systems has attracted some attention recently @xcite , but is still a relatively poorly understood topic , reflecting the general difficulty of non - equilibrium physics in quantum condensed matter .
entanglement dynamics has also gained interest from the quantum information perspective , _
e.g. _ , in the context of adiabatic quantum computing @xcite and in the context of quantum state transfer ( ` quantum communication ' ) through spin chains @xcite .
obviously , the number of different conceivable non - equilibrium situations is large ; we are forced to restrict the types of dynamics considered . in this article
we will describe the entanglement dynamics after a sharp change ( _ quench _ ) of interaction parameter . after describing in sec .
[ sec_partitns - n - entanglements ] the various entanglement measures and partitions to be used , in sec .
[ sec_model ] we introduce the hamiltonian and its condensed - matter context .
[ sec_model ] also describes the level - crossings and compares to other examples of similar crossings known to us .
the next sections present entanglement results , starting in sec .
[ sec_corrfn - concurr ] with the quantities containing the most traditional information , namely , the concurrences which are readily related to correlation functions .
sections [ sec_multispin_derivatives ] and [ sec_fidelities ] present quantum - information quantities which are more distinct from condensed - matter measures . sec .
[ sec_dynamics ] describes our results on entanglement dynamics .
( a - c ) various partitions of closed chain for calculating bipartite entanglement entropy . spins belonging to partition @xmath2 ( @xmath3 ) are represented by open ( full ) circles .
( d ) geometry for exploiting the sublattice entanglement @xmath4 , _
e.g. _ , in an experimental realization . in this setup @xmath4 is simply the entanglement between inner and outer rings .
dashed and full lines represent @xmath5 and @xmath6 interactions respectively . ] we will mostly measure entanglement using the _ entanglement entropy_. the entanglement entropy between partitions @xmath2 and @xmath3 is defined using the reduced density matrix of one part ( e.g. , @xmath7 obtained by tracing over @xmath3 degrees of freedom ) to calculate the von neumann entropy , @xmath8 $ ] . since the purpose of using entanglement in many - particle systems is to provide viewpoints not present in usual condensed - matter measures , we want to consider more ` non - local ' entanglements .
the partitions we choose for this purpose are shown in figure [ fig_partition - cartoons ] .
the most obvious choice is for @xmath2 and @xmath3 to be spatially connected blocks , in particular for each partition to be a connected half of the ring ( fig.[fig_partition - cartoons]a ) .
this partitioning is suggested by many existing results on block entanglement entropies in one dimension @xcite .
another choice is to take all the odd sites as partition @xmath2 and the even sites as partition @xmath3 this is then the entanglement between @xmath2 and @xmath3 _ sublattices _ , in usual quantum magnetism language ( fig.[fig_partition - cartoons]b ) .
the entanglement between sublattices has been studied for a number of spin chains and lattices , and extrema of @xmath9 have been claimed to characterize some quantum phase transitions @xcite .
note also that since it is natural to think of the @xmath5-@xmath6 chain as a ` zigzag ladder ' ( _ e.g. _ , @xcite ) , the sublattice entanglement is simply the entanglement between the two legs of such a ladder . in the mesoscopic ( finite - size ) context that we are interested in , a @xmath5-@xmath6 ring is likely to be implemented as shown in fig .
[ fig_partition - cartoons]d ; the two ` legs ' are now two @xmath6 rings ( either inner or outer rings , or an upper and a lower layer ) , connected by zigzag @xmath5 interactions . from an experimental quantum information processing perspective , the sublattice entanglement is natural as it is simply the entanglement between inner and outer ( or upper and lower ) rings .
since the antiferromagnetic interaction in our case is the _ next_-nearest - neighbor interaction , we extend the idea of the sublattice partition ( @xmath10 ) and also use the @xmath11 partition shown in fig.[fig_partition - cartoons]c .
we call the entanglement entropies corresponding to the above three partitions respectively @xmath12 , @xmath4 , @xmath13 .
another nonlocal quantity motivated by quantum information theory is the fidelity , the overlap between ground state wavefunctions at slightly different system parameters @xcite .
we describe fidelities in our spin rings in section [ sec_fidelities ] .
we also present results for more ` local ' measures of entanglement , namely , ( 1 ) entanglement entropy for 2-site partitions , and ( 2 ) the concurrence @xmath14 measuring the entanglement between sites @xmath15 and @xmath16 in the environment provided by the rest of the chain @xcite . the concurrence is defined as @xmath17 \ ; , \ ] ] where @xmath18 are the eigenvalues in decreasing order of the matrix @xmath19 , and @xmath20 is the reduced density matrix of the two - site subsystem . since we will consider singlet states , two - site reduced density matrices are strongly constrained , and the concurrence is a simple function of the spin - spin correlation functions .
two spins are entangled ( @xmath21 ) if the correlations between them are sufficiently antiferromagnetic , @xmath22 , and in that case @xmath23 ( derived in , _
e.g. _ , @xcite ; see also ref . 22 in @xcite ) .
the behavior of concurrence in various spin models is briefly reviewed from the literature in sec .
[ sec_corrfn - concurr ] in order to provide context for our concurrence results .
we are concerned with closed chains of @xmath24 localized spin-1/2 objects , _
i.e. _ , a finite - size spin chain , interacting via the hamiltonian @xmath25 where @xmath26 @xmath27 , and @xmath15 is the site index obeying periodic boundary conditions appropriate to a ring .
we define @xmath28 and @xmath29 .
this spin chain is known for any even @xmath24 to have a singlet ground state for @xmath30 and a @xmath31-fold degenerate ferromagnetic ground state manifold for @xmath32 @xcite . at @xmath33 a singlet ground state
is degenerate with the ferromagnetic manifold ; these degenerate states are known exactly @xcite . we will confine ourselves to the singlet region , @xmath34 .
the numerical results are presented for ring sizes divisible by 4 , in order to avoid complications due to the parity of @xmath35 .
the rings for other even sizes have very similar behavior .
although far less studied compared to the @xmath36 case ( majumdar - ghosh chain ) , the frustrated ferromagnet hamiltonian has already been known for some time to possess some peculiar properties @xcite .
more recently , this model has attracted a fresh surge of theoretical attention @xcite , partly due to the appearance of materials whose spin physics is described by this model ( see references in @xcite ) .
most studies concentrate on finite magnetic fields or finite temperatures and on properties in the thermodynamic limit @xcite , or very near the transition point at @xmath33 @xcite .
_ proliferation of ground state level crossings _ an unusual feature of the singlet region @xmath37 is the number of ground state level crossings , which proliferate as we increase the number of spins @xcite . for @xmath38 8 , 12 , 16 , 20 , ... spins , there are respectively one , two , three , four , ... level crossings in this region .
a very likely interpretation is that the natural correlations in the chain for @xmath34 are incommensurate with wavevector dependent on @xmath39 .
this incommensurability is resolved in different ways by the different ground states . for larger rings ,
there are a greater number of possibilities to accommodate the incommensurate correlations into the finite system , and hence the larger number of level crossings .
the ground state lattice momentum alternates between @xmath40 and @xmath41 in the regions between successive level crossings .
very similar level crossings appear in other spin models known or thought to have spiral correlations .
the examples known to us are ( 1 ) the majumdar - ghosh chain for @xmath42 @xcite ; ( 2 ) the ` sawtooth ' chain of ref .
@xcite ; ( 3 ) the frustrated ferrimagnetic chain of ref .
this phenomenon of proliferating level crossings in finite - size rings as a result of incommensurate correlations , however , is not widely known or discussed in the condensed matter literature .
correlation functions and related quantities for @xmath43 .
( a ) correlation function at @xmath44 . ( b ) the structure factor , @xmath45 , at @xmath44 . ( c )
position of the maximum of @xmath45 , as a function of @xmath46 .
( d ) concurrences between two spins at distance @xmath47 2 , 3 , 8 .
( e ) entropy of entanglement between partition consisting of sites ( @xmath15 , @xmath16 ) , and the remaining @xmath48 sites . shown also for @xmath47 2 , 3 , 8 . in ( c - e ) , nonlinear horizontal scale highlights the region near @xmath49 . ] in this section we present more local quantities , in particular traditional condensed matter quantities like spin - spin correlation functions and structure factors , and closely related entanglements like concurrences . for singlet states ,
the correlation function @xmath50 is three times @xmath51 ; we will discuss the latter as a function of @xmath52 . between @xmath53 and the first crossing , the signs of this quantity has the structure @xmath54 for @xmath55 , indicating that the state in this region has a ` spin - density - wave'-like @xmath56 structure .
an example is shown in fig .
[ fig_strcturefactor - n - qmax]a for @xmath43 .
the oscillating magnitude / sign pattern also suggests ` nematic'-like correlations @xcite ; detailed distinctions are inappropriate since this is not a macroscopic phase .
the sign pattern of @xmath57 changes at each level crossing ; for @xmath43 the patterns in the four regions ( separated by level crossings ) are @xmath58 the structure factor @xmath45 , which is the fourier transform of the real - space correlation function , has peaks at @xmath59 , with the integer @xmath60 decreasing in unit steps at each level crossing , from @xmath61 at @xmath53 to @xmath62 near the ferromagnetic transition point @xcite . similar behavior ( fig.[fig_strcturefactor - n - qmax]c ) is also seen in the other examples known to us where finite - size spin systems have level crossings due to spiral correlations @xcite .
we now consider the concurrences @xmath14 between spins @xmath15 and @xmath16 on the rings . for most simple spin models ,
the concurrence does not extend significantly beyond the next - nearest - neighbor site , _ i.e. _ , @xmath63 for many spin models , such as the @xmath64 model in transverse field @xcite , the heisenberg model ground states on 1d chains and various 2d lattices @xcite , and even in the majumdar - ghosh ( @xmath36 , @xmath27 ) chain @xcite . in the xyz model in a magnetic field , the concurrence extends to several lattice sites , but still has finite range @xcite .
long - range concurrences are rare ; the known examples correspond to specific phase transitions @xcite or to carefully constructed models and geometries @xcite . in this context , it is noteworthy that our highly frustrated spin rings display nonzero concurrences at all length scales , depending on the interaction parameter @xmath46 ( fig .
[ fig_strcturefactor - n - qmax]d ) . for any particular value of @xmath46 ,
only a couple of concurrences are nonzero , but not necessarily the ones with small @xmath65 .
the @xmath65 values at which @xmath14 are nonzero , increases as @xmath46 changes from 0 to 4 , from @xmath66 to @xmath67 . in the region near the ferromagnetic transition @xmath33
, there is nonzero concurrence between spins at opposite ends of the ring ( @xmath68 ) , reflecting the unusual long - range quantum correlations of the so - called udrvb state @xcite near @xmath33 .
the behavior we have found indicates that the interaction here acts as a knob to tune the distance between spins with nonzero entanglement .
this tunable distance is quite different from the ` entanglement range ' of ref.@xcite ; our concurrences are not decaying with distance but are nonzero only at a certain distance .
[ fig_strcturefactor - n - qmax]e shows the entanglement between a two - spin partition ( consisting of site @xmath15 and site @xmath16 ) and the rest of the ring . loosely speaking
, this entanglement entropy is large ( small ) when the concurrence _ between _ the two spins is small ( large ) .
this reflects the intuitive idea that objects highly entangled with each other are generally weakly entangled with their environment .
16-spin and 20-spin rings . top : half - block and sublattice entanglements , @xmath12 ( upper dashed ) , @xmath4 ( solid ) , @xmath13 ( lower dashed , magnified ) , against @xmath46 in the singlet region . for the @xmath69 curves , smaller values of @xmath46 are omitted , to focus on the level crossings .
middle and bottom : derivatives of @xmath70 , @xmath71 , respectively . ]
figure [ fig_entnglmnts - a ] displays entanglements @xmath12 , @xmath4 , @xmath13 for the partitions explained in figure [ fig_partition - cartoons ] , and also the derivatives with respect to @xmath46 for the first two cases .
the block entanglement entropy @xmath12 is large compared to , say , the singlet ground state of the nearest - neighbor heisenberg ground state , in which case @xmath721.28(1.354 ) for @xmath38 16(20 ) .
this is not surprising , as the longer - range interaction causes larger entanglement .
the majumdar - ghosh chain ( positive @xmath5 and @xmath6 ) similarly has relatively large block entanglement @xcite .
the sublattice entanglement @xmath4 starts from zero at @xmath53 , because the two sublattices are decoupled in the @xmath73 limit .
after the first level crossing , @xmath4 also reaches larger values . note that @xmath4 is quite large in the heisenberg and majumdar - ghosh chains ; in comparison the ground - state singlets of our system have somewhat lower sublattice entanglement . the derivatives of @xmath70 and @xmath74 ( fig .
[ fig_entnglmnts - a ] lower panels ) are similar and have some striking features .
first , at each level crossing except the leftmost , the derivatives have shapes reminiscent of a resonance , although there is no real divergence .
in other words , the entanglement curves @xmath75 become steep but not completely vertical near each crossing .
second , the @xmath76 curves appear to respond ( curve upward or downward ) already some distance away from the level crossing . while phase transition language is not completely appropriate for our mesoscopic arrangements , the crossings in this sense are loosely speaking `` second - order '' like rather than `` first - order '' like . finally , the @xmath76 curves at the leftmost crossing lacks the resonance - like feature , indicating that the first level crossing is different in nature . the @xmath13 entanglement entropy is peculiarly small ( shown 4-fold magnified in fig .
[ fig_entnglmnts - a ] ) , although the relevant reduced density matrix has the same dimensions as the other two partitions .
this peculiarity remains unexplained at present .
fidelity @xmath77 with @xmath78 , and its derivative .
since @xmath79 is generally very close to unity , it is more convenient to plot @xmath80 on a logarithmic scale . note that the discontinuity at the leftmost level crossing is pronounced in the fidelity but almost invisible in its derivative . ]
the ground - state fidelity is defined as the overlap between ground state wavefunctions at nearby parameter values : @xmath81 recent studies indicate that the fidelity provides useful signatures of quantum phase transitions @xcite . since we concentrate on mesoscopic spin structures , our purpose is not to study phase transitions but to present fidelity signatures of the level crossings .
figure [ fig_fidelities ] presents fidelity data for @xmath78 .
the derivatives @xmath82 again show striking resonance - like features .
the special nature of the leftmost crossing is manifested even more dramatically here the derivatives match almost perfectly across this crossing , even though the fidelities themselves are discontinuous .
we now turn to the _ dynamics _ of entanglement in the frustrated ferromagnetic chain .
while each entanglement measure has its own dynamics when the quantum state of the spin ring evolves outside equilibrium , a full study of the dynamics of all such measures is outside the scope of the present investigation , especially since such comprehensive entanglement dynamics studies are yet to be completed for simpler spin systems .
we therefore restrict the presentation to the dynamics of the sublattice entanglement entropy ( @xmath4 ) . the fact that our system has true level crossings , place them in the complete opposite limit to what is necessary for adiabatic quantum computing @xcite .
it is therefore perhaps natural to study the opposite limit of adiabatic parameter changes , namely parameter _ quenches _ where @xmath46 changes as a step function . in an isolated finite spin ring , a quench of this type leads to oscillatory behaviors of most quantities , including entanglement entropies for various partitions ; there is no relaxation mechanism for the system to reach its ground state .
this is shown in fig .
[ fig_time - evolution ] ( center panel ) . studying the evolution after a parameter quench from @xmath83 to @xmath84
involves following the wavefunction @xmath85 explicitly in time , where the initial state @xmath86 is the ground state of @xmath87 .
we performed this calculation by expanding the operator @xmath88 to sufficiently high order for each time step .
we also explore relaxation issues , specifically the effect of the @xmath40-@xmath41-@xmath40-@xmath41- crossings , by adding an artificial dissipation to the temporal evolution ( fig .
[ fig_time - evolution ] rightmost panel ) . instead of evolving in time @xmath89
, we evolve in @xmath90 .
this form of damping ( ` cooling ' ) preserves the lattice momentum but not the energy .
we display entanglement evolutions for the large - damping case of @xmath91 , and also a single example with smaller damping , @xmath92 . with damping , the sublattice entanglement @xmath93 relaxes to the ground - state @xmath4 corresponding to the final @xmath46 only if the starting state has the same lattice momentum @xmath94 . for @xmath95 ,
two level crossings divides the parameter space into regions i and iii with the same @xmath96 and an intermediate region ii with @xmath97 .
the entanglement entropy after a quench to @xmath44 ( region i ) relaxes to the ` correct ' value if the starting ground state is in region i or iii , but not if it is in region ii .
the structure of level crossings and alternating ground - state momenta thus allow selective targeting of quantum ground states , and therefore could serve as the basis of selective variations of the basic idea behind adiabatic quantum computing .
left panel shows sublattice entanglement entropy for @xmath95 .
center [ rightmost ] panel shows evolution of @xmath4 after a quench , without [ with ] dissipation .
initial values are @xmath98 ( solid ) , @xmath99 ( dotted ) , @xmath100 ( dashed ) .
final value is @xmath101 in each case .
damping constant in rightmost panel is shown for @xmath91 for all three @xmath102 . for @xmath98 ,
the thick solid line has smaller damping , @xmath92 . in the presence of dissipation
, @xmath4 relaxes to the ground state value when the initial state has same lattice momentum ( @xmath96 ) but not when the initial state has different lattice momentum ( @xmath97 ) . ]
we have identified a mesoscopic system with unusual entanglement properties .
our system is the finite - size analog of the intriguing ferromagnetic - antiferromagnetic chain .
the surprising long - distance concurrence we have found , with tunable distance , makes this system ideal for various quantum information processing tasks .
for example , _ entanglement transport _ is here simply a matter of tuning the interaction parameter while remaining in the ground state .
our calculations of multi - qubit entanglement entropies and the fidelities have revealed unexpected aspects of the level crossings present in these systems .
we have also presented entanglement dynamics results , again indicating possible quantum information processing applications .
our work opens up a number of questions .
first , our focus on the @xmath40-@xmath41-@xmath40-@xmath41- level crossings raises the question : which of the entanglement behaviors reported here also carry over to the other systems where such level crossings have been observed ? to the best of our knowledge , no unified study of such crossings exists .
our entanglement results call for a general investigation of finite - size level crossing sequences occurring as a result of incommensurate correlations .
second , our study of entanglement dynamics treats only a minute fraction of the possible kinds of dynamics , and only a single entanglement measure out of many . since entanglement measures are sensitive to a variety of wavefunction characteristics not necessarily present in traditional condensed - matter measures such as correlation functions , the dynamics of these new measures present rich opportunities for characterizing non - equilibrium phenomena .
again , comprehensive investigations are called for , but the study of the evolution of various entanglements would be appropriately first done in simpler ( perhaps exactly solvable ) many - body systems .
g. vidal , j. i. latorre , e. rico and a. kitaev , phys .
90 * , 2279021 ( 2003 ) .
j. i. latorre , e. rico , and g. vidal , quantum info . and comput . , * 4 * , 48 ( 2004 ) .
korepin ; phys .
lett . * 92 * , 096402 ( 2004 ) .
p. calabrese and j. cardy ; j. stat .
* 0406 * , 002 ( 2004 ) .
d. gioev and i. klich , phys .
lett . * 96 * , 100503 ( 2006 ) .
m. m. wolf , phys .
lett . * 96 * , 010404 ( 2006 ) .
t. barthel , m .- c .
chung , and u. schollwck , phys .
a * 74 * , 022329 ( 2006 ) .
a. kitaev and j. preskill ; phys .
lett . * 96 * , 110404 ( 2006 ) .
m. levin and x. g. wen ; phys .
lett . * 96 * , 110405 ( 2006 ) .
e. fradkin and j. e. moore , phys .
* 97 * , 050404 ( 2006 ) .
h. m. wiseman and j. a. vaccaro , phys .
lett . * 91 * , 097902 ( 2003 ) .
m. r. dowling , a. c. doherty , and h. m. wiseman , phys .
a * 73 * , 052323 ( 2006 ) .
y. shi , j. phys .
a * 37 * , 6807 ( 2004 ) .
m. haque , o. zozulya , and k. schoutens , phys .
* 98 * , 060401 ( 2007 ) .
h. katsura and y. hatsuda , j. phys . a * 40 * , 13931 ( 2007 ) .
r. santachiara , f. stauffer , and d. cabra , j. stat . mech .
l05003 ( 2007 ) .
o. zozulya , m. haque , and k. schoutens , phys .
a * 78 * , 042326 ( 2008 ) . a. m. luchli and c. kollath ; j. stat .
p05018 ( 2008 ) .
p. calabrese and j. cardy ; j. stat .
p06008 ( 2007 ) .
g. de chiara , s. montangero , p. calabrese , r. fazio ; j. stat .
p03001 ( 2006 ) .
barmettler _ et al_. ;
arxiv:0803.1643 . l. amico _ et al_. ; phys . rev .
a * 69 * 022304 ( 2004 ) .
p. zanardi & n. paunkovic ; phys .
e * 74 * , 031123 ( 2006 ) .
p. buonsante & a. vezzani ; phys .
98 * , 11601 ( 2007 ) .
m. cozzini , p. giorda , and p. zanardi , phys .
b * 75 * , 014439 ( 2007 ) .
m. cozzini , r. ionicioiu , and p. zanardi , phys .
b * 76 * , 104420 ( 2007 ) .
yang , phys .
b * 76 * , 180403(r ) ( 2007 ) | we study the entanglement content of a class of mesoscopic tunable magnetic systems .
the systems are closed finite spin-1/2 chains with ferromagnetic nearest - neighbor interactions frustrated by antiferromagnetic next - nearest - neighbor interactions .
the finite chains display a series of level crossings reflecting the incommensurate physics of the corresponding infinite - size chain .
we present dramatic entanglement signatures characterizing these unusual level crossings .
we focus on multi - spin and global measures of entanglement rather than only one - spin or two - spin entanglements . we compare and contrast the information obtained from these measures to that obtained from traditional condensed matter measures such as correlation functions . |
a basic ( or _ vanilla _ ) option is a financial product which provides the holder of the option with the right to buy or sell a specified quantity of an underlying asset at a fixed price on or before the expiration date of the option .
there are many more complex options ( called _ exotic _ options ) in use today ; these are often of more practical interest and are harder to deal with . in many scenarios of practical interest and
as we shall assume in this article , the value of the underlying asset can be described by a diffusion process ; in a complete market , the value of the option can be expressed as the expectation under the risk neutral probability of a functional of the paths of the underlying diffusion process . in general
these expectations can not be calculated analytically .
the monte carlo ( mc ) method is a standard approach used to approximate these quantities and it has been extensively used in option pricing since @xcite .
subsequently , a wide variety of monte carlo approaches have been applied ( @xcite provides a thorough introduction ) .
+ the importance of monte carlo for option pricing against other numerical approaches is its ability to deal with high - dimensional integrals .
this is either in the time parameter of the option ( path - dependent options ) or in the dimension of the underlying ( basket of options ) , and more generally in both .
however , it has been noted in the option pricing literature that standard monte carlo estimates can suffer from high variability .
it has been seen in @xcite that , in many situations of practical interest , sequential monte carlo ( smc ) approaches can vastly improve over more standard monte carlo techniques . + sequential monte carlo methods are a general class of methods to sample from a sequence of distributions of increasing dimensions which have extensively been used in engineering , statistics , physics and other domains .
@xcite provides an introduction and shows how essentially all methods for particle filtering can be interpreted as some special instances of a generic smc algorithm .
smc methods make use of a sequence of proposal densities to sequentially approximate the targets via a collection of @xmath0 samples , termed particles . in most scenarios
it is not possible to use the distribution of interest as a proposal .
therefore , one must correct for the discrepancy between proposal and target via importance weights . in the majority of cases of practical interest , the variance of these importance weights increases with algorithmic time .
this can , to some extent , be dealt with a resampling procedure consisted of sampling with replacement from the current weighted samples and resetting them to @xmath1 ( adaptive resampling ) .
the variability of the weights is often measured by the effective sample size ( ess ) .
several convergence results , as @xmath0 grows , have been proved @xcite .
smc methods have also recently been proven to be stable in certain high - dimensional contexts @xcite .
+ the main contributions of this paper are as follows .
we develop the formal framework of _ weighting functions _ ; this technique has already been used , implicitly or explicitly , in @xcite . exploiting this framework ,
we develop tailored methods for pricing of barrier options in high dimensional settings .
it is also applied to the pricing of target accrual redemption note ( tarn ) which are another widely traded kind of path dependent options that are notoriously difficult to accurately value . on the theoretical side , we provide with a proof of the unbiasedness of the smc
estimates when an adaptive resampling scheme is used .
+ this paper is structured as follows . in section
[ sec : options ] we provide background details on option pricing . in section [ sec : smc ] we give a basic summary of smc methods . in section [ sec : weighting ] we give the weighting functions framework and its application in the context of our option pricing problems . in section [ sec : numerics ] our methods are illustrated numerically .
the appendix gives the proof of unbiasedness of our smc estimate in the adaptive resampling case .
+ in the remainder of this article , we use the notation @xmath2 to denote the @xmath3-dimensional euclidean space and @xmath4 . a normal distribution with mean @xmath5 and variance @xmath6 is denoted by @xmath7 and its density at @xmath8 is denoted by by @xmath9 .
@xmath10 denotes the @xmath3-dimensional identity matrix .
@xmath11 denotes expectation .
options come in two basic kinds - call and put .
call options give the right to buy and put options give the right to sell . in this context , there are two main kinds of options - american and european .
american options can be exercised at any time prior to expiration whereas european options can be exercised only at expiration .
we focus on european options in this paper .
european call / put options are known as vanilla options since they are relatively simple in structure .
an exotic option is an option which has features making it more complex than commonly traded vanilla options .
path - dependent options are an example , in which case the payoff depends on the value of the underlying at some ( or all ) time points prior to the expiration date .
we consider two kinds of path - dependent options in this paper , namely barrier options and tarn s , which we shall describe shortly .
consider a collection of @xmath3 underlying assets ; this is also known as a basket .
we denote by @xmath12 the value of the assets in the basket .
@xmath13 is typically modelled by a diffusion process .
one such process is a black - scholes model with a drift and a volatility @xmath14 where @xmath15 is the drift function , @xmath16 is the volatility and @xmath17 denotes a brownian motion in @xmath2 with mean @xmath18 and covariance matrix @xmath19 .
it is reasonable to assume that @xmath17 is normalized , that is , @xmath20 for all @xmath21 ; this assumption is valid because the scale factor can be included in the volatility term .
there is an interest rate @xmath22 , which can depend on time as well .
+ in general , it is hard to analytically work with except in simple scenarios .
this has lead to several discretization methods being available in literature ( @xcite ) with varying levels of accuracy and complexity .
one of the most widely used discretization methods is the euler - maruyama discretization and we work with it in this paper ; however other discretization schemes could also be used which could lead to a lower bias .
consider discretized time points @xmath23 . by letting @xmath24
denote the logarithm @xmath25 and writing @xmath26 in place of @xmath27 , the euler - maruyama discretization of is @xmath28 for @xmath29 , where @xmath30 , @xmath18 denotes the origin in @xmath2 , and @xmath31 and @xmath32 to denote @xmath33 and @xmath34 respectively . ] .
+ we do not focus on the level of discretization here .
given a particular discretization , we apply our methods to it .
the methods developed here could however be used in a multilevel setup ( as in @xcite ) ; we do not explore this further in this paper .
+ we assume the drift @xmath35 in order to keep things simple . is a constant other than @xmath18 , then it is trivial to extend the methods we propose .
if it is a function of the asset value , we could do things similar to what we do in the local volatility model considered later .
] we work with two cases , one where the volatility @xmath36 is a constant , and another where it depends on the price of the underlying .
we now describe two kinds of path - dependent options and we shall later demonstrate our methods on these .
a barrier option is an exotic derivative , typically an option , on the underlying asset(s ) whose value(s ) on reaching pre - set barrier levels either springs the option into existence or extinguishes an already existing option .
barrier options exist for baskets as well and the barrier conditions may in general be defined as a function of the underlying assets .
for example , the function of the underlying assets could be the mean or could be the maximum of their values .
there are two kinds of barrier options : * when the option springs into existence if a function of the underlying asset values breaches prespecified barriers , it is referred to as being ` knocked - in ' , and * when the option is extinguished if a function of the underlying asset values breaches prespecified barriers , it is referred to as being ` knocked - out ' .
we consider knocked - out options . these are options that are ` alive ' as long as the values of the underlying satisfy barrier conditions at some ( or all ) time points prior to the expiration .
if the barrier condition is breached , the option ` dies ' leading to a zero payoff .
if the option is still ` alive ' at the expiration date , then it gives a payoff that is akin to either a call or a put option ( depending on the option type ) .
+ we consider a basket of @xmath3 underlying assets and a barrier option based on these .
barrier options are hard to price using standard mc in the sense that most ( if not all ) of the particles lead to a zero payoff and this contributes to a high variance of the final estimate . because of the sequential nature of the evolution of the asset values over time , this is a natural example of a setting where smc methods can be applied ; indeed @xcite talks about how smc methods can be used in this context .
we extend their method and show that one can obtain significant gains by choosing the @xmath37 s in section 5 of their paper even by heuristic methods . a target accumulation redemption note ( tarn )
provides a capped sum of payments over a period with the possibility of early termination ( knockout ) determined by target levels imposed on the accumulated amounts .
a certain amount of payment is made on a series of cash flow dates ( referred to as fixing dates ) until the target level is breached .
the payoff function of a tarn is path dependent in that the payment on a fixing date depends on the spot value of the asset as well as on the accumulated payment amount up to the fixing date .
typically , commercial software solutions for pricing tarn s are based on the mc method .
+ there are different versions of tarn products used in fx trading . for simplicity , we consider here a specific form of tarn s .
consider a sequence of fixing dates @xmath38 and a function @xmath39 .
the function @xmath40 is decomposed into its positive and negative parts as @xmath41 .
gain and loss processes are defined as follows : @xmath42 these are the amounts of positive and negative cashflows respectively .
there are two cashflow cutoffs @xmath43 and @xmath44 . stopping times @xmath45 and @xmath46
are defined as @xmath47 these are the first times when the positive and negative cash flows cross their cut - offs respectively .
the overall stopping time @xmath48 is defined as @xmath49 which is the first time either the positive or the negative cash flows cross a cut - off .
the price of the tarn is the expected value of the overall cash flow , @xmath50 $ ] .
here we have assumed that the interest rate is @xmath51 , if it was nt then the expectation would be a weighted sum with the weights corresponding to discounting factors .
the reason we assume the interest rate is @xmath51 will be explained in section 4.3 .
the main difficulty in applying standard mc methods in this scenario arises from the fact that the function @xmath40 may be discontinuous .
smc methods can be used instead , which we shall show later .
smc methods are a general class of mc methods that sample sequentially from a sequence of target probability densities @xmath52 of increasing dimension , where each distribution @xmath53 is defined on the product space @xmath54 .
writing @xmath55 it is required only that @xmath56 be known pointwise . in particular , the normalizing constants @xmath57 may be unknown and an estimate is obtained by the smc method .
smc provides an approximation of @xmath58 and an estimate of @xmath59 at time 1 , then an estimate of @xmath60 and an estimate of @xmath61 at time 2 , and so on .
these methods work by propagating a collection of @xmath0 particles using a sequence of importance sampling and resampling steps . in what follows
, superscript @xmath62 shall denote the @xmath63-th particle .
when we write an operation with superscript @xmath62 , we mean that it happens for all @xmath0 particles . a proposal density @xmath64 is selected and particles are proposed from this .
the proposal density has the following structure : @xmath65 after proposing particles , associated _ unnormalized _ weights are computed recursively using the decomposition @xmath66 these can be written in the form @xmath67 where the _ incremental weight _
function @xmath68 is given by @xmath69 these are computed for each particle .
the weights @xmath70 s are unnormalized because they do not add up to 1 .
they are normalized by dividing by their sum and the normalized weights are denoted by @xmath71 s .
+ once we obtain a collection of @xmath0 weighted particles , they are _ resampled _ according to their weights @xmath72 .
this is a crucial step which ensures that the system does nt collapse to very few particles with very high weights .
resampling indices @xmath73 are chosen such that @xmath74 .
the system is then updated by setting @xmath75 and @xmath76 for @xmath77 .
resampling is expensive and in practice is done only if the variance of the weights is high .
one such method is to use the effective sample size ( ess ) , defined at time @xmath78 as @xmath79 and to perform resampling at time @xmath78 if the ess falls below a certain threshold , usually taken to be @xmath80 or @xmath81 .
algorithm [ algo : smc ] describes a general smc algorithm . set initial weights @xmath82 and initial estimate of normalizing constant @xmath83 .
sample @xmath84 , compute unnormalized weights @xmath85 , update estimate of normalizing constant @xmath86 , compute normalized weights @xmath87 , resample @xmath88 to obtain @xmath0 equally - weighted particles @xmath89 and set @xmath90 , set @xmath91 .
sample @xmath92 and set @xmath93 , compute incremental weights @xmath94 and unnormalized weights @xmath95 , update estimate of normalizing constant @xmath96 , compute normalized weights @xmath97 , resample @xmath98 to obtain @xmath0 equally - weighted particles @xmath99 and set @xmath100 , set @xmath101 .
we have two approximations of @xmath102 : @xmath103 which are equal if no resampling is used at time @xmath78 . here
@xmath104 denotes the dirac delta function .
we can also estimate @xmath105 through @xmath106
the goal of using weighting functions is to create a sequence of intermediate target densities that try to approximate the optimal importance sampling density to guide particles towards regions of interest .
consider a discrete time stochastic process @xmath107 on a general space @xmath108 .
suppose that we want to estimate the expectation of a function of this process @xmath109 .
the basic monte carlo method simulates @xmath0 independent realizations of the process @xmath110 , where @xmath111 denotes the @xmath63-th realization .
@xmath112}$ ] is estimated by @xmath113 if @xmath114 is of the form where @xmath109 is zero over most of the space @xmath108 and non - zero only on a small subset , most of the @xmath115 s would be zero .
this subsequently leads to a high variance of the resulting estimate .
+ more specifically , since @xmath108 is the sample space , we write @xmath116 where @xmath117 and @xmath118 if @xmath119 is much larger than @xmath120 , simulating @xmath121 independent realizations will lead to most of them lying in @xmath122 .
our goal is to use smc to simulate more particles from @xmath123 . in order to do that
, we consider a sequence of positive potential functions @xmath124 , @xmath125 , such that @xmath126 and write @xmath127 h(x_{1:n}).\ ] ] + the goal is to choose the @xmath128 s such that they guide the particles towards being in @xmath123 through the weighting and resampling steps of smc .
+ when this is done on a path space , the resulting algorithm is sometimes known as tempering .
@xcite considers this and shows how one can construct an artificial sequence of intermediate target densities on the path space which guide particles towards regions of interest .
doing it on the path space however makes it computationally expensive and we do not work on the path space .
we consider a discretely monitored barrier option monitored on a series of monitoring dates @xmath129 and a sequence of lower and upper barriers @xmath130 and @xmath131 respectively .
we suppose that the barrier conditions are that all the underlying asset values lie inside their respective barriers at the monitoring times .
this is a simplistic assumption and makes it easier for us to demonstrate our methods ; more complicated barrier conditions could also be used .
+ for ease of notation , we remove the @xmath132 from the barriers and replace it simply by @xmath21 .
let @xmath133 for @xmath134 and let @xmath114 denote the payoff function at time @xmath135 .
the sequence of random variables @xmath107 then forms a markov chain .
the price of the barrier option is @xmath136 } \footnote{we have assumed here that the interest rate is 0 .
if the interest rate was $ r$ , then there would be a factor of $ e^{\int_{0}^{t } r(t ) d t } $ multiplied with $ q_{d}$. this is a constant and affects the variance of the estimate only upto a ( known ) scale factor.},\ ] ] where @xmath137 @xmath138 and @xmath139 denote the @xmath140-th components of @xmath141 and @xmath142 respectively . in this case , @xmath143 the authors in @xmath144 introduce an smc algorithm to estimate the price .
the proposal density @xmath145 is chosen to be density with respect to the underlying discretization . for each time interval
@xmath146 they simulate forward till @xmath132 and then resample particles that breach the barrier condition at time @xmath132 from among the particles that are still inside the barrier .
their algorithm is algorithm [ algo : smc_barrier ] .
it is commented that they do not use an adaptive version of resampling and instead always resample
. set initial weights @xmath82 and initial estimate of normalizing constant @xmath83
. sample @xmath84 .
sample @xmath147 and set @xmath93 .
compute unnormalized weights @xmath148 , update estimate of normalizing constant @xmath149 , compute normalized weights @xmath150 , resample @xmath98 to obtain @xmath0 equally - weighted particles @xmath99 and set @xmath100 .
the estimated price is @xmath151 moreover , since essentially only the normalizing constant is being computed , there is no issue of path degeneracy .. this causes estimates based on the entire paths being unreliable . ]
resampling paths outside the barriers at the monitoring times from paths which are inside the barriers improves the efficiency of the estimator with respect to the standard mc estimator .
it is remarked that for constant volatility , in a black - scholes context , one can sample the price to ensure , in one step , that the process always survives ; see @xcite .
+ if , however , very few particles satisfy the barrier condition at time @xmath132 , then this estimate will also have a high variance .
this can happen for example if : the dimension @xmath3 is high ; the barrier condition is a narrow one , i.e , @xmath152 and @xmath142 are close to each other ; the volatility @xmath36 is high ; the time intervals @xmath146 are large .
our goal is to introduce a sequence of positive weighting functions so that they resolve this issue . in order to do so ,
we write @xmath153 \label{eq : h_n } \nonumber \\ & = & h_{0}({\mathbf s}_{0 } ) \times { { \ensuremath{\operatorname{e}}}\left [ \prod_{n=1}^{n } g_{n}(x_{n } ) \right ] } , \label{eq : price_potential_functions}\end{aligned}\ ] ] where @xmath154 the @xmath37 s , @xmath155 , are the sequence of positive weighting functions . in algorithm
[ algo : smc_barrier ] , they were simply 1 .
we attempt to choose them more carefully in order to approximate the optimal importance sampling densities .
the algorithm is algorithm [ algo : smc_barrier_wt_fn ] .
+ set initial weights @xmath82 and initial estimate of normalizing constant @xmath83 .
sample @xmath84 , compute unnormalized weights @xmath156 update estimate of normalizing constant @xmath157 , compute normalized weights @xmath158 , resample @xmath88 to obtain @xmath0 equally - weighted particles @xmath89 and set @xmath90 , set @xmath91 .
sample @xmath147 and set @xmath93 , compute unnormalized weights @xmath159 update estimate of normalizing constant @xmath160 , compute normalized weights @xmath161 , resample @xmath98 to obtain @xmath0 equally - weighted particles @xmath99 and set @xmath100 , set @xmath101 .
the estimated price is @xmath162 path degeneracy is again not an issue because we are still essentially estimating the normalizing constant .
this estimate is unbiased and a proof is provided in the appendix .
+ since @xmath163 is the indicator of @xmath164 being inside the barriers at time @xmath132 , paths which are outside the barriers at the monitoring times are discarded with probability 1 in this case as well .
however unlike in algorithm [ algo : smc_barrier ] , here we seek to give higher weights to particles which we think have a higher chance of being inside the barriers at the monitoring times .
what is being sought while using the weighting functions is an approximation to the optimal importance sampling density , that is , the density of @xmath165 ( at time @xmath166 ) conditional on it surviving at times @xmath132 , @xmath167 . in the case of the barrier options being considered , this corresponds to @xmath168 .
this is aachieved by giving higher weights to particles which have a higher chance of surviving at times @xmath132 .
for example , particles which are far away from the barriers have a lower chance of survival than particles which are closer to the barriers .
this is the intuitive idea behind our choice of weighting functions , and this is illustrated in section [ numerics : barrier ] .
we consider a tarn based on a single underlying asset .
since the main problem arises when the function @xmath40 is discontinuous , we consider a discontinuous @xmath40 to illustrate our methods .
let @xmath169 + this function has two big jumps at @xmath170 and @xmath171 .
the negative and positive cashflow cutoffs are @xmath172 and @xmath173 .
we recall that we had earlier assumed the interest rate is @xmath51 .
this is now justified .
the main reason why standard mc can not be used efficiently in this scenario is as follows . using mc ,
most of the particles stay inside @xmath174 for the first five fixing dates .
this leads to the contribution of the particle in the mc estimate being @xmath175 . however , an occasional particle escapes @xmath174 within the first five fixing dates and contributes a value that is significantly different from @xmath175 because of the big jumps in @xmath40 .
this causes the variance of the mc estimate to be high .
even if the interest rate was positive , this difficulty would still remain .
therefore for simplicity we assume the interest rate to be @xmath51 .
+ in order to go back to the previous notation , let @xmath176 and define @xmath177 .
then let @xmath178 be the new payoff function , where we have written @xmath179 in place of @xmath180 . by defining @xmath181 in this way , the problem has been transformed into the format that was being using before . in this case , @xmath182 a particle lies in @xmath122 if ( and only if ) it stays within @xmath174 for the first five fixing dates .
@xmath183 if , for example , the volatility is low or the time intervals @xmath184 are small . it is noted that this is the opposite of the barrier option case ( in which we considered the time intervals and volatility to be large ) .
we again consider a sequence of positive weighting functions @xmath185 and write @xmath186 \times \frac { g(s_{1:n } ) } { h_{t_{5 } } ( s_{t_{5 } } ) } , \ ] ] where @xmath187 ; we are basically doing the same thing as in the barrier option case .
the goal is again to guide particles towards being in @xmath123 , and we show that this can be achieved through the usage of some simple weighting functions in algorithm [ algo : smc_barrier_wt_fn ] .
in this section we demonstrate numerically the benefit of using weighting functions . in order to compare the standard deviations of algorithms [ algo : smc_barrier ] and [ algo : smc_barrier_wt_fn ] , we run them 100 times with @xmath188 particles in each run .
we then look at the standard deviations of the 100 estimates and report the relative standard deviations ( the ratio of the standard deviations ) .
we consider a barrier option whose asset values evolve independently of each other .
this translates to @xmath189 .
the option type is call .
we assume that we can simulate forward one day at a time .
a common monitoring strategy is to monitor the underlying assets after every @xmath190 days for a total of @xmath191 time periods . in that case ,
@xmath192 for @xmath193 and @xmath194 .
we choose @xmath195 and @xmath196 .
algorithm [ algo : smc_barrier ] resamples at the end of @xmath191 days and so resets the system of particles by choosing all resampled particles being inside the barriers .
therefore the gain that we expect by using algorithm [ algo : smc_barrier_wt_fn ] can only be before the system is reset and this is why we choose @xmath197 . in what follows
, we refer to algorithm [ algo : smc_barrier ] as ` mc ' and algorithm [ algo : smc_barrier_wt_fn ] as ` smc ' .
+ the algorithms are run on different values of the dimension @xmath3 and the volatility @xmath198 . recalling , the targeted density at any time @xmath78 is proportional to @xmath199 , where @xmath200 denotes the density of @xmath201 .
this implies that the targeted marginal density at time @xmath78 is proportional to @xmath202 , where @xmath203 denotes the marginal density of @xmath204 at time @xmath78 .
we denote the targeted density at time @xmath78 by @xmath205 .
since our basket consists of @xmath3 independent assets , we choose the marginal targeted densities ( at different time points ) to be the product of @xmath3 ( unnormalized ) densities .
that is , @xmath206 is such that @xmath207 , where @xmath208 . in the constant volatility case ,
the marginal density of @xmath209 at time @xmath78 is known and is denoted by @xmath210 ; this is simply a product of @xmath3 gaussians .
therefore the weighting functions are @xmath211 $ ] .
+ we choose @xmath212 .
since our goal is to push particles towards being inside the barriers at time @xmath190 , we consider a ( one dimensional ) brownian bridge with volatility @xmath213 tied down at @xmath214 and @xmath215 .
however , since the variance of the brownian bridge goes to 0 as the time approaches @xmath190 , we can not directly use it as it would lead to a high variance .
so we add @xmath216 to the standard deviation of the brownian bridge .
it is remarked that @xmath216 is somewhat an arbitrary choice .
this is not necessarily the best choice , and we do not claim so ; this choice just serves to illustrate the benefit of using weighting functions .
we have noticed gains even while using other values , all we need to do is to ensure that the targeted densities do nt tend towards being degenerate . in practice
, we introduce the weighting functions only after time @xmath217 because we want to let particles initially explore the space and then give higher weights to particles which are more likely to be inside the barriers at time @xmath190 .
+ ) for a barrier option with @xmath218 and @xmath195 for a constant volatility model when we target a brownian bridge ; in the left figure , the dimension @xmath219 ; in the right figure , the volatility @xmath220.,scaledwidth=100.0% ] the results are in figure [ fig : fig2 ] .
we observe a gain in the standard deviation by using these functions . in general ,
if the volatilities , initial prices and strike prices were different for the different chains , we could have considered a product of @xmath3 dissimilar brownian bridge densities as the weighting functions .
+ since the actual goal is to estimate the optimal importance density , we try to do it in our simple setup to see how much better it does .
the marginal densities of a particle given that it survives at time @xmath190 can be approximated by gaussian densities .
in fact , this is what we were doing earlier when we choose the targeted distribution to be a brownian bridge . instead of working with a brownian bridge
, we can try to estimate the means and the variances of the gaussian densities by first simulating particles in one dimension and then looking at the means and variance of those that survive .
+ in this regard , we simulate @xmath221 particles in one dimension with volatility @xmath220 .
we then look at the particles that survive at the end . given the particles that survive , we look at their marginal means and variances .
this approximates the marginal densities ( in one dimension ) of a particle given that it survives .
since we are in one dimension , the number of particles surviving is much higher than what we would have in higher dimensions ( about 3900 out of 10000 in our simulations ) .
this means that the probability of survival in one dimension is about @xmath222 for @xmath220 , which for example in 10 dimensions would mean a probability of survival of @xmath223 ( since the chains are independent ) .
we call this the ` optimal ' target .
+ the results are in figure [ fig : fig3 ] .
we can observe an improvement against the standard mc method , relative to figure [ fig : fig2 ] .
however , the gains are not very drastic and serve to illustrate that even the simple intuitive use of a brownian bridge does well .
it is worth noting that even by choosing the volatility to be @xmath224 and running the chain in one dimension , we observe a gain for other values of the volatility as well .
this suggests that even if the volatilities were different for the different chains , we could have simply run a chain in one dimension with a value of the volatility that is in the range of the volatilities and still get decent results .
+ ) for a barrier option with @xmath218 and @xmath195 for a constant volatility model when we target the ` optimal ' target ; in the left figure , the dimension @xmath219 ; in the right figure , the volatility @xmath220.,scaledwidth=100.0% ] a note on running times : we have observed that the running time for mc is less than three times that of smc ( even without having optimized the codes too much ) . thus if we were to take running times into account , we could have run mc with @xmath225 particles in the same time in which we run smc with @xmath0 particles . the standard deviation for mc using @xmath225 particles would be @xmath226 times less than that of mc with @xmath0 particles .
so even by taking running times into account , smc outperforms mc . in the last example ,
in addition we run mc with @xmath227 particles .
we should take this time into account as well , but we run it for only one value of the volatility and it is fast .
the reason we do not consider the running times in our figures is that we believe it might be possible to optimize the codes even further .
the intention is to demonstrate the benefit of using weighting functions in smc .
local volatility models are in practice used more frequently as they are typically more accurate . the local volatility function considered here
is a linear interpolation between the values of the volatility @xmath213 at @xmath228 and the values of the underlying @xmath229 at @xmath230 . in this case
, we do not know the marginal densities of @xmath209 . for simplicity
, we still assume that the volatility functions are the same for all the chains and approximate it in the following way .
consider a single chain .
if we knew @xmath231 , @xmath232 , then we could approximate @xmath233 by @xmath234 .
we approximate @xmath235 iteratively as follows ( assuming that the time step of the discretization @xmath236 is constant ) : @xmath237 } & = & s_{n-1 } - \frac{1}{2 } \sigma^{2}(s_{n-1 } ) \delta t \nonumber \\ \rightarrow { \ensuremath{\operatorname{e}}}s_{n } & = & { \ensuremath{\operatorname{e}}}s_{n-1 } - \frac{1}{2 } \delta t { { \ensuremath{\operatorname{e}}}\left [ \sigma^{2}(s_{n-1 } ) \right ] } \nonumber \\ & \approx & { \ensuremath{\operatorname{e}}}s_{n-1 } - \frac{1}{2 } \delta t \sigma^{2 } \left ( { \ensuremath{\operatorname{e}}}s_{n-1 } \right ) ; \label{eq : approx_expectation}\end{aligned}\ ] ] therefore we approximate @xmath235 and @xmath238 iteratively .
we first approximate @xmath231 by ; given this estimate , we approximate @xmath239 by and keep doing this .
we approximate @xmath240 by @xmath241 iteratively . 1 .
we begin with a brownian bridge weighting function using the previously estimated values of @xmath242 in the target density and add @xmath243 to the standard deviation as before .
2 . as in the constant volatility model ,
we simulate @xmath221 particles ( for the local volatility model ) in one dimension and look at the marginal means and variance of the particles that survive .
we call this the ` optimal ' target and target it in the smc algorithm .
the results are in figure [ fig : barrier_lv ] .
we observe a significant gain , which as expected , gets more significant as the dimension increases .
this is because the probability of a particle surviving gets smaller as the dimension increases .
the gains are more significant when we target the ` optimal ' .
as far as running times are concerned , the observations are the same as in the case of the constant volatility model .
thus algorithm [ algo : smc_barrier_wt_fn ] outperforms algorithm [ algo : smc_barrier ] even upon taking running times into consideration . ) for a barrier option with @xmath218 and @xmath195 for local volatility model ; the left figure is when we target a brownian bridge ; the right figure is when we target the ` optimal'.,scaledwidth=100.0% ] we again assume that we can simulate forward one day at a time .
we take the number of fixing dates @xmath244 and the interval between fixing dates @xmath245 days .
this roughly corresponds to a tarn where there is a cash flow at the end of every month for a maximum of two years .
we consider different values of the volatility .
a particle leads to a zero payoff if it stays within @xmath246 for the first five months and to a positive payoff if it escapes by the end of the fourth month .
we again compare standard mc and algorithm [ algo : smc_barrier_wt_fn ] ( referring to this as ` smc ' ) .
we note that contrary to barrier options , in this case we expect smc to do better than usual mc for _ low _ values of the volatility .
this is because in this case the region @xmath123 has low probability for lower values of the volatility unlike in the barrier option case where @xmath123 had low probability for higher values of the volatility .
this is why we restrict ourselves to lower values of the volaitility in this section than in section [ numerics : barrier ] .
when the underlying follows black - scholes dynamics with a constant volatility @xmath213 , we can directly simulate forward @xmath190 days at a time using the euler - maruyama discretization .
we run the algorithms different values of the volatility and report the results .
we try three ( increasingly more sophisticated , yet intuitive ) weighting functions to try and get particles to escape .
+ in the most naive version , we choose @xmath247 to simply be @xmath248 . this just gives higher weights to particles that are further away from @xmath249 at the end of the @xmath78-th month .
we obtain the results in figure [ fig : tarn_cv ] . + ) using most naive weighting functions for tarn in the constant volatility case.,scaledwidth=100.0% ] we then notice that the targeted density at time @xmath78 is proportional @xmath250 for @xmath251 .
if we choose @xmath247 to be of the form @xmath252 , then the marginal of the targeted density at time @xmath78 is proportional to @xmath253 , where @xmath254 is the marginal density of @xmath179 at time @xmath78 .
since we target @xmath248 , we choose @xmath255 and this leads to the results in figure [ fig : tarn_cv1 ] .
it is seen that smc with weighting functions does much better in most cases , except in the last few cases .
the last few cases correspond to higher volatilities , in which case the probability of a particle escaping increases .
smc does better when the probability of a particle escaping is low , which is what we wanted . + ) using simple weighting functions for tarn in the constant volatility case.,scaledwidth=100.0% ] looking into the problem
further , we observe that particles which escape can do so in one of two ways : either to the left or to the right .
we fix a value of @xmath256 and run @xmath227 copies of the chain .
our goal is to look at the exact marginal distribution of particles of particles which have escaped .
we observe that approximately @xmath257 of particles escape to the left and @xmath258 escape to the right .
this is of course for a fixed @xmath213 , but we use these marginal approximations as our target ( we use a mixture of normals ) and run the smc algorithm for different values of @xmath213 .
the results are in figure [ fig : tarn_cv2 ] .
significant gains are observed in this case as well , and are more than the gains before . in this case
as well , we are trying to approximate the optimal importance density . + ) using a mixture of normal densities as the target for tarn in the constant volatility case.,scaledwidth=100.0% ] in all of the above three examples in this section ,
we have observed that the running time for smc is less than twice that of mc .
we can thus have similar conclusions as in the case of barrier options , which shows that smc outperforms mc even upon taking running times into consideration . in this case , the volatility @xmath213 is a function of the price of the underlying @xmath259 .
we can not simulate forward @xmath190 days at a time any more and we simulate forward one day at a time now . in this case the weighting functions are introduced each day for the first 5 time periods . + we consider a volatility function which is minimum at @xmath260 and increases on either side of @xmath260 .
the value at @xmath260 is @xmath261 .
the values in the @xmath174 are less than @xmath262 and are higher outside it .
we choose this because a low value of @xmath213 would mean less particles escaping , but for the ones that do escape we let explore the space more freely .
the local volatility function is a linear interpolation between the values of the volatility @xmath213 at @xmath263 and the values of the underlying @xmath229 at @xmath264 .
+ we start off with the simplest weighting function @xmath265 . in this case , we observe that algorithm [ algo : smc_barrier_wt_fn ] has a standard deviation which is 1.83 times lower than that of mc .
+ since the marginal density of @xmath266 is unknown , we choose a crude approximation of the marginal and use it .
we choose @xmath267 to be the approximation .
this is the density of @xmath268 if the volatility had been a constant @xmath262 .
we run @xmath227 copies of the chain and look at the ones which escape towards the left and towards the right .
we then choose a mixture of normals ( in this case , the proportions @xmath269 and @xmath270 ) and repeat the experiments .
we observe that smc has a standard deviation which is 2.33 times less than that of mc .
+ the running times for smc are twice that of usual mc in this case .
this implies that while there are some gains to be achieved in using smc in this example , they are not as significant as before .
this is because of the fact that the local volatility function is low ( in the region of around @xmath271 ) only for a few values of the underlying , which suggests that one shouldd actually give even higher weights to force particles to escape .
nevertheless , this does serve to illustrate the potential benefits of using smc over mc .
we have formally presented the idea of weighting functions in smc .
the main idea is to try to estimate the sequence of optimal importance densities using a relatively heuristic technique of weighting functions , giving more weights to potentially favourable particles .
we demonstrated this on two examples from finance , but this can be extended to other cases ( for instance those in @xcite ) .
we have also seen that as we get closer to the optimal importance density the gains keep increasing .
however even using an approximation leads to significant gains .
since the idea is quite general , it can potentially be used in finance to price other kinds of path dependent options , for instance asian options as in @xcite .
it can also be used in the context of the estimation of marginal expectations w.r.t
. laws of jump - diffusions ( and hence some control problems ) ; see for instance @xcite .
the basic notion of weighting functions is also applied in @xcite for high - dimensional filtering problems . in summary the ideas in the article provide a simple formalisation of many that have previously appeared in the literature .
+ there are many extensions of the work in the article . for instance
one can adaptively determine the potential functions in the sequence , albeit at the cost of statistical bias .
other ideas include further applications and adaptation to problems where the multi - level monte carlo method can be used ( e.g. @xcite ) .
we provide a proof of the unbiasedness of the estimate when the resampling is done adaptively and multinomially .
let @xmath272 be the successive resampling times and let @xmath273 be the most recent resampling time before time @xmath78 , with @xmath274 if no resampling has been carried out before time @xmath78 . also define @xmath275 and @xmath276 .
define @xmath277 where @xmath278
if resampling is carried out at time @xmath78 , then @xmath279 are the resampling weights . + + for any test function @xmath280 , we estimate @xmath281 } } $ ] by @xmath282 and we estimate the normalizng constant by @xmath283 .
define : @xmath284 where @xmath285 + observe that @xmath286 define the likelihod ratio as @xmath287 and let @xmath288 \nonumber \\ & = & \frac { \overline{v}_{1 } \cdots \overline{v}_{\tau_{r } } \overline{v}_{\tau_{r+1 } } } { z_{n } } \sum_{{m}=1}^{{m } } \frac { v_{n } \left ( x_{1:n}^{({m})}\right ) } { { m}\overline{v}_{n } } } { \psi \left ( x_{1:n}^{({m } ) } \right ) \nonumber \\ & = & \frac { z_{n}^{{m } } } { z_{n } } \widehat{\psi}_{or}. \nonumber \end{aligned}\ ] ] the third equality is because @xmath276 .
define @xmath289 iteratvely as @xmath290 and @xmath291 , where we recall that @xmath292 is the resampled index of the @xmath63-th particle at the @xmath293-th resampling step .
let @xmath294 these are the @xmath213-fields generated by the random variables associated with the @xmath121 particles just before and just after the @xmath166-th resampling step respectively .
let @xmath295 and define for @xmath296 , @xmath297 } } , \ ] ] where @xmath298 denotes expectation under the proposal density . then @xmath299 } } $ ]
let @xmath300 h_{\tau_{s-1}}^{({m } ) } , \nonumber \\
z_{2s}^{({m } ) } & = & \widetilde{f}_{\tau_{s } } \left ( \overline{x}_{1:\tau_{s}}^{({m } ) } \right ) h_{\tau_{s}}^{({m } ) } - \sum_{j=1}^{{m } } v_{\tau_{s}}^{(j ) } \widetilde{f}_{\tau_{s } } \left ( x_{1:\tau_{s}}^{(j ) } \right ) \widetilde{h}_{\tau_{s}}^{(j)}. \nonumber \end{aligned}\ ] ] it follows from the proposition that @xmath303 } & = & \psi_{n } \nonumber \\ \rightarrow { { \ensuremath{\operatorname{e}}}\left [ \frac { z_{n}^{{m } } } { z_{n } } \widehat{\psi}_{or } \right ] } & = & \psi_{n } \nonumber \\ \rightarrow { { \ensuremath{\operatorname{e}}}\left [ z_{n}^{{m } } \widehat{\psi}_{or } \right ] } & = & z_{n } \psi_{n}. \nonumber\end{aligned}\ ] ] taking @xmath304 yields the desired result .
we observe that @xmath305 \nonumber \\ & & + \hspace{0.1 in } \sum_{s=1}^{r+1 } \left [ \widetilde{f}_{\tau_{s } } \left ( x_{1:\tau_{s}}^{({m } ) } \right ) - \widetilde{f}_{\tau_{s-1 } } \left ( \overline{x}_{1:\tau_{s-1}}^{({m } ) } \right ) \right ] h_{\tau_{s-1}}^{({m } ) } \nonumber \\ & = & - \widetilde{f}_{\tau_{0 } } \left ( \overline{x}_{\tau_{0}}^{({m } ) } \right ) h_{\tau_{0}}^{({m } ) } + \sum_{s=1}^{r+1 } \widetilde{f}_{\tau_{s } } \left ( x_{1:\tau_{s}}^{({m } ) } \right ) h_{\tau_{s-1}}^{({m } ) } - \sum_{s=1}^{r } \sum_{j=1}^{{m } } v_{\tau_{s}}^{(j ) } \widetilde{f}_{\tau_{s } } \left ( x_{1:\tau_{s}}^{(j ) } \right ) \widetilde{h}_{\tau_{s}}^{(j ) } \nonumber\end{aligned}\ ] ] thus , @xmath306 @xmath307 @xmath308 @xmath309 @xmath310 the second inequality is from [ eq:3.new ] , and the last equality is because @xmath275 and @xmath311 } } = \psi \left ( x_{1:\tau_{r+1 } } \right ) l_{n } \left ( x_{1:\tau_{r+1 } } \right ) \nonumber \\
\rightarrow \widetilde{f}_{\tau_{r+1 } } \left ( x_{1:\tau_{r+1}}^{({m } ) } \right ) h_{\tau_{r}}^{({m } ) } & = & l_{n } \left ( x_{1:n}^{({m } ) } \right ) \psi \left ( x_{1:n}^{({m } ) } \right ) h^{({m})}_{\widetilde{\tau}(n ) } \hspace{0.2 in } \textrm{as } x_{1:\tau_{r+1}}^{({m } ) } = x_{1:n}^{({m } ) } \textrm { and } \widetilde{\tau}(n ) = \tau_{r}. \nonumber\end{aligned}\ ] ] @xmath313 } & = & { { \ensuremath{\operatorname{e}}}_{{m}}\left[\left .
\widetilde{f}_{\tau_{1 } } \left ( \overline{x}_{1:\tau_{1}}^{({m } ) } \right ) h_{\tau_{1}}^{({m } ) } - \sum_{j=1}^{{m } } v_{\tau_{1}}^{(j ) } \widetilde{f}_{\tau_{1 } } \left ( x_{1:\tau_{1}}^{(j ) } \right ) \widetilde{h}_{\tau_{1}}^{(j ) } \ , \right| { \mathcal{f}}_{1 } \right ] } \nonumber \\ & = & { { \ensuremath{\operatorname{e}}}_{{m}}\left[\left .
\widetilde{f}_{\tau_{1 } } \left ( \overline{x}_{1:\tau_{1}}^{({m } ) } \right ) h_{\tau_{1}}^{({m } ) } - \frac{1}{{m } } \sum_{j=1}^{{m } } \widetilde{f}_{\tau_{1 } } \left ( x_{\tau_{1}}^{(j ) } \right ) h_{\tau_{0}}^{(j ) } \ , \right| { \mathcal{f}}_{1 } \right ] } \hspace{0.3 in } \textrm{from } \eqref{eq:3.new } \nonumber \\ & = & { { \ensuremath{\operatorname{e } } } _ { { m}}\left[\left .
\widetilde{f}_{\tau_{1 } } \left ( \overline{x}_{1:\tau_{1}}^{({m } ) } \right ) h_{\tau_{1}}^{({m } ) } - \frac{1}{{m } } \sum_{j=1}^{{m } } \widetilde{f}_{\tau_{1 } } \left ( x_{\tau_{1}}^{(j ) } \right ) \ , \right| { \mathcal{f}}_{1 } \right ] } \hspace{0.57 in } \textrm{as } h_{0}^{(j ) } = 1 \nonumber \\ & = & 0 . \nonumber\end{aligned}\ ] ] the last equality is because the conditional distribution of @xmath314 given @xmath315 is that of @xmath0 i.i.d .
random vectors which take the value @xmath316 with probability @xmath317 .
also , @xmath318 } & = & { \ensuremath{\operatorname{e}}}\left [ \left \ { \widetilde{f}_{\tau_{2 } } \left ( x_{1:\tau_{2}}^{({m } ) } \right ) - \widetilde{f}_{\tau_{1 } } \left ( \overline{x}_{1:\tau_{1}}^{({m } ) } \right ) \right \ } h_{\tau_{1}}^{({m } ) } \bigg| { { \mathcal{f}}_{2 } } \right ] \nonumber \\ & = & { \ensuremath{\operatorname{e}}}\left [ \left \ { \widetilde{f}_{\tau_{2 } } \left ( x_{1:\tau_{2}}^{({m } ) } \right ) - \widetilde{f}_{\tau_{1 } } \left ( \overline{x}_{1:\tau_{1}}^{({m } ) } \right ) \right \ } \bigg| { { \mathcal{f}}_{2 } } \right ] h_{\tau_{1}}^{({m } ) } \nonumber \\ & = & 0 . \nonumber\end{aligned}\ ] ] the last equality is because @xmath319 } & = & { \ensuremath{\operatorname{e}}}_{{m } } \left [ { \ensuremath{\operatorname{e}}}_{q } \left ( \psi ( x_{1:n } ) l_{n}(x_{1:n } ) \bigg| x_{1:\tau_{s } } = x_{1:\tau_{s}}^{({m } ) } \right ) \bigg| { \mathcal{f}}_{2(s-1 ) } \right ] \nonumber \\ & = & { \ensuremath{\operatorname{e}}}_{q } \left ( \psi \left ( x_{1:n } \right ) l_{n } \left ( x_{1:n } \right ) \bigg| x_{1:\tau_{s-1 } } = x_{1:\tau_{s-1}}^{({m } ) } \right ) \nonumber \\ & = & \widetilde{f}_{\tau_{s-1 } } \left ( \overline{x}_{1:\tau_{s-1}}^{({m } ) } \right ) .
\nonumber \end{aligned}\ ] ] the last equality is by the tower property of conditional expectations .
proceeding in this way , it is seen that @xmath320 is a martingale difference sequence . | pricing options is an important problem in financial engineering . in many scenarios of practical interest , financial option prices associated to an underlying asset
reduces to computing an expectation w.r.t . a diffusion process .
in general , these expectations can not be calculated analytically , and one way to approximate these quantities is via the monte carlo method ; monte carlo methods have been used to price options since at least the 1970 s .
it has been seen in @xcite that sequential monte carlo ( smc ) methods are a natural tool to apply in this context and can vastly improve over standard monte carlo . in this article , in a similar spirit to @xcite we show that one can achieve significant gains by using smc methods by constructing a sequence of artificial target densities over time . in particular , we approximate the optimal importance sampling distribution in the smc algorithm by using a sequence of weighting functions .
this is demonstrated on two examples , barrier options and target accrual redemption notes ( tarn s ) .
we also provide a proof of unbiasedness of our smc estimate .
+ * key words : * diffusions ; sequential monte carlo ; option pricing + * ams subject classification : * primary 91g60 ; secondary 65c05 . * some contributions to sequential monte carlo methods for option pricing
* by deborshee
sen , ajay jasra & yan zhou department of statistics & applied probability , national university of singapore , singapore , 117546 , sg .
email:`[email protected] ; [email protected] ; [email protected] ` + aj was supported by a singapore ministry of education academic research fund tier 1 grant ( r-155 - 000 - 156 - 112 ) and is affiliated with the rmi and cqf at nus .
yz was supported by a singapore ministry of education academic research fund tier 2 grant ( r-155 - 000 - 143 - 112 ) . |
the transport of large molecules and small particles that are geometrically confined within zeolites @xcite , biological@xcite as well as designed nanopores@xcite , channels or other quasi - one - dimensional systems attracted attention in the last decade .
this activity stems from the profitableness for shape and size selective catalysis @xcite , particle separation and the dynamical characterization of polymers during their translocation @xcite . in particular ,
the latter theme which aims at the experimental determination of the structural properties and the amino acid sequence in dna or rna when they pass through narrow openings or the so - called bottlenecks , comprises challenges for technical developments of nanoscaled channel structures @xcite .
along with the progress of the experimental techniques the problem of particle transport through corrugated channel structures containing narrow openings and bottlenecks has give rise to recent theoretical activities to study diffusion dynamics occurring in such geometries @xcite .
previous studies by jacobs @xcite and zwanzig @xcite ignited a revival of doing research in this topic . the so - called _ fick - jacobs approach _
@xcite , accounts for the elimination of transverse stochastic degrees of freedom by assuming a ( infinitely ) fast equilibration in those transverse directions .
the theme found its application for biased particle transport through periodic @xmath1d planar channel structures @xcite as well as for tubes @xcite exhibiting smoothly varying side - walls . beyond the fick - jacobs ( fj ) approach , which is suitably applied to channel geometries with smoothly varying side - walls , there exist yet other methods for describing the transport through varying channel structures like cylindrical septate channels @xcite , tubes formed by spherical compartments @xcite or channels containing abrupt changes of cross diameters @xcite . in a recent work@xcite
we have provided a systematic treatment by using a series expansion of the stationary probability density in terms of a geometric parameter which specifies the channel corrugation for _ biased _ particle transport proceeding along a _
planar three - dimensional channel _ exhibiting periodically varying , axis symmetric side - walls .
we have demonstrated that the consideration of the higher order corrections to the stationary probability density leads to a substantial improvement of the commonly employed fick - jacobs approach towards extreme corrugate channels .
the object of this work is to provide an analytic treatment to biased brownian motion in _ cylindrical three - dimensional tubes _ with periodically varying radius . in @xmath2 [ sec :
structure ] we introduce the model system : a brownian particle in a confined tube geometry with periodically modulated boundaries .
the central findings , namely the analytic expressions for the probability density are presented in @xmath2 [ sec : longwave ] .
it turns out that the latter scales linearly with the particle velocity derived within the fj approach , cf .
[ sec : part_sinustube ] . in @xmath2 [ sec : corr ] we calculate the particle mobility and the diffusion coefficient and employ our analytical results to a tube with sinusoidal varying cross - section .
section [ sec : conclusion ] summarizes our findings .
the present paper deals with biased transport of overdamped brownian particles in a cylindrical tube with periodically varying cross - section , respectively , radius @xmath3 .
a sketch of a tube segment with period @xmath4 is shown in @xmath5 [ fig : channelprob ] .
the particles budge within a static fluid with constant friction coefficient @xmath6 .
as the particle radius ( assumed to be point - like ) is small compared to the tube radius , hydrodynamic particle - particle interactions as well as hydrodynamic particle - wall interactions can safely be neglected .
further the particles are subjected to an external force with static magnitude @xmath7 acting along the longitudinal direction of the tube @xmath8 , i.e. the corresponding potential is @xmath9 .
that is confining the motion of the overdamped , point - like brownian particle .
the periodicity of the tube structures is @xmath4 , the minimal and maximal tube widths are @xmath10 and @xmath11 , respectively . the constant force @xmath7 pointing in the direction of the tube is applied on the particles . ]
the evolution of the probability density @xmath12 of finding the particle at the local position @xmath13 at time @xmath14 is governed by the three - dimensional smoluchowski equation @xcite , i.e. , [ eq : sm ] @xmath15 is the probability current @xmath16 associated to the probability density @xmath12 .
the boltzmann constant is @xmath17 and @xmath18 refers to the environmental temperature . at the tube wall
the probability current obeys the no - flux boundary condition ( bc ) caused by the impenetrability of the tube walls , viz .
@xmath19 where @xmath20 is the out - pointing normal vector at the tube walls . for a tube with radius
@xmath3 the bc read @xmath21 [ eq : bc ] further the probability density satisfies the normalization condition @xmath22 as well as the periodicity condition @xmath23 . since the external force acts only in longitudinal direction @xmath12 is radial symmetric .
this allows a reduction of the problem s dimensionality from @xmath1d to @xmath24d by integrating @xmath25 over the angle @xmath26 ; yielding @xmath27 \notag \\ & + \frac{k_{b}t}{\eta}\,\frac{1}{r}\partial_r\left[r\,\partial_r p(x , r , t)\right]\ , , \label{eq : smoluchowski2 } \intertext{where the two - point probability density is defined as } p(x , r , t)=&\,\frac{1}{2\pi } { \int\limits_{0}^{2\pi } } d\phi\,p{\left(x , r,\phi , t\right)}\ , .
\label{eq : pdf_xr}\end{aligned}\ ] ] integrating @xmath25 further over the cross - section and taking the boundary conditions eqs . into account , one gets @xmath28 .
\label{eq : statsmo } \intertext{thereby the marginal probability density reads } p(x , t)=&\frac{1}{2\pi } { \int\limits_{0}^{r(x)}}dr\,r\,{\int\limits_{0}^{2\pi}}d\phi\,p{\left(x , r,\phi , t\right ) } \label{eq : marg_p } \,.\end{aligned}\ ] ] in ref .
@xcite we present a perturbation series expansion in terms of a geometric parameter for the problem of biased brownian dynamics in a planar three - dimensional channel geometry .
below we apply this method for brownian motion in cylindrical three - dimensional tubes . in doing
so we introduce dimensionless variables .
we measure the longitudinal length in units of the period length @xmath4 , viz .
@xmath29 . for the rescaling of the @xmath30-coordinate
, we introduce the dimensionless aspect parameter @xmath31 , i.e. the difference of the widest cross - section of the tube , i.e. @xmath32 , and the most narrow constriction at the bottleneck , i.e. @xmath10 , in units of the period length , yielding @xmath33 the dimensionless parameter @xmath31 characterizes the deviation of the boundary from the straight tube corresponding to @xmath34 .
several authors considered different choices for the expansion parameter like the averaged half width @xcite or the ratio of an imposed anisotropy of the diffusion constants @xmath35 @xcite .
we next measure , for the case of finite corrugation @xmath36 , the radius @xmath30 in units of @xmath37 , i.e. @xmath38 and , likewise , the boundary function @xmath39 .
time is measured in units of @xmath40 which is twice the time the particle requires to overcome diffusively , at zero bias @xmath41 , the distance @xmath4 , i.e. @xmath42 .
the potential energy is rescaled by the thermal energy @xmath43 , i.e. , for the considered situation with a constant force component in longitudinal direction , one gets @xmath44 , with the dimensionless force magnitude @xcite : @xmath45 after scaling the probability distribution reads @xmath46 .
below we shall omit the overbar in our notation .
further we concentrate only on the steady state , i.e. , @xmath47 , which is in fact , the only state necessary for deriving the key quantities of particle transport like the average particle velocity @xmath48 @xmath49 and the effective diffusion coefficient @xmath50 in force direction .
the latter is given by @xmath51 and can be calculated by means of the stationary probability density @xmath52 using an established method taken from ref .
@xcite . at steady state ,
the smoluchowski equation @xmath25 in dimensionless units becomes @xmath53+\frac{1}{r}\partial_r\left[r\,\partial_r p_\mathrm{st}(x , r)\right]=\,0\ , , \label{eq : smoluchowski_scaled}\end{aligned}\ ] ] and the no - flux boundary conditions eqs .
read @xmath54_{r = h(x ) } , \label{eq : bc2_u}\\ 0=&\,\partial_r p_\mathrm{st}(x , r)|_{r=0}\ , .
\label{eq : bc2_l } \end{aligned}\ ] ] [ eq : bc2 ]
we apply the asymptotic analysis @xcite to the problem stated by @xmath25 and eqs . .
in doing so , we use for the stationary probability density @xmath55 ( the index @xmath56 will be omitted in the following ) the ansatz @xmath57 in the form of a formal perturbation series in even orders of the parameter @xmath58 . substituting these expressions into @xmath25 ,
we find @xmath59+\sum_{n=1}^\infty \varepsilon^{2n } \left\{\frac{1}{r}\partial_r\left[r \partial_r p_n(x , r)\right ] \right . & \\
+ \partial_x\left[e^{-u } \partial_x{\left(e^{u}p_{n-1}(x , r)\right)}\right]\right\}\ , . &
\label{eq : longwave_pn } \end{split}\end{aligned}\ ] ] the no - flux bc at the tube walls @xmath60 , cf .
@xmath25 , turns into @xmath61 [ eq : longwave_bc ] we claim that the normalization condition for the probability density @xmath62 corresponds to the zeroth solution @xmath63 that is normalized to unity , @xmath64 consequently the higher orders in the perturbation series have zero average , @xmath65 .
further each order @xmath66 has to obey the periodic boundary condition @xmath67 .
in@xmath2 [ subsect : fj ] , we demonstrate that the zeroth order of the perturbation series expansion coincides with the fick - jacobs equation @xcite .
referring to @xcite an expression for the average velocity @xmath68 is known . moreover , in @xmath2 [ subsect : higherorders ] , the higher order corrections to the probability density are derived . using those results we are able to obtain corrections , see in @xmath2 [ subsect : mean_vel ] , to the average velocity beyond the zeroth order fick - jacobs approximation presented in the next section . for the zeroth order ,
@xmath25 read @xmath69 \intertext{supplemented with the corresponding no - flux boundary condition at $ r=0 $ as well as at $ r = h(x)$ } 0=&\,\partial_r p_0(x , r)\,.\end{aligned}\ ] ] we make the ansatz @xmath70 where @xmath71 is an unknown function which has to be determined from the second order @xmath72 balance given by @xmath25 : @xmath73.\end{aligned}\ ] ] integrating the latter over the radius @xmath30 and taking the no - flux boundary conditions eqs . into account
, one immediately obtains @xmath74 the effective entropic potential @xmath75 is defined by @xmath76 note , that upon an irrelevant additive constant , i.e. @xmath77 , the effective entropic potential corresponds to that given in ref .
@xcite . the normalized stationary probability density within the zeroth order reads @xcite @xmath78 expressing next @xmath71 by @xmath79 , see @xmath25
, then yields the celebrated stationary fick - jacobs equation @xmath80 \label{eq : fj_pdf}\end{aligned}\ ] ] derived previously in ref .
@xcite .
thus , we demonstrate that the leading order term of the asymptotic analysis is equivalent to the fj - equation . in the fj equation
the problem of biased brownian dynamics in a confined @xmath1d geometry is replaced by brownian motion in the tilted periodic one - dimensional potential @xmath75 . in general ,
the stationary probability density of finding an overdamped brownian particle budging in a tube with periodically varying cross - section is sufficiently described by @xmath25 as long as the extension of the bulges of the tube structures is small compared to the periodicity , i.e. @xmath81
. then the average particle current is calculated by integrating the probability flux @xmath82 over the unit - cell @xcite @xmath83 in the spirit of linear response theory , the mobility in dimensionless units is defined by the ratio of the mean particle current @xmath25 and the applied force @xmath84 , yielding @xmath85 note , that in order to obtain the mobility in physical units one has to multiply @xmath86 with the mobility of unconfined particles , i.e. @xmath87 . further , resulting from the normalization condition @xmath25 , the average particle velocity @xmath25 simplifies to @xmath88 therefore we derive that the average particle current is composed of ( i ) the fick - jacobs result @xmath68 , cf .
@xmath25 , and ( ii ) becomes corrected by the sum of the averaged derivatives of the higher orders @xmath89 .
we next address the higher order corrections @xmath89 of the probability density which become necessary for more corrugated structures . according to @xmath25
, one needs to iteratively solve @xmath90=&\,\mathfrak{l}\,p_{n-1}(x , r)\,,\quad n\geq 1\,,\end{aligned}\ ] ] under consideration of the boundary conditions eqs . .
in @xmath25 , we make use of the operator @xmath91 , reading @xmath92 . each solution of the second order partial differential equation @xmath25 possesses two integration constants @xmath93 and @xmath94 . the first one , @xmath93 , is determined by the no - flux bc at the centerline @xmath95 , cf .
@xmath25 , while the second provides the zero average condition @xmath96 . for the first order correction
, the determining equation reads @xmath97=&\,2\,{\left\langle \dot{x } \right\rangle}_0\,\partial_x { \left(\frac{1}{h^2(x)}\right)}\ , , \intertext{and after integrating twice over $ r$ , we obtain } p_1(x , r)=&\,-{\left\langle \dot{x } \right\rangle}_0\,{\left(\frac{h'(x)}{h^3(x)}\right)}\,r^2\ , .
\label{eq : finalsol_p1}\end{aligned}\ ] ] hereby , as requested above , the first integration constant @xmath98 is set to @xmath99 in order to fulfill the no - flux bcs , and the second must provide the normalization condition @xmath25 , i.e. @xmath100 . one notices that the first correction to the probability density becomes positive if the confinement is constricting , i.e. for @xmath101 and @xmath102 .
in contrast , the probability density becomes less in unbolting regions of the confinement , i.e. for @xmath103 .
please note , that the first order correction scales linearly with the average particle current @xmath68 .
overall , the break of spatial symmetry observed within numerical simulations in previous works @xcite is reproduced by this very first order correction .
particularly , with increasing forcing , the probability for finding a particle close to the constricting part of the confinement increases , cf .
@xcite . upon recursively solving , the higher order corrections , @xmath104 ,
are given by @xmath105 note that each single order @xmath89 , cf .
@xmath25 , satisfies the normalization condition , the bc at the centerline but does not obey the bc at the tube wall , see @xmath25 . the stationary probability density @xmath62 is obtained by summing all correction terms , cf . @xmath25 , yielding @xmath106 inserting @xmath25 into the equation for the no - flux bc at the tube wall , cf .
@xmath25 , results to @xmath107 according to @xmath25 the latter equals zero in the steady state . summing up
, the exact solution for the stationary probability density of finding a biased brownian particle in tube is given by @xmath25 .
the latter solves the corresponding smoluchowski equation @xmath25 under satisfaction of the normalization as well as the periodicity requirements .
more importantly the solution , cf .
@xmath25 , obeys the no - flux boundary conditions at the centerline @xmath95 as well as at the tube wall @xmath60 .
further one notices that @xmath62 is fully determined by the fick - jacobs results @xmath63 .
caused by @xmath108 , the contribution of the higher order corrections to the @xmath24d probability density scales linear with the average particle current in the fj limit @xmath68 , cf .
@xmath25 .
the latter is determined by the break of spatial symmetry induced by the external force .
consequently , in the absence of the external force @xmath109 the stationary probability density equals the zeroth order contribution @xmath110 despite the value of @xmath31 . according to @xmath25
, one recognizes that the average particle current scales with the average particle current obtained from the fick - jacobs formalism @xmath68 for all values of @xmath31 . therefore , in order to validate the obtained results for @xmath62 as well as to derive correction to the mean particle current it is required to calculate @xmath68 first .
in the following we study the key transport quantities like the particle mobility @xmath111 and the effective diffusion coefficient @xmath112 of point - like brownian particles moving in a sinusoidally - shaped@xcite tube . the dimensionless boundary function @xmath113 reads @xmath114 and is illustrated in @xmath5 [ fig : channelprob ] .
the function @xmath113 is solely governed by the aspect ratio of the minimal and maximum tube width @xmath115 .
obviously different realizations of tube geometries can possess the same value of @xmath116 .
the number of orders have to taken into account in the perturbation series @xmath25 , respectively , the applicability of the _ fick - jacobs approach _ to the problem , depends only on the value of the geometric parameter @xmath117 for a given aspect ratio @xmath116 .
first , we obtain the particle mobility @xmath118 within the zeroth - order ( fick - jacobs approximation ) . referring to eqs . and the dimensionless mobility is given by @xmath119 for the considered sinusoidal boundary function , cf .
@xmath25 , we obtain @xmath120 with the substitute @xmath121 . caused by the reflection symmetry of the boundary function @xmath113 the particle mobility obeys @xmath122 .
thus it is sufficient to discuss only the behavior for @xmath123 . in the limiting case of infinite large force strength
, the mobility goes to @xmath124 in the diffusion dominated regime , @xmath125 , the sutherland - einstein relation emerges @xcite and thus the dimensionless mobility equals the dimensionless effective diffusion coefficient : @xmath126 in the limit of vanishing bottleneck width , i.e. @xmath127 , the mobility , respectively , @xmath50 tends to @xmath99 .
in contrast , for straight tubes corresponding to @xmath128 , i.e. @xmath129 , the mobility as well as the effective diffusion coefficient equal their free values which are one in the considered scaling . in a recent work @xcite
, we studied the biased brownian motion in a planar three - dimensional channel geometry with periodically varying width @xmath130 .
we found that the asymptotic value for the mobility is given by @xmath131 for @xmath132 , cf .
( 45 ) in ref .
@xcite . comparing this result with the asymptotic value @xmath25 one
notices that the mobility and the effective diffusion coefficient in a tube , respectively , are less compared to case of a planar channel geometry . .
the maximum tube width is kept fixed , viz .
@xmath133 , while the aspect ratio is varied @xmath134 , respectively , the corresponding values for @xmath31 are @xmath135 .
the symbols correspond to the numerical obtained mobility , respectively , the effective diffusion coefficient . in subfigure (
a ) the lines correspond to the analytic result , cf .
@xmath25 . ] in @xmath5 [ fig : mobdeff ] we depict the dependence of the mobility @xmath111 and the effective diffusion coefficient @xmath112 on the external force magnitude @xmath84 .
the numerical results are obtained by solving the stationary smoluchowski equation @xmath25 using finite element method @xcite and subsequently calculating the average particle current according to @xmath25 . in order to determine the effective diffusion coefficient @xmath112
, one has to solve numerically the reaction - diffusion equation for the _ b - field _
@xcite . referring to @xmath5 [
fig : mobdeff](a ) one notices that the analytic predictions for the particle mobility @xmath25 are corroborated by numerics .
further , one observes that for the case of smoothly varying tube geometry , i.e. @xmath136 , the analytic result is in excellent agreement with the numerics for a large range of dimensionless force magnitudes @xmath84 , indicating the applicability of the fick - jacobs approach .
as long as the extension of the bulges of the tube structures is small compared to the periodicity , sufficiently fast transversal equilibration , which serves as fundamental ingredient for the validity of the fick - jacobs approximation is taking place . the effective diffusion coefficient @xmath112 exhibits a non - monotonic dependence versus the dimensionless force @xmath84 , see @xmath5 [ fig : mobdeff](b ) .
it starts out with a value that is less than the free diffusion constant in the diffusion dominated regime , i.e .
@xmath125 . according to the sutherland - einstein - relation the value equals the mobility value , cf .
@xmath25 .
then it reaches a maximum with increasing @xmath84 and finally approaches the value of the free diffusion from above .
further one notices that the location of the diffusion peak as well as the peak height depends on the aspect ratio @xmath116 .
with decreasing width at the bottleneck , while keeping the maximum width @xmath32 fix , the diffusion peak is shifted towards larger force magnitude @xmath84 .
simultaneously the peak height grows . in the limit of a straight tube ,
i.e. @xmath137 , as expected the effective diffusion coefficient coincide with its free value which is one in the considered scaling . on the particle mobility
is presented .
the value of @xmath31 is varied , viz .
@xmath138 , while the aspect ratio is kept fixed , viz . @xmath139 .
the solid line corresponds to the analytic result @xmath25 while the dash - dotted line indicates the asymptotic value one . ] in @xmath5 [ fig : mobdeff_eps ] we present the impact of the expansion parameter @xmath31 on the particle mobility .
it turns out that for values of @xmath140 the fick - jacobs approximation is in very good agreement with the simulation . with increasing geometric parameter
@xmath31 the difference between the fj - result ( solid line ) and the numerics is growing .
the more available space in the tube leads to a decrease of the particle mobility , respectively , of the effective diffusion coefficient .
consequently , the higher order corrections to the stationary probability density @xmath62 , see @xmath25 , respectively , the corrections to the mobility , cf .
@xmath25 , need to be included in order to provide a better agreement .
a common used way to include the corrugation of the channel structure bases on the concept of the spatially dependent diffusion coefficient @xmath0 which was introduced by zwanzig @xcite and subsequently supported by the study of reguera and rubi @xcite .
zwanzig obtained the fj equation , cf .
@xmath25 , from the full @xmath1d smoluchowski equation upon eliminating the transverse degrees of freedom supposing infinitely fast relaxation . in a more detailed view
, we have to notice that diffusing particles can flow out from/ or towards the wall in @xmath30-direction only at finite time .
these finite relaxation processes are included by scaling the diffusion constant in longitudinal direction by a position dependent function , viz . @xmath0 .
the latter substitutes the constant diffusion coefficient - which is one in the considered scaling - in the common stationary fj equation , cf .
@xmath25 , yielding @xmath141 .
\label{eq : fj_pdf_spatiald}\end{aligned}\ ] ] according to @xmath25 an expression for the particle mobility that is similar to the stratonovich formula @xcite for the mobility in titled periodic energy landscapes , but now including the the spatial diffusion coefficient @xmath0 , can be derived @xcite . in the diffusion dominated regime ,
i.e. @xmath125 , this expression simplifies to the lifson - jackson formula @xcite @xmath142 with the period average @xmath143 .
unfortunately , for many boundary functions @xmath113 it is impossible to analytically evaluate the expression @xmath25 .
a first systematically treatment taking the finite diffusion time into account was presented by kalinay and percus ( kp ) @xcite .
their suggested mapping procedure enables the derivation of higher order corrections in terms of an expansion parameter @xmath144 , which is the ratio of the diffusion constants in the longitudinal and transverse directions . within this scaling
, kp have shown that the fast transverse modes ( transients ) separate from the slow longitudinal ones and therefore the transients can be projected out by integration over the transverse directions . in what follows
we present a derivation for the spatially dependent diffusion coefficient @xmath0 which based on our previously considered perturbation series expansion for the stationary probability density @xmath2 [ sec : longwave ] . according to @xmath25 the marginal probability current @xmath145 , equivalent to @xmath25 ,
can be derived in an alternative way using the stationary two - point probability density @xmath62 , yielding @xmath146 the second equality determines the sought - after spatial dependent diffusion coefficient @xmath0 .
one immediately notices that the relation @xmath25 simplifies to @xmath147 in the force dominated regime @xmath148 .
then it follows that the spatially diffusion coefficient equals the free one , which is one in the considered scaling , @xmath149 in the opposite limit of small force strengths , i.e. for @xmath150 , diffusion is the dominating process
. then @xmath25 simplifies to @xmath151 inserting our result for the stationary probability density @xmath62 , cf .
@xmath25 , into @xmath25 we determine an expression for @xmath0 . in compliance with ref .
@xcite , we make the ansatz that all but the first derivative of the boundary function @xmath113 are negligible .
then the @xmath152-times applied operator @xmath91 , cf .
@xmath25 , simplifies to @xmath153 , yielding , @xmath154 inserting the latter into @xmath25 and calculating the complete sum , one finds @xmath155 for the spatially dependent diffusion coefficient @xmath0 in the diffusion dominated regime , i.e. for @xmath125 . using our above presented series expansion for the stationary probability density @xmath62
we confirm the expression for the spatially dependent diffusion coefficient @xmath0 previously derived by reguera and rubi @xcite and kp @xcite .
next , we derive an estimate for the mean particle current @xmath156 based on the higher expansion orders @xmath89 . referring to @xmath25
, the average particle current is composed of ( i ) the fick - jacobs result @xmath68 , cf .
@xmath25 , and ( ii ) becomes corrected by the sum of the averaged derivatives of the higher orders @xmath89 .
immediately one notices that the integration constant @xmath94 , resulting from the normalization condition @xmath25 , does not influence the result for the average particle velocity , cf .
@xmath25 .
we concentrate on the diffusion dominated limit @xmath125 and further we make the ansatz that all but the first derivative of the boundary function @xmath113 are negligible @xcite
. then the partial derivative of @xmath89 with respect to @xmath157 simplifies to @xmath158 inserting the latter into @xmath25 and integrating the results over one unit - cell , results in @xmath159 we obtain that the average transport velocity is obtained as the product of the zeroth - order fick - jacobs result and the expectation value of a complicated function including the slope of the boundary for @xmath125 . in the previously studied case of
biased brownian motion in a @xmath1d planar channel geometry @xcite we have found that the average transport velocity is obtained as the product of the zeroth - order fick - jacobs result and the expectation value of the spatially dependent diffusion coefficient @xmath160 .
in contrast to the @xmath1d planar geometry , for biased transport in extreme corrugated tubes the corrections term to the particle velocity does not coincide with the expectation value of @xmath161 , see @xmath25 . calculating the expectation value in @xmath25 for the considered tube geometry , cf .
@xmath25 , yields to the estimate @xmath162 where @xmath163 is the first hypergeometric function .
we derive that in the diffusion dominated regime the average velocity is obtained as the product of the zeroth - order fick - jacobs result and the correction term @xmath163 including the corrugation of the tube structure . referring to the sutherland - einstein relation , cf .
@xmath25 , if the average current @xmath164 ( or the effective diffusion coefficient @xmath165 ) is known in the zeroth order , the higher order corrections to both quantities can be obtained according to @xmath25 .
we have to emphasize that considering only the first derivative of the boundary function @xmath166 results in an additional term proportional to @xmath167 .
taking further the second derivative @xmath168 into accounts indicates that this second term is negligible compared to @xmath169 for arbitrarily value of @xmath31 . in units of the maximum channel width @xmath32 .
the latter is varied @xmath170 ( from top to bottom ) while the external bias is kept fixed @xmath171 ( corresponding to the diffusion dominated regime ) .
the symbols correspond to the numerical obtained mobility ( triangles ) and the effective diffusion coefficient ( circles ) .
the solid lines correspond to analytic higher order result , cf .
@xmath25 .
the zeroth order - fick - jacobs results given by @xmath25 collapse to a single curve hidden by the solid line for @xmath172 .
in addition the numerical evaluation of @xmath25 is represented by the dash - dotted lines . ] in @xmath5 [ fig : mobdeffperi_eps ] , we present the dependence of the @xmath111 ( triangles ) and @xmath112 ( circles ) on the slope parameter @xmath31 for @xmath171 .
one observes that the numerical results for the effective diffusion coefficient @xmath173 and the mobility @xmath111 coincide for all values of @xmath31 , thus corroborating the sutherland - einstein relation .
in addition , the fick - jacobs result , given by @xmath25 , the higher order result ( solid lines ) , see @xmath25 , and the numerical evaluation of the lifson - jackson formula using @xmath161 ( dash - dotted lines ) , cf . @xmath25 ,
are depicted in @xmath5 [ fig : mobdeffperi_eps ] .
for the case of smoothly varying tube geometry , i.e. @xmath136 , all analytic expressions are in excellent agreement with the numerics , indicating the applicability of the fick - jacobs approach . in virtue of @xmath25 ,
the geometric parameter is defined by @xmath174 and hence the maximal value of @xmath31 equals @xmath175 .
consequently the influence of the higher expansion orders @xmath176 on the average velocity @xmath25 and on the mobility , respectively , becomes negligible if the maximum tube s width @xmath175 is small . with increasing maximum width the difference between the fj - result @xmath25 and the numerics
is growing .
specifically , the fj - approximation overestimates the mobility @xmath177 and the effective diffusion coefficient @xmath178 .
consequently the corrugation of the tube geometry needs to be included . the consideration of @xmath161 , cf .
@xmath25 provides a good agreement for a wide range of @xmath31-values as long as the maximum width @xmath175 is on the scale to the period length of the tube , i.e. @xmath179 . upon further increasing the maximum width @xmath175 diminishes the range of applicability of the presented concept . in detail
the expression @xmath25 drastically underestimate the numerical results due to the neglect of the higher derivatives of the boundary function @xmath113 . put differently , the higher derivatives of @xmath113 become significant for @xmath180 .
in contrast , one notices that the result for the mobility and the diffusion coefficient , respectively , based on the higher order corrections to the stationary probability density @xmath25 is in very good agreement with the numerics . for tube geometries where the maximum width @xmath175 is on the scale to the period length ,
i.e. @xmath179 , the correction estimate matches perfectly with the numerical results .
further increase of the tube width results in a small deviation from the simulation results .
in summary , we have considered the transport of point - sized brownian particles under the action of a constant and uniform force field through a @xmath1d tube .
the cross - section , respectively , the radius of the tube varies periodically .
we have presented a systematic treatment of particle transport by using a perturbation series expansion of the stationary probability density in terms of a smallness parameter which specifies the corrugation of the tube walls . in particular
, it turns out that the leading order term of the series expansion is equivalent to the well - established _ fick - jacobs approach _ @xcite .
the higher order corrections to the probability density become significant for extreme bending of the tube s side - walls .
analytic results for each order of the perturbation series have been derived .
similar to biased brownian motion in a @xmath1d planar channel all higher order corrections to the stationary probability density scale with the average particle current obtained from the fick - jacobs formalism .
moreover , for the diffusion dominated regime , i.e. for small forcing @xmath125 , we calculate the correction to the mean particle velocity originated by the tube s corrugation using the series expansion for the stationary probability density . according to the sutherland - einstein relation ,
the obtained relation is also valid for the effective diffusion coefficient . in addition , by using the higher order corrections we present an alternative derivation for the spatially dependent diffusion coefficient @xmath0 which substitutes the constant diffusion coefficient present in the common fick - jacobs equation based on similar assumptions as those suggested by kalinay and percus as well as by rubi and reguera .
finally , we have applied our analytic results to a specific example , namely , the particle transport through a tube with sinusoidally varying radius @xmath3 .
we corroborate our theoretical predictions for the mobility and the effective diffusion coefficient with precise numerical results of a finite element calculation of the stationary smoluchowski - equation . in conclusion
, the consideration of the higher order corrections leads to a substantial improvement of the fick - jacobs - approach , which corresponds to the zeroth order in our perturbation analysis , towards more winding boundaries of the tube .
notably , we have shown that the common approach using the the spatially dependent diffusion coefficient @xmath0 fails for extreme corrugated tube geometries .
this article is dedicated to the memory of dr .
frank moss .
frank was an outstanding and enthusiastic scientist who was a true pioneer in the field of stochastic resonance @xcite and in the application of stochastic nonlinear dynamics to biological systems @xcite .
we will always remember him .
this work has been supported by the vw foundation via project i/83903
( l.s .- g .
, s.m . ) and i/83902 ( p.h .
, g.s . ) .
p.h . acknowledges the support the excellence cluster
nanosystems initiative munich
( nim ) .
51ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physreve.83.051135 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.95.164505 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.044501 [ * * , ( ) ] @noop _ _ , ed .
( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) \doibase doi:10.1063/1.3009621 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.86.3188 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.118103 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.238104 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/revmodphys.81.387 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) \doibase doi:10.1063/1.3527546 [ * * , ( ) ] link:\doibase 10.1103/physreve.74.051114 [ * * , ( ) ] link:\doibase 10.1063/1.3431756 [ * * , ( ) ] link:\doibase 10.1063/1.3561680 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreve.82.041121 [ * * , ( ) ] link:\doibase 10.1021/jp112393q [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreve.82.031143 [ * * , ( ) ] \doibase doi : 10.1016/j.chemphys.2010.04.012 [ * * , ( ) ] @noop _ _ , ed .
( , ) \doibase doi : 10.1016/0370 - 1573(82)90045-x [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.2710894 [ * * , ( ) ] @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreve.82.051128 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link : www.freefem.org [ `` , '' ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physreva.32.695 [ * * , ( ) ] link:\doibase 10.1103/physreva.33.4459 [ * * , ( ) ] \doibase doi : 10.1016/j.mbs.2006.11.014 [ * * , ( ) ] http://dx.doi.org/10.1140/epjst/e2008-00628-8 [ * * , ( ) ] \doibase doi : 10.1006/jtbi.2001.2445 [ * * , ( ) ] | biased brownian motion of point - size particles in a three - dimensional tube with smoothly varying cross - section is investigated . in the fashion of our recent work
@xcite we employ an asymptotic analysis to the stationary probability density in a geometric parameter of the tube geometry .
we demonstrate that the leading order term is equivalent to the fick - jacobs approximation .
expression for the higher order corrections to the probability density are derived . using this expansion orders we obtain that in the diffusion dominated regime the average particle current equals the zeroth - order fick - jacobs result corrected by a factor including the corrugation of the tube geometry . in particular
we demonstrate that this estimate is more accurate for extreme corrugated geometries compared to the common applied method using the spatially dependent diffusion coefficient @xmath0 .
the analytic findings are corroborated with the finite element calculation of a sinusoidal - shaped tube .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ particle transport in micro- and nanostructured channel structures exhibits peculiar characteristics which differs from other transport phenomena occurring for energetic systems .
the theoretical modelling involves fokker - planck type dynamics in three dimensions which can not be solved for arbitrary boundary conditions imposed by the geometrical restrictions .
recently , much effort is drawn on a reduction of the complexity of the problem resulting in the so - called fick - jacobs approximation in which ( infinitely ) fast equilibration in certain spatial directions is assumed .
within the present manuscript we derive a reduction method which ( i ) corresponds in zeroth order in the expansion parameter , which describes the corrugation of the tube wall , to the celebrated fick - jacobs result and ( ii ) extends the validity of the fick - jacobs approximation towards extreme corrugated tube structures . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ |
extinction in astronomy describes how much light is absorbed ( or scattered ) by dust grains in the medium between the source and the observer .
locally , extinction is measured by comparing the light emitted from stars of similar spectral type , for which differential variations from interstellar dust can be assessed .
local measurements in the milky way reveal a general trend in the extinction law although with significant scatter ( see , e.g. * ? ? ?
the behaviour of the extinction law reveals the overall properties of the dust grains in the interstellar medium @xcite .
closer to home , light scattered by the small air molecules in the atmosphere follow a typical @xmath3 law that explains the blueness of the day - time sky .
the typical milky way extinction law implies a shallower scaling dependence within the optical and near - infrared window , @xmath4@xcite .
this effective law results from the integration of the contribution from all grains in the dust component . a distribution with smaller grain sizes would make this law steeper , and a distribution with larger grains would induce a shallower dependence with wavelength , approaching the mie regime @xcite , where there is no wavelength dependence ( i.e. clouds are white ) .
one classical parameterisation of this smooth wavelength dependence is the ratio of total - to - selective extinction , @xmath5 , where @xmath6 is the extinction , measured in magnitudes , in the @xmath7 band , and @xmath8 is the colour excess , i.e. the additional reddening , measured as a @xmath9 colour , caused by dust .
in addition to this smooth trend in the extinction law , there are some features characteristic of a specific component . in our analysis extended to the near uv and optical spectral window
the most remarkable feature is the nuv bump at 2,175 @xcite , see fig .
[ fig : atten ] . although the component responsible for this feature is not confirmed , it is possibly carbonaceous material , graphite , or pah molecules @xcite , although a more complex mixture is possible @xcite .
nuv observations of the milky way extinction curve consistently find this feature ( see , e.g. * ? ? ? * ) .
in contrast , both lower metallicity regions , such as the small magellanic cloud @xcite , and starbursting galaxies @xcite , lack the nuv bump .
an extinction law is a function that takes into account the wavelength - dependent obscuration due to absorption and scattering processes as light traverses a distribution of dust particles .
one can therefore think of it as a foreground screen between the source and the observer .
such a scenario is a valid representation when observing stars in the milky way .
however , when we extend the analysis of dust to distant , mostly unresolved , galaxies , this process is complicated by the distribution of the dust , the variation of the composition within the observed region , and the underlying stellar populations .
it is common to refer to this integrated effect of dust in galaxies as attenuation .
galaxies with the same dust properties , hence with the same extinction , can yield different attenuation laws ( see , e.g. * ? ? ?
for instance , the absence of the nuv bump in the attenuation law of starbursting galaxies @xcite can either be caused by an intrinsic lack of the component that causes this bump , or by the superposition of several stellar populations with different dust geometries @xcite .
although it is beyond the scope of this paper to disentangle extinction and attenuation , one should bear in mind this subtlety in the interpretation of the observational constraints .
this work focuses on the dust component in the nearby starburst galaxy m82 ( ngc3034 ) located at a distance of @xmath103.3 - 3.5mpc @xcite , with a dynamical mass of @xmath11m@xmath12 @xcite . the ongoing strong star formation in this galaxy is believed to have been triggered by a close encounter with the more massive galaxy m81 , located in the same group .
uv light extends out from the central regions of the galaxy where star formation is the strongest , revealing the presence of dust entrained in the supernov - driven wind . in (
* hereafter hf14 ) we found that the dust obscuration in the nuv required the presence of an nuv bump , with the puzzling result that the prototypical starburst attenuation law of @xcite which features no nuv bump gave poorer fits than the standard milky way extinction of @xcite .
this paper extends this issue by probing in more detail the dust properties of the interstellar medium ( ism ) of m82 , as a function of galactocentric distance . literally , as the accepted hf14 paper went to press , an enthusiastic and diligently - led team of students at ucl s university of london observatory discovered sn2014j , a type ia supernova , within 1kpc of the central starburst of m82 @xcite .
a number of papers followed the evolution of the supernova ( see , e.g. , * ? ? ? * ; * ? ? ?
* ) , and , most relevant to this paper , the properties of the dust ( see , e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we probe here the dust properties in the ism of m82 pre - sn2014j , an issue of relevance to the understanding of the progenitor system of this supernova event .
lccr + observable & parameter & range & steps + age & log(t@xmath13/gyr ) & @xmath14 $ ] & 32 + metallicity & @xmath15 & @xmath16 $ ] & 8 + total to selective extinction ratio & @xmath0 & @xmath17 $ ] & 32 + nuv bump strength & b & @xmath18 $ ] & 32 + colour excess & e(b - v ) & @xmath19 $ ] & 32 + & 8,388,608 + the structure of the paper is as follows : the nuv - to - optical data used for the analysis are presented in
[ sec : data ] , followed by the modelling of the dust attenuation law in [ sec : dustlaw ] .
the methodology followed to constrain the dust parameters via stellar population synthesis is explained in
[ sec : method ] , with the results shown in [ sec : m82dust ] , along with comparisons with the recent constraints on the dust properties from sn2014j .
finally , [ sec : conc ] closes with the conclusions of our analysis .
we use the same dataset as in hf14 . in a nutshell ,
deep nuv images were retrieved from the heasarc archive of _ swift_/uvot .
the ultraviolet / optical telescope ( uvot , * ? ? ?
* ) is a 30 cm instrument working as a photon counter , with three passbands in the nuv region , straddling the position of the 2,175 bump .
we refer the reader to @xcite and @xcite for details of the properties , mode of operation and data reduction procedure of uvot imaging .
we retrieved imaging data of m82 with total exposure times of 10.2 , 13.9 and 8.7ks in the uvw2 , uvm2 , and uvw1 passbands , respectively .
the data were taken between 2008 and 2012 , therefore our photometry corresponds to the galaxy before the sn2014j event .
in addition to the uvot data , we retrieved optical images from the sdss dr8 server @xcite .
all images were registered to a common grid with a 0.5arcsec pixel size ( i.e. the uvot reference ) .
the optical ( sdss ) images were convolved to a common spatial resolution with a moffat profile .
we note that the effective resolution of the uvot images is @xmath20arcsec .
photometry is performed within 5arcsec ( radius ) apertures , following the standard calibration of the uvot camera @xcite .
the apertures were chosen to follow the disc of m82 ( see fig . 2 of hf14 ) .
we also retrieved the galex/@xmath21 images of m82 from the mast archive , creating new sets of photometric measurements by convolving the uvot data to the ( poorer ) resolution of galex colour ] ( @xmath22arcsec ) .
therefore , our combined dataset comprises photometry in the @xmath23 passbands .
we refer the interested reader to hf14 for further details about the dataset .
there are , however , two differences with respect to the hf14 data . here , we make use of an updated estimate of the foreground ( milky way ) reddening towards m82 . in hf14
we took the value derived by @xcite , as given in the ned database , @xmath24mag . following the suggestion of @xcite , updated by @xcite , we choose a lower estimate of @xmath25mag .
this correction is due to the fact that m82 affects the reading of the large scale dust maps used for the determination of foreground reddening .
this change affects the correction made to the observed photometry , and the derivation of the dust parameters , since in this paper , the amount of correction from the milky way which follows a @xmath26 @xcite law is significantly reduced .
nevertheless , we find that our results are compatible with hf14 .
in addition , we use 3.3mpc for the distance to m82 , for a more consistent comparison with recent papers ( e.g. * ? ? ? * ) .
this change only affects the translation from angular distances to projected physical distances .
the photometric apertures adopted in this paper map into a projected 160pc diameter .
we follow the parameterisation of the attenuation law of ( * ? ? ?
* hereafter csb10 ) .
this function depends on two basic parameters : @xmath0 ( defined in [ sec : intro ] ) , and @xmath1 , which controls the strength of the nuv bump .
[ fig : atten ] shows a few realizations of the csb10 function , along with the standard milky way attenuation law from @xcite , assuming @xmath27 .
notice that within the nuv spectral window ( left - hand panels ) , the uvot filters sample the bump , with uvm2 located in the middle of the feature and uvw2/uvw1 straddling it on the blue and red sides , respectively .
in addition , the galex / fuv filter includes an additional measurement blueward of the bump .
the case @xmath1=1 corresponds very nearly to the standard milky way extinction curve , by construction .
we follow a simple one - zone model for the dust component , assuming the light from the stellar populations is affected by a foreground dust screen
. other methods ( e.g. , * ? ? ?
* ) would result in a more complex set of parameters , whereas we focus here on an `` effective '' attenuation law integrated to all the light within the aperture ( see also * ? ? ? * ) . in hf14
, we compared the nuv and optical photometry of m82 with a range of stellar population models , and two typical attenuation curves , the above mentioned milky way standard of @xcite , and the law of @xcite usually applied to star - bursting systems such as m82 .
we found that the latter which lacks the presence of an nuv bump gave poorer fits to the nuv photometry with respect to a milky - way law . in this paper , we explore in more detail this issue by use of the parameterization of csb10 .
in addition to the nuv bump , the photometric data can be used to constrain the parameter @xmath0 , which depends on the composition of the dust grains . a standard milky way law assumes @xmath27 , although with significant scatter , roughly between 2.2 and 5.8 @xcite .
lower ( higher ) values of @xmath0 will require a distribution of smaller ( larger ) dust grains , with respect to the milky way standard . in hf14 , we studied the extraplanar light originated from the reflection of the stellar light in the galaxy , that impinged on the dust grains above and below the galactic plane .
these dust grains were entrained in the material ejected by the strong galactic winds triggered by the starburst . we fitted the photometric data assuming the absorption / scattering processes are described by a power law ( @xmath28 ) , finding a value ( @xmath29 ) .
note , however , that this case pertains to the scattering of light from dust grains in the extraplanar wind region , as they are illuminated by the galaxy .
in contrast , we focus in this paper on the efective attenuation law in the main body of the galaxy . the right - hand panels of fig .
[ fig : atten ] show a wide nuv / optical spectral window , focusing on variations of @xmath0 .
note how higher values of @xmath0 as expected in an ism with larger dust grains result in a flatter attenuation law , closer to the mie regime .
therefore , the use of a wide spectral range ( nuv and optical photometry ) will allow us to constrain this parameter .
in order to constrain the dust attenuation parameters , we need to quantify the effect of the other , `` nuisance '' , parameters related to the underlying stellar population .
our methodology involves a comparison of the observed nuv and optical photometry with a large grid of synthetic populations from the models of @xcite . tab .
[ tab : params ] shows the grid of 8,388,608 models used in this paper .
they correspond to simple stellar populations ( ssp , representing a population with a single age and metallicity ) for a standard initial mass function @xcite . within our wavelength interval
, dust simply introduces a smooth , long ( spectral ) range behaviour ( aside from the nuv bump ) .
therefore it is possible to factor out the attenuation law from the different trends introduced in the photometry by the age and metallicity distribution of the populations . to prove this point , we ran three sets of simulations , each one with a different assumption about the star formation history : i ) ssp ( as above ) ; ii ) exp : a composite of stellar populations following an exponentially decaying star formation rate , choosing the formation epoch , decay timescale and metallicity as free parameters of the populations ; iii ) 2ssp , a superposition of two ssps , where the age of the young and old components , the fractional contribution , by mass , of the young component , and the metallicity are free parameters .
these three choices encompass a wide range of both smooth and bursty star formation histories .
this exercise serves to understand whether systematic biases are being introduced because of a methodology where the analysis is based on a grid of ssps .
the derivation of the model parameters follows a bayesian approach , using a @xmath30-based likelihood as a probability distribution function ( @xmath31 ) from which we derive the errors as the 68% confidence level of this distribution .
@xmath32 is defined as @xmath33 , and @xmath30 is given by : @xmath34 ^ 2}{\sigma^2(m_j)},\ ] ] where @xmath35 represents all possible choices of the five parameters listed in tab .
[ tab : params ] .
@xmath36 and @xmath37 are the observed and model magnitudes , respectively , as follows : \{fuv , uvw2 , uvm2 , uvw1 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42}. finally , @xmath43 is the observational errors for each magnitude .
in addition to the parameters listed in tab . [ tab : params ] , we include a `` normalization '' parameter that assumes the model sdss @xmath39 band magnitude lies somewhere between @xmath44mag ; where @xmath45 is the observed magnitude in the @xmath39 band for each aperture .
the extracted parameters are then computed using the likelihood as a probability distribution function , namely : @xmath46 the priors applied in the analysis are simply the range of parameters explored , as shown in tab .
[ tab : params ] ( i.e. one can consider them as `` flat priors '' , giving equal probability within the range stated ) . + fig . [
fig : sims ] shows the results for 3@xmath471,000 simulations corresponding to each of the three formation histories .
the simulated data explore a wide range of ages , metallicities , and dust properties , comparable to those found in m82 ( see fig . 5 of hf14 ) .
our analysis marginalizes over the stellar population properties , and produces the extinction - related constraints on from left to right
@xmath0 , @xmath1 and @xmath48 .
the first two parameters are intrinsic to the effective dust attenuation law , whereas the latter can be considered as a normalization factor , reflecting the amount of dust present .
the panels on the left compare input value ( abscissa ) versus retrieved value ( ordinate ) .
an orange line in each panel gives the 1:1 correspondence .
the grey dots are results from individual estimates , whereas the black dots and error bars mark the median and root - mean square values , respectively , when binned at fixed number of data points per bin .
note that no large bias is obtained , regardless of the use of different star formation histories ( top to bottom ) .
the derivation of the nuv bump strength only uses the uv photometry for the definition of the likelihood , since the optical photometry is fully insensitive to the presence of this bump ( fig .
[ fig : atten ] right - hand panel ) , and would therefore `` wash - out '' the constraining power in the comparison .
we use the comparison with the simulated data to derive a mean linear correction between input and output parameters .
the results of the fits are shown in tab .
[ tab : corr ] .
since one can not determine the best star formation history corresponding to the real data , we decided to choose for the general correction the one that gave the best compromise among the three sets of star formation histories , i.e. we used the results from all three sets of simulations for the fit .
the right - hand panels of fig .
[ fig : sims ] illustrate the validity of this correction and the retrieved error bars .
they compare the difference between the `` true '' and the extracted value ( `` input''@xmath49``output '' in the figure ) .
this comparison allows us to determine whether the quoted error bars are valid .
the shaded regions correspond to the original data , whereas the blue lines with error bars give the results for the corrected parameters .
note that the corrections are relatively minor with respect to the error bar , and no significant bias is apparent when the corrections are taken into account .
nevertheless , we will show in the next figures our results with and without this correction .
the comparison with simulated data therefore confirms that it is possible to factor out the attenuation law from the properties of the underlying , unresolved stellar populations , by use of a large library of simple stellar populations . as a function of projected galactocentric distance .
the grey dots represent individual measurements , with a characteristic error bar shown on the left - hand side of the figure .
the red stars are measurements within 30arcsec ( i.e. 480pc ) of the position of sn2014j .
the blue dots and error bars are the median and root mean square values of the data , binned at a fixed number of data points per bin .
the black line shows the same result if the retrieved parameters are not calibrated with respect to the simulations ( see tab . [
tab : corr ] ) .
the orange cross and error bar correspond to the measurement in the region of sn2014j after the event @xcite , width=321 ] we also need to assess a possible bias between the dust related parameters ( @xmath0,@xmath1,@xmath48 ) and the underlying stellar populations . in fig .
[ fig : simstz ] we compare the bias @xmath50 where @xmath51 represents any of the three dust - related parameters as a function of the retrieved age and metallicity .
note that the input parameters are , by construction , uncorrelated .
the figure shows that most of the comparisons give negligible dependence on age and metallicity , or on the functional form of the star formation history .
this result is very important for the interpretation of the radial gradients of the effective attenuation law presented in the next section .
we also note that the values we retrieve for the ssp ages are consistent with @xcite , where only two standard extinction laws were considered .
these results are also consistent with independent estimates , giving ages around 0.41gyr within the central 3kpc , with older ages in the outer regions ; and a colour excess @xmath52 with a significant increase in the central regions ( see , e.g. * ? ? ?
* ; * ? ? ?
as regards to metallicity , we emphasize that broadband photometric analyses have large degeneracies with respect to this parameter .
our simulations in fact show that of all five parameters given in tab . [
tab : params ] , metallicity is the only one that fails at retrieving the input values , within a 0.5dex uncertainty .
however , we emphasize that such a result does not prevent us from deriving unbiased estimates of the other parameters ( see fig . [
fig : sims ] ) , and consider age and metallicity simply as nuisance parameters in the derivation of the dust - related properties .
in addition , we have done tests , imposing the typical metallicities expected in m82 ( e.g. * ? ? ? * ) as priors , and the extracted values of @xmath0 , @xmath1 , and @xmath48 are unaffected .
a partial version of the parameters derived from this analysis are shown in tab .
[ tab : results ] .
the complete version can be retrieved online .
finally , in order to assess the possible systematics caused by the specific population synthesis model used , we create an additional set of 1,000 simulations with the models of @xcite .
we note that these models apply widely different prescriptions , including a different treatment of the thermally - pulsing agb phase , which can be prominent in populations of ages @xmath53gyr , therefore relevant for the study of m82 .
we emphasize here that the aim in this paper is _ not _ to constrain the age or metallicity composition of the populations , but to factor out the degeneracies caused by these parameters . in order to show that our results are robust against different prescriptions for the population synthesis models , we treat the simulated data
which use the @xcite models , with the same grid of ssps which are based on the @xcite models .
[ fig : cm05_comp ] shows the histograms of the difference between the `` true '' ( i.e. the assumed value ) parameter , and the retrieved one , in both cases ( bc03 : for the simulated data based on @xcite models ; cm05 for the data based on the @xcite models ) .
we conclude that our derivation of the dust - related parameters is robust . ,
are plotted with respect to the projected galactocentric distance .
the symbols are the same as in fig .
[ fig : m82_rv].,width=321 ]
[ fig : maps ] shows the spatial distribution of the dust - related parameters constrained with the photometry .
the results are shown as colour - maps .
the missing apertures in the distribution are masked out because of the presence of foreground stars .
the bottom - right panel includes an sdss-@xmath39 band image of the galaxy , shown at the same scale .
the radial profile of the constraints on the dust properties of m82 are shown in fig .
[ fig : m82_rv ] ( @xmath0 ) ; fig .
[ fig : m82_b ] ( nuv bump strength ) ; and fig .
[ fig : m82_ebv ] ( colour excess ) . in all figures
the grey dots correspond to individual measurements , with a typical error bar shown on the left - hand side . the median and root - mean - square values of the data , binned at fixed number of data points per bin ,
are shown as black solid dots with error bars .
finally , the solid line without error bars show the result if the calibration presented in tab . [
tab : corr ] is not applied . note the significant trend with galacto - centric distance .
the central region of m82 which due to projection effects include _ both _ the starbursting region and the outer areas has a more standard attenuation law , compatible wih the @xmath26 and @xmath54 values corresponding to the @xcite case . as we probe further out , the total - to - selective extinction ratio parameter , @xmath0 , decreases , and so does the nuv bump strength , although we never find measurements compatible with a weak or non - existent bump .
the colour excess also depends on galacto - centric distance , although in this case the correlation is more tightly linked to the variation in the metallicity and star formation distribution .
we note that this result is robust against a possible bias caused by the inherent radial trend in age and metallicity found in galaxies .
the previous section explored through simulations the possibility of a bias in the derivation of the parameters related to the dust composition , to find no significant systematic trends with age , metallicity or the different functional form of the star formation history .
therefore , these trends are intrinsic radial variations of the effective dust attenuation law ( @xmath0,@xmath1 ) , and the amount of dust ( @xmath48 ) . , are shown against galactocentric distance .
the symbols are the same as in fig .
[ fig : m82_rv].,width=321 ] fig .
[ fig : m82_paramcorr ] shows the correlations among the three dust - related parameters ( using the same symbols as in the previous figures ) .
the range of values found _ within _
m82 is comparable with the average estimates of @xcite for a general sample of galaxies ( @xmath55 ; @xmath56 ) .
our values are also consistent with the sample of nearby galaxies of @xcite , who find a range of values for the nuv bump strength , weaker than milky way standard , in the range to the @xmath1 parameterisation of @xcite by a simple comparison of the effective attenuation laws . ]
note the significant trend between @xmath0 and @xmath1 ( top - left panel of fig .
[ fig : m82_paramcorr ] ) , suggestive of a strong link between the overall dust grain size distribution ( larger grains causing higher values of @xmath0 ) and the carrier of the nuv bump strength .
although changes in the attenuation law can be produced by changes in the age - dependent extinction of the geometry of the stellar / dust components , our study based on a single galaxy
would suggest that most of the variations are caused by the intrinsic extinction law .
a similar trend is found between @xmath1 and colour reddening , although in this case the data mostly show that regions with high reddening ( @xmath58 ) have high nuv bump strengths ( whereas less dusty regions can have a wide range of @xmath1 , including @xmath54 ) . in a similar way ,
regions with high reddening are associated with higher values of @xmath0 ( top - right panel ) , corresponding to the presence of larger dust grains .
lccc + @xmath51 & @xmath59 & @xmath60 & rms + @xmath0 & @xmath61 & @xmath62 & @xmath63 + @xmath1 & @xmath64 & @xmath65 & @xmath66 + @xmath48 & @xmath67 & @xmath68 & @xmath69 + [ cols="^,^,^,^,^,^,^ " , ] this is an excerpt from the central region .
the full version of this table can be retrieved online .
our data are especially relevant to explore the region around the recent sn2014j event .
we use the position of the supernova from @xcite to locate the relevant apertures in our dataset . in fig .
[ fig : m82_rv ] , [ fig : m82_b ] , [ fig : m82_ebv ] , we show with red star symbols the estimates of @xmath0,@xmath1,@xmath48 , respectively , measured in the apertures located within 30arcsec of the sn2014j event , corresponding to a projected distance of 480pc . notice that those regions feature higher values of @xmath0(@xmath103 - 4 ) than the recent estimates determined post - sn2014j ( @xmath70 , * ? ? ?
* orange cross ) .
in general , photometric studies of the environment of type ia supernova favour low values of @xmath0 @xcite , however the dispersion is quite large @xcite . for this particular case , even within the error bars and the dispersion of the data , it is not possible to assume that post - sn2014j the dust properties in this region are compatible .
therefore , our results indicate that pre - sn2014j , this region had a dust content in the ism similar to the milky way standard .
such a result implies the post - sn2014j observations should be explained either by a combination both from dust in the ism surrounding the supernova and the intrinsic contribution from a circumstellar disk ( @xcite , although see @xcite ) , or by a significant change in the ism towards the observed region .
alternatively , one could also expect that supernovae have intrinsic variations in colour that are not accounted for , or that the distribution of dust within the line of sight towards the supernova can produce the mismatch .
we note that the trends presented in this paper are at odds with some recent results targeting our galaxy and m31 . in @xcite and @xcite ,
the dust attenuation properties of the milky way are probed towards the galactic bulge , finding a steeper law
i.e. lower @xmath0 .
in addition , @xcite find similar trends in the andromeda galaxy .
we note that the populations towards the centre of m82 are significantly different from these two galaxies .
the presence of a strong starburst will affect the properties of the ism , and the dust .
in addition , the nearly edge - on orientation of m82 along with its dusty ism results in a complex interpretation of the results towards small _ projected _ galactocentric radii .
another interesting issue is the strong correlation between @xmath0 and bump strength ( top - left panel of fig .
[ fig : m82_paramcorr ] ) . a recent study of the nuv bump strength in high redshift galaxies
@xcite suggests the opposite trend , namely a _ decreasing _ bump strength with increasing @xmath0 .
in contrast , detailed studies in our galaxy do not seem to find a correlation @xcite , so there is still no clear indication whether such a correlation should be expected in general .
we emphasize here that our methodology is unlikely to produce a spurious correlation between these two parameters .
the simulations show that the ( uncorrelated ) dust parameters are retrieved without any significant bias or artificial correlation among parameters .
in addition , we note that our study mainly differs from @xcite in the sense that we explore _ internal _ variations within the same galaxy , rather than trends among galaxies .
we revisit the nuv and optical photometry of m82 , presented in hf14 , to constrain in more detail the properties of the dust attenuation law . through the analysis of simulated data ,
we prove that it is possible to factor out the contribution from the underlying stellar populations , setting constraints on the nuv bump strength and the total - to - selective attenuation ratio @xmath0 .
we find a significant gradient with projected galactocentric distance ( fig .
[ fig : m82_rv ] ) , with the central regions showing a standard milky way attenuation law , progressing outwards towards an extinction with a weaker nuv bump and steeper @xmath0 , more characteristic of a smaller grain size distribution , effectively tending towards the standard attenuation law of the smc @xcite .
we should note here that our results correspond to a one - zone model , where the dust only contributes as a foreground screen .
changes in the geometry of the dust distribution can therefore cause variations on the dust parameters presented here ( see , e.g. , * ? ? ?
.,width=321 ] as regards to the nuv bump strength , we find a systematic decrease outwards .
the weaker bump in the outer regions can be indicative of changes in the dust properties @xcite , although an age - dependent attenuation law can produce similar effects @xcite .
however , our observational constraints go in the opposite way with respect to the models .
a trend towards a weaker bump is found at large galacto - centric distances , _ away _ from the central starburst , where the dust content is lower , and where @xmath0 is low , indicative of smaller dust grains ( see fig . [
fig : m82_paramcorr ] ) .
our results confirm hf14 in the sense that a milky way law , with a significant nuv bump is favoured with respect to a @xcite law , in the central starburst .
our work is most relevant to the analysis of the environment of the recent type ia supernova sn2014j .
our data were taken before the event , and therefore probe the properties of the dust of the ism . at the position of sn2014j , we find a standard dust component , typical of the milky way , with @xmath71 and @xmath72 .
therefore , the recent measurements with lower values of @xmath0 reflect either the presence of a circumstellar component , or a significant change in the dust properties within the volume probed by the post - sn2014j measurements .
we would like to thank the anonymous referee for very valuable comments and suggestions .
this research has made use of data and software provided by the high energy astrophysics science archive research center ( heasarc ) , which is a service of the astrophysics science division at nasa / gsfc and the high energy astrophysics division of the smithsonian astrophysical observatory .
we acknowledge use of imaging data from the sloan digital sky survey ( sdss ) .
funding for sdss - iii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , and the u.s .
department of energy office of science .
the sdss - iii web site is http://www.sdss3.org/. we also thank the mikulski archive for space telescopes ( mast ) .
stsci is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
support for mast for non - hst data is provided by the nasa office of space science via grant nnx13ac07 g and by other grants and contracts . | we use near ultraviolet and optical photometry to investigate the dust properties in the nearby starburst galaxy m82 . by combining imaging from the swift / uvot instrument and optical data from the sloan digital sky survey
, we derive the extinction curve parameterized by the standard @xmath0 factor , and the strength of the nuv 2175 feature quantified by a parameter @xmath1 out to projected galactocentric distances of 4kpc .
our analysis is robust against possible degeneracies from the properties of the underlying stellar populations .
both @xmath1 and @xmath0 correlate with galactocentric distance , revealing a systematic trend of the dust properties .
our results confirm previous findings that dust in m82 is better fit by a milky way standard extinction curve ( hutton et al . ) , in contrast to a calzetti law .
we find a strong correlation between @xmath0 and @xmath1 , towards a stronger nuv bump in regions with higher @xmath0 , possibly reflecting a distribution with larger dust grain sizes .
the data we use were taken before sn2014j , and therefore can be used to probe the properties of the interstellar medium before the event .
our @xmath0 values around the position of the supernova are significantly higher than recent measurements post - sn2014j ( @xmath2 ) .
this result is consistent with a significant change in the dust properties after the supernova event , either from disruption of large grains or from the contribution by an intrinsic circumstellar component .
intrinsic variations among supernovae not accounted for could also give rise to this mismatch .
= 10000 [ firstpage ] galaxies : individual : m82 galaxies : starburst galaxies : stellar content galaxies : ism ism : dust , extinction supernovae : individual : sn 2014j |
according to steinitz paper @xcite from 1906 , the cone spanned by the @xmath3-vectors of @xmath6-dimensional polytopes ( with apex at the @xmath3-vector of the @xmath6-simplex ) is given by @xmath7 in conjunction with euler s formula , @xmath8 ; see also ( * ? ? ?
moreover , the cone of flag vectors of @xmath6-polytopes is also given by these two inequalities , together with the usual dehn
sommerville relations , such as @xmath9 and @xmath10 . here
the extreme cases , polytopes whose @xmath3- or flag vectors lie at the boundary of the cone , are given by the simplicial polytopes ( for which steinitz first inequality is tight ) and the simple polytopes ( second inequality tight ) .
moreover , _ all _ the integer points that satisfy the above conditions do correspond to actual polytopes . for @xmath1-dimensional polytopes ,
such a complete and simple answer is not to be expected ; compare grnbaum ( * ? ? ?
10.4 ) , bayer @xcite and hppner & ziegler @xcite
. however , we may now be getting close to a full description of the convex cone spanned by the @xmath3-vectors ( with apex at the @xmath3-vector of the @xmath1-simplex ) : in the projective coordinates introduced in @xcite , @xmath11 we can write the known necessary conditions as @xmath12 where the first two conditions are trivial , the second two have simplicial resp . simple polytopes as extreme cases , and the last condition is a lower bound that follows from `` @xmath13 . '' this is indeed a complete description of the cone if and only if there are polytopes for which the sum @xmath14 is arbitrarily small , that is , for which the _ fatness _ parameter @xmath15 is arbitrarily large @xcite .
this observation has sparked a certain race for `` fat '' @xmath1-dimensional polytopes .
the following table summarizes the main steps .
most of the examples that appear there are 2-simple and 2-simplicial , with a symmetric @xmath3-vector ; the first infinite family of such polytopes was constructed by eppstein , kuperberg & ziegler @xcite ; a simple construction appears in paffenholz & ziegler @xcite .
we call a polytope `` even '' if its graph is bipartite . @xmath16 } & 5.048 & 2-simple , 2-simplicial & eppstein et al.\ \cite{z80 } & 2003\\ $ e(c_m\times c_n)$ & $ 6-{\varepsilon}$ & 2-simple , 2-simplicial & paffenholz \cite{paffenholz - pc } & 2004\\ projected products & $ 9-{\varepsilon}$ & even & here & \\ \hline \end{tabular}\ ] ] a flag vector parameter that is similar to fatness , called _ complexity _ , was also introduced in @xcite .
this is an invariant of the flag vector , defined by @xmath17 where @xmath18 and @xmath19 are components of the toric @xmath20-vector of the polytope , while @xmath21 and @xmath22 refer to @xmath20-entries of the dual polytope ; compare @xcite and @xcite .
all @xmath1-polytopes satisfy @xmath23 .
fatness and complexity are roughly within a factor of @xmath0 : @xmath24 and @xmath25 . in particular , it is not known whether @xmath26 can be arbitrarily large . previously , the polytopes with the largest known complexity were the `` neighborly cubical polytopes '' of joswig & ziegler @xcite , of complexity @xmath27 . our present construction yields `` neighborly cubical polytopes '' in the special case @xmath28 ( possibly special cases of those of @xcite , with a simpler description ) , but for @xmath29 it yields complexity values as large as @xmath30 .
the concept of `` strictly preserving a face '' used in the following theorem will be explained in section [ sec : projections ] .
( compare the concept of faces in the `` shadow boundary '' of a projection , e.g. in @xcite . )
[ thm : ppp ] let @xmath31 be even and @xmath32 .
then there is a @xmath33-polytope @xmath34 , combinatorially equivalent to a product of @xmath35 @xmath36-gons , @xmath37 , such that the projection @xmath38 to the last four coordinates preserves the @xmath39-skeleton as well as all the `` @xmath36-gon @xmath0-faces '' of @xmath40 . in the following ,
we describe in brief the main ingredients for the construction and sketch the proof for its correctness .
detailed proofs , the complete combinatorial characterization of the resulting polytopes , possible extensions , further remarkable aspects ( such as the polyhedral surfaces of high genus embedded in the @xmath0-skeleta of the resulting @xmath1-polytopes ; cf .
@xcite ) as well as necessity of the restrictions imposed here ( e.g. , that @xmath36 must be even ) are topics of current research and will be presented later . the intuition for the construction given here grew from previous joint work and current discussions with nina amenta , michael joswig , and thilo schrder .
the combinatorial structure of the products of polygons @xmath41 is easy to describe : these are simple @xmath33-polytopes , with @xmath42 vertices , @xmath43 edges , and @xmath44 facets . in general , its non - empty faces are products of non - empty faces of the polygons , so @xmath45 the @xmath0-dimensional faces of @xmath41 , and thus of any polytope combinatorially equivalent to @xmath41 , may be split into two classes : there are @xmath46 faces that are @xmath36-gons , which we refer to _ polygons _ , which arise as products of one of the @xmath36-gons with a vertex from each of the other factors ; and there are @xmath47 _ quadrilaterals _ that ( in @xmath41 ) arise as products of edges from two of the factors with vertices from the others .
thus , in total @xmath41 has @xmath48 @xmath0-faces . in the case
@xmath28 , the polygon @xmath0-faces of @xmath41 are @xmath1-gons , but we nevertheless treat the @xmath49 polygons and the @xmath50 quadrilaterals separately in this case . an inequality description for such a product polytope may be obtained as @xmath51 assuming that @xmath52 is a correct description for an @xmath36-gon : for this it is necessary and sufficient that the row vectors @xmath53 of @xmath54 are non - zero and distinct and that they positively span @xmath55 , that the components @xmath56 of @xmath57 are positive , and that the rescaled vectors @xmath58 are in convex position ( the vertices of the polar of the polygon ) . for this
we say that a finite set of vectors @xmath59 _ positively spans _ if it satisfies the following equivalent conditions : every vector @xmath60 is a linear combination of the vectors @xmath53 , with non - negative coefficients , every vector @xmath60 is a linear combination of the vectors @xmath53 , with positive coefficients , the vectors @xmath53 span @xmath61 , and @xmath62 is a linear combination of the vectors @xmath53 , with positive coefficients ( that is , the vectors @xmath53 are _ positively dependent _ ) . in the following , we will need `` deformed products '' ( as described in amenta & ziegler @xcite ) of polygons . for this
, we look at systems of the form @xmath63 given any such left - hand side matrix for such a system , we can adapt the right - hand side so that the resulting polytope is _ combinatorially equivalent _ to @xmath41 : for this all components of @xmath64 have to be sufficiently large compared to @xmath65 , for @xmath66 .
( compare @xcite and @xcite . )
we work with a rather restrictive concept of faces `` being preserved under projection . ''
[ def : strictly]let @xmath67 be a projection of polytopes . then a face @xmath68 is _ strictly preserved _
if its image @xmath69 is a face of @xmath70 , the map @xmath71 is a bijection , and the pre - image of the image is @xmath72 , that is , @xmath73 . in the definition ,
conditions ( ii ) and ( iii ) are both needed .
indeed , in the projection `` to the second coordinate '' displayed in our figure , the vertex @xmath74 is strictly preserved , but the vertex @xmath75 and the edge @xmath76 are not : for @xmath75 condition ( iii ) fails , while for @xmath76 condition ( ii ) is violated .
@xmath77 for simplicity , the following characterization result is given only in a coordinatized version , for the projection `` to the last @xmath78 coordinates . ''
we say that a vector @xmath79 _ defines _ the face @xmath68 given by all the points of @xmath80 that have maximal scalar product with @xmath79 .
this describes exactly all the vectors in the relative interior of the _ normal cone _ of @xmath72 .
if @xmath80 is full - dimensional , this interior of the normal cone consists of all the positive combinations of outer facet normals @xmath81 to the facets @xmath82 that contain @xmath72 .
( compare ( * ? ? ?
* sects . 2.1 , 3.2 , 7.1 ) . )
[ prop : preserve]let @xmath83 , @xmath84 be the projection to the last @xmath78 coordinates , and let @xmath85 be an @xmath86-dimensional polytope , and let @xmath72 be a face of @xmath80 .
then the following three conditions are equivalent : 1 .
@xmath72 is strictly preserved by the projection @xmath87 .
any @xmath88 arises as the first @xmath76 components of a vector @xmath89 that defines @xmath72 .
the vectors @xmath90 , given by the first @xmath76 components of the normal vectors @xmath91 to facets @xmath92 of @xmath80 that contain @xmath72 , positively span @xmath93 . here
we only establish `` @xmath94 , '' which is used in the following .
if the vectors @xmath90 are positively dependent , then some positive combination of the vectors @xmath91 yields @xmath95 .
a point @xmath96 lies in the face @xmath68 if and only if its scalar product with each facet normal @xmath97 is maximal .
this happens if and only if @xmath98 is maximal , that is , iff @xmath99 is maximal under the restriction @xmath100 .
thus we have established that under the assumption ( 3 ) , @xmath101 is a face of @xmath102 , and @xmath103 ; that is , conditions ( i ) and ( iii ) of definition [ def : strictly ] are satisfied . since the vectors @xmath104 additionally are positively spanning , we know that if @xmath105 , then @xmath106 for @xmath107 : every such @xmath108 arises as a positive combination of the vectors @xmath104 , and since the scalar product with the corresponding combination of the @xmath109 is maximized over @xmath80 at @xmath110 , a scalar product with @xmath111 will be larger than the maximum over @xmath80 .
this implies that the projection @xmath112 is injective .
in particular @xmath71 is injective , and this establishes part ( ii ) of definition [ def : strictly ] .
for @xmath31 even and @xmath32 , let @xmath40 be defined by the linear inequality system @xmath113 here the left - hand side coefficient matrix @xmath114 contains blocks of size @xmath115 , where @xmath116 with @xmath117 the block @xmath118 arises from @xmath54 by an @xmath119-perturbation : @xmath120 for a sufficiently small @xmath121 .
all entries of @xmath122 outside the @xmath123 blocks of types @xmath118 , @xmath124 and @xmath125 are zero .
let the right - hand side vector be such that @xmath126 is given by @xmath127 for even @xmath128 , and @xmath129 for odd @xmath128 , and by @xmath130 for sufficiently large @xmath131
. then @xmath40 has the properties claimed by theorem [ thm : ppp ] .
in particular , it is a deformed product of @xmath35 @xmath36-gons , and all its polygon @xmath0-faces survive the projection to the last @xmath1 coordinates .
the rows @xmath132 of @xmath118 are indeed in cyclic order : @xmath133 moreover , rescaled as @xmath134 for odd @xmath128 and as @xmath135 for even @xmath128 they are in convex position , if @xmath119 is small ; so @xmath136 defines a convex @xmath36-gon .
thus for sufficiently small @xmath119 and sufficiently large @xmath131 , the polytope @xmath137 is indeed a deformed product of polygons , as discussed in section [ sec : products ] .
now we show that for sufficiently small @xmath119 , all the polygon @xmath0-faces of @xmath40 survive the projection to the last @xmath1 coordinates . for this
, we verify that the left - hand side matrix with @xmath54-blocks instead of @xmath118-blocks , which we denote by @xmath138 , satisfies the linear algebra condition dictated by proposition [ prop : preserve](3 ) .
this is sufficient , since the `` positively spanning '' condition is stable under perturbation by a small @xmath119 .
any polygon @xmath0-face @xmath72 of the simple @xmath33-polytope @xmath139 is defined by the facet normals to the @xmath140 facets that contain @xmath72 .
the facet normals correspond to the rows of the inequality system , and thus for the facet normals of a polygon @xmath0-face one has to choose two cyclically adjacent rows from each block ( corresponding to a vertex from each factor polygon ) , except from one of the blocks no row is taken .
moreover , due to the structure of the matrices @xmath125 , @xmath54 , and @xmath124 , in which rows alternate , any choice of two cyclically - adjacent rows from a block yields the same pair of rows ( only the order is not clear , but it also does not matter ) .
thus , to apply proposition [ prop : preserve](3 ) we have to show : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ if one of the @xmath35 pairs of rows is deleted from the reduced matrix @xmath141 then the remaining @xmath140 rows + ( a ) span @xmath142 , and + ( b ) have a linear dependence with strictly positive coefficients . _
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ let us establish ( b ) first . for this , let @xmath143 these sequences are designed to be non - negative , @xmath144 for all @xmath145 , with equality only for @xmath146 .
thus for ( b ) it suffices to verify _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for any @xmath147 , the rows of @xmath148 are positively dependent with coefficients @xmath149 for the even - index row from the @xmath150-th block , and @xmath151 for the odd - index row from the @xmath150-th block . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ since the ( two ) vectors in the @xmath150-th block thus get zero coefficients , so they may be deleted from any linear dependence ( with otherwise positive coefficients ) .
thus we are led to the condition @xmath152 which is needed to hold for @xmath153 , but which we impose for all @xmath145 .
the choice of vectors @xmath154 is designed to satisfy this condition .
indeed , except for the choice of a basis , which we took to be @xmath155 and @xmath156 , the configuration of five vectors @xmath154 are uniquely determined by the condition . for property ( a ) , we have to show that if one of the @xmath35 pairs of rows is deleted from the matrix @xmath148 , then the resulting matrix still has full rank .
if the first or the second pair of rows is deleted , then we still have the last @xmath157 rows , and they form a block upper triangular matrix , which has full rank since its diagonal block @xmath158 is non - singular . if a later pair of rows is deleted , then we are faced with the task to show that the @xmath159 matrices @xmath160 of the form @xmath161 are non - singular . to verify this ( without proving explicitly that @xmath162 , which might need combinatorial ingenuity ) we use our knowledge about row combinations of @xmath160 .
indeed , if we sum the rows of @xmath160 with coefficients @xmath163 , then this will result in the linear combination of the three rows of the matrix @xmath164 with the coefficients @xmath165 , since @xmath166 .
similarly , if we sum the rows of @xmath160 with coefficients @xmath167 , then we get a linear combination of the same three rows , with coefficients @xmath168 . and
if we use coefficients @xmath169 to sum the rows of @xmath160 , then the result will be a sum with coefficients @xmath170 .
the coefficient matrix @xmath171 is non - singular for @xmath172 : its determinant is @xmath173 .
thus the full row - space of @xmath174 is contained in the row space of @xmath160 . in particular , we find the unit vectors @xmath175 in the row space of @xmath174 , and thus of @xmath160 , and this allows us to complete the argument by induction .
[ cor : flag ] the @xmath1-polytope @xmath176 has the flag vector @xmath177 we obtain @xmath42 and @xmath43 from the products @xmath41 , which are simple @xmath33-polytopes with @xmath178 vertices . with the abbreviation @xmath179 this yields @xmath180 vertices and @xmath181 edges for @xmath176 .
the projected polytope has two types of facets : there are `` prism '' facets , which involve two of the polygons , as well as `` cube '' facets , which in @xmath41 arise as products of three edges and @xmath183 vertices , but contain no polygon @xmath0-faces .
thus each prism facet is bounded by two polygons , and each polygon lies in two prism facets .
hence there are @xmath184 prism facets , as well as some number @xmath185 of cube facets . , _ deformed products and maximal shadows _ , in advances in discrete and computational geometry ( south hadley , ma , 1996 ) , b. chazelle , j. e. goodman , and r. pollack , eds . , vol . 223 of contemporary mathematics , providence ri , 1998 , amer .
soc . , pp .
5790 . , _ fat @xmath1-polytopes and fatter @xmath6-spheres _ , in discrete geometry : in honor of w. kuperberg s 60th birthday , a. bezdek , ed . ,
253 of pure and applied mathematics , marcel dekker inc .
, new york , 2003 , pp .
arxiv : math.co/0204007 . ,
_ theorie der vielfachen kontinuitt _ , denkschriften der schweizerischen naturforschenden gesellschaft , vol .
38 , pp . 1237 , zrcher und furrer , zrich , 1901 .
written 1850 - 1852 ; reprinted in : ludwig schlfli , 1814 - 1895 , gesammelte mathematische abhandlungen vol .
i , birkhuser , basel 1950 , pp . 167387 . ,
_ generalized @xmath193-vectors , intersection cohomology of toric varieties , and related results _ , in commutative algebra and combinatorics , m. nagata and h. matsumura , eds .
11 of advanced studies in pure mathematics , kinokuniya , tokyo , 1987 , pp .
187213 . , _ face numbers of @xmath1-polytopes and @xmath6-spheres _ , in proceedings of the international congress of mathematicians ( icm 2002 , beijing ) , l. tatsien , ed . ,
iii , beijing , china , 2002 , higher education press , pp .
arxiv : math.mg/0208073 . | we construct a @xmath0-parameter family of @xmath1-dimensional polytopes @xmath2 with extreme combinatorial structure : in this family , the `` fatness '' of the @xmath3-vector gets arbitrarily close to @xmath4 , the `` complexity '' ( given by the flag vector ) gets arbitrarily close to @xmath5 .
the polytopes are obtained from suitable deformed products of even polygons by a projection to four - space . |
the virmos ( visible and infrared multi - object spectrograph ) project consists of two spectrographs with enhanced survey capabilities to be installed on two unit telescopes of eso very large telescope ( chile ) : vimos ( 0.37 - 1 @xmath1 ) and nirmos ( 0.9 - 1.85 @xmath1 ) , each one having a large field of view ( 14x16 ) split into 4 quadrants and a high multiplexing factor ( up to approximately 800 spectra per exposure ) . to easily exploit such a potential a dedicated tool ,
the virmos mask preparation software ( mps ) , has been implemented .
it provides the astronomer with tools for the selection of the objects to be spectroscopically observed , and for automatic slit positioning .
the output of mps is used to build the slit masks to be mounted in the instrument for the spectroscopic observations .
at a limiting magnitude @xmath2 , the density of objects in the sky is such that more than 1000 galaxies are visible in a vimos quadrant . of course , not all these objects can be spectroscopically observed , as some requirements imposed by data quality have to be taken into account when placing slits : the minimum slit length will depend on the object size , since the slit must contain some area of `` pure sky '' to allow for a reliable sky subtraction ; spectra must not overlap either along the dispersion or the spatial direction ; as each first order spectrum is coupled with a second order spectrum which will contaminate the first order spectrum of the slit above , a good sky subtraction can be performed only if slits are aligned in columns ( same spatial coordinate ) and , within the same column , have the same length .
all these factor lead to a theoretical maximum number of spectra per quadrant of approximately 200 .
another requirement for mps is set by the very good vlt seeing which allows the use of slits widths of 0.3 - 0.4 arcsec .
such narrow slits imply an extremely precise slit positioning , with maximum uncertainties of the order of 0.1 arcsec .
thus the need for some ( 1 - 2 per quadrant ) manually selected reference objects ( possibly bright and point - like ) to be used for mask alignment .
moreover , the user must have the possibility to manually choose some particularly interesting sources to be included ( compulsory objects ) and some others to be excluded ( forbidden objects ) from the spectroscopic sample .
also a tool for manual definition of curved or tilted slits , to better follow the shape of particularly interesting objects , has to be provided . the mps works starting from a vimos image , to which a catalogue of objects is associated .
the catalogue can be derived from the image itself or from some other astronomical data - set . in this second case ,
a way to correlate the celestial coordinates of the objects in the catalogue with image coordinates is to be provided .
some catalog handling capabilities , to allow for the selection of classes of sources among which to operate the choice of spectroscopic targets , some image display and catalogue overlay capabilities have to be provided by the package .
as mps will be distributed to the astronomical community , it should be based on some already known package ( not yet another system ) .
it was therefore decided to base the mps gui on the skycat tool distributed by eso ( see ` http://archive.eso.org:8080/skycat/ ` ) .
this tool allows astronomers to couple vimos images and catalogues on which to operate selections of objects over which to place slits .
a new panel for catalogue display , for object selection ( reference , compulsory , forbidden object ) has been implemented . for each type of catalogue
objects a different overlay symbol has been defined .
a dedicated zoom panel allows the definition of curved / tilted slits .
curved slits are defined by fitting a bezier curve to a set of points chosen by the user by clicking on the zoom display .
the fitted curve is then automatically plotted .
the slit width is chosen through a scale widget .
tilted slits can be defined as curved slits and then straightened .
if the astronomer wants to have slits of a width different from the one chosen for the automatic slit placements , he can define them as tilted slits and then align them to the other automatically placed slits .
the core of mask preparation software is the slit positioning optimization code ( spoc ) . given a catalog of objects
, spoc maximizes the number of observable objects in a single exposure and computes the corresponding slit positions .
spoc places slits on the field of view taking into account : special objects ( reference , compulsory , forbidden ) , special slits ( curved , tilted or user s dimension defined ) , spectral first order superposition , spectral higher order superposition and sky region parameter ( the minimum amount of sky to be added to an object size when defining a slit ) .
the issue to be solved is a combinatory computational problem . because of the constraint of slits aligned in the dispersion direction ,
the problem can be slightly simplified : the quadrant area can be considered as a sum of strips which are not necessarily of the same width in the spatial direction .
slits within the same strip have the same length and the alignment of orders is fully ensured .
the problem is thus reduced to be mono - dimensional .
it is easy to show that the number of combination is roughly given by : @xmath3 the slit length ( or strip width ) can vary from a minimum of 4 arcsec ( 20 pixels , i.e. twice the minimum sky region required for the sky subtraction ) to a maximum of 30 arcsec ( 150 pixels , limit imposed by the slit laser cutting machine ) .
the average number of strips can be estimated as the spatial direction size of the fov divided by the most probable slit length : assuming the latter to be 50 pixels ( 10 arcsec ) , we would have @xmath4 strips .
the number of combinations would then be : @xmath5 .
computing these many combinations would correspond to @xmath6 years of cpu work !
the problem is similar to the well known traveling salesman problem : in the standard approach , this is solved by randomly extracting a `` reasonable '' number of combinations and maximizing over this subsample . in our case ,
due to computational time , the `` reasonable '' number of combinations can not be higher than @xmath7 , so small with respect to the total number of combination that the result is not guaranteed to be near the real maximum .
our approach has been to consider only the most `` probable '' combinations , i.e. the ones that have the highest probability to maximize the solution .
step 1 : for each spatial coordinate , we can vary the strip width from the given minimum to the given maximum , count how many objects we can place in the strip , and build the diagram of the number of slits in a strip divided by the strip width as a function of the strip width . for each spatial coordinate ,
only the strip widths corresponding to peaks in this histogram are worth considering , as they correspond to local maxima of the number of slits per strip .
the exact positioning of the peaks varies for each spatial coordinate , but the shape of the function remains the same .
the position of the peaks can be easily found in no more than 6 - 7 trials ( using a partition exchange method ) .
step 2 for each spatial coordinate we have k ( where k is the number of peaks ) possible strips , each with its own length and number of slits . although the number of combinations to be tested is decreased it is still too big in terms of computational time .
step 3 a further reduction can be obtained if , instead of considering all the strips simultaneously , we consider sequentially m subsets of n consecutive strips , which together cover the whole fov . at this point , we should vary n ( and consequently m ) to find the best solution . in practice ,
when n is higher than 8 - 10 , nothing changes in terms of number of observable objects . for n=10 , thus m=4 ( i.e. @xmath8 ) ,
the number of combination is reduced to only @xmath9 , which means a few seconds of cpu work . unfortunately , as a consequence of the optimization process
, small size objects are favored against the big ones . a second ,
less optimized algorithm , has been implemented within spoc .
this alternative algorithm does not optimize all strips simultaneously but builds the @xmath10 function strip by strip without considering object sizes , and takes only the maximum of the distribution .
then it enlarges each strip width by taking into account object sizes . in this way
the number of placed slits decreases by a few percent but the object dimension bias disappears .
a dedicated panel for spoc set up has been implemented within skycat .
trough this panel , users can select the grism , the slit width , and the sky region parameter , the number of masks to be obtained for the given field , and the type of spoc maximization . the slit catalogue produced by spoc
can be loaded as a normal skycat catalog with overlay symbols defined for all kinds of objects , and it is also possible to plot the slit and spectrum overlay for all spoc catalog objects . | the main scientific characteristic of vimos ( visible multi object spectrograph , to be placed on eso very large telescope ) is its high multiplexing capability , allowing astronomers to obtain up to 800 spectra per exposure . to fully exploit such a potential a dedicated tool , the vimos mask preparation software ( mps ) , has been designed .
the mps provides the astronomer with tools for the selection of the objects to be spectroscopically observed , including interactive object selection , handling of curved slits , and an algorithm for automatic slit positioning that derives the most effective solution in terms of number of object selected per field .
the slit positioning algorithm must take into account both the initial list of user s preferred objects , and the constraints imposed either by the instrument characteristics or by the requirement of having easily reduceable data .
the number of possible slit combinations in a field is in any case very high ( @xmath0 ) , and the task of slits maximization can not be solved through a purely combinational approach .
we have introduced an innovative approach , based on the analysis of the function ( number of slits)/(slit length ) vs. ( slit length ) .
the algorithm has been fully tested with good results , and it will be distributed to the astronomical community for the observation preparation phase . |
after experimental observation and the study of bright solitons in bose - einstein condensates ( bec ) @xcite , recent observations @xcite and experimental @xcite and theoretical @xcite studies of a degenerate fermi gas ( dfg ) by sympathetic cooling in the presence of a second boson or fermion component suggest the possibility of soliton formation @xcite in a degenerate fermion - fermion mixture ( dffm ) . apart from the observation of a dfg in the degenerate boson - fermion mixtures ( dbfm ) @xmath0li @xcite , @xmath1na-@xmath2li @xcite and @xmath3rb-@xmath4k @xcite , there have been studies of the following spin - polarized dffm @xmath4k-@xmath4k @xcite and @xmath2li-@xmath2li @xcite . the dimensionless one - dimensional nonlinear schrdinger ( nls ) equation in the attractive ( self - focussing ) case @xcite @xmath5 sustains the following bright soliton @xcite : @xmath6 e^{ivy -i(v^2 -a^2)t/2+i\sigma}\end{aligned}\ ] ] the parameter @xmath7 represents the amplitude as well as pulse width , @xmath8 the velocity , and @xmath9 is a phase constant .
these bright solitons are possible only in one dimension . in two and three dimensions
they are allowed in the presence of a transverse trap @xcite , or in the presence of an oscillating nonlinearity @xcite .
apart from bright solitons in the attractive case , gap solitons are possible in the repulsive ( self - defocussing ) case in the presence of a periodic potential @xcite .
bright solitons in a bec are formed due to a nonlinear atomic attraction @xcite .
as the interaction in a pure dfg at short distances is repulsive due to strong pauli blocking , there can not be bright solitons in a dfg .
however , bright solitons can be formed @xcite in a dbfm in the presence of a sufficiently strong boson - fermion attraction which can overcome the pauli repulsion among identical fermions .
bright solitons can also be formed in a binary mixture of repulsive bosons supported by interspecies attraction @xcite .
we demonstrate the formation of stable fermionic bright solitons in a dffm for a sufficiently attractive interspecies interaction . in a dffm
, the coupled system can lower its energy by forming high density regions , the bright solitons , when the attraction between the two types of fermions is large enough to overcome the pauli repulsion .
we use a coupled time - dependent mean - field - hydrodynamic model for a dffm and consider the formation of axially - free localized bright solitons in a quasi - one - dimensional cigar - shaped geometry using numerical and variational solutions .
the present model is inspired by the success of a similar model used recently in the investigation of collapse @xcite and bright @xcite and dark @xcite solitons in a dbfm , and black solitons @xcite and mixing - demixing @xcite in a dffm .
we study the condition of modulational instability of a constant - amplitude solution of the present model under a plane - wave perturbation and demonstrate the possibility of the formation of bright solitons by a linear stability analysis .
we find that for a sufficiently strong interspecies interaction , under the plane - wave perturbation the constant - amplitude solution becomes unstable and localized solitonic solutions may appear .
we present a numerical stability analysis of these bright solitons by introducing different small perturbations when the solitons undergo stable and sustained breathing oscillation .
we also consider and study the formation of a soliton train in a dffm by a large sudden jump in the interspecies fermion - fermion scattering length realized by manipulating a background magnetic field near a feshbach resonance , experimentally observed in two hyperfine states of the following spin - polarized dffms : @xmath2li-@xmath2li and @xmath4k-@xmath4k @xcite .
this procedure effectively and rapidly turns a repulsive dffm to a highly attractive one and generates bright solitons .
an experimental realization of bright solitons in a dffm could be tried in two hyperfine states in samples such as @xmath2li-@xmath2li or @xmath4k-@xmath4k using a feshbach resonance
. however , there is already experimental evidence @xcite and theoretical conjecture @xcite that the maximum attractive force , that can be created by a feshbach resonance in fermionic atoms in two hyperfine states is limited by quantum mechanical constraints of unitarity .
although there is a limit to the creation of attraction in a two - component spin - polarized dffm in two hyperfine states , there is no such limit in a multi - component spin - polarized dffm @xcite or a dffm composed of atoms of distinct mass @xcite .
so , if bright solitons and soliton trains can not be efficiently created in a two - component dffm in two hyperfine states , a better and more efficient dffm for the creation of solitons could be the mixture of fermionic atoms of distinct mass , where one can avoid the problem of a possible suppression of interspecies attraction .
one such system , among many others , could be the mixture of spin - polarized @xmath2li-@xmath4k mixture : both @xmath2li @xcite and @xmath4k @xcite have been trapped and studied in laboratory .
the formation of bright solitons in a dffm by turning the effective interspecies interaction among spin - polarized fermions from repulsion to strong attraction seems within the reach of present experimental possibilities .
here we shall be interested in the formation of bright solitons in a spin - polarized dffm of different fermionic atoms in the presence of strong interspices attraction and we shall ignore the possibility of the formation of a bardeen - cooper - schreiffer ( bcs ) condensate @xcite through cooper pairing . a bcs condensate is usually formed in the presence of a weak attraction among identical fermions with fermion pairs in the singlet ( spin parallel ) state .
the formation of a bcs condensate should not be favored in a spin - polarized ( spin antiparallel ) two - component dffm with strong attraction among the different types of atoms , as a strong interspecies attraction is not the domain of bcs condensation . by choosing a strong interspecies attraction
our study stays in the bec region of the bcs - bose crossover problem @xcite strongly favoring molecule formation .
the solitons once formed will decay via molecule formation of two different types of fermionic atoms @xcite as in the case of bosonic solitons .
nevertheless , such molecule formation is a slow process and can be accounted for by a three - body recombination term as introduced in a previous study of collapse in a dffm @xcite .
the molecule formation will eventually destroy the solitons in the dffm , albeit , at a slow rate .
the same is true in the formation of soliton and soliton train in an attractive bec , where the essential features of the dynamics have been well explained within the mean - field gross - pitaevskii model by neglecting molecule formation @xcite .
a subsequent study including the effect of molecule formation @xcite has not essentially changed the conclusions of @xcite .
hence this pioneering study on the formation of soliton and soliton trains in a dffm of different atoms using mean - field hydrodynamic equations with the neglect of molecule formation is expected to explain the essential features of the dynamics . however , it would be of interest to include the effect of molecule formation on the dffm solitons in a future study .
in section 2 we present our mean - field - hydrodynamic model and its reduction to a quasi - one - dimensional form in a cigar - shaped geometry . in section 3
we show that bright solitons can appear in this model through modulational instability of a constant amplitude solution . in section 4
we perform a variational analysis of the mean - field equations .
numerical results for isolated bright solitons are presented in section 5 and are compared with variational results .
then we study the generation of a train of solitons by modulational instability .
finally , we present a summary in section 6 .
we use a simplified mean - field - hydrodynamic lagrangian for a dfg used successfully to study a dbfm @xcite . to develop a set of time - dependent mean - field - hydrodynamic equations for the interacting dffm , we use the following lagrangian density @xcite @xmath10+\end{aligned}\ ] ] @xmath11 where @xmath12 represents the two components , @xmath13 the probability amplitude , @xmath14 the probability density
, @xmath15 denotes complex conjugate , @xmath16 the mass , @xmath17 the interspecies coupling @xmath18 with @xmath19 the reduced mass , and @xmath20 the interspecies fermion - fermion scattering length . the number of fermionic atoms @xmath21 is given by @xmath22 .
the trap potential with axial symmetry is taken as @xmath23 where @xmath24 and @xmath25 are the angular frequencies in the radial ( @xmath26 ) and axial ( @xmath27 ) directions with @xmath28 the anisotropy .
the @xmath29 limit corresponds to a cigar - shaped geometry and allows a reduction of the three - dimensional equations to a quasi - one - dimensional form appropriate for freely moving solitons .
the interaction between identical intra - species fermions in spin - polarized state is highly suppressed due to pauli blocking terms @xmath30 and has been neglected in ( [ yy ] ) .
the kinetic energy terms @xmath31 @xmath32 in ( [ yy ] ) contribute little to this problem compared to the dominating pauli - blocking terms .
however , its inclusion leads to an analytic solution for the probability density everywhere @xcite . with the lagrangian density ( [ yy ] ) , the following euler - lagrange equations : @xmath33 with @xmath34 being the three space components ,
become @xmath35\psi_j=0,\end{aligned}\ ] ] where @xmath36 .
this is a time - dependent version of a similar time - independent hydrodynamic model for fermions @xcite . for large @xmath37 ,
both lead to @xcite the thomas - fermi result @xcite @xmath38^{3/2}$ ] with @xmath39 the chemical potential .
they yield identical results for time - independent stationary states .
however , the present time - dependent model can be used in the study of nonequilibrium dynamics , as in the study of soliton trains . for @xmath40 , ( [ e ] )
can be reduced to an effective one - dimensional form by considering solutions of the type @xmath41 where @xmath42 with @xmath43 corresponds to the ground state wave function in the radial trap alone in the absence of nonlinear interactions . here to have an algebraic simplification we have taken the masses of two types of fermions to be equal ( @xmath44 ) .
these wave functions satisfy @xmath45 with normalization @xmath46 now the dynamics is carried by @xmath47 and the radial dependence is frozen in the ground state @xmath48 . in the quasi - one - dimensional cigar - shaped geometry
the linear fermionic probability densities are given by @xmath49 .
averaging over the radial mode @xmath50 , i.e. , multiplying ( [ e ] ) by @xmath51 and integrating over @xmath26 , we obtain the following one - dimensional dynamical equations @xcite : @xmath52
\phi_j(z , t)=0 , \nonumber \\\end{aligned}\ ] ] where @xmath53 @xmath54^{2/3}.\nonumber \end{aligned}\ ] ] in ( [ i ] ) the normalization is given by @xmath55 . in these equations
we have set the anisotropy parameter @xmath56 to remove the axial trap and thus to generate axially - free quasi - one - dimensional solitons . to reduce three - dimensional equations ( [ i ] ) to a dimensionless form ,
following @xcite , we consider variables @xmath57 , @xmath58 , @xmath59 , with @xmath60 , while ( [ i ] ) becomes @xmath61{\varphi}_{{j}}({y},\tau)=0 , \nonumber \\\end{aligned}\ ] ] where @xmath62 and @xmath63 . here
we employ a negative @xmath64 corresponding to attraction , and @xmath65 in ( [ m ] ) a sufficiently strong attractive fermion - fermion coupling @xmath66 can overcome the pauli repulsion @xmath67 and form bright solitons . for the usual gross - pitaevskii with the cubic nonlinearity
, the reduction of the full three - dimensional equation to its one - dimensional counterpart was performed , in different forms , by many authors @xcite .
the nonlinearity with 4/3 power in ( [ m ] ) was obtained starting from a similar nonlinearity in a three - dimensional dfg .
a strict one - dimensional ( two - dimensional ) dfg will give rise to a quintic ( cubic ) nonlinearity . whatever be the nonlinearity in the intraspecies fermions , bright fermionic solitons will be generated , provided that the interspecies attraction is strong enough to overcome the intraspecies repulsion due to pauli blocking .
however , in this paper we shall consider only the pauli - blocking nonlinearity with the 4/3 power as in ( [ m ] ) .
the two coupled equations ( [ m ] ) could be simplified for @xmath68 , while @xmath69 , and these equations reduce to the following single equation @xmath70{\varphi}({y},\tau)=0,\end{aligned}\ ] ] where @xmath71 and @xmath72 .
equation ( [ o ] ) maintains the essential features of ( [ m ] ) , e.g. , a quadratic nonlinear attraction and a pauli blocking repulsion .
the one - dimensional equation ( [ o ] ) is quite similar in structure to the nls equation ( [ nls ] ) apart for the pauli - blocking repulsive term @xmath73 .
for a small @xmath73 and large @xmath74 the solitons of ( [ nls ] ) survive in ( [ o ] ) . however , they disappear in the opposite limit of large @xmath73 and small @xmath74 .
we find that ( [ o ] ) allows a constant - amplitude solution which exhibits modulational instability leading to a modulation of the solution .
we perform a stability analysis of this solution and study the possibility of generation of solitons by modulational instability in the symmetric case .
we consider the constant - amplitude solution @xcite @xmath76\ ] ] of ( [ o ] ) , where @xmath77 is the constant amplitude and @xmath78 a phase .
the time evolution of solution ( [ so ] ) maintains the constant amplitude @xmath77 but acquires an amplitude - dependent phase .
now we study if this solution is stable against small perturbations by performing a linear stability analysis .
we consider a small perturbation of the constant - amplitude solution ( [ so ] ) given by : @xmath79 where @xmath80 is the small perturbation .
substituting the perturbed solution ( [ per ] ) in ( [ o ] ) , and for small perturbations retaining only the linear terms in @xmath81 we get @xmath82 we consider the complex plane - wave perturbation @xmath83 in ( [ p ] ) , where @xmath84 and @xmath85 are the amplitudes of the real and imaginary parts , respectively , and @xmath86 is a frequency parameter and @xmath87 a wave number . then separating the real and imaginary parts we get @xmath88 and eliminating @xmath84 and @xmath85 we obtain the dispersion relation @xmath89^{1/2}.\ ] ] the constant - amplitude solution ( [ so ] ) is stable if perturbations at any wave number @xmath87 do not grow with time .
this is true as long as frequency @xmath86 is real . from ( [ di ] )
we find that @xmath86 remains real for any @xmath87 provided that @xmath90 or @xmath91 .
however , @xmath86 can become imaginary for @xmath92 and the plane - wave perturbations can grow exponentially with time @xmath93 .
this is the domain of modulational instability of a constant - intensity solution @xcite .
the perturbation then grows exponentially with the intensity given by the growth rate or the modulational instability gain @xmath94 defined by @xmath95^{1/2},\ ] ] where @xmath96 denotes imaginary part .
the presence of modulational instability is closely connected with the appearance of a bright soliton @xcite .
localized bright solitons are possible only when the constant - amplitude solution is unstable .
now we consider the possibility of modulational instability of a similar constant - intensity solution in coupled equations ( [ m ] ) .
we consider the constant - amplitude solutions @xmath98 , \nonumber\end{aligned}\ ] ] of ( [ m ] ) , where @xmath99 is the amplitude and @xmath100 a phase for component @xmath101 .
the constant - amplitude solution develops an amplitude dependent phase on time evolution .
we consider a small perturbation @xmath102 to these solutions via @xmath103 where @xmath104 . substituting this perturbed solution in ( [ m ] ) , and for small perturbations retaining only the linear terms in @xmath105 we get @xmath106@xmath107 we consider the complex plane - wave perturbation @xmath108 in ( [ q ] ) for @xmath12 , where @xmath109 and @xmath110 are the amplitudes for the real and imaginary parts , respectively , and @xmath86 and @xmath87 are frequency and wave numbers . then separating the real and imaginary parts we get @xmath111 for @xmath112 , and @xmath113 for @xmath114 .
eliminating @xmath115 between ( [ p1 ] ) and ( [ p2 ] ) we obtain @xmath116= 2{\cal a}_{21}n_{12}a_{10}a_{20 } \omega^2 , \nonumber \\ \end{aligned}\ ] ] and eliminating @xmath117 between ( [ p3 ] ) and ( [ p4 ] ) we obtain @xmath118= 2{\cal a}_{11}n_{21}a_{10}a_{20 } \omega^2 .
\nonumber \\ \end{aligned}\ ] ] eliminating @xmath119 and @xmath120 from ( [ p5 ] ) and ( [ p6 ] ) , finally , we obtain the following dispersion relation @xmath121@xmath122^{1/2}.\nonumber \end{aligned}\ ] ] for stability of the plane - wave perturbation , @xmath86 has to be real . for any @xmath87 this happens for @xmath123 @xmath124 or for @xmath125 however , for @xmath126 @xcite , @xmath86 can become imaginary and the plane - wave perturbation can grow exponentially with time . this is the domain of modulational instability of a constant - intensity solution signalling the possibility of coupled fermionic bright soliton to appear . in the symmetric case @xmath75 and @xmath127 , consequently , @xmath128 and @xmath129 and we recover the condition of modulational instability @xmath130 derived in section 3.1 .
next we present a variational analysis of ( [ o ] ) based on the gaussian trial wave function @xcite @xmath131,\ ] ] where @xmath81 is the amplitude , @xmath132 is the width , @xmath133 the chirp , and @xmath134 the phase .
the lagrangian density for ( [ o ] ) is the one - term version of ( [ yy ] ) , e.g. , @xmath135 + \biggr|\frac{\partial \varphi}{\partial y } \biggr|^2 -\frac{1}{2 } \gamma n ^2 + \frac{3}{5}\beta n^ { 5/3},\nonumber \end{aligned}\ ] ] which is evaluated with this trial function and the effective lagrangian @xmath136 becomes @xmath137 @xmath138 where the overhead dot denotes time derivative .
the variational lagrange equations @xmath139 where @xmath140 stands for @xmath134 , @xmath81 , @xmath132 and @xmath133 can then be written as @xmath141 @xmath142 the constant in ( [ con ] ) is fixed by the normalization condition .
eliminating @xmath143 from ( [ x ] ) and ( [ y ] ) we obtain @xmath144 the use of ( [ con ] ) , ( [ z ] ) and ( [ eli ] ) leads to the following differential equation for the width @xmath132 : @xmath145 . \end{aligned}\ ] ] the quantity in the square bracket is the effective potential of the equation of motion .
small oscillation around a stable configuration is possible when there is a minimum in this potential .
the variational result for width @xmath132 follows by setting the right hand side of ( [ min ] ) to zero corresponding to a minimum in this effective potential , from which the variational profile for the soliton can be obtained @xcite .
the above variational analysis can be extended to the asymmetric case .
however , the algebra becomes quite involved if we take a general variational trial wave function with chirp and phase parameters .
as we are interested mostly in the density profiles , we consider the following normalized gaussian trial wave function for fermion component @xmath101 of ( [ m ] ) @xmath146 , \quad j=1,2.\ ] ] using essentially the lagrangian density ( [ yy ] ) in this case we obtain the following effective lagrangian @xmath147 with @xmath148 .
the variational lagrange equations ( [ la ] ) for @xmath149 and @xmath150 now become @xmath151 equations ( [ cp ] ) can be solved for variational widths @xmath152 and consequently the variational profile of the wave functions obtained from ( [ vx ] ) . when @xmath75 , in ( [ cp ] ) @xmath153 , @xmath154 and @xmath155 ; and in this case ( [ cp ] ) yields the same variational widths as from the result obtained in the symmetric case in section 4.1 given by ( [ min ] ) ; e. g. , @xmath156
we solve ( [ m ] ) for bright solitons numerically using a time - iteration method based on the crank - nicholson discretization scheme @xcite .
we discretize the coupled partial differential equations ( [ m ] ) using time step @xmath157 and space step @xmath158 in the domain @xmath159 .
the second derivative in @xmath160 is discretized by a three - point finite - difference rule and the first derivative in @xmath93 by a two - point finite - difference rule .
we perform a time evolution of ( [ m ] ) introducing an harmonic oscillator potential @xmath161 in it and setting the nonlinear terms to zero , and starting with the eigenfunction of the harmonic oscillator problem : @xmath162 the extra harmonic oscillator potential , set equal to zero in the end , only aids in starting the time evolution with an exact analytic form . with this initial solution
we perform time evolution of ( [ m ] ) . during the time evolution
the nonlinear terms are switched on slowly and the harmonic oscillator potential is switched off slowly and the time evolution continued to obtain the final converged solutions .
in addition to solving the coupled equations ( [ m ] ) , we also solved the single equation ( [ o ] ) in the symmetric case : @xmath75 . for given values of @xmath163 and @xmath164 ,
solitons can be obtained for @xmath165 above a certain value . in the coupled - channel case solitons
are easily obtained for a smaller @xmath165 when @xmath163 and @xmath164 are not very different from each other . of ( [ m ] ) vs. @xmath27 for @xmath166 , @xmath167 , @xmath168 nm , while @xmath169 , @xmath170 , @xmath171 , and @xmath172 .
the corresponding variational solutions @xmath173 calculated from the coupled equations ( [ cp ] ) are also shown . ] in our numerical study we take @xmath174 @xmath175 m and consider a dffm consisting of two electronic states of @xmath4k atoms .
this corresponds to a radial trap of frequency @xmath176 hz .
consequently , the unit of time is @xmath177 ms . for another fermionic atom the @xmath24 value
gets changed accordingly for @xmath174 @xmath175 m . from a solution of coupled equations ( [ m ] )
we find that these equations permit solitonic solutions provided that @xmath165 is larger than a certain threshold value consistent with the analysis of section 3 .
first we solve ( [ m ] ) for @xmath178 , and @xmath179 nm . in this case no solitons are allowed for @xmath180 nm .
the solitonic solution suddenly appears as @xmath165 increases past @xmath181 nm .
the solitons in this case are shown in figure 1 , where we also plot the variational solutions of coupled equations ( [ cp ] ) . in this case
the nonlinearity parameters are @xmath169 , @xmath170 , @xmath171 , and @xmath172 . substituting these values of the nonlinearities in ( [ cp ] ) and solving we obtain the variational widths @xmath182 and @xmath183 . from ( [ v ] ) and ( [ con ] )
we then obtain the variational soliton profiles @xmath184 and @xmath185 . from figure 1
we find that the variational results agree well with the numerical solutions of the coupled equations . in this case
we also found the variational result in the symmetric case ( @xmath186 ) using ( [ sim ] ) , which lies close to the above two variational solutions . and ( b ) @xmath187 of figure 1 vs. @xmath27 and @xmath188 for @xmath166 and @xmath167 . at @xmath189 ms ( marked by arrows ) the bright solitons are set into small breathing oscillation by suddenly changing the fermion numbers to @xmath190 and @xmath191.,title="fig : " ] and ( b ) @xmath187 of figure 1 vs. @xmath27 and @xmath188 for @xmath166 and @xmath167 . at @xmath189 ms ( marked by arrows )
the bright solitons are set into small breathing oscillation by suddenly changing the fermion numbers to @xmath190 and @xmath191.,title="fig : " ] to test the robustness of these solitons we inflicted different perturbations on them and studied the resultant dynamics numerically .
first , after the formation of the solitons of figure 1 with @xmath166 and @xmath167 we suddenly changed the particle numbers to @xmath190 and @xmath191 at time @xmath192 ms . this corresponds to a sudden change of nonlinearities from @xmath193 , @xmath194 , @xmath195 , and @xmath196 to @xmath197 , @xmath198 , @xmath199 , and @xmath200 .
the resultant dynamics is shown in figures 2a and 2b .
due to the sudden change in nonlinearities , the fermionic bright solitons are set into stable non - periodic small - amplitude breathing oscillation .
this demonstrates the robustness of the solitons .
after the formation of the solitons of figure 1 with @xmath166 and @xmath167 we suddenly changed the interspecies scattering length @xmath64 from @xmath201 nm to @xmath202 nm at time @xmath192 ms .
this can be realized by manipulating a background magnetic field near a fermion - fermion feshbach resonance @xcite this corresponds to a sudden change of nonlinearities from @xmath193 , @xmath194 , @xmath195 , and @xmath196 to @xmath193 , @xmath203 , @xmath204 , and @xmath196 .
due to the sudden change in nonlinearities , the fermionic bright solitons are set into stable non - periodic small - amplitude breathing oscillation .
instead of plotting the soliton profile in this case we plot the root mean square ( rms ) size of the solitons in figure 3 which demonstrates the stable nonperiodic breathing oscillation .
we also ( i ) gave a small displacement between the centers of these solitons and ( ii ) suddenly changed @xmath205 and @xmath206 .
in both cases after oscillation and dissipation the solitons continue stable propagation which shows their robust nature . and @xmath187 of figure 1 vs. @xmath188 .
at @xmath189 ms the solitons of figure 1 are set into small breathing oscillation by suddenly changing @xmath64 from @xmath201 nm to @xmath202 nm . ] upon removing the harmonic trap @xmath161 and jumping the nonlinearities at @xmath207 from @xmath208 to ( a ) @xmath209 ( b ) @xmath210 and to ( c ) @xmath211 respectively , corresponding to a jump in scattering length @xmath64 from @xmath201 nm to @xmath212 nm , @xmath213 nm , and @xmath214 nm . for this purpose we solved the coupled equations ( [ m]).,title="fig : " ] upon removing the harmonic trap @xmath161 and jumping the nonlinearities at @xmath207 from @xmath208 to ( a ) @xmath209 ( b ) @xmath210 and to ( c ) @xmath211 respectively , corresponding to a jump in scattering length @xmath64 from @xmath201 nm to @xmath212 nm , @xmath213 nm , and @xmath214 nm . for this purpose we solved the coupled equations ( [ m]).,title="fig : " ] upon removing the harmonic trap @xmath161 and jumping the nonlinearities at @xmath207 from @xmath208 to ( a ) @xmath209 ( b ) @xmath210 and to ( c ) @xmath211 respectively , corresponding to a jump in scattering length @xmath64 from @xmath201 nm to @xmath212 nm , @xmath213 nm , and @xmath214 nm . for this purpose we solved the coupled equations ( [ m]).,title="fig : " ] during the time evolution of ( [ m ] ) if the nonlinearities are changed by a small amount or changed slowly , usually one gets a single stable soliton when the final nonlinearities are appropriate .
however , if the interspecies attraction is increased suddenly by a large amount by jumping @xmath64 from a positive ( repulsive ) value to a large negative ( attractive ) value , a soliton train is formed as in the experiment with bec @xcite because of modulational instability @xcite .
to illustrate the formation of soliton train in a fermion - fermion mixture in numerical simulation , we consider the solution of ( [ m ] ) for @xmath215 and @xmath179 nm with an added axial harmonic trap @xmath161 corresponding to nonlinearities @xmath216 and @xmath217 . after the formation of the solitons in the axial trap we suddenly jump the scattering length to @xmath212 nm , corresponding to off - diagonal nonlinearities @xmath218 , and also switch off the harmonic trap at time @xmath207 . although the initial value of scattering length ( @xmath179 nm ) in this case corresponds to an interatomic attraction , because of strong pauli - blocking repulsion the effective fermion - fermion interaction is repulsive in this case .
however , the final value of scattering length ( @xmath219 nm ) corresponds to a stronger interatomic attraction which can overcome the pauli repulsion so that an effective fermion - fermion attraction emerges in this case which might allow the formation of soliton(s ) . in our numerical simulation
we find that this is indeed the case . upon a jump of the scattering length to @xmath219 nm , we find that after some initial noise and dissipation the time evolution of ( [ m ] ) generates a single stable bright soliton as shown in figure 4a
. however , more solitons in the form of a soliton train can be formed for a larger jump in the scattering length . in figure 4b , from the same initial state in figure 4a we consider a larger jump of @xmath64 to @xmath220 nm corresponding to off - diagonal nonlinearities @xmath221 .
the final value of scattering length ( @xmath222 nm ) in this case corresponds to a stronger interatomic attraction than considered in figure 4a which might allow the formation of soliton trains .
this is verified in numerical simulation . upon a jump of the scattering length to @xmath222 nm , from @xmath179 nm
, we find that after some initial noise and dissipation the time evolution of ( [ m ] ) generates three slowly receding bright solitons as shown in figure 4b .
more solitons can be generated when the jump in the scattering length @xmath64 or off - diagonal nonlinearities is larger . in figure 4c
we show the generation of five receding solitons of each component upon a sudden jump of the scattering length @xmath64 to @xmath214 nm from the initial state of figure 4a .
this corresponds to a jump of the off - diagonal nonlinearities to @xmath223 .
the formation of soliton trains from a stable initial state is due to modulational instability @xcite .
the sudden jump in the off - diagonal nonlinearities could be effected by a jump in the interspecies scattering length @xmath64 obtained by manipulating a background magnetic field near a feshbach resonance @xcite .
we used a coupled mean - field - hydrodynamic model for a dffm to study the formation of bright solitons and soliton trains in a quasi - one - dimensional geometry by numerical and variational methods .
we find that an attractive interspecies interaction can overcome the pauli repulsion and form fermionic bright solitons in a dffm .
this is similar to the formation of bright solitons in a coupled boson - boson @xcite and boson - fermion @xcite mixtures supported by interspecies interaction .
we show by a linear stability analysis that when the interspecies attraction is larger than a threshold value , bright solitons can be formed due to modulational instability of a constant - amplitude solution of the nonlinear equations describing the dffm . the stability of these solitons is demonstrated numerically through their sustained breathing oscillation initiated by a sudden small perturbation .
we also illustrate the creation of soliton trains upon a sudden large jump in interspecies attraction by manipulating a background magnetic field near a feshbach resonance @xcite resulting in a sudden jump in the off - diagonal nonlinearities .
this jump transforms an effectively repulsive dffm into an effectively attractive one , responsible for the formation of a soliton train due to modulational instability .
bright solitons and soliton trains have been created experimentally in attractive becs in the presence of a radial trap only without any axial trap @xcite in a similar fashion by transforming a repulsive bec into an attractive one . in view of this
, fermionic bright solitons and trains could be created in laboratory in a dffm in a quasi - one - dimensional configuration .
here we used a set of mean - field equations for the dffm .
a proper treatment of a dfg or dffm should be done using a fully antisymmetrized many - body slater determinant wave function @xcite as in the case of scattering involving many electrons @xcite . however , in view of the success of a fermionic mean - field model in studies of collapse @xcite , bright @xcite and dark @xcite soliton in a dbfm and of mixing and demixing in a dffm @xcite , we do not believe that the present study on bright solitons in a dffm to be so peculiar as to have no general validity .
99 hulet r g 2006 private communication .
i thank dr .
hulet for a discussion , which clarified this point , suggesting the enhanced possibility of the formation of soliton train in a dffm of different atoms . | we employ a time - dependent mean - field - hydrodynamic model to study the generation of bright solitons in a degenerate fermion - fermion mixture in a cigar - shaped geometry using variational and numerical methods . due to a strong pauli - blocking repulsion among identical spin - polarized fermions at short distances there
can not be bright solitons for repulsive interspecies interactions . employing a linear stability analysis we demonstrate the formation of stable solitons due to modulational instability of a constant - amplitude solution of the model equations for a sufficiently attractive interspecies interaction .
we perform a numerical stability analysis of these solitons and also demonstrate the formation of soliton trains by jumping the effective interspecies interaction from repulsive to attractive .
these fermionic solitons can be formed and studied in laboratory with present technology . |
gamma - ray burst ( grb ) is one of the most mysterious phenomena in the universe .
although there was a great revolution in our understanding on grbs in 1997 that some of them were proved to be of extragalactic origins @xcite , there are still a lot of questions and puzzles left on the phenomena .
that is why grbs have been the fascinating targets for astronomers .
in fact , there are a lot of interesting and valuable discussions under debate . for example
, we have not known what is the origin of grb .
although this problem has never been settled and still controversial , there are some reports that suggest the physical connections between grbs and supernovae ( sne ) @xcite .
if the sn / grb connection is correct , grbs will be born from the death of massive stars .
then the grb event rate should trace the star formation history in the universe @xcite .
the central engine of a grb is an important but difficult problem to be solved .
although a fireball model @xcite is a promising and attractive one , we do not know how the initial condition is .
we merely gave it artificially .
in the fireball model , since the ratio of baryon number to photon number is required to be extremely small as an initial condition , only a peculiar environment could meet such a condition .
as long as we believe the grb / sn connection , however , a collapsar ( failed supernova ) model @xcite will be one of the convincing candidates as the central engine of a grb . in the model ,
much energy is generated by the neutrino heating from an accretion disk surrounding a black hole which is made by the core collapse of a massive star .
on the other hand , total explosion energy of a grb is also one of the most important problems in the physics of grbs .
observationally it is well known that there is a large dispersion of total energy of gamma - rays emitted from grbs @xcite .
although the dispersion may become small when we consider the beaming effects @xcite , after all we do not know the total explosion energy of the system that generates a grb only by the gamma - ray observations . in the theoretical side ,
the total explosion energy in the collapsar model is also not estimated clearly , and only a rough estimate is reported as a result of the numerical computations @xcite .
therefore , to know the total explosion energy , we require the multi - dimension numerical hydrodynamical computations including the microphysics of neutrino heating in future , which is , of course , a laborious task to be performed .
are there observations which give us new clues to know how a grb is produced ? in this study , we point out that the grb neutrino background may be detected in a large water cherenkov neutrino detector such as super - kamiokande ( sk ) and titand which is proposed as a next - generation plan of the multi - megaton water cherenkov detector @xcite .
we also stress that this signal should give us valuable informations on grbs , such as the total explosion energy and/or event rate of grbs .
the reason is as follows . in this study
, we assume that the origins of grbs are massive stars .
thus , the event rate of grbs should trace the star formation history in the universe . since we adopt the collapsar model here , we can assume that the central engine of a grb is the neutrino heating from an accretion disk surrounding the central black hole .
we show that the total number of neutrinos emitted from a grb is very sensitive to the mass - accretion rate @xmath2 which is difficult to estimate only from the existing theoretical models .
however , we can obtain the informations of the mass - accretion rate by observing the grb neutrino background .
we also derive an approximate relationship between the total explosion energy and the mass - accretion rate of a grb .
namely , although we can not determine the total explosion energy only from the gamma - ray observations , we can obtain informations of it by observing the grb neutrino background .
that is an important clue to understand grbs . in this study
we find that the grb neutrino background may be detected within @xmath0 10 yrs by titand as long as the average mass - accretion rate among grbs is @xmath3 a few @xmath4 s@xmath5 , and the probability that one collapsar generates a grb is @xmath6 @xmath0 0.5 1.0 .
here we outline the organization of this paper . in section [ formula ] , we explain the formulations in this study .
results are shown in section [ results ] .
discussions are given in section [ discussion ] .
summary and conclusion are presented in section [ summary ] .
first we assume that the rate of grbs traces the global star formation history of the universe , @xmath7 .
although a number of workers have modeled the expected evolution of the cosmic sfr with redshift , there are some uncertainties ( in particular , at high redshift @xmath8 ) , and further observations are required to determine the cosmic sfr . in this study , we use the three different parameterizations of the global star formation rate per comoving volume in an einstein - de sitter universe to take into account the uncertainty of the cosmic sfr which is derived by ref .
they are @xmath9 @xmath10 and @xmath11 where @xmath12 . assuming that massive stars with masses larger than @xmath13 explode as grbs @xcite @xcite , grb - rate density @xmath14
can then be estimated by multiplying the selected sfr by the coefficient @xmath15 where @xmath16 is the initial mass function ( imf ) and @xmath17 is the stellar mass in solar units . in this study , the salpeter s imf ( @xmath18 ) is adopted throughout .
here we assumed that one tenth of the collapsars whose mass range is in ( 35 125)@xmath4 are accompanied with grbs ( @xmath19 ) .
the validity of this assumption and the detailed discussions are presented in section [ discussion ] .
we label these models as grbr1 , grbr2 , and grbr3 , respectively . for comparison
, we use the sn - rate density @xmath20 by multiplying the selected sfr by the coefficient @xmath21 as for the energy spectrum of @xmath22 of sne , the fermi - dirac distribution with zero chemical potential is used ( see figs [ flux1] [ flux3 ] ) , which have the same luminosity with the numerical model @xcite . in the flat universe , i.e. , when the density parameter is @xmath23 with the non - relativistic matter density ( @xmath24 ) and the @xmath25 term ( @xmath26 ) , the star formation rate is given by @xcite @xmath27 where @xmath28 is redshift . in fig .
[ fig1 ] , we show the grb rates as a function of the redshift
. in the collapsar model , neutrinos are mainly generated at the inner - most region of an accretion disk which is formed around the central black hole .
thus we estimate the energy spectrum of neutrinos using an analytical shape of the accretion disk .
popham et al .
( 1999 ) proposed an accretion - disk model including the effects of general relativity as a collapsar model @xcite .
the model in ref .
@xcite and the numerically computed collapsar model @xcite are fitted well by the analytical shape of the accretion disk in ref @xcite .
we use it in this study .
then , the density [ g @xmath29 , the temperature [ mev ] and the disk thickness [ cm ] are fitted as follows ( see ref .
@xcite for details ) .
@xmath30 where @xmath31 is the mass of the central black hole , @xmath32 is the radial coordinate , and @xmath33 (= 10@xmath34 cm ) is the core radius , respectively .
note that the schwarzshild radius is @xmath35 . in ref .
@xcite , they assumed that the mass of the central black hole is constant . in this study , however , to estimate the total energy of neutrinos we consider the effects that the mass of the central black hole becomes larger with time by accretions .
when we know the density and the temperature in the neutrino emitting region , we can estimate the emissivity of neutrinos ( @xmath36 ) in the accretion disk .
the total emissivity of @xmath22 is represented by two terms : @xmath37 where @xmath38 is energy of @xmath22 . for @xmath39 , the emissivity of @xmath22
is represented by @xmath40 where @xmath41 is fermi coupling constant , @xmath42 is the axial vector coupling constant of the nucleon which is normalized by the experimental value of neutron lifetime @xmath43 s @xcite , @xmath44 is number density of neutron , q @xmath45 1.29 mev , and @xmath46 is electron mass . for @xmath47 , we obtain @xmath48 where @xmath49 , @xmath50 = 1/2 , and @xmath51 is weinberg angle @xcite , and we assume @xmath52 .
we plot them in fig .
[ fig2 ] in the case that @xmath53 5mev and @xmath54 .
using eqs .
( [ eq : rho_n ] ) , ( [ eq : temperature ] ) and ( [ eq : disk_thicknss ] ) , we obtain the integrated energy spectrum of @xmath22 , @xmath55 where @xmath56 denotes the integration in the volume of the emitting region in accretion disk where the temperature is high enough to produce neutrinos , i.e. , @xmath57 mev , @xmath58 denotes the integration of time @xmath59 , and @xmath60 is the duration of the accretion ( @xmath61 is the mass of the progenitor ) . here
we assume that the time evolution of the mass of the central black hole is given by @xmath62 where @xmath63 is the initial value of @xmath31 .
in this study , @xmath63 is set to be 3@xmath4 . in fig .
[ fig3 ] we plot the obtained energy spectrum of @xmath22 emitted from the accretion disk in unit energy [ @xmath64 . here
we define the time - dependent luminosity of @xmath65 and @xmath22 by @xmath66 where @xmath67 is the total emissivity of @xmath65 with their energy @xmath68 , and we assume it is approximately equal to @xmath69 . in addition , we define the total energy of @xmath65 s and @xmath22 s emitted from a collapsar , @xmath70 we see from eq .
( [ eq : temperature ] ) that as the mass of the central black hole grows , the temperature becomes lower , and the flux of neutrinos decreases . in fig .
[ duration ] , we show the duration ( @xmath71 ) which is defined as the e - folding time of the time - dependent luminosity in eq .
( [ eq : l_nu1 ] ) .
this almost corresponds to the timescale in which the temperature at the innermost region of the accretion disk becomes lower than @xmath0 1 mev .
namely this will approximately reflect the timescale of a grb .
now we investigate the dependence of the energy spectrum of anti - electron neutrino on the mass - accretion rate ( @xmath2 ) .
the mass - accretion rate is theoretically not known while the numerical simulations have been performed , because there are uncertainties on the viscosity of the accretion disk and angular momentum of the progenitor . in fig .
[ fig : a ] , we show the energy spectrum for the case of @xmath2 = 0.01@xmath4 s@xmath5 , 0.1@xmath4 s@xmath5 , 1@xmath4 s@xmath5 , 5@xmath4 s@xmath5 , and 10@xmath4 s@xmath5 .
the energy spectrum is represented as @xmath72 [ mev ] because the cross section of the reaction @xmath73 is proportional to the square of the energy .
it is clearly shown that the energy spectrum becomes hard , and intensity becomes large as the mass - accretion rate increases .
this is because the temperature at the inner region of the accretion disk becomes higher .
here we have to give some comments on the application limit for the present accretion disk model in this study .
high energy electron - positron pairs whose energies are greater than @xmath74mev produce muons and charged pions , which , in turn , decay into low energy leptons .
such microphysics is not included in the present model .
the influence might become important for the high accretion rate model .
in addition , as the density becomes large , the assumption that the neutrinos are optically thin might become to break down @xcite .
for example , the opacity of electron neutrino ( @xmath75 ) at the inner most region of the accretion disk is estimated as @xcite @xmath76 where @xmath77 is the avogadro s number , @xmath78 is the cross section of the neutrino off a nucleon , @xmath79 is the density normalized by 10@xmath80 g @xmath81 .
therefore , @xmath75 may comparably become unity if @xmath2 is approximately @xmath82 s@xmath5 , @xmath83 , and @xmath84 . from the above considerations , models with high accretion rate ( @xmath85 s@xmath5 )
may be the extreme cases . in fig .
[ fig4 ] , we show the average neutrino luminosity which is defined by @xmath86 for comparison , in the figure we also show the result of popham et al . ( 1999 ) .
the uncertainty of their results comes from the kerr parameter of the central black hole ( @xmath87 0.5 ) . from the figure
, we find that our simple formula reproduce their results fairly well in wide range of mass - accretion rate . at high mass - accretion rate ( @xmath88 s@xmath5 ) ,
our formula may overestimate the neutrino luminosity by a factor of three .
as shown below ( see section [ results ] ) , however , we also compute the grb neutrino background in such models as extreme cases . in future
, improvements of the accretion model will enhance the predictability of the flux of grb neutrino background .
the event rates of @xmath22 expected in water cherenkov detectors is represented by @xmath89 where @xmath90 is the energy of the positron which is scattered through @xmath91 in the detector , @xmath92 is the cross section of the process , and @xmath93 is the distance from the earth to the collapsar .
@xmath94 is the number of proton in the detector , e.g. , @xmath95 for sk .
thus the total event rate is estimated by the integration , @xmath96 it is reported that the energy deposition rate becomes as large as @xmath97 erg s@xmath5 by the @xmath98 pair annihilation @xcite . in their studies , they computed the energy deposition rate considering the shape of the accretion disk .
their results are shown in fig .
[ fig6 ] as a shaded region . here
the efficiency of @xmath99 pair annihilations is defined as @xmath100 , where @xmath101 is the total deposited energy .
their results are fitted well as @xmath102 which is also shown in fig . [ fig6 ] . in this study
, we use this fitting formula to estimate the explosion energy of a grb .
then , we obtain the total deposited energy by @xmath103 = @xmath104 . in fig .
[ qtotqee ] , the total energy of emitted neutrinos ( @xmath105 ) and the total deposited energy ( @xmath101 ) are shown as a function of the mass - accretion rate .
since the total explosion energy of sn1998bw is reported to be @xmath106 erg @xcite , we find that it corresponds to the mass - accretion rate @xmath107 in fig .
[ qtotqee ] . for simplicity , we ignored the heating process from @xmath108 and @xmath109 pair annihilations in this computation because the effect is much smaller than @xmath110 pair annihilation .
the reason is as follows . in the electromagnetic thermal bath , @xmath108 and @xmath109 pairs
are produced from @xmath111 pair annihilations only through the neutral current interaction .
in addition , @xmath112 and @xmath109 pairs annihilate into @xmath111 pairs only through the neutral current interaction . on the other hand @xmath110 pairs
are produced through both the charged and neutral currents and annihilate into @xmath111 pairs correspondingly through the same process .
then , compared with the case of @xmath110 , the efficiency of the heating from @xmath112 and @xmath109 pairs in the electromagnetic thermal plasma is just within @xmath113 of the total , even if the @xmath114 pair production from @xmath111 pair annihilation dominates all the other processes of neutrino production . in order to get the differential number flux of back ground neutrinos ,
first we compute the present number density of the background neutrinos per unit neutrino energy , @xmath115 in eq .
( [ eq : dnde ] ) .
the contribution of the neutrinos emitted in the interval of the redshift @xmath116 is given as @xmath117 where @xmath118 .
it is noted that @xmath14 is grb rate per comoving volume , and the effect of the expansion of the universe is taken into account in eq .
( [ eq : a ] ) . the friedmann equation gives the relation between @xmath59 and @xmath28 as follows : @xmath119 we now obtain the differential number flux of grb , @xmath120 , by using the relation @xmath121 : @xmath122
the differential number flux of grb neutrinos is shown in figs .
[ flux1 ] , [ flux2 ] and [ flux3 ] . as for the grb formation history , grbr1 , grbr2 , and grbr3 are assumed , respectively .
it is found that the differential number flux is similar with each other even if the different star formation history is adopted .
that is because the energy of neutrinos from high - z grbs is redshifted as @xmath123 , and the volume containing high - z grbs is relatively small @xmath124 as given in eq .
( [ eq : c ] ) .
thus , we show the results only for grbr3 below . on the other hand
, we can see clearly that the differential number flux becomes larger as the mass - accretion rate increases .
in particular , in high accretion model such as @xmath125 , the intensity of the grb neutrino background is larger than that of sn neutrino background in the energy region @xmath1mev . as discussed below , in such a high energy region , the neutrino background will be obscured by the noise from the atmospheric neutrino background ( anb ) .
as the event number increases , however , we will have statistically sufficient data , and the signals from grbs will dominate the statistical error of anb .
in the next section we estimate the time required to detect the signals from grb ( see also table [ tab : table2 ] ) .
note that at the energy which is larger than @xmath126mev , the signals from grbs dominate the noise of electrons from the decay of invisible muons because their spectrum has a steep cut - off at @xmath126mev . in the next section
, we will discuss them in detail .
the event rates at sk are shown in fig .
[ event1 ] in units of yr@xmath5 mev@xmath5 for various mass - accretion rates . for event rates , increasingly we can not discriminate among the three models of the star formation history because the cross section of @xmath127 is proportional to the square of the neutrino energy ( @xmath128 ) , and the relative contribution of low energy neutrinos to the event rate is small .
therefore , irrespective of the differences among the models of the star formation history , we can expect the event rate of grb neutrino background as a function of mass - accretion rate .
namely the informations on the average mass - accretion rate will be obtained from such an observation , which , in turn , should give us informations on the total explosion energy of a grb ( see fig .
[ qtotqee ] ) .
they are , of course , never obtained only by the gamma - ray observations . in fig .
[ event2 ] , we show the dependence of the event rates on the probability @xmath6 that one collapsar generates a grb in the case of @xmath129 s@xmath5 . it is clear that the signals from grbs dominate those from sn for @xmath130mev in the case of @xmath131 .
however , the noise from the atmospheric neutrinos obscures the signals of grbs even if we adopt the high value of @xmath6 ( = 1.0 ) . in the next section ,
we discuss the detectability of the signals of the grb neutrino background as a significant deviation from the statistical error of anb . to estimate the event rates more realistically , we should consider the effects of neutrino oscillation .
the flavor eigen states @xmath132 ( @xmath133 and @xmath75 ) are related with the mass eigen states @xmath134 ( `` @xmath135 '' = 1 , 2 , and 3 ) by the mixing matrix @xmath136 as @xmath137 .
then , we parameterize it by @xmath138 with @xmath139 and @xmath140 , where we ignored the cp phase in the lepton sector . in general , the probability of the vacuum oscillation from a flavor @xmath132 to a flavor @xmath141 is represented by @xmath142 where @xmath143 is the distance from the source to the earth (= @xmath144 ) , @xmath145 is the neutrino energy , and the mass difference is defined by @xmath146 . for simplicity , here we assume the normal hierarchy of neutrino masses , @xmath147 .
the results of recent oscillation experiments are as follows .
the solar neutrino data favor the large mixing angle solution ( @xmath148 ) , and @xmath149 at 3@xmath78 c.l .
. the atmospheric neutrino data also imply the large mixing ( @xmath150 ) , and @xmath151 at 99 @xmath152 c.l .
the chooz reactor experiment shows that @xmath153 at 95 @xmath152 c.l @xcite .
then , the above data tells us that @xmath154 . adopting the above data
, we can assume that the mixing matrix is approximately simplified into @xmath155 because neutrinos propagate for long distances , the second term in eq .
( [ eq : prob1 ] ) is averaged and becomes zero .
then , we have @xmath156 if we choose the inverted hierarchy , i.e. , @xmath157 , we only interchange the roles of @xmath158 and @xmath159 in eq . ( [ eq : mixing_gen ] ) . namely we only interchange the first and third columns of @xmath136 .
even then , it is remarkable that the results in eq .
( [ eq : nuebar1 ] ) , ( [ eq : nuebar2 ] ) , and ( [ eq : nuebar3 ] ) are not changed at all . in fig .
[ osc ] , we show the event rate as a function of @xmath38 at sk in which we consider the effects of the neutrino oscillation for @xmath160 . it is noted that the flux of @xmath22 becomes lower as a result of the neutrino oscillation .
that is because @xmath161 and @xmath162 are produced only through the @xmath163 pair annihilation .
we consider the detectability of the neutrino background from grbs . in the energy range @xmath164mev ,
the neutrino background is obscured by the events of the electrons which come from the decay of invisible muons @xcite .
the spectrum of the electrons is given by @xmath165 where @xmath166 is the muon mass , and @xmath167 is the electron ( positron ) energy .
the event rate of the electrons from the invisible muons at sk is about @xmath0 1 in units of yr@xmath5 mev@xmath5 in the energy range @xmath168mev @xcite .
thus , signals of grb neutrino background are smaller than those of electrons from invisible muons in the range . in addition
, the signals from sn neutrino background will basically dominate those of grb neutrino background in the energy range . on the other hand
, the spectrum of electrons from the invisible muons has a steep cut - off at about 50 mev because the maximum energy of the emitted electrons is about half of the muon mass .
thus , as mentioned above , the signal from grbs becomes larger than the noise from invisible muons in the energy range @xmath1mev . in the case of the high mass - accretion rate such as @xmath169 ,
the signals of the grb neutrino background are larger than those of sn neutrino background in the energy region .
however , the severest noise comes from the atmospheric neutrinos . in table
[ tab : table0 ] , we show the event rate ( yr@xmath5 ) of anb and its statistical error in sk / titand in the energy regions @xmath170 = ( 50 60)mev and
@xmath171 50mev @xcite .
titand is proposed as a next - generation plan of the multi - megaton water cherenkov detector @xcite .
the first phase of titand is planed to be with 2 mt water inside .
thus , the event rate becomes approximately one hundred times larger than sk . for comparison
, we show the event rates from the grb neutrino background at sk / titand in table [ tab : table1 ] for various models .
the third column of the table represents whether we consider the effects of the neutrino oscillation or not .
yes means that the effects of the neutrino oscillation are taken into account .
it is apparent that the signals from grb are smaller than the noise from anb .
as the event number increases , however , we will have statistically sufficient data , and the signals from grbs will dominate the statistical error of anb .
namely , then we can definitely detect the signals of the grb neutrino background as a significant deviation from the statistical error of anb .
the time ( yr ) required to marginally distinguish the signals from the noise is computed by @xmath172 , where @xmath173 is the noise ( yr@xmath5 ) from anb and @xmath174 is the signals from grbs ( yr@xmath5 ) . in table
[ tab : table2 ] , we show the time required to detect the signal of grbs at sk / titand for the energy range @xmath175 = ( 50 60 ) mev and @xmath176 mev . from the table , we find that we can detect the signals within @xmath010 years at titand in the case that the mass - accretion rate is ( 5
10@xmath4 s@xmath5 ) and @xmath6 is ( 0.5 1 ) .
although it seems to take a long time to detect the signals from grbs for the cases in which the conservative parameter sets are adopted , constraints for optimistic parameter sets will be obtained by the observations of titand .
of course , a larger water cherenkov detector than titand will detect the signals from grbs more immediately even if the grbs obey the conservative parameters . in this study
, we have mainly assumed that one tenth of collapsars whose mass range are in ( 35 125)@xmath4 are accompanied with grbs .
here we investigate the validity of the assumption .
the locally observed value for the sn rate is in the range @xmath177 yr@xmath5 mpc@xmath178 , where @xmath179(100 km s@xmath5 mpc@xmath5 ) @xcite . on the other hand
, it is reported that the grb rate at @xmath180 is @xmath181 yr@xmath5 mpc@xmath178 using the batse data catalog @xcite . on the other hand , we assumed that the ratio of the collapsar rate to sn rate is @xmath182 using eqs .
( [ eq:4 ] ) and ( [ eq:4.2 ] ) .
thus , we estimate that the collapsar rate at @xmath180 is @xmath183 yr@xmath5 mpc@xmath178 . adopting the beaming factor
is @xmath184 , and the probability that one collapsar generates a grb is @xmath185 , we estimate the grb rate at @xmath180 : @xmath186 in units of @xmath187 which is consistent with the previous work . at present , the average beaming angle is not determined precisely except for some observations that suggest it is several degree @xcite .
therefore , we leave the value of @xmath6 as a free parameter at present . in addition , the previous works on the grb formation rate would have a lot of uncertainties since only the batse data catalog is used in these analyses , and the average luminosity of grbs is adopted to fit the model .
more precise data on the beaming angle and on the grb formation rate ( for example , by using the peak - lag relation @xcite ) will give us informations on the probability @xmath6 .
then , our predictions of grb neutrino background become more correct . in this study , we have assumed that grb formation rate is proportional to the star formation rate in the universe .
if this assumption is not correct , of course , the resulting energy spectrum will also be changed .
therefore , on the contrary , when we observe the differential flux of grb neutrino background , we will also be able to know whether the grb formation history traces the star formation history or not .
we used the model presented by popham et al .
( 1999 ) as an accretion model .
we have to investigate the model dependence of the neutrino luminosity in future .
in particular , we have to investigate the neutrino - dominated accretion disk ( ndaf ) for the high accretion rate cases @xcite . in this study
, we have considered the effects of vacuum neutrino oscillations . however , the effects of msw effects are not taken into account .
as long as the collapsar model is adopted , the density profile of the progenitor should depend on the zenith angle and time .
we are planning to investigate these effects in the near future . when the mechanism of the central engine is not the neutrino heating but the blandford - znajek mechanism @xcite , the observations of grb neutrino background will give a lower limit of the total explosion energy of a grb .
we have estimated the flux of the grb neutrino background and computed the event rate at sk / titand in the collapsar models , assuming that grb formation history traces the star formation history .
we have found that the flux and the event rate depend sensitively on the mass - accretion rate although the detection of signals from grbs seems to be difficult by sk . on the other hand , however , we have found that the grb neutrino background may be detected by titand within @xmath0 10 yrs as long as the average mass - accretion rate is high ( @xmath188 a few @xmath4 s@xmath5 ) , and the probability that one collapsar generates a grb is high ( @xmath6 = 0.5 1.0 ) .
thus , we conclude that the informations on the mass - accretion rate of collapsar will be obtained by the observations of the grb neutrino background , which in turn , should give us informations on the total explosion energy of grbs in future .
they are , of course , never obtained only by the gamma - ray observations .
although there are some simplifications in our analyses and some uncertainties on the present observations , our proposal to determine the total explosion energy of a grb is very challenging and interesting .
this paper is the maiden attempt about the grb neutrino background .
we will revise our model to enhance the predictability , and our suggestions will be tested by the future observations .
is grateful to r. narayan for useful comments .
k.k . is grateful to t. onogi for useful comments .
s.n . and k.k .
thank the yukawa institute for theoretical physics at kyoto university , where this work was initiated during the yitp - w-01 - 06 on grb2001 and completed during the yitp - w-99 - 99 on blackholes , gravitational lens , and gamma - ray bursts. the authors are grateful to y. totsuka , t.suzuki , m. nakahata , y. fukuda , and a. suzuki for useful comments on the neutrino background at super - kamiokande .
this research has been supported in part by a grant - in - aid for the center - of - excellence ( coe ) research ( 07ce2002 ) and for the scientific research fund ( 199908802 ) of the ministry of education , culture , sports , science and technology and by japan society for the promotion of science postdoctoral fellowships for the research abroad .
t. j. galama , p. m. vreeswijk , j. van paradijs , c. kouveliotou , t. augusteijn , h. bohnhardt , j. p. brewer , v. doublier et al .
nature * 395 * , 670 ( 1998 ) ; d. e. reichart astrophys .
j. * 521 * , 111 ( 1999 ) ; j. s. bloom , s. r. kulkarni , s. g. djorgovski , a. c. eichelberger et al .
nature * 401 * , 453 ( 1999 ) | we estimate the flux of the grb neutrino background and compute the event rate at sk and titand in the collapsar model , assuming that grb formation rate is proportional to the star formation rate .
we find that the flux and the event rate depend sensitively on the mass - accretion rate .
although the detection of signals from grbs seems to be difficult by sk , we find that we can detect them by titand ( optimistically @xmath0 20 events per year in the energy range @xmath1mev ) .
therefore , we show that we will obtain the informations on the mass - accretion rate of collapsar by observing the grb neutrino background , which , in turn , should give us informations on the total explosion energy of grbs . |
the studies of stable spatial solitons is an important problem of contemporary physics @xcite .
there are numerous manifestations of stable solitonic solutions of nonlinear equations in two and three spatial dimensions in nonlinear optics @xcite , solid state @xcite and plasma physics @xcite . in plasma physics it was established @xcite that the nonlinear electron - ion interaction leads to the modulation instability and results in the collapse of a langmiur wave @xcite .
in contrast to the electron - ion interaction , the nonlinear electron - electron interactions , studied in refs .
@xcite , were shown to stabilize the evolution of a langmuir wave packet , making possible the existence of stable spatial plasma structures .
it is convenient to describe the evolution of nonlinear langmuir wave packets on the basis of a nonlinear schrdinger equation ( nlse ) @xcite .
the nonlinear terms studied in refs .
@xcite are local since their contributions to nlse contain only a certain power of the electric field amplitude .
the nonlocal terms in nlse , derived in refs .
@xcite , are also important when , e.g. , a wave packet is very steep . under certain conditions these nonlocal electron nonlinearities
can arrest the lamgmuir collapse .
it is suggested in ref .
@xcite that a nonlocal nlse is a theoretical model for stable spatial plasma structures obtained in a laboratory @xcite . besides the nonlocal electron nonlinearities taken into account in ref .
@xcite there are analogous contributions to nlse originating from the electron pressure term .
these terms have the same order of magnitude as those considered in ref .
@xcite and are important for plasma with nonzero electron temperature .
note that nonlinear waves in warm plasma were also studied in ref .
@xcite . in the present work
we carefully study these additional nonlinearities and examine their contribution to nlse .
this paper is organized as follows . in sec .
[ dyn ] , on the basis of the system of nonlinear plasma equations , we examine electrostatic plasma waves having axial and spherical symmetry .
then , in sec . [ locnon ] , we briefly review the previous studies of the influence of electron nonlinearities on the dynamics of langmuir waves in plasma .
the new nlse , describing stable spatial plasma structures , taking into account the nonlocal nonlinear terms is derived in sec .
[ nonlocnon ] . in sec .
[ num ] we analyze solutions of this nlse numerically .
we consider the possible application of our results to the theoretical explanation of the existence of atmospheric and ionospheric plasmoids in sec .
finally we summarize our results in sec .
[ concl ] .
the description of plasma waves in frames of lagrange variables is presented in appendix [ lagrange ] .
in this section we study axially and spherically symmetric waves in plasma accounting for local and nonlocal electron nonlinearities .
we derive a new nlse which is shown to have solitonic solutions . to describe electrostatic waves in isotropic warm plasma , in which the magnetic field is equal to zero , @xmath0 , we start from the system of nonlinear hydrodynamic equations , @xmath1 where @xmath2 are the densities of electrons and ions , @xmath3 are their velocities , @xmath4 is the amplitude of the electric field , @xmath5 is the electron mass , @xmath6 is the proton charge , and @xmath7 is the pressure tensor which is calculated using the electron distribution function @xmath8 .
note that eq
. follows from more general vlasov kinetic equation for the function @xmath8 .
it is known that eq .
allows small amplitude langmuir oscillations on the frequency @xmath9 , where @xmath10 is the unperturbed electron density .
however , since eq . is nonlinear , the higher harmonics generation is possible
. therefore we can look for the solution of eq . in the following form : @xmath11 where we separate different time scales , @xmath12 the functions @xmath13 , @xmath14 , and @xmath15 in eq . and the amplitude functions @xmath16 ,
@xmath17 and @xmath18 in eq .
are supposed to vary slowly on the @xmath19 time scale .
moreover we suggest that , e.g. , @xmath20 etc , i.e. slowly varying functions , marked with index @xmath21 " , and the amplitudes of the second harmonic are much smaller than the corresponding amplitudes of the main oscillation .
we neglect the velocity of ions in eq . and suggest that ion density is represented as @xmath22 , where the perturbation @xmath23 is also a slowly varying function on the @xmath19 time scale .
note that @xmath23 does not necessarily coincide with @xmath13 .
we will study electrostatic plasma oscillations in two or three dimensions .
thus we assume radially symmetric quantities in eq . , @xmath24 where @xmath25 is the radial coordinate and @xmath26 is the basis vector in spherical or cylindrical coordinate system .
the contribution of electron nonlinearities to the evolution of langmiur waves in frames of the model - was taken into account in refs .
@xcite and the following equation for the description of the main oscillation amplitude @xmath27 was obtained : @xmath28 where @xmath29 is the debye length , @xmath30 is the electron temperature , and @xmath31 , @xmath32 is the dimension of space .
eq . should be supplied with the wave equation for the ion motion @xcite , @xmath33 where @xmath34 is the sound velocity , @xmath35 is the ions temperature , @xmath36 is the heat capacity ratio for ions , @xmath37 is the ion mass , and @xmath38 is the laplace operator . at the absence of electron nonlinearities [ the last term in eq . ]
the system and corresponds to the zakharov equations @xcite .
it should be noted that the contribution of local electron nonlinearities is washed out from eq . in one dimensional case @xmath39 .
the zakharov equations are known to reveal the collapse of a langmuir wave packet @xcite : the size of a wave packet is contracting and the amplitude of the electric field is growing .
it was shown in refs .
@xcite that langmuir collapse can be arrested and stable spatial plasma structures can appear in two and three dimensions , @xmath40 , since the second nonlinear term in eq . is defocusing .
the influence of electron nonlinearities on the dynamics of a langmuir collapse was further studied in ref .
@xcite . using the relation between slowly varying electron density @xmath13 and the perturbation of ion density @xmath23 , @xmath41 which was obtained in ref .
@xcite , we can approximately find @xmath13 as @xmath42 note that the last term in eq .
, @xmath43 , which is important at rapidly varying ion density , was omitted in refs .
@xcite . using eqs . , and
supposing that @xmath44 , one obtains the new nonlinear term in the left hand side of eq .
( see ref .
@xcite ) , @xmath45 this new contribution was shown in ref .
@xcite to arrest the langmuir collapse .
moreover the nonlocal nonlinearity is more effective in preventing the collapse compared to that found in refs . @xcite .
it should be also noted that the new nonlinear term predicted in ref .
@xcite does not disappear in one dimensional case .
the nonlinearity analogous to that in eq .
, @xmath46 , appears in the zakharov equations with quantum effects @xcite , while taking into account the quantum bohm potential .
the nonlocal electron nonlinearities , analogous to that studied in ref .
@xcite , can follow not only from eq . .
we can consider the contribution of the slowly varying electron density to the electron pressure .
analogously to eqs . and
one can discuss the decomposition of the distribution function @xmath8 proposed in ref .
@xcite , @xmath47 where @xmath48 is the equilibrium distribution function . for classical plasma
it can be , e.g. , a maxwell distribution corresponding to @xmath30 .
all the quantities in the expansion series depend on @xmath49 . using the result of ref .
@xcite we represent @xmath50 as , @xmath51 + \dotsb,\end{aligned}\ ] ] where we drop terms containing @xmath52 , @xmath53 , and higher power of the first harmonic amplitudes .
the function @xmath54 in eq . was also found in ref .
@xcite for the case of isotropic equilibrium distribution , @xmath55 where @xmath56 is the thermal velocity of electrons .
the terms which do not contain the slowly varying density @xmath13 are omitted in eq . .
now we can express the nonlinear term @xmath57 in eq . as @xmath58 e^{-\mathrm{i } \omega_p t}.\end{aligned}\ ] ] using the maxwell equilibrium distribution function
, @xmath59 normalized on the unperturbed electron density , we can express the gradient of pressure in eq . in the following form
: @xmath60,\end{aligned}\ ] ] where we use the relations between the amplitudes of the main harmonic , @xmath61 established in ref .
@xcite . the leading term in eq . was derived in refs .
the next - to - leading terms , proportional to the derivatives of @xmath13 , are important when one has rapidly varying in space wave packets . using eqs . and we can obtain the generalization of eq . which takes into account the nonlocal electron nonlineariries due to the interaction of a langmuir wave with the low - frequency perturbation of electron density , @xmath62 - \nabla_i ( n \nabla_i \mathbf{e } )
\notag \\
& - \nabla_i ( n \nabla e_{i } ) - \nabla [ ( \mathbf{e } \cdot \nabla ) n ] + ( \nabla \cdot \mathbf{e } ) \nabla n \notag \\ & + 3 n \nabla ( \nabla \cdot \mathbf{e } ) \big ] = 0,\end{aligned}\ ] ] where for simplicity we omit the index @xmath63 " : @xmath64 .
eq . should be supplied with eq .
governing the evolution of ion density perturbation @xmath23 .
the nonlocal term @xmath65 in eq . was derived in ref .
the remaining nonlocal nonlinearities , which are of the same order of magnitude as the term @xmath65 , were omitted in that work .
let us study the evolution of the system and in the subsonic regime , when one can neglect the second time derivative of the ion density in eq . .
considering axially symmetric , @xmath66 , or spherically symmetric , @xmath67 , cases [ see eq . ] and using the dimensionless variables , @xmath68 we can represent the dynamics of the system and in the form of a single nlse , @xmath69 - \phi |\phi|^2 + \beta(d ) \frac{\phi |\phi|^2}{x^2 } \notag \\ & - \frac{3}{2 } \left [ \phi \frac{d-1}{x } - 3 \frac{\partial \phi}{\partial x } \right ] \frac{\partial}{\partial x } |\phi|^2 = 0.\end{aligned}\ ] ] it should be noted that eqs . or does not conserve the number of plasmons , @xmath70 this fact is because of the presence of the term @xmath71 in eq . .
to analyze the integrals of eq .
let us separate the variables in eq .
: @xmath72 . then making the following nonlinear gauge transformation of the nonhermitian " eq . : @xmath73 we can cast it in the equivalent form , @xmath74 - \psi |\psi|^2 + \beta(d ) \frac{\psi |\psi|^2}{x^2 } \notag \\ & - \frac{3}{2 } \left [ \frac{3}{2 } \delta_x |\psi|^2 + \frac{d-1}{x } \frac{\partial |\psi|^2}{\partial x } \right]\psi,\end{aligned}\ ] ] where @xmath75 is the radial part of the laplace operator . to derive eq .
we use the decomposition of the transformed wave function " squared , @xmath76 , and keep only cubic terms in eq . .
the modified nlse is hermitian " , i.e. it conserves the number of plasmons , @xmath77 where we restore the @xmath21 dependence , @xmath78 , and @xmath79 , for @xmath80 , or @xmath81 , for @xmath82 , is the solid angle
. it should be also noticed that the nonlocal nonlinearities does not disappear in one dimensional case as the local ones studied in ref .
the dynamics of langmuir waves in warm plasma in arbitrary dimensions is analyzed in appendix [ lagrange ] , using lagrange variables .
it is shown there that in case of nonzero electron temperature the nonlinear terms does not disappear in higher dimensions @xmath83 .
we can notice that eq . is analogous to nlse equation derived in ref .
it also contains @xmath84 term , although the coefficient is different .
the main discrepancy is the presence of the term @xmath85 . in sec .
[ num ] we analyze the influence of this new contribution numerically .
it is difficult to construct other conserved integrals of eq . ,
e.g. , a hamiltonian , independent of the number of plasmons .
therefore one has to analyze the behaviour of solutions of this equation numerically .
eq . should be supplied with the boundary conditions , @xmath86 , and thus treated as a boundary condition problem .
we have found numerical solutions of this problem using a boundary condition problem solver , incorporated in the matlab 7.6 program .
it requires an initial guess " function which was chosen as @xmath87 where @xmath88 is the width " of the function , @xmath89 , for @xmath80 , and @xmath90 , for @xmath82 .
the function is now normalized on the initial number of plasmons @xmath91 , to be compares with the actual number of plasmons @xmath92 obtained from a numerical solution .
the best convergence of the numerical procedure is achieved when @xmath88 corresponds to the minimal value of @xmath93 in eq . at the given @xmath91 .
this analysis is analogous to the trial function method @xcite for minimizing of the hamiltonian of nlse . in fig . [
2dfig](a ) , for @xmath80 , and in fig .
[ 3dfig](a ) , for @xmath82 , we present the dependence of @xmath93 versus @xmath88 for different values of @xmath91 .
one can see that the function @xmath94 has a minimum if @xmath95 .
in two dimensional case the critical number of plasmons can be easily found : @xmath96 .
note that one can expect the stability of a soliton with respect to the collapse if @xmath97 for @xmath80 , whereas in 3d case there is some range of positive @xmath93 which corresponds to uncollapsing solutions .
the solutions which correspond to various values of @xmath93 , @xmath88 , and @xmath92 are shown in fig .
[ 2dfig](c , d ) , for @xmath98 , and in fig .
[ 3dfig](c , d ) , for @xmath82 .
a guess " function which does not correspond to the minimum of the function @xmath94 also gives some solitonic solution of eq .
, however the convergence is much worse than in the minimal @xmath93 case .
indeed , @xmath92 significantly differs from @xmath91 for such a solution .
when the deviation from the minimal @xmath93 is big , although @xmath93 remains to be negative , no regular solutions can be found and the system capsizes into chaos .
thus these nonoptimal " solutions seem to be unstable . to analyze the stability of the found solutions we present the @xmath99 dependence in fig .
[ 2dfig](b ) , for @xmath80 , and in fig .
[ 3dfig](b ) , for @xmath82 .
note these curves were built for guess " functions corresponding to a minimal @xmath93 . applying the vakhitov
kolokolov criterion @xcite to the results shown on these plots one can conclude that the presented solutions are stable in 2d case , whereas some unstable solitons can exist in three dimensions . for @xmath82
, a stable solution can be generated starting from a threshold plasmon number @xmath100 and at @xmath101 ( see also the discussion in sec . [ appl ] ) .
we show an example of a unstable soliton in fig . [ unstable ] .
an example of the unstable solution of eq .
[ left - handed branch in fig .
[ 3dfig](b ) ] obtained for @xmath102 .
it corresponds to @xmath103 , @xmath104 , and @xmath105 . ] in fig . [
2dfig](b ) , for @xmath80 , one can see that the new term in eq . ,
@xmath106(b ) , the difference is just quantitative : the critical plasmon number and critical frequency are shifted .
therefore our results are in agreement with ref .
@xcite where nlse with a nonlocal term was analyzed .
one should also notice that the numerical curve in fig .
[ 2dfig](b ) ( solid line ) is in a good agreement with the analytical " @xmath107 dependence ( dashed line ) , @xmath108 , found from eq . with help of the guess " function .
stable spatial solitons , involving nonlocal nonlinearities , similar to plasma structures described in the present work , were reported to be obtained in various laboratory experiments in plasma physics , condescended matter , and nonlinear optics ( see , e.g. , reviews @xcite and references therein ) . another opportunity for the physical realization of plasmoids described in secs .
[ dyn ] and [ num ] consists in their implementation as a rare natural atmospheric electricity phenomenon called a ball lightning ( bl ) @xcite . according to the bl observations ,
most frequently it has a form of a rather regular sphere with the diameter of @xmath109 @xcite .
however big bls with the size more than one meter were reported to exist @xcite .
besides spherical bl , snake - like objects were observed @xcite .
the lifetime of bl can be up to several minutes @xcite .
the estimates of energy of bl were obtained only in cases when it produced some destruction while disappearing .
these estimates give for the energy of bl the value in the range ( several kj several mj ) @xcite .
however , since in many cases bl disappears just fading , one should assume that a very low energy bl can exist .
numerous bl models are reviewed in ref .
despite very exotic theoretical descriptions of bl were proposed , it is most probable that this object is a plasma based phenomenon . in refs . @xcite we developed a bl model based on radial oscillations of electron gas in plasma , studying these oscillations in both classical and quantum approaches .
the present work is a development of our bl model . as in our previous studies ,
here we also treat bl on the basis of radial oscillations of electrons . however , as we show in secs .
[ dyn ] and [ num ] , local and nonlocal electron nonlinearities as well as the interaction of electrons with ions play an important role for the stability of an axially or spherically symmetric plasmoid .
for example , the inhomogeneity of ion density was not accounted for in ref .
it should be noted that the idea that spherically symmetric oscillations of electrons underlie the bl phenomenon was independently put forward in ref .
@xcite . in fig .
[ blprop ] we present the characteristics of axially symmetric ( snake - like bl @xcite ) and spherically symmetric ( spherical bl ) plasmoids calculated on the basis of the results of sec .
[ num ] .
the total energy of electric field inside a plasmoid ( energy density in 2d case ) and the effective plasmoid radius are defined as @xmath110 the frequency of the total oscillation , including the main harmonic oscillation which was separated in the derivation of eq . , is @xmath111 , where @xmath112 is the langmuir frequency .
here we take into account that frequency shift @xmath113 for a stable plasmoid is negative [ see figs . [ 2dfig](a ) and [ 3dfig](a ) ] .
one can see in fig .
[ blprop ] that in frames of our bl model we predict the existence of low energy ( @xmath114 in 2d case and @xmath115 in 3d case ) plasma structures with the typical size @xmath116 .
this kind of plasmoids should appear in low density @xmath117 and hot @xmath118 plasma .
such a density of plasma can well exist in the earth ionosphere @xcite . a linear lightning can provide the plasma heating up to @xmath119 since , e.g. , @xmath120 in the lightning channel @xcite .
thus we can assume that such high temperatures of plasma can be present in some localized area of ionosphere during a thunderstorm , making possible the existence of plasma structures described in the present work .
it is interesting to notice that several meters objects were reported to appear in the bl observations near airplanes @xcite . to demonstrate the stability of the described plasmoids with respect to decay , in fig . [ stability ]
we present the ratio @xmath121 versus @xmath122 , where @xmath123 is the total thermal energy of a plasmoid ( energy density in 2d case ) and @xmath124 is the effective volume of a plasmoid , which is equal to @xmath125 for @xmath66 and to @xmath126 for @xmath67 . indeed , hot electrons with @xmath119 could just escape the plasmoid volume making it unstable .
one can however see in fig .
[ stability ] that for stable plasma structures the ratio @xmath121 has a tendency to increase reaching the unit value at a certain soliton energy ( @xmath127 for @xmath66 and @xmath128 for @xmath67 ) .
it means that hot electrons will not escape form the plasmoid volume as soon its energy has this critical value .
it should be noticed that an unstable plasmoid , corresponding to the lower branch in fig .
[ stability](b ) , will lose electrons since @xmath129 always . comparing the predicted plasmoid properties with the characteristics of circumterrestrial ( not ionospheric ) bl
one can say that our model is not directly applicable for the description of such an object since its energy @xmath130 and the size @xmath131 are beyond our predictions .
we can however use our model at the initial stages of the plasmoid formation when the amplitude of the electric field is so high since in eq .
we keep only cubic nonlinearity .
higher nonlinear terms , which seem to be important for denser plasma , can explain smaller size and bigger energy content .
a natural plasmoid appearing in circumterrestrial atmosphere is surrounded by the neutral gas .
it means that plasma in the interior of a plasmoid should be maintained in the state with proper ionization during its lifetime .
it is however known @xcite that plasma of a low energy plasmoid , without an internal energy source , will lose energy and recombine back to a neutral gas in the millisecond time scale at the atmospheric pressure .
thus the lifetime of such an object will be extremely short .
it was suggested that under certain conditions plasma can reveal superconducting properties @xcite .
this mechanism could prevent the energy losses and thus the recombination , providing the long lifetime of a low energy plasmoid .
in conclusion we mention that in the present work we have studied axially and spherically symmetric langmiur solitons in warm plasma . in sec .
[ dyn ] we started with the discussion of radially symmetric oscillations of electrons in plasma and then separated the motions on different time scales . in sec .
[ locnon ] we briefly reviewed the previous works on the theory of stable plasma structures which involved local electron nonlinearities .
then , in sec . [ nonlocnon ] we have calculated the contribution of the slowly varying electron density to the electron pressure and derived , together with the results of ref .
@xcite , the new nlse which accounts for nonlocal electron nonlineariries .
these additional nonlinear terms do not disappear in one dimensional case .
this fact was also demonstrated in appendix [ lagrange ] using lagrange variables .
the solutions of this new nlse , rewritten in the equivalent form for the radially symmetric case , have been analyzed numerically in sec .
we presented the examples of some of the solutions of eq .
[ see fig .
[ 2dfig](c , d ) and fig .
[ 3dfig](c , d ) ] and analyzed their stability .
it has been found that solutions in 2d case are stable , whereas in tree dimensions some unstable solitons can exist .
we also compared out results with ref .
@xcite , where analogous nlse was considered . in sec .
[ appl ] we suggested that the described axially and spherically symmetric solitons can be realized in the form of a natural plasmoid , a ball lightning .
we have computed the characteristics of such a plasma structure , like energy and radius ( see fig . [ blprop ] ) , predicted in frames of our model . comparing these characteristics with the properties of a natural plasmoid
one concludes that our model describes a low energy bl which can exist in hot ionospheric plasma
. note that in many cases bls were observed form airplanes @xcite at high altitudes .
we also examined the stability of plasmoids with respect to decay because of the escape of hot electrons .
we have already considered a bl model based on radial oscillations of electrons in our previous publications @xcite . in the present work we performed more detailed analysis of electron nonlinearities and showed that they play a crucial role for the stability of a plasmoid .
although plasma structures described in the present work does not reproduce all the properties of circumterrestrial ( not ionospheric ) bl @xcite , we can use our results for the description of the bl formation , when the amplitude of the electric field is not so big . as we have found in sec .
[ num ] [ see also fig .
[ blprop](a , c ) ] , the total frequency of electron oscillations is less than langmuir frequency , since the frequency shift @xmath113 is negative .
it is important fact for the future experimental studies of bl .
for example , the electron density in a linear lightning discharge can be @xmath132 @xcite , giving for the plasma frequency a huge value of @xmath133 . in a laboratory
it is extremely difficult to create a strong electromagnetic field of such a frequency to generate bl .
however , if the frequency of electron oscillations has a tendency to decrease , it can facilitate the plasmoid generation .
this work has been supported by conicyt ( chile ) , programa bicentenario psd-91 - 2006 .
the author is thankful to s. i. dvornikov , e. a. kuznetsov , l. stodolsky , and a. a. sukhorukov for helpful discussions and to deutscher akademischer austausch dienst for a grant .
the propagation of waves in plasma was analyzed in sec .
[ locnon ] using euler variables which are more convenient for the practical purposes
. there is , however , a lagrange approach for the treatment of plasma waves . instead of describing of plasma characteristics , like velocity , density etc , in a certain point of space , one can study the dependence of these quantities on the initial coordinates of plasma particles .
let us change the variables @xmath134 in eqs . and as @xmath135 where @xmath136 is the initial coordinate of an electron , @xmath137 is deviation of an electron from the equilibrium , and @xmath138 is the new temporal variable .
an electron is supposed to be in the equilibrium initially , @xmath139 . in eq .
we use the definition of electron velocity , @xmath140 . unlike the euler picture ,
the continuity equation in lagrange variables can be integrated and it does not contain the time derivative , @xmath141 where @xmath142 is the initial ( unperturbed ) electron density , which is supposed to be uniform . using eqs . and we can obtain from eq . a single nonlinear equation for electron velocity , @xmath143 = 0,\end{aligned}\ ] ] where a dot " and a prime " mean the derivatives with respect to @xmath138 and @xmath136 . to derive eq .
we assume that electrons obey the adiabatic equation @xmath144 , where @xmath145 is electron pressure [ @xmath146 , see eq . ] and @xmath147 since we study electrostatic oscillations @xcite .
analogous assumptions were taken in ref .
@xcite to study the nonlinear one - dimensional plasma waves in warm plasma within the lagrange picture .
note that eq .
is an exact one , which does not suppose any expansion over a small parameter .
we can , however , discuss the small deviations from the equilibrium position , @xmath148 \notag \\ & + v_e \dot{v}_e \frac{d-1}{\rho } % - % v_e \dot{v}_e \xi \frac{d-1}{\rho^2 } + 3 v_\mathrm{t}^2 \bigg [ 2(\xi '' v_e ' + v_e '' \xi ' ) \notag \\ & + ( d-1)(d-3)\frac{v_e \xi}{\rho^3 } + ( d-1 ) \left ( 2\frac{v_e'\xi'}{\rho } - \frac{v_e \xi''}{\rho } \right ) \notag \\ & - ( d-1)^2 \frac{v_e \xi'}{\rho^2 } % - % 6 v_e ' \xi ' \xi '' - 3 v_e '' \xi'^2 % \notag % \\ % & + % ( d-1 ) % \left ( % 2 \frac{v_e \xi ' \xi''}{\rho } + % \frac{v_e \xi \xi''}{\rho^2 } - % 3 \frac{v_e ' \xi'^2}{\rho } % \right ) % \notag % \\ % & + % ( d-1)^2 \frac{v_e}{\rho^2 } % \left ( % \xi'^2 + \frac{\xi \xi'}{\rho } % \right ) - % ( d-1)(2d-5 ) \frac{v_e \xi^2}{\rho^4 } \bigg ] = 0,\end{aligned}\ ] ] keeping only quadratic nonlinearities . as in sec
. [ locnon ] , @xmath149 , stays for the perturbation of the ion density in eq . .
it can be noticed that eq . has analogous structure as eq . before the separation of the main
harmonic ( for the details see , e.g. , ref .
we can see , however , that nonlinear terms do not disappear completely at @xmath39 if we consider warm plasma with @xmath150 .
one can also notice that these nonvanishing nonlinear terms arise from the pressure term in the initial plasma hydrodynamics equations .
analogous conclusion was obtained in sec .
[ nonlocnon ] using euler coordinates .
note that the nonlinear plasma waves were also studied in refs .
it was found in ref .
@xcite that the nonlinear terms are important even for one dimensional plasma oscillations for the case of nonzero electron temperature , which is in agreement with our results [ see eqs . and ] .
v. l. bychkov , a. i. nikitin , and g. c. dijkhius , in _ the atmosphere and ionosphere : dynamics , processing and monitoring _ , ed . by v.
l. bychkov , g. v. golubkov , and a. i. nikitin , ( heidelberg , springer , 2010 ) , p. 204 .
m. dvornikov and s. dvornikov , in _ advances in plasma physics research _ , vol . 5 , ed .
by f. gerard ( ny , nova science , 2007 ) , pp . 197212 , physics/0306157 ; m. dvornikov , s. dvornikov , and g. smirnov , appl . math .
* 1 * , 9 ( 2001 ) , physics/0203044 . | we study the existence of stable axially and spherically symmetric plasma structures on the basis of the new nonlinear schrdinger equation ( nlse ) accounting for nonlocal electron nonlinearities .
the numerical solutions of nlse having the form of spatial solitions are obtained and their stability is analyzed .
we discuss the possible application of the obtained results to the theoretical description of natural plasmoids in the atmosphere . |
there is an increased interest for understanding the mechanisms that can destabilize the aperiodic or weakly propagating modes in anisotropic plasmas . of these
the purely growing ( aperiodic , i.e. , @xmath6 ) ordinary ( o ) mode instability has recently received particular attention owing to its potential applications in space plasmas @xcite . in a plasma at rest the o - mode instability ( omi )
can develop only if the plasma beta is sufficiently high ( @xmath7 ) , and thermal energy is higher in the direction parallel to the uniform magnetic field ( @xmath8 ) , i.e. , @xmath9 @xcite . in the presence of streams , propagating along the ordered magnetic field , the activity of this instability extends to low beta @xmath10 states @xcite ( although the instability is inhibited by the magnetic field by limiting the range of unstable wavenumbers @xcite ) .
thus , for two symmetric counterstreams of electrons ( subscript @xmath11 ) the conditions necessary for the o - mode instability are @xcite @xmath12 and @xmath13 , where @xmath14 is the electron thermal velocity in parallel direction , @xmath15 , and @xmath16 is the streaming velocity .
( symmetric counterstreams enable to analyze the o - mode decoupled from the extraordinary ( x ) mode , which is less susceptible to the instability .
[ f1 ] presents a schematic with the possible configurations of symmetric counterstreams . )
these conditions simply show that instability is also predicted for low values of @xmath17 ( if streaming velocity is large enough @xmath18 ) , and for high values of the anisotropy @xmath19 ( given by an excess of perpendicular temperature ) .
growth rates are of the order of electron cyclotron frequency @xmath20
. a plasma system with counterstreaming ions can be more susceptible to the o - mode instability than one in which only electrons are streaming @xcite .
the streaming ions enlarge the range of unstable wavenumbers but affect only slightly the maximum growth rates .
however , for large growth rates ( of the order of @xmath20 ) , high electron beta @xmath21 and temperature anisotropy in parallel direction , i.e. , @xmath22 , are needed . on the other hand , only sufficiently high streaming velocities
@xmath23 can predict the occurrence of instability at large @xmath19 . recently , a number of studies have been devoted to express analytically the marginal condition of the o - mode instability @xcite .
it is now straightforward to determine the instability conditions for the whole range of plasma beta , including the low - beta regime where the o - mode is driven unstable only by the relative motion of the plasma streams @xcite . for low @xmath24 it is shown for the first time the existence of a new regime , where the anisotropy thresholds steeply decrease with lowering @xmath5 if the streaming velocity is sufficiently high .
the exact instability thresholds have been derived numerically for small but finite growth rates @xcite , in order to confirm the marginal instability condition in analytical forms , and implicitly the existence of the new regime in the low - beta limit .
counterstreaming plasmas are also subject to the electrostatic two - stream instabilities ( tsi ) , e.g. , electron - electron , electron - ion and ion - ion , of which , the instabilities driven by electrons ( with a growth rate of the order of electron plasma frequency @xmath25 ) appear to be faster @xcite .
moreover , the electrostatic two - stream instability is in most scenarios faster than the o - mode instability ( omi ) , except for streaming velocities very near or below the threshold for the onset of the two - stream instability @xcite .
the instability of the o - mode remains to be established only for streaming velocities below the threshold of the two - stream instability even for low @xmath26 . in this paper
we propose to delimitate these regimes on the basis of the results in ref . 13 , where the marginal condition has been derived systematically for different types and characteristics of two - stream instabilities .
thus , in sec . ii we revisit the instability conditions of the o - mode for three specific cases of symmetric counterstreams .
these are confronted in sec .
iii with the instability conditions of the electrostatic two - stream modes , providing the existence conditions for the o - mode streaming instability .
the new criteria are discussed along with our final conclusions in the last section .
we first reanalyze the o - mode instability invoking the recent results in refs . 2 and 4 .
these results are here applied for three specific cases of symmetric counterstreams with each component modeled by a drifting bi - maxwellian distribution function .
a schematic of these plasma systems is presented in fig .
[ f1 ] : i. the streams are neutral with electrons and ions having the same streaming velocity @xmath27 ; ii .
the electron streaming velocity is higher than the ion streaming velocity , @xmath28 ; and iii .
ions are at rest , @xmath29 , and only electrons are streaming .
symmetric counterstreams minimize the number of employed parameters , and enable to analyze the o - mode decoupled from the extraordinary ( x- ) mode , which is less susceptible to being unstable
. only counterstreams of the same species need to be symmetric to satisfy this condition , but the electron and ion properties ( e.g. , streaming velocity , parallel or perpendicular temperature ) are not necessary the same .
we are dealing with a single species of ions , namely , protons . ; ii .
the electron streaming velocity is higher than the ion streaming velocity @xmath28 ; iii .
ions are stationary @xmath29.,width=321 ] we assume two - component streams of electron - proton plasmas , and start their stability analysis from the analytical eq .
( 49 ) in ref
. 4 @xmath30y \notag \\ & = \beta_\parallel -1 \left(1 + { \mu \over \nu } \right)p_e , \label{e0}\end{aligned}\ ] ] with the same streaming parameters , p_e _
e1 1-_e1 , introduced in ref . 4 , and @xmath31 being the limit value of the squared normalized wavenumber @xmath32 required by the marginal condition of instability
( @xmath33 ) , this equation is obtained based on the improved approximations ( 36 ) and ( 37 ) in the same reference , and here in the next will be refined by neglecting @xmath34 ( or @xmath35 ) , where @xmath36 is the proton - electron mass ratio , and removing the restriction to very high values of the parameter @xmath37 . because of the symmetry of the counterstreams of each species ( with the same relative density @xmath38 , @xmath39 , and the same streaming velocity @xmath40 , @xmath41 ) the quantities @xmath37 and @xmath42 simplify as follows = v_e^2 v_p^2 , p_e = _ pe^2 _ e^2 .
[ e5 ] also for simplicity , the electron and ion temperatures are assumed equal ( @xmath43 , @xmath44 ) , implying @xmath45 .
equation ( [ e0 ] ) then reads y^2 + ay + b = 0 , [ e1 ] with a = 1- _ 2 - p_e ( 1 + 1 ) , b = 1 ^2 , this equation admits a positive solution @xmath46 when at least one of the two coefficients @xmath47 or @xmath48 is negative . in a low @xmath49 regime
this condition is satisfied when the terms depending on @xmath42 are large enough .
thus , @xmath50 is satisfied if p_e > 1-_2 1 + 1 , and @xmath51 if p_e > 1-_1 + .
[ e7]for a low @xmath52 > 1-_1 + , leading to the necessary condition @xmath51 , also found in ref .
4 , eq . ( 50 ) . looking to the solutions of eq .
( [ e1 ] ) we can easily observe that this condition @xmath51 is also sufficient to have at least one solution positive y = -a + 2 > 0 , [ e9 ] that yields , explicitly , @xmath53 \notag \\ & + { 1 \over 2 } \left [ \left({\beta_\parallel \over 2 } -1 \right)^2 + p_e^2 \left(1 + { 1 \over \nu \mu } \right)^2 - 2 p_e \left(1- { \beta_\parallel \over 2}\right ) \right . \notag \\ & \left . + { 2 p_e \over \nu
\mu}\left(1+{\beta_\parallel \over 2}\right)\right]^{0.5}. \label{e10}\end{aligned}\ ] ] otherwise , for a high @xmath0 , @xmath54 is satisfied for any value of @xmath42 , and the same solution ( [ e9 ] ) remains positive .
here we analyze in detail the necessary condition @xmath51 for particular cases , when only the electrons are counterstreaming and ions are stationary ( forming just a neutralizing background ) , and for neutral beams when both the electrons and ions have the same streaming velocities . when ions are at rest , @xmath29 , then @xmath55 , and the instability condition ( [ e7 ] ) becomes ( also see eq .
( 50 ) from ref .
4 ) p_e > 1- _ [ e11 ] . using the explicit form in eq .
( [ e5 ] ) , the instability condition ( [ e11 ] ) requires @xmath56 or _
> 1 1 + v_e^2 u_e,^2 which is less constrained than the condition derived in ref .
9 _ > 2 1 + 2v_e^2 u_e,^2 > 1 1 + v_e^2 u_e,^2 for cold beams we recover the same condition derived in ref
. 14 @xmath57 . when both the electrons and ions are streaming with the same velocity @xmath58 , implying @xmath59 , the same condition ( [ e7 ] ) becomes p_e > 1- _ 1 + / = 1 - _ .
[ e14 ] writing @xmath42 in terms of alfven speed @xmath60 p_e = v_e^2 v_a^2implies in condition ( [ e14 ] ) v > v_a ( 1 - _ ) ^1/2 , which is the same with condition ( 7 ) from ref
. 11 . for cold beams
we find necessary @xmath61 , the same condition derived in ref .
15 . since @xmath62
, it follows that a system with counterstreams of protons ( ions ) is much more susceptible to the instability than one with only counterstreams of electrons .
+ + + + with @xmath63 derived in eq .
( [ e10 ] ) the marginal instability condition is readily found from eq .
( 34 ) in ref .
[ 4 ] a < w ( x_0 ) + w ( x_0 ) + 2p_e _ - 2x_0 _ , [ e17 ] where @xmath64 .
we use this condition to derive the marginal stability against the o - mode instability .
this condition is displayed with solid lines in figs .
[ f2 ] and [ f3 ] for a number of relevant cases .
the o - mode must be unstable in the gray shading below the solid lines . in fig .
[ f3 ] counterstreams are chosen to be neutral , i.e. , with @xmath65 ( case i in fig .
[ f1 ] ) , while in fig .
[ f2 ] the electron streams are faster than ions ( @xmath66 , i.e. , case ii in fig . [ f1 ] ) with @xmath67 .
for the third case , when ions are stationary and only electrons are counterstreaming ( case iii in fig .
[ f1 ] ) , the parameter @xmath68 becomes very large , reducing the expression in eq .
( [ e10 ] ) @xmath69 and the instability condition from eq .
( [ e17 ] ) a < w ( x_0 ) + w ( x_0 ) + 2 _ [ p_e w ( x_0 ) -x_0 ] .
[ e19 ] notice that by comparison to ref . 2 , where the instability condition is derived by neglecting the ion effects , here the ion nonstreaming effects are still present in the right - side ( second term ) of inequality ( [ e19 ] ) .
this new form in eq .
( [ e19 ] ) is used in the present paper to derive the marginal stability of the electron counterstreams , which is displayed with solid lines in fig .
[ f4 ] .
counterstreaming plasmas are also subject to the electrostatic streaming instabilities , i.e. , electron - electron ( @xmath70 ) , electron - proton ( @xmath71 ) , and proton - proton ( @xmath72 ) instabilities , with a maximum growth in the streaming direction . of these ,
the most efficient ( or faster ) are the instabilities driven by electrons , i.e. , @xmath70 or @xmath71 two - stream instabilities @xcite .
notice that no acoustic mode can be excited if the plasma populations and components are isothermal ( @xmath73 ) .
the two - stream instability is also faster than the o - mode instability @xcite , except for streaming velocity very near or below the threshold for the onset of the two - stream instability .
the electrostatic instability is only inhibited by the thermal spread of plasma particles in the streaming direction ( i.e. , the parallel temperature ) , and it is therefore expected to become even more competitive against the o - mode instability in the low @xmath49 regime , where either the two - stream instability is enhanced by the low temperature , or the o - mode instability is inhibited by the intense magnetic field @xcite .
for that reason , the existence of the o - mode instability can be well established only below the marginal condition for the two - stream instability . here
we propose to delimitate these regimes based on the results in ref .
[ 13 ] , where the marginal condition has been derived systematically for different two - stream instabilities .
the symmetry conditions imposed in our present study for the counterstreaming plasmas ( enables decoupling of the o - mode from the x - mode ) along with @xmath74 ( ubiquitous in space plasmas ) lead to a reduced number of two - stream instabilities specified in table [ t1 ] .
relevant for us here are only the plasma counterstreams @xmath75 of the same species @xmath76 or different species @xmath77 , @xmath78 , but satisfying ( _ a t_b _ b t_a ) ^1/2 = 1 , [ e20 ] ( since @xmath79 , and @xmath80 , see previous section ) . in this case
the marginal condition of _ stability _ for two generic counterstreams @xmath75 is @xcite 0.92 ( 1 + _ p , b_p , a ) , [ e21 ] we analyze this condition for each type of two - stream instability found relevant in table [ t1 ] , and then compare them to derive the lowest threshold condition . in this case
@xmath81 , the relative velocity reads @xmath82 and the marginal condition ( [ e21 ] ) becomes @xcite v_e 0.92 u_e , .
[ e22 ] now , to adapt within a @xmath83-dependence ( temperature anisotropy vs. inverse plasma beta ) we use definition ( [ e5 ] ) , and express condition ( [ e22 ] ) in terms of the ( electron ) streaming parameter @xmath42 and the parallel plasma beta p_e 0.85 _ .
[ e23 ] if the instability is driven only by protons , i.e. , @xmath84 , the marginal condition ( [ e21 ] ) provides a lower threshold for the streaming velocity of protons ( also discussed by stringer @xcite ) v_p 0.92 u_p , .
[ e24 ] since @xmath85 ( @xmath36 ) .
the first two cases in fig .
[ f1 ] include conditions when the proton streaming velocity is high enough to drive the two - stream @xmath72 instability .
however , developing of this instability ( with a growth rate of the order of proton plasma frequency ) is not realistic since the o - mode instability ( with a growth rate of the order of electron gyrofrequency ) is expected to be much faster . in this case a buneman - like instability is induced @xcite , and the lowest threshold is found from the same condition ( [ e21 ] ) if we consider the case @xmath86 and @xmath87 v_e 0.92 u_e , .
[ e25 ] furthermore , we can distinguish between two growing modes distinctively assigned to `` @xmath88 '' in the denominator .
thus , the instability can be driven by electrons and protons streaming either in the same direction , when the relative drifting speed is only @xmath89 , or in opposite directions , when the relative drifting speed is higher @xmath90 and determines a lower threshold , asigned to `` @xmath91 '' in the denominator in eq .
( [ e25 ] ) . in terms of the streaming parameter @xmath42 and the parallel plasma beta
this condition reads p_e 0.85 _ ( 1 + 1 ^ 1/2 1 + 1 ^ 1/2)^2 .
[ e26 ] now , comparing conditions ( [ e23 ] ) and ( [ e26 ] ) , we find that the lowest threshold is defined by ( [ e23 ] ) only if @xmath92 , condition well satisfied when ions are almost stationary , e.g. , the limit case iii . indeed , in the limit case iii , the ions are stationary @xmath55 , and the two - stream @xmath93 instability presents the lowest marginal condition given by ( [ e23 ] ) . in the other limit of neutral counterstreams , when the electrons and ions move with the same streaming velocity , @xmath59 and
the lowest marginal condition is that against the electron - proton instability reduced to p_e 0.21 _ ( 1 + 1 ^ 1/2)^2 0.22 _
< 0.92 _ . [ e27 ] the lowest marginal conditions of stability against the electrostatic two - stream instabilities are summarized in table [ t1 ] , indicating three different cases , which are not necessarily related to our first classification , i.e. , cases i , ii and iii .
to conclude , the conditions where the o - mode instability can operate efficiently are markedly constrained by the electrostatic instabilities in both the low @xmath52 and large @xmath0 plasmas .
thus the existence of the o - mode instability becomes possible only for sufficiently high values of the plasma beta parameter in parallel direction @xmath94 .
this restriction is particularly important for the low - beta plasmas , since the existence of this instability was claimed in the previous studies for any small value ( without limit ) of the parallel plasma beta @xmath95 .
the instability of the o - mode can develop in the low - beta plasmas but only for @xmath96 and a streaming energy density less than the magnetic energy density ( @xmath97 ) .
further comparison is now possible with the solar wind streaming conditions to establish whether this instability can explain the observed limits of the anisotropy or not . for a supra - unitary anisotropy @xmath98 , the stimulation of this instability by the energetic streams seems to be impossible ( or unrealistic ) in the low - beta plasmas .
these situations appear to be resolved by a more realistic dissipation of the streaming free energy by the electrostatic instabilities .
the authors acknowledge support from the katholieke universiteit leuven , grant nr .
sf/12/003 , and from the ruhr - universitt bochum , the deutsche forschungsgemeinschaft ( dfg ) , grants schl 201/25 - 1 and sh 21/3 - 2 .
these results were obtained in the framework of the projects goa/2015 - 014 ( ku leuven ) , g.0729.11 ( fwo - vlaanderen ) and c 90347 ( esa prodex 9 ) .
the research leading to these results has also received funding from the european commission s seventh framework programme ( fp7/2007 - 2013 ) under the grant agreements solspanet ( project n 269299 , www.solspanet.eu ) and eheroes ( project n 284461 , www.eheroes.eu ) .
a. stockem , i. lerche and r. schlickeiser , astrophys .
j. * 651 * , 584 ( 2006 ) .
d. ibscher , m. lazar , m. j. michno and r. schlickeiser , phys .
plasmas * 20 * , 012103 ( 2013 ) .
f. hadi , m. f. bashir , a. qamar , p. yoon and r. schlickeiser , phys .
plasmas * 21 * , 052111 ( 2014 ) .
r. schlickeiser and p. yoon , phys .
plasmas * 21 * , 072119 ( 2014 ) .
d. ibscher , m. lazar , and r. schlickeiser , phys .
plasmas * 19 * , 072116 ( 2012 ) .
m. lazar , s. poedts , r. schlickeiser and d. ibscher , solar phys .
289 , 369 ( 2013 ) .
m. bornatici and k. f. lee , phys .
fluids * 13 * , 3007 ( 1970 ) .
k. f. lee and j. c. armstrong , phys .
rev . a * 4 * , 2087 ( 1971 ) . | in plasmas where the thermal energy density exceeds the magnetic energy density ( @xmath0 ) , the aperiodic ordinary mode ( o - mode ) instability is driven by an excess of parallel temperature @xmath1 ( where @xmath2 and @xmath3 denote directions relative to the uniform magnetic field ) .
when stimulated by parallel plasma streams the instability conditions extend to low beta states , i.e. , @xmath4 , and recent studies have proven the existence of a new regime , where the anisotropy threshold decreases steeply with lowering @xmath5 if the streaming velocity is sufficiently high . however , the occurrence of this instability is questionable especially in the low - beta plasmas , where the electrostatic two - stream instabilities are expected to develop much faster in the process of relaxation of the counterstreams .
it is therefore proposed here to identify the instability conditions for the o - mode below those required for the onset of the electrostatic instability .
an hierarchy of these two instabilities is established for both the low @xmath4 and large @xmath0 plasmas .
the conditions where the o - mode instability can operate efficiently are markedly constrained by the electrostatic instabilities especially in the low - beta plasmas . |
the large hadron collider ( lhc ) will soon test the mechanism of spontaneous electroweak symmetry breaking .
it is widely agreed that this phenomenon is associated to the expectation value of a scalar field which transforms as a doublet with respect to the weak su(2)-symmetry .
the origin and status of this order parameter wait , however , for experimental clarification .
in particular , an effective description in terms of a scalar field does not tell us if this scalar is fundamental " in the sense that it constitutes a dynamical degree of freedom in a microscopic theory which is formulated at momentum scales much larger than the fermi scale .
alternatively , no fundamental higgs scalar may be present , and the order parameter rather involves an effective composite field " . in this note
we investigate the second possibility and therefore consider the fermions of the standard model without a fundamental higgs scalar . as long as we include only the gauge interactions of the standard model , the microscopic or classical action of such a theory does not involve any mass scale .
an effective mass scale @xmath2 will only be generated by the running of the strong gauge coupling , which is asymptotically free @xcite .
confinement removes then the gluons and quarks from the massless spectrum .
in addition , the spontaneous chiral symmetry breaking by quark - antiquark condensates would also imply electroweak symmerty breaking by a composite order parameter , in this case the chiral condensate . in this
setting all particle masses would be of the order @xmath2 or zero , such that this scenario can not explain why the top quark mass or the @xmath3,@xmath4-boson masses are much larger than 1gev .
furthermore , all leptons would remain massless .
one concludes that any realistic model needs further interactions beyond the standard model gauge interactions .
since the strong and electroweak gauge couplings remain actually quite small at the fermi scale of electroweak symmetry breaking , they are expected to produce only some quantitative corrections to the dominant mechanism of electroweak symmetry breaking .
we will therefore neglect the gauge couplings in this paper .
our knowledge about the effective interactions between the quarks and leptons at some microscopic or
ultraviolet " scale @xmath5 is very limited .
furthermore , it is not known which scale @xmath5 has to be taken .
typically , one may associate @xmath5 with the scale where further unification takes place , as a grand unified scale or the planck scale for the unification with gravity .
this would suggest a very high scale , @xmath6 gev and we will have this scenario in mind for our discussion . however , much smaller values of @xmath5 are also possible . in practice
, we will only assume here that @xmath5 is sufficiently above the fermi scale ( say @xmath7 tev ) , such that an effective description involving only the fermions of the standard model becomes possible in the momentum range @xmath8 .
we will formulate our model in terms of effective fermion interactions at the scale @xmath5 and restrict the discussion to a four - fermion interaction involving only the right - handed top quark and the left - handed top and bottom quarks .
this is motivated by the observation that only the top quark has a mass comparable to the fermi scale .
interactions with the other quarks and leptons are assumed to be much smaller than the top quark interactions - typically their relative suppression is reflected in the much smaller masses of the other fermions .
for the discussion in this paper we omit all light " fermions and the gauge bosons . since we do not know the effective degrees of freedom at the scale @xmath5 , the effective interaction is not necessarily local .
non - localities involving inverse powers of the exchanged momenta are typically generated by the propagators of exchanged massless particles . with the inclusion of possibly non - local interactions the limitation to an effective four - fermion interaction
poses no severe restriction .
many models with additional degrees of freedom can be effectively described in this way .
local four - fermion interactions have already been investigated earlier , for example in the models of top quark condensation " @xcite . by simple dimensional analysis
a local interaction involves a coupling @xmath9(mass)@xmath10 .
such models therefore exhibit an explicit mass scale in the microscopic action .
it is indeed possible to obtain spontaneous electroweak symmetry breaking in this way - the prototype is the nambu - jona - lasinio model @xcite . without a tuning of parameters
the top quark mass @xmath11 turns out , however , to be of the same order as @xmath5 , in contradiction to the assumed separation of scales . by a tuning of parameters it is possible to obtain @xmath12 , but
the issue is now similar to the gauge hierarchy problem " in presence of a fundamental scalar field . in order to obtain @xmath12
the microscopic effective action must be close to an ultraviolet fixed point .
the necessity of tuning arises from a relevant parameter " in the vicinity of the fixed point ( in the sense of statistical physics for critical phenomena ) which has a dimension not much smaller than one .
a rather extensive search for possible ultraviolet fixed points for pointlike four - fermion interactions has been performed in @xcite .
many fixed points have been found , but all show a relevant direction with substantial dimension , and therefore the need for a parameter tuning for @xmath12
. we will therefore concentrate in this paper on non - local effective interactions . for such interactions
the coupling @xmath13 is replaced by @xmath14 , with @xmath15 the square of some appropriate exchanged momentum and @xmath0 a dimensionless coupling .
interactions of this type do not involve a mass scale on the level of the microscopic action - the classical action exhibits dilatation symmetry .
still , the quantum fluctuations typically induce an anomaly for the scale symmetry , associated to the running of the dimensionless coupling @xmath0
. this may be responsible for the generation of the fermi scale , in analogy to the confinement scale " @xmath2 for qcd .
since the running of dimensionless couplings is only logarithmic , this offers a chance for a large natural hierarchy @xmath16 , without tuning of parameters .
we will present a model of this type .
let us first discuss the possible tensor structures for a non - local four - fermion interaction @xmath17 .
the basic building block is a color singlet fermion bilinear @xmath18 , where the color indices are contracted . the tensor structure with respect to the lorentz symmetry
is determined by @xmath19 such that @xmath18 is a scalar or pseudoscalar , a vector or pseudovector , or a second rank antisymmetric tensor .
interactions in the vector or pseudovector channels involve bilinears with two left - handed or two right - handed fermions , @xmath20 or @xmath21 .
they conserve chiral flavour symmetries which act separately on @xmath22 and @xmath23 and therefore forbid mass terms for the fermions @xmath24 .
interactions capable of producing masses for the top quark and w / z - bosons of comparable magnitude must therefore involve the ( pseudo)scalar or tensor channels . a non - local scalar interaction @xmath25 , with @xmath15
the squared momentum in the scalar exchange channel , has very similar properties as a model with a fundamental higgs scalar which is massless at the scale @xmath5 .
we therefore expect the usual necessity of parameter tuning if we want to achieve a small ratio @xmath26 .
a local coupling @xmath27 is allowed by the symmetries and will be generated by quantum fluctuations .
the interesting remaining candidate is a tensor interaction , with @xmath28 $ ] . for chiral tensors no local interaction in this channel
is consistent with the su(3)@xmath29su(2)@xmath29u(1 ) symmetries of the standard model as well as lorentz symmetry .
we will therefore investigate a model with a non - local interaction of this type .
we define the microscopic or classical action for a lorentz invariant theory of massless interacting fermions by @xmath30 , defined in momentum space as @xmath31 and @xmath32 \ ; \left [ \bar{t}(p')\sigma_-^l t(p'+q)\right ] \\
\nonumber & + & \left [ \bar{t}(q+p)\sigma_+^k b(p)\right ] \ ; \left [ \bar{b}(p')\sigma_-^l t(p'+q)\right ] \bigg\}. \end{aligned}\ ] ] here @xmath33 and @xmath34 are dirac spinor fields describing the top and bottom quark , respectively .
( the theory can be easily extended to all three generations of quarks and also to leptons . )
contracted indices are summed .
the @xmath35 matrix @xmath36 involves the spacelike indices @xmath37,@xmath38 and is defined by @xmath39 it has the properties @xmath40 the non - local character of the interaction arises from the factor @xmath41 . in a standard spinor basis in which @xmath42 , @xmath43 , the @xmath44 matrices @xmath45 are defined in terms of the pauli matrices @xmath46 , @xmath47 the fermion bilinears @xmath48 and @xmath49 therefore mix left- and right - handed spinors and
violate the chiral symmetry that would protect the top quark from acquiring a mass .
the matrices @xmath50 correspond to appropriate projections of the commutator @xmath51 $ ] on left ( right ) handed spinors , such that eq .
( [ a2 ] ) indeed describes a tensor exchange interaction ( cf . the appendix a ) .
the lorentz - invariance of the interaction can be checked explicitly .
furthermore , the action ( [ a1 ] ) , ( [ a2 ] ) has a global su(2)@xmath29u(1 ) symmetry - the remnant of the electroweak gauge symmetry in the limit of neglected gauge couplings .
this symmetry forbids mass terms for the quarks , such that a mass term can be generated only by spontaneous symmetry breaking .
similarly , the interaction is invariant under the color symmetry su(3 ) , with implicitly summed color indices in the bilinears .
we note that only the left - handed bottom quarks are involved in the interaction and we will therefore omit the right - handed bottom quark together with the other light quarks and the leptons .
similar to the photon - exchange description of the non - local coulomb interaction we may obtain @xmath52 from the exchange of chiral tensor fields @xcite , according to the feynman diagrams in fig . 1 .
the lorentz symmetry of @xmath52 becomes more apparent in this formulation .
a summary of the properties of the associated tensor fields and a proof of equivalence of a local theory with massless chiral tensor fields with our non - local fermion interaction ( [ a2 ] ) is provided in appendix a. in this paper we will not use tensor fields and concentrate on the purely fermionic action ( [ a1 ] ) , ( [ a2 ] ) .
in particular , this avoids the delicate issues of consistency of the quantization of chiral antisymmetric tensors @xcite .
one can write a consistent functional integral and therefore a consistent quantum field theory based on the action ( [ a1 ] ) , ( [ a2 ] ) .
( anomaly cancellation involves the omitted light quarks and leptons . ) ( 400,120)(0,0 ) ( 50,20)(70,60 ) ( 70,60)(50,100 ) ( 150,20)(130,60 ) ( 130,60)(150,100 ) ( 70,60)(130,60)3 ( 55,100)[tl]@xmath53 ( 55,20)[bl]@xmath54 ( 145,100)[tr]@xmath54 ( 145,20)[br]@xmath53 ( 100,65)[b]@xmath55 ( 250,20)(270,60 ) ( 270,60)(250,100 ) ( 350,20)(330,60 ) ( 330,60)(350,100 ) ( 270,60)(330,60)3 ( 255,100)[tl]@xmath56 ( 255,20)[bl]@xmath54 ( 345,100)[tr]@xmath54 ( 345,20)[br]@xmath56 ( 300,65)[b]@xmath57 this paper is organized as follows . in sect .
2 we compute the running of the dimensionless coupling @xmath0 and demonstrate that it is asymptotically free . in consequence , a non - perturbative chiral " scale @xmath1
is generated where the running coupling @xmath0 grows large .
3 evaluates the effective interactions in the scalar and vector channels which are induced in one loop - order .
we find that the induced interaction in the scalar channel @xmath58 also gets large once the chiral scale is approached . as a consequence of this large interaction spontaneous electroweak symmetry breaking can be generated similar to the njl - model @xcite .
we compute the resulting top quark mass @xmath11 in sect .
4 , using a schwinger - dyson @xcite or gap equation .
we find @xmath59 . in sect .
5 we introduce a composite higgs field by partial bosonization " of the interaction in the scalar channel .
we compute the running of its yukawa coupling @xmath60 to the top quark .
the latter is directly related to the mass ratio @xmath61 , with @xmath62 and @xmath63 the mass and gauge coupling of the @xmath3 boson . in sect .
6 we present our conclusions .
the dimensionless coupling constant @xmath0 is the only free parameter in our model
. we will show in this section that the corresponding renormalized running coupling is asymptotically free in the ultraviolet . on the other side
, there is a characteristic infrared scale @xmath1 where the coupling @xmath0 grows large .
this is in complete analogy to qcd , where @xmath2 corresponds to the scale where the strong gauge coupling grows large . by
dimensional transmutation " we may therefore trade @xmath0 for the mass scale @xmath1 . besides this mass scale
the model has no free dimensionless parameter . in particular
, we will argue in sect . 4 that our model leads to spontaneous breaking of the su(2)@xmath29u(1 ) symmetry .
the fermi scale @xmath64 characterizing the electroweak symmetry breaking must be proportional to @xmath1 , @xmath65 , with a dimensionless proportionality coefficient @xmath66 that is , in principle , calculable without involving a further free parameter .
if the model leads to a composite higgs scalar , its mass in units of @xmath64 , @xmath67 , as well as its effective yukawa coupling to the top quark , @xmath68 , are calculable quantities in our model . in order to investigate the running of @xmath0
, we wish to compute the effective action @xmath69 corresponding to the classical action ( [ a1]),([a2 ] ) , to one - loop order .
the one - loop correction reads @xmath70 where the field dependent inverse propagator @xmath71 is defined as the second functional derivative of the action with respect to the quark fields , @xmath72 here @xmath73 are the quark fields , with color indices @xmath74 , flavor indices @xmath75 ( taking values @xmath76 ) , spinor indices @xmath77 and momenta @xmath78 .
the trace tr involves an integral over momentum and summation over all kinds of indices .
we write @xmath79 , where @xmath80 is the free " part of the propagator , derived from @xmath81 , @xmath82\ ; \delta_{cc'}\delta(p - p'),\ ] ] and @xmath83 .
we treat @xmath84 as a perturbative correction due to @xmath52
. then @xmath85 reads , up to an
infinite constant " , @xmath86 where the dots stand for neglected six - quark and higher interactions .
we display the explicit expressions for @xmath87 , as well as the formal expressions for the first two terms on the r.h.s .
( [ c4 ] ) , in appendix b. ( 400,120)(0,0 ) ( 40,10)(40,35 ) ( 40,35)(40,85 ) ( 40,85)(40,110 ) ( 40,60)(25,270,90)3 ( 35,20)[r]@xmath54 ( 35,60)[r]@xmath53 ( 35,100)[r]@xmath54 ( 70,60)[l]@xmath55 ( 140,10)(140,35 ) ( 140,35)(140,85 ) ( 140,85)(140,110 ) ( 140,60)(25,270,90)3 ( 135,20)[r]@xmath54 ( 135,60)[r]@xmath56 ( 135,100)[r]@xmath54 ( 170,60)[l]@xmath57 ( 240,10)(240,35 ) ( 240,35)(240,85 ) ( 240,85)(240,110 ) ( 240,60)(25,90,270)3 ( 235,20)[r]@xmath53 ( 235,60)[r]@xmath54 ( 235,100)[r]@xmath53 ( 270,60)[l]@xmath55 ( 340,10)(340,35 ) ( 340,35)(340,85 ) ( 340,85)(340,110 ) ( 340,60)(25,90,270)3 ( 335,20)[r]@xmath56 ( 335,60)[r]@xmath54 ( 335,100)[r]@xmath56 ( 370,60)[l]@xmath57 ( 320,130)(0,0 ) ( 20,20)(20,60 ) ( 20,60)(20,100 ) ( 120,20)(120,60 ) ( 120,60)(120,100 ) ( 20,60)(50,60)3 ( 90,60)(120,60)3 ( 70,60)(20,0,180 ) ( 70,60)(20,180,0 ) ( 15,40)[tr]@xmath53 ( 15,80)[br]@xmath54 ( 125,40)[tl]@xmath54 ( 125,80)[bl]@xmath53 ( 35,63)[b]@xmath55 ( 105,63)[b]@xmath55 ( 70,83)[b]@xmath54 ( 70,37)[t]@xmath53 ( 190,20)(190,60 ) ( 190,60)(190,100 ) ( 290,20)(290,60 ) ( 290,60)(290,100 ) ( 190,60)(220,60)3 ( 260,60)(290,60)3 ( 240,60)(20,0,180 ) ( 240,60)(20,180,0 ) ( 185,40)[tr]@xmath56 ( 185,80)[br]@xmath54 ( 295,40)[tl]@xmath54 ( 295,80)[bl]@xmath56 ( 205,63)[b]@xmath57 ( 275,63)[b]@xmath57 ( 240,83)[b]@xmath54 ( 240,37)[t]@xmath56 the nonlocal factors @xmath88 appearing in the interaction ( [ a2 ] ) are attached in different ways to the fermion lines .
we represent them as dashed lines in the corresponding feynman diagrams in figs . 2 - 4 .
our notation recalls the one - to - one correspondence with the exchange of the associated chiral tensor fields .
the diagrams in fig . 2 correspond to the first term @xmath89 in eq .
( [ c4 ] ) , while figs .
3,4 reflect the terms @xmath90 .
+ our one - loop calculation results in an effective action @xmath91 with a kinetic term @xmath92 and three different types of quartic interactions @xmath93 .
( note the relative minus sign between @xmath69 and the classical action @xmath94 which is chosen in order to make analytic continuation to the euclidean effective action straightforward by replacing @xmath95 . )
the tensorial part " , @xmath96 , corresponds to the exchange of a tensor field and has the form of the classical interaction @xmath52 .
the other two parts , @xmath97 and @xmath98 , correspond to the exchange of vectors and scalars , respectively , and will be further discussed in sect .
the momentum intergals in the loop expansion ( [ c4 ] ) are logarithmically divergent , both in the ultraviolet and the infrared .
we regularize our model in the ultraviolet by a suitable momentum cutoff @xmath5 . in order to investigate the flow of effective couplings we also introduce an effective infrared cutoff @xmath37 . the effective action @xmath99 depends then on the scale @xmath37 , resulting in an effective coupling @xmath100 . instead of the infrared scale induced by non - vanishing external momenta for the vertices ( as most common for perturbative renormalization )
we introduce the cutoff by modifying the quark propagators + @xmath101 .
this is a procedure known from functional renormalization .
indeed , @xmath99 may be considered as the average action " or flowing action " @xcite .
the precise implementation of the infrared cutoff is not important and does not influence the one - loop beta function for the running coupling @xmath100 that we will derive next .
the fermion anomalous dimensions arise from eq .
( [ c9 ] ) or fig . 2 .
our computation in the purely fermionic model yields the same result as in ref .
@xcite , where it was computed in the equivalent model with chiral tensors , namely @xmath102 only the terms visualized diagramatically in fig . 3 , which are @xmath103 in eq . ( [ c10 ] ) ,
generate the same tensor structure as the classical interaction ( [ a2 ] ) .
they provide the one - loop correction to the inverse chiron propagator , i.e. to the matrix @xmath104 , and one obtains @xmath105 again our fermionic computation reproduces the computation in ref .
@xcite with chiral tensors .
the correction results in an anomalous dimension for the chiron , @xmath106 there are no further one loop corrections in the tensor exchange channel . among the quartic corrections shown in fig . 4 , the first four terms in eq . ( [ c10 ] )
contribute to an interaction channel which is equivalent to the exchange of a vector particle .
these diagrams will be evaluated in the next section .
similarly , the fifth and sixth term in eq .
( [ c10 ] ) contribute to an interaction with a tensor structure different from the classical action ( [ a2 ] ) .
it can be interpreted as an effective scalar exchange and will also be discussed in the next section .
the running of the renormalized coupling @xmath107 ( to one - loop order ) is therefore given by the anomalous dimensions of the fermions and the correction to @xmath104 , @xmath108 this implies that the interaction is asymptotically free .
the solution to the renormalization group equation ( [ c13 ] ) is @xmath109 where the chiral scale " @xmath1 is the scale at which @xmath110 becomes much larger than one .
this is completely equivalent to the result of ref .
the asymptotic freedom of the chiral fermion - tensor interaction is simply taken over to the non - local four - fermion interaction .
( 450,240)(0,0 ) ( 30,130)(30,155 ) ( 30,155)(30,205 ) ( 30,205)(30,230 ) ( 110,130)(110,155 ) ( 110,155)(110,205 ) ( 110,205)(110,230 ) ( 30,205)(70,180)3 ( 110,205)(70,180)3 ( 70,180)(110,155)3 ( 70,180)(30,155)3 ( 25,140)[r]@xmath53 ( 25,180)[r]@xmath54 ( 25,220)[r]@xmath53 ( 115,140)[l]@xmath53 ( 115,180)[l]@xmath54 ( 115,220)[l]@xmath53 ( 40,202)[bl]@xmath55 ( 100,202)[br]@xmath55 ( 170,130)(170,155 ) ( 170,155)(170,205 ) ( 170,205)(170,230 ) ( 250,130)(250,155 ) ( 250,155)(250,205 ) ( 250,205)(250,230 ) ( 170,205)(210,180)3 ( 250,205)(210,180)3 ( 250,155)(210,180)3 ( 170,155)(210,180)3 ( 165,140)[r]@xmath54 ( 165,180)[r]@xmath53 ( 165,220)[r]@xmath54 ( 255,140)[l]@xmath54 ( 255,180)[l]@xmath53 ( 255,220)[l]@xmath54 ( 180,162)[tl]@xmath55 ( 240,162)[tr]@xmath55 ( 310,130)(310,155 ) ( 310,155)(310,205 ) ( 310,205)(310,230 ) ( 390,130)(390,155 ) ( 390,155)(390,205 ) ( 390,205)(390,230 ) ( 310,205)(350,180)3 ( 390,205)(350,180)3 ( 350,180)(390,155)3 ( 350,180)(310,155)3 ( 305,140)[r]@xmath53 ( 305,180)[r]@xmath54 ( 305,220)[r]@xmath56 ( 395,140)[l]@xmath56 ( 395,180)[l]@xmath54 ( 395,220)[l]@xmath53 ( 320,202)[bl]@xmath57 ( 380,202)[br]@xmath55 ( 30,10)(30,35 ) ( 30,35)(30,85 ) ( 30,85)(30,110 ) ( 110,10)(110,35 ) ( 110,35)(110,85 ) ( 110,85)(110,110 ) ( 30,85)(70,60)3 ( 110,85)(70,60)3 ( 70,60)(110,35)3 ( 70,60)(30,35)3 ( 25,20)[r]@xmath56 ( 25,60)[r]@xmath54 ( 25,100)[r]@xmath56 ( 115,20)[l]@xmath56 ( 115,60)[l]@xmath54 ( 115,100)[l]@xmath56 ( 40,82)[bl]@xmath57 ( 100,82)[br]@xmath57 ( 170,10)(170,35 ) ( 170,35)(170,85 ) ( 170,85)(170,110 ) ( 250,10)(250,35 ) ( 250,35)(250,85 ) ( 250,85)(250,110 ) ( 250,85)(170,85)3 ( 170,35)(250,35)3 ( 165,20)[r]@xmath54 ( 165,60)[r]@xmath53 ( 165,100)[r]@xmath54 ( 255,20)[l]@xmath53 ( 255,60)[l]@xmath54 ( 255,100)[l]@xmath53 ( 210,88)[b]@xmath55 ( 210,38)[b]@xmath55 ( 310,10)(310,35 ) ( 310,35)(310,85 ) ( 310,85)(310,110 ) ( 390,10)(390,35 ) ( 390,35)(390,85 ) ( 390,85)(390,110 ) ( 390,85)(310,85)3 ( 310,35)(390,35)3 ( 305,20)[r]@xmath54 ( 305,60)[r]@xmath56 ( 305,100)[r]@xmath54 ( 395,20)[l]@xmath56 ( 395,60)[l]@xmath54 ( 395,100)[l]@xmath56 ( 350,88)[b]@xmath57 ( 350,38)[b]@xmath57 in this section we evaluate the diagrams in fig . 4 , representing the terms @xmath111 in eq . ( [ c10 ] ) .
we begin with the fifth diagram with all external momenta set to zero .
( the contribution of the sixth diagram is equivalent , with @xmath53 substituted by @xmath56 . ) the contribution to @xmath69 is @xmath112 \left[\bar{t}\sigma^n_- \frac{\slash\hspace{-0.2 cm } q}{q^2 } \sigma^k_+ t \right].\ ] ] this may be rewritten in terms of weyl spinors @xmath113 \left [ t^\dagger_l \tau^n \frac{\slash\hspace{-0.2 cm } \bar{q}}{q^2 } \tau^k t_l \right],\ ] ] where now @xmath114 with the identity @xmath115_{\alpha\beta } \left [ \tau^n \slash\hspace{-0.2 cm } \bar{q } \tau^k \right]_{\lambda\eta } = 5 q^4 \left [ \slash\hspace{-0.2 cm } q \right]_{\alpha\beta } \left [ \slash\hspace{-0.2 cm } \bar{q } \right]_{\lambda\eta } + 4 q^6 \delta_{\alpha\eta}\delta_{\beta\lambda}\ ] ] this simplifies to @xmath116 [ t^\dagger_l \slash\hspace{-0.2 cm } \bar{q } t_l ] + 4 q^2 [ t^{c \dagger}_r t^{c'}_l][t^{c ' \dagger}_l t^c_r ] \right).\ ] ] the relative minus sign is due to an exchange of grassmann variables . note the different color structure of the second term .
the momentum integral can be evaluated by analytic continuation to euclidean space , @xmath117 . introducing the infrared cutoff in the quark propagators ( i.e. substituting @xmath118 ) ,
one obtains @xmath119 [ t^\dagger_l \bar{\tau}^\nu t_l ] - 16 [ t^{c \dagger}_r t^{c'}_l][t^{c ' \dagger}_l t^c_r ] \right).\ ] ] using the identities @xmath120 and @xmath121 ( where @xmath122 are the eight gell - mann matrices ) this can be further simplified to @xmath123[\bar{t}_l t_r ] + \frac{3 f^4}{4 \pi^2 k^2}[\bar{t}^c_r t^z_{cd } t^d_l ] [ \bar{t}^{c'}_l t^z_{c'd ' } t^{d'}_r].\ ] ] the total contribution to @xmath69 from the fifth and sixth diagram in fig .
4 reads @xmath124 the first term in eq .
( [ d8 ] ) is equivalent to the four - fermion interaction generated at tree level by a yukawa interaction with a scalar field @xmath125 , which has a mass @xmath37 and a yukawa coupling @xmath126 given by @xmath127 this scalar transforms as a singlet under color and a doublet with respect to the electroweak interactions .
it therefore has the same quantum numbers as a ( composite ) higgs doublet . the second term in eq .
( [ d8 ] ) corresponds to the exchange of a second scalar which is an octet with respect to color .
finally we evaluate the first four diagrams of fig . 4 , which generate an interaction equivalent to the exchange of a vector particle .
the expression for the first diagram is @xmath128 \left[\bar{t}\sigma^n_- \frac{\slash\hspace{-0.2 cm } q}{q^2 } \sigma^k_+ t \right].\ ] ] this can again be rewritten in terms of weyl spinors @xmath129 \left [ t^\dagger_l \tau^n \frac{\slash\hspace{-0.2 cm } \bar{q}}{q^2 } \tau^k t_l \right].\ ] ] with the identities @xmath130_{\alpha\beta } \left [ \tau^n \slash\hspace{-0.2 cm } \bar{q } \tau^k \right]_{\lambda\eta } = 5 q^4 \left [ \slash\hspace{-0.2 cm } \bar{q } \right]_{\alpha\beta } \left [ \slash\hspace{-0.2 cm } \bar{q } \right]_{\lambda\eta } + 4 q^6 ( \delta_{\alpha\beta}\delta_{\eta\lambda}-\delta_{\alpha\eta}\delta_{\beta\lambda})\ ] ] and @xmath131 the expression ( [ e2 ] ) simplifies to @xmath132 [ t^\dagger_l \slash\hspace{-0.2 cm } \bar{q } t_l ] -2 q^2 [ t^\dagger_l \bar{\tau}^\mu
t_l][t^\dagger_l \bar{\tau}_\mu t_l ] \right).\ ] ] with an infrared cutoff @xmath37 in the quark propagator the momentum integral yields @xmath133 [ \bar{t}_l \gamma_\mu t_l].\ ] ] this is equivalent to the four - fermion interaction generated at tree level by the exchange of a vector field with mass @xmath37 and coupling @xmath134 given by @xmath135 the evaluation of the diagrams 3 and 4 is equivalent , with two or four external @xmath53 fields substituted by @xmath56 .
the expression for the second diagram , in terms of weyl spinors , is @xmath136 \left [ t^\dagger_r \tau^k \frac{\slash\hspace{-0.2 cm } q}{q^2 } \tau^n t_r \right].\ ] ] identities similar to eqs .
( [ e4 ] ) and ( [ e7 ] ) result in @xmath137 [ \bar{t}_r \gamma_\mu t_r].\ ] ] the vector exchange " interaction generated at one loop level can be summarized as @xmath138 [ \bar{t}_l \gamma_\mu t_l ] + 2 [ \bar{b}_l \gamma^\mu t_l ] [ \bar{t}_l \gamma_\mu b_l ] \qquad \\ \nonumber & + [ \bar{b}_l \gamma^\mu b_l ] [ \bar{b}_l \gamma_\mu b_l ] + 2 [ \bar{t}_r \gamma^\mu t_r ] [ \bar{t}_r \gamma_\mu t_r]\bigg).\end{aligned}\ ] ]
the presence of the effective yukawa interaction in eq .
( [ d8 ] ) indicates the possibility of spontaneous symmetry breaking and an analogue of the higgs mechanism .
if we neglect for @xmath37 close to @xmath1 all interactions except for the scalar singlet exchange channel in the first term of eq .
( [ d8 ] ) , we may characterize the effective action by a
scalar four - fermion coupling " @xmath139 , @xmath140 = 2 \lambda ( \bar{\psi}_l \psi_r)(\bar{\psi}_r \psi_l)\ ] ] with @xmath141 in the limit where the momentum dependence of the scalar four - fermion interactions can be neglected this can be interpreted as an effective njl model . here
the role of the uv cutoff in the effective njl model is played by the scale @xmath37 , since the computation of @xmath142 has involved only fluctuations with momenta @xmath143 ( due to the infrared cutoff ) such that the remaining fluctuations with @xmath144 still have to be included .
it is well known that for @xmath145 exceeding a critical value the njl model leads to spontaneous symmetry breaking . for a rough estimate of the top quark mass @xmath11 induced by spontaneous electroweak symmetry breaking we consider the schwinger dyson equation @xmath146 setting @xmath147 and performing the momentum integral yields the gap equation for the top quark mass @xmath148 @xmath149 we have indicated the scale @xmath37 for @xmath148 in order to recall that the solution of the gap equation will depend on the choice of the scale @xmath37 which we use as an uv cutoff for the effective njl model
of course , for an exact treatment the physical top quark mass should no longer depend on @xmath37 .
let us next discuss the solution of eq .
( [ g2 ] ) . obviously , @xmath150 is always one possible solution .
the expression in brackets is always @xmath151 .
if we write @xmath152 we see that non - zero solutions for @xmath148 ( indicating spontaneous symmetry breaking ) occur for @xmath153 once the condition ( [ g4 ] ) is obeyed , one finds indeed a non - zero @xmath148 obeying @xmath154 since @xmath155 grows arbitrarily large as @xmath37 approaches @xmath1 and @xmath100 diverges , the condition ( [ g4 ] ) is always fulfilled for @xmath37 sufficiently close to @xmath1 . for a qualitative investigation we first use the one loop result @xmath156 and replace in eq .
( [ c14 ] ) @xmath157 with @xmath158 of order 1 .
this is motivated by the effective infrared cutoff @xmath159 which stops or slows down the running of @xmath107 . in fig .
5 we show @xmath160 as a function of @xmath161 for @xmath162 . after the onset of a nonzero @xmath148 for @xmath163 we find first a very rapid increase until @xmath148 settles at @xmath164 for @xmath165 .
it is obvious that @xmath1 sets the scale for the top quark mass .
this coincides with the result of a two - loop schwinger - dyson equation in a formulation with chiral tensor fields in @xcite .
indeed , since the generation of @xmath166 is a one - loop effect , and the gap equation ( [ g1 ] ) involves a further loop , the generation of the top quark mass consists of two nested one - loop integrals , which are equivalent to a two - loop integral . in the njl approximation . ] the solution of eq .
( [ xx ] ) for @xmath165 , @xmath167 corresponds to an asymptotic value which is obtained from the condition @xmath168 . as an alternative to the ad hoc insertion of the quark mass cutoff in eq .
( [ c14 ] ) we may take into account the additional infrared cutoff due to @xmath148 by replacing in the quark propagator @xmath169 . as a consequence ,
the anomalous dimensions involve a threshold function @xmath170 , @xmath171 given by @xmath172 the one loop expression for @xmath173 ( [ g1b ] ) can be replaced by a flow equation @xmath174 ( for @xmath150 , @xmath175 and constant @xmath0 this reproduces eq .
( [ g1b ] ) . )
nonzero @xmath148 results in a threshold function for the running of @xmath166 , @xmath176 @xmath177 we show the running of @xmath107 and @xmath166 in fig . 6 , for different values of @xmath178 .
again we stop the flow at some value of @xmath179 and solve the schwinger - dyson equation with this uv cutoff . the value of @xmath37 for which the sde yields the given @xmath178 is indicated in fig . 6 by a dot . the dots in fig .
5 show the corresponding @xmath37 dependence of @xmath160 .
( red curves ) and @xmath166 ( blue curves ) .
solid ( dashed , dotted ) lines correspond to @xmath180 ( 0.6 , 0.9 ) @xmath1.,title="fig : " ] ( red curves ) and @xmath166 ( blue curves ) .
solid ( dashed , dotted ) lines correspond to @xmath180 ( 0.6 , 0.9 ) @xmath1.,title="fig : " ] we are aware that our treatment of the infrared cutoff is only qualitative . while the one - loop form of the flow equation can be maintained if we interpret the flow as an approximation to the exact functional renormalization group equation for the average action @xcite , the approximation of the exact inverse quark and chiron propagators by @xmath181 and @xmath182 with momentum independent @xmath4-factors becomes questionable in the presence of large anomalous dimensions .
an elegant method for the description of composite particles in the context of functional renormalization uses partial bosonization @xcite .
it is based on the observation that an interaction of the type ( [ g1a ] ) can be described by the exchange of a composite scalar field .
indeed , one may add to the flowing action @xmath99 a scalar piece , with @xmath183 a complex doublet scalar field @xmath184 integrating out " the scalar field by solving its field equation as a functional of @xmath185 and reinserting the solution into eq .
( [ s1 ] ) yields eq .
( [ g1a ] ) , with @xmath173 dependent on the squared exchanged momentum in the scalar channel @xmath186 as long as the momentum dependence of the effective scalar exchange vertex is not resolved ( as in our computation where the vertex is only evaluated for @xmath187 ) , one may take arbitrary @xmath188 .
we will therefore replace @xmath142 in eq .
( [ g1a ] ) by eq .
( [ s1 ] ) , and replace the flow of @xmath173 in eq .
( [ x1 ] ) by @xmath189 at this stage our reformulation precisely reproduces the results in sect .
the rule for the replacement of the flow of @xmath190 by a running of the renormalized yukawa coupling and dimensionless mass term @xmath191 is to adjust the flow of @xmath192 and @xmath193 such that the flow of @xmath190 is reproduced , with @xmath194 for the approximately scale invariant situation for small coupling one expects that the relative split into @xmath15-dependent and @xmath15-independent parts does not depend much on @xmath37 .
we therefore make the approximation that the flow of @xmath193 receives no contribution from bosonization , such that @xmath195 or @xmath196 the effective initial value @xmath197 can be computed by evaluating the momentum dependence of the four - fermion interaction in the scalar channel . at present
, it remains a free parameter . in a more complete calculation
one should also evaluate the diagrams in fig . 4 for non - zero external momenta and
choose the flow of @xmath192 and @xmath193 such that the flow of the vertex @xmath190 is well approximated by eq .
( [ s3a ] ) .
we expect that the resulting flow for @xmath193 will be attracted to an approximate fixed point .
we note that for a fixed ( point ) value of @xmath193 the mass term @xmath198 decreases roughly @xmath199 .
( 320,120)(0,0 ) ( 20,40)(45,40 ) ( 45,40)(95,40 ) ( 95,40)(120,40 ) ( 45,40)2 ( 95,40)2 ( 70,40)(26,180,0 ) ( 70,40)(24,180,0 ) ( 20,37)[tl]@xmath53 ( 45,37)[t]@xmath126 ( 70,37)[t]@xmath54 ( 95,37)[t]@xmath126 ( 120,37)[tr]@xmath53 ( 70,70)[b]@xmath64 ( 300,10)(300,35 ) ( 300,35)(300,85 ) ( 300,85)(300,110 ) ( 300,35)(250,35)3 ( 250,85)(300,85)3 ( 250,60)(25,90,270 ) ( 250,60)(25,270,90 ) ( 304,22)[tl]@xmath53 ( 304,60)[l]@xmath54 ( 304,98)[bl]@xmath53 ( 275,32)[t]@xmath55 ( 275,88)[b]@xmath55 ( 279,60)[l]@xmath53 ( 222,60)[r]@xmath54 ( 10,115)[tl](a ) ( 190,115)[tl](b ) ( 160,60)[t]@xmath200 the formulation in terms of a composite scalar field allows an inclusion of the scalar field fluctuations as well .
this adds new diagrams , which can be viewed as a partial resummation of the box diagrams in fig .
4 . in particular , the quark wave function renormalizations @xmath201 get additional contributions from the scalar exchange diagrams in fig . 7 ( with scalars represented as double lines ) , as well known from computations in the standard model . in consequence ,
the anomalous dimensions acquire a scalar contribution with opposite sign to the tensor contribution .
@xmath202 @xmath203 with @xmath204 and @xmath205 } { 1+\tilde{m}_t^2+\tilde{m}_\varphi^2 } .\ ] ] for @xmath206 the scalar correction to @xmath207 becomes approximately @xmath208 .
as long as @xmath107 remains small , the effective two - loop contribution corresponding to scalar exchange ( cf . fig .
7b ) remains subleading .
however , for large @xmath107 the scalar contribution to @xmath207 may become important and modify the anomalous dimensions towards positive values .
this could contribute to a final stop of the increase of @xmath107 which obeys now @xmath209 with @xmath207 given by eq .
( [ s4 ] ) .
( 200,100)(0,0 ) ( 75,50)2 ( 125,50)2 ( 100,50)(25,180,0 ) ( 100,50)(25,0,180 ) ( 35,51)(75,51 ) ( 35,49)(75,49 ) ( 125,51)(165,51 ) ( 125,49)(165,49 ) ( 145,55)[b]@xmath64 ( 55,55)[b]@xmath64 ( 72,47)[tr]@xmath126 ( 128,47)[tl]@xmath126 ( 100,22)[t]@xmath54 ( 100,78)[b]@xmath210 we next turn to the scalar contributions to the running of @xmath198 and @xmath188 .
the fermion loop correction to the inverse scalar propagator , as depicted in fig . 8 ,
results in the standard result for the anomalous dimension of the scalar field , @xmath211 due to the yukawa coupling @xmath212 a positive non - vanishing @xmath188 is generated , even if it is absent at some microscopic scale .
this produces a pole in the scalar propagator for @xmath213 , such that the scalar behaves as a dynamical particle .
the contribution of the quark loop shown in fig .
8 to the flow of the scalar mass term is @xmath214 .
within functional renormalization it has been investigated in @xcite and one finds with our cutoff in the fermion propagator @xmath215 @xmath216 the momentum integral for the contribution of fig .
8 to the flow of @xmath198 has been cut at some effective scale @xmath217 . here
@xmath217 is a characteristic scale below which the description of the flow in terms of scalar fluctuations becomes a reasonable approximation
. this should be somewhat above @xmath1 , but the precise value remains somewhat arbitrary without additional computations .
( in any case the use of an improved infrared cutoff within functional renormalization would remove this spurious dependence on a scale . ) for sufficiently large @xmath192 the flow ( [ s8 ] ) drives @xmath193 to negative values , indicating the onset of spontaneous symmetry breaking . this is the same physics that is responsible for the nontrivial solutions of the schwinger - dyson equation for @xmath11 in the preceding section . indeed , for a qualitative picture we can take for @xmath218 a constant @xmath212 and @xmath219 , and neglect @xmath220 as well as the contribution from bosonization .
this replaces eq .
( [ s8 ] ) by @xmath221 for the solution one finds the critical value @xmath222 for which @xmath198 reaches zero for @xmath165 .
inserting @xmath223 and using @xmath224 this indeed corresponds to @xmath225 .
the vacuum expectation value @xmath226 , with @xmath227 the location of the minimum of the scalar effective potential , differs from zero if @xmath198 gets negative . for the computation of its value , which should be @xmath228175 gev in a realistic model ,
one further needs to compute the quartic scalar self interaction @xmath229 .
we can adjust the value of @xmath1 ( or the ultraviolet value @xmath230 ) such that the fermi scale @xmath231 has the correct value .
a particularly interesting quantity is the renormalized yukawa coupling @xmath60 ( [ s5 ] ) .
its value for @xmath232 determines the top quark mass in terms of the fermi scale @xmath233 in other words , the knowledge of @xmath234 is equivalent to a determination of the mass ratio @xmath235 , where we use @xmath236 , with @xmath63 the known weak gauge coupling ( @xmath237 ) .
the observational value is @xmath238 .
while the scale @xmath231 is set by dimensional transmutation and therefore a free parameter , a computation of @xmath239 is equivalent to a parameter - free pre "- diction for the mass ratio @xmath235 .
in our approximation we can infer the flow equation for @xmath192 from eq .
( [ s3 ] ) , @xmath240 -\frac{f^4}{\pi^2 } \tilde{m}_\varphi^2 \tilde{s}(\tilde{m}_t^2 ) .\end{aligned}\ ] ] ( red curves ) , @xmath60 ( blue curves in the left picture ) and @xmath193 ( picture on the right ) . at large values of @xmath37
we started with @xmath241 ( solid lines ) and @xmath242 ( dashed lines).,title="fig : " ] ( red curves ) , @xmath60 ( blue curves in the left picture ) and @xmath193 ( picture on the right ) . at large values of @xmath37
we started with @xmath241 ( solid lines ) and @xmath242 ( dashed lines).,title="fig : " ] we have solved the flow equations numerically until the onset of spontaneous symmetry breaking at @xmath243 where @xmath244 . for @xmath245 one
has @xmath206 such that many threshold functions equal one .
we display the running of @xmath60 and @xmath0 in fig .
9 . for @xmath246
one has to continue the flow in the regime with spontaneous symmetry breaking and non - zero @xmath247 , as well known from many studies of functional flow equations @xcite .
we will not do so here . in the present approximation to the flow equations we observe an increase of @xmath0 and @xmath60 to very large values as @xmath248 approaches zero .
we do not expect our approximations to remain valid for large couplings , even though the one - loop form of the functional flow equations is exact @xcite .
one of the main shortcomings is the inaccurate truncation of the general form of momentum dependence for the fermion propagators and non - local interactions in a region where the anomalous dimensions @xmath249 are of the order one .
for example , the quartic interactions for the bare quarks in the tensor channel replaces in our approximation @xmath250 in eq .
( [ a2 ] ) .
we may consider an effective momentum dependence of @xmath251 given by the anomalous dimension @xmath252 evaluated for @xmath253 , i.e. @xmath254 - this would indeed be a valid approximation for small @xmath255 and @xmath256 .
however , for @xmath257 the momentum dependence of @xmath251 would effectively cancel the nonlocality @xmath258 of the interaction .
for a quasi - local interaction in the tensor channel we expect strong modifications of the flow of @xmath0 - for example , it could reach a fixed point @xmath259 , replacing in the interaction @xmath260 .
( in the language of chiral tensor fields this would correspond to the generation of a non - local mass for the chirons " @xcite . ) for a realistic model of electroweak interactions the increase of the yukawa coupling should stop or be substantially slowed down in the vicinity of its final value at @xmath232 , say for @xmath261 . for the corresponding region in the scale @xmath37 ,
@xmath262 we find values @xmath263 .
( we use @xmath264 for the quantitative estimates . ) in view of our discussion , it seems not unreasonable that the non - perturbative infrared physics stops the further increase of @xmath0 and @xmath60 in this range . if this happens , the ratio @xmath235 may come out in a reasonable range . a typical value of the scale for non - perturbative physics where the increase of @xmath0 and @xmath60 stops , may be @xmath265 .
at this scale the quantity @xmath166 ( cf .
( [ s8a ] ) with @xmath266 has reached a value @xmath267 , not too far from the critical value in the schwinger - dyson approach .
it is well conceavable that the top quark mass is substantially below the scale @xmath268 , such that the region of strong interactions may correspond to the multi - tev - scales and not disturb too much the lep precision tests of the electroweak theory .
our computation of the flow of @xmath0 and the yukawa coupling @xmath60 has further substantial quantitative uncertainties .
for example , we have neglected effects from the interactions in sect . 3 that correspond to the exchange of color - octet scalars or vectors .
this may be motivated for the region of @xmath37 where @xmath248 is already small , since a resonance type behavior and spontaneous symmetry breaking is only expected in the scalar singlet sector .
on the other hand , these contributions may still play a role in the interesting region where the increase of @xmath0 may stop .
we conclude that models for quarks and leptons with a non - local four - fermion interaction in the tensor channel appear to be promising candidates for an understanding of the electroweak symmetry breaking .
no fundamental higgs scalar is needed .
the theory is asymptotically free and generates an exponentially small chiral scale " @xmath1 where the dimensionless coupling @xmath0 grows large .
furthermore , a strong interaction in the scalar channel is generated at scales where @xmath0 is large .
a solution of schwinger - dyson equations suggests that this interaction triggers the spontaneous breaking of the electroweak symmetry at a scale determined by @xmath1 .
this would solve the gauge hierarchy problem .
we have also introduced a composite higgs scalar and investigated the flow of its mass .
we find indeed spontaneous symmetry breaking with a fermi scale somewhat below the chiral scale .
our first attempt of an estimate of the yukawa coupling of the top quark is encouraging , yielding a reasonable range for the ratio @xmath235 .
the model has also other interesting features .
it was shown @xcite that the flavor and cp - violation is completely described by the ckm matrix @xcite .
masses of the light quarks and leptons arise from appropriate four - fermion couplings @xcite .
first phenomenological constraints from lep precision tests and the anomalous magnetic moment of the muon have been computed @xcite . at the present stage
the understanding of the strong interactions around the scale @xmath1 , which would be a few tev in a realistic model , is still in its infancy . for this reason
our estimate of @xmath235 is only very crude .
nevertheless , no free parameter enters in the determination of this ratio in our model .
if the strong interactions can be understood quantitatively , our model leads to a unique pre "- diction of @xmath235 .
it would also predict the masses and interactions of the composite scalar fields .
* appendix a : non - local four - fermion interaction and chiral tensor fields * + in this appendix we show the equivalence of the microscopic action with a theory of chiral tensor fields @xcite .
this is similar to the equivalence of quantum electrodynamics to a theory with a non - local four - fermion interaction , generated by integrating out the photon .
we may start from the non - local four - fermion interaction and obtain the chiral tensor theory by a hubbard - stratonovich transformation .
we proceed here in the opposite way , starting with a chiral tensor theory and integrating out the chiral tensors .
starting point is the theory of chiral antisymmetric tensor fields investigated in ref .
we consider a complex antisymmetric tensor field @xmath269 which is a doublet of weak isospin and carries hypercharge @xmath270 .
the field can be decomposed into two parts which correspond to irreducible representations of the lorentz group .
the two parts are the
chiral " components of @xmath271 : @xmath272 these components can be written as @xmath44 matrices acting in the space of dirac spinors via @xmath273 where @xmath274 $ ] .
the matrix @xmath275 ( @xmath276 ) acts only on left - handed ( right - handed ) fermions , i.e. @xmath277 one may introduce an interaction between the fermions and the chiral tensors : @xmath278 here the chiral couplings @xmath279 are @xmath35 matrices in generation space , @xmath280 are the left - handed quark doublets , @xmath281 ( @xmath282 ) are the right - handed up - type ( down - type ) quarks , @xmath283 are the left - handed lepton doublets , @xmath284 are the right - handed electron - type leptons , and we defined @xmath285 where the transposition @xmath286 and the pauli matrix @xmath287 act in weak isospin space , i.e. on the two components of the weak doublet @xmath288 . the fields @xmath289 can be represented by three - vectors @xmath290 , @xmath291 rewriting the kinetic term @xmath292 in terms of the @xmath293-fields and in fourier space gives @xmath294.\ ] ] the propagator @xmath295 has the properties @xmath296 in the following we ignore for simplicity all couplings except @xmath297 , i.e. we assume @xmath298 in this case , only the @xmath299 fields couple to top quarks @xmath33 and left - handed bottom quarks @xmath34 .
all other fields are free if we ignore gauge couplings .
we denote the two weak isospin components of @xmath299 as @xmath57 and @xmath55 , since they correspond to electric charge @xmath300 and @xmath301 after electroweak symmetry breaking .
the action for the chiral interactions is then @xmath302,\ ] ] in a spinor basis in which @xmath303 the matrices @xmath45 are defined in terms of the pauli matrices @xmath46 as @xmath304 in the action @xmath305 , a summation over quark color is understood .
the classical field equations are obtained by varying @xmath306 with respect to the @xmath299 fields .
they determine these fields as functionals of the quark fields . in momentum space , these relations are @xmath307 we then insert these relations into @xmath308 and obtain the action @xmath52 of the non - local four - fermion interaction given in eq .
( [ a2 ] ) . in a functional integral formulation
this procedure is equivalent to performing the gaussian integral over the @xmath299 fields .
+ * appendix b : one loop expressions for the effective action * + in this appendix we evaluate the interaction contribution @xmath87 to the inverse propagator @xmath71 in eq .
( [ c2 ] ) for the different quark types separately , ( @xmath309 ) .
one obtains @xmath310 \left ( -\frac{p_{kl}^*(q)}{q^4}\right)\\ \nonumber & + \delta_{cc'}\bar{t}^{c''}_\gamma(q+p')t^{c''}_\eta(q+p ) \left [ \sigma^k_{+ \alpha\beta } \sigma^l_{- \gamma\eta } + \sigma^l_{- \alpha\beta } \sigma^k_{+ \gamma\eta}\right ] \frac{p_{kl}^*(p'-p)}{(p'-p)^4}\\ \nonumber & + \bar{b}^{c'}_\gamma(q+p')b^c_\eta(q+p ) \sigma^k_{+ \alpha\eta } \sigma^l_{- \gamma\beta } \left ( -\frac{p_{kl}^*(q)}{q^4}\right)\bigg\},\end{aligned}\ ] ] @xmath311 @xmath312 \sigma^k_{+\alpha\beta } \frac{p_{kl}^*(p'-p)}{(p'-p)^4},\ ] ] @xmath313 \sigma^k_{-\alpha\beta } \frac{p_{kl}^*(p'-p)}{(p'-p)^4},\ ] ] @xmath314 \frac{p_{kl}^*(q)}{q^4},\ ] ] @xmath315 \frac{p_{kl}^*(q)}{q^4},\ ] ] @xmath316 @xmath317 @xmath318 inserting this into the expression ( [ c4 ] ) , we find an anomalous dimension term for @xmath33 and for the left - handed @xmath34 @xmath319 the term @xmath90 in eq .
( [ c4 ] ) produces an effective four fermion interaction @xmath320 with @xmath321 and @xmath322 color indices are suppressed , since all pairs of fermions in large brackets are color singlets .
gross and f. wilczek , phys .
* 30 * , 1343 ( 1973 ) ; + h.d .
politzer , phys .
30 * , 1346 ( 1973 ) v.a .
miransky , m. tanabashi and k. yamawaki , phys .
* b 221 * , 177 ( 1989 ) ; mod .
phys . lett . *
a 4 * , 1043 ( 1989 ) ; w. bardeen , c. hill and m. lindner , phys .
* d 41 * , 1647 ( 1990 ) y. nambu and g. jona - lasinio , phys .
* 122 * , 345 ( 1961 ) h. gies , j. jaeckel and c. wetterich , phys . rev . *
d 69 * , 105008 ( 2004 ) c. wetterich , phys .
* d 74 * , 095009 ( 2006 ) ; mod .
lett . * a 23 * , 677 ( 2008 ) c. wetterich , int .
j. mod .
phys . * a 23 * , 1545 ( 2008 ) f.j .
dyson , phys .
rev . * 75 * , 1736 ( 1949 ) ; + j.s .
schwinger , proc .
sci . * 37 * 452 ( 1951 ) c. wetterich , phys .
* b 301 * , 90 ( 1993 ) h. gies and c. wetterich , phys . rev .
* d 65 * , 065001 ( 2002 ) c. wetterich , z. phys . * c 48 * , 693 ( 1990 ) j. berges , n. tetradis and c. wetterich , phys .
rept . * 363 * , 223 ( 2002 ) m. kobayashi and t. maskawa , prog .
. phys . * 49 * , 652 ( 1973 ) ; + n. cabibbo , phys .
* 10 * , 531 ( 1963 ) c. wetterich , arxiv:0709.1102 [ hep - ph ] j .- m . schwindt and c. wetterich , int .
j. mod .
phys . * a 23 * , 4345 ( 2008 ) b. de wit and j.w .
van holten , nucl .
* b 155 * , 530 ( 1979 ) ; e. bergshoeff , m. de roo and b. de wit , nucl .
* b 182 * , 173 ( 1981 ) ; i.t .
todorov , m.c .
mintchev and v.b .
petkova , _ conformal invariance in quantum field theory _ , p.89 ( ets , pisa , 1978 ) ; m.v .
chizhov , mod .
lett . * a 8 * , 2753 ( 1993 ) ; l.v .
avdeev and m.v .
chizhov , phys .
* b 321 * , 212 ( 1994 ) n. kemmer , proc .
r. soc . * a 166 * , 127 ( 1938 ) ; v.i .
ogievetsky and i.v .
polubarinov , yad .
fiz . * 4 * , 216 ( 1966 ) ; a. salam and e. sezgin ( eds . ) , _ supergravities in diverse dimensions _ ( north holland and world scientific , 1989 ) ; m. kalb and p. ramond , phys . rev .
* d 9 * , 2273 ( 1974 ) ; e. cremmer and j. scherk , nucl .
* b 72 * , 117 ( 1974 ) ; s.a .
frolov and a.a .
slavnov , phys .
* b218 * , 461 ( 1989 ) | we investigate the fermions of the standard model without a higgs scalar . instead , we consider a non - local four - quark interaction in the tensor channel which is characterized by a single dimensionless coupling @xmath0 .
quantization leads to a consistent perturbative expansion for small @xmath0 .
the running of @xmath0 is asymptotically free and therefore induces a non - perturbative scale @xmath1 , in analogy to the strong interactions .
we argue that spontaneous electroweak symmetry breaking is triggered at a scale where @xmath0 grows large and find the top quark mass of the order of @xmath1 .
we also present a first estimate of the effective yukawa coupling of a composite higgs scalar to the top quark , as well as the associated mass ratio between the top quark and the w boson .
_ institut fr theoretische physik , philosophenweg 16 , 69120 heidelberg , germany _ _ e - mail : [email protected] , [email protected]_ |
hubble space telescope ( hst ) observations and the development of new powerful high - resolution spectrographs at large ground - based telescopes demonstrate that gcs do not consist of stars of a single age and metallicity ( see the review paper by gratton et al .
( 2012 ) and references therein ) .
in addition to stars having a chemical composition similar to that of the galactic halo , gcs contain stars built of the matter that has passed through the full @xmath8-cycle and proton capture of light nuclei .
contrary to the field stars , they often show sodium - oxygen anti - correlations ( gratton et al .
( 2012 ) ) .
anomalies of chemical and stellar composition are found for many galactic gcs .
explaining the origin and chemical evolution of stars in gcs is very important for understanding the processes of nucleosynthesis and chemical evolution in the early universe .
although deep stellar photometry is available for many galactic gcs thanks to hst high - resolution images , high resolution spectroscopy in a wide spectral range for many faint stars in gcs is still very expensive .
this is why we decided to explore the possibility of studying the detailed chemical composition in gcs using medium - resolution spectra . in the present paper ,
we fit medium - resolution spectra of ngc2419 integrated along different slit positions with stellar atmosphere models taking into account the stellar mass function .
ngc2419 is a unique massive gc ( @xmath9 ) in the outer galactic halo , at a distance of 84 kpc .
its large half - light radius ( @xmath10 pc , ibata et al . ,
2011 ) is similar to that of ultra - compact dwarf galaxies .
the stellar content and mass function , the chemical composition and the structure of ngc 2419 have been studied actively . the blue horizontal branch ( bhb ) of the cluster is very extended ( di criscienzo et al .
( 2011 ) ) .
the stellar mass function and the properties of horizontal branch ( hb ) stars , hertzsprung gap stars and blue stragglers have been explored in a number of papers using deep color - magnitude diagrams ( @xmath11 ) ( harris et al .
1997 , dalessandro et al .
2008 , di criscienzo et al .
2011 , bellazzini et al .
these studies gave information about the age and metallicity of ngc 2419 .
the helium and alpha - element abundances were estimated in the aforementioned papers by comparison of the @xmath11 of ngc 2419 with model isochrones ( pietrinferni et al .
( 2004 , 2006 ) ) .
an unusually high mass - to - luminosity ( m / l ) and the possible presence of a black hole in the center of the cluster were suggested by investigators of its stellar dynamics ( ibata 2011a , b ) .
measurements of stellar velocities at higher spectral resolution ( baumgardt et al . , 2009 )
did not favor the last hypothesis .
so far detailed chemical compositions are known only for brightest red giant stars of ngc 2419 ( cohen et al .
( 2011 ) , shetrone et al .
( 2001 ) ) . in the latter papers
the unusual chemical composition and other properties of the cluster were interpreted as a signature of its possible extragalactic origin . in this paper
we carry out population synthesis modeling of medium - resolution spectra of ngc2419 .
our main aim was to obtain theoretical spectra in good agreement with the observed ones .
this task is solved for the first time .
the lowest resolution used in the literature for such a task was @xmath12 ( colucci et al . , 2012 ) .
at variance with the population systhesis modeling of schiavon et al .
( 2002 ) , we use theoretical spectra of stars calculated using stellar atmosphere models .
medium resolution allows us to extend the method to extragalactic gcs at distances of @xmath13 mpc . in section 2
we describe our observational material and procedure of spectra reduction . in section 3
we explain our method of theoretical population synthesis . in section 4
we derive the age and metallicity of ngc2419 and a mean helium abundance by comparison of theoretical and observed balmer line profiles . in section 5
we describe how we derive chemical abundances for the main chemical elements by fitting absorption - line features in the observed spectra with theoretical ones .
our conclusions are formulated in section 6 .
medium - resolution long slit ( @xmath14 ) spectra of ngc 2419 were obtained with the spectrograph carelec ( lemaitre et al .
( 1990 ) ) at the 1.93-m telescope of the haute - provence observatory .
the total exposure time was 2.6 h ( table 1 ) , with a seeing of @xmath15 .
the grism 300 grooves / mm and the slit width @xmath16 provided a resolution of @xmath17 pix@xmath18 , fhhm@xmath19 and spectral range @xmath20 .
he and ne lamps were exposed at the beginning and the end of each night for wavelength calibration .
the spectrophotometric standard star hr1544 ( hamuy et al . , 1992 , 1994 ) was observed for relative calibration of the spectral sensitivity of the spectrograph .
lick standard stars ( worthey et al .
( 1994 ) , table 1 ) were observed each night for calibration of radial velocities and spectrophotometric indices .
note that our observing program also included other distant galactic gcs .
separate papers will be devoted to their spectroscopic analysis .
data reduction and analysis were carried out using midas ( european southern observatory munich image data analysis system ) ( banse et al . , 1983 ) and iraf ( image reduction and analysis facility ) . cosmic - ray rejection was made using the program _
filter / cosmic _ in midas .
a map of hot pixels was prepared prior to the masking process .
the accuracy of the dispersion solution obtained using the he - ne lamp is @xmath21 .
possible systematic uncertainties may arise due to differences in optical paths from the lamp to the spectrograph and from a position on the sky to the spectrograph .
such uncertainties were corrected using wavelengths of prominent emission night - sky lines , e.g. @xmath22 5577 $ ] .
one - dimensional spectra were extracted using the procedure _ apsum _ in iraf .
this program allows one to correct the curvature of spectra along the dispersion , which may arise due to a slight slope of the slit , or as a result of the influence of atmosphere dispersion if the slit position angle has a large parallactic angle ( pa > > 0 ) .
the signal - to - noise ratio ( s / n ) achieved in the summed one - dimensional spectrum is @xmath23 in the middle of the spectral range .
the heliocentric radial velocity of ngc 2419 derived by cross - correlation using radial velocity standard stars and the program _
fxcor _ in iraf agrees well with the literature one : @xmath24 km s@xmath18 ( harris , 1996 ) . for calculating this value using bright night - sky emission lines we took into account the heliocentric correction ( stumpff , 1980 ) and possible systematics of our wavelength scale with respect to the standard one .
modeling the gc integrated light was based on synthetic spectra for stars with physical parameters corresponding to the selected isochrone .
we use isochrones of bertelli et al .
( 2009 ) having a wide ranges of age ( @xmath25 ) , heavy element abundance ( @xmath26 ) and helium abundance ( @xmath27 ) .
after initialization of a specified isochrone , its optimization was carried out to reduce the amount of calculations .
for the stars ( places of isochrones ) which contribute less than @xmath28 to the cluster light we use the following maximum steps of temperature and surface gravity : @xmath29 and @xmath30 . for
the other points the steps were twice as large .
the places of isochrones with the initial stellar masses which contribute less than @xmath31 to the cluster mass were excluded from the calculations .
finally the resulting isochrones contained 55 - 65 points .
so possible errors of modeling caused by the optimization procedure were reduced to @xmath32 of the continuum flux .
we used a modified @xmath33 package ( shimansky et al . , 2003 ) for calculating synthetic blanketed spectra of stellar atmospheres .
plane - parallel , hydrostatic stellar atmosphere models for a given initial set of parameters ( @xmath34 , @xmath35 , @xmath36 ) were obtained by interpolation of a model grid of castelli & kurucz ( 2004 ) using a method described in suleimanov ( 1996 ) .
we used _ the solar _ chemical composition and isotopic abundance of some chemical elements ( anders & grevesse , 1989 ) . for each stellar atmosphere we derived fluxes propagating in three basic directions with angles to the stellar surface
@xmath37 , @xmath38 , @xmath39 .
then the stellar surface was divided into sectors with a step of @xmath40 for all orienting angles .
for each corner the area and the angle of visibility were calculated .
the radiation intensity of the sector reaching an observer was found by interpolation of the specific intensities in the three basic angles according to the actual visibility angle .
the integrated stellar radiation was obtained by integration of the radiation from each sector according to the area and local radial velocities caused by the rotation of the star and radial tangential macro - turbulence .
finally , the obtained spectra were broadened according to the instrumental spectral - line profile of the spectrograph , using a gaussian function with a half - width corresponding to the resolution . before normalization , spectra with and without atomic lines and molecular bands
were calculated simultaneously in the spectral range @xmath41@xmath42 .
modeling the spectra with different wavelength steps @xmath43 has shown that @xmath44 is the maximum value when errors in the line profiles are less than @xmath32 of the continuum flux .
we took into account all sources of continuous opacity tabulated in the packages @xmath45 ( suleimanov , 1992 ) and @xmath33 ( sakhibullin & shimansky , 1997 ) , approximately 570000 atomic lines from the list of kurucz ( 1994 ) and 28 molecular bands in the optical spectral range calculated according to the theory of nersisyan et al .
( 1989 ) kindly provided by ya.v .
we used the empirical oscillator strengths calculated by shimansky ( 2011 ) for the 750 most intense lines .
additionally , we found that it was necessary to take into consideration the @xmath46 molecule , powerful bands of which are seen in the spectrum of ngc 2419 in the range @xmath47@xmath48 .
therefore we included calculation of the opacities for 75000 molecular bands @xmath49 and @xmath50 from the catalogs of kurucz ( 1994 ) in the package @xmath33 .
modeling of the balmer - line profiles was done using of the stark broadening theories of vidal , cooper & smith ( 1973 ) and griem ( 1960 ) . for the other lines we calculated the standard voigt profile taking into account the following information : doppler broadening of lines due to thermal motion of atoms , microturbulent velocity @xmath51 , natural damping , stark broadening according to the approximation of kurucz & furenlid ( 1979 ) and the van der waals broadening constants defined using the classical formula unsold ( 1955 ) with the the scaling factor in the range @xmath52@xmath53 ( shimansky , 2011 ) .
we modeled the line profiles of chemical elements taking into account departures from local thermodynamic equilibrium ( lte ) according to the methods developed by ivanova et al .
( 2004 ) and shimansky et al .
( 2008 ) for binary stellar systems .
the non - lte atomic populations for each atmosphere model were calculated using the linearization method by auer & heasley ( 1976 ) with the package @xmath54 of sakhibullin ( 1983 ) .
we used : @xmath55-level model of the @xmath56 atom ( shimansky et al .
, 2008 ) , @xmath57-level model of the @xmath58 atom ( mashonkina et al . , 2000 ) , @xmath59-level model of the @xmath60 atom ( shimansky et al . , 2000 ) , @xmath61- level model of the @xmath62 ion ( shimansky , 2002 ) , @xmath63-level model of the @xmath64 atom ( menzhevitski et al . ,
2012 ) , @xmath65-level model of the @xmath66 atom ( ivanova et al . , 2000 ) and @xmath67-level model of the @xmath68 ion ( ivanova et al . , 2002 ) .
the influence of non - lte effects on line profiles for different chemical elements in the summed spectrum of ngc 2419 is shown in fig.1 for the physical parameters and chemical composition obtained by us ( see sec . 4 ) .
note that non - lte effects do not affect the hi line intensities ( increase by @xmath69 of the central line depth ) due to the dominance of hi in almost all stellar atmospheres of the cluster .
lines of heavy elements are enhanced with very different amplitudes taking into account non - lte effects .
the intensity of the resonance doublet @xmath70grows a lot ( up to @xmath71 of the central depth ) . as a result ,
the sodium abundance is reduced by @xmath72 dex .
the non - blended resonance line @xmath73 is significantly influenced by non - lte effects ( up to @xmath74 of the central depth ) . on the other hand , the lines @xmath60 and @xmath62 do not change .
unfortunately , we have no near - ir spectra containing the lines @xmath75 , @xmath76 having significant non - lte corrections .
the calculated spectra for stars belonging to the selected isochrone were summed according to their radii and initial mass function @xmath77 . in this work the chabrier ( 2005 )
initial mass function ( imf ) was used @xmath78 @xmath79 where @xmath80 .
it gives a good approximation of the stellar luminosity function ( lf ) found in the literature for ngc 2419 ( bellazzini et al .
the chabrier imf matches well the theoretical mass distributions of padoan & nordlund ( 2004 ) for lowest mass stars .
finally , normalization of the integrated spectrum was done by division of the spectrum with its lines and molecular bands by the corresponding continuum .
the isochrone used for the cluster is defined by its parameters ( age @xmath25 , helium abundance @xmath27 and metallicity @xmath26 ) , where @xmath81 corresponds to the hydrogen abundance , @xmath27 stands for the helium abundance , and @xmath26 means the heavier elements abundance .
the values for the sun are @xmath82 and @xmath83 ( anders & grevesse , 1989 ) ] which may be derived by comparison of the observed and theoretical spectra .
the index @xmath26 of heavy - element abundance is related to the chemical composition in the atmospheres of stars belonging to the cluster ( @xmath84 $ ] ) : @xmath85}$ ] . it can be obtained with an accuracy of @xmath86 dex from the analysis of synthetic spectra ( see below ) . to define the initial @xmath25 and @xmath27 we adopt the metallicity from the literature @xmath87 ( bellazzini et al . , 2012 ) .
note that chemical elements synthesized in deep stellar layers of some red giant and blue horizontal branch ( bhb ) stars may be delivered to the surface and change the initial chemical composition ( e.g. smiljanic et al . , 2007 ) .
however , our analysis shows that an upper limit for the contribution of these stars to the total flux in the spectral range @xmath88@xmath42 is only @xmath89 .
we show in fig . 2 the contribution of different parts of the isochrone to the total spectrum .
if we assume that chemical anomalies due to the mixing effect do not exceed @xmath90 dex , the errors @xmath91 are @xmath92 ( pereira et al . , 2011 ) and
the determined abundances of chemical elements for the cluster are close to their initial values .
we derive @xmath27 and @xmath26 as follows .
we vary them until the theoretical profiles for the lines @xmath93 , @xmath94 , @xmath95 fit the observed ones , because the @xmath27 content influences the shape of the balmer lines .
additionally , it is important to ensure that the theoretical profiles for the lines @xmath96 and @xmath97 fit well the observed ones for a given ca content .
if the ionization balance between @xmath98 and @xmath99 is reached , our calculations are considered to be near completion . in fig . 2 ( a ) we present isochrones for different ages and @xmath27 .
the best - fit age and he abundance for ngc 2419 are @xmath100 and @xmath101 . in fig . 2 ( b ) we show the theoretical line profiles @xmath93 and
@xmath95 compared to the observed ones . when reducing the age , the temperature of stars increases , and the main sequence turn - off point ( msto ) gets bluer .
this leads to an increased contribution of early - type main sequence @xmath102-dwarfs to the integrated light of the cluster .
these stars have strong balmer lines . as a result of reducing the age ,
all @xmath103 lines appear to be more intense simultaneously in the doppler cores and in the wings . on the other hand ,
increasing the he abundance does not influence strongly the shape of the isochrone ( position of the main sequence , sub - giant and giant branches ) but it increases the luminosity of hot hb stars . it is known that the doppler cores of the balmer lines are deep in the spectra of such stars , because gas pressure is not strong in their atmospheres . therefore
, increasing @xmath27 leads to a sharp increase of the central intensities of the balmer lines and a weakening of their wings .
note that @xmath95 increases less than @xmath93 ( see fig . 2 ( b ) ) . in summary , @xmath25 and @xmath27 influence
differently the @xmath103 line profiles , and this fact allows us to disentangle these two parameters .
finally , we obtained @xmath104 and @xmath105 . with these values
the central depths and the equivalent widths ( @xmath106 ) for the observed and theoretical hi line profiles agree with an accuracy of @xmath107 and @xmath74 , respectively .
the obtained age of ngc 2419 agrees well with the results of published photometric studies .
bellazzini et al .
, 2012 analyzed narrow - band @xmath108 stellar photometry .
they argue that the age of the cluster is in the range @xmath109@xmath110 using isochrones of dotter et al .
( 2007 ) with [ fe / h]@xmath111 , @xmath4=0.4 $ ] , @xmath112 and of pietrinferni et al .
( 2004 ) with [ fe / h]@xmath113 , @xmath4=0.0 $ ] , @xmath114 . the same age range @xmath112@xmath115 was obtained earlier by harris et al .
( 1997 ) using hst stellar photometry data .
we determined the mean chemical abundances in stellar atmospheres in a regular way by fitting the observed line profiles with theoretical ones .
however the moderate spectral resolution did not allow us to determine the abundances using individual lines and molecular bands merging in blends .
this is why we determined the abundance of each element taking into account the agreement between the observed and theoretical spectra in all blends containing its lines in the studied spectral range . in this fitting process
all chemical elements were divided into 5 groups : \1 ) this group is composed of elements ( @xmath116 , @xmath117 , @xmath118 ) with lines and molecular bands dominating in spectra .
it is possible to derive unambiguous estimates of the abundances using these lines .
they are accurate ( up to @xmath119 dex ) .
comparison of the observed and the theoretical spectra calculated for different @xmath120 $ ] is shown in fig .
\2 ) this group consists of elements ( @xmath121 , @xmath122 , @xmath123 , @xmath124 , @xmath125 , @xmath126 , @xmath127 ) with weak and blended lines and molecular bands .
their abundances were derived taking into account the abundances for the elements of the first group derived by us .
the accuracy of abundances in this case is lower ( @xmath128 dex ) .
\3 ) this group consists of elements with spectroscopic features too weak to be detected in the studied spectral range .
however , they influence the molecular and ionization equilibrium of other elements .
it is possible to obtain their approximate abundances from the analysis of this equilibrium .
for example , we argue that part of the c atoms forms the @xmath129 molecule if the o abundance is enhanced .
this reduces the intensity of the molecular bands @xmath130 and @xmath46 .
therefore , testing the equilibrium between different observed and theoretical intensities of these bands allows one to estimate abundances of all @xmath8 group elements .
the @xmath8 abundances are interdependent .
\4 ) sodium ( @xmath131 ) with lines strongly distorted by interstellar components was considered in this group .
\5 ) this group consists of elements ( @xmath132 , @xmath133 , @xmath134 , @xmath135 ) which do not have lines detectable at medium resolution .
however , they influence the equilibrium of electrons in stellar atmospheres . abundances of these elements were considered to be equal to that of other elements of this group .
the microturbulent velocity @xmath51 was considered to be equal for all stars of the cluster .
we derived it from the best fit of weak and strong fe lines in the observed spectra .
comparison between the observed and theoretical spectra calculated with different @xmath51 and @xmath120 $ ] is shown in fig .
4 . as a result of this analysis ,
a mean microturbulent velocity was defined to be @xmath136 km / s .
note that it is not quite correct to use such value @xmath51 , bearing in mind a variety of stellar luminosities and temperatures in the cluster .
nevertheless , our calculations have shown that the hottest msto stars ( @xmath137k , @xmath138 ) and red giant branch ( rgb ) stars ( @xmath139@xmath140k , @xmath141@xmath142 ) are the main contributors to the observed spectrum . for such stars
our adopted value of @xmath51 corresponds well to the typical values found in the literature ( cayrel et al .
, 2004 , luck & heiter , 2005 ) .
a final comparison of the observed and theoretical spectra is shown in fig .
the obtained abundances are presented in fig . 6 and table 2 in comparison with the data from cohen
( 2011 ) and shetrone et al .
( 2001 ) .
one may conclude that the derived chemical composition of ngc 2419 agrees qualitatively with the data published for individual stars in the cluster .
cohen et al .
( 2011 ) obtained on average @xmath90 dex higher o and mg abundance for red giants in ngc 2419 .
the abundances of other elements agree with our estimates within the errors ( @xmath143 dex ) .
it is worth noting that the o abundance was estimated indirectly by us , using the influence of this element on the molecular bands @xmath46 , @xmath130 .
this estimate may not be very accurate .
note that cohen et al .
( 2011 ) determined @xmath144 $ ] studying the infrared triplet @xmath145 which is influenced strongly by non - lte effects .
we suggest that to obtain an accurate o abundance it is important to resort to non - lte analysis .
we derived the abundance of mg using a strong not - blended line @xmath146and a total flux distribution in the range of the index @xmath147 ( @xmath148 ) . the abundance of mg in cohen et al .
( 2011 ) was determined at high spectral resolution using strong lines including @xmath146 .
we calculated non - lte corrections for @xmath149 in atmospheres of different stars of the cluster and obtained values less than @xmath150=0.1 $ ] dex .
the reason for the high @xmath151 $ ] found in the literature was not clear to us .
it is worth noting that there are significant variations of the mg abundance in the red giants studied by cohen et al .
( 2011 ) , shetrone et al .
( 2001 ) and differences in the methods of measuring and the theoretical calculation of strong @xmath152 lines in the literature and in our analysis . when we finished writing this paper , papers by mucciarelli et al .
( 2012 ) and cohen & kirby ( 2012 ) were published .
a large dispersion of @xmath153 and @xmath118 abundances for rgb stars in ngc 2419 was discovered by these authors and may partially explain the reason of the low @xmath151 $ ] derived by us .
the chemical composition of ngc 2419 has features distinguishing it from typical old gcs , or from a number of field stars in our galaxy ( borkova & marsakov , 2005 , pritzl & venn ( 2005 ) , venn et al . ,
2004 , see also discussion in cohen et al .
( 2011 ) ) .
according to the present theory of chemical evolution for our galaxy , ( see the classical results of samland ( 1998 ) and later studies of chemical evolution of alibes et al . , 2001 and kobayashi et al .
, 2006 ) there should be overabundances of @xmath154 $ ] , @xmath1 $ ] , @xmath2 $ ] , @xmath155 $ ] , @xmath3 $ ] , and @xmath156 $ ] of an order @xmath157@xmath158 dex for stars with metallicity [ fe / h]@xmath159 dex .
additionally , a deficit of [ na / fe ] , [ al / fe ] of the order @xmath143 dex is also expected . other elements ( including @xmath160 ) have solar x / fe abundance ratios .
table 2 shows that both estimates of the c abundance for ngc 2419 indicate its extreme deficit @xmath154 < 0.7 $ ] dex which can not be explained by supply of the material synthesized in the last stages of evolution on its surface .
overabundances of @xmath3 $ ] , @xmath156 $ ] in stars of the cluster ( @xmath119@xmath143 dex ) appear to be much less than the ones suggested by samland ( 1998 ) .
@xmath2 $ ] has a value close to the solar one .
we have carried out theoretical population synthesis modeling for ngc 2419 , an unusual gc in the outer galactic halo , using medium - resolution spectra . to solve the task we developed a new method which may be applied to medium - resolution spectra ( @xmath161 ) with high signal - to - noise ratio ( @xmath162 ) in a wide spectral range @xmath163 .
the efficiency of this method is higher for higher spectral resolution , wider spectral range and larger @xmath164 .
the example considered in this paper is almost limiting for these parameters .
we used _ as an initial approximation _ the isochrone , the imf , and distribution the cluster stars according to their masses , radii , and @xmath165 using the published theoretical isochrone corresponding best to the observed @xmath11 ( dalessandro et al .
, 2008 ; bellazzini et al . , 2012 ) .
we summed the synthetic blanketed stellar spectra calculated using the models of stellar atmospheres according to the chabrier ( 2005 ) imf which gives a good approximation of the stellar lf found in the literature for ngc 2419 ( bellazzini et al . , 2012 ) .
fitting of the observed spectra by the theoretical ones was made iteratively in several stages . finally , we adopted the chemical composition of stars corresponding to the selected isochrone . to determine the age and helium content of the stars we analyzed mainly the depth and the shape of the wings of hi lines , because these are independent of other parameters .
then we varied the fe abundance , since these lines dominate the optical spectrum , even at low metallicity .
after optimization of the aforementioned main parameters ( age , @xmath165 , he and fe abundance ) , we fitted the c , mg , and ca abundances .
we found that molecular bands @xmath166 , @xmath167 and @xmath168 lines are some of the most prominent features in the spectrum of ngc 2419 . to derive abundances of other chemical elements it was necessary to analyze blends of many lines . finally , we managed to estimate abundances of 14 elements including 3 alpha - elements ( ca , o , mg ) and ti .
we obtained age=12.6 gyr , [ fe / h]@xmath169 dex , @xmath101 and [ @xmath5/fe]@xmath170 as the mean of @xmath1 $ ] , @xmath2 $ ] and @xmath3 $ ] . the derived chemical composition ( except for the alpha - elements )
agrees well with the published one , which was obtained using high - resolution spectra for 8 rgb stars of the cluster ( cohen et al .
( 2011 ) , shetrone et al .
( 2001 ) ) , and confirms the conclusion about the chemical peculiarity of ngc 2419 .
the abundance of mg ( @xmath2=-0.05 \pm 0.15 $ ] ) derived by us agrees within the errors with a mean value for 49 rgb stars obtained by mucciarelli et al .
( 2012 ) ( @xmath2=0.05 \pm 0.08 $ ] ) using high - resolution spectroscopy .
the method presented here employs information from all the stars in the cluster , and may be successfully applied to integrated - light spectroscopic studies of extragalactic gcs .
it is still impossible to obtain deep @xmath171 , or high - resolution , high @xmath164 spectra for such objects . in these cases , an initial approximation for the age
may be derived using lick indices , or full spectral fitting using simple stellar population models ( see e.g. sharina et al .
( 2010 ) and references therein )
. acknowledgements .
s.m.e . acknowledges partial support by the russian foundation for basic research ( rfbr ) grant 11 - 02 - 00639 , and a russian - ukrainian rfbr grant 11 - 02 - 90449 ; support of the ministry of education and science of the russian federation , the contract 14.740.11.0901 and proposal 2012 - 1.5 - 12 - 000 - 1011 - 004 .
s.v.v . acknowledges support by the rfbr grant 10 - 02 - 01145-a .
we thank th .
h. puzia for interesting comments .
a. alibes , j. labay , r. canal , 2001 , a&a , * 370 * , 1103 e. anders , n. grevesse , geochimica et cosmochimica acta , 1989 , * 53 * , 197 l.h .
auer , j. heasley , 1976 , apj , * 205 * , 165 k. banse , ph . crane , ch
. ounnas , d. ponz 1983 , // midas , in proc . of decus ,
zurich , p. 87
h. baumgardt , p. cote , m. hilker , m. rejkuba , s. mieske , s.g .
djorgovski , p. stetson , 2009 , mnras , * 396 * , 2051 m. bellazzini , e. dalessandro , a. sollima , r. ibata , 2012 , mnras , * 423 * , 844 g. bertelli , e. nasi , l. girardi , p. marigo , 2009 , a&a , * 508 * , 355 t. v. borkova , v. a. marsakov , 2005 , astronomy reports , * 49 * , 405 f. castelli , r.l .
kurucz , 2004 , iau symp .
210 `` modeling of stellar atmospheres '' .
r. cayrel , e. depagne , m. spite et al . , 2004 ,
a&a , * 416 * , 1117 g. chabrier , 2005 , the initial mass function 50 years later , astrophysics and space science library , * 327 * , 41 j.g .
cohen , e.n .
kirby , 2012 , apj * 760 * , 86 j.g .
cohen , w. huang , e.n .
kirby , 2011 , apj * 740 * , 60 j. e. colucci , r. a. bernstein , s. a. cameron , a. mcwilliam , apj , 2012 , * 746 * , 29 e. dalessandro , b. lanzoni , f.r .
ferraro , f. vespe , m. bellazzini , r.t .
rood , 2008 , apj * 681 * , 311 a. dotter , b. chaboyer , d. jevremovic et al .
, 2007 , aj , * 134 * , 376 m. di criscienzo , f. dantona , a. p. milone , p. ventura , v. caloi , r. carini , a. dercole , e. vesperini , g. piotto , 2011 , mnras * 414 * , 3381 r. gratton , e. carretta , a. bragaglia , 2012 , a&arv , 20 , 50 h.r .
griem , 1960 , apj , * 132 * , 883 m. hamuy , n.b .
suntzeff , s. r. heathcote , a. r. walker , p.gigoux , m. m. phillips , 1992 , pasp , * 104 * , 533 m. hamuy , n.b .
suntzeff , s. r. heathcote , a. r. walker , p.gigoux , m. m. phillips , 1994 , pasp , * 106 * , 566 w.e .
harris , 1996 , aj , 112 , 1487 ( 2010 edition ) w.e .
harris et al .
, 1997 , apj , * 114 * , 1030 r. ibata , a. sollima , c. nipoti , m. bellazzini , s.c .
chapman , e. dalessandro , 2011 , apj , * 738 * , 186 r. ibata , a. sollima , c. nipoti , m. bellazzini , s.c .
chapman , e. dalessandro , 2011 , apj , * 743 * , 43 d.v .
ivanova , n.a .
sakhibullin , v.v .
shimansky , 2004 , astronomy reports , * 48 * , 476 d.v .
ivanova , n.a .
sakhibullin , v.v .
shimansky , 2002 , astronomy reports , * 46 * , 390 d.v .
ivanova , v.v .
shimansky , astronomy reports , * 44 * , 376 c. kobayashi , h. umeda , k. nomoto et al . , 2006 , apj , * 653 * , 1145 r.l .
kurucz , i. furenlid , sample spectral atlas for sirius , 1979 , sao special rep . , * 387 * , 1 r.l .
kurucz , sao cd - roms , ( cambridge , ma02138 , usa ) g. lemaitre , d. kohler , d. lacroix , j .-
meunier , a. vin , astron .
astrophys . , 1990 , * 228 * , 546 r.e .
luck , u. heiter , 2005 , aj * 129 * , 1063 l.i .
mashonkina , v.v .
shimansky , n.a .
sakhibullin , 2000 , astronomy reports , * 44 * , 790 v.s .
menzhevitski , v.v .
shimansky , n.n .
shimanskaya , 2012 , astrophysical bulletin , * 67 * , 245 a. mucciarelli , m. bellazzini , r. ibata , t. merle , s. c. chapman , e. dalessandro , a. sollima , 2012 , mnras , * 426 * , 2889 s.e .
nersisyan , a.v .
shavrina , a.a .
yaremchuk , 1989 , astrofizika , * 30 * , 247 p. padoan , a. nordlund , 2004 , apj , * 617 * , 559 c.b .
pereira , j.v .
sales silva , c. chavero et al .
, 2011 , a&a , * 533 * , 51 a. pietrinferni , s. cassisi , m. salaris , f. castelli , 2006 , apj , * 642 * , 797 a. pietrinferni , s. cassisi , m. salaris , f. castelli , 2004 , apj , * 612 * , 168 b.j .
pritzl , k.a .
venn , m. irwin , 2005 , aj , * 130 * , 2140 m. samland , 1998 , apj , * 496 * , 155 n.a .
sakhibullin , v.v .
shimansky , astronomy reports , * 41 * , 378 n.a .
sakhibullin , 1983 , trudy kazan .
* 48 * , 9 r.p .
schiavon , s. m. faber , b.v .
castilho , j.a . rose , apj , 2002 , * 580 * , 850 m.e .
sharina , r. chandar , t.h .
puzia , p. goudfrooij , e. davoust , 2010 , mnras , * 405 * , 839 m.d .
shetrone , p. cote , w.l .
sargent , 2001 , apj * 548 * , 592 v.v .
shimansky , s.a .
pozdnyakova n.v .
borisov , i.f .
bikmaev et al , 2008 , astronomy letters , * 34 * , 423 v.v .
shimansky , n.v .
borisov , n.n .
shimanskaya , astronomy reports , 2003 , * 47 * , 763 v.v .
shimansky , l.i .
mashonkina , n.a .
sakhibullin , 2000 , astronomy reports , * 44 * , 530 v.v .
shimansky , 2002 , astronomy reports , * 46 * , 127 n.n .
shimanskaya , i.f .
bikmaev , v.v .
shimansky , 2011 , astrophysical bulletin , * 66 * , 255 r. smiljanic , g.f .
porto de mello , l. da silva , 2007 , a&a , * 468 * , 679 p. stumpff , 1980 , a&a suppl .
ser . , 1980 , *
41 * , 1 v.f .
suleimanov , 1996 , astronomy letter , * 22 * , 92 v.f .
suleimanov , astron .
trans . , 1992 , * 2 * , 197 a. unsold , 1955 , _ physik der sternatmospheren _ ,
( springer , berlin gottingen heidelberg , 1955 k.a .
venn , m. irwin , m.d .
shetrone , c.a .
tout , v. hill , e. tolstoy , 2004 , aj , * 128 * , 1177 c.r .
vidal , j. cooper , e.w .
smith , 1973 , a&a suppl . ser . , * 25 * , 37 g. worthey , s.m .
faber , j.j .
gonzalez , d. burnstein , 1994 , apj suppl .
ser . , * 94 * , 687 | ngc 2419 is the most distant massive globular cluster in the outer galactic halo .
it is unusual also due to the chemical peculiarities found in its red giant stars in recent years .
we study the stellar population of this unusual object using spectra obtained at the 1.93-m telescope of the haute - provence observatory . at variance with commonly used methods of high - resolution spectroscopy applicable only to bright stars , we employ spectroscopic information on the integrated light of the cluster .
we carry out population synthesis modeling of medium - resolution spectra using synthetic stellar atmosphere models based on a theoretical isochrone corresponding accurately to the observed color - magnitude diagram .
we study the influence of non - local thermodynamic equilibrium for some chemical elements on our results .
the derived age ( 12.6 gyr ) , [ fe / h]=-2.25 dex , helium content y=0.25 , and abundances of 12 other chemical elements are in a good qualitative agreement with published high - resolution spectroscopy estimates for red giant members in the cluster . on the other hand , the derived element abundance , [ @xmath0=0.13
$ ] dex ( the mean of @xmath1 $ ] , @xmath2 $ ] and @xmath3 $ ] ) , differs from the published one ( @xmath4 = 0.4 $ ] dex ) for selected red giants in the cluster and may be explained by a large dispersion in the @xmath5-element abundances recently discovered in ngc 2419 .
we suggest that studies of the _ integrated _ light in the cluster using high - resolution spectrographs in different wavelength regions will help to understand the nature of these chemical anomalies .
+ @xmath6 kazan federal state university , kremlevskaya street 18 , kazan , 420008 , russia + @xmath7 irap , universit de toulouse , cnrs , 14 avenue edouard belin , 31400 toulouse , france + |
high - energy heavy ion collisions at the relativistic heavy ion collider ( rhic ) and the large hadron collider ( lhc ) create matter that is well described as an equilibrated system with initial temperatures in excess of 340420 mev @xcite . in this regime ,
the matter is understood to be a quark - gluon plasma ( qgp ) with bound hadronic states no longer in existence as the temperatures far exceed the transition temperature of approximately 155 mev calculated by lattice quantum chromodynamics ( qcd ) @xcite .
this qgp follows hydrodynamical flow behavior with extremely small dissipation , characterized by the shear viscosity to entropy density ratio @xmath7 and is thus termed a near - perfect fluid @xcite .
charm and bottom quarks ( @xmath8 gev/@xmath9 and @xmath10 gev/@xmath9 ) are too heavy to be significantly produced via the interaction of thermal particles in the qgp .
thus the dominant production mechanism is via hard interactions between partons in the incoming nuclei , i.e. interactions that involve large momentum transfer , @xmath11 .
once produced , these heavy quarks are not destroyed by the strong interaction and thus propagate through the qgp and eventually emerge in heavy flavor hadrons , for example @xmath12and @xmath13mesons .
early measurement of heavy flavor electrons from the phenix collaboration in collisions at rhic indicated that although the total heavy flavor production scales with the number of binary collisions within uncertainties @xcite , the momentum distribution of these heavy quarks is significantly modified when compared with that in collisions @xcite .
these results indicate a large suppression for high-@xmath14 electrons and a substantial elliptic flow for @xmath15@xmath16 electrons from heavy quark decays . here , and throughout the paper , we use `` electrons '' to refer to both electrons and positrons .
the suppression of the charm quark has since been confirmed through the direct reconstruction of @xmath12mesons by the star collaboration @xcite . in collisions at the lhc at
= 2.76 tev , similar momentum distribution modifications of heavy flavor electrons and @xmath12mesons have been measured @xcite .
recently , the cms experiment has reported first measurements of @xmath17 @xcite and b - jets @xcite in collisions .
in contrast to this suppression pattern found in collisions , and peripheral collisions at exhibit an enhancement at intermediate electron in the heavy flavor electron spectrum @xcite that must be understood in terms of a mechanism that enhances the spectrum , e.g. the cronin effect @xcite .
that mechanism potentially moderates the large suppression observed in collisions at .
it is notable that in central collisions at = 62 gev an enhancement is also observed at intermediate @xcite .
the possibility that charm quarks follow the qgp flow was postulated early on @xcite , and more detailed langevin - type calculations with drag and diffusion of these heavy quarks yield a reasonable description of the electron data @xcite .
many of these theory calculations incorporate radiative and collisional energy loss of the heavy quarks in the qgp that are particularly important at high- , where qgp flow effects are expected to be sub - dominant .
the large suppression of heavy flavor electrons extending up to @xmath18 has been a particular challenge to understand theoretically , in part due to an expected suppression of radiation in the direction of the heavy quarks propagation often referred to as the `` dead - cone '' effect @xcite .
this observation of the high- suppression @xcite is all the more striking because perturbative qcd ( pqcd ) calculations indicate a substantial contribution from bottom quark decays for @xmath14 @xcite .
first measurements in collisions at 200 gev via electron - hadron correlations confirm this expected bottom contribution to the electrons that increases as a function of @xcite . to date , there are no direct measurements at rhic of the contribution of bottom quarks in collisions . for the specific purpose of separating the contributions of charm and bottom quarks at midrapidity , the phenix collaboration has added micro - vertexing capabilities in the form of a silicon vertex tracker ( vtx ) . the different lifetimes and kinematics for charm and bottom hadrons decaying to electrons
enables separation of their contributions with measurements of displaced tracks ( i.e. the decay electron not pointing back to the collision vertex ) . in this paper
, we report on first results of separated charm and bottom yields via single electrons in minimum bias ( mb ) collisions at .
as detailed in ref .
@xcite , the phenix detector was originally designed with precision charged particle reconstruction combined with excellent electron identification . in 2011 ,
the vtx was installed thus enabling micro - vertexing capabilities .
the dataset utilized in this analysis comprises collisions at .
a set of global event - characterization detectors are utilized to select events and eliminate background contributions .
two beam - beam counters ( bbc ) covering pseudorapidity @xmath19 and full azimuth are located at @xmath20 1.44 meters along the beam axis and relative to the nominal beam - beam collision point .
each of the bbcs comprises 64 erenkov counters . based on the coincidence of the bbcs ,
collisions are selected via an online mb trigger , which requires at least two counters on each side of the bbc to fire .
the mb sample covers @xmath21% of the total inelastic cross section as determined by comparison with monte carlo glauber models @xcite .
the bbc detectors also enable a selection on the @xmath22-vertex position of the collision as determined by the time - of - flight difference between hits in the two sets of bbc counters .
the @xmath22-vertex resolution of the bbc is approximately @xmath23 cm in central collisions .
a selection within approximately @xmath24 cm of the nominal detector center was implemented and @xmath25 85% of all collisions within that selection were recorded by the phenix high - bandwidth data acquisition system .
electrons ( @xmath26 and @xmath27 ) are reconstructed using two central spectrometer arms as shown in fig .
[ fig : phenix](a ) , each of which covers the pseudorapidity range @xmath28 and with azimuthal angle @xmath29 .
the detector configuration of the central arms is the same as in previous phenix collaboration heavy flavor electron publications @xcite .
charged particle tracks are reconstructed outside of an axial magnetic field using layers of drift chamber ( dc ) and multi - wire proportional pad chambers ( pc ) .
the momentum resolution is @xmath30 0.7% @xmath31 0.9% @xmath2 ( ) . for central arm charged particle reconstructions the trajectory is only measured for radial positions @xmath32 meters , and the momentum vector is calculated by assuming the track originates at the collision point determined by the bbc detectors and assuming 0 radial distance .
electron identification is performed by hits in a ring imaging erenkov detector ( rich ) and a confirming energy deposit in an electromagnetic calorimeter ( emcal ) . the rich uses co@xmath33 gas at atmospheric pressure as a erenkov radiator . electrons and pions begin to radiate in the rich at @xmath34 20 mev/@xmath4 and @xmath34 4.9 , respectively .
the emcal is composed of four sectors in each arm .
the bottom two sectors of the east arm are lead - glass and the other six are lead - scintillator .
the energy resolution of the emcal is @xmath35 4.5% @xmath31 8.3@xmath36 and @xmath35 4.3% @xmath31 7.7@xmath36 for lead - scintillator and lead - glass , respectively .
( color online ) ( a ) a schematic view of the phenix detector configuration for the 2011 run .
( b ) a schematic view of the vtx detector with the individual ladders shown.,title="fig : " ] ( color online ) ( a ) a schematic view of the phenix detector configuration for the 2011 run .
( b ) a schematic view of the vtx detector with the individual ladders shown.,title="fig : " ] in 2011 , the central detector was upgraded with the vtx detector as shown in fig . [ fig : phenix ] . in addition , a new beryllium beam pipe with 2.16 cm inner diameter and 760 @xmath37 m nominal thickness was installed to reduce multiple - scattering before the vtx detector . the vtx detector @xcite consists of four radial layers of silicon detectors as shown in fig .
[ fig : phenix](b ) .
the detector is separated into two arms , each with nominal acceptance @xmath38 centered on the acceptance of the outer phenix central arm spectrometers .
the detector covers pseudorapidity @xmath39 1.2 for collisions taking place at @xmath40 .
the vtx can precisely measure the vertex position of a collision within @xmath41 cm range of the center of the vtx .
the two inner layers , referred to as b0 and b1 , of the vtx detector comprise silicon pixel detectors , as detailed in ref .
@xcite . b0
( b1 ) comprises 10 ( 20 ) ladders with a central radial position of 2.6 ( 5.1 ) cm .
the silicon pixel technology is based on the sensor - readout chip @xcite , which was developed at cern .
each ladder is electrically divided into two independent half - ladders .
each ladder comprises four sensor modules mounted on a mechanical support made from carbon - fiber composite .
each sensor module comprises a silicon pixel sensor with a pixel size of 50 @xmath37m(@xmath42 ) @xmath43 425 @xmath37m(@xmath22 ) bump - bonded with four pixel readout chips .
one pixel readout chip reads 256 @xmath44 32 @xmath45= 8192 pixels and covers approximately 1.3 cm @xmath46 1.4 cm @xmath47 of the active area of the sensor .
the position resolution is @xmath48 = 14.4 @xmath37 m in the azimuthal direction .
the two outer layers of the vtx detector , referred to as b2 and b3 , are constructed using silicon stripixel sensors , as detailed in ref .
the b2 ( b3 ) layer comprises 16 ( 24 ) silicon stripixel ladders at a central radial distance of 11.8 ( 16.7 ) cm .
the stripixel sensor is a novel silicon sensor , and is a single - sided , n - type , dc - coupled , two - dimensional ( 2-d ) sensitive detector @xcite .
one sensor has an active area of approximately 30 mm @xmath43 60 mm , which is divided into two independent sectors of 30 mm @xmath43 30 mm .
each sector is divided into 384 @xmath43 30 pixels .
each pixel has an effective size of 80 @xmath37 m ( @xmath42 ) @xmath43 1000 @xmath37 m ( @xmath22 ) , leading to a position resolution of @xmath48=23 @xmath37 m .
a pixel comprises two implants ( a and b ) interleaved such that each of the implants registers half of the charge deposited by ionizing particles .
there are 30 a implants along the beam direction , connected to form a 30 mm long x - strip , and 30 b implants are connected with a stereo angle of 80 mrad to form a u - strip .
x - strip and u - strip are visualized in @xcite . when a charged particle hits a pixel , both the x- and the u - strip sharing the pixel register a hit
. thus the hit pixel is determined as the intersection of the two strips .
the stripixel sensor is read out with the svx4 chip developed by a fnal - lbnl collaboration @xcite .
the total number of channels in the vtx pixel and stripixel layers is 3.9 million pixels and 0.34 million strips .
the compositions of the pixel and strip are illustrated in @xcite .
the main characteristics of the vtx detector are summarized in table [ tab : vtx ] . [ cols="<,^,^,^,^,^,^,^,^,^,^,^,^ " , ] when performing the unfolding procedure , only the statistical uncertainties on the electron and spectra are included . in this section
we describe how we consider the systematic uncertainties on both the measured data and the unfolding procedure .
we take the following uncertainties into account as uncorrelated uncertainties : 1 .
systematic uncertainty in the heavy flavor electron invariant yield 2 .
uncertainty in the high - multiplicity background 3 .
uncertainty in the fraction of nonphotonic electrons ( @xmath49 ) 4 .
uncertainty in @xmath50 normalization 5 .
regularization hyperparameter @xmath51 6 .
uncertainty in the form of @xmath52 the uncertainty in @xmath49 ( see sec .
[ sec : fnp ] ) , and @xmath50 are propagated to the unfolded hadron yields by varying each independently by @xmath53 , and performing the unfolding procedure with the modified background template .
the difference between the resulting hadron yields and the central values is taken as the systematic uncertainty .
the same procedure is used to determine the uncertainty in the result due to the regularization parameter , which is varied by @xmath54 based on where the summed likelihood from both the data and regularization drops by 1 from the maximum value .
the uncertainty in the high - multiplicity background includes two components .
the first is the uncertainty on the normalization of the high - multiplicity background distribution , as determined in sec .
[ sec : dca_sideband ] and shown in fig .
[ fig : dca0 ] .
this is propagated to the unfolded hadron yields by varying the normalization by @xmath53 and performing the unfolding procedure with the modified background template , as with the @xmath49 and @xmath50 uncertainties .
the second component addresses the small excess in the embedded primary electron distribution observed in fig .
[ fig : embedsideband ] and not accounted for by using the distribution for large .
we parametrize the excess , which is more than two orders of magnitude below the peak , and apply it to the background components , re - performing the unfolding procedure to find its effect on the hadron yield .
both effects combined are small relative to the dominant uncertainties .
incorporating the correlated systematic uncertainty on the heavy flavor electron invariant yield is more difficult .
ideally one would include a full covariance matrix encoding the correlations into the unfolding procedure . in practice ,
the methodology employed in @xcite does not provide a convenient description of the @xmath55 correlations needed to shape the covariance matrix .
instead we take a conservative approach by considering the cases which we believe represent the maximum correlations .
we modify the heavy flavor electron invariant yield by either tilting or kinking the spectrum about a given point .
tilting simply pivots the spectra about the given point so that , for instance , the first point goes up by a fraction of the systematic uncertainty while the last point goes down by the same fraction of its systematic uncertainty , with a linear interpolation in between .
kinking simply folds the spectra about the given point so that that the spectrum is deformed in the form of a * v*. we implement the following modifications and re - perform the unfolding procedure : 1 .
tilt the spectra about @xmath56 by @xmath53 of the systematic uncertainty .
2 . tilt the spectra about @xmath57 by @xmath53 of the systematic uncertainty .
3 . kink the spectra about @xmath56 by @xmath53 of the systematic uncertainty .
4 . kink the spectra about @xmath57 by @xmath53 of the systematic uncertainty .
the points about which the spectra were modified were motivated by the points in at which analysis methods and details changed , as discussed in @xcite .
we then take the rms of the resulting deviations on the hadron yield from the central value as the propagated systematic uncertainty due to the systematic uncertainty on the heavy flavor electron invariant yield .
the effect of our choice of @xmath52 on the charm and bottom hadron yields is taken into account by varying @xmath52 , as discussed in section [ sec : prior ] .
the differences between each case and the central value are added in quadrature to account for the bias introduced by @xmath52 .
the uncertainties on the unfolded hadron yields due to the six components described above and the uncertainty determined from the posterior probability distributions are added in quadrature to give the uncertainty shown in fig .
[ fig : hadron - pt ] .
( color online ) the relative contributions from the different components to the uncertainty on the fraction of electrons from bottom hadron decays as a function of .
the shaded red band in each panel is the total uncertainty . ] due to the correlations between charm and bottom yields , the relative contributions from the different uncertainties depend on the variable being plotted . to give some intuition for this ,
we have plotted the relative contributions from the different uncertainties to the fraction of electrons from bottom hadron decays as a function of ( discussed in sec .
[ sec : bfraction ] ) in fig .
[ fig : bfrac - sys ] .
one can see that the dominant uncertainties come from the statistical uncertainty on the and heavy flavor electron invariant yield , the systematic uncertainty on the heavy flavor electron invariant yield , and @xmath49 .
we remind the reader that for @xmath58 we no longer have information to directly constrain the unfolding , and all information comes dominantly from the heavy flavor electron invariant yield , leading to the growth in the uncertainty band in this region .
the final result of the unfolding procedure applied simultaneously to the heavy flavor electron invariant yield vs ( shown in fig .
[ fig : ept - refold ] ) and the five electron distributions ( shown in fig .
[ fig : dca - refold ] ) is the invariant yield of charm and bottom hadrons , integrated over all rapidity , as a function of . as a reminder , the hadron yields are integrated over all rapidity by assuming the rapidity distribution within is accurate and that it is unmodified in , as detailed in sec . [
sec : decay_matrix ] .
the unfolded results for mb ( 0%96% ) collisions at = 200 gev are shown in fig .
[ fig : hadron - pt ] .
the central point represents the most likely value and the shaded band represents the @xmath59 limits on the combination of the uncertainty in the unfolding procedure and the systematic uncertainties on the data , as described in sec .
[ sec : unfolding_systematics ] .
the uncertainty band represents point - to - point correlated uncertainties , typically termed type b in phenix publications .
there are no point - to - point uncorrelated ( type a ) , or global scale uncertainties ( type c ) , from this procedure . the uncertainties on the hadron invariant yields shown in fig .
[ fig : hadron - pt ] grow rapidly for charm and bottom hadrons with @xmath60 .
this is due to the lack of information for @xmath61 . above @xmath61 ,
the unfolding is constrained by the heavy flavor electron invariant yield only .
this provides an important constraint on the shape of the hadron distributions , but the distributions provide the dominant source of discriminating power between the charm and bottom .
however , due to the decay kinematics , even high hadrons contribute electrons in the range @xmath62<5.0 $ ] .
we find that charm(bottom ) hadrons in the range @xmath63<20 $ ] contribute 18.2%(0.3% ) of the total electron yield in the region @xmath64<5.0 $ ] .
this explains the larger uncertainties in the bottom hadron yield compared to the charm hadron yield at high .
the yield of @xmath65 mesons over @xmath66 as a function of has been previously published in collisions at = 200 gev by star @xcite . in order to compare our unfolded charm hadron results over all rapidity to the star measurement
, we use to calculate the fraction of @xmath65 mesons within @xmath66 compared to charm hadrons over all rapidity . since the measurement by star is over a narrower centrality region ( 0%80% vs 0%-96% ) , we scale the star result by the ratio of the values .
this comparison is shown in fig .
[ fig : stard0 ] . for added clarity , we have fit the star measurement with a levy function modified by a blast wave calculation given by @xmath67 where @xmath68 is a standard gaussian function , and @xmath69 are the parameters of the fit .
the ratio of the data to the fit is shown in the bottom panel of fig .
[ fig : stard0 ] .
we find that , within uncertainties , the unfolded @xmath65 yield agrees with that measured by star over the complementary range .
the unfolded yield hints at a different trend than the star data for @xmath58 . however , we note that the @xmath70 of charm(bottom ) hadrons which contribute electrons in the range @xmath71<5.0 $ ] is 7.2(6.4 ) .
this means that the yields of charm and bottom hadrons have minimal constraint from the measurements in the high - regions , which is represented by an increase in the uncertainties .
mesons as a function of for @xmath66 inferred from the unfolded yield of charm hadrons integrated over all rapidity compared to measurements from star @xcite .
see the text for details on the calculation of the @xmath65 yield inferred from the unfolded result . to match the centrality intervals ,
the star result has been scaled by the ratio of values .
the bottom panel shows the ratio of the data to a fit of the star @xmath65 yield . ]
the fraction of heavy flavor electrons from bottom hadrons ( @xmath72 ) is computed by re - folding the charm and bottom hadron yields shown in fig .
[ fig : hadron - pt ] to get the invariant yield of electrons from charm and bottom decays at midrapidity ( @xmath73 ) . here
the electrons from bottom hadron decays include the cascade decay @xmath74 . the resulting bottom electron fraction is shown as a function of in fig .
[ fig : bfrac ] .
the central values integrated over the range of each distribution are also quoted in fig .
[ fig : dca - refold ] . as in the hadron yields , the band represents the @xmath59 limits of
the point - to - point correlated ( type b ) uncertainties . also shown in fig .
[ fig : bfrac ] is the bottom electron fraction predictions from @xcite for collisions at = 200 gev .
we find a bottom electron fraction which is encompassed by the calculation uncertainties .
the shape of the resulting bottom electron fraction shows a steeper rise in the region @xmath75<4.0 $ ] with a possible peak in the distribution compared to the central calculation .
the fraction of electrons from bottom decays has been previously measured in collisions at = 200 gev by both phenix @xcite and star @xcite .
these measurements are made through electron - hadron or electron-@xmath12 meson correlations .
these are very different analyses than the one presented here , and have their own model dependencies . in fig .
[ fig : bfrac - pp ] we compare the bottom electron fraction between our unfolded result and the electron - hadron correlation measurements in . for @xmath76
we find agreement between and within the large uncertainties on both measurements .
this implies that electrons from bottom hadron decays are similarly suppressed to those from charm . for reference , included in fig .
[ fig : bfrac - pp ] is the central calculation which , within the large uncertainties , is consistent with the measurements . with the additional constraints on the bottom electron fraction in from the correlation measurements and the measured nuclear modification of heavy flavor electrons , we can calculate the nuclear modification of electrons from charm and bottom hadron decays separately .
the nuclear modifications , @xmath77 and @xmath78 , for charm and bottom hadron decays respectively are calculated using @xmath79 where @xmath80 and @xmath81 are the fractions of heavy flavor electrons from bottom hadron decays in and respectively and @xmath82 is the nuclear modification of heavy flavor electrons ( combined charm and bottom ) . rather than combining all measurements for the bottom electron fraction in , which introduces a further extraction uncertainty , we have chosen to calculate @xmath77 and @xmath78 using only the six star electron - hadron @xmath81 values . when performing the calculation we determine the full probability distributions assuming gaussian uncertainties on @xmath80 , @xmath81 and @xmath82 . as when determining the charm and bottom hadron yields , we take the median of the distribution as the central value , and the 16% and 84% of the distribution as the lower and upper @xmath59 uncertainties .
the resulting values are shown in fig .
[ fig : hfraa](a ) .
we find that the electrons from bottom hadron decays are less suppressed than electrons from charm hadron decays for @xmath83 .
to further clarify this statement , we calculate the ratio of @xmath84 , shown in fig .
[ fig : hfraa](b ) . in this ratio , the uncertainty on @xmath82 cancels . here
again we calculate the full probability distributions and use the same procedure as above to determine the central values and uncertainties .
we find that the probability distributions for @xmath84 are highly nongaussian , which leads to the large asymmetric uncertainty band shown in fig .
[ fig : hfraa](b ) .
it is clear from the ratio that @xmath85 is less suppressed than @xmath86 at the @xmath59 level up to @xmath87 .
, @xmath85 and combined heavy flavor @xcite as a function of .
the @xmath86 and @xmath85 is calculated using eq .
[ eq : craa]-[eq : braa ] where @xmath80 uses the unfolded result determined in this work and @xmath81 determined from star @xmath88 correlations @xcite .
( b ) the ratio @xmath89 as a function of . ]
there are a number of theoretical calculations in the literature for the interaction of charm and bottom quarks with the qgp .
many of these models have predictions for the nuclear modification factor @xmath5 for electrons from charm decays and , separately , @xmath5 for electrons from bottom decays . for consistency
, we have assumed the @xcite yields for electrons from charm ( bottom ) decays calculated for at = 200 gev and then scaled them by the heavy - ion model results for the @xmath5 of electrons from charm ( bottom ) .
+ + figure [ fig : bfrac - models](a ) compares the bottom electron fraction from one class of calculations modeling only energy loss of these heavy quarks in medium . in an early pqcd calculation by djordjevic , gyulassy , vogt , and wicks @xcite ,
the authors apply the dglv theory of radiative energy loss .
they find that even for extreme opacities with gluon rapidity densities up to 3500 , the bottom quark decay electrons dominate at high - and that limits the single electron to the range @xmath90@xmath91 for @xmath58 .
although this result is known to be higher than the phenix measured heavy flavor electron @xmath5 @xcite , we show the @xmath92 predictions for gluon rapidity densities of 1000 and 3500 in fig .
[ fig : bfrac - models](a ) .
however , we do note that the calculations are for 0%10% central collisions compared to the mb data , although the calculations span a factor of 3.5 range in the gluon density .
we find that the calculations for both gluon rapidity densities are in good agreement with our results for @xmath93 , but are slightly above and outside the uncertainty band on the unfolded result at higher .
more recent calculations in the same framework , but with the inclusion of collisional energy loss @xcite , result in a heavy flavor electron high-@xmath5 closer to 0.3 and in reasonable agreement with previous phenix published results @xcite . this updated prediction for the bottom electron fraction , also shown in fig .
[ fig : bfrac - models ] , gives a similar value to their previous result , but is only published for @xmath58 .
figure [ fig : bfrac - models](b ) compares the bottom electron fraction from a calculation using a t - matrix approach by van hees , mannarelli , greco , and rapp @xcite .
the authors provided us with different results for 0%10% central collisions depending on the coupling of the heavy - quark to the medium .
the coupling is encapsulated in the diffusion parameter @xmath12 , where smaller values yield a stronger coupling . shown in fig .
[ fig : bfrac - models](b ) are three results corresponding to three values of the parameter @xmath94 .
the largest @xmath12 value , corresponding to the weakest coupling , yields almost no deviation from the reference result , and the successively stronger coupling pushes the bottom fraction contribution higher and higher .
we find that the calculations with @xmath95 are in good agreement with our result for @xmath93 , but begin to diverge where the calculation stops at 5 .
figure [ fig : bfrac - models](c ) compares the bottom electron fraction from another class of calculations which employ a combination of langevin , or transport type modeling of heavy - quarks , in the bulk qgp with energy loss mechanisms that dominate at higher . in ref .
@xcite , alberico _ et al . _
employ a langevin calculation where a good match to the phenix heavy flavor electrons is found .
it is notable that this calculation has a very strong suppression of charm decay electrons such that bottom contributions dominate even at modest @xmath96 2 .
the calculations are consistent with the data for @xmath93 and over - predict the bottom contribution for higher values .
figure [ fig : bfrac - models](c ) also compares the bottom electron fraction from another variant of the langevin calculation by cao _
et al . _ , as detailed in ref .
@xcite . for this calculation
, we show two results corresponding to two different input values @xmath97 and 6 . for the lower parameter , again stronger heavy - quark to medium coupling , there is a sharp rise in the bottom contribution which then flattens out .
this feature is due to the increased collisional energy loss , which has a larger effect on the charm quarks , coupled with the strong radial flow effects enabling the heavier bottom quarks to dominate even at @xmath98 .
these calculations use an impact parameter of @xmath99 fm , which should roughly correspond to mb collisions .
we find that the calculation using the larger value of @xmath100 is in reasonable agreement with the data across the calculated range .
lastly , fig .
[ fig : bfrac - models](d ) shows a more recent calculation by he _
_ employing a t - matrix approach similar to that shown in fig .
[ fig : bfrac - models](b ) , but with a number of updates as described in ref . @xcite . in this case
the authors provided a calculation of the bottom electron fraction in both and at = 200 gev , and we therefore do not calculate the bottom fraction using as a baseline . the calculation is performed for the 20%40% centrality bin , which the authors find well represents mb .
we find that the calculation under - predicts the bottom fraction for @xmath101 , although it is worth noting that the calculation in is also below the curve across the full range . above @xmath102
the calculation is in agreement with the measurement .
it is also worth noting that , of the models presented here , this is the only one that shows in a slight decrease in the bottom fraction at high .
there are numerous other calculations in the literature @xcite that require mapping charm and bottom hadrons to electrons at midrapidity to make direct data comparisons .
we look forward to soon being able to test these calculations with analysis of new phenix data sets .
this article has detailed the measurements of electrons as a function of and from data taken at = 200 gev in 2011 with the enhanced vertexing capabilities provided by the vtx detector . in conjunction with previous phenix results for the heavy flavor electron invariant yield as a function of @xcite ,
we perform an unfolding procedure to infer the parent charm and bottom hadron yields as a function of .
we find that this procedure yields consistent agreement between the heavy flavor electron invariant yield and the newly measured electron distributions .
we find that the extracted @xmath65 yield vs is in good agreement with that measured by star @xcite over the complimentary region . without a proper baseline extracted from a similar analysis
it is difficult to make any quantitative statements about the charm or bottom hadron modification .
we compare the extracted bottom electron fraction to measurements in collisions and find agreement between and for @xmath76 within the large uncertainties on both measurements . the agreement between and coupled with the measured heavy flavor electron @xmath5 strongly implies that electrons from charm and bottom hadron decays are suppressed . using these components we calculate the nuclear modification for electrons from charm and bottom hadron decays and find that electrons from bottom hadron decays are less suppressed than those from charm hadron decays in the range .
we further compare the bottom electron fraction to a variety of model calculations employing variously energy loss , langevin transport , and t - matrix approaches .
we find that there are a number of models which are in reasonable agreement with the extracted bottom electron fraction within the relatively large uncertainties .
we note that a significantly larger data set of collisions at = 200 gev was collected in 2014 with an improved performance of the vtx detector .
the 2014 data coupled with the data taken in 2015 should yield both an important baseline measurement of the bottom electron fraction and a more precise measurement in .
we thank the staff of the collider - accelerator and physics departments at brookhaven national laboratory and the staff of the other phenix participating institutions for their vital contributions .
we acknowledge support from the office of nuclear physics in the office of science of the department of energy , the national science foundation , abilene christian university research council , research foundation of suny , and dean of the college of arts and sciences , vanderbilt university ( u.s.a ) , ministry of education , culture , sports , science , and technology and the japan society for the promotion of science ( japan ) , conselho nacional de desenvolvimento cientfico e tecnolgico and fundao de amparo pesquisa do estado de so paulo ( brazil ) , natural science foundation of china ( p. r. china ) , croatian science foundation and ministry of science , education , and sports ( croatia ) , ministry of education , youth and sports ( czech republic ) , centre national de la recherche scientifique , commissariat lnergie atomique , and institut national de physique nuclaire et de physique des particules ( france ) , bundesministerium fr bildung und forschung , deutscher akademischer austausch dienst , and alexander von humboldt stiftung ( germany ) , national science fund , otka , kroly rbert university college , and the ch .
simonyi fund ( hungary ) , department of atomic energy and department of science and technology ( india ) , israel science foundation ( israel ) , basic science research program through nrf of the ministry of education ( korea ) , physics department , lahore university of management sciences ( pakistan ) , ministry of education and science , russian academy of sciences , federal agency of atomic energy ( russia ) , vr and wallenberg foundation ( sweden ) , the u.s .
civilian research and development foundation for the independent states of the former soviet union , the hungarian american enterprise scholarship fund , and the us - israel binational science foundation .
this appendix details the calculation of the normalizations for the background components : using the bootstrap method described in sec .
[ sec : norm ] .
we first determine the fraction of nonphotonic electrons , @xmath49 .
we then calculate the normalization of dalitz and conversion components followed by the normalization of @xmath50 and quarkonia components .
we first determine @xmath49 , the fraction of nonphotonic electrons to inclusive electrons after the application of all analysis cuts , including the conversion veto cut .
note that nonphotonic electrons include contributions from heavy flavor semi - leptonic decays , quarkonia decays , and kaon decays .
photonic electrons are from @xmath103 and @xmath104 dalitz decays and photon conversions .
@xmath49 in the 2011 data can be determined using the published 2004 result @xcite as follows .
let @xmath105 be the yield of nonphotonic electrons and @xmath106 the yield of electrons from dalitz decays .
note that both @xmath105 and @xmath106 are independent of the year of data taking . in the phenix 2004 data run
, the ratio of the nonphotonic electron yield to the photonic electron yield ( @xmath107 ) was measured .
the relation of @xmath105 and @xmath106 is as follows : @xmath108 where @xmath109 represents the ratio of conversion electron yield to dalitz electron yield in the 2004 phenix detector .
it is calculated as @xmath110 here @xmath111 is the ratio of conversion electrons to electrons from dalitz decays in the 2004 phenix detector calculated by a full simulation .
the factors are the fractional contributions of @xmath115 , @xmath104 , and direct photon contribution to the total dalitz decays , respectively ( @xmath104 ) is caused by internal conversion of one of two decay photons in @xmath116 . ] .
we only consider the contributions of @xmath115 , @xmath104 , and @xmath117 ( direct photon ) since the sum of other contributions is small ( 5% or less ) .
thus they are normalized such that figure [ fig : dalitz_ratio ] shows @xmath119 for @xmath115 , @xmath104 , and direct photon as a function of transverse momentum of the electrons for mb collisions at 200 gev .
the ratios are calculated from the invariant yield of @xmath115@xcite , @xmath104@xcite , and direct photons@xcite . in the 2011 data set the observed electron yields from conversion and dalitz decays
are modified by the electron survival probability after the conversion veto cut is applied .
the yield of photonic electrons which pass the conversion veto @xmath120 is @xmath121 where @xmath122 is the survival probability of conversion electrons , @xmath123 are survival probabilities of dalitz decay electrons from @xmath115 , @xmath104 , and direct photons , respectively , as shown in fig .
[ fig : survival_rate ] .
@xmath124 ( @xmath125)is the ratio of conversion electrons to dalitz electrons for particle @xmath126 in the 2011 phenix detector after the addition of the vtx and the replacement of the beam pipe .
it is determined to be @xmath127 from full simulations .
the resulting @xmath49 as a function of and the calculated systematic uncertainties due to the uncertainties on the input yields is shown in fig .
[ fig : fnp_method1 ] . with @xmath49 in hand
, we obtain the number of photonic electrons , @xmath129 , and the number of nonphotonic electrons , @xmath130 as where @xmath136 and @xmath137 are the number of electrons from conversions and dalitz from particle @xmath126 after the conversion veto cut , respectively ; @xmath138 is the common reduction factor of tracks due to random hits in the windows of the conversion veto cut ; and @xmath139 is the efficiency and acceptance without the conversion veto cut .
since the number of photonic electron is @xmath140 , the fraction of conversions and dalitz decays in the photonic electrons are where @xmath145 is the relative contribution of electrons from ( conversion + dalitz decay ) for particle @xmath126 after application of conversion veto cut .
figure [ fig : photonic_ratio ] shows @xmath146 @xmath147 as a function of .
the conversion contributions are nearly the same for @xmath148 and @xmath149 , and effectively cancel when calculating the ratio .
therefore , @xmath150 ( fig . [ fig : photonic_ratio ] ) is almost identical with @xmath119 ( fig .
[ fig : dalitz_ratio ] ) .
the ratio of electrons from kaons to all nonphotonic electrons before the application of the conversion veto cut , @xmath151 , is calculated from the ratio of the nonphotonic electron yield to the electron yield from kaons @xcite .
compared to ref .
@xcite , we find that @xmath152 of electrons from kaon decays are removed by and cuts as well as the method used to subtract random background , which contains some real electrons from kaon decays .
the ratio of electrons from @xmath153 decays to all nonphotonic electrons before the application of the conversion veto cut , @xmath154 , is taken from ref .
the survival rate for electrons from @xmath153 decays , @xmath155 , is taken to be unity , while the survival rate for @xmath50 decays , @xmath156 , is taken to be the same value as that for electrons from charm and bottom decays ( namely , @xmath157 ) .
see sec .
[ sec : photonic_electrons ] for details .
69ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx
[ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1016/j.nuclphysa.2005.03.086 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2005.03.085 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.104.132301 [ * * , ( ) ] link:\doibase 10.1142/s0218301310014613 [ * * , ( ) ] link:\doibase 10.1146/annurev - nucl-102212 - 170540 [ * * , ( ) ] link:\doibase 10.1103/physrevd.90.094503 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.111601 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2004.10.034 [ * * , ( ) ] in @noop _ _ , p. link:\doibase 10.1103/physrevlett.94.082301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.88.192303 [ * * , ( ) ] link:\doibase 10.1103/physrevc.84.044905 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.172301 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.142301 [ * * , ( ) ] link:\doibase 10.1103/physrevc.90.034904 [ * * , ( ) ] @noop
`` , '' @noop `` , '' @noop * * , ( ) , link:\doibase 10.1103/physrevc.90.034903 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.242301 [ * * , ( ) ] link:\doibase 10.1103/physrevd.19.764 [ * * , ( ) ] link:\doibase 10.1103/physrevc.91.044907 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 2693(03)00175 - 8 [ * * , ( ) ] link:\doibase 10.1103/physrevc.71.064904 [ * * , ( ) ] link:\doibase 10.1103/physrevc.88.044907 [ * * , ( ) ] @noop _ _ ( ) link:\doibase 10.1103/physrevlett.100.192301 [ * * , ( ) ] link:\doibase 10.1088/0954 - 3899/37/9/094019 [ * * , ( ) ] link:\doibase 10.1103/physrevc.90.024911 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 2693(01)01130 - 3 [ * * , ( ) ] link:\doibase 10.1103/physrevc.90.034910 [ * * , ( ) ] link:\doibase 10.1088/0954 - 3899/32/12/s41 [ * * , ( ) ] link:\doibase
10.1103/physrevlett.95.122001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.082002 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.202301 [ * * , ( ) ] link:\doibase 10.1016/s0168 - 9002(02)01950 - 2 [ * * , ( ) ] link:\doibase 10.1146/annurev.nucl.57.090506.123020 [ * * , ( ) ] link:\doibase 10.2172/15007383;%% [ `` , '' ] link:\doibase 10.1016/j.nuclphysa.2013.02.099 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2013.02.033 [ * * , ( ) ] @noop * * , ( ) , in link:\doibase 10.1016/s0168 - 9002(01)00590 - 3 [ _ _ ] , vol .
( ) p. link:\doibase 10.1016/j.nima.2003.11.189 [ * * , ( ) ] @noop * * , ( ) , link:\doibase 10.1016/s0168 - 9002(03)01786 - 8 [ * * , ( ) ] link:\doibase 10.1088/1674 - 1137/38/9/090001 [ * * , ( ) ] @noop link:\doibase 10.1103/physrevd.44.3501 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.232301 [ * * , ( ) ] link:\doibase 10.1103/physrevc.82.011902 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.152302 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.232301 [ * * , ( ) ] @noop * * , ( ) @noop `` , ''
@noop * * , ( ) link:\doibase 10.1086/670067 [ * * , ( ) ] @noop `` , '' link:\doibase 10.1016/j.nuclphysa.2006.06.031 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.152301 [ * * , ( ) ] @noop * * , ( ) , link:\doibase 10.1103/physrevc.74.024902 [ * * , ( ) ] link:\doibase 10.1016/j.physletb.2005.09.087 [ * * , ( ) ] link:\doibase 10.1140/epjc / s10052 - 008 - 0838 - 5 [ * * , ( ) ] link:\doibase 10.1140/epjc / s10052 - 011 - 1666 - 6 [ * * , ( ) ] link:\doibase 10.1088/0954 - 3899/40/8/085103 [ * * , ( ) ] link:\doibase 10.1016/j.physletb.2014.05.050 [ * * , ( ) ] link:\doibase 10.1088/0954 - 3899/34/8/s92 [ * * , ( ) ] link:\doibase 10.1103/physrevd.91.085019 [ * * , ( ) ] link:\doibase 10.1016/j.nuclphysa.2009.10.003 [ * * , ( ) ] | the phenix collaboration at the relativistic heavy ion collider has measured open heavy flavor production in minimum bias au@xmath0au collisions at @xmath1 gev via the yields of electrons from semileptonic decays of charm and bottom hadrons .
previous heavy flavor electron measurements indicated substantial modification in the momentum distribution of the parent heavy quarks due to the quark - gluon plasma created in these collisions . for the first time , using the phenix silicon vertex detector to measure precision displaced tracking , the relative contributions from charm and bottom hadrons to these electrons as a function of transverse momentum are measured in au@xmath0au collisions .
we compare the fraction of electrons from bottom hadrons to previously published results extracted from electron - hadron correlations in @xmath2@xmath0@xmath2 collisions at @xmath1 gev and find the fractions to be similar within the large uncertainties on both measurements for @xmath3 gev/@xmath4 .
we use the bottom electron fractions in au@xmath0au and @xmath2@xmath0@xmath2 along with the previously measured heavy flavor electron @xmath5 to calculate the @xmath5 for electrons from charm and bottom hadron decays separately .
we find that electrons from bottom hadron decays are less suppressed than those from charm for the region @xmath6 gev/@xmath4 . |
over last couple of decades , unique physical properties due to quantum confinement effects have been reported for a wide range of semiconductor nanocrystalline materials @xcite . since in these systems surface to volume ratio
is high compared to that of corresponding bulk materials , surface states play a crucial role in determining their physical properties .
the surface of nanocrystals is made of atoms that are not fully coordinated .
hence , they are highly active and these surface atoms act like defect states unless they are passivated by either organic ligands or higher bandgap semiconductor materials @xcite .
thus , the current direction in this field of research includes modification of given size - quantized semiconductor particles by means of surface chemistry . in the literature
, we find reports on quantum dot - quantum well ( qdqw ) and core - shell nanocrystalline structures due to surface modifications of the particles .
qdqw is a three layered structure consisting of a size quantized particle acting as a core [ marked a in fig .
1(a ) ] and a complete layer of another material on the surface [ marked b in fig .
1(a ) ] of this core , which is again covered by core material as an outermost shell .
this core - shell - core structure is marked as a - b - a in fig .
colloquially , these particular structures are also called nano - onions. recently , dorfs and eychmller have outlined the preparation and characterization of a series of qdqw systems containing two wells @xcite .
coating nanoparticles by another material yields core - shell nanocrystals [ see fig .
1(b ) ] @xcite .
= 4.5 in in core - shell strucure , like inorganic epitaxial shell growth , the organic ligands can not passivate both cationic and anionic surface sites of the core @xcite .
the particles passivated by inorganic shell structures are more robust than organically passivated nanocrystals and have greater tolerance to processing conditions necessary for incorporation into solid structures @xcite . for effective surface passivation ,
the core particles having certain band gap are capped with a higher band gap material . moreover ,
the conduction band energy of the capping ( shell ) material is usually higher than that of the core material with the valence band energy of the capping material being lower .
this energetic situation is called a type- i structure @xcite . due to presence of higher band gap capping material ,
the photogenerated excitons in the core remain localized in the same region and are forced to recombine while spatially confined in the core .
confinement of electrons in the nanocrystals gives rise to blue shift in optical absorption and pl spectra of the material .
as the non - radiative decay channels through surface states are not accessible for these electrons , core - shell structures thus formed show higher luminescence quantum yields @xcite lower fluorescence life time @xcite and many other benefits @xcite related to tuning of band gap in two materials . like electrons ,
phonons are also confined in nanocrystals . in bulk crystals ,
the phonon eigenstate is a plane wave and wavevector selection rule for the first order raman scattering requires @xmath0 .
in contrast , the spatial correlation function of the phonon becomes finite due to its confinement in the nanocrystal and hence the @xmath0 selection rule gets relaxed .
in general , the phonon - dispersion curves of bulk crystals show the frequency @xmath1 to be a decreasing function of wavevector @xmath2 .
hence , the first order raman line shifts and broadens towards the low frequency side for the nanocrystals .
this has been proposed and explained by phonon confinement model @xcite .
confinement of phonons for si , ge , bn , cds , cds@xmath3se@xmath4 and many other nanocrystals have been reported in the literature @xcite .
as mentioned before , surface states play an important role in deciding different physical properties of nanocrystals . for a plane wave propagating in the x - direction in a bulk crystal , the temporal and spatial variation of the wave
is described by the factor @xmath5 $ ] , where the wavevector @xmath6 ; @xmath7 is the dielectric constant of the crystal . in the frequency range between bulk longitudinal optical ( lo ) phonon frequency , @xmath8 , and transverse optical ( to ) mode frequency , @xmath9 , @xmath10
, @xmath11 is imaginary .
therefore , in this frequency range the wave decays exponentially in the medium , i.e. it can not propagate in bulk crystals and only surface modes exist @xcite .
because of enhanced surface to volume ratio these modes appear for nanocrystals and it provides relevant information on surface states .
recently , baranov et al have reported the surface phonon modes in zns shelled cdse particles @xcite .
characteristics of bare cdse nanocrysals are now well established .
cdse nanocrystals passivated with long organic chain ( like topo ) have room temperature photoluminescence ( pl ) ( quantum yield 10@xmath12 ) with very long fluorescence time @xcite . the inorganic shelling of cdse core by cds has also been explored and shown to be better for surface passivation .
previous hrtem and xps measurements have shown that shell growth in this system does not form an alloy @xcite .
the lattice mismatch of 3.9@xmath12 between cds and cdse is small enough to allow epitaxial growth while still preventing alloying .
moreover , the difference in band gap ( 2.42 ev for cds and 1.74 ev for cdse at room temperature ) is large enough for shell growth .
the room temperature pl yield for cdse - cds core - shell system has been reported to be upto 50@xmath12 @xcite .
the electronic structure in this material can be understood from molecular orbital model and particle in a box model , as discussed in ref .
the same has also been discussed from the analysis of photoelectron spectra from core - shell structure @xcite .
the high photostability of the system has been explained by confinement of electrons and holes in the core - shell region . in this article , we have discussed the optical properties of marcaptoacetic acid stabilized bare cdse particles and cds capped cdse particles .
the nanoparticles have been characterized by their absorption spectra as well as by high resolution transmission electron microscope ( hrtem ) images .
the optical transition is modeled using effective bond order model ( ebom ) .
average particle size and size distribution of the particles have been estimated from both hrtem images and optical absorption measurements ; which have been further supported by raman measurements .
we have quantitatively analyzed asymmetric raman line shapes including both confined optic modes and surface phonon modes .
the observed frequencies of the sp modes in these systems have been compared with the calculations based on the dielectric response function model . in section
ii of this article , we have briefly discussed the sample preparation procedure , which we have followed .
section iii deals with analysis of hrtem images of the samples .
section iv demonstrates the optical properties of the samples , studied by optical absorption and pl spectroscopy .
detail analysis of phonon spectra is reported in section v. finally , in section vi , we have summarized our results with a few concluding remarks .
for sample preparation , cadmium perchlorate , [ cd(clo@xmath13)@xmath14 and marcaptoacetic acid [ hs - ch@xmath15-cooh ] were purchased from aldrich , usa and loba chemie , india respectively .
anhydrous sodium sulfite [ na@xmath15so@xmath16 and sodium sulfide [ na@xmath15s ] were from merck ltd . ,
india and s. d. fine chem . ltd .
, india respectively . solid selenium powder[se ] and sodium hydroxide [ naoh ] were used as received .
1.0 gm of se powder was added in 200 ml ( 1.1 m ) hot na@xmath15so@xmath17 solution under stirring condition and then boiled for half an hour .
subsequently , the solution was cooled at room temperature and the aqueous layer was filtered using whatman filter paper .
this solution is used as the stock solution .
na@xmath15so@xmath17 + se @xmath18 na@xmath15seso@xmath17 following liu _
et al _ @xcite , 150 ml of cd(clo@xmath13)@xmath15 ( 2 x 10@xmath19 m ) and 150 ml of hs - ch@xmath15-cooh ( 2 x10@xmath19 m ) were mixed and stirred vigorously for about 5 min . here , the latter acts as a stabilizing agent .
the ph of the solution was adjusted to 9.0 by adding aqueous naoh ( 1.0 m ) solution and heated to 100@xmath20c under n@xmath15 atmosphere .
subsequently , 0.5 ml of na@xmath15seso@xmath17 was added drop wise to the solution and boiled again for about half an hour under stirring condition .
the solution turned orange in color indicating the formation of cdse particles .
for the preparation of coated cdse - cds particle , the above cdse solution was cooled to 50@xmath20c and 20 @xmath21l of cd(clo@xmath13)@xmath15 ( 0.1 m ) and 20 @xmath21l of na@xmath15s ( 0.1 m ) were added drop wise alternatively under stirring condition in n@xmath15 atmosphere . after 30 min of stirring the solution turned orange - red in color due to formation of cdse - cds core - shell particles
= 5.5 in samples for transmission electron microscopy were deposited onto 300 mesh copper tem grids coated with 50 nm carbon films .
samples a and b are suspended in water ; thus , directly added drop - wise on the grid .
the excess water was allowed to evaporate in air .
the grids were examined in jeol 2010 microscope with ultra - high resolution ( uhr ) microscope using a lab@xmath22 filament operated at 200 kv .
hrtem images of many nanocrystals for each sample were measured and analyzed .
tem has been used to determine the particle size and size distribution of the afore - mentioned nanocrsytals .
also , using the high resolution tem , it is possible to confirm the core - shell structure of nanocrystals @xcite .
low magnification tem micrographs are shown in fig .
2 ( a ) and ( c ) for sample a and b , respectively . fig .
2(b ) and ( d ) are the histogram plots , obtained by measuring sizes of many particles per sample .
size distribution for the nanoparticles are usually found to be log - normal : @xmath23 here @xmath24 is the average size and @xmath25 is related to the size distribution of the particles . by fitting the frequency plot using eqn .
1 ( solid lines in fig .
2 ) , we have estimated the average particle size ( @xmath24 ) and @xmath25 of the particles , which are listed in table i. the size distribution of the particles is estimated to be more spread out for bare particles ( sample a ) compared to the other sample . = 5.5 in transmission
electron microscopy measurements provide the convincing proof of the crystalline core - shell structure of the sample .
3 shows high resolution micrograph of selected nanocrystals along the line profiles .
the contrast in the image depends on the electron density in the object forming the image .
hence , cds is expected to show less contrast than cdse since the former has fewer electrons per unit cell .
the line profile of the image contrast for sample a and b corresponding the the images are presented in fig . 3 ( b ) and ( d ) . for sample
a , we observe a smooth drop in the contrast near the edge of the nanocrystal . in comparison ,
the line profile of core - shell nanocrystals ( sample b ) , exhibit a trend of a stepwise - drop .
the contrast can also be explained by the change in thickness of nanocrystals , but it is not very probable that it would occur in such a step - like way . from the hrtem images the average shell thickness of the particles in sample b
has been estimated to be 1.2 nm . from the micrograph [ fig .
3 a ] , we have determined the lattice spacing in sample a to be 3.72 , which corresponds to ( 100 ) plane of cdse in hexagonal phase @xcite .
.comaparison of the average size and size distribution of the particles in sample a and b as obtained from blue shift in tem images , optical absorption bands and confinement of phonons [ cols="^,^,^,^,^,^,^",options="header " , ]
our goal in this article was to show electron and phonon confinement in the cdse - cds core - shell nanostructure .
to this end , we have first confirmed the existence of the core - shell structure in our samples , prepared by soft chemical route , using hrtem images .
subsequently , to analyze the electronic transitions we focus on optical absorption measurements , supported by theoretical considerations of the effective bond order model .
furthermore , the first order raman line shape of the cdse - cds nanocrystals is quantitatively explained by taking into account both confined phonon modes and sp modes .
the frequencies of sp modes are shown to match well with their calculated values as obtained from the dielectric response function model .
we also demonstrate that due to passivated surface states cds capped samples exhibit clear sp modes separated from lo modes in core - shell type nanocrystallites . in summery , we have presented a technique to confirm the core - shell structure in this type of semiconductor nanostructure . while analysing the optical absorption of the cdse - cds core - shell structure using effective bond order model , we have assumed the expected electronic transitions for cdse and cds particles .
the phonon confinement in the shell structure is explained by taking the shell as a thin layer on cdse core .
the above assumptions may or may not play a crucial role in analyzing the data
. however , more accurate theoretical models are necessary to understand core - shell strucure better .
ar thanks dst and brns in india , for financial support .
66 nanoparticles : from theory and application ed .
gunter schmid wiley - vch verlag gmbh @xmath26 co. kgaa , weinheim , 2004 and the references therein .
l. spanhel , m. haase , h. weller and a. henglein , j. am .
soc . * 109 * , 5649 ( 1987 ) .
d. dorfs and a. eychmller , nano lett .
* 1 * , 663 ( 2001 ) .
x. peng , m.c .
schlamp , a.v .
kadavanich and a. p. alivisatos , j. am .
soc . * 119 * , 7019 ( 1997 ) . w. l. wilson , p. j. szajowski , l.e .
brus , science * 262 * , 1242 ( 1993 ) .
kortan , r. hull , r.l .
opila , m.g .
bawendi , m.l .
steigerwald , p.j .
carroll , l.e .
brus j. am .
soc . * 112 * , 1327 ( 1990 ) .
hoener , k.a .
allan , a.j .
bard , a. campion , m.a .
fox , t.e .
mallouk , s.e .
webber and j.m .
white , j. phys .
chem . * 96 * , 3812 ( 1992 ) .
a.mews , a.eychmuller , m.giersig , d.schooss,and h.j.weller,phys.chem . * 98 * , 934 ( 1994 ) .
m. danek , k.f .
jensen , b.c .
murray , and g.m .
bawendi , chem.mater.*8 * , 173 ( 1996 ) .
m.a.hines , and p.g.sionnest , j.phys.chem.100 , 468 ( 1996 ) .
n. pradhan , b. katz and s. efrima , j. phys . chem . *
b107 * , 13843 ( 2003 ) .
m. danek , k.f .
jensen , c.b .
murray , m.g .. bawendi appl .
65 * , 2795 ( 1994 ) .
i.h . campbell and p.m. fauchet , solid state commun .
* 58 * , 739 ( 1986 ) .
h. richter , z.p .
wang and l. ley , solid state commun .
* 39 * , 625 ( 1981 ) .
nemanich , s.a .
solin and r.m .
martin , phys .
rev . * b23 * , 6348 ( 1981 ) .
r. tsu , h. shen and m. dutta , appl .
* 16 * , 112 ( 1992 ) .
sood , k. jayaram and d.v.s .
muthu , j. appl . phys . * 72 * , 4963 ( 1992 ) .
a. roy , k. jayaram and a.k .
sood , solid state commun . * 89 * , 229 ( 1994 ) .
a. roy and a. k. sood , phys . rev .
* b53 * , 12127 ( 1996 ) .
r. fuchs and k.l .
kliewer , j. opt .
am * 58 * , 319 ( 1968 ) .
baranov , yu .
p. rakovich , j.f .
donegan , t.s .
perova , r.a .
moore , d.v .
talapin , a.l .
rogach , y. masumoto and i. nabiev , phys . rev . *
b68 * , 165306 ( 2003 ) .
norris , a. scara , c.b .
murray , m.g .
bawendi , phys .
72 * , 2612 ( 1994 ) .
w. hoheisel , v.l .
colvin , c.s .
johnson and a.p .
alivisatos , j. chem . phys .
* 101 * , 8455 ( 1994 ) .
j.nanda , b.a.kuruvilla , and d.d.sarma , phys.rev . *
b59 * , 7473 ( 1999 ) .
liu , h. guo , z - h zhang , r. li , w. chen and z - g .
wang physica e * 8 * 174 ( 2000 ) .
x - ray powder diffraction , jcpds file .
l.m.ramaniah , and s.v.nair , phys.rev .
* b47 * , 7132 ( 1993 ) .
m. nirmal , d.j .
norris , m. kuno and m.g .
bawendi , phys .
lett . * 75 * , 3728 ( 1995 ) . c. trallero - giner , a. debernardi , m. cardona , e. menendez proupin , a. i. ekimov , phys
* b57 * , 4664 ( 1998 ) . h. bliz and w. kress , _ phonon dispersion relations in insulators _ ( springer verlag , p. 97
, 1979 ) ; also see reference [ 2 ] p.63 & p.143 .
r. ruppin , r. englman , rep .
. phys . * 33 * , 146 ( 1970 ) .
fedorov , a.v .
baranov , k. inoue , phys .
rev . * b56 * , 7491 ( 1997 ) .
f. comas , c. trallero - giner , n. studart and g.e .
marques , phys . rev . *
b65 * 073303 ( 2002 ) .
handbook of chemistry and physics , david r. lide , ed . in chief , crc press . | optical and vibrational properties of bare and cds shelled cdse nanocrystalline particles are investigated . to confirm the formation of such nanocrystals in our samples we estimate their average particle sizes and size distributions using tem measurements . from the line profile analysis of the images the core - shell structure in the particles has been confirmed .
the blue shift in optical absorption spectra , analyzed using theoretical estimates based on the effective bond order model , establishes the electron confinement in the nanoparticles .
unique characteristics of the nanocrystals ( which are absent in the corresponding bulk material ) , such as confinement of optical phonons and the appearance of surface phonons , are then discussed .
making use of the dielectric response function model we are able to match the experimental and theoretical values of the frequencies of the surface phonons .
we believe that our studies using optical probes provide further evidence on the existence of core - shell structures in cdse - cds type materials . |
the study of the properties of young massive star clusters in our local group can be done by means of the census of the stars that belong to the corresponding object and the analysis of their properties using a colour - magnitude diagram . in the case of those objects
whose stellar population can not be resolved with the facilities that are nowadays available , it is necessary to appeal to other tecniques . a possible approach is to study the massive stars by inspecting their effects on the surrounding interstellar medium ( ism ) , including the supply of mechanical energy from stellar winds and supernovae explosions , the ejection of new metals produced in the interiors of the stars , or the excitation and ionization of the atoms in the ism by the energetic photons emitted by the stars .
these factors make the optical spectrum of an hii region to be frequently characterized by the presence of bright prominent recombination lines emitted by hydrogen and helium and of collisional lines emitted by the metals .
although the shape of the stellar continuum and some stellar features are sometimes detectable ( _ e.g. _ the wolf - rayet features ) , the most important source of information about the properties of the ionizing stellar clusters can only be the relative intensities of these lines .
these intensities are mainly dominated by the so - called functional parameters , including the metallicity ( @xmath2 ) , the ionization parameter ( @xmath3 , _
i.e. _ the ratio between ionizing photons and the density of particles ) , and the equivalent effective temperature ( @xmath0 ) .
this last is the unique parameter that only depends on the spectral energy distribution ( sed ) of the ionizing star or cluster .
this parameter , denoted by the greek letter @xmath1 , was introduced by vlchez & pagel ( 1988 ) to derive @xmath0 using only the information from emission - line intensities in the optical spectrum .
it is defined as : @xmath4 3727 \aa/[oiii ] 4959,5007 \aa } { [ sii ] 6717,6731 \aa/[siii ] 9069,9532 } + o(z)\ ] ] depending mainly on the ionic abundances of oxygen and sulphur ions .
this ratio is basically a function of the ratio between the number of ionizing photons of o@xmath5 ( i.p .
= 35.1 ev ) and s@xmath5 ( i.p .
= 23.3 ev ) and has also a certain dependence on the overall metallicity [ @xmath6 when it is derived directly from the emission - line intensities . in the ideal case of a blackbody
it only has a linear relation with the slope of the sed , but for young massive star atmospheres it noticeably changes for different model atmospheres . in upper panels of fig .
1 we show the seds of wm - basic ( pauldrach et al . , 2001 with an expanding spherical geometry ) at left and tlusty ( hubeny & lanz , 1995 , with a plane - paralel geometry ) at right for a same value of @xmath0 ( 40000 k ) and different values of the metallicity
. the slope and the shape of the seds can be very different in the energetic range covered by the optical @xmath1 parameter ( simn - daz & stasiska , 2008 ) .
an alternative to the optical @xmath1 parameter was proposed for the mid - infrared spectral range with emission lines of [ neii ] ( 12.8 @xmath7 m ) , [ neiii ] ( 15.6 @xmath7 m ) , [ siii ] ( 18.7 @xmath7 m ) , and [ siv ] ( 10.5 @xmath7 m ) ( martn - hernndez et al . , 2002 ;
morisset et al . 2004 ) , but this involves an even more energetic range between the ionization potentials of s@xmath8 ( i.p .
= 34.8 ev ) and ne@xmath9 ( i.p .
= 40.5 ev ) , where discrepancies between the different synthesis model atmospheres are larger ( see prez - montero & vlchez , 2009 , for a more detailed discussion ) . in the lower panels of fig .
1 we show the relation between the optical emission - line ratios of oxygen ( [ oii]/[oiii ] ) and sulphur ( [ sii]/[siii ] ) for a sample of objects with emission - line like spectra compiled by prez - montero & vlchez ( 2009 ) .
the solid lines are quadratical fits to different sets of cloudy ( ferland et al . , 1998 )
photoionization models with different values of @xmath0 ( from 35 kk to 50 kk ) . at
left , these models are calculated using wm - basic stellar atmospheres and , at right , tlusty ( see prez - montero & vlchez to see a deeper descripion of these models ) . as can be seen
, there are substantial differences between the two studied synthesis model atmospheres .
these differences prevent an accurate determination of the absolute value of @xmath0 in hii regions but , apparently , allows a good relative characterization for different families of objects . in both panels , circumnuclear star forming regions ( cnsfrs ) and hii
galaxies have the higher temperatures and , in contrast , smaller regions in our galaxy and the magellanic clouds have lower temperatures .
upper panels : seds of synthesis model atmospheres from wm - basic ( left ) and tlusty ( right ) for different values of the metallicity and @xmath0 = 40 kk .
the vertical solid lines represent the ionization potentials of s@xmath5 and o@xmath5 . in the lower panels we show the relation between the ratios of optical emission line - intensities [ oii]/[oiii ] and [ sii]/[siii ] for different objects .
the solid lines show the quadratical fits to sets of photoionization models with stars of different equivalent effective temperatures from wm - basic ( left ) and tlusty ( right ) .
, title="fig : " ] upper panels : seds of synthesis model atmospheres from wm - basic ( left ) and tlusty ( right ) for different values of the metallicity and @xmath0 = 40 kk .
the vertical solid lines represent the ionization potentials of s@xmath5 and o@xmath5 . in the lower panels we show the relation between the ratios of optical emission line - intensities [ oii]/[oiii ] and [ sii]/[siii ] for different objects .
the solid lines show the quadratical fits to sets of photoionization models with stars of different equivalent effective temperatures from wm - basic ( left ) and tlusty ( right ) .
, title="fig : " ] upper panels : seds of synthesis model atmospheres from wm - basic ( left ) and tlusty ( right ) for different values of the metallicity and @xmath0 = 40 kk .
the vertical solid lines represent the ionization potentials of s@xmath5 and o@xmath5 . in the lower panels we show the relation between the ratios of optical emission line - intensities [ oii]/[oiii ] and [ sii]/[siii ] for different objects .
the solid lines show the quadratical fits to sets of photoionization models with stars of different equivalent effective temperatures from wm - basic ( left ) and tlusty ( right ) .
, title="fig : " ] upper panels : seds of synthesis model atmospheres from wm - basic ( left ) and tlusty ( right ) for different values of the metallicity and @xmath0 = 40 kk .
the vertical solid lines represent the ionization potentials of s@xmath5 and o@xmath5 . in the lower panels we show the relation between the ratios of optical emission line - intensities [ oii]/[oiii ] and [ sii]/[siii ] for different objects .
the solid lines show the quadratical fits to sets of photoionization models with stars of different equivalent effective temperatures from wm - basic ( left ) and tlusty ( right ) .
, title="fig : " ]
one ot the possible caveats when using the @xmath1 parameter to find out @xmath0 in massive star clusters is due to that the information comes from the ionized gas .
this implies that possible variations in the properties of the gas other than the functional parameters , such as gas geometry or density , or dust structure could also affect to the @xmath1 parameter .
we used as a test to check to what extent these variation can affect to the determination of @xmath0 the cloudy photoionization models of the 2d structure of the giant hii region ngc 595 ( prez - montero et al . , 2011 ) .
the observational inputs for these models come from integral field spectroscopy taken with the pmas instrument at the caha 3.5 m telescope in the optical range between 3700 - 6800 ( relao et al . , 2010 ) and photometric information from spitzer space observatory at 8 and 24 @xmath7 m filters . in these models
it was assumed a single ionizing source for all the hii region according to the properties of the cmd diagram derived by malumuth et al .
these models reproduce the variations in the optical and mid - ir properties for different elliptical aperture regions with similar observational properties by assuming different matter - bounded geometries combined with different dust - to - gas ratios , in good agreement to the dust - to - gas ratio derived in the integrated region ( prez - montero et al . , 2011 ) . in fig .
2 , we show the comparison between the models and the observations in each elliptical aperture . variations of the 8 to 24 @xmath7 m ratio ( at left ) and optical [ oii]/[oiii ] at right for different elliptical apertures around the central ionizing cluster in the giant hii region ngc 595 .
black points represent the observed properties and the red solid line the results from tailored photoionization models for each of these apertures ( from prez - montero et al . ,
2011 ) , title="fig : " ] variations of the 8 to 24 @xmath7 m ratio ( at left ) and optical [ oii]/[oiii ] at right for different elliptical apertures around the central ionizing cluster in the giant hii region ngc 595 .
black points represent the observed properties and the red solid line the results from tailored photoionization models for each of these apertures ( from prez - montero et al . , 2011 ) ,
title="fig : " ] in fig .
3 we show the relation between the two optical emission - line ratios [ oii]/[oiii ] and [ sii]/[siii ] involved in the @xmath1 parameter with the fits to the photoionization models with tlusty atmospheres shown in lower right panel of fig .
1 . the black circles represent the values derived from the tailored models for each region in ngc 595 . as can be seen , with the exception of some models ,
the most part of the elliptical apertures lie in regions consistent with a single value of @xmath0 across the nebula , independently on the inner variations in the geometry of the gas and in the dust - to - gas ratio .
the value of @xmath0 ( between 35 kk and 40 kk ) is consistent to the value associated with the properties of the cluster derived by malumuth using a cmd diagram , which is around 40 kk .
it must be noted , however , that neither the slope of the grid of models for ngc 595 nor the slopes of the grids of models for different @xmath0 are equal to 1 , which is the value of the slope for the lines with different @xmath1 across the diagram , so this implies that this parameter must not be used as one - dimensional .
in contrast , its analysis must be undertaken using a 2d approach as in this case .
for instance , in the models of ngc 595 , there is a variation in @xmath1 in a factor 2 which is not contradictory with an homogeneous value of @xmath0 in the models .
relation between the optical emission - line ratios [ oii]/[oiii ] and [ sii]/[siii ] .
solid lines represent the same fits to grids of photoionization models of different @xmath0 for tlusty atmospheres represented in fig .
black circles represent the models for ngc 595 derived by prez - montero et al .
the determination of properties of a massive young stellar cluster can be obtained by using information from the emission line spectrum from the ionized gas in those cases where the stellar population of the cluster can not be resolved .
the @xmath1 parameter as analyzed in 2 dimensions is a robust method to derive the equivalent effective temperature of a cluster , even in those cases where there are variations in the geometry of the gas or in the dust - to - gas ratio .
the most important caveat appears as a consequence of the disagreements in the involved energetic ranges between different synthesis model atmospheres , but it is possible to use this to study the relative variations of @xmath0 in homogeneous samples of objects .
this is the case of the variation of @xmath0 across spiral disks found by prez - montero & vlchez ( 2009 ) , which points to a hardening of the ionizing radiation above all in spiral galaxies of low dynamical mass and luminosity with late morphological type .
this work has been supported by the projects aya2007- 67965-c03 - 02 of the spanish national plan for astronomy and astrophysics and csd2006 00070 1st science with gtc of the spanish ministry of science and innovation ( micinn ) . | the softness parameter is based on the relative intensity of several optical emission lines emitted by the gas ionized by young massive star clusters and can be used to derive the equivalent effective temperature ( @xmath0 ) in those objects whose stellar population can not be resolved .
this method has several uncertainties due to the disagreement between different synthesis model atmospheres but it is robust to study the relative variations between objects . following the 2d photoionization models of the giant hii region ngc 595 ( prez - montero et al .
2011 ) we show that the determination of @xmath0 with the @xmath1 parameter is also robust in different regions of a same object with large variations in the geometry of the gas and in the dust - to - gas ratio . |
we establish explicitly disentangled real and imaginary amplitudes for pp elastic scattering based on a qcd motivated model . with impact parameter representation @xmath1 and its fourier transform in @xmath2 space both represented by simple analytical forms
, we are able to control unitarity and dispersion relation constraints , and provide geometric interpretation of the interaction range .
the regularity obtained in the description of the data and the physical interpretation give reliability to the proposed amplitudes .
the amplitudes of pp elastic scattering originally constructed through profile functions in @xmath3 -space are written @xmath4with the usual gaussian forms plus the characteristic shape functions @xmath5~.
\label{shape - b}\]]the label @xmath6 @xmath7 indicates either the real or the imaginary part of the complex amplitude . for large @xmath3 ,
corresponding to peripheral collisions , the amplitudes fall down with a yukawa - like tail , @xmath8 , that reflects effects of virtual partons ( modified gluon field ) at large distances in the stochastic vacuum model @xcite .
the comparison with @xmath9 data and determination of parameters are made with the amplitudes in @xmath10-space .
the quantities @xmath11 obtained by fourier transform of eq .
( [ b - amplituden ] ) are written @xmath12 and the shape functions converted to @xmath13 space take the form @xmath14~.
\label{psi_st } \end{aligned}\ ] ] the complete analysis of elastic scattering @xcite requires also the contributions from the coulomb interaction at small @xmath15 and from perturbative 3-gluon exchange at large @xmath15 .
the fixed parameter @xmath16 is given by the correlation length of the gluon condensate .
all parameters have been determined as smooth functions of @xmath17 and the properties of the amplitudes ( magnitudes , signs , zeros ) have been described in detail @xcite . in our normalization the elastic differential and the total cross sections
are written @xmath18 = \frac{d\sigma^i(s , t)}{dt } + \frac{d\sigma^r(s , t)}{dt } ~ ~ ~ , ~ ~ ~ \sigma(s ) = ( \hbar c)^2~ 4\sqrt{\pi } ~ t_{i}^{n}(s , t=0)~.\end{aligned}\ ] ] [ [ differential - cross - sections - and - amplitudes - in - the-1.8-to-14-tev - range - energy_range ] ] differential cross sections and amplitudes in the 1.8 to 14 tev range [ energy_range ] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ in fig . [ pilhas - fig ] we show the predictions of @xmath9 for the lhc energies 2.76 , 8 , 13 and 14 tev . in the rhs
we use the energy @xmath19 tev as an example to show the imaginary and real nuclear amplitudes @xmath20 , @xmath21 as functions of @xmath15 as predicted by eq.([t_space],[psi_st ] ) .
the interplay of the imaginary and real amplitudes at mid values of @xmath15 is responsible for the dip - bump structure of the differential cross section , that was shown before @xcite for @xmath22 tev . for @xmath23
the real part becomes dominant , with positive sign .
our description @xcite of the elastic scattering data at 7 tev from the totem collaboration @xcite reproduces n=165 points in @xmath24 with an impressive squared average relative deviation @xmath25 .
characteristic quantities at this energy are @xmath26 mb , @xmath27 mb , @xmath28 gev@xmath29 , that compare extremely well with the values @xmath30 mb , @xmath31 mb , @xmath32 gev@xmath29 published by totem .
the preliminary totem data of @xmath34 at 8 tev seem to be regular enough for our analysis .
the data discussed below are taken ( reading by eye ) from slides of talks by members of the totem collaboration @xcite .
as far as we can read from the presentation slides , we identify 212 data points in three sets : 1 ) n=97 points in the forward interval @xmath35 gev@xmath36 ; 2 ) n=45 points in an intermediate interval @xmath37 gev@xmath36 ; 3 ) n=70 points in a mid @xmath15 range @xmath38 gev@xmath36 .
this information is transferred to the plot in fig .
[ data_8tev - fig ] , together with our calculation .
the quality of the representation seems to be equivalent to that of our treatment of the 7 tev data . for the inelastic cross section we assume the difference @xmath43 and
then we have 73.26 mb at 7 tev .. published values of the totem coll . using different methods are @xmath44 @xcite , @xmath45 @xcite and @xmath46 totem_7c .
alice coll .
@xcite gives @xmath47 mb , and atlas coll .
@xmath48 mb @xcite .
there are also cms results @xcite with non - informed missing contributions . in these measurements
there are extrapolations using monte carlo models to include diffractive events of low mass .
all these results are compatible with our calculations .
the analysis of compatibility for the 1.8 tev measurements of @xmath51 by cdf and e811 in fermilab @xcite suggests the value @xmath52 mb , that with our @xmath53 value gives @xmath54 mb .
we have 58.89 mb for 1.8 tev , once more in very good agreement . at 57 tev
the auger cosmic ray experiment @xcite , using other models for pp input , obtains @xmath55 mb , while our extrapolation gives 101 mb .
this measurement @xcite is discussed together with other cr extended air showers ( eas ) experiments , using our inputs and a basic glauber method to connect pp and p - air processes .
our calculation reproduces well all cr data for p - air cross sections with energies @xmath0 ( in the pp system ) up to 100 tev .
the proposed amplitudes have simple forms , evaluated with few operations with elementary functions .
the shape of the dip - bump behavior results from a delicate interplay of the imaginary and real amplitudes .
all intervening quantities and derived properties present smooth energy dependences .
the zeros of the real and imaginary parts have very regular displacements , converging to finite limits as the energy increases , showing remarkable connection with positions and heights of dips , bumps and inflections in @xmath34 .
the slopes @xmath39 and @xmath40 at the origin , with their characteristic difference in values , together with the ratio @xmath53 , are essential quantities in the definition , through the unique analytical forms of the amplitudes , of the properties of the observed @xmath34 in the whole @xmath10 range .
their values are thus fixed with high accuracy .
it is important that the slopes show quadratic dependence in @xmath65 , instead of the linear dependence suggested by regge phenomenology .
the properties of ratios ( with respect to the total cross section ) of slopes and of integrated elastic cross sections , that tend to finite asymptotic limits , show that the hypothesis of a black disk limit in the behaviour of the pp interaction seems to be excluded by phenomenology .
we also discuss the geometrical interpretation of our amplitudes , showing that the effective interaction radius in @xmath3-space increases with the energy .
our amplitudes obey a geometric scaling in asymptotic energies , and indicate that the profile function @xmath66 tends to a universal ( energy independent ) function with respect to a scaling variable , @xmath67 .
this universal function exhibits a considerable diffused surface , indicating a scenario different from the commonly accepted black disk . at lhc energies
, the saturation seems to start ( the central value of @xmath68 is almost unity ) , but the asymptotic profile is still far and only can be reached for @xmath69 tev .
the connection between the diffused surface of long range and inelastic diffractive processes will be an interesting line of investigation .
we believe that our analytic representation of the scattering amplitudes will serve as important guidance for the future measurements in lhc , and also for a theoretical understanding of the intermediate region of partonic saturation phenomena . at this
diffraction 2014 conference related work on the black - disk behaviour @xcite and on the determination of pp amplitudes at lhc energies @xcite were presented , showing that this is an important field of research at the present .
dosch , _ phys .
b _ * 190 * , 177 ( 1987 ) ; h.g .
dosch , e. ferreira , a. kramer _ phys .
d _ * 50 * , 1992 ( 1994 ) .
e. ferreira and f. pereira , _ phys .
d _ * 59 * , 014008 ( 1998 ) ; _ phys .
d _ * 61 * , 077507 ( 2000 ) .
a. kendi kohara , e. ferreira and t. kodama , _ eur .
j. c _ , * 73 * , 2326 ( 2013 ) .
a. kendi kohara , e. ferreira and t. kodama , phys .
d * 87 * , 054024 , ( 2013 ) .
fagundes , m.j .
menon , p.v.r.g .
silva , j. phys .
g40 ( 2013 ) 065005 ; ibid : arxiv : 1410 - 4423[hep - ph ] .
c. bourrely , j.m .
myers , j.soffer and t.t .
wu , _ phys .
rev.d _ * 85 * , 096009 ( 2012 ) .
selyugin , _ eur .
j. c_(2012 ) , 72:2073 ; talk presented at diffraction 2014 . | using a unified analytic representation for the elastic scattering amplitudes of pp scattering valid for all high energy region , the behavior of observables in the lhc collisions in the range @xmath0 = 2.76 - 14 tev is discussed .
similarly to the case of 7 tev data , the proposed amplitudes give excellent description of the preliminary 8 tev data .
we discuss the expected energy dependence of the observable quantities , and present predictions for the experiments at 2.76 , 13 and 14 tev .
[email protected] , address = instituto de fsica , universidade federal do rio de janeiro , c.p .
68528 , rio de janeiro 21945 - 970 , rj , brazil [email protected] [email protected] |
establishing the properties of the chiral phase transition in qcd is one of the outstanding problems in lattice simulations of qcd thermodynamics .
although several attempts have been undertaken to establish the unique scaling behavior in the limit of vanishing light quark masses no clear - cut evidence for the expected @xmath4 scaling , related to the underlying chiral symmetry of the qcd lagrangian , has been found .
this may have several reasons related to common problems in lattice calculations : cut - off effects , finite volume effects , explicit flavor symmetry breaking and/or too large quark masses .
we will discuss here yet another effect that influences the scaling behavior of thermodynamic quantities in the vicinity of the finite temperature chiral transition of qcd and may compete with the universal scaling in the vicinity of @xmath2 .
this is related to singularities in chiral observables , e.g. the chiral susceptibility , that are induced at any temperature in the chiral symmetry broken phase .
the existence of these singularities induced by fluctuations of goldstone modes are well known from the analysis of 3 and 4 dimensional statistical models and field theories with global @xmath5 ) symmetries that are spontaneously broken at low temperature @xcite .
at vanishing temperature properties of qcd in the light quark sector are described by an effective 4-d , @xmath7 symmetric lagrangian ; chiral symmetry is broken and , what is of interest here , at non - zero quark mass the light quark chiral condensate , @xmath8 , receives logarithmic corrections proportional to @xmath9 that are due to the presence of light goldstone modes @xcite .
this leads to a logarithmic divergence of the chiral susceptibility , @xmath10 . in three dimensions the analog to these logarithmic terms
are corrections that are proportional to the square root of the quark mass @xcite . as a consequence @xmath11 will diverge in the chiral limit like @xmath12 . in the vicinity of the qcd phase transition a 3-d ,
@xmath7 symmetric effective theory is expected to describe the universal properties of the transition @xcite which will also control the properties of the chiral condensate for temperatures close to @xmath2 . we thus expect that in the chiral limit @xmath11 will diverge for all values of the temperature @xmath13 , @xmath14 m_l^{1/\delta -1 } & , \ ;
t = t_c \end{cases } \label{scaling}\ ] ] where @xmath15 for 3-d @xmath7 as well as @xmath16 symmetric spin models .
the appearance of a square root singularity has , in fact , been established in the vicinity of the chiral phase transition of an @xmath17 gauge theory coupled to adjoint fermions @xcite where it is apparent already for moderately small values of the quark mass .
similar evidence for the role of light goldstone modes in qcd , however , has been missing so far . in fig .
[ fig : pbp ] we show the light quark chiral condensate , @xmath8 calculated in ( 2 + 1)-flavor qcd with improved staggered fermions ( p4-action ) on lattices of size @xmath18 for various values of the gauge coupling .
the figures show results for @xmath8 , plotted versus the light quark mass ( @xmath1 ) expressed in units of the strange quark mass ( @xmath19 ) ( left ) and the square root of this ratio ( right ) .
the right hand figure indeed suggests that the strong curvature of @xmath8 as function of @xmath1 is well accounted for by a @xmath20 dependence for @xmath0 . in any case , the weaker than linear dependence of @xmath8 on @xmath1 leads to a strong quark mass dependence of the chiral susceptibility , @xmath11 .
this susceptibility receives two contributions , usually referred to as disconnected and connected parts , @xmath21 , with @xmath22 here @xmath23 denotes the staggered fermion matrix for light quarks . in general , both contributions are sensitive to the goldstone modes and will receive divergent contributions , @xmath24 @xcite .
we will restrict our analysis here to the disconnected part . in fig .
[ fig : chi ] we show @xmath25 calculated for various values of @xmath1 . in these calculations the strange quark mass is close to its physical value and @xmath1 reaches values well below the physical one ; a ratio @xmath26 corresponds in the continuum limit to the physical quark mass ratio .
the strong quark mass dependence in a wide temperature interval is obvious from fig .
[ fig : chi](left ) ; its dominant contribution in the low temperature regime is indeed well described by @xmath12 as can be seen from the rescaled data shown in fig .
[ fig : chi](right ) .
we note that this seems to hold for all quark masses , @xmath27 , although at present our data sample for the lightest quark mass is not yet good enough to substantiate this statement also for this light mass value , which corresponds to a light pseudo - scalar mass of about @xmath28 mev .
we furthermore , note that the square root behavior also dominates the quark mass dependence in the peak region , _
i.e. _ there is no evidence for a stronger singularity , which is expected to arise from the singular part of the partition function of 3-d , o(4 ) ( or o(2 ) ) symmetric models , @xmath29 .
the general features of the quark mass dependence of the disconnected part of the chiral susceptibility are also apparent in calculations currently performed by the hotqcd collaboration on lattice with temporal extent @xmath30 @xcite , _ i.e. _ closer to the continuum limit . in fig .
[ fig : chi8 ] we show @xmath25 and the rescaled susceptibility .
the left hand part of the figure shows that for temperatures above the transition region the susceptibilities quickly become quark mass independent .
this is expected as the chiral condensates will be linear in the quark mass in the chirally symmetric phase of qcd .
below the transition region we again observe a strong quark mass dependence that is consistent with @xmath31 .
the weak quark mass dependence of @xmath11 in the high temperature regime suggest that the susceptibility can already be described by the limiting form for the @xmath32 limit , @xmath33 a \left(\frac{t - t_c}{t_c}\right)^{-\gamma } & , \ ; t > t_c \end{cases } \label{chiral}\ ] ] a consistent picture for the chiral susceptibility thus seems to emerge despite the missing evidence for @xmath4 scaling for its peak height .
the numerical calculations presented here have been carried out on the qcdoc supercomputers of the riken - bnl research center , the bluegene / l computers at the new york center for computational sciences and the llnl .
the work of fk has been supported by doe under contract no .
de - ac02 - 98ch10886 .
d.j.wallace and r.k.p .
zia , phys . rev .
* b12 * ( 1975 ) 5340 j. gasser and h. leutwyler , phys . lett .
* b184 * ( 1987 ) 83 p. hasenfratz and h. leutwyler , nucl . phys .
* b343 * ( 1990 ) 241 a. v. smilga , phys . lett . *
b318 * ( 1993 ) 531 ; + a. v. smilga and j. j. m. verbaarschot , phys .
* d54 * ( 1996 ) 1087 r. pisarski and f. wilczek , phys . rev . * d29 * ( 1984 ) 338 f. karsch and m. ltgemeier , nucl
b550 * ( 1999 ) 449 ; + j. engels , s. holtmann and t. schulze , nucl .
* b724 * ( 2005 ) 357 m. cheng et al .
( rbc - bielefeld ) , phys . rev .
* d74 * ( 2006 ) 054507 c. detar , pos 001 ( lat2008 ) + r. gupta [ for the hotqcd collaboration ] , pos 170 ( lat2008 ) , arxiv:0810.1764 [ hep - lat ] . | we provide evidence for the influence of thermal fluctuations of goldstone modes on the chiral condensate at finite temperature .
we show that at fixed temperature , @xmath0 , in the vicinity of the chiral transition temperature this leads to a characteristic dependence of the chiral condensate on the square root of the light quark mass ( @xmath1 ) , which is expected for 3-dimensional models with broken o(n ) symmetry .
as a consequence the chiral susceptibility shows a strong quark mass dependence for all temperatures below @xmath2 and diverges like @xmath3 in the chiral limit .
qcd thermodynamics , chiral symmetry breaking , goldstone modes 11.15.ha , 11.10.wx , 12.38.gc , 12.38.mh |
by large - zenith - angle observations we mean targeting an astronomical source with a ground - based telescope when it is more than 45@xmath0 away from the zenith .
many authors have stressed the potential of using imaging atmospheric cherenkov telescopes ( cts ) for observations of sources of high - energy @xmath1-rays at large zenith angles ( see e.g. @xcite and references therein ) .
several successful observations have already been made ( e.g. @xcite , @xcite , @xcite , @xcite , @xcite ) . since the effective depth of the atmosphere increases to a good approximation as @xmath2 ( where @xmath3 denotes the zenith angle ) , the distance of the telescope to the shower maximum increases proportionally due to the curvature of the earth .
the cherenkov light produced in the shower illuminates a larger area on the ground which leads to an increase in the effective collection area of the ct .
this increase should very roughly be as large as @xmath4 but is in reality even larger due to changes in the shape of the lateral distribution of the cherenkov photons density .
the expected drawbacks of large-@xmath3 observations are an increased energy threshold and possibly worse background suppression since the average cherenkov photon density caused by the shower at the telescope decreases with increasing @xmath3 roughly as @xmath5 .
furthermore , the angular extension of the shower image decreases .
this means that with a given ct photomultiplier camera , the image is fainter and contained in fewer pixels .
there is less information for distinguishing @xmath1-induced showers from the hadronic background showers .
it is a priori not obvious by how much the gain in collection area will compensate for the decreasing image quality .
a camera with higher angular resolution should suffer less from the effects of shrinking images .
it is therefore especially interesting to investigate the performance of the new granite iii camera on the whipple telescope ( mt .
hopkins , arizona ) which has an angular resolution of 0.12@xmath0 in its central region , the best resolution ever available on this ct .
this camera is described in finley et al .
( these proceedings ) . in this article
we derive the performance of granite iii as a function of the zenith angle @xmath3 and investigate possible improvements to the present standard data analysis which is optimized for small @xmath3 .
we do this by examining the performance of the standard analysis at different zenith angles using simulated @xmath1-showers and real telescope data from off - source observations .
the improvements are then verified using observations of mkn 421 .
it is especially helpful that mkn 421 was in a high state during the observations .
the monte carlo ( mc ) data used in this study was specially produced using the simulation code described by @xcite .
the agreement of this mc data with the data obtained from the whipple telescope in the 2000/2001 observing season was thoroughly tested .
special measurements of the telescopes point spread function , photomultiplier efficiency , light - cone collection efficiency and the photoelectron - to - digital - count conversion were made and the results included in the simulation ( krennrich et al .
realistic noise was added to the pixel values in order to simulate the night sky background light .
since all shower image parameters are expected to change linearly with the average distance to the shower maximum which is in turn expected to scale with @xmath2 , the mc data was produced at equidistant values of @xmath2 .
table [ tab - mc ] gives an overview of the production parameters .
only gamma primaries were produced . for studies of hadronic showers ,
off - source observations with the whipple telescope were used .
.[tab - mc ] production parameters of the monte carlo data for @xmath1-induced showers used in this study .
@xmath6 denotes the primary energy , @xmath7 the impact parameter . [
cols="<,^,^,^,^,^,^ " , ] comparing figs .
[ fig - efcoltrig ] and [ fig - efcolscuts ] , improvements seem possible mainly at large energies and not so much at large zenith angles .
however , as fig.[fig - effplot3 ] ( left plot ) shows , the background rejection of the supercuts2001 deteriorates dramatically with increasing @xmath3 .
we examine the individual image parameters in order to identify the reasons .
the efficiency ( eff . ) of the supercuts2001 for simulated @xmath1s ( solid symbols ) and real background data ( open symbols ) . left : eff .
of all cuts after the max1 & max2 cut .
right : ( a , b ) eff . of the upper length and width cuts after all other cuts , ( c , d ) eff . of the dist cut after all other cuts , ( e , f ) eff . of the length / size cut after all other cuts.,width=264 ] the alpha distribution after cuts has a shape which is independent of @xmath3 .
the cut at 15@xmath0 is optimal for all @xmath3 .
the dist cut ( @xmath8dist@xmath9 ) has ( after all other cuts ) a nearly @xmath3-independent efficiency both for @xmath1s and background ( see fig.[fig - effplot3 ] ) .
the peak of the dist distribution for @xmath1s before cuts moves towards smaller values as @xmath3 increases , but for the @xmath1s below @xmath10 the alpha distribution is so broad that more than 60% are eliminated by the alpha cut ( for all @xmath3 ) .
also the energy resolution for events with small dist is poor and since the solid angle is small , they constitute a small part of the total events .
the upper dist cut is determined by the field of view of the camera .
it is therefore not necessary to vary the dist cut with @xmath3 .
comparison of the width and length distributions ( after all other cuts ) of simulated @xmath1 events at small and large zenith angles .
see text.,width=283 ] comparison of the width and length distributions ( after all other cuts ) for observations of mkn 421 and off - source positions at small and large zenith angles .
see text.,width=283 ] the lower cuts on width and length ( width@xmath11 and length@xmath12 ) are dictated by the optics of the telescope and therefore nearly independent of @xmath3 .
this is illustrated by figs .
[ fig - widthlength ] and [ fig - m4widthlength ] .
while their lower end remains at the same values , the width and length distributions become , however , more narrow as @xmath3 increases and a @xmath3-independent cut has too large background efficiency ( see also fig.[fig - effplot3 ] ) .
the upper cuts on width and length ( width@xmath13 and length@xmath14 in the supercuts2001 ) are therefore the prime candidate for improving the background rejection .
finally , the cut on length / size ( @xmath15/photoelectron ) ) shows a constant @xmath1-efficiency and a decreasing background efficiency with increasing @xmath3 ( see fig.[fig - effplot3 ] ) . here , little improvement seems possible . the upper width and length cut values with @xmath1-efficiency 84% as a function of @xmath2 .
the cuts are calculated by adding one standard deviation to the mean of the distribution .
see text.,width=241 ] in order to improve the background rejection of the upper cuts on width and length , we examine as a function of @xmath3 the cut value with a @xmath1-efficiency of 84% .
since both the width and the length distributions for pure @xmath1s after cuts are gaussian to a very good approximation , we determine the mean and the standard deviation of the simulated distributions at each value of @xmath3 by fitting a gaussian .
the sum of mean and standard deviation is then plotted versus @xmath2 ( fig.[fig - widthxvsza ] ) . as one would expect from the geometry of the shower observation
, there is a linear dependence of the shower image size , and therefore also the width and length values under discussion , on @xmath2 .
the result of a linear fit to the two groups of points is given in fig.[fig - widthxvsza ] . in order to obtain an expression for @xmath3-dependent upper cuts on width and length ,
we scale the two lines such that they assume the supercuts2001 value at the zenith angle where the supercuts were optimized which is @xmath16 .
the resulting cuts are : @xmath17 we replace the corresponding @xmath3-independent cuts in the supercuts2001 by the cuts in equation [ equ - newcuts ] and compare the performance of this new set of cuts ( the `` zenith - angle - dependent supercuts '' ) with that of the standard supercuts2001 .
( a ) ( left ) the efficiency of the upper cuts on width and length after all other cuts for simulated @xmath1s and real background data in the case the supercuts2001 and the zenith - angle - dependent supercuts derived here .
( b ) ( right ) the total quality factor of the whole set of cuts ( upper two curves ) and the quality factor of only the upper cuts on width and length after all other cuts.,width=264 ] the distribution of the orientation angle alpha for observations of mkn 421 ( symbols ) and off - source observations ( lines ) after application of supercuts2001 ( upper curves ) and the zenith - angle - dependent supercuts derived here.,width=264 ] fig.[fig - effq ] shows that the @xmath3-dependent cuts improve the background rejection by up to a factor 2 at the largest zenith angles while they reduce the @xmath1-efficiency by only up to 10% .
furthermore , the overall quality factor ( @xmath1-efficiency divided by the square - root of the background efficiency ) drops from 4.7 to 2.7 for the standard supercuts2001 when going from @xmath18 to @xmath19 , while it drops only to 3.2 when the @xmath3-dependent cuts are used . the improvement in sensitivity is verified using the mkn 421 observations .
fig.[fig - alphademo ] shows the result for the data at the largest @xmath3 available in this study .
the expected improvement in significance from the q - factors in fig.[fig - effq ] is @xmath20 , the observed improvement is @xmath21 . for intermediate zenith angles ( 45@xmath0 - 56@xmath0 )
the expected improvement is @xmath22 , in the mkn 421 dataset we observe @xmath23 . at smaller @xmath3
the two sets of cuts converge and no significant improvement is expected or observed .
we have quantified the performance of the granite iii camera on the whipple telescope and the current standard analysis at zenith angles @xmath24 and identified improvements to the standard data analysis for @xmath25 .
aharonian , f. et al . , 1999 , a&a 349 , 29 chadwick , p.m. , et al . , 1999 , j.phys .
g 25 , 1749 konopelko , a. , et al . , 1999 ,
j.phys.g 25 , 1989 krennrich , f. , et al . , 1999 ,
apj 511 , 149 krennrich , f. , et al . , 2001 , veritas internal note . | the granite iii camera of the whipple cherenkov telescope at the fred lawrence whipple observatory on mount hopkins , arizona ( 2300 m a.s.l . ) has the highest angular resolution of all cameras used on this telescope so far .
the central region of the camera has 379 pixels with an individual angular diameter of 0.12@xmath0 .
this makes the instrument especially suitable for observations of gamma - induced air - showers at large zenith angles since the increase in average distance to the shower maximum leads to smaller shower images in the focal plane of the telescope .
we examine the performance of the telescope for observations of gamma - induced air - showers at zenith angles up to 63@xmath0 based on observations of mkn 421 and using monte carlo simulations .
an improvement to the standard data analysis is suggested . |
despite several years of studies , the chemical evolution history of the large magellanic cloud ( lmc ) as well its star formation history ( sfh ) is still poorly understood .
the lmc has experienced a complex sfh due to the interaction occurred both with the small magellanic cloud ( smc ) and the galaxy @xcite . a clear evidence of this quite complex evolution can be recognized in the lmc star cluster system , characterized by a wide range of age and metallicities .
the study of these cluster stellar populations reveals the existence of at least three components : the old ( @xmath1 13 gyr , @xcite , @xcite ) metal - poor component ( [ fe / h ] @xmath3 1.5 dex , @xcite , @xcite , @xcite ) probably formed during the first episode of star formation ( sf ) , the intermediate age component ( @xmath1 13 gyr , @xcite , @xcite ) which is the dominant one , and a young component ( @xmath4 1 gyr , @xcite , @xcite ) that include the last formed clusters .
the sfh of field stellar populations is less known .
@xcite found that the dominant stellar population in the central bar was formed between 4 - 6 gyr and 1 - 2 gyr ago .
this result was substantially confirmed by the simulations of @xcite who found that the formation of young stellar populations in the lmc bar is associated with efficient sf in the last few gyr ( in particular , @xmath1 2 gyr ago ) .
a comprehensive spectroscopic study of field populations was performed by @xcite measuring the infrared caii triplet of 373 giant stars located around the center of the bar .
they found a distribution peaked at [ fe / h ] @xmath1 0.4 dex with a low metallicity tail reaching [ fe / h ] @xmath4 2.0 dex . @xcite
have measured the infrared caii triplet in four fields at different radial distances ( @xmath1 3@xmath5 , 5@xmath5 , 6@xmath5 and 8@xmath5 , that is between 2.6 and 7 kpc ) from the center of lmc .
they found an average [ fe / h ] @xmath1 0.5 dex for the closest fields , a lower [ fe / h ] @xmath1 0.8 dex was found only in the more distant one .
a detailed spectroscopic analysis of 59 giant stars located in the inner disk at @xmath1 1.2 kpc from the center was performed by @xcite , who found a distribution peaked at [ fe / h ] @xmath1 0.75 dex .
this paper is part of a project devoted to investigate the kinematic and chemical properties of the stellar populations in the lmc through the use of high - resolution spectra , able to provide accurate information about the kinematics , the metallicity and the chemical abundance of individual elements of these stars .
previous papers of the project have discussed the properties of old @xcite , intermediate - age @xcite and young @xcite globular clusters ( gc ) in the lmc .
this paper is the first of the project dedicated to the kinematic and chemical characterization of the lmc field stellar populations : we present chemical patterns for a sample of 89 red giant branch ( rgb ) stars members of the lmc and located in the field around the old metal - poor gc ngc 1786 .
this represents to date the largest sample of field giants in the lmc for which high - resolution spectra have been obtained .
we have observed 91 stars located in the region surrounding the old gc ngc 1786 at @xmath1 2@xmath5 nw from the center of the lmc @xcite .
the spectra have been acquired with the multi - object spectrograph flames @xcite at the kueyen eso - vlt telescope .
the spectroscopic targets have been selected by using the near - infrared ( j , h and k bands ) photometric catalog , obtained by combining the sofi catalog for the inner @xmath1 2.5 arcmin , corresponding to the area covered by the gc ngc 1786 ( see @xcite for details ) and the two micro all sky survey ( 2mass ) catalog for the outermost region , in order to sample the surrouding field population .
the final catalog was placed onto the 2mass photometric and absolute astrometric system by following the standard procedure used in @xcite .
photometric uncertainties are of about 0.01 - 0.02 and 0.03 - 0.05 for sofi and 2mass target , respectively .
the targets have been selected along the rgb and in the magnitude range @xmath6 @xmath1 12.314.0 , in order to reach a sufficient snr ( @xmath7 30 - 40 ) and excluding stars brighter than the magnitude level ( @xmath812.3 , see * ? ? ?
* ) of the rgb tip of the intermediate - age lmc population .
observations consists of a series of 45-minute - long exposures obtained with hr11 ( 56005840 @xmath9 and r = 24200 ) and hr13 ( 61206400 @xmath9 and r = 22500 ) gratings by using the giraffe / medusa configurations .
the data reduction was performed with the giraffe eso pipeline that includes bias subtraction , flat fielding , wavelength calibration and spectrum extraction .
individual stellar spectra have been cleaned from the sky contribution , by subtracting the corresponding median sky spectrum .
finally , multiple spectra of each target have been coadded , reaching a snr per pixel of @xmath1 40 in the faintest stars and up to @xmath1 100 in the brightest ones .
identification number , right ascension and declination of each target are listed in table [ taba ] .
radial velocities ( @xmath10 ) have been measured by means of the daospec code @xcite . for each star , the spectra from the two different gratings
have been analysed independently and the derived @xmath10 averaged together by using the individual uncertainty as weight .
typical internal errors ( computed as @xmath11/@xmath12 ) are @xmath1 0.150.20 km s@xmath13 . finally , we applied the heliocentric correction to each @xmath10 .
the chemical analysis has been performed by using the suite of codes developed by r.l.kurucz ( see @xcite ) aimed to computing abundances from the observed equivalent widths ( ews ) and synthetic spectra .
the model atmospheres were computed with the atlas9 code , assuming plane - parallel geometry , local thermodynamical equilibrium for all species and without the inclusion of the approximate overshooting in the computation of the convective flux .
the atlas9 model atmospheres were calculated with the new set of opacity distribution functions by @xcite .
the line list has been selected starting from the kurucz / castelli data and updated with recent laboratory data from vald and nist databases .
we included only transitions with theoretical / laboratory atomic data and checked against spectral blendings through the inspection of suitable synthetic spectra convolved at the giraffe resolution .
in particular , we adopted for the fe lines the atomic data from the critical compilations by @xcite and @xcite , that represent the most updated dataset of theoretical / laboratory log gf for the iron lines . for some unblended transitions for which theoretical / laboratory oscillator strengths are not available ( or for those lines not well reproduced in the solar spectrum ) we derived astrophysical oscillator strengths ( labelled as sun in table [ tabb ] ) by using the solar flux spectra of @xcite and the model atmosphere for the sun computed by f. castelli adopting the solar abundances of @xcite .
we estimated for these oscillator strengths an accuracy @xmath3 15% , according to the uncertainties in the line profile fitting and in the continuum placement .
we decide to include in our final linelist some fe i transitions listed by @xcite and @xcite with high quoted uncertainties after the verification that these lines are well reproduced in the solar spectrum of @xcite and providing an iron abundance within @xmath14 0.1 dex from the value of @xcite . concerning the van der waals damping constants , we adopted , whenever possible , the damping values by @xcite , while for the other transitions the van der waals constants were calculated according to @xcite .
the complete line list used is available in table [ tabb ] , including wavelength , element code , oscillator strength ( and corresponding accuracy ) , lower excitational potential and reference source .
the atmospheric parameters have been derived as follows : _ ( 1 ) _
a preliminary estimate of @xmath15 for each star was obtained from photometric @xmath16 and @xmath17 colors adopting an average color excess of @xmath18 = 0.12 @xcite , the extinction law by @xcite and the color - temperature calibrations by @xcite based on the infra - red flux method .
the 2mass magnitudes of our targets have been converted in the tcs photometric system ( where the @xcite relations are defined ) by using the transformations provided by @xcite .
we used the photometric @xmath15 as _ first - guess values _
and then we refined them by imposing the lack of any trend between fe i abundances and the excitation potential @xmath19 .
basically , we find a good agreement between the photometric and spectroscopic @xmath15 , with an average difference of @xmath20-@xmath21= 139 k ( @xmath11 = 173 k ) . finally , the spectroscopic @xmath15 have been adopted for the following analysis .
typical uncertainties in the spectroscopic temperatures ( calculated according to the slope in the plane a(fe i ) vs @xmath19 ) ranging from @xmath180 up to @xmath1130 k. _ ( 2 ) _
microturbulent velocities have been derived by requiring the lack of any trend between fe i abundances and the reduced ew ( defined as @xmath22 ) .
typical uncertainty in the @xmath23 determination is 0.1 - 0.2 km / s .
_ ( 3 ) _
surface gravities have been estimated from the photometry , because the classical method to infer log g from the comparison between fe i and fe ii lines can not be applied , due to the small number of available fe ii lines in our spectra .
surface gravities have been derived from the stefan - boltzmann relation , using for each target the corresponding spectroscopic @xmath15 .
we adopted the distance modulus @xmath24 = 18.50 @xcite and a typical evolutionary mass ( 1.5 @xmath25 ) for all the targets derived from a basti @xcite isochrone with age of 2 gyr , z= 0.008 and solar - scaled chemical mixture , according to the values estimated for the bulk of the lmc disk population by @xcite and @xcite .
the luminosities are derived from the @xmath26-band magnitude using the bolometric corrections of @xcite .
internal errors in the derived log g ( calculated by taking into account the uncertainties in the spectroscopic @xmath15 , mass and luminosity ) are of about 0.1 - 0.15 dex .
the photometric estimates of temperatures and gravities together with those obtained from spectroscopy are listed in table [ tab1 ] .
ews were measured with daospec .
the abundances of fe and si have been obtained from the ews .
o abundances have been derived from the forbidden line at 6300.31 @xmath9 by employing spectral synthesis technique .
the abundances have been determined from @xmath27-minimization between the observed spectrum and a grid of synthetic spectra computed with the corresponding atmospheric parameters and convolved at the giraffe resolution .
mg abundances have been derived with the same technique by using the strong line 5711.09 @xmath9 ; we decide to exclude the mg triplet at @xmath1 6318 @xmath9 because these weak lines are located on the red wing of a close auto - ionization ca line .
this produce a decrease of the continuum level which was not easily measurable with daospec .
the final abundances are expressed in the usual spectroscopic formalism ( logarithmic scale in unit of solar abundances ) and adopting the solar values from @xcite except for oxygen , for which we used the value by @xcite .
all the abundance ratios of our sample are listed in table [ tab2 ] .
for each abundance ratio we take into account two source of uncertainties , the internal error relative to the ew measurements and that arising from the uncertainties in the atmospheric parameters .
uncertainties for each abundance ratio due to the uncertainties in the ew measurements have been estimated ( for each star ) by dividing the line - to - line scatter by the square root of the number of used lines .
daospec provides also for each measured spectral line an error computed from the standard deviation of the local flux residuals ( representing a 68% confidence interval on the derived value of the ew ) .
these uncertainties have been used only in the computation of the slopes in the a(fe i ) vs @xmath19 and a(fe i ) vs ewr planes ( see section 3 ) .
recently , @xcite pointed out that the internal ew errors provided by daospec are under - estimated of a factor of 2 with respect to those derived with the classical formula by @xcite , probably due to the pixel correlation during the wavelength calibration .
we checked this effect in our data , assuming the fwhm estimated by daospec and the proper snr for each spectrum . on average
, we found a general agreement between the two sets of uncertainties ( within @xmath14 10 - 20% ) but not a systematic effect .
however , the daospec errors provide an internal ranking among the measured lines in a given star and possible systematic under / over - estimates do not afflict our results . for o and mg , for which we derived the abundances from spectral synthesis of only one line each ,
the uncertainty in the measurement has been estimated by resorting to monte carlo simulations .
we study some stars as representative of the different snr of our targets .
we injected poisson noise into the best - fit synthetic spectrum according to the snr of the observed spectrum and we repeated the fit procedure ; for each line we performed a total of 500 monte carlo events , assuming as uncertainty in the fitting procedure 1@xmath11 of the abundances distribution .
for the o line ( falling in the hr13 grating ) , the derived uncertainties are of 0.04 , 0.03 and 0.02 dex at snr= 30 , 50 and 80 respectively , while for the mg line ( falling in the hr11 grating ) we obtained uncertainties of 0.13 , 0.09 and 0.04 dex for snr= 30 , 40 and 70 .
we estimated also the uncertainty due to the continuum location , by using the root mean square of the residual provided by daospec , after removing all the fitted lines from the observed spectrum .
for our stars the typical dispersion of the residual spectra is of @xmath14 3 - 4% , corresponding to an error in the abundances less than 0.1 dex . keeping in mind that daospec estimates a global continuum along the entire spectrum ( and not a local continuum for each individual transition )
, this uncertainty can be considered as a systematic errorbar for all the lines in a given spectrum .
the uncertainty arising from the atmospheric parameters has been computed following the approach described by @xcite .
the usual method to derive the errors due to the stellar parameters is to vary one only parameter each time , keeping the other ones fixed and finally adding in quadrature the derived abundance variations .
obviously , this method does not take into account the correlations among the atmospheric parameters , providing only a conservative value for the uncertainty . in our case , @xmath15 and @xmath23
have been optimized spectroscopically ( and log g has been derived according to the spectroscopic temperature ) , thus all the parameters are not independent each other . for each star ,
the temperature has been varied by @xmath14 1@xmath28 , because the uncertainty in @xmath15 dominates the derived abundances as pointed out by @xcite .
then , we repeated the optimization procedure described above by keeping the temperature fixed and deriving new values for log g and @xmath23 .
the advantage of this method is to naturally take into account the covariance terms among the parameters ( see also * ? ? ?
the procedure has been repeated independently for each star by considering the corresponding @xmath29 uncertainty computed from the error of the slope in the a(fe)@xmath19 plane .
table [ tab3 ] lists the results of this procedure for a representative star of our sample : the second column shows the final uncertainty for each abundance ratio , while the other columns are the results obtained with the usual approach to vary independently each parameter .
last column of table [ tab3 ] is the sum in quadrature of the uncertainties due to one only parameter : these values are larger than those obtained with the @xcite approach because the covariances among the parameters is neglected .
typical abundance uncertainties due to the atmospheric parameters are of the order of [ fe / h ] = @xmath14 0.03 dex , [ o / fe ] = @xmath14 0.04 dex , [ mg / fe ] = @xmath14 0.06 dex and [ si / fe ] = @xmath14 0.09 dex . finally , the total internal error for each abundance ratio was obtained by adding in quadrature the error associated to ew measurements and atmospheric parameters .
for [ fe / h ] it turns out to be of the order of @xmath1 0.04 - 0.05 dex because of the large number of measured lines .
the heliocentric radial velocity distribution of the stars in our sample is shown in the left panel of figure [ histogram ] .
stars with radial velocity in the range 170 @xmath30 @xmath10 @xmath30 380 km s@xmath13 are considered lmc members , according to @xcite : only 2 stars with @xmath31100 km s@xmath13 have been excluded ( likely belonging to the galaxy ) .
the mean velocity of the sample is @xmath10= 259.3 km s@xmath13
( @xmath32 = 33.9 km s@xmath13 ) , in good agreement with previous measurements in other samples of lmc disk by @xcite , @xcite and @xcite .
the metallicity distribution of our sample is shown in the right panel of figure [ histogram ] .
the entire sample has an average [ fe / h ] = 0.58 dex ( @xmath33}$ ] = 0.25 dex ) .
also , two main components can be distinguished : 1 . a principal component ( hereafter lmc - r )
comprising @xmath1 84@xmath2 of the observed sample with [ fe / h ] @xmath7 0.7 dex , peaked at [ fe / h]= 0.48 dex and with a quite small dispersion ( @xmath33}$ ] = 0.13 dex ) .
2 . a secondary component ( hereafter lmc - p ) ,
comprising @xmath1 16@xmath2 of the sample , peaked at [ fe / h ] = 1.06 dex ( @xmath33}$ ] = 0.18 dex ) with an extended tail up to [ fe / h ] @xmath1 1.5 dex .
the peak of the metallicity distribution of lmc - r ( see fig . [ mult_hist ] , a ) is in good agreement with those obtained by @xcite ( [ fe / h ] = 0.45 dex , @xmath33}$ ] = 0.31 dex ) ( see fig . [ mult_hist ] , b ) from the central bar and @xcite ( [ fe / h ] = 0.47 dex ( @xmath33}$ ] = 0.30 dex ) , [ fe / h ] = 0.50 dex ( @xmath33}$ ] = 0.44 dex ) and [ fe / h ] = 0.45 dex ( @xmath33}$ ] = 0.31 dex ) for the fields located at @xmath1 3@xmath5 , 5@xmath5 and 6@xmath5 at north of lmc center , respectively ) .
a distribution peaked at a slightly lower value ( [ fe / h ] = 0.75 dex , @xmath33}$ ] = 0.23 dex ) was found by @xcite ( see fig .
[ mult_hist ] ) .
figure [ omgsi ] shows the behaviour of o , mg and si , as a function of [ fe / h ] .
we plotted for comparison the abundance ratios measured in other environments : the galaxy ( including data for thin disk , thick disk and halo , by * ? ? ?
* ; * ? ? ?
* ) , the lmc disk @xcite , the nearby dwarf spheroidals @xcite , the sagittarius dwarf galaxy @xcite and the old and intermediate - age lmc gcs @xcite1 - 2 gyr . ] .
the [ o / fe ] ratio appears systematically lower than those measured in the galaxy and basically consistent with those by @xcite , even if they provide a few of measures for this abundance ratio .
the [ mg / fe ] ratio turns out to be subsolar in the entire metallicity range and ever lower than the milky way stars ; also , our targets show lower mg abundances with respect to those measured by @xcite and we attribute such a discrepancy to the different values of oscillator strengths and damping constants for the mg line at 5711 @xmath9 employing in the two analysis . finally , the [ si / fe ] ratio seems to be barely consistent with the galactic stars , despite a larger dispersion .
figure [ alpha ] shows the behaviour of the average [ @xmath0/fe ] ratio , obtained by averaging the abundances of o , mg and si , as a function of [ fe / h ] .
we compared our targets with the [ @xmath0/fe ] ratios of the other samples calculated by averaging the available @xmath0-element abundance ratios ; only for the old lmc gcs we excluded o and mg , because of the intrinsic star - to - star variations of these elements due to the self - enrichment process occurred in the early stages of these clusters ( at variance to the intermediate - age lmc gcs where intrinsic spread in o and mg content are not detected ) .
the overall trend of the [ @xmath0/fe ] ratio with the [ fe / h ] abundance shows a decrease at increasing metallicity . the most metal - poor stars with [ fe / h ]
@xmath3 0.7 dex show an [ @xmath0/fe ] ratios larger than solar value , while stars with [ fe / h ] @xmath7 0.7 dex show [ @xmath0/fe ] ratio from solar to sub - solar values . for a given [ fe / h
] , the [ @xmath0/fe ] ratio measured in the lmc appears systematically lower than those measured in the galaxy .
a substantial agreement of our abundances was generally found with the values measured in dsphs , in particular with the abundance ratios observed in the metal - poor component of the sagittarius dwarf galaxy @xcite : such a similarity of @xmath0-elements , metallicity distribution and fraction of metal - poor stars has been already suggested by @xcite .
we determined kinematic and chemical properties for 89 giant stars members of the disk of the lmc .
this sample increases significantly the number of stars analysed so far through high - resolution spectroscopy ( the largest sample observed by @xcite includes 59 stars ) .
the derived metallicity distribution is dominated by lmc - r , a metal - rich and narrow component peaked at [ fe / h ] = 0.48 dex , with a secondary component lmc - p peaked at [ fe / h ] = @xmath1 1.0 dex and reaching [ fe / h ] = @xmath1 1.5 dex ( right panel of fig .
[ histogram ] ) .
as shown in figure [ feh_vrad ] , the two components show similar kinematic properties , with similar average @xmath34 ( 265.9 km s@xmath13 for lmc - r and 261 km s@xmath13 for lmc - p ) and with a small increase of the velocity dispersion decreasing the metallicity ( @xmath32 = 25 km s@xmath13 and @xmath32 = 29.7 km s@xmath13 , respectively ) .
the kinematic and abundance distributions can be now compared to those observed in the stellar population of the gc ngc 1786 .
all the stars of our sample belong to the rgb , hence their age can range from @xmath1 1 - 2 gyr up to @xmath1 12 - 13 gyr , thus excluding very recent ( last 500 myr ) burst of star formation . at variance to the stars belonging to a stellar cluster ,
the determination of the age for field stars through isochrone fitting is a dangerous and uncertain technique , because the position of a rgb star in the color - magnitude diagram is weakly sensitive to the age and highly affected by uncertainties in the color excess , evolutive mass and distance .
even if precise ages for each target star can not be determined , we can draw the timeline of the chemical evolution in this region of the lmc , by comparing our results with the recent simulations by @xcite and the sfh inferred by @xcite , @xcite and @xcite in different regions of the lmc ; unfortunately , there are no determinations of sfh in the region of our targets .
( 1 ) ngc 1786 is an old lmc gc generated during the first burst of sf occurred @xmath1 1213 gyr ago .
following @xcite a loose stellar halo of old stars with a velocity dispersion of about @xmath1 40 km s@xmath13 and a broad metallicity distribution is also expected to be formed during that event .
it is interesting to check whether some of the stars observed in our sample could belong to ngc 1786 or to the lmc halo .
note that @xcite found for ngc 1786 @xmath35 = 264.3 km s@xmath13 ( @xmath32 = 5.7 km s@xmath13 ) , [ fe / h ] = 1.75 dex and [ @xmath0/fe ] = 0.37 dex .
figure [ feh_vrad ] shows that none of the observed stars has kinematic and chemical properties compatible with the population of ngc 1786 .
in particular , the observed sample has a systematically higher [ fe / h ] ( @xmath36 1.0 dex ) and a lower [ @xmath0/fe ] ( @xmath4 0.20 dex ) , thus demonstrating that even lmc - p did not originated in the cluster .
this evidence could be used to put some constraints on the presence of old metal - poor @xmath0-enhanced stars in the sampled lmc disk field , which turns out to be less then @xmath1 1@xmath2 .
our outcome provides a substantial confirmation of the results by @xcite , who found @xmath1 3@xmath2 of their sample with [ fe / h ] @xmath3 1.5 dex ( but no measurements of [ @xmath0/fe ] ratio are available ) and by @xcite , who found @xmath1 2@xmath2 , although the abundance patterns are incompatible with old lmc gcs .
since these samples cover different regions of the galaxy , the measured similar fraction of metal - poor stars in these three samples seems to suggest that the old , metal - poor component is distributed in a quite homogeneous way along the lmc disk .
( 2 ) after the initial burst , the lmc has undergone a long period characterized by a continuous sf with low efficiency .
the occurence of this quiescent period in the lmc field is clearly visible in the sfh provided by @xcite , @xcite and @xcite , where a prolonged phase with very low or lacking sf is appreciable between @xmath1 4 - 5 gyr and @xmath1 12 gyr ago .
this quiescent period is predicted also by the theoretical models proposed by @xcite that derived a typical sf efficiency in this age range of @xmath1 0.1 m@xmath37 yr@xmath13 . during this period ,
the lmc evolved in isolation , without gravitational interactions with the galaxy and the smc @xcite .
lmc - p likely formed during this period , as also suggested by the lower ( by @xmath1 0.20.3 dex ) [ @xmath0/fe ] ratio with respect to the galactic values at similar metallicity , which indicate that during this long period mainly sneia contribute to the gas enrichment .
if this scenario is correct , these stars were born during the period in which no gc formed ( the so - called _ age gap _ , @xcite ) , hence they are unique tracers of the lmc chemical evolution history between @xmath1 12 and 3 gyr ago .
we also notice that the lmc - p fraction ( @xmath1 16@xmath2 ) is larger than that ( @xmath4 10@xmath2 ) found by @xcite .
this is probably due to different location of the two fields .
the sample analysed by @xcite is located around the central bar ( at variance with our targets , located in the inner disk at @xmath1 1.8 kpc from the lmc center ) .
infact , @xcite suggest that the central bar formed in the last 34 gyr with a marginal fraction of metal - poor stars in its stellar content .
also , it is interesting to notice that the [ @xmath0/fe ] ratio of this stellar component well resembles the mean locus defined by the dsph stars ( grey triangles in the lower panel of fig .
[ alpha ] , but see also fig.11 in @xcite ) , suggesting similar chemical enrichment histories .
( 3 ) lmc - r formed during the relevant burst of sf occurred in the last few gyr probably due to the tidal capture of the smc by the lmc .
numerical simulations @xcite predict that first close encounter between the two clouds occurred @xmath1 34 gyr ago , when the lmc and smc become a gravitational bound system .
this strong interaction triggered the sfr up to @xmath1 0.4 m@xmath37 yr@xmath13 .
the occurrence of this sf enhancement has been confirmed by @xcite by using color - magnitude diagrams of different regions of the lmc .
however , the binary system status of magellanic clouds and its timescales is still a matter of debate since there is no unambiguous consensus between dynamical simulation and photometric studies .
for instance , @xcite analysed new photometric dataset from the vista magellanic survey , finding an enhancement of star formation with a peak at @xmath1 2 gyr and in some regions a second peak at @xmath1 5 gyr and interpreted as the epoch of tidal interaction between the lmc and the galaxy .
conversely , @xcite suggested that the lmc - smc system is on its first close passage about the galaxy , being entered in the milky way virial radius only in the last 1 - 3 gyr .
lmc - r distribution is very narrow ( @xmath33}$]= 0.13 dex ) , suggesting that the episode of sf associated to the first close encounter between the two clouds has been very efficient ( in order to produce the majority of the lmc disk stars ) and fast , because the stars have not had time to furtherly enrich in iron .
these stars show [ @xmath0/fe ] abundance ratios that are close to the solar value , in nice agreement with the values measured in the intermediate - age gcs @xcite with age of 13 gyr , as shown in figure [ alpha ] .
the observed [ @xmath0/fe ] ratios agree with those derived in the sagittarius dsph giants ( asterisks in lower panel of fig .
[ alpha ] ) , while the nearby dsphs studied so far do not reach such metallicity . the most metal - rich stars discussed by @xcite in fornax show [ fe / h]@xmath10.6 dex , consistent with the metal - poor edge of the lmc - r population .
these evidences ( the similarity with sgr and the difference with the nearby dsphs ) seem to confirm that the recent sfh of the lmc has been quite complex and turbulent , and the stars of the lmc - r population ( that dominates the metallicity distribution of our sample ) are the products of the tidal interactions occurring among lmc , smc and our galaxy .
the intrinsic dispersion of each individual abundance ratio of the lmc - r population suggests a high degree of homogeneity , pointing out that an efficient and homogeneous mixing of the sn ejecta occurs .
we note that the dispersion of the [ @xmath0/fe ] ratios in these stars is larger than the range of values covered by the intermediate - age gcs .
thus , the scatter in [ @xmath0/fe ] abundances of the lmc - r stars reflects the larger age range in which these stars formed .
however , we find a general agree between the values observed in the field and gc stars in this metallicity range , suggesting in any case a common origin of these stars from similar burst of sf .
it is important to recall that there is an offset between the age of the onset of the sf in the lmc field and that of the gc formation .
in fact , if the sf in the field started about 4 - 5 gyr ago ( the precise age depends on the location in the disk and the magnitude of the tidal interactions with the smc ) , the gc formation restarts about 2 gyr ago , as suggested also by @xcite .
this difference is confirmed by the measured ages in the intermediate - ages gcs ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and the lack of globulars in the age range between 2 and 5 gyr .
furthermore , the lower metallicity edge of lmc - r stars ( [ fe / h ] @xmath38 0.7 dex ) , if compared to that of gcs , seems to support that the sf of field stars becomes efficient slightly before .
recently , @xcite discussed the unusual low [ mg / fe ] ( 0.9 dex ) ratio measured by @xcite in the intermediate - age , metal - rich gc ngc 1718 , suggesting that this cluster and a fraction of lmc field stars were formed from the metal - poor gas acquired by infall from the smc about 1 - 2 gyr ago ( the so - called _ magellanic squall _ , @xcite ) . in our sample
we did not detect metal - rich stars characterized by such low mg or @xmath0-depletion , pointing out that the stars formed from metal - poor gas accreted by the smc are ( if any ) a negligible ( @xmath31@xmath2 ) fraction of the lmc disk .
rrrccc & & & & & + 353 & 74.8429031 & -67.7529373 & 14.55 & 13.80 & 13.61 + 1415 & 74.7569885 & -67.7277374 & 14.52 & 13.75 & 13.57 + 1593 & 74.7403107 & -67.7479858 & 14.34 & 13.56 & 13.36 + 1954 & 74.6953583 & -67.7596817 & 14.11 & 13.28 & 13.12 + 1980 & 74.6899948 & -67.7497635 & 14.52 & 13.74 & 13.58 + 2353 & 74.7275314 & -67.7423553 & 12.56 & 11.87 & 11.71 + 2363 & 74.7172394 & -67.7453918 & 14.67 & 13.97 & 13.82 + 2376 & 74.7086487 & -67.7528458 & 14.62 & 13.93 & 13.81 + 90012 & 74.2391357 & -67.7102280 & 13.91 & 13.05 & 12.93 + 90119 & 74.5899658 & -67.9048157 & 14.27 & 13.38 & 13.23 + [ taba ] cccccc 5586.756 & fei & -0.144 & @xmath310% & 3.37 & @xcite + 5607.664 & fei & -2.270 & @xmath350% & 4.15 & @xcite + 5611.356 & fei & -2.990 & @xmath350% & 3.64 & @xcite + 5618.632 & fei & -1.276 & @xmath318% & 4.21 & @xcite + 5619.225 & fei & -3.270 & @xmath350% & 3.69 & @xcite + 5619.595 & fei & -1.670 & @xmath750% & 4.39 & @xcite + 5624.022 & fei & -1.140 & @xmath315% & 4.39 & sun + 5633.946 & fei & -0.320 & @xmath750% & 4.99 & @xcite + 5636.696 & fei & -2.560 & @xmath750% & 3.64 & @xcite + 5638.262 & fei & -0.840 & @xmath350% & 4.22 & @xcite + 5640.307 & fei & -1.700 & @xmath315% & 4.64 & sun + 5646.684 & fei & -2.500 & @xmath350% & 4.26 & @xcite + 5650.690 & fei & -0.960 & @xmath750% & 5.09 & @xcite + 5651.469 & fei & -2.000 & @xmath350% & 4.47 & @xcite + 5652.318 & fei & -1.920 & @xmath750% & 4.26 & @xcite + 5653.867 & fei & -1.610 & @xmath750% & 4.39 & @xcite + 5661.021 & fei & -2.430 & @xmath350% & 4.58 & @xcite + 5661.345 & fei & -1.756 & @xmath310% & 4.28 & @xcite + 5662.516 & fei & -0.573 & @xmath310% & 4.18 & @xcite + 5677.684 & fei & -2.700 & @xmath350% & 4.10 & @xcite + 5678.601 & fei & -4.670 & @xmath350% & 2.42 & @xcite + 5680.240 & fei & -2.540 & @xmath750% & 4.19 & @xcite + 5691.497 & fei & -1.490 & @xmath750% & 4.30 & @xcite + 5693.640 & fei & -0.680 & @xmath315% & 4.96 & sun + 5704.733 & fei & -1.250 & @xmath315% & 5.03 & sun + 5705.465 & fei & -1.355 & @xmath310% & 4.30 & @xcite + 5714.551 & fei & -1.770 & @xmath315% & 5.09 & sun + 5720.886 & fei & -1.950 & @xmath350% & 4.55 & @xcite + 5731.762 & fei & -1.270 & @xmath350% & 4.26 & @xcite + 5732.296 & fei & -1.560 & @xmath325% & 4.99 & @xcite + 5741.848 & fei & -1.670 & @xmath325% & 4.26 & @xcite + 5760.345 & fei & -2.440 & @xmath750% & 3.64 & @xcite + 5767.972 & fei & -3.200 & @xmath315% & 4.29 & sun + 5775.081 & fei & -1.298 & @xmath318% & 4.22 & @xcite + 5776.224 & fei & -3.400 & @xmath315% & 3.69 & sun + 5778.453 & fei & -3.430 & @xmath325% & 2.59 & @xcite + 5793.915 & fei & -1.660 & @xmath750% & 4.22 & @xcite + 5805.757 & fei & -1.590 & @xmath325% & 5.03 & @xcite + 5806.725 & fei & -1.030 & @xmath350% & 4.61 & @xcite + 5811.915 & fei & -2.430 & @xmath350% & 4.14 & @xcite + 6120.249 & fei & -5.970 & @xmath37% & 0.91 & @xcite + 6151.618 & fei & -3.299 & @xmath310% & 2.18 & @xcite + 6165.360 & fei & -1.474 & @xmath318% & 4.14 & @xcite + 6187.989 & fei & -1.670 & @xmath750% & 3.94 & @xcite + 6200.313 & fei & -2.437 & @xmath318% & 2.61 & @xcite + 6226.734 & fei & -2.220 & @xmath350% & 3.88 & @xcite + 6246.319 & fei & -0.877 & @xmath310% & 3.60 & @xcite + 6252.555 & fei & -1.687 & @xmath310% & 2.40 & @xcite + 6322.686 & fei & -2.426 & @xmath318% & 2.59 & @xcite + 6330.849 & fei & -1.190 & @xmath315% & 4.73 & sun + 6335.330 & fei & -2.177 & @xmath318% & 2.20 & @xcite + 6336.824 & fei & -0.856 & @xmath310% & 3.69 & @xcite + 6380.743 & fei & -1.376 & @xmath325% & 4.19 & @xcite + 6300.304 & oi & -9.717 & @xmath32% & 0.00 & @xcite + 5711.095 & mgi & -1.724 & @xmath310% & 4.35 & nist + 5665.555 & sii & -2.040 & 20% & 4.92 & @xcite + 5666.677 & sii & -1.710 & @xmath315% & 5.62 & sun + 5690.425 & sii & -1.870 & 20% & 4.93 & @xcite + 5701.104 & sii & -2.050 & 20% & 4.93 & @xcite + 5793.073 & sii & -2.060 & 20% & 4.93 & @xcite + 6155.144 & sii & -0.870 & @xmath315% & 5.62 & sun + 6237.319 & sii & -1.100 & @xmath315% & 5.61 & sun + [ tabb ] rcccccrc & & & & & & & + 353 & 4074 & 1.12 & 269.7 & 4150 & 1.10 & 0.50 & 1.50 + 1415 & 4043 & 1.10 & 223.6 & 4200 & 1.10 & 0.50 & 1.80 + 1593 & 3989 & 1.01 & 254.2 & 4150 & 1.10 & 0.50 & 1.60 + 1954 & 3922 & 0.91 & 223.3 & 3850 & 0.90 & 0.00 & 1.50 + 1980 & 4018 & 1.10 & 272.3 & 4100 & 1.10 & 0.50 & 1.60 + 2353 & 4255 & 0.39 & 299.4 & 4350 & 0.40 & 0.50 & 1.80 + 2363 & 4227 & 1.23 & 265.1 & 4350 & 1.30 & 0.50 & 1.70 + 2376 & 4308 & 1.24 & 230.8 & 4600 & 1.30 & 1.00 & 2.10 + 90012 & 3906 & 0.83 & 243.5 & 4000 & 0.80 & 1.00 & 2.00 + 90119 &
3795 & 0.93 & 251.1 & 4450 & 1.20 & 0.50 & 2.10 + [ tab1 ] rrrrr & & & & + 353 & 0.42 @xmath14 0.03 & 0.01 @xmath14 0.05 & 0.16 @xmath14 0.08 & 0.07 @xmath14 0.06 + 1415 & 0.62 @xmath14 0.03 & 0.19 @xmath14 0.04 & 0.21 @xmath14 0.02 & 0.16 @xmath14 0.03 + 1593 & 0.58 @xmath14 0.02 & 0.04 @xmath14 0.06 & 0.06 @xmath14 0.09 & 0.14 @xmath14 0.05 + 1954 & 0.45 @xmath14 0.03 & 0.18 @xmath14 0.09 & 0.13 @xmath14 0.10 & 0.17 @xmath14 0.05 + 1980 & 0.46 @xmath14 0.02 & 0.06 @xmath14 0.04 & 0.22 @xmath14 0.10 & 0.13 @xmath14 0.02 + 2353 & 0.37 @xmath14 0.02 & 0.38 @xmath14 0.08 & 0.16 @xmath14 0.09 & 0.09 @xmath14 0.03 + 2363 & 0.50 @xmath14 0.03 & 0.03 @xmath14 0.07 & 0.12 @xmath14 0.10 & 0.03 @xmath14 0.04 + 2376 & 1.08 @xmath14 0.03 & 0.28 @xmath14 0.05 & 0.14 @xmath14 0.11 & 0.18 @xmath14 0.06 + 90012 & 1.24 @xmath14 0.02 & 0.38 @xmath14 0.09 & 0.02 @xmath14 0.07 & 0.37 @xmath14 0.07 + 90119 & 0.64 @xmath14 0.02 & 0.24 @xmath14 0.09 & 0.11 @xmath14 0.13 & 0.04 @xmath14 0.05 + [ tab2 ] lccccc fe & @xmath140.01 & @xmath140.03 & @xmath140.02 & @xmath140.04 & @xmath140.05 + o & @xmath140.02 & @xmath140.01 & @xmath140.05 & @xmath140.02 & @xmath140.05 + mg & @xmath140.04 & @xmath140.03 & @xmath140.01 & @xmath140.02 & @xmath140.04 + si & @xmath140.06 & @xmath140.07 & @xmath140.03 & @xmath140.02 & @xmath140.08 + [ tab3 ] | we have used high - resolution spectra obtained with the multifiber facility flames at the very large telescope of the european southern observatory to derive kinematic properties and chemical abundances of fe , o , mg and si for 89 stars in the disk of the large magellanic cloud ( lmc ) .
the derived metallicity and [ @xmath0/fe ] , obtained as the average of o , mg and si abundances , allow us to draw a preliminary scheme of the star formation history occurred in this region of the lmc .
the derived metallicity distribution shows two main components : one component ( comprising @xmath1 84@xmath2 of the sample ) is peaked at [ fe / h ] = 0.48 dex and it shows an [ @xmath0/fe ] ratio slightly under solar ( [ @xmath0/fe ] @xmath1 0.1 dex ) .
this population was probably originated by the main star formation event occurred 34 gyr ago ( possibly triggered by tidal capture of the small magellanic cloud ) . the other component ( comprising @xmath1 16@xmath2 of the sample )
is peaked at [ fe / h ] @xmath1 1 dex and it shows an [ @xmath0/fe ] @xmath1 0.2 dex .
this population was probably generated during the long quiescent epoch of star formation in between the first episode and the most recent bursts .
indeed , in our sample we do not find stars with chemical properties similar to the old lmc globular clusters nor to the iron - rich and @xmath0-poor stars recently found in the lmc globular cluster ngc 1718 and predicted to be also in the lmc field , thus suggesting that both these components are small ( @xmath3 1@xmath2 ) in the lmc disk population . |
nonlinear optical phenomena arise from the coupling between two or more photons , mediated by their interaction with matter , to produce a photon with a frequency that is a linear combination of the original photon frequencies .
these processes are responsible for many significant advances in laser - based technologies @xcite , most of which rely on phase - matching of intense electromagnetic fields in extended bulk crystalline media to enable efficient frequency conversion .
now , as the accessibility of nanostructured materials offered by modern nanofabrication techniques continues to increase , so does the interest in mastering nonlinear optics on subwavelength scales .
indeed , nonlinear optical nanomaterials find diverse applications , including optical microscopy @xcite , biological imaging / detection @xcite , and signal conversion in nanoscale photonic devices @xcite .
inherently , nanostructured materials possess small volumes that limit their interaction with optical fields .
fortunately , this can be compensated by high oscillator strengths provided by either quantum confinement effects @xcite or the intense near - fields generated by localized surface plasmons @xcite . in particular , the near - field enhancement associated with plasmons in noble metal nanostructures has been widely used with the purpose of enhancing the nonlinear response of surrounding dielectric materials @xcite .
plasmons also display strong intrinsic optical nonlinearities , clearly observed in metal nanoparticles @xcite .
the plasmonic response of a metal nanostructure can be tailored by its size , shape , and surrounding environment @xcite , as well as by combining two or more interacting nanostructures . in this manner ,
a metallic nanocomposite can be engineered to possess multiple resonance frequencies , enabling the optimization of nonlinear wave mixing between multifrequency optical fields @xcite .
unfortunately , plasmons are hardly tunable in metals after a structure has been fabricated @xcite , thus severely limiting the choice of frequency combinations . as an alternative plasmonic material
, electrically doped graphene has been found to support long - lived plasmonic excitations that efficiently couple to light and are actively tunable by changing the density of charge carriers @xcite .
strong intrinsic nonlinearities have also been observed in this material @xcite , which could be further enhanced by plasmons @xcite .
recently , plasmon - assisted second - harmonic generation ( shg ) and down conversion with good efficiencies at the few - photon level have been shown to be possible when the fundamental and the second - harmonic are simultaneously resonant with plasmons in a graphene nanoisland @xcite .
we have predicted that a similar mechanism can lead to unprecedentedly intense shg and third - harmonic generation ( thg ) in nanographene @xcite .
although experimental studies have so far demonstrated strong plasmons at mid - infrared and terahertz frequencies @xcite , further reduction in the size of the islands down to less than 10 nm should allow us to reach the visible and near - infrared ( vis - nir ) regimes @xcite .
in particular , commercially - available polycyclic aromatic molecules sustain plasmon - like resonances that are switched on and off by changing their charge state @xcite .
additionally , recent progress in the chemical synthesis of nanographene @xcite provides further stimulus for the use of this material to produce electrically tunable vis - nir plasmons , as well as their application to nonlinear optics at the nanoscale . in this work ,
we investigate nonlinear optical wave mixing in doped nanographene .
specifically , we propose a scheme for optimizing wave mixing among multifrequency optical fields that utilizes the plasmons supported by a doped graphene nanoisland .
using a tight - binding description for the electronic structure combined with density - matrix quantum - mechanical simulations , we demonstrate that efficient wave mixing is achieved in an island when the incident and/or the mixed frequencies are coupled with one or more of its plasmons . by actively tuning the doping level in a graphene nanoisland of fixed geometry , a wide range of plasmon - enhanced input / output mixing frequency combinations can be realized .
we study wave mixing of coincident light pulses in fig .
[ fig1 ] for a triangular armchair - edged graphene nanoisland containing @xmath0 carbon atoms ( side length pf @xmath1 nm ) and doped with three electrons ( doping density @xmath2 , corresponding to an equivalent extended graphene fermi energy @xmath3ev ) . additionally , we assume a conservative inelastic lifetime @xmath4fs (
i.e. , linewidth @xmath5mev ) . from the linear absorption cross - section of the nanoisland , we identify several prominent plasmons for incident photon energies .
in particular , we concentrate on @xmath6ev and @xmath7ev .
the first of these plasmons is highly tunable upon electrical doping , as it is switched on when moving from @xmath8 to @xmath9 charge states ( see lower panel in fig .
[ fig1]b ) , and its frequency increases with @xmath10 ( see fig .
[ figs1 ] in the appendix ) .
the plasmon at @xmath11 is comparatively less tunable .
plasmon - enhanced optical wave mixing in the nanoisland is demonstrated in fig .
[ fig1]b , where the spectral decomposition of the induced dipole moment is shown for excitation by collinear gaussian pulses with central frequencies @xmath12 and @xmath11 .
we show the response for incident polarization aligned with an edge of the nanoisland ( upper panels ) or perpendicular to an edge ( middle panels ) , and in each case the dipole is calculated along the direction of polarization .
further , in fig . [ fig1]b we superimpose the spectra obtained from coincident pulses with the spectra for excitation of the nanoisland by each of the individual pulses in isolation , which exhibit polarization features that oscillate at harmonics of the fundamental frequencies ( @xmath13 and @xmath14 ) .
the dual - pulse spectrum exhibits these features in addition to polarization components produced via sum and difference frequency generation ( @xmath15 ) and degenerate four - wave mixing ( @xmath16 or @xmath17 ) , along with various other combinations of harmonic generation and wave mixing .
we note that for polarization along the edge of a nanotriangle , inversion symmetry in this direction prevents even - ordered nonlinear processes from occurring , such as shg and sum / difference frequency generation .
incidentally , mixing of less tunable plasmons ( e.g. , @xmath18 with the resonance at @xmath19ev ) produces less intense nonlinear features ( see fig .
[ figs2 ] in the appendix ) . while the response of a graphene nanoisland to ultrashort pulses can provide information on the relative strengths of the nonlinear processes for excitation at specific frequencies , a quantitative analysis of each nonlinear process , along with its optimal plasmonic enhancement , is best provided by studying the response under continuous - wave ( cw ) illumination . in fig .
[ fig2 ] we consider sum - frequency generation ( sfg ) in a nanoisland containing @xmath20 atoms ( side length of @xmath21 nm ) , for doping with either three ( fig .
[ fig2]a ) or six ( fig . [ fig2]b ) additional charge carriers . the upper panels in fig .
[ fig2 ] show the linear response of the nanoisland for the two doping levels considered , which show multiple plasmon resonance peaks at low doping that converge to a single , stronger feature as the doping level increases .
the sfg polarizability @xmath22 is presented as a function of the two applied field frequencies , enabling exploration of all possible frequency combinations that may result in plasmon - enhanced wave mixing
. the input frequencies at which sfg is enhanced are found to coincide with the plasmons , corresponding to the horizontally and vertically aligned features in the contour plot of @xmath23 in fig .
strong enhancement is observed where these features intersect , driven by plasmons excited at both fundamental frequencies .
prominent features also follow frequencies satisfying @xmath24 , where @xmath25 denotes one of the plasmon frequencies of the nanoisland , indicating plasmonic enhancement at the output ( sum ) frequency . finally ,
since shg can be considered as a special case of sfg , we also note plasmonic enhancement of shg for frequencies satisfying @xmath26 .
similar conclusions are drawn from frequency - difference generation ( i.e. , @xmath27 ) , where we observe enhancement when either one of the incident frequencies or the output frequency resonates with a plasmon ( see fig .
[ figs3 ] in the appendix ) .
plasmonic enhancement of four - wave mixing is investigated in fig .
[ fig3 ] for the same graphene nanoisland and doping conditions considered for fig .
[ fig2 ] , with an output frequency @xmath28 .
we find a similar enhancement in @xmath29 when either of the fundamental frequencies of the incident fields are resonant with plasmons in the nanoisland , although in this case the enhancement favors @xmath30 , as @xmath12 is involved twice in these specific wave mixing processes . here
we find increased four - wave mixing following the frequencies @xmath31 , once again indicating enhancement at the output frequency .
for @xmath32 we have fully - degenerate four - wave mixing , where we are actually investigating the third - order polarizability @xmath33 contributing to the linear response ( i.e. , the kerr effect ) . in this case , frequencies satisfying the condition @xmath26 produce extremely high polarizabilities , as the three input frequencies and the output frequency are all equally amplified by the same plasmon resonance .
we have also examined four - wave mixing with output frequency @xmath34 , leading to similar conclusions on plasmon - assisted enhancement ( see fig .
[ figs4 ] in the appendix ) . as noble metal nanoparticles possess the highest nonlinear polarizabilities per atom measured experimentally @xcite , we compare them with the above results for doped nanographene nonlinear polarizabilities . in particular , we contrast data available in the literature for metal third - order polarizabilities @xmath35 with those calculated here for graphene . in fig . [
fig4]a we show the maximum polarizability @xmath36 for graphene nanoislands of increasing size , and for a fixed doping ratio of one electron per every hundred carbon atoms , along with experimental data obtained from degenerate four - wave mixing experiments on gold , silver , and copper nanoparticles of comparable or greater sizes .
even through metal particles have much larger volumes than graphene nanoislands for a given lateral size @xmath37 , we find the latter to exhibit much larger nonlinear response , reaching two orders of magnitude higher values .
obviously , we are comparing theory for graphene with experiments for metals , so this is a preliminary conclusion , which should be examined experimentally .
nonetheless , we are using a conservative inelastic lifetime for nanographene in our calculations ( @xmath4fs ) , whereas much longer lifetimes could be actually encountered in practice , based upon recently measured plasmon resonances @xcite , and it should be noted that the nonlinear polarizability of order @xmath38 scales roughly as @xmath39 @xcite .
the linear absorption spectra for the nanoislands that we compare with metal nanoparticles are presented in fig .
[ fig4]b , illustrating how the plasmon resonances vary as the nanoisland size increases while a fixed doping density is maintained .
in conclusion , graphene nanoislands can be used to realize nonlinear wave mixing on the nanoscale with extraordinarily high efficiencies , surpassing those of metal nanoparticles of similar lateral sizes .
the large magnitudes of the predicted nonlinear polarizabilities are attributed to plasmonic enhancement in these nanostructures .
wave mixing is further enhanced by simultaneously exploiting multiple plasmonic resonances , enabling two incident fields with distinct frequencies to independently couple to plasmons , and additionally tuning the mixed frequency to yet another plasmon .
these plasmons can be tuned by changing the number of doping charge carriers , adding another advantage with respect to conventional plasmonic metal response .
interestingly , by considering islands formed by a few hundreds or thousands of carbon atoms , we predict tunable plasmonic response and plasmon - induced enhanced nonlinearities within the visible and near - infrared spectral ranges .
our results configure a new platform for the development of nanoscale nonlinear optical devices based upon graphene nanostructures with lateral dimensions of only a few nanometers , such as those that are currently produced by chemical synthesis @xcite .
we describe the low - energy ( @xmath40ev ) optical response of graphene nanoislands within a density - matrix approach , using a tight - binding model for the @xmath41-band electronic structure @xcite .
one - electron states @xmath42 are obtained by assuming a single @xmath43 orbital per carbon site , oriented perpendicular to the graphene plane , with a hopping energy of @xmath44ev between nearest neighbors . in the spirit of the mean - field approximation @xcite ,
a single - particle density matrix is constructed as @xmath45 , where @xmath46 are time - dependent complex numbers .
an incoherent fermi - dirac distribution of occupation fractions @xmath47 is assumed in the unperturbed state , characterized by a density matrix @xmath48 , whereas the time evolution under external illumination is governed by the equation of motion @xmath49 - \frac{1}{2\tau } \left ( \rho - \rho^{0 } \right).\ ] ] the last term of eq .
[ rho_eom ] describes inelastic losses at a phenomenological decay rate @xmath50 .
we set @xmath5mev throughout this work ( i.e. , @xmath4fs ) , corresponding to a conservative drude - model graphene mobility @xmath51@xmath52v@xmath53s@xmath53 for a characteristic doping carrier density @xmath54@xmath55 ( i.e. , one charge carrier per every 100 carbon atoms ) .
the system hamiltonian @xmath56 consists of the tight - binding part @xmath57 ( i.e. , nearest - neighbors hopping ) and the interaction with the self - consistent electric potential @xmath58 , which is in turn the sum of external and induced potentials .
the latter is simply taken as the hartree potential produced by the perturbed electron density , while the former reduces to @xmath59 for an incident electric field @xmath60 .
the induced dipole moment is then calculated from the diagonal elements of the density matrix in the carbon - site representation as @xmath61\rb_l$ ] , where the factor of 2 accounts for spin degeneracy and @xmath62 runs over carbon sites .
we use two different methods to solve eq .
[ rho_eom ] and find @xmath63 ( direct time - domain numerical integration and a perturbative approach @xcite ) , which we find in excellent mutual agreement under low - intensity cw illumination .
direct time integration allows us to simulate the response to short light pulses for arbitrarily large intensity , while the perturbative method yields the nonlinear polarizabilities under multifrequency cw illumination ( @xmath64 , with the same amplitude @xmath65 at both frequencies for simplicity ) , for which we can express the dipole moment as a power series in the electric field strength @xmath65 according to @xmath66 here , @xmath38 is the scattering order , whereas @xmath67 and @xmath68 give the harmonic orders of the incident frequencies @xmath12 and @xmath11 .
we denote polarizabilities according to @xmath69 , where @xmath70 is the frequency generated by a particular @xmath71-order process . considering terms up to third order ( @xmath72 ) , eq .
[ pt ] then defines the linear polarizability @xmath73 ( @xmath74 or 2 ) , the polarizabilities for shg and thg , @xmath75 and @xmath76 , respectively , the wave - mixing polarizabilities corresponding to sum and difference frequency generation @xmath77 ( @xmath78 or 2 , @xmath79 ) , and the four - wave mixing polarizabilities @xmath80 and @xmath81 .
we obtain these polarizabilities by expanding the density matrix in eq .
[ rho_eom ] as @xmath82 which leads to self - consistent equations in @xmath83 ( for the relevant combinations of @xmath38 with harmonics @xmath67 and @xmath68 ) .
we solve eq .
[ rhons1s2 ] following a procedure inspired in the random - phase approximation ( rpa ) formalism , as discussed elsewhere @xcite for shg and thg .
further details on the extension of this formalism to cope with wave mixing are given in the appendix .
this work has been supported in part by the european commission ( graphene flagship cnect - ict-604391 and fp7 -ict-2013 - 613024-grasp ) .
the perturbative method used to simulate the optical response of nanographene to multifrequency continuous wave illumination is presented in detail .
we also include additional results from numerical simulations performed in the time - domain , and provide further wave - mixing polarizability spectra for graphene nanoislands with various sizes and dopings .
following a procedure similar to that described in previous work for the analysis of multiple - harmonic generation @xcite , we express the single - particle density matrix equation of motion for a graphene nanoisland in the basis set of its tight - binding electronic states as @xmath84 where @xmath85 is the energy of state @xmath42 and @xmath86 are the matrix elements of the total electric potential in the basis set of the 2p orbitals @xmath87 located at the carbon sites @xmath88 .
the single - electron states and the carbon site orbitals are related through @xmath89 give the amplitude of orbitals @xmath87 in states @xmath42 .
these states are orthonormal ( @xmath90 ) and form a complete set ( @xmath91 ) , thus facilitating transformations of the density matrix elements between site and state representations according to @xmath92 and @xmath93 . in what follows , we use indices @xmath94 to label carbon sites and @xmath95 for single - electron states . in what follows ,
we solve eq .
( [ dmjj ] ) for graphene nanoislands exposed to multifrequency continuous - wave ( cw ) illumination described by the electric field @xmath96 where @xmath65 is the field amplitude and @xmath97 the polarization unit vector .
assuming that @xmath65 is weak , we expand the density matrix as @xmath98 where @xmath99 indicates the perturbation order ( i.e. , terms proportional to @xmath100 , see eq .
( [ ee0 ] ) ) and @xmath67 ( @xmath68 ) is the harmonic index of @xmath12 ( @xmath11 ) . at @xmath101 order , eq .
( [ dmjj ] ) is trivially satisfied with @xmath102 .
following ref .
@xcite , we insert eq .
( [ rhons ] ) into eq .
( [ dmjj ] ) and collect terms with the same perturbation order and @xmath103 dependence , from which we obtain the density matrix at order @xmath104 as @xmath105 where @xmath106 and @xmath107 is the total potential . in eq .
( [ phinsrho ] ) , the first term represents the contribution from the external field ( nonzero only for order @xmath108 ) , while the second term describes the hartree potential produced by the perturbed electron density ( here @xmath109 indicates the spatial dependence of the coulomb interaction between electrons in orbitals @xmath87 and @xmath110 ) .
the latter quantity ensures a linear dependence on @xmath83 for the first term on the right - hand side of eq .
( [ rhons2 ] ) , whereas lower perturbation orders are contained within @xmath111 . for each order @xmath38 with harmonics
@xmath67 and @xmath68 satisfying @xmath112 , we are thus dealing with a self - consistent system in @xmath113 , which is treated using the approach described in ref .
@xcite . specifically , we proceed by first using the identity @xmath114 in the sum of eq .
( [ rhons2 ] ) , and then moving from state to site representation to obtain the diagonal density - matrix elements as @xmath115 where @xmath116 is the noninteracting rpa susceptibility at frequency @xmath117 . in summary ,
each new iteration order @xmath38 is computed from the results of previous orders following a precedure similar to ref .
@xcite , here generalized to deal with multifrequency cw illumination : 1 .
we first calculate @xmath111 using eq .
( [ etajj ] ) .
we then combine eqs .
( [ phinsrho ] ) and ( [ rhonsll2 ] ) to find a self - consistent equation for @xmath118 , which reduces in matrix form to @xmath119^{-1}\cdot\beta^{ns_1s_2},\ ] ] using site labels @xmath94 as matrix indices and having defined @xmath120 3 .
we use the calculated values of @xmath121 and @xmath122 to obtain @xmath123 using eq .
( [ rhonsll2 ] ) , and from here the induced charge at site @xmath94 at order @xmath38 associated with the harmonics @xmath67 and @xmath68 as @xmath124 .
computation and storage demand are reduced by using the property @xmath125 .
finally , the polarizability for wave mixing among the harmonics @xmath67 and @xmath68 is calculated from @xmath126 upon iteration of this procedure up to order @xmath38 .
extension to more frequencies is straightforward .
the linear absorption spectrum of an armchair - edged , doped graphene nanoisland exhibits numerous plasmonic resonance features , some of which display strong electrical tunability . in particular , the lowest - energy prominent peak ,
which is absent when the nanoisland is undoped , undergoes dramatic energy shifts upon adding only a few additional charge carriers ( see fig .
[ figs1 ] ) . as this mode is tunable , and also tends to produce large nonlinearities @xcite , we excite it in the nanoisland considered in fig . 1 of the main text with one of two collinear ultrashort pulses to study wave mixing .
[ figs2 ] we investigate wave mixing of collinear pulses in the same nanoisland , but here one of the pulses is tuned to a different plasmon from that of the main text .
interestingly , a different qualitative behavior is observed when the island is either undoped ( fig .
[ figs2]a ) or doped ( fig . [ figs2]b ) .
the former indicates that the second - order response vanishes for polarizations both parallel ( upper panel ) and perpendicular ( lower panel ) to an edge of the nanoisland when it is undoped , while a strong third - order response remains , although it is weaker than that of fig .
when the nanoisland is doped , the second - order response is recovered for the perpendicular polarization , but overall the nonlinear response is weaker than that of fig . 1b , where the highly - tunable mode participates in wave mixing .
in figs . [ figs3 ] and [ figs4 ] we present the wave - mixing polarizabilities corresponding to difference frequency generation , @xmath127 and four - wave mixing , @xmath128 , respectively , for the @xmath20 graphene nanoisland considered in figs . 2 and 3 of the main text doped with ( a ) three or ( b ) six electrons .
as before , we find similar plasmonic enhancement of the input frequencies in both cases ( see horizontal and vertical features in the contour plots ) , while in fig .
[ figs3 ] enhancement at the output frequency follows the curve @xmath129 and in fig .
[ figs4 ] the output enhancement follows @xmath130 , @xmath25 being any of the plasmonic resonances in an island . in fig .
[ figs3 ] we note a very large difference frequency generation polarizability near @xmath131ev , @xmath132ev ( or vice versa ) , where in fact we have a triple - resonance condition : simultaneously , both of the input frequencies and the output frequency are resonant with plasmons in the nanoisland . in fig .
[ figs4 ] , the cases where @xmath26 correspond to plasmonic enhancement of third - harmonic generation ( thg ) . for comparison , the third - harmonic susceptibility
@xmath133 has been measured in 10 nm silver nanoparticles as @xmath134esu @xcite , corresponding to a third - harmonic polarizability @xmath135 of @xmath136esu .
boyd , r. w. _ nonlinear optics _
, 3rd ed . ; academic press : amsterdam , 2008 garmire , e. nonlinear optics in daily life . _
express _ * 2013 * , _ 21 _ , 3053230544 wang , y. ; lin , c .- y . ; nikolaenko , a. ; raghunathan , v. ; potma , e. o. four - wave mixing microscopy of nanostructures .
photon . _ * 2011 * , _ 3 _ , 152 huang , l. ; cheng , j .- x .
nonlinear optical microscopy of single nanostructures .
res . _ * 2013 * , _ 43 _ , 213236 pu , y. ; grange , r. ; hsieh , c .- l . ;
psaltis , d. nonlinear optical properties of core - shell nanocavities for enhanced second - harmonic generation .
lett . _ * 2010 * , _ 104 _ , 207402 pantazis , p. ; maloney , j. ; wu , d. ; fraser , s. e. second harmonic generating ( shg ) nanoprobes for in vivo imaging .
sci . _ * 2010 * , _ 107 _ , 1453514540 novotny , l. ; van hulst , n. antennas for light .
photon . _ * 2011 * , _ 5 _ , 8390 chen , p .- y . ; argyropoulos , c. ; al , a. enhanced nonlinearities using plasmonic nanoantennas . _ nanophotonics _ * 2012 * , _ 1 _ , 221233 aouani , h. ; rahmani , m. ; navarro - ca , m. ; maier , s. a. third - harmonic - upconversion enhancement from a single semiconductor nanoparticle coupled to a plasmonic antenna .
nanotech . _ * 2014 * , _ 9 _ , 290294 metzger , b. ; hentschel , m. ; schumacher , t. ; lippitz , m. ; ye , x. ; murray , c. b. ; knabe , b. ; buse , k. ; giessen , h. doubling the efficiency of third harmonic generation by positioning ito nanocrystals into the hot - spot of plasmonic gap - antennas .
_ nano lett . _
* 2014 * , _ 14 _ , 28672872 alivisatos , a. p. semiconductor clusters , nanocrystals , and quantum dots .
_ science _ * 1996 * , _ 271 _ , 933937 stockman , m. i. nanoplasmonics : the physics behind the applications .
_ phys . today
_ * 2011 * , _ 64 _ , 3944 kauranen , m. ; zayats , a. v. nonlinear plasmonics .
. photon . _
* 2012 * , _ 6 _ , 737748 danckwerts , m. ; novotny , l. optical frequency mixing at coupled gold nanoparticles .
lett . _ * 2007 * , _ 98 _ , 026104 palomba , s. ; danckwerts , m. ; novotny , l. nonlinear plasmonics with gold nanoparticle antennas .
_ j. opt .
a : pure appl
. opt . _
* 2009 * , _ 11 _ , 114030 valev , v. k. characterization of nanostructured plasmonic surfaces with second harmonic generation .
_ langmuir _ * 2012 * , _ 28 _ , 1545415471 harutyunyan , h. ; volpe , g. ; quidant , r. ; novotny , l. enhancing the nonlinear optical response using multifrequency gold - nanowire antennas .
lett . _ * 2012 * , _ 108 _ , 217403 nordlander , y. z. f. w. y .- r
. z. p. ; halas , n. j. coherent fano resonances in a plasmonic nanocluster enhance optical four - wave mixing .
sci . _ * 2013 * , _ 110 _ , 92159219 novotny , l. ; hecht , b. _ principles of nano - optics _ ; cambridge university press : new york , 2006 ju , l. ; geng , b. ; horng , j. ; girit , c. ; martin , m. ; hao , z. ; bechtel , h. a. ; liang , x. ; zettl , a. ; shen , y. r. et al .
graphene plasmonics for tunable terahertz metamaterials .
nanotech . _ * 2011 * , _ 6 _ , 630634 fei , z. ; andreev , g. o. ; bao , w. ; zhang , l. m. ; mcleod , a. s. ; wang , c. ; stewart , m. k. ; zhao , z. ; dominguez , g. ; thiemens , m. et al .
infrared nanoscopy of dirac plasmons at the graphene
sio@xmath137 interface .
_ nano lett . _ * 2011 * , _ 11 _ , 47014705 chen , j. ; badioli , m. ; alonso - gonzlez , p. ; thongrattanasiri , s. ; huth , f. ; osmond , j. ; spasenovi , m. ; centeno , a. ; pesquera , a. ; godignon , p. et al .
optical nano - imaging of gate - tunable graphene plasmons .
_ nature _ * 2012 * , _ 487 _ , 7781 fei , z. ; rodin , a. s. ; andreev , g. o. ; bao , w. ; mcleod , a. s. ; wagner , m. ; zhang , l. m. ; zhao , z. ; thiemens , m. ; dominguez , g. et al .
gate - tuning of graphene plasmons revealed by infrared nano - imaging .
_ nature _ * 2012 * , _ 487 _ , 8285 yan , h. ; li , x. ; chandra , b. ; tulevski , g. ; wu , y. ; freitag , m. ; zhu , w. ; avouris , p. ; xia , f. tunable infrared plasmonic devices using graphene / insulator stacks . _ nat . nanotech . _ * 2012 * , _ 7 _ , 330334 yan , h. ; li , z. ; li , x. ; zhu , w. ; avouris , p. ; xia , f. infrared spectroscopy of tunable dirac terahertz magneto - plasmons in graphene .
_ nano lett . _ * 2012 * , _ 12 _ , 37663771 fang , z. ; thongrattanasiri , s. ; schlather , a. ; liu , z. ; ma , l. ; wang , y. ; ajayan , p. m. ; nordlander , p. ; halas , n. j. ; garca de abajo , f. j. gated tunability and hybridization of localized plasmons in nanostructured graphene .
_ acs nano _ * 2013 * , _ 7 _ , 23882395 brar , v. w. ; jang , m. s. ; sherrott , m. ; lopez , j. j. ; atwater , h. a. highly confined tunable mid - infrared plasmonics in graphene nanoresonators .
_ nano lett . _ * 2013 * , _ 13 _ , 25412547 fang , z. ; wang , y. ; schlather , a. ; liu , z. ; ajayan , p. m. ; garca de abajo , f. j. ; nordlander , p. ; zhu , x. ; halas , n. j. active tunable absorption enhancement with graphene nanodisk arrays .
_ nano lett . _ * 2014 * , _ 14 _ , 299304 yan , h. ; low , t. ; zhu , w. ; wu , y. ; freitag , m. ; li , x. ; guinea , f. ; avouris , p. ; xia , f. damping pathways of mid - infrared plasmons in graphene nanostructures .
photon . _ * 2013 * , _ 7 _ , 394399 freitag , m. ; low , t. ; zhu , w. ; yan , h. ; xia , f. ; avouris , p. photocurrent in graphene harnessed by tunable intrinsic plasmons .
commun . _ * 2014 * , _ 4 _ , 1951 grigorenko , a. n. ; polini , m. ; novoselov , k. s. graphene plasmonics .
photon . _ * 2012 * , _ 6 _ , 749758 , f. j. graphene plasmonics : challenges and opportunities .
_ acs photon . _ * 2014 * , _ 1 _ , 135152 hendry , e. ; hale , p. j. ; moger , j. ; savchenko , a. k. ; mikhailov , s. a. coherent nonlinear optical response of graphene .
_ phys . rev .
lett . _ * 2010 * , _ 105 _ , 097401 zhang , h. ; virally , s. ; bao , q. ; ping , l. k. ; massar , s. ; godbout , n. ; kockaert , p. z - scan measurement of the nonlinear refractive index of graphene . _
lett . _ * 2012 * , _ 37 _ , 18561858 kumar , n. ; kumar , j. ; gerstenkorn , c. ; wang , r. ; chiu , h .- y . ; smirl , a. l. ; zhao , h. third harmonic generation in graphene and few - layer graphite films .
b _ * 2013 * , _ 87 _ , 121406(r ) mikhailov , s. a. theory of the giant plasmon - enhanced second - harmonic generation in graphene and semiconductor two - dimensional electron systems .
b _ * 2011 * , _ 84 _ , 045432 gullans , m. ; chang , d. e. ; koppens , f. h. l. ; garca de abajo , f. j. ; lukin , m. d. single - photon nonlinear optics with graphene plasmons .
lett . _ * 2013 * , _ 111 _ , 247401 manzoni , m. t. ; silveiro , i. ; garca de abajo , f. j. ; chang , d. e. second - order quantum nonlinear optical processes in graphene nanostructures .
_ arxiv:1406.4360_. cox , j. d. ; garca de abajo , f. j. electrically tunable nonlinear plasmonics in graphene nanoislands .
commun . _ * 2014 * , _ 5 _ , 5725 thongrattanasiri , s. ; manjavacas , a. ; garca de abajo , f. j. quantum finite - size effects in graphene plasmons .
_ acs nano _ * 2012 * , _ 6 _ , 17661775 manjavacas , a. ; thongrattanasiri , s. ; garca de abajo , f. j. plasmons driven by single electrons in graphene nanoislands .
_ nanophotonics _ * 2013 * , _ 2 _ , 139151 manjavacas , a. ; marchesin , f. ; thongrattanasiri , s. ; koval , p. ; nordlander , p. ; snchez - portal , d. ; garca de abajo , f. j. tunable molecular plasmons in polycyclic aromatic hydrocarbons .
_ acs nano _ * 2013 * , _ 7 _ , 36353643 wu , j. ; pisula , w. ; mllen , k. graphenes as potential material for electronics . _ chem .
rev . _ * 2007 * , _ 107 _ , 718747 feng , x. ; liu , m. ; pisula , w. ; takase , m. ; li , j. ; mllen , k. supramolecular organization and photovoltaics of triangle - shaped discotic graphenes with swallow - tailed alkyl substituents . _ adv . mater . _ * 2008 * , _ 20 _ , 26842689 feng , x. ; pisula , w. ; mllen , k. large polycyclic aromatic hydrocarbons : synthesis and discotic organization .
_ pure appl .
chem . _ * 2009 * , _ 81 _ , 22032224 cai , j. ; ruffieux , p. ; jaafar , r. ; bieri , m. ; braun , t. ; blankenburg , s. ; muoth , m. ; seitsonen , a. p. ; saleh , m. ; feng , x. et al .
atomically precise bottom - up fabrication of graphene nanoribbons .
_ nature _ * 2010 * , _ 466 _ , 470473 denk , r. ; hohage , m. ; zeppenfeld , p. ; cai , j. ; pignedoli , c. a. ; sde , h. ; fasel , r. ; feng , x. ; mllen , k. ; wang , s. et al .
exciton - dominated optical response of ultra - narrow graphene nanoribbons .
commun . _ * 2014 * , _ 5 _ , 4253 hache , f. ; ricard , d. ; flytzanis , c. ; kreibig , u. the optical kerr effect in small metal particles and metal colloids : the case of gold .
phys . a - mater .
sci . process .
_ * 1988 * , _ 47 _ , 347357 uchida , k. ; kaneko , s. ; omi , s. ; hata , c. ; tanji , h. ; asahara , y. ; ikushima , a. j. ; tokizaki , t. ; nakamura , a. optical nonlinearities of a high concentration of small metal particles dispersed in glass : copper and silver particles .
_ j. opt .
b _ * 1994 * , _ 11 _ , 12361243 yang , l. ; becker , k. ; smith , f. m. ; iii , r. h. m. ; r.f .
haglund , j. ; yang , l. ; dorsinville , r. ; alfano , r. r. ; zuhr , r. a. size dependence of the third - order susceptibility of copper nanoclusters investigated by four - wave mixing .
_ j. opt .
b _ * 1994 * , _ 11 _ , 457461 sato , r. ; ohnuma , m. ; oyoshi , k. ; takeda , y. experimental investigation of nonlinear optical properties of ag nanoparticles : effects of size quantization .
b _ * 2014 * , _ 90 _ , 125417 woessner a. _ et al .
_ highly confined low - loss plasmons in graphene - boron nitride heterostructures .
_ arxiv_:1409.5674 .
wallace , p. r. the band theory of graphite .
rev . _ * 1947 * , _ 71 _ , 622634 , a. h. ; guinea , f. ; peres , n. m. r. ; novoselov , k. s. ; geim , a. k. the electronic properties of graphene .
phys . _ * 2009 * , _ 81 _ , 109162 hedin , l. ; lundqvist , s. effects of electron - electron and electron - phonon interactions on the one - electron states of solids . in _ solid state physics _ ; frederick seitz , d. t. , ehrenreich , h. , eds . ; academic press , 1970 ; vol .
23 , pp 1 181 liu , t .-
m . ; tai , s .- p . ; yu , c .- h . ; wen , y .- c . ; chu , s .- w . ; chen , l .- j . ; prasad , m. r. ; lin , k .- j . ; sun , c .- k .
_ * 2006 * , _ 89 _ , 043122 | localized plasmons in metallic nanostructures have been widely used to enhance nonlinear optical effects due to their ability to concentrate and enhance light down to extreme - subwavelength scales . as alternatives to noble metal nanoparticles , graphene nanostructures can host long - lived plasmons that efficiently couple to light and are actively tunable via electrical doping . here
we show that doped graphene nanoislands present unique opportunities for enhancing nonlinear optical wave - mixing processes between two externally applied optical fields at the nanoscale .
these small islands can support pronounced plasmons at multiple frequencies , resulting in extraordinarily high wave - mixing susceptibilities when one or more of the input or output frequencies coincide with a plasmon resonance . by varying the doping charge density in a nanoisland with a fixed geometry , enhanced wave mixing can be realized over a wide spectral range in the visible and near infrared .
we concentrate in particular on second- and third - order processes , including sum and difference frequency generation , as well as on four - wave mixing .
our calculations for armchair graphene triangles composed of up to several hundred carbon atoms display large wave mixing polarizabilities compared with metal nanoparticles of similar lateral size , thus supporting nanographene as an excellent material for tunable nonlinear optical nanodevices . |
the quest for the origin of cosmic rays is one of the most fundamental problems in astroparticle physics @xcite . since the discovery of these highly energetic particles more than a century ago ,
numerous measurements of several of their properties have been made , using sophisticated instruments ( see e.g. ref .
@xcite for a review ) .
however , the exact nature of their sources still remains an open question .
the search is mainly hindered due to the fact that cosmic rays , being charged particles , are scattered or deflected by the galactic and inter - galactic magnetic fields during their propagation to the earth , making it extremely difficult to reconstruct the direction of their sources
. nevertheless , observed cosmic - ray properties like the energy spectrum and composition have been used to understand and characterize the properties of the sources such as their galactic or extragalactic nature , the cosmic - ray production spectrum and the power injected into cosmic rays ( see e.g. refs .
@xcite for recent reviews ) .
lofar , the low frequency array , is an astronomical radio telescope @xcite .
it has been designed to measure the properties of cosmic rays above @xmath2 ev by detecting radio emission from extensive air showers in the frequency range of @xmath3 mhz @xcite .
one of the main goals of the lofar key science project cosmic rays is to provide an accurate measurement of the mass composition of cosmic rays in the energy range between @xmath2 and @xmath4 ev , a region where the transition from galactic to extragalactic cosmic rays is expected .
this is being carried out by measuring the depth of the shower maximum ( @xmath5 ) , using a technique based on the reconstruction of the two - dimensional radio intensity profile on the ground @xcite .
another focus of the lofar cosmic - ray measurements is to understand the nature and production mechanisms of the radio emission from air showers .
this is done by measuring various properties of the radio signals in great detail such as their polarization properties , the radio wave front and relativistic time compression effects on the emission profile @xcite . in order to assist the radio measurement of air showers with lofar ,
we have built a particle detector array lora ( lofar radboud air shower array ) in the center of lofar @xcite .
its main objectives are to trigger the read - out of the lofar radio antennas to register radio signals from air showers , and to provide basic air shower parameters such as the position of the shower axis as well as the energy and the arrival direction of the incoming cosmic - ray .
these parameters are used to cross - check the reconstruction of air shower properties , based on the measured radio signals . currently , given the lack of an absolute calibration of the radio signals , the cosmic - ray energy is estimated through the reconstruction of the particle data .
therefore , an accurate energy reconstruction with lora is essential for a proper understanding of the air showers measured with lofar . in this article
, we describe in detail the various steps of the energy reconstruction and present the cosmic - ray energy spectrum above @xmath2 ev as measured with lora .
the article is organized as follows .
a short description of the set - up will be given in section [ sec : lora ] followed by a description of the data analysis technique in section [ sec : analysis ] .
the various steps involved in the monte - carlo simulation studies of the array will be described in section [ sec : simulation ] , and a comparison between measurements and simulations for some of the air shower properties will be given in section [ sec : comparison ] . in sections
[ sec : energy - calibration ] and [ sec : energy - resolution ] , the energy calibration , the uncertainties in the reconstructed energies , and reconstructed cosmic - ray intensity will be described .
the measured cosmic - ray spectrum and a comparison with the measurements of other experiments will be presented in section [ sec : cosmic - ray - spectrum ] , followed by a short conclusion and a future outlook .
m , which is used in the analysis . ]
lora ( the lofar radboud air shower array ) consists of an array of 20 plastic scintillation detectors of size @xmath6 m each , distributed over a circular area with a diameter of @xmath7 m in the center of lofar @xcite .
the array is subdivided into @xmath8 units , each comprising of 4 detectors .
the detectors have a spacing between @xmath9 m , and have been designed to measure cosmic rays with energies above @xmath2 ev .
the array is co - located with six lofar stations high - band antennas , operating in the frequency range of @xmath10 mhz and @xmath11 mhz respectively . ] .
the layout of the array is shown in figure [ fig : lora - array ] .
the data acquisition in each unit is controlled locally .
a local trigger condition of 3 out of 4 detectors is set for each unit , and an event is accepted for a read - out of the full array when at least one unit has been triggered .
a high - level trigger for the lofar radio antennas is formed when at least 13 out of the 20 detectors have measured a signal above threshold .
more technical details can be found in ref .
[ sec : analysis ] data collected with the lora array since its first science operation in june 2011 until october 2014 are used . only data collected in periods with all 20 detectors in operation will be considered .
this amounts to a total of @xmath12 days of data . for the analysis ,
only showers that trigger a minimum of @xmath8 detectors will be considered , which corresponds to a total of @xmath13 air showers . for every measured shower ,
the signal arrival time and the energy deposit in each detector are recorded .
the relative signal arrival times between the detectors are used to reconstruct the arrival direction of the primary cosmic ray .
the energy deposits are used to reconstruct the position of the shower axis and the shower size ( the effective number of charged particles at the ground ) .
the latter is determined in terms of the number of vertical equivalent charged particles , which may also include converted photons in addition to the dominant charged particles - electrons and muons . the shower axis position and the shower size
are determined simultaneously by fitting a lateral density distribution function to the measured two - dimensional distribution of particle densities , projected into the shower plane . the particle density in each detector
is obtained by first dividing the track - length - corrected factor where @xmath14 is the zenith angle of the reconstructed arrival direction of the primary cosmic ray . ]
energy deposition by the energy deposition of a single particle obtained from calibration , and then by further dividing by the projected area of the detector in the shower plane .
the lateral density distribution of an air shower is generally described by the nishimura - kamata - greisen ( nkg ) function which is given by @xcite @xmath15 where @xmath16 represents the particle density in the shower plane at a radial distance @xmath17 from the shower axis , @xmath18 is the shower size , @xmath19 is shower age or lateral shape parameter and @xmath20 is the radius parameter which is basically a measure of the lateral spread of the shower
. the function @xmath21 is given by @xmath22 in the case of lora , the value of @xmath20 is determined from the fit along with @xmath18 and the position of the shower axis .
the parameter @xmath19 is kept constant at a value of @xmath23 throughout the fitting process .
simultaneous fitting of both @xmath20 and @xmath19 results in fits of poorer quality .
simulation studies have shown that keeping @xmath19 constant gives better results than keeping @xmath20 constant @xcite .
the fitting procedure is repeated three times with the output of each fit taken as starting values for the next iteration .
details about the minimization procedure and the choice of starting values , as well as the reconstruction of the arrival direction of the primary particle are described in ref .
@xcite .
all showers that trigger at least @xmath8 detectors with a minimum of @xmath24 particle m@xmath25 are allowed to pass through the reconstruction algorithm , and their shower parameters are calculated .
furthermore , only showers whose reconstructed position of the shower axis falls within @xmath26 m from the center of the array are selected .
the normalized distribution of @xmath20 values for the selected showers with reconstructed sizes @xmath27 and reconstructed zenith angles in the range of @xmath28 are shown in figure [ average - lateral ] ( left panel ) .
the inset shows a closer view of the distribution around the maximum value between @xmath29 and @xmath30 m , and a gaussian fit to the distribution .
the fit gives a peak value of @xmath31 m. figure [ average - lateral ] ( right panel ) shows the averaged lateral distributions of the measured showers for different reconstructed size bins in the range of @xmath32 for zenith angle between @xmath33 and @xmath34 .
the distributions include events that passed through the same selection cuts applied in the left panel of figure [ average - lateral ] and have @xmath20 values in the range of @xmath35 m. the averaged distributions are obtained by stacking together the lateral distributions of all individual showers contained in each size bin .
the lines in the plot represent the fits to the data using ( [ eq : lateral - density ] ) . to avoid clumsiness of the plots
, uncertainties are shown only for the size bins of @xmath36 and @xmath37 .
however , all the respective uncertainties are taken into account in the fitting procedure .
the values of @xmath20 obtained from the fits are in the range of @xmath38 m. the shower size gives a good measure of the energy of the primary cosmic - ray particle , initiating the air shower .
therefore , the shower size distribution should reflect the energy distribution of the cosmic rays at size values where the primary energy is above the detector threshold .
figure [ size - distribution ] shows the distribution of reconstructed shower sizes for all the measured showers that passed through the various trigger and quality cuts applied in the analysis .
this corresponds to a total of @xmath39 air showers .
the distribution shows a steep rise as @xmath18 increases which is due to the sharp increase in the detector acceptance ( see section [ sec : simulation ] ) as function of the primary energy . after reaching a maximum
, the distribution falls off steeply which is due to the power - law behavior of the cosmic - ray spectrum .
the peak of the total distribution gives the shower size threshold of the detector array .
fitting a gaussian function around the peak gives a value of @xmath40 .
also shown in figure [ size - distribution ] are the reconstructed size distributions for four zenith angle bins : @xmath28 , @xmath41 , @xmath42 , and @xmath43 .
the parameterization of the cosmic - ray energy will be determined separately for each zenith angle bin ( see section [ sec : energy - calibration ] ) .
all cuts applied in this analysis are summarized in table [ quality - cut ] .
measured size distribution of the air showers that have passed the quality cuts given in table [ quality - cut ] .
the line represents a gaussian fit to the total distribution around the peak , giving a shower size threshold of @xmath40 for the array . ]
[ quality - cut ] [ cols= " < , < " , ]
the measured shower size can be converted into the energy of the primary particle using a conversion relation obtained from simulations .
simulated showers are stored in a two dimensional log - log histogram in reconstructed size and true energy .
such a histogram is shown in figure [ size - energy - plot ] for showers induced by protons ( left panel ) and iron nuclei ( right panel ) for the zenith angle range of @xmath28 .
the color profile represents the weight of the distribution .
the distribution is broader for proton showers , which is mainly due to the large intrinsic fluctuations of proton showers . from simulations ,
the fluctuations in the true shower size for proton showers within the @xmath28 zenith angle bin are found to be @xmath44 while the uncertainty due to the reconstruction is in the range of @xmath45 for the energy region of our interest .
for iron induced showers , the intrinsic size fluctuation is only @xmath46 , while the reconstruction accuracy remains almost the same as that of the proton induced showers .
another major difference is that for the same reconstructed shower size , iron showers have higher energies than the protons .
this is related to the shallower penetration depth of iron induced showers in the atmosphere , which leads to an increased attenuation of electrons before they can reach the ground .
the distributions in figure [ size - energy - plot ] are binned in @xmath18 , taking a logarithmic bin size of @xmath47 , and profile plots of the true energy as function of @xmath18 are generated .
the profile plots are represented by the solid points in figure [ size - energy - plot ] .
each point in the plots represents the peak of the energy distribution for each size bin , and the uncertainty on each point corresponds to the spread of the energy distribution which is described in detail in the following .
figure [ energy - dis ] shows the energy distribution for the bin slice of @xmath48 for both , showers induced by protons ( left panel ) and iron nuclei ( right panel ) .
the distribution for protons is not symmetric about its mean and is found to be more extended to lower energies .
this can be understood as more contamination from low - energy showers in a given size bin than from higher energies which is caused by the larger intrinsic fluctuations of low - energy showers .
the level of contamination depends on the assumed slope of the primary cosmic - ray spectrum in the simulation .
the peaks of the distributions are obtained by fitting with a skewed gaussian function . the skewed gaussian distribution function used in the present analysis
is given by , @xmath49,\ ] ] where @xmath50 , @xmath51 is the normalisation constant , @xmath52 and @xmath53 represent measures of the spread and the position of the distribution respectively , and @xmath54 is the skewness parameter of the distribution function . the lines in figure [ energy - dis ] represent the fitted functions .
the important fit parameters @xmath55 are also shown .
for the distribution of iron - induced showers , it can be noticed that the value of @xmath54 is close to zero , indicating that the distribution closely resembles a normal gaussian distribution .
the uncertainties in the profile plots shown in figure [ size - energy - plot ] are obtained by taking the difference between the energies corresponding to the full width at half maximum ( fwhm ) and the peak energy of the energy distribution for each size bin .
the uncertainties obtained are asymmetric for the proton distribution while for iron induced showers , they are almost symmetric .
the @xmath18 values of the profile plots shown in figure [ size - energy - plot ] are calculated as the weighted mean of the size distribution within each size bin .
these size values are found to be slightly smaller than the bin centers . to obtain the size - energy relation , each profile plot
is fitted using the following function , @xmath56 where @xmath57 denotes the true energy , and @xmath58 and @xmath59 are the fit parameters .
the fit is performed only in the size region where a reliable fit of the true energy distribution , as shown in figure [ energy - dis ] , could be performed .
this corresponds to a size region of @xmath60 for both the type of particles .
the profile plots as well as the fitted functions for all the zenith angle ranges are shown in figure [ size - energy - all - angles ] for protons ( left panel ) and iron nuclei ( right panel ) . from these figures
, it can be noticed that for the same shower size , primary energies at larger zenith angles are larger than at smaller angles . in other words , it requires a higher energy at larger zenith angles to generate the same number of particles on the ground as at lower zenith angles .
this is due to higher attenuation of air shower particles at larger zenith angles as the showers pass through a longer column depth of air in the atmosphere .
the values of the @xmath58 and @xmath59 parameters obtained from the fits for the four zenith angle ranges are listed in table [ fit - parameters ] . using these values , for any simulated or measured shower for which the reconstructed arrival direction and the reconstructed size are known
, the primary cosmic - ray energy can be reconstructed using the relation @xmath61 where @xmath62 denotes the reconstructed energy .
in this section , we present details about the accuracy of the reconstructed energies and the uncertainties that have to be considered for the reconstruction of the cosmic - ray intensity .
the accuracy depends on the variation of the true shower size caused by the intrinsic shower - to - shower fluctuations in the atmosphere and also on the accuracy in the reconstruction of the shower size . for each size bin in the @xmath63 profile plot , reconstructed energies @xmath64
are obtained for every simulated shower , and a distribution of the differences between the true energies and the reconstructed energies @xmath65 is generated .
the distribution obtained is similar to the one shown in figure [ energy - dis ] , except for a shift in the peak position to the left by an interval equal to the value of the reconstructed energy .
the peak position and the spread of these distributions are obtained correspondingly .
the peak represents the systematic uncertainty due to energy calibration , while the spread corresponds to the energy resolution .
their values expressed as fraction of the reconstructed energies are shown in figure [ energy - accu ] as function of the shower size for the zenith angle range of @xmath66 .
the resolution is in the range of @xmath67 for proton induced showers and @xmath68 for showers induced by iron nuclei .
the systematics are within 12% for protons and within 8% for iron nuclei . at @xmath69 ,
the uncertainty in energy for protons increases to the range of @xmath70 in resolution and to @xmath71 in systematics . for iron nuclei ,
the uncertainty remains almost the same up to @xmath72 .
accuracy in the reconstructed energies ( @xmath62 ) for showers induced by protons ( squares ) and iron nuclei ( circles ) as a function of the reconstructed size @xmath18 for the zenith angle bin of @xmath28 .
the filled points represent the energy resolution and the the empty points are the systematic uncertainties resulting from the energy calibration .
see section [ energy - resolution ] for details . ] the systematic uncertainty in energy shown in figure [ energy - accu ] is associated with the energy calibration performed using equation [ eq : energy - calibration ] .
other main sources of systematic uncertainty in energy include the assumed slope of the primary cosmic - ray spectrum in the corsika simulation , the vem peak obtained from the detector simulation and the hadronic interaction models .
thus , the calibration parameters , listed in table [ fit - parameters ] , also depend on the choice of simulation parameters .
a part of the systematic uncertainties are obtained by changing the values of the slope and the vem peak in the simulations within reasonable limits , and by comparing the newly reconstructed energies with the energies obtained using the fixed parameters given in table [ fit - parameters ] . for the slope of the energy spectrum ,
simulated showers with an original slope @xmath73 are weighted to generate distributions for @xmath74 and @xmath75 .
then , following the same procedure as described in section [ sec : energy - calibration ] , energy calibrations are performed separately for the two different slopes and calibration parameters are obtained .
the differences between the energies reconstructed with the new parameters and the ones reconstructed using the parameters given in table [ fit - parameters ] gives the systematic uncertainty due to the spectral slope .
the uncertainties are found to be within @xmath76 for protons and within @xmath77 for iron nuclei . from the detector simulation
, it has been observed that adding noise to the deposited energy in the detector at the level of @xmath24 mev ( see section [ detector - simulation ] ) leads to around @xmath78 positive shift in the value of the most probable energy deposition @xmath79 in the detector for vertical incident muons .
the 10% increase in @xmath79 will lead to a decrease in the shower size and subsequently to an increase in the reconstructed energy by @xmath80 .
the average systematic shift in the reconstructed energy due to this uncertainty in vem calibration is obtained to be @xmath81 for showers induced by either protons or iron nuclei .
the different systematic uncertainties obtained are shown in figure [ energy - sys ] as a function of the reconstructed energy for showers induced by protons ( left panel ) and iron nuclei ( right panel ) . for proton showers , the total systematic uncertainty , obtained by adding the individual systematic components in quadrature ,
is found to be within @xmath82 and for iron showers , the total systematic is within @xmath83 . at larger zenith angles ,
the total systematic for protons increases slightly , reaching @xmath84 at @xmath69 , while for iron nuclei , the uncertainty remains almost unchanged . any systematic uncertainty in energy results in a systematic shift in the reconstructed cosmic - ray flux intensity .
to estimate the systematic uncertainty in intensity due to the energy calibration , the reconstructed energies are determined using equation [ eq : energy - calibration ] for all simulated showers with @xmath85 that pass through all selection and quality cuts as listed in table [ quality - cut ] .
the distribution of the reconstructed energies is compared to the distribution of the true energies , and the systematic uncertainty in intensity is calculated as @xmath86 for each energy bin , where @xmath87 and @xmath88 represent the number of showers per bin in the true and reconstructed energy distributions respectively . for the systematic effect due to the uncertainties in the spectral slope and the vem calibration , the energy calibration determined in their respective cases
are applied to the simulated showers for @xmath85 and the distributions of the newly reconstructed energies are compared with the old distribution obtained using the parameters given in table [ fit - parameters ] .
figure [ intensity - sys ] shows the different systematic uncertainties in intensity that have been obtained for protons ( left panel ) and iron nuclei ( right panel ) .
the thick solid line represents the systematic uncertainty due to the energy calibration , the blue band represents the contribution due to the spectral slope , the dashed line is the vem contribution , and the shaded - striped region represents the total systematic uncertainty . for energies above @xmath89 ,
the systematic uncertainty due to the energy calibration is found to be within @xmath90 for proton showers and within @xmath82 for showers induced by iron nuclei . the systematic uncertainty due to the spectral slope is within @xmath91 for protons , and within @xmath92 for iron nuclei except at @xmath93 where the uncertainty reaches @xmath94 .
the systematic uncertainty associated with the vem calibration is found to be within @xmath95 for both types of nuclei .
a contribution of @xmath96 due to the uncertainty in the hadronic interaction model @xcite is also included in figure [ intensity - sys ] . for protons , the total systematic uncertainty above @xmath89 is within @xmath97 , and for iron nuclei , the total uncertainty is within @xmath98 .
for all high - quality lora data , reconstructed energies are determined on shower - by - shower basis , and a distribution of reconstructed energies is built taking a logarithmic bin size of 0.15 . from the distribution ,
the differential cosmic - ray spectrum @xmath99 is obtained by folding in the total acceptance of the lora array @xmath100 ( figure [ total - effective - area ] ) and the total observation time @xmath101 as follows , @xmath102 where the subscript @xmath103 denotes the @xmath104 energy bin and @xmath105 is the number of showers in an energy bin of width @xmath106 . for constructing the spectrum , only the energy region that has a total ( trigger and reconstruction ) efficiency greater than @xmath107 is used .
this corresponds to an energy of @xmath108 gev for protons and @xmath109 gev for iron nuclei ( see figure [ total - efficiency ] ) .
figure [ cr - spectrum ] ( left panel ) shows the reconstructed energy spectrum multiplied by @xmath110 , assuming that cosmic rays are only protons or iron nuclei .
the spectrum is given in the energy range of @xmath111 gev for protons , and in the range of @xmath112 gev for iron nuclei .
the measured values along with the uncertainties are listed in table [ energy - spectra ] .
the measured spectra can not be described by single power laws over the full energy range because of the structures present in the spectra , particularly the dip at @xmath113 gev .
a power law fit to the measured spectra data below @xmath114 gev gives spectral index values of @xmath115 for protons and @xmath116 for iron nuclei . in figure [ cr - spectrum ] ( right panel )
, our measured spectra are compared with the all - particle spectra measured with the icetop @xcite and kascade - grande @xcite experiments .
both their spectra lie between our reconstructed spectra , which is expected in the case of a mixed cosmic - ray composition .
they are close to our proton spectrum at @xmath117 gev , and become closer to our iron spectrum as the energy increases .
this might be an indication of a change in the mass composition of cosmic rays in the energy region between @xmath118 and @xmath119 gev , which is expected as due to a transition from a galactic to an extragalactic origin of cosmic rays .
we have conducted a detailed energy reconstruction study for the extensive air showers measured with the lora particle detector array .
important parameters such as the energy resolution of the array and the systematic uncertainty of the reconstructed energy have been obtained .
the energy resolution is found to be in the range of @xmath120 for showers induced by protons and @xmath121 for iron nuclei .
the total systematic uncertainty of the reconstructed energy is within @xmath82 for protons and within @xmath122 for iron nuclei . applying the reconstruction method to the measured data ,
the all - particle cosmic - ray energy spectrum has been obtained , assuming that cosmic rays are only constituted of protons or iron nuclei for energies above @xmath2 ev with a systematic uncertainty in intensity of @xmath123 .
our future effort will concentrate on combining the energy measurement of lora with the composition measurement from the lofar radio antennas to determine an all - particle energy spectrum , taking into account the actual cosmic - ray composition . especially the primary energy determined using the energy calibration given here
is being used in the reconstruction of air shower properties with the radio data from lofar .
calculation of energy calibration parameters for higher zenith angles above @xmath124 is underway .
this is particularly important for the lofar radio measurements where a significant fraction of showers have been observed at larger zenith angles . at present
, the small size of the lora array effectively limits the effective area of lofar .
efforts are ongoing to expand the size of the array to exploit the full potential of lofar .
[ energy - spectra ] ccc & + ' '' '' ( gev ) & protons & iron nuclei + ' '' '' @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 @xmath158 + ' '' ''
@xmath159 & @xmath160 @xmath161 & @xmath162 @xmath163 + ' '' ''
@xmath164 & @xmath165 @xmath166 & @xmath167 @xmath168 + ' '' ''
@xmath169 & @xmath170 @xmath171 & @xmath172 @xmath173 + ' '' ''
@xmath174 & @xmath175 @xmath176 & @xmath177 @xmath178 + ' '' ''
@xmath179 & @xmath180 @xmath181 & @xmath182 @xmath183 + ' '' ''
@xmath184 & @xmath185 @xmath186 & @xmath187 @xmath188 + ' '' ''
@xmath189 & @xmath128 & @xmath190 @xmath191 +
we would like to thank the technical support from astron . in particular , we are grateful to j. nijboer , m. norden , k. stuurwold and h. meulman for their support in the installation and in the maintenance of lora in the lofar core .
we are also grateful to the kascade - grande collaboration for generously lending us the scintillator units .
we acknowledge funding from the samenwerkingsverband noord - nederland ( snn ) , the netherlands research school for astronomy ( nova ) and from the european research council ( erc ) under the european unions seventh framework program ( fp/2007 - 2013 ) / erc grant agreement no .
lofar , the low frequency array designed and constructed by astron , has facilities in several countries , that are owned by various parties ( each with their own funding sources ) , and that are collectively operated by the international lofar telescope ( ilt ) foundation under a joint scientific policy .
nagano , m. , @xmath192 watson , a. a , 2000 , rev .
phys . , 72 , 689 blmer , j. , engel , r. , @xmath192 hrandel , j , 2009 , prog . part .
63 , 293 hrandel , j. r. , 2008 , rev .
20 , 203 hrandel , j. r. , 2006 , jphcs , 47 , 41 thoudam & hrandel , j. r. , 2014 , a&a , 567 , a33 hrandel , j. r. , 2004 , aph , 21 , 241 hillas , a.m. , j. phys .
g : nucl . part .
phys . , 2005 , 31 , r95 berezhko , e. g. , 2009 , apj , 698 , l138 hrandel , j. r. , 2008 , adspr , 2008 , 41 , 442 blasi , p. , 2013
, a&arv , 21 , 70 blasi , p. , 2014 ,
brjph , 44 , 426 van haarlem , m. , 2013 , a&a , 556 , a2 schellart , p. et al . , 2013 , a&a , 560 , a98 buitink , s. et al . , 2014 ,
prd , 90 , 082003 nelles , a. et al .
, 2015 , jcap , 05 , 018 schellart , p. et al . , 2014 , jcap , 10 , 014 corstanje , a. et al . , 2015 ,
aph , 61 , 22 nelles , a. et al . , 2015 , aph , 65 , 11 thoudam , s. et al . , 2014 , nimpa , 767 , 339 kamata , k. , & nishimura , j. , 1958 , prog .
6 , 93 greisen , k. , 1960 , ann .
, 10 , 63 antoni , t. , et al .
2001 , aph , 14 , 245 heck , d. , 1998 , report fzka , 6019 ostapchenko , s. s. , 2011 , prd , 83 , 014018 .
fass , a. , et al . , 2005 , cern-2005 - 10 , infn / tc-05/11 , slac - r-773 nelson , w. r. , hirayama , h. & rogers , d. w. o , 1985 , slac-0265 kobal , m. et al . , 2001 ,
aph , 15 , 259 agostinelli , s. et al . , 2003 , nimpa , 506 , 250 aartsen , m. g. et al , 2013 , prd , 88 , 042004 apel , w. d. et al . , 2012 ,
aph , 36 , 183 kang , d. et al . , 2013 , jphcs , 409 , 012101 | the energy reconstruction of extensive air showers measured with the lofar radboud air shower array ( lora ) is presented in detail .
lora is a particle detector array located in the center of the lofar radio telescope in the netherlands .
the aim of this work is to provide an accurate and independent energy measurement for the air showers measured through their radio signal with the lofar antennas .
the energy reconstruction is performed using a parameterized relation between the measured shower size and the cosmic - ray energy obtained from air shower simulations . in order to illustrate the capabilities of lora ,
the all - particle cosmic - ray energy spectrum has been reconstructed , assuming that cosmic rays are composed only of protons or iron nuclei in the energy range between @xmath0 and @xmath1 ev .
the results are compatible with literature values and a changing mass composition in the transition region from a galactic to an extragalactic origin of cosmic rays . cosmic rays , air showers , energy spectrum , lora |
in order to characterize our transfer function , we measure rabi oscillations of the spin signal under microwave pulses across a set of offset frequencies . by numerically fitting the measured oscillations to a theoretical model , the amplitude and phase of the transfer function are obtained .
first , we consider the hamiltonian of the single - qubit system in the lab frame : @xmath128 where @xmath129 is the zeeman interaction between the electron spin and the static magnetic field and @xmath130 is the interaction between the electron spin and the microwave field .
using the rotating - wave approximation ( rwa ) , the hamiltonian in the rotating frame with frequency @xmath131 has the following form : @xmath132 where @xmath133 , @xmath134 , and @xmath135 are the offset frequency , amplitude , and phase of the microwave pulse , respectively .
here , we consider a constant @xmath134 for simplicity which means the microwave pulse has a single frequency component @xmath131 .
@xmath136 and @xmath137 are always referred to as in - phase part ( denoted as @xmath138 ) and quadrature part ( denoted as @xmath139 ) .
this allows the waveform to be conveniently written in a complex form of @xmath140 .
distortion caused by the finite bandwidth of the resonator and imperfections in microwave generation and transmission makes @xmath141 different from the intended waveform @xmath142 .
the ratio @xmath143 is the value of the transfer function @xmath67 at offset frequency @xmath144 .
therefore , @xmath67 can be obtained by measuring @xmath143 at different offset frequencies .
equation ( [ rotateh ] ) shows that in the rotating frame with frequency @xmath131 , the spin polarization is rotating at the frequency @xmath145 around the axis @xmath146 ( fig .
[ sphere ] ) .
the initial state of @xmath87 evolves as : @xmath147 where @xmath148 and @xmath149 . in our experiments , it is more convenient to work in the rotating frame of frequency @xmath150 , where the evolution of expectation values @xmath151 and @xmath152 can be expressed as : @xmath153 we applied constant amplitude pulses at frequency @xmath131 and fitted the experimental oscillations of @xmath151 and @xmath152 to eqs .
( [ sigmax ] ) and ( [ sigmay ] ) ( see fig . [ rabi ] ) to extract @xmath135 and @xmath134 . by varying @xmath131
, we can obtain estimates of @xmath143 at different frequencies .
there are two issues of the method described above that need to be addressed .
first , the method assumes the input microwave pulse with constant amplitude and single frequency component @xmath131 .
however , in real systems the pulses will have a rising and falling at the beginning and end , during which the spin experiences varying microwave field . if the rising and falling time is very short compared to the total pulse length , the effect can be neglected to a good approximation . in our system , due to the finite bandwidth of the resonator as well as limitations in the transmission components ( _ e.g. _ , iq mixer and amplifier ) ,
the rising and falling time of a pulse is around 5@xmath11410 ns , which is non - negligible for short pulses . to reduce the effect of the rising and falling of the pulse , we used relatively long pulses ( > 120 ns ) . by fitting rabi oscillations
assuming the pulses are perfect pulses with constant amplitude , we obtained an initial guess for the transfer function ( denoted as @xmath154 ) . in the next fitting iteration , we simulated the rabi oscillations due to the pulses distorted by @xmath154 . in this way
, we minimize the effect of imperfect pulses and obtained a refined transfer function .
another issue is that the fitting results of the phase part of the transfer function , @xmath155 , strongly depend on the choice of the starting time point @xmath156 . from eq .
( [ hamiltonian ] ) , it is not difficult to prove that with a temporal shift @xmath157 relative to @xmath156 , @xmath155 will get an additional slope which is @xmath158 . due to the distortion of the pulses , there is uncertainty in determining @xmath156 . to compensate , we used the refined fitting result of the transfer function as an initial guess to modify our optimal control ( oc ) pulses and measured the average gate fidelity using randomized benchmarking ( rb ) .
we then performed a few iterations of feed - back control where we slightly adjusted the slope of the phase of the transfer function ( and our oc pulses accordingly ) until we maximized our average gate fidelity and got the measured transfer function @xmath72 as shown in fig .
[ transfer ] .
the gate set tomography ( gst ) results of unmodified oc pulses , the oc pulses modified by @xmath72 , and the oc pulses modified by @xmath159 are given in table [ gst ] .
the table clearly shows both the amplitude and phase of the transfer function are important in improving the fidelities of the oc pulses .
it should be noted that @xmath72 becomes inaccurate when the offset frequency @xmath144 is large , as the fitting of the experimental rabi oscillations is less sensitive to the microwave pulse amplitude and phase when @xmath160 is small .
however , because the excitation bandwidths of our oc pulses are about 100 mhz or less , the imperfections of @xmath72 for @xmath161 larger than the bandwidths does not affect the controls of the oc pulses significantly . from the experimental results shown in the main text and the results listed in table [ gst ] , correcting the oc pulses using @xmath72 obtained by the method described above improves the control fidelities greatly .
therefore , we conclude that our @xmath72 is a good estimate of the system transfer function @xmath67 within the bandwidths of the oc pulses . , @xmath162 , @xmath163 ) of the spin polarization under the hamiltonian in eq .
( [ rotateh]).,scaledwidth=25.0% ] and @xmath152 at offset frequency @xmath164 mhz.,scaledwidth=42.0% ] , as a function of offset frequency relative to the resonant frequency @xmath150=9.996 ghz.,scaledwidth=40.0% ]
figure [ schematic ] shows the schematic diagram of the home - built pulsed x - band electron spin resonance ( esr ) spectrometer used for the experiments presented in the paper .
first , the signal source ( rhode @xmath165 schwarz ) generates continuous - wave reference microwave signal of @xmath11410 ghz , which is split into two by the two - way splitter ( marki ) . for the transmission ,
pulses with desired amplitudes and phases are constructed by up - converting the reference using the iq mixer ( marki ) and the arbitrary waveform generator ( tektronix ) which has maximum resolution of 1 ns . typical intermediate frequency ( if ) fed into the iq ports for the up - convertion is from 150 to 200 mhz .
next the pulses enter the microwave resonator located at the center of the water - cooled electromagnet through a series of attenuator ( advanced technical materials ) and amplifiers ( pre - amplifier : miteq ; pulsed traveling wave tube amplifier : applied system engineering ) to achieve appropriate amplifications , followed by circulators ( ditom ) to minimize unwanted reflections .
spin signals are directed to the receiver which consists of the limiter ( pasternack ) , fast pin didoe switch ( advanced technical materials ) , filters ( mini - circuits ) , low - noise amplifier ( miteq ) , circulator ( ditom ) , and front - end receiver mixer ( miteq ) .
the limiter and pin switch are implemented to prevent too much power flooding into the low - noise amplifier , and the high - pass filter ( mini - circuits ) blocks the low - frequency transient responses of the switching on and off .
the circulator also protects the low - noise amplifier from unwanted reflections as well as possible leakage of the reference through the receiver mixer .
finally another filtering stage ( mini - circuits ) removes artifacts from the receiver mixer , and the spin signals , down - converted to the same if frequency used in the up - conversion scheme , are captured by a fast digital oscilloscope ( lecroy ) .
further signal processing , _
i.e. _ , down - convertion to d.c . ,
is performed by a computer .
more details about the spectrometer can be found in ref .
as we reported in ref .
@xcite , the raw output from the pulsed traveling wave tube ( twt ) amplifier in our setup can be non - linear ( both in amplitude and phase ; mainly due to heating of the amplifier ) across the desired pulse sequence period . in rb and purity benchmarking ( pb ) experiments where we extend the sequences nearly to the maximum pulse sequence duration ( @xmath11415@xmath62s ) limited by the pulsed twt amplifier
, we corrected the non - linear amplitude fluctuation and phase droop in order to achieve the best control fidelities .
the non - linearity was corrected using a technique similar to how the transfer function was used to pre - distort the oc pulses .
we first measure the twt amplifier s output when a long square pulse was input in order to observe the amplitude non - linearity and phase droop as a function of time .
we can feed back this information to pre - distort the input pulse sequences . after this process
, the majority of non - linearity and phase droop is corrected as shown in fig .
[ twtcorrection ] .
s after unblanking the twt is employed to allow the output to stabilize ( not shown).,scaledwidth=42.0% ] m1.2 cm m1.2 cm m1.2 cm m1.2 cm m1.2 cm m1.2 cm & & & & & + & & & & & + & & & & & + & & & & & + & & & & & + & & & & & + & & & & & +
the larmor frequency distribution and @xmath77 field distribution , measured by detecting the spin signals , are the properties of the combined system of the sample and the resonator ( figs .
[ spectrum ] and [ b1 ] ) .
in particular , the asymmetric shape of the larmor frequency distribution is mainly due to the anisotropy of the @xmath167-value of the electron spins in the irradiated fused quartz sample .
pulse that is not robust to larmor frequency distribution is repeated eight times and each separated by an order of @xmath6 before the spin echo readout.,scaledwidth=42.0% ] distribution profile obtained by taking the fourier transform of the experimental rabi oscillations , and refined by comparing the experimental and simulated rabi oscillations @xcite .
the fwhm is about 1.7 mhz.,scaledwidth=45.0% ]
we designed three oc pulse shapes to realize @xmath73 and @xmath74 rotations around @xmath75 and @xmath76-axes and an identity operation .
different rotation axes for the @xmath73 and @xmath74 pulses were realized through phase shifting .
x90 and y90 denote a @xmath73 pulse about the @xmath75- and @xmath76-axes , x180 and y180 denote a @xmath74 pulse about the @xmath75- and @xmath76-axes , and @xmath9 denotes an identity pulse .
we used the gradient ascent pulse engineering algorithm @xcite to design oc pulses . in our system
, @xmath168 is the shortest time scale at which decoherence occurs . to make the oc pulses robust to @xmath0 noise , the weighted average fidelity over the larmor frequency distribution ( fig . [ spectrum ] ) is numerically optimized .
furthermore , oc pulses are made to be robust to the inhomogeneity in the microwave field ( fig . [ b1 ] ) using the same method .
table [ numerical ] gives the weighted average fidelity ( eq .
( 1 ) ) of each oc pulse under different numerical simulation conditions ( @xmath5 and @xmath6 effects are not considered here ) , and it shows the ideal oc pulses should be very robust to @xmath0 noise and power mis - calibration .
however , in the experiments , when the oc pulses arrive at the spins , they are distorted by the system transfer function . to investigate the effect of the system transfer function on the oc pulse fidelities
, we use @xmath169 to approximately simulate the distorted oc pulses , and the fidelities are presented in table [ numerical ] .
it shows that the coherent error of the pulses increases greatly , and the incoherent error also increases because the pulses lose part of the designed robustness to the @xmath0 noise and @xmath77 field distribution due to the transfer function .
the experimental fidelities of the oc pulses are given in table [ gst ] .
after using the measured transfer function @xmath72 to modify the pulses to compensate the effect of the system transfer function in experiment , despite the imperfection of @xmath72 , the pulses remain somewhat robust to these inhomogeneities .
m1.5 cm m1.2 cm m1.2 cm m1.2 cm m1.2 cm m1.2 cm & & & & & + & & & & & + & & & & & + & & & & & + & & & & & + & & & & & + is @xmath87 and the measurements @xmath170 are spin echo detection sequences for measuring @xmath171 . in @xmath170 ,
the @xmath74 pulse is a 35 ns gaussian pulse around the _ y_-axis , and @xmath111=700 ns represents a delay .
the @xmath74 pulse acts as the refocusing pulse .
the spin echo appears at time @xmath111 after the refocusing pulse and the spin magnetization in the @xmath75-@xmath76 plane is detected inductively.,scaledwidth=25.0% ]
m1.2 cm m0.6 cm m1.6 cm m1.5 cm m1.5 cm m1.5 cm & & & & & + [ 2ex ] & & & & & + [ 2ex ] & & & & & + & & & & & + & & & & & + & & & & & + & & & & & + the rb and pb protocols assume that the noise has little or no gate dependence , which is likely violated in real systems . to study the robustness of rb and pb to gate - dependent noise
, we also implement gst @xcite to reconstruct the process matrices of the oc pulses and use them to compare with the benchmarking results .
compared to standard quantum process tomography ( qpt ) , gst is more robust to state preparation and measurement ( spam ) errors and thus provides more accurate information about the target gates .
gst is implemented by applying three operations @xmath172 \{x90 , y90 , x180 , y180 , @xmath9 } to @xmath173 and measuring @xmath174 and @xmath102 ( see fig .
[ gst ] ) .
the pauli process matrices @xcite of the oc pulses are then estimated by minimizing the variance between the estimated and experimental values of @xmath174 and @xmath102 @xcite .
let @xmath175 = \{@xmath176 , @xmath177 , @xmath178 , @xmath179 , @xmath180 } denote the pauli process matrices of the experimentally - implemented \{x90 , y90 , x180 , y180 , @xmath9 } pulses and @xmath181 denote the expectation value of @xmath182 after applying @xmath183 .
we estimate the pauli process matrices by measuring the expectation values for each of the 125 combinations of the oc pulses and minimizing @xmath184 under the constraint that the noise is completely - positive and trace - preserving ( cptp ) @xcite . here ,
@xmath185 and @xmath186 are the vector forms of the measurement operator @xmath187 ( @xmath188 ) and initial state @xmath189 , whose elements are @xmath190 $ ] and @xmath191/2 $ ] with @xmath192 ( @xmath193 ) .
the estimated process matrices for the oc pulses are given in the next section and their fidelities are given in table [ gst ] .
we then use these measured process matrices to construct the process matrices of clifford gates from @xmath80 ( @xmath81 } , @xmath82 , and @xmath83 ) and calculate the benchmarking average error per gate ( bepg ) @xmath10 , which are listed in the ` gst estimates a ' column of table [ tab : parameters ] .
we found that the calculated @xmath10 values are closer to the fitting results of the first few gates of the experimental data within error bars than the fitting results obtained by including all gates in the rb and pb sequences ( _ i.e. _ values listed in table i of the main text ) .
we believe this is due to the gst sequences containing only three gates , which are not appropriate to predict the asymptotic behaviour of the rb and pb sequences ( also see fig . [ gstrbdiff ] )
. between experimental rb results and gst simulations of the expectation values over 150 random sequences .
the results of the three conditons , using the unmodified oc pulses , the transfer function modified oc pulses and the transfer function modified oc pulses with sel , are presented by green triangles , orange diamonds , and purple circles , respectively .
the increase of @xmath194 with the sequence length can be attributed to the fact that gst was performed using sequences of three pulses , which limits the accuracy . up to the first few gates ,
the gst simulations agree closely with each individual sequence of the rb experiments for the modified oc pulses , but demonstrate substantial disagreement for the unmodified pulses .
we attribute this disagreement to inaccuracies of the gst results in characterizing the large coherent errors in the individual gates resulting from the uncorrected distortion . for the oc pulses modified by the transfer function with sel , the @xmath194 values are unexpectedly higher ( compared to no sel ) for the first few gates .
we attribute this mainly to a smaller signal - to - noise ratio in the case of sel .
note that the roughly linear increases in @xmath194 versus sequence length suggest that discrepancies are accumulating stochastically rather than coherently . ]
we also use two different methods to calculate the incoherent error per gate ( iepg ) @xmath195 using the process matrices derived from gst .
the pb protocol has only been analyzed under the assumption that the noise is gate - independent .
one immediate question is whether the decay parameter of pb should be the average of the unitarities of the gate - dependent noise , @xmath196,\end{aligned}\ ] ] or the unitarity of the average noise , @xmath197.\end{aligned}\ ] ] these two quantities are not equal in general , as the unitarity is not a linear function of quantum channels .
the iepg for both quantities are listed in columns ` gst estimates a ' and ` gst estimates b ' of table [ tab : parameters ] .
although the table shows that @xmath198 is closer to the experimentally observed decay than @xmath199 , the error bars are large and we can not get statistically meaningful conclusion whether the pb protocol measures the average of the incoherent errors of the gate - dependent noise or the incoherent error of the noise averaged over the gates being benchmarked .
therefore , we use the help of simulation to further study the performance of pb and rb .
we use the experimental gst process matrices as realistic noise models to simulate rb and pb with the same sequences used in the experiment ( up to 55 clifford gates ) .
the derived decay rates of the simulations are listed in the column ` gst simulation of rb / pb ' of table [ tab : parameters ] . the values in the ` gst simulation of rb / pb ' and in the ` gst es- timates a ' agree to within the fitting uncertainty ( note that the uncertainties in the ` gst estimates a ' and ` gst estimates b ' are not relevant as our simulations are of the average gst reconstructions ) , indicating that under realistic noise models consistent with our system , both the rb and pb work very well in estimating the bepg and iepg . furthermore , our simulations suggest the pb protocol measures the average of the incoherent errors of the gate - dependent noise , which is sensitive to individual calibration errors ( even if these gate - dependent calibration errors average into an incoherent process ) .
here we list the reconstructed pauli transfer matrices obtained from the gst experiments .
there are three cases : ( i ) oc pulses are modified by @xmath72 and sel is used ; ( ii ) oc pulses are modified by @xmath72 and no sel ; ( iii ) unmodified pulses and no sel . in case ( i ) : @xmath200 @xmath201 @xmath202 @xmath203 @xmath204 in case ( ii ) : @xmath205 @xmath206 @xmath207 @xmath208 @xmath209 in case ( iii ) : @xmath210 @xmath211 @xmath212 @xmath213 @xmath214
we generated 150 random sequences for each length of clifford gates for the rb and pb experiments , up to 55 clifford gates .
each sequence had randomly selected clifford gates such that the clifford group was sampled uniformly .
for the experiment , we only used clifford gate sequences of length @xmath89=1,@xmath215 , 7 , 9 , 11 , 16 , 19 , 23 , 28 , 30 , 33 , 36 , 39 , 40 , 44 , 45 , 48 , 49 , 53 , 55 for the rb without sel ; length @xmath89=1,@xmath215 , 7 , 9 , 14 , 17 , 20 , 22 , 25 , 27 , 31 , 35 , 37 , 40 , 41 , 42 , 45 , 49 , 54 , 55 for the pb without sel ; length @xmath89=1,@xmath215 , 7 , 10 , 13 , 14 , 16 , 17 , 21 , 23 , 26 , 27 , 31 , 33 , 37 , 41 , 43 , 45 for the rb with sel ; length @xmath89=1,@xmath215 , 8 , 9 , 13 , 17 , 18 , 19 , 22 , 27 , 29 , 32 , 34 , 38 , 40 , 43 , 47 for the pb with sel .
longer sequences could not be sampled mainly due to the limitation of the twt amplifier .
we also numerically simulate our benchmarking experiments .
the simulation evolves an initial density matrix according to the linblad model using the same pulse sequences as in the experiments .
experimentally measured @xmath5=160 @xmath216 and @xmath6=30 @xmath216 are used in the simulation .
we incorporate the effects of @xmath0 and local @xmath77 field inhomogeneities by averaging the simulation over these distributions ( see fig .
[ spectrum ] for the larmor frequency distribution , which corresponds to the @xmath0 effect , and see fig .
[ b1 ] for the local @xmath77 distribution ) .
the sel sequence is used in the experiments to effectively extend @xmath21780 ns .
this sequence is comprised of eight 2@xmath74 rotation pulses , each separated by an order of @xmath6 to allow the transverse components to dephase @xcite .
this sel sequence selects a subset of the spin ensemble within narrower larmor frequency distribution while dephasing most of the off - resonance spin packets as shown in fig .
[ spectrum ] , but we found that it does not have much effect on the @xmath77 distribution .
the sel step is incorporated in the simulation by using a distribution that matches the larmor frequency distribution measured in the experiments with sel sequences ( @xmath218160 ns ) .
figures [ fig_rb_sim ] and [ fig_unit_sim ] show the simulation results of rb and pb protocols when taking and not taking sel sequences into account .
these simulations only take into account @xmath0 distribution since we simulated that our measured @xmath77 field inhomogeneity has little effect on our bepg ( @xmath219 ) .
the values from the simulation with no sel sequences are @xmath220 and @xmath221 .
the values from the simulation using the larmor frequency distribution with sel sequences are @xmath222 and @xmath223 .
as simulations assume no pulse distortions ( _ i.e_. , perfect knowledge of @xmath67 ) , the only coherent error source in the simulations is the imperfection in the oc pulse design which is very small ( see table [ numerical ] ) . in the case with short @xmath0 ,
the simulated decays slightly deviate from a single exponential decay , while the simulated decays in the case with long @xmath0 can be fitted very well to a single exponential decay up to @xmath114 50 gates , which agrees with the experimental results .
this indicates the @xmath0 effect contributes to the non - markovian noise .
160 ns and @xmath0 @xmath114 80 ns , respectively .
the solid and dotted lines are least squares fits to @xmath224 where @xmath10 is the bepg for the clifford group .
the error bars indicate the standard error of the mean.,scaledwidth=45.0% ] @xmath114 160 ns and @xmath0 @xmath114 80 ns , respectively .
the solid and dotted lines are least squares fits to @xmath225 where @xmath11 is the iepg for the clifford group .
the error bars indicate the standard error of the mean.,scaledwidth=45.0% ] | to exploit a given physical system for quantum information processing , it is critical to understand the different types of noise affecting quantum control .
distinguishing coherent and incoherent errors is extremely useful as they can be reduced in different ways .
coherent errors are generally easier to reduce at the hardware level , e.g. by improving calibration , whereas some sources of incoherent errors , e.g. @xmath0 processes , can be reduced by engineering robust pulses . in this work ,
we illustrate how purity benchmarking and randomized benchmarking can be used together to distinguish between coherent and incoherent errors and to quantify the reduction in both of them due to using optimal control pulses and accounting for the transfer function in an electron spin resonance system .
we also prove that purity benchmarking provides bounds on the optimal fidelity and diamond norm that can be achieved by correcting the coherent errors through improving calibration . a key obstacle to realizing scalable quantum information processing ( qip )
is implementing quantum gates sufficiently precisely so that errors can be detected and corrected @xcite .
this requires both the intrinsic noise and the noise in the control to be characterized .
the combined noise can be completely characterized using either quantum process tomography ( qpt ) @xcite or gate set tomography ( gst ) @xcite . however , these methods are time - consuming and scale exponentially in the number of qubits . instead of completely characterizing a system
, we can efficiently quantify how noisy the experimental operations are .
the most prominent method along these lines is randomized benchmarking ( rb ) @xcite , which gives an efficient estimate of the benchmarking error per gate ( bepg ) defined as @xmath1 where @xmath2 is the noise channel and the integral ( the channel fidelity @xmath3 ) is over all pure states @xmath4 according to the haar measure .
however , the bepg is , by construction , insensitive to many of the particular details of the noise mechanism . as errors due to different noise mechanisms
can be corrected in different ways and have different impacts on qip , understanding the noise characteristics in quantum systems is of critical importance .
noise characteristics can be broadly grouped as either coherent ( unitary ) or incoherent ( statistical ) .
coherent noise is usually due to systematic control errors in , for example , imperfect rotation angles or axes @xcite , which may be easier to reduce than incoherent noise such as @xmath5 and @xmath6 processes .
the bepg for coherent noise accumulates quadratically with the number of gates whereas incoherent noise accumulates linearly .
furthermore , coherent and incoherent noises with the same bepg may lead to dramatically different thresholds as quantified by the worst - case error per gate ( wepg ) , also known as the diamond distance , @xcite @xmath7\otimes \mathcal{i}(\psi)\|_1,\end{aligned}\ ] ] where @xmath8 and @xmath9 is the identity channel acting on an ancillary system of the same size to account for the effect of the noise on entangled inputs .
therefore , identifying whether the noise is primarily coherent or incoherent is essential for determining an appropriate error threshold when evaluating a physical system and for determining whether experimental effort should prioritize improving calibration or suppressing incoherent error processes .
several approaches have been developed to provide more information about the noise than just the bepg while retaining the advantages of rb @xcite .
in particular , purity benchmarking ( pb ) @xcite enables the quantification of the coherence of a noise process without assuming a specific noise model , which can be used to obtain an improved estimate of the wepg @xcite , whereas the method of ref .
@xcite detects additive coherent errors under specific assumptions about the noise model . in this paper
, we show that pb can be used to quantify the best achievable bepg and wepg under optimal control for single - qubit systems .
we then test pb in a specific modality , namely , a solid - state electron spin resonance ( esr ) system .
bulk esr samples consist of an ensemble of ( nearly ) identical spins , which can mimic the behaviour of a fixed number of qubits depending on the structure of the solid and the species of the spins .
esr provides one path to scalable qip using techniques such as algorithmic cooling and distributed node quantum information processing @xcite , which are viable because electron spins have larger thermal polarization and faster relaxation rates than nuclear spins , and hyperfine - coupled nuclear spins can also be efficiently controlled using esr techniques @xcite .
the quantum control techniques developed in qip are also very useful for modern esr spectroscopy @xcite .
achieving high fidelity quantum control in esr is challenging due to the limited bandwidth of a conventional microwave resonator . in this work , rb and pb protocols
are used to assess the control accuracy of an ensemble single - qubit system .
we demonstrate the reduction in both the coherent and incoherent errors obtained by first using the transfer function of the microwave control system to correct numerically - derived optimal control ( oc ) pulses @xcite and then using a spin - packet selection technique to effectively reduce the inhomogeneous spectral broadening @xcite .
the lowest values we obtained for bepg ( @xmath10 ) and the incoherent error ( @xmath11 , defined below ) for clifford gates are @xmath12 and @xmath13 , respectively .
_ _ the incoherent error per gate__the primary characteristic of a coherent noise process is that it can be corrected by directly reversing the unitary process with perfect control .
we therefore define the incoherent error per gate ( iepg ) of a noise channel @xmath2 to be the optimal bepg that can be achieved by correcting @xmath2 with perfect unitary operations , that is , @xmath14 for any unitary operations @xmath15 and @xmath16 . for a general @xmath17-dimensional system , the incoherent error satisfies @xmath18,\label{noneq}\end{aligned}\ ] ] where the unitarity is @xcite @xmath19 ^ 2.\end{aligned}\ ] ] we now show that the lower bound on the incoherent error in eq .
is saturated to @xmath20 $ ] in the single - qubit case .
let @xmath21/2 $ ] be the process matrix of @xmath2 , where @xmath22 .
the process matrix of any completely - positive and trace - preserving ( cptp ) noise channel can be written in block form as @xmath23 the unitarity and bepg of @xmath2 are @xmath24 any single - qubit noise channel can be corrected to another channel @xmath25 such that @xmath26 and @xmath27 for some @xmath28 and some real diagonal @xmath29 by applying suitable ( perfect ) unitary operators @xcite , which leaves the unitarity unchanged , that is , @xmath30 . by von neumann s trace inequality , @xmath31 for any unitary operations @xmath15 and @xmath16 , so @xmath32 .
writing @xmath33 where @xmath34 is nonnegative for any cptp map @xcite and @xmath35 from eq .
, we have @xmath36 for a single qubit .
the minimum and maximum values of @xmath37 subject to @xmath38 , @xmath35 , and the cptp constraints @xcite are 0 and 2 , attained when @xmath39 and @xmath40 respectively .
therefore the incoherent error for a single qubit satisfies @xmath41 to within @xmath42 as claimed .
now we consider the part of error that is removed by the optimal unitary corrections . with @xmath43 and @xmath44 , from eq .
( [ define2 ] ) the bepg of @xmath2 is @xmath45\end{aligned}\ ] ] where the order of the higher - order term comes from @xmath29 being diagonal and the diagonal elements of a generic cptp map @xmath46 being @xmath47 @xcite .
we can regard @xmath15 and @xmath16 as coherent errors and so the bepg of the ( composite ) coherent error is @xmath48 which is also equivalent [ to @xmath49 to the bepg removed by the optimal unitary corrections .
the iepg also provides an improved bound on the optimal wepg @xmath50 that can be achieved by applying unitary corrections .
let @xmath51 be the unital part of @xmath25 , that is , the channel such that @xmath52 for all @xmath53 .
we then have @xmath54 where @xmath55 is the partial trace over the first system . by the triangle inequality and submultiplicativity of the diamond norm , @xmath56 where the maximization is achieved by any maximally entangled state . as @xmath51 is a pauli channel @xcite , @xmath57 and with a lower bound on the wepg in terms of the bepg @xcite , we have @xmath58.\end{aligned}\ ] ] both these constraints are linear in @xmath59 and so give reasonable estimates as @xmath59 decreases compared to the gap between the optimal scalings for the lower and upper bounds in terms of @xmath60 alone , which diverge by orders of magnitude as @xmath60 decreases @xcite . _
_ experimental implementation__our x - band pulsed esr spectrometer was custom - built for qip experiments and includes arbitrary waveform generation and a loop - gap resonator for sub - millimeter sized samples that allows for relatively broadband control @xcite .
for an ensemble single - qubit system , we use a sample of gamma - irradiated fused quartz , a paramagnetic sample in powder form where the primary defect is a spin-1/2 unpaired electron at an oxygen vacancy @xcite , with @xmath61 @xmath62s , @xmath63 @xmath62s , and @xmath64 ns . a pulse generated with an initial waveform @xmath65 in the frequency - domain representation will be distorted to a new waveform @xmath66 seen by the spins due to the system s transfer function @xmath67 , which is the frequency - domain representation of the impulse response of the system @xcite , so that @xmath68 where @xmath69 denotes the point - wise product of @xmath67 and @xmath70
. the transfer function includes contributions from the resonator s transfer function and other imperfections in the pulse generation and transmission .
one method to correct @xmath66 is to distort the initial waveform to be @xmath71 .
the accuracy of this method is limited by the accuracy with which @xmath67 can be determined .
we measure @xmath67 by detecting rabi oscillations of the electron spins as a function of the microwave frequency @xcite .
this measured transfer function , denoted by @xmath72 , is then used to modify the input oc pulse so that the distorted pulse seen by the spins will approximate the desired waveform .
we use three oc pulses : @xmath73 and @xmath74 rotations ( denoted by x90 , x180 , y90 and y180 for rotations around the @xmath75- and @xmath76-axes respectively ) and an identity operation ( denoted by @xmath9 ) .
the pulses are each 150 ns long and designed to be robust to distributions of larmor frequency and microwave ( @xmath77 ) field that closely mimic the measured properties of the combined system of our sample and resonator @xcite .
the design fidelity of each pulse exceeds 99.7% when averaged over these distributions @xcite .
the experimental results span three different conditions for implementing the oc pulses : ( 1 ) not taking the system transfer function into account , _
i.e. _ , assuming @xmath78 for all frequencies , ( 2 ) modifying the input pulses using @xmath79 , and ( 3 ) the same as ( 2 ) but also implementing a spin - packet selection ( sel ) state preparation sequence @xcite which effectively increases @xmath0 by a factor of 2 .
we implement the 24 elements of the clifford group as @xmath80 where @xmath81 } , @xmath82 , and @xmath83 .
@xmath84 and @xmath85 are implemented using the numerically derived @xmath9 , @xmath73 and @xmath74 pulses and altering the phase as needed to achieve @xmath75- and @xmath76-axis rotations .
the operations in @xmath86 are implemented virtually by changing the reference frame @xcite .
the initial state in all experiments is represented by the deviation density matrix @xmath87 .
we can estimate the bepg and iepg averaged over the set of operations @xmath88 via rb and pb as follows @xcite ( see the quantum circuits in fig . [ circuit ] ) .
( 1 ) prepare the state @xmath87 .
( 2 ) apply a sequence of @xmath89 uniformly - random operations from @xmath88 , which maps @xmath87 to @xmath90 .
( 2.1 ) for rb , apply a recovery gate @xmath91 that maps @xmath90 back to @xmath92 . when the final state is @xmath93 we change the sign to be positive in post - processing , that is , implement a virtual x gate .
( 3 ) for rb , estimate the expectation value @xmath94 . for pb , estimate the purity @xmath95 of the final state @xmath90 .
averaging over random sequences of length @xmath89 and fitting to @xmath96 for rb and pb , respectively , under trace - preserving noise , allows @xmath10 and the unitarity @xmath97 ( and hence @xmath11 via eq . ) to be estimated where the constants absorb the state preparation and measurement ( spam ) errors and the non - unitality of the noise .
in particular , @xmath98 ( @xmath99 ) with @xmath100 where the summation is over the single - qubit clifford group and the equality follows from the fact that the clifford group is a unitary 2-design and hence is also a unitary 1-design @xcite .
we can therefore estimate both constant off - sets by performing a single clifford gate , measuring the expectation values of @xmath94 , @xmath101 , and @xmath102 and averaging over all clifford gates .
the expectation values are measured by the corresponding spin echo detection sequences in figs .
[ circuit](c ) and ( d ) .
we sample 150 random sequences for each sequence length @xmath89 of rb and pb independently .
is @xmath87 and the measurements @xmath103 are spin echo detection sequences for measuring @xmath94 for rb and @xmath104 for pb .
@xmath105 in ( a ) is the recovery gate that returns the state to @xmath92 .
a total of 150 random sequences with @xmath106 \{@xmath9 , x90 , y90 } and @xmath107 \{@xmath9 , x180 } ( and virtual @xmath108-axis rotations ) are applied for each sequence length @xmath89 for rb and pb .
( c ) and ( d ) are the spin echo detection sequences for measuring @xmath109 and @xmath110 , respectively .
the @xmath73 and @xmath74 pulses are 35 ns gaussian pulses around the _
y_-axis , and @xmath111=700 ns represents a delay.,scaledwidth=48.0% ] _ _ results and discussion__the results of the rb and pb experiments are presented in fig .
[ regular ] , with the corresponding estimates for the bepg , iepg , coherent error rate and optimal wepg listed in table [ tab : parameters ] .
l c24 mm c24 mm c24 mm & @xmath78 & @xmath112 & @xmath112 + & no sel & no sel & sel + @xmath10 & 0.0234(11 ) & 0.0073(2 ) & 0.0063(2 ) + @xmath11 & 0.0105(10 ) & 0.0066(2 ) & 0.0054(2 ) + @xmath113 & 0.0129(21 ) & 0.0007(4 ) & 0.0009(4 ) + @xmath50 & 0.040(26 ) & 0.024(15 ) & 0.020(12 ) + pulse distortion due to the system transfer function is significant , as the transfer function bandwidth of @xmath114100 mhz is comparable to the pulse excitation bandwidth .
the improvement between results from the unmodified oc pulses ( @xmath78 ) and those modified by taking into account @xmath79 in table [ tab : parameters ] demonstrates substantial reduction in @xmath113 from @xmath115 to @xmath116 .
the @xmath11 is also reduced by approximately a factor of two , from @xmath117 to @xmath118 .
this shows that the pulse distortion is non - negligible , and causes both coherent and incoherent errors .
the larger incoherent error for the unmodified oc pulses is largely due to the oc pulses losing their engineered robustness to larmor frequency and @xmath77 inhomogeneities when assuming @xmath78 . although the decay rates of both the rb and pb experimental results are substantially reduced by using @xmath72 to improve the oc pulses , the decays seem to deviate from a single exponential decay ( i.e. , see the oscillating deviations of the orange data points from the orange solid lines in fig . [ regular ] ) , implying the existence of non - markovian noise . in our system ,
the larmor frequency distribution for different spin - packets ( @xmath0 effect ) results in a significant non - markovian effect @xcite .
the benchmarking pulse sequences act like filters , in that the spectral line - width of the part of the spin - packet that contributes to the signal decreases with the number of gates .
this means the effective @xmath0 lifetime is not constant but increases with the number of gates that are implemented .
therefore , the error rates estimated using the single - exponential decay model are the averaged values over this non - constant noise .
lindblad numerical simulations ( where the @xmath0 process is simulated by averaging over multiple simulations with different larmor frequencies ) give non - exponential decays for rb and pb @xcite , agreeing with our experimental results . to reduce the non - markovianity due to @xmath0
, we implement sel sequences before each of the benchmarking sequences , which selects a narrower line - width so the benchmarking experiments have a longer @xmath0 ( @xmath114 160 ns ) to begin with @xcite . after incorporating the sel sequences ,
the experimentally observed decays fit to a single exponential better ( see the purple data points and purple dotted lines in fig . [ regular ] ) .
the lindblad simulation results with the longer @xmath0 also exhibit single exponential decays up to @xmath114 50 gates @xcite .
using the sel sequence improves @xmath119 from @xmath120 to @xmath121 , but has no statistically significant effect on @xmath122 , which is @xmath123 and @xmath124 with and without sel , respectively .
this implies the @xmath0 effect mainly contributes to the incoherent error . in the lindblad simulations of the benchmarking sequences using the extended @xmath0 , @xmath119 caused by @xmath5 , @xmath6 , and @xmath0
is @xmath125 , and @xmath126 caused by the imperfection in the oc pulse design is @xmath127 @xcite .
we attribute the discrepancy between the simulated and experimental values of @xmath119 and @xmath126 to possible inaccuracy in the measured decoherence times , fluctuations in the control mechanisms , and imperfect knowledge of the transfer function . _
conclusions _ we have demonstrated how rb and pb can be used together to go beyond quantifying average gate fidelities by distinguishing coherent and incoherent contributions to the error .
this allows improvements in calibration and engineering pulses to suppress incoherent errors to be implemented and diagnosed independently
. pulse distortion due to the system transfer function @xmath67 is the dominant error source in our system and contributes greatly to the coherent part of the gate error .
our measurement of @xmath67 helps improve the oc pulse fidelities significantly .
the incoherent error is primarily due to @xmath5 , @xmath6 and @xmath0 processes . by effectively extending @xmath0
we reduce the non - markovian effect and improve the control fidelity further .
results from gate set tomography included in the supplemental material indicate that our system has substantial gate - dependent noise .
the pb protocol has only been analyzed under the assumption of gate - independent noise .
simulations using the estimates from gate set tomography indicate that pb can distinguish between gate - dependent coherent errors that look incoherent when averaged over the gates and a gate - independent incoherent process , at least for some physically - realistic error models .
however , we leave the general behavior of pb under gate - dependent noise as an open problem . * acknowledgement * this research was supported by nserc , the canada foundation for innovation , cifar , the province of ontario , industry canada , the gerald schwartz & heather reisman foundation , and the u.s . army research office through grant w911nf-14 - 1 - 0103 .
we thank david cory and troy borneman for providing the esr sample and for stimulating discussions ; colm ryan , yingjie zhang , and jeremy chamilliard for their contributions to the spectrometer ; ming lyu and ian hincks for the discussions on the transfer function measurement method ; robabeh rahimi for help with preparing the esr sample ; roberto romero and hiruy haile for help with machining ; hemant katiyar for help with the simulations .
10 e. knill , and r. laflamme , phys .
rev .
a 55 ( 2 ) , 900 ( 1997 ) .
e. knill , r. laflamme , and w. h. zurek , science 279 ( 5349 ) , 342345 ( 1998 ) .
j. preskill , proceedings of the royal society of london .
series a : mathematical , physical and engineering sciences 454 ( 1969 ) , 385410 ( 1998 ) .
e. knill , nature 434 ( 7029 ) , 3944 ( 2005 ) .
p. aliferis , d. gottesman , and j. preskill , quantum information & computation 8 ( 3 ) , 181244 ( 2008 ) .
d. gottesman , stabilizer codes and quantum error correction .
caltech ph .
d , ph.d .
thesis , thesis , eprint : quant - ph/9705052 ( 1997 ) .
i. l. chuang , and m. a. nielsen , j. mod .
opt .
44 , 2455 ( 1997 ) .
j. f. poyatos , j. i. cirac , and p. zoller , phys .
rev .
lett .
78 , 390 ( 1997 ) .
s. t. merkel , j. m. gambetta , j. a. smolin , s. poletto , a. d. crcoles , b. r. johnson , c. a. ryan , and m. steffen , phys .
rev .
a 87 , 062119 ( 2013 ) .
r. blume - kohout , j. k. gamble , e. nielsen , j. mizrahi , j. d. sterk , and p. maunz , arxiv:1310.4492 ( 2013 ) .
s. kimmel , m. p. da silva , c. a. ryan , b. r. johnson , and t. ohki , phys .
rev .
x 4 , 011050 ( 2014 ) .
j. emerson , r. alicki , and k. zyczkowski , journal of optics b : quantum and semiclassical optics 7 ( 10 ) , s347 ( 2005 ) .
e. knill , d. leibfried , r. reichle , j. britton , r. b. blakestad , j. d. jost , c. langer , r. ozeri , s. seidelin , and d. j. wineland , phys .
rev .
a 77 , 012307 ( 2008 ) . c. dankert , r. cleve , j. emerson , and e. livine , phys .
rev .
a 80 ( 1 ) , 012304 ( 2009 ) .
e. magesan , j. m. gambetta , and j. emerson , phys .
rev .
lett .
106 , 180504 ( 2011 ) .
e. magesan , j. m. gambetta , and j. emerson , phys .
rev .
a 85 , 042311 ( 2012 ) .
j. j. wallman , and s. t. flammia , new j. phys .
16 , 103032 ( 2014 ) .
j. j. l. morton , a. m. tyryshkin , a. ardavan , k. porfyrakis , s. a. lyon , and g. a. d. briggs , phys .
rev . a 71 , 012332 ( 2005 ) .
s. kimmel , g. h. low , and t. j. yoder , phys .
rev .
a 92 , 062315 ( 2015 ) .
y. r. sanders , j. j. wallman , and b. c. sanders , new j. phys .
18 , 012002 ( 2015 ) .
j. j. wallman , m. barnhill , and j. emerson , phys .
rev .
lett .
115 , 060501 ( 2015 ) .
p. j. j. omalley _ et al .
_ , phys .
rev .
appl . 3 , 044009 ( 2015 ) .
j. j. wallman , c. granade , r. harper , and s. t. flammia , new j. phys .
17 , 113020 ( 2015 ) .
s. sheldon _ et al .
_ , phys .
rev .
a 93 , 012301 ( 2016 ) .
j. j. wallman , arxiv:1511.00727 ( 2015 ) .
r. kueng , d. m. long , a. c. doherty , and s. t. flammia , arxiv:1510.05653 ( 2015 ) .
t.w .
borneman , c.e .
granade , d.g .
cory , phys .
rev .
lett .
108 ( 14 ) , 140502 ( 2012 ) .
j. s. hodges , j. c. yang , c. ramanathan , and d. g. cory , phys .
rev .
a 78 ( 1 ) , 010303 ( 2008 ) .
y. zhang , c. a. ryan , r. laflamme , and j. baugh , phys .
rev .
lett .
107 ( 17 ) , 170503 ( 2011 ) .
d. k. park , g. feng , r. rahimi , s. labruyre , t. shibata , s. nakazawa , k. sato , t. takui , r. laflamme , and j. baugh , quantum information processing 14 ( 7 ) , 24352461 ( 2015 ) .
t. prisner , m. rohrer , and f. macmillan , annu .
rev .
phys .
chem .
52 , 279 ( 2001 ) .
p.e .
spindler , y. zhang , b. endeward , _ et al .
_ , j. magn .
reson .
218 , 49 ( 2012 ) .
n. khaneja , t. reiss , c. kehlet , t. schulte - herbrggen , and s. j. glaser , journal of magnetic resonance 172 , 296305 ( 2005 ) . d. k. park , g. feng , r. rahimi , j. baugh , and r. laflamme , journal of magnetic resonance 267 , 68 ( 2016 ) .
c. king , and m. b. ruskai , j. math .
phys .
42 , 87 ( 2001 ) .
m. b. ruskai , s. szarek , and e. werner , linear algebr .
appl .
347 159 ( 2002 ) .
s. eaton , and g. eaton , journal of magnetic resonance , series a 102 354 356 ( 1993 ) .
t. kaufmann _
et al . _ , j. magn .
reson .
235 , 95 ( 2013 ) .
s. gustavsson , o. zwier , j. bylander , f. yan , f. yoshihara , y. nakamura , t. p. orlando , and w. d. oliver , phys .
rev .
lett .
110 , 040502 ( 2013 ) .
see supplemental material , which includes ref .
@xcite , for details of the transfer function measurement , optimal control pulse design , gate set tomography results , and simulations . j. m. chow _
et al .
_ , phys .
rev .
lett .
109 , 060501 ( 2012 ) . c. a. ryan , m. laforest , and r. laflamme , new j. of phys . 11 ( 1 ) , 013034 ( 2009 ) . m. a. fogarty , m. veldhorst , r. harper , c. h. yang , s. d. bartlett , s. t. flammia , and a. s. dzurak , phys . rev . a 92 , 022326 ( 2015 ) .
url # 1`#1`urlprefixhref # 1#2#2 # 1#1 |
radial velocity measurements of the non degenerate donors in x ray transients at quiescence combined with information about donor spectra and optical light curves allow us to measure the masses of accreting compact objects ( e.g. , charles 1998 ) , as well as other binary properties ( e.g. , orbital periods , donor masses and spectral types , kinematic properties ) . at present , measured masses for nine x ray transients exceed the optimum maximum neutron mass of @xmath1(e.g .
, kalogera & baym 1996 ) and the binaries are thought to harbor black holes ( bh ) .
these observed bh x ray transients seem to form a rather homogeneous sample and studies of its properties as a whole can shed a light to their evolutionary history and the process of bh formation .
black hole x ray transients are similar to low mass x
ray binaries with neutron stars in that mass transfer is driven by roche lobe overflow and the donors are less massive than the bh
. this maximum mass ratio of about unity ( see kalogera & webbink 1996 ) allows the donor to transfer mass stably to the compact object .
however , for the majority of the bh x ray transients ( including six bh candidate systems based on their spectral properties , see chen , shrader , & livio 1997 ) the donors are less massive than @xmath2 , much less massive than the typical bh masses .
only two systems , j1655 - 40 and 4u1543 - 47 , have intermediate mass donors ( more massive than @xmath3 ) of @xmath4 and @xmath5 , respectively .
further , orbital periods are found in the range of a few hours to days , much smaller than orbital periods of binary progenitors that could accommodate the radial expansion of evolved massive stars ( bh progenitors ) . in what follows ,
we consider a number of the observed characteristics of bh x ray transients and use them as clues and windows to the evolutionary history of these binaries , bh formation , and helium star evolution .
currently observed orbital periods vary from just a few hours to a couple of days typically ( with the exception of v404 cyg , which has an orbital period of about 6 days ) .
these orbital periods along with the measured bh masses and the estimated donor masses ( based on spectral type classifications ) clearly indicate that the orbital separations of the x - ray transients are of order @xmath6(even for v404 cyg where the donor is on the giant branch , we derive @xmath7 ) . since evolved massive stars typically expand to @xmath8(even for stars that lose their hydrogen envelopes though winds , see schaller et al .
1992 ) , the progenitors of the observed binaries must have experienced a drastic orbital contraction . based on this simple observation it is generally accepted that a common - envelope phase is necessary to explain the current short orbital periods .
we can look into this question in some more detail .
let us for a moment assume that the primordial binary was wide enough to avoid common envelope evolution .
then the only possible way to achieve the required orbital contraction is through natal kicks imparted to black holes .
the magnitude of these kicks may be quite modest , since the effects of kicks on binary characteristics are more important when their magnitude is comparable to orbital velocity of the binary . for massive binaries with orbital separations
well in excess of @xmath9 , orbital velocities lie in the range @xmath10kms@xmath11 .
however , the required degree of orbital contraction is rather large , with a ratio of circularized post supernova to pre supernova orbital separations : @xmath12 under the assumption of angular momentum conservation during circularization , it is : @xmath13 where @xmath14 , @xmath15 is the immediate ( non circularized ) post supernova separation , and @xmath16 is the post
supernova eccentricity .
further , a strict lower limit on @xmath17 can be derived analytically ( e.g. , flannery & van den heuvel 1975 ; kalogera 1996 ) : @xmath18 combining equations ( 1 ) , ( 2 ) , and ( 3 ) we derive that the required degree of post supernova orbital contraction can occur only for binaries with post
supernova eccentricities in excess of 0.99 ( ! ) .
analysis of post supernova binary characteristics , for cases of significant mass loss and kicks comparable to the pre supernova orbital velocity shows that only a negligible fraction ( less than 0.1% ) of surviving ( bound ) systems acquire eccentricities that high ( see figures 4 and 10 in kalogera 1996 ) .
such a low survival probability in the absence of a common envelope phase would lead to a negligible formation rate for bh x ray transients ( see also portegies zwart , verbunt , & ergma 1997 ) .
therefore , we conclude that a common envelope phase is _ necessary _ for the formation of all of the observed bh x ray transients .
given that the progenitors of bh binaries must experience a common envelope ( ce ) phase , which leads to both orbital contraction and loss of hydrogen envelope of the bh progenitor , it is quite possible that the evolutionary history is very similar to that of low mass x
ray binaries with neutron stars ( e.g. , van den heuvel 1983 ) . in
what follows , we consider structural and evolutionary requirements for such a formation path . in the next section
we examine how these requirements determine the donor properties of the bh binaries formed ( for more details , see kalogera 1999 )
. the binary primary must be massive enough so that its helium core exposed at the end of the ce phase collapses into a bh .
the x ray phase is initiated when the donor fills its roche lobe because of orbital shrinkage through magnetic braking ( for low mass donors ) or of radial expansion through nuclear evolution on the main sequence ( for intermediate mass donors ) .
black hole binary progenitors evolve through this path provided that the following constraints are satisfied : * the orbit is small enough that the primary fills its roche lobe and the binary enters a ce phase .
* at the end of the ce phase the orbit is wide enough so that both the helium rich primary and its companion fit within their roche lobes .
the constraint for the companion turns out to be stricter . *
the system remains bound after the collapse of the helium star . in the case of small or zero kicks imparted to the bh , this sets an upper limit on the mass of the bh progenitor .
* after the collapse , the orbit must be small enough so that mass transfer from the donor starts before it leaves the main sequence and within 10@xmath19yr .
* mass transfer from the donor proceeds stably and at sub - eddington rates .
this sets an upper limit to the donor mass on the zero - age main sequence and to the orbital size for more evolved donors .
for a specific value of the bh mass , the above constraints translate into limits on the properties , circularized post
collapse orbital sizes ( @xmath20 ) and donor masses ( @xmath21 ) , of bh binaries with roche lobe filling donors .
the relative positions of these limits on the @xmath22 plane and the resulting allowed @xmath21 ranges are exactly determined by three well constrained model parameters : * the amount of mass loss from the binary during bh formation , characterized by the ratio @xmath23 , where @xmath24 is the mass of the helium rich bh progenitor at the time of the collapse . for the post
collapse system to remain bound it must be @xmath25 . * the amount of mass lost in the helium
star wind between the end of the ce phase and the bh formation , characterized by the ratio @xmath26 , where @xmath27 is the initial helium
star mass ( at the end of the ce phase ) .
this ratio must lie in the range @xmath28 . * the ce efficiency , @xmath29 , defined as the ratio of the ce binding energy to the orbital energy released during the spiral in of the companion . although the absolute normalization of @xmath29 is not well determined ( see kalogera 1999 ) , values higher than unity imply the existence of energy sources other than the orbit ( ionization or nuclear burning energy ) .
note that the last two of the evolutionary constraints ( 3 ) depend _ only _ on the bh mass , while @xmath29 affects only the upper limit on @xmath20 ( first of the constraints 3 ) . for different values of these three parameters , the positions of the limits on the @xmath22 plane change and three different outcomes with respect to the donor masses are possible : bh binaries can be formed with ( i ) _ only _ low mass ; ( ii ) _ only _ intermediate mass ; ( iii ) _ both _
low and intermediate mass donors .
the donor types as a function of the three parameters , @xmath24 , @xmath26 , and @xmath30 , are shown in fig . 1 , for a 7 bh . for @xmath30
smaller than @xmath31 , the orbital contraction is so high that the donor stars can not fit in the post ce orbits , and hence no bh x - ray binaries are formed . as @xmath30 increases ,
ce ejection without the need of strong orbital contraction becomes possible for the more massive of the donors , while formation of binaries with low
mass donors occurs only if @xmath32 .
the results become independent of @xmath30 for values in excess of @xmath33 , when the upper limit for ce evolution ( first of the constraints ) lies at high enough values of @xmath20 that it never interferes with the other limits .
the dependence of these results on the two mass loss parameters ( wind and collapse ) are determined by their association with orbital expansion . for strong helium
star wind mass loss ( @xmath34 ) , the progenitor orbits expand so much that donors less massive than the bh can never fill their roche lobes on the main sequence .
both low and intermediate
mass donors are formed only if less than 50% of the initial helium - star mass is lost in the wind .
mass loss at bh formation is limited to bh progenitors less massive than about twice the bh mass so that post collapse systems with low
mass donors remain bound .
note that amounts of mass lost in helium - star winds and in bh formation are actually anti - correlated .
if one is close to the maximum allowed then the other must be minimal ( see fig . 1 ) . the dependence on @xmath24 of the orbital expansion during helium
star wind mass loss and bh formation is such that the ratio of circularized post collapse over post
ce orbital separations becomes independent of @xmath24 .
this means that , for a specific bh mass , the position of the limits on the @xmath22 plane depend only on the initial helium star mass and the ce efficiency . indeed , in fig . 1
, the change of donor types occurs along straight lines in the @xmath26@xmath24 plane , or else along lines of constant @xmath27 .
this simplifying property allows us to combine the panels in fig . 1 into one plot ( fig . 2 ) .
it is evident that formation of 7 bh x ray binaries with both low and intermediate mass donors ( as required by the observed sample ) constrains the common
envelope efficiency to relatively high values and the initial helium - star progenitors at most twice as massive as the bh ( corresponding to initial primaries in the range 2545 ) .
additional constraints can be obtained by examining the relative numbers of systems with low and intermediate
mass donors formed for the parameters in the black - shaded areas in figs . 1 and 2 .
the lifetimes for the two different types are determined by the process that drives mass transfer .
the magnetic braking time scale , for low
mass donors is comparable to the nuclear evolution time scale of intermediate mass stars ( kalogera 1999 ) .
the number ratio then becomes equal to the ratio of birth rates .
the latter can be calculated using the derived limits on @xmath20 and @xmath21 and assumed distributions of mass ratios and orbital separations of primordial binaries .
the results indicate that even when low mass companions in primordial binaries are strongly favored , bh binaries with intermediate
mass donors are much more easily formed because of the larger range of orbital separations allowed to their progenitors ( see figure 3 ) .
models predict a small fraction of intermediate mass donors ( as seen in the current observed sample ) only for rather high @xmath30 values ( @xmath35 ) or for moderate ( but still higher than unity ) @xmath30 values ( @xmath36 ) and bh progenitors either slightly more massive or twice as massive as the bh .
we note that these results are quite robust and do not depend on the assumed bh mass nor the properties of primordial binaries ( see kalogera 1999 ) .
more careful consideration of the above analysis shows that the constraints derived for helium star wind mass loss depend on the timing of the ce phase with respect to the evolution of the massive bh progenitor .
current single
star evolutionary models ( schaller et al.1992 ; hurley , pols , & tout 2000 ) for massive stars ( @xmath37 ) of solar metallicity with wind mass loss imply that a roche lobe overflow ( and hence a ce phase ) can occur only either _ before _ or _
after _ the star s core helium burning phase . during this
phase wind mass loss accelerates , leading to significant orbital expansion ( jeans mode of mass loss ) , while the stellar radius remains almost constant .
also , for stars more massive than about 25 ( about 35 for z=0.002 ) , the evolutionary models show a radial contraction until the star reaches core collapse ( see also kalogera 1999 ) . based on the above considerations ,
bh progenitors can fill their roche lobe only _ before _ core helium ignition ( case b mass transfer ) .
the exposed helium cores evolve through the complete core helium burning phase before they reach core collapse . taking into account this implicit constraint , we are able to derive the associated constraints on helium star wind mass loss
. current models of helium star evolution through core helium burning ( woosley 1995 ) predict amounts of mass lost in the wind significantly larger than the maxima allowed for bh x - ray transient formation ( @xmath38% ) .
in fact , the final helium - star masses in these models are @xmath39 , far too small to explain the bh mass measurements .
therefore , if the ce phase is initiated early in the core helium burning phase of the primary , then helium - star winds must be much weaker than thought until now .
it is worth noting that more recent empirical estimates of wind mass loss rates show a downward trend ( hamann & koesterke 1998 ) . however , if the models for massive star evolution with winds do not represent reality , and radial expansion ( and hence roche lobe overflow and ce evolution ) is possible after the end of core helium burning , then it is possible that the progenitors of bh binaries evolve through case c mass transfer ( see wellstein & langer 1999 ) . in this case , the strength of helium
star winds becomes irrelevant to the process of bh x ray binary formation .
the reason is that the helium core is exposed only through its core carbon and later burning phases , and the total duration of these phases is so short that the wind mass loss is insignificant and the helium
star mass remains essentially constant ( woosley 1995 ) .
given the uncertainties in models of massive star evolution and the calculation of stellar radii , such a modification can not be excluded at present . in any case , it becomes clear that the existence of bh x ray transients requires that either the hydrogen rich massive star models or the strength of helium star winds be modified .
we have shown that the observed properties of the current sample of bh x ray transients provide us with clues to their evolutionary history and bh formation .
specifically , we find that ( i ) common envelope evolution is necessary to account for the present tight orbits , ( ii ) orbital contraction during the common envelope phase must be moderate and therefore ce efficiencies must be relatively high ( depending on the exact radii of massive stars and their density profiles , significant contributions from energy sources other than the orbit may be required ) , ( iii ) helium stars that form black holes are at most twice more massive than the black holes at the time of collapse .
all these constraints do not depend on the details of the radial evolution of massive stars .
we further find that current evolutionary models for massive stars losing mass in winds appear to be in conflict with models of wind mass loss from helium stars . assuming that models massive star evolution are more accurate than estimates of helium star winds
, we find that wind mass loss from helium stars must be limited so that these stars lose at most half of their mass at the beginning of core helium burning .
current stellar evolution models face one additional challenge posed by the bh mass of v404 cyg , if the upper end of the measured range is confirmed ( @xmath40 ; see , e.g. , bailyn et al .
1998 ) . even if we ignore
_ any _ mass loss from helium stars , predicted masses for helium cores of massive stars can not account for such a high bh mass ( see wellstein & langer 1999 ; hurley et al .
2000 ) . an alternative model for the formation of bh x ray binaries that invokes the evolution of hierarchical triples
has actually been suggested by eggleton & verbunt ( 1986 ) .
the basic idea is that the progenitor consists of a high mass inner binary with an outer low
mass companion .
the inner binary evolves to a high mass x
ray binary , where roche lobe overflow from the companion to the neutron star leads to a complete merger and possibly the formation of a `` thorne zytkow '' object . during the merger
the neutron star is expected to collapse into a bh , which can continue to grow through accretion .
if the massive envelope around the bh expands as giants do then it is possible that this envelope will eventually engulf the outer companion .
the resulting spiral in and envelope ejection would then lead to the formation of a tight binary with a bh and a roche lobe filling low
mass companion .
such an alternative formation path can very easily overcome difficulties with ( initial and final ) helium star masses and requirements for large ( possibly unphysical ) common envelope efficiencies .
furthermore , a significant fraction of stars appear to be members of multiple systems .
these considerations could motivate a detailed analysis of this triple star formation path for bh binaries .
such an analysis would eventually include the study of a number of very interesting problems : orbital stability of the triple system throughout the long term phases of its evolution , the effects on the outer orbit of wind mass loss and supernova explosions occurring in the inner binary , the unsettled question of the stability of `` thorne zytkow '' objects ( see , e.g. , cannon et al.1992 , but also fryer , benz , & herant 1996 ) and their subsequent evolution , the frequency and orbital properties of triple systems .
* acknowledgments .
* i would like to thank n. langer , r. taam , r. webbink , and s.wellstein for useful discussions .
support by the smithsonian astrophysical observatory through a harvard - smithsonian center for astrophysics postdoctoral fellowship and a clay fellowship is also acknowledged . | studies of the observed characteristics of black hole ( bh ) x ray binaries can be provide us with valuable information about the process of bh formation . in this paper
i address some of the aspects of our current understanding of bh formation in binaries and point out some of the existing problems of current theoretical models .
in particular , the measured orbital periods and donor star properties indicate that a common envelope phase appears to be a necessary ingredient of the evolutionary history of observed bh x ray transients , and that it must be associated only with a modest orbital contraction . the timing of this common
envelope phase is crucial in determining the final bh masses and current evolutionary models of mass losing massive stars place strong constraints on the possible masses for immediate bh progenitors and wind mass loss from helium stars .
last , it is interesting that , even in the absence of any source of mass loss , the highest helium star masses predicted by current evolutionary models are still not high enough to account for the measured bh mass in v404 cyg ( @xmath0 ) .
an alternative for the formation of relatively massive bh may be provided by the evolutionary sequence proposed by eggleton & verbunt ( 1986 ) , which invokes hierarchical triples as progenitors of bh x ray binaries with low mass companions .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.